Cancer in Patients With Gabapentin (GPRD)

• Sponsor: GlaxoSmithKline (Industry)
• Overall Status: Completed
• CT.gov ID: NCT01236053
• Collaborator: (none)
• Enrollment: 2,323,608
• Duration (Months): 3

Study Details

Study Description

Brief Summary

High doses of gabapentin are associated with pancreatic acinar cell tumors in rats, but there has been no post marketing pancreatic carcinogenicity signal with gabapentin as reported by spontaneous reports in AERS or in the published literature. In a published case-control screening study of the association of gabapentin with 55 cancers, the only cancer that met the screening criteria for possibly increased cancer risk with gabapentin exposure was renal (including renal pelvis) cancer. This association was judged to be likely due to or substantially accentuated by confounding by cigarette smoking, hypertension, and lifestyle (Cancer Causes Control 2009;20:1821-1835).

The relationship between gabapentin exposure and pancreatic cancer and renal cancer is studied in NCT01138124, and supplemental analyses for these cancers are performed in the current study. The FDA recommended GSK also study the relationship between gabapentin and all-cancer sites, as well as cancer at the following specific sites: 1) stomach, 2) anus, anal canal, and anorectum, 3) lung and bronchus, 4) bones and joints, 5) breast, 6) penis, 7) urinary bladder, and 8) other nervous system.

The primary objective of this study is to determine whether exposure to gabapentin is associated with an increased risk of developing all-cancer, and these specific cancers in the United Kingdom (UK) General Practice Research Database (GPRD). Each member of the UK population is registered with a General Practice, which centralizes the medical information not only from the general practitioners themselves but also from specialist referrals and hospital attendances. Over 487 General Practices contribute data to the GPRD.

The study cohort from which cases and controls are drawn is all subjects in the GPRD 1993-2008. Gabapentin was approved in the UK in May 1993. Entry into the study cohort begins Jan 1, 1993 for all those who are registered in GPRD before that time, and at the time of registration if later than Jan 1, 1993. Subjects are excluded from the GPRD cohort if they have a cancer diagnosis or a history of cancer prior to the cohort entry date. Patients with a first diagnosis of the respective cancer 1995-2008 are risk set matched with up to 10 controls within the same General Practice for age at cohort entry (within two years), sex, and year of entry into the study cohort (within one year). For cases, the index date is the date of first diagnosis of the respective cancer. The index date for controls is set as the date at which the follow-up time from cohort entry is the same as the case. The index date is chosen so as to give the control equal follow-up time to that of the case for ascertainment of use of gabapentin. Cases and controls will be required to have at least 2 years of follow-up in the study cohort before their index date. Cases must have no history of any other cancer diagnosis prior to the index date. Controls are required to be free of cancer diagnosis in the database up to the control's index date.

Data on gabapentin prescriptions are obtained for cases and controls from study cohort entry to the index date. Gabapentin exposure will be assessed as ever/never, number of prescriptions, cumulative dose, and cumulative duration, with a 2 year lag period incorporated to control for protopathic bias (gabapentin prescription for initial pain symptoms of undiagnosed cancer) and latency (time between cancer onset and specific GPRD cancer diagnosis).

Crude and adjusted odds ratios and 95% confidence intervals (CI) will be produced from conditional logistic regression models, with additional analyses evaluating for dose-response. Covariates include indications for gabapentin use and risk factors for each cancer.

Condition or Disease	Intervention/Treatment	Phase
Pain, Neuropathic; Epilepsy; Renal Pelvis Cancer; Pancreatic Cancer; Breast Cancer; Nervous System Cancer; Chronic Pancreatitis; Stomach Cancer; Renal Cell Carcinoma; Diabetes; Bladder Cancer; Bone and Joint Cancer; Penis Cancer; Anal Cancer; Cancer; Renal Cancer	Drug: Gabapentin prescriptions

Detailed Description

Actual number of patients may be less, as it is possible for a patient to be represented in more than one of the 22 arms (See "Participant Flow: Overall Study" Table) because of the risk set sampling.

Study Design

• Study Type: Observational
• Actual Enrollment: 2323608 participants
• Observational Model: Case-Control
• Time Perspective: Retrospective
• Official Title: Risk of Cancer in Patients Exposed to Gabapentin in the GPRD
• Study Start Date: Jun 1, 2010
• Actual Primary Completion Date: Sep 1, 2010
• Actual Study Completion Date: Sep 1, 2010

Arms and Interventions

Arm

Intervention/Treatment

UK GPRD 1993-2008 The study cohort from which cases and controls are drawn is all subjects in the United Kingdom (UK) General Practice Research Database (GPRD) 1993-2008. Each member of the UK population is registered with a General Practice, which centralizes the medical information not only from the general practitioners themselves but also from specialist referrals and hospital attendances. Over 487 General Practices contribute data to the GPRD. Entry into the study cohort begins Jan 1, 1993 for all those who are registered in GPRD before that time, and at the time of registration if later than Jan 1, 1993. Subjects are excluded from the GPRD cohort if they have a cancer diagnosis or a history of cancer prior to the cohort entry date.

Drug: Gabapentin prescriptions; The exposure of interest is gabapentin use as defined by prescriptions recorded by the GPRD general practitioner (British National Formulry codes). Data on prescriptions for gabapentin will be extracted for each case and control from entry into the study cohort up to two years prior to the index date (the exposure window). A two year lagged exposure is incorporated to account for latency between cancer onset and GPRD diagnosis, and for protopathic bias (gabapentin prescription for initial pain symptoms of undiagnosed cancer). Gabapentin exposure will be parameterized as follows: (1) Ever versus never exposed; (2) Number of prescriptions; (3) Duration of exposure; and (4) Cumulative dose.

Outcome Measures

Primary Outcome Measures

1. Number of All-Cancer Cases and Matched Controls With the Indicated Exposure to Gabapentin [The case index date was the date of incident cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.]

   Incidence of all cancers. Gabapentin Exposure Description: Without 2 year lag = Gabapentin prescription from cohort entry to index date. With 2 year lag = Gabapentin prescription from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer).

2. Number of All-Cancer Cases and Matched Controls With the Indicated Number of Gabapentin Prescriptions [The case index date was the date of incident cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.]

   Incidence of all cancers. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date. Tertile's with 2 year lag: Tertile 1 (1-2 prescriptions), Tertile 2 (3-7 prescriptions), Tertile 3 (8-298 prescriptions). Tertiles without 2 year lag: Tertile 1 (1-2 prescriptions), Tertile 2 (3-7 prescriptions), and Tertile 3 (8-388 prescriptions).

3. Number of All-Cancer Cases and Matched Controls With the Indicated Duration of Exposure to Gabapentin [The case index date was the date of incident cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.]

   Incidence of all cancers. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date. Tertile's with 2 year lag: Tertile 1 (0.01 - 1.38 months), Tertile 2 (1.39 - 5.56 months), and Tertile 3 (5.57 - 105.82 months). Tertile's without 2 year lag: Tertile 1 (0.01 - 1.38 months), Tertile 2 (1.39 - 5.72 months), and Tertile 3 (5.73 - 123.70 months).

4. Number of All-Cancer Cases and Matched Controls With the Indicated Long Duration of Exposure to Gabapentin [The case index date was the date of incident cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.]

   Incidence of all cancers. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date.

5. Number of All-Cancer Cases and Matched Controls With the Indicated Cumulative Dose of Gabapentin [The case index date was the date of incident cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.]

   Incidence of all cancers. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date. Tertile's with 2 year lag: Tertile 1 (0.1 - 30.0 grams), Tertile 2 (30.1 - 189.0 grams), and Tertile 3 (189.1 - 9600.0 grams). Tertile's without 2 year lag: Tertile 1 (0.1 - 30.0 grams), Tertile 2 (30.1 - 189.0 grams), and Tertile 3 (189.1 - 11610.0 grams).

6. Number of Stomach Cancer Cases and Matched Controls With the Indicated Exposure to Gabapentin [The case index date was the date of incident stomach cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case]

   Incident stomach cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin prescription from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin prescription from cohort entry to index date.

7. Number of Stomach Cancer Cases and Matched Controls With the Indicated Number of Gabapentin Prescriptions [The case index date was the date of incident stomach cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case]

   Incident stomach cancer. Gabapentin (Gaba.) Exposure Description: With 2 year (yr) lag=Gaba. exposure from cohort entry to 2 yr prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 yr lag=Gaba. exposure from cohort entry to index date. Tertiles (T) with 2 yr lag: T 1 (1-2 prescriptions [prescrip]), T 2 (3-7 prescrip), T 3 (8-298 prescrip). Tertiles without 2 yr lag: T 1 (1-2 prescrip), T 2 (3-7 prescrip), and T 3 (8-388 prescrip). Inestimable OR and 95% CI when no gaba.-exposed cancer cases or controls at the exposure level.

8. Number of Stomach Cancer Cases and Matched Controls With the Indicated Duration of Exposure to Gabapentin [The case index date was the date of incident stomach cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case]

   Incident stomach cancer. Gabapentin (Gaba.) Exposure Description: With 2 year (yr) lag=Gaba. exposure from cohort entry to 2 yr prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 yr lag=Gaba. exposure from cohort entry to index date. Tertiles (T) with 2 yr lag: T 1 (0.01-1.38 months [mo.]), T 2 (1.3-5.56 mo.), and T 3 (5.57-105.82 mo.). Tertiles without 2 yr lag: T 1 (0.01-1.38 m.), T 2 (1.39-5.72 mo.), and T 3 (5.73-123.70 mo.). Inestimable OR and 95% CI when no gaba.-exposed cancer cases or controls at the exposure level.

9. Number of Stomach Cancer Cases and Matched Controls With the Indicated Long Duration of Exposure to Gabapentin [The case index date was the date of incident stomach cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case]

   Incident stomach cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date.

10. Number of Stomach Cancer Cases and Matched Controls With the Indicated Cumulative Dose of Gabapentin [The case index date was the date of incident stomach cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case]

    Incident stomach cancer. Gabapentin (Gaba.) Exposure Description: With 2 year (yr) lag=Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 yr lag=Gaba. exposure from cohort entry to index date. Tertiles (T) with 2 yr lag: T 1 (0.1-30.0 grams [g]), T 2 (30.1-189.0 g), and T 3 (189.1-9600.0 g). Tertiles without 2 yr lag: T 1 (0.1-30.0 g), T 2 (30.1-189.0 g), and T 3 (189.1-11610.0 g). Inestimable OR and 95% CI when no gaba.-exposed cancer cases or controls at the exposure level.

11. Number of Anal Cancer Cases and Matched Controls With the Indicated Exposure to Gabapentin [The case index date was the date of incident anal cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.]

    Incident Anal cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin prescription from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin prescription from cohort entry to index date.

12. Number of Anal Cancer Cases and Matched Controls With the Indicated Number of Gabapentin Prescriptions [The case index date was the date of incident anal cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.]

    Incident anal cancer. Gabapentin (Gaba.) Exposure Description: With 2 year (yr) lag=Gaba. exposure from cohort entry to 2 yr prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 yr lag=Gaba. exposure from cohort entry to index date. Tertiles (T) with 2 yr lag: T 1 (1-2 prescriptions [prescrip.]), T 2 (3-7 prescrip.), T 3 (8-298 prescrip.). Tertiles without 2 yr lag: T 1 (1-2 prescrip.), T 2 (3-7 prescrip.), and T 3 (8-388 prescrip.). Inestimable OR and 95% CI when no gaba.-exposed cancer cases or controls at the exposure level.

13. Number of Anal Cancer Cases and Matched Controls With the Indicated Duration of Exposure to Gabapentin [The case index date was the date of incident anal cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.]

    Incident anal cancer. Gabapentin (Gaba.) Exposure Description: With 2 year (yr) lag=Gaba. exposure from cohort entry to 2 yr prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 yr lag=Gaba. exposure from cohort entry to index date. Tertiles (T) with 2 yr lag: T 1 (0.01-1.38 months [mo.]), T 2 (1.39-5.56 mo.), and T 3 (5.57 -105.82 mo.). Tertiles without 2 yr lag: T 1 (0.01-1.38 mo.), T 2 (1.39-5.72 mo.), and T 3 (5.73-123.70 mo.). Inestimable OR and 95% CI when no gaba.-exposed cancer cases or controls at the exposure level.

14. Number of Anal Cancer Cases and Matched Controls With the Indicated Long Duration of Exposure to Gabapentin [The case index date was the date of incident anal cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.]

    Incident anal cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date.

15. Number of Anal Cancer Cases and Matched Controls With the Indicated Cumulative Dose of Gabapentin [The case index date was the date of incident anal cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.]

    Incident anal cancer. Gabapentin (Gaba.) Exposure Description: With 2 yr lag=Gaba. exposure from cohort entry to 2 yr prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 yr lag=Gaba. exposure from cohort entry to index date. Tertiles (T) with 2 yr lag: T 1 (0.1-30.0 grams [g]), T 2 (30.1-189.0 g), and T 3 (189.1-9600.0 g). Tertiles without 2 yr lag: T 1 (0.1-30.0 g), T 2 (30.1-189.0 g), and T 3 (189.1-11610.0 g). Inestimable OR and 95% CI when no gaba.-exposed cancer cases or controls at the exposure level.

16. Number of Lung Cancer Cases and Matched Controls With the Indicated Exposure to Gabapentin [The case index date was the date of incident lung cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.]

    Incident lung cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin prescription from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin prescription from cohort entry to index date.

17. Number of Lung Cancer Cases and Matched Controls With the Indicated Number of Gabapentin Prescriptions [The case index date was the date of incident lung cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.]

    Incident lung cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date. Tertile's with 2 year lag: Tertile 1 (1-2 prescriptions), Tertile 2 (3-7 prescriptions), Tertile 3 (8-298 prescriptions). Tertiles without 2 year lag: Tertile 1 (1-2 prescriptions), Tertile 2 (3-7 prescriptions), and Tertile 3 (8-388 prescriptions).

18. Number of Lung Cancer Cases and Matched Controls With the Indicated Duration of Exposure to Gabapentin [The case index date was the date of incident lung cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.]

    Incident lung cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date. Tertile's with 2 year lag: Tertile 1 (0.01 - 1.38 months), Tertile 2 (1.39 - 5.56 months), and Tertile 3 (5.57 - 105.82 months). Tertile's without 2 year lag: Tertile 1 (0.01 - 1.38 months), Tertile 2 (1.39 - 5.72 months), and Tertile 3 (5.73 - 123.70 months).

19. Number of Lung Cancer Cases and Matched Controls With the Indicated Long Duration of Exposure to Gabapentin [The case index date was the date of incident lung cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.]

    Incident lung cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date.

20. Number of Lung Cancer Cases and Matched Controls With the Indicated Cumulative Dose of Gabapentin [The case index date was the date of incident lung cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.]

    Incident lung cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date. Tertile's with 2 year lag: Tertile 1 (0.1 - 30.0 grams), Tertile 2 (30.1 - 189.0 grams), and Tertile 3 (189.1 - 9600.0 grams). Tertile's without 2 year lag: Tertile 1 (0.1 - 30.0 grams), Tertile 2 (30.1 - 189.0 grams), and Tertile 3 (189.1 - 11610.0 grams).

21. Number of Bone/Joint Cancer Cases and Matched Controls With the Indicated Exposure to Gabapentin [The case index date was the date of incident bone/joint cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as for the case.]

    Incident bone/joint cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin prescription from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin prescription from cohort entry to index date. Inestimable OR and 95% CI when no gabapentin-exposed cancer cases or no gabapentin-exposed controls at the exposure level.

22. Number of Bone/Joint Cancer Cases and Matched Controls With the Indicated Number of Gabapentin Prescriptions [The case index date was the date of incident bone/joint cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as for the case.]

    Incident bone/joint cancer. Gabapentin (Gaba.) Exposure Description: With 2 year (yr) lag=Gaba. exposure from cohort entry to 2 yr prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 yr lag=Gaba. exposure from cohort entry to index date. Tertiles (T) with 2 yr lag: T 1 (1-2 prescriptions [prescrip.]), T 2 (3-7 prescrip.), T 3 (8-298 prescrip.). Tertiles without 2 yr lag: T 1 (1-2 prescrip.), T 2 (3-7 prescrip.), and T 3 (8-388 prescrip.). Inestimable OR and 95% CI when no gaba.-exposed cases or controls at the exposure level.

23. Number of Bone/Joint Cancer Cases and Matched Controls With the Indicated Duration of Exposure to Gabapentin [The case index date was the date of incident bone/joint cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as for the case.]

    Incident bone/joint cancer. Gabapentin (Gaba.) Exposure Description: With 2 year (yr) lag=Gaba. exposure from cohort entry to 2 yr prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 yr lag=Gaba. exposure from cohort entry to index date. Tertiles (T) with 2 yr lag: T 1 (0.01-1.38 months [mo.]), T 2 (1.39-5.56 mo.), and T 3 (5.57-105.82 mo.). Tertiles without 2 yr lag: T 1 (0.01-1.38 mo.), T 2 (1.39-5.72 mo.), and T 3 (5.73-123.70 mo.). Inestimable OR and 95% CI when no gaba.-exposed cases or controls at the exposure level.

24. Number of Bone/Joint Cancer Cases and Matched Controls With the Indicated Long Duration of Exposure to Gabapentin [The case index date was the date of incident bone/joint cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as for the case.]

    Incident bone/joint cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date.

25. Number of Bone/Joint Cancer Cases and Matched Controls With the Indicated Cumulative Dose of Gabapentin [The case index date was the date of incident bone/joint cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as for the case.]

    Incident bone/joint cancer. Gabapentin (Gaba.) Exposure Description: With 2 year (yr) lag=Gaba. exposure from cohort entry to 2 yr prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 yr lag=Gaba. exposure from cohort entry to index date. Tertiles (T) with 2 yr lag: T 1 (0.1-30.0 grams [g]), T 2 (30.1-189.0 g), and T 3 (189.1-9600.0 g). Tertiles without 2 yr lag: T 1 (0.1-30.0 g), T 2 (30.1-189.0 g), and T 3 (189.1-11610.0 g). Inestimable OR and 95% CI when no gaba.-exposed cancer cases or controls at the exposure level.

26. Number of Breast Cancer Cases and Matched Controls With the Indicated Exposure to Gabapentin [The case index date was the date of incident breast cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.]

    Incident breast cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin prescription from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin prescription from cohort entry to index date.

27. Number of Breast Cancer Cases and Matched Controls With the Indicated Number of Gabapentin Prescriptions [The case index date was the date of incident breast cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.]

    Incident breast cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date. Tertile's with 2 year lag: Tertile 1 (1-2 prescriptions), Tertile 2 (3-7 prescriptions), Tertile 3 (8-298 prescriptions). Tertiles without 2 year lag: Tertile 1 (1-2 prescriptions), Tertile 2 (3-7 prescriptions), and Tertile 3 (8-388 prescriptions).

28. Number of Breast Cancer Cases and Matched Controls With the Indicated Duration of Exposure to Gabapentin [The case index date was the date of incident breast cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.]

    Incident breast cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date. Tertile's with 2 year lag: Tertile 1 (0.01 - 1.38 months), Tertile 2 (1.39 - 5.56 months), and Tertile 3 (5.57 - 105.82 months). Tertile's without 2 year lag: Tertile 1 (0.01 - 1.38 months), Tertile 2 (1.39 - 5.72 months), and Tertile 3 (5.73 - 123.70 months).

29. Number of Breast Cancer Cases and Matched Controls With the Indicated Long Duration of Exposure to Gabapentin [The case index date was the date of incident breast cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.]

    Incident breast cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date.

30. Number of Breast Cancer Cases and Matched Controls With the Indicated Cumulative Dose of Gabapentin [The case index date was the date of incident breast cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.]

    Incident breast cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date. Tertile's with 2 year lag: Tertile 1 (0.1 - 30.0 grams), Tertile 2 (30.1 - 189.0 grams), and Tertile 3 (189.1 - 9600.0 grams). Tertile's without 2 year lag: Tertile 1 (0.1 - 30.0 grams), Tertile 2 (30.1 - 189.0 grams), and Tertile 3 (189.1 - 11610.0 grams).

31. Number of Penile Cancer Cases and Matched Controls With the Indicated Exposure to Gabapentin [The case index date was the date of incident penile cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.]

    Incident penile cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin prescription from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin prescription from cohort entry to index date.

32. Number of Penile Cancer Cases and Matched Controls With the Indicated Number of Gabapentin Prescriptions [The case index date was the date of incident penile cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.]

    Incident penile cancer. Gabapentin (Gaba.) Exposure Description: With 2 year (yr) lag=Gaba. exposure from cohort entry to 2 yr prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 yr lag=Gaba. exposure from cohort entry to index date. Tertiles (T) with 2 yr lag: T 1 (1-2 prescriptions [prescrip.]), T 2 (3-7 prescrip.), T 3 (8-298 prescrip.). Tertiles without 2 yr lag: T 1 (1-2 prescrip.), T 2 (3-7 prescrip.), and T 3 (8-388 prescrip.). Inestimable OR/95% CI when no gaba.-exposed cancer cases or controls at the exposure level.

33. Number of Penile Cancer Cases and Matched Controls With the Indicated Duration of Exposure to Gabapentin [The case index date was the date of incident penile cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.]

    Incident penile cancer. Gabapentin (Gaba.) Exposure Description: With 2 year (yr) lag=Gaba. exposure from cohort entry to 2 yr prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 yr lag=Gaba. exposure from cohort entry to index date. Tertiles (T) with 2 yr lag: T 1 (0.01-1.38 months [mo.]), T 2 (1.39-5.56 mo.), and T 3 (5.57-105.82 mo.). Tertiles without 2 yr lag: T 1 (0.01-1.38 mo.), T 2 (1.39-5.72 mo.), and T 3 (5.73-123.70 mo.). Inestimable OR and 95% CI when no gaba.-exposed cancer cases or controls at the exposure level.

34. Number of Penile Cancer Cases and Matched Controls With the Indicated Long Duration of Exposure to Gabapentin [The case index date was the date of incident penile cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.]

    Incident penile cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date.

35. Number of Penile Cancer Cases and Matched Controls With the Indicated Cumulative Dose of Gabapentin [The case index date was the date of incident penile cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.]

    Incident penile cancer. Gabapentin (Gaba.) Exposure Description: With 2 year (yr) lag=Gaba. exposure from cohort entry to 2 yr prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 yr lag=Gaba. exposure from cohort entry to index date. Tertiles (T) with 2 yr lag: T 1 (0.1-30.0 grams [g]), T 2 (30.1-189.0 g), and T 3 (189.1-9600.0 g). Tertiles without 2 yr lag: T 1 (0.1-30.0 g), T 2 (30.1-189.0 g), and T 3 (189.1-11610.0 g). Inestimable OR and 95% CI when no gaba.-exposed cancer cases or controls at the exposure level.

36. Number of Bladder Cancer Cases and Matched Controls With the Indicated Exposure to Gabapentin [The case index date was the date of incident bladder cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case]

    Incident bladder cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin prescription from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin prescription from cohort entry to index date.

37. Number of Bladder Cancer Cases and Matched Controls With the Indicated Number of Gabapentin Prescriptions [The case index date was the date of incident bladder cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case]

    Incident bladder cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date. Tertile's with 2 year lag: Tertile 1 (1-2 prescriptions), Tertile 2 (3-7 prescriptions), Tertile 3 (8-298 prescriptions). Tertiles without 2 year lag: Tertile 1 (1-2 prescriptions), Tertile 2 (3-7 prescriptions), and Tertile 3 (8-388 prescriptions).

38. Number of Bladder Cancer Cases and Matched Controls With the Indicated Duration of Exposure to Gabapentin [The case index date was the date of incident bladder cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case]

    Incident bladder cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date. Tertile's with 2 year lag: Tertile 1 (0.01 - 1.38 months), Tertile 2 (1.39 - 5.56 months), and Tertile 3 (5.57 - 105.82 months). Tertile's without 2 year lag: Tertile 1 (0.01 - 1.38 months), Tertile 2 (1.39 - 5.72 months), and Tertile 3 (5.73 - 123.70 months).

39. Number of Bladder Cancer Cases and Matched Controls With the Indicated Long Duration of Exposure to Gabapentin [The case index date was the date of incident bladder cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case]

    Incident bladder cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date.

40. Number of Bladder Cancer Cases and Matched Controls With the Indicated Cumulative Dose of Gabapentin [The case index date was the date of incident bladder cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case]

    Incident bladder cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date. Tertile's with 2 year lag: Tertile 1 (0.1 - 30.0 grams), Tertile 2 (30.1 - 189.0 grams), and Tertile 3 (189.1 - 9600.0 grams). Tertile's without 2 year lag: Tertile 1 (0.1 - 30.0 grams), Tertile 2 (30.1 - 189.0 grams), and Tertile 3 (189.1 - 11610.0 grams).

41. Number of Other Nervous System Cancer Cases and Matched Controls With the Indicated Exposure to Gabapentin [The case index date was the date of incident other nervous system cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his cohort entry was the same as case.]

    Incident other nervous system cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin prescription from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin prescription from cohort entry to index date. Inestimable OR and 95% CI when no gabapentin-exposed cancer cases or no gabapentin-exposed controls at the exposure level.

42. Number of Other Nervous System (ONS) Cancer Cases and Matched Controls With the Indicated Number of Gabapentin Prescriptions [The case index date was the date of incident other nervous system cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his cohort entry was same as for case.]

    Incident ONS cancer. Gabapentin (Gaba.) Exposure Description: With 2 year (yr) lag=Gaba. exposure from cohort entry to 2 yr prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 yr lag=Gaba. exposure from cohort entry to index date. Tertiles (T) with 2 yr lag: T 1 (1-2 prescriptions [prescrip.]), T 2 (3-7 prescrip.), T 3 (8-298 prescrip.). Tertiles without 2 yr lag: T 1 (1-2 prescrip.), T 2 (3-7 prescrip.), and T 3 (8-388 prescrip.). Inestimable OR and 95% CI when no gaba.-exposed cancer cases or controls at the exposure level.

43. Number of Other Nervous System (ONS) Cancer Cases and Matched Controls With the Indicated Duration of Exposure to Gabapentin [The case index date was the date of incident other nervous system cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was same as for case]

    Incident ONS cancer. Gabapentin (Gaba.) Exposure Description: With 2 yr lag=Gaba. exposure from cohort entry to 2 yr prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 yr lag=Gaba. exposure from cohort entry to index date. Tertiles (T) with 2 yr lag: T 1 (0.01-1.38 months [mo.]), T 2 (1.39-5.56 mo.), and T 3 (5.57-105.82 mo.). Tertiles without 2 yr lag: T 1 (0.01-1.38 mo.), T 2 (1.39-5.72 mo.), and T 3 (5.73-123.70 mo.). Inestimable OR and 95% CI when no gaba.-exposed cancer cases or controls at the exposure level.

44. Number of Other Nervous System Cancer Cases and Matched Controls With the Indicated Long Duration of Exposure to Gabapentin [The case index date was the date of incident other nervous system cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was same as for case]

    Incident other nervous system cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date.

45. Number of Other Nervous System (ONS) Cancer Cases and Matched Controls With the Indicated Cumulative Dose of Gabapentin [The case index date was the date of incident other nervous system cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was same as for case]

    Incident ONS cancer. Gabapentin (Gaba.) Exposure Description: With 2 year (yr) lag=Gaba. exposure from cohort entry to 2 yr prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 yr lag=Gaba. exposure from cohort entry to index date. Tertiles (T) with 2 yr lag: T 1 (0.1-30.0 grams [g]), T 2 (30.1-189.0 g), and T 3 (189.1-9600.0 g). Tertiles without 2 yr lag: T 1 (0.1-30.0 g), T 2 (30.1-189.0 g), and T 3 (189.1-11610.0 g). Inestimable OR and 95% CI when no gaba.-exposed cancer cases or controls at the exposure level.

46. Number of Pancreatic Cancer Cases and Matched Controls With the Indicated Exposure to Gabapentin [The case index date was the date of incident pancreatic cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as for the case.]

    Incident pancreatic cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin prescription from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin prescription from cohort entry to index date.

47. Number of Pancreatic Cancer Cases and Matched Controls With the Indicated Number of Gabapentin Prescriptions [The case index date was the date of incident pancreatic cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as for the case.]

    Incident pancreatic cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date. Tertile's with 2 year lag: Tertile 1 (1-2 prescriptions), Tertile 2 (3-7 prescriptions), Tertile 3 (8-298 prescriptions). Tertiles without 2 year lag: Tertile 1 (1-2 prescriptions), Tertile 2 (3-7 prescriptions), and Tertile 3 (8-388 prescriptions).

48. Number of Pancreatic Cancer Cases and Matched Controls With the Indicated Duration of Exposure to Gabapentin [The case index date was the date of incident pancreatic cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as for the case.]

    Incident pancreatic cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date. Tertile's with 2 year lag: Tertile 1 (0.01 - 1.38 months), Tertile 2 (1.39 - 5.56 months), and Tertile 3 (5.57 - 105.82 months). Tertile's without 2 year lag: Tertile 1 (0.01 - 1.38 months), Tertile 2 (1.39 - 5.72 months), and Tertile 3 (5.73 - 123.70 months).

49. Number of Pancreatic Cancer Cases and Matched Controls With the Indicated Long Duration of Exposure to Gabapentin [The case index date was the date of incident pancreatic cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as for the case]

    Incident pancreatic cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date.

50. Number of Pancreatic Cancer Cases and Matched Controls With the Indicated Cumulative Dose of Gabapentin [The case index date was the date of incident pancreatic cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as for the case.]

    Incident pancreatic cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date. Tertile's with 2 year lag: Tertile 1 (0.1 - 30.0 grams), Tertile 2 (30.1 - 189.0 grams), and Tertile 3 (189.1 - 9600.0 grams). Tertile's without 2 year lag: Tertile 1 (0.1 - 30.0 grams), Tertile 2 (30.1 - 189.0 grams), and Tertile 3 (189.1 - 11610.0 grams).

51. Number of Renal Cancer Cases and Matched Controls With the Indicated Exposure to Gabapentin [The case index date was the date of incident pancreatic cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as for the case.]

    Incident renal cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin prescription from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin prescription from cohort entry to index date.

52. Number of Renal Cancer Cases and Matched Controls With the Indicated Number of Gabapentin Prescriptions [The case index date was the date of incident renal cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.]

    Incident renal cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date. Tertile's with 2 year lag: Tertile 1 (1-2 prescriptions), Tertile 2 (3-7 prescriptions), Tertile 3 (8-298 prescriptions). Tertiles without 2 year lag: Tertile 1 (1-2 prescriptions), Tertile 2 (3-7 prescriptions), and Tertile 3 (8-388 prescriptions).

53. Number of Renal Cancer Cases and Matched Controls With the Indicated Duration of Exposure to Gabapentin [The case index date was the date of incident renal cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.]

    Incident renal cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date. Tertile's with 2 year lag: Tertile 1 (0.01 - 1.38 months), Tertile 2 (1.39 - 5.56 months), and Tertile 3 (5.57 - 105.82 months). Tertile's without 2 year lag: Tertile 1 (0.01 - 1.38 months), Tertile 2 (1.39 - 5.72 months), and Tertile 3 (5.73 - 123.70 months).

54. Number of Renal Cancer Cases and Matched Controls With the Indicated Long Duration of Exposure to Gabapentin [The case index date was the date of incident renal cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.]

    Incident renal cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date.

55. Number of Renal Cancer Cases and Matched Controls With the Indicated Cumulative Dose of Gabapentin [The case index date was the date of incident renal cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.]

    Incident renal cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date. Tertile's with 2 year lag: Tertile 1 (0.1 - 30.0 grams), Tertile 2 (30.1 - 189.0 grams), and Tertile 3 (189.1 - 9600.0 grams). Tertile's without 2 year lag: Tertile 1 (0.1 - 30.0 grams), Tertile 2 (30.1 - 189.0 grams), and Tertile 3 (189.1 - 11610.0 grams).

Eligibility Criteria

Criteria

• Ages Eligible for Study: N/A and Older
• Sexes Eligible for Study: All
• Accepts Healthy Volunteers: No

Inclusion Criteria:

• The study cohort from which cases and controls are drawn is all subjects in the UK GPRD 1993-2008. Entry into the study cohort begins Jan 1, 1993 for all those who are registered in GPRD before that time, and at the time of registration if later than Jan 1, 1993. Follow-up ends Dec 31, 2008, or earlier if the subject leaves the GPRD for any reason including death.

Exclusion Criteria:

• Subjects are excluded from the GPRD cohort if they have a cancer diagnosis or a history of cancer prior to the cohort entry date. Cases and controls will be required to have at least 2 years of follow-up in the study cohort before their index date (For cases, the index date is the date of first diagnosis of the respective cancer. The index date for controls is set as the date at which the follow-up time from cohort entry is the same as the case.) Cases must have no history of any other cancer diagnosis prior to the index date. Controls are required to be free of cancer diagnosis in the database up to the control's index date.

Contacts and Locations

Locations

No locations specified.

Sponsors and Collaborators

• GlaxoSmithKline

Investigators

• Study Director: GSK Clinical Trials, GlaxoSmithKline

Study Documents (Full-Text)

None provided.

More Information

Publications

None provided.

• Responsible Party: GlaxoSmithKline
• ClinicalTrials.gov Identifier: NCT01236053
• Other Study ID Numbers: 114790; WEUSRTP4931
• First Posted: Nov 7, 2010
• Last Update Posted: Jul 6, 2017
• Last Verified: Jun 1, 2017

• Keywords provided by GlaxoSmithKline: renal cancer; pancreatic cancer; epidemiology; renal cell carcinoma; nervous system cancer; lung cancer; breast cancer; case-control; GPRD; Gapabentin; cancer; anal cancer; bone and joint cancer; renal pelvis cancer; penis cancer; stomach cancer; bladder cancer
• Additional relevant MeSH terms: Pancreatic Neoplasms; Urinary Bladder Neoplasms; Carcinoma, Renal Cell; Anus Neoplasms; Kidney Neoplasms; Stomach Neoplasms; Pelvic Neoplasms; Nervous System Neoplasms; Penile Neoplasms; Pancreatitis; Pancreatitis, Chronic; Neuralgia; Gabapentin

Study Results

Participant Flow

Recruitment Details	Patients were not recruited for nor enrolled in this study. This study is a retrospective observational study. Data from medical records or insurance claims databases are anonymized and used to develop a patient cohort. All diagnoses and treatments are recorded in the course of routine medical practice.
Pre-assignment Detail	Actual number of patients may be less, as it is possible for a patient to be represented in more than 1 of the 22 arms (See "Participant Flow: Overall Study" Table) because of the risk set sampling.

Arm/Group Title	All-Cancer: Cases	All-Cancer: Controls	Stomach Cancer: Cases	Stomach Cancer: Controls	Anal Cancer: Cases	Anal Cancer: Controls	Lung Cancer: Cases	Lung Cancer: Controls	Bone/Joint Cancer: Cases	Bone/Joint Cancer: Controls	Breast Cancer: Cases	Breast Cancer: Controls	Penile Cancer: Cases	Penile Cancer: Controls	Bladder Cancer: Cases	Bladder Cancer: Controls	Other Nervous System Cancer: Cases	Other Nervous System Cancer: Controls	Pancreatic Cancer: Cases	Pancreatic Cancer: Controls	Renal Cancer: Cases	Renal Cancer: Controls
Arm/Group Description	Incidence of all cancers defined as first time incident cancer diagnosis (READ/Oxford Medical Information System [OXMIS] codes) in the General Practice Research Database (GPRD) study cohort. Entry into the study cohort began January 1, 1993, or at the time of GPRD registration if after January 1, 1993. Follow-up ended December 31, 2008, or earlier if the respective cancer was diagnosed or if the participant left the GPRD for any reason including death. Participants with a prior history of cancer were excluded.	Cancer cases were risk set matched with up to 10 controls for sex, age at GPRD study cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site. The index date for controls was set as the date at which the follow-up time from cohort entry was the same as the case. The index date was chosen so as to give the control equal follow-up time to that of the case for ascertainment of use of gabapentin. Participants with a history of cancer prior to the index date were excluded as controls.	Incidence of stomach cancer defined as first time incident cancer diagnosis (READ/Oxford Medical Information System [OXMIS] codes) in the General Practice Research Database (GPRD) study cohort. Entry into the study cohort began January 1, 1993, or at the time of GPRD registration if after January 1, 1993. Follow-up ended December 31, 2008, or earlier if the respective cancer was diagnosed or if the participant left the GPRD for any reason including death. Participants with a prior history of cancer were excluded.	Cancer cases were risk set matched with up to 10 controls for sex, age at GPRD study cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site. The index date for controls was set as the date at which the follow-up time from cohort entry was the same as the case. The index date was chosen so as to give the control equal follow-up time to that of the case for ascertainment of use of gabapentin. Participants with a history of cancer prior to the index date were excluded as controls.	Incidence of anal cancer defined as first time incident cancer diagnosis (READ/Oxford Medical Information System [OXMIS] codes) in the General Practice Research Database (GPRD) study cohort. Entry into the study cohort began January 1, 1993, or at the time of GPRD registration if after January 1, 1993. Follow-up ended December 31, 2008, or earlier if the respective cancer was diagnosed or if the participant left the GPRD for any reason including death. Participants with a prior history of cancer were excluded.	Cancer cases were risk set matched with up to 10 controls for sex, age at GPRD study cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site. The index date for controls was set as the date at which the follow-up time from cohort entry was the same as the case. The index date was chosen so as to give the control equal follow-up time to that of the case for ascertainment of use of gabapentin. Participants with a history of cancer prior to the index date were excluded as controls.	Incidence of lung cancer defined as first time incident cancer diagnosis (READ/Oxford Medical Information System [OXMIS] codes) in the General Practice Research Database (GPRD) study cohort. Entry into the study cohort began January 1, 1993, or at the time of GPRD registration if after January 1, 1993. Follow-up ended December 31, 2008, or earlier if the respective cancer was diagnosed or if the participant left the GPRD for any reason including death. Participants with a prior history of cancer were excluded.	Cancer cases were risk set matched with up to 10 controls for sex, age at GPRD study cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site. The index date for controls was set as the date at which the follow-up time from cohort entry was the same as the case. The index date was chosen so as to give the control equal follow-up time to that of the case for ascertainment of use of gabapentin. Participants with a history of cancer prior to the index date were excluded as controls.	Incidence of bone/joint cancer defined as first time incident cancer diagnosis (READ/Oxford Medical Information System [OXMIS] codes) in the General Practice Research Database (GPRD) study cohort. Entry into the study cohort began January 1, 1993, or at the time of GPRD registration if after January 1, 1993. Follow-up ended December 31, 2008, or earlier if the respective cancer was diagnosed or if the participant left the GPRD for any reason including death. Participants with a prior history of cancer were excluded.	Cancer cases were risk set matched with up to 10 controls for sex, age at GPRD study cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site. The index date for controls was set as the date at which the follow-up time from cohort entry was the same as the case. The index date was chosen so as to give the control equal follow-up time to that of the case for ascertainment of use of gabapentin. Participants with a history of cancer prior to the index date were excluded as controls.	Incidence of breast cancer defined as first time incident cancer diagnosis (READ/Oxford Medical Information System [OXMIS] codes) in the General Practice Research Database (GPRD) study cohort. Entry into the study cohort began January 1, 1993, or at the time of GPRD registration if after January 1, 1993. Follow-up ended December 31, 2008, or earlier if the respective cancer was diagnosed or if the participant left the GPRD for any reason including death. Participants with a prior history of cancer were excluded.	Cancer cases were risk set matched with up to 10 controls for sex, age at GPRD study cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site. The index date for controls was set as the date at which the follow-up time from cohort entry was the same as the case. The index date was chosen so as to give the control equal follow-up time to that of the case for ascertainment of use of gabapentin. Participants with a history of cancer prior to the index date were excluded as controls.	Incidence of penile cancer defined as first time incident cancer diagnosis (READ/Oxford Medical Information System [OXMIS] codes) in the General Practice Research Database (GPRD) study cohort. Entry into the study cohort began January 1, 1993, or at the time of GPRD registration if after January 1, 1993. Follow-up ended December 31, 2008, or earlier if the respective cancer was diagnosed or if the participant left the GPRD for any reason including death. Participants with a prior history of cancer were excluded.	Cancer cases were risk set matched with up to 10 controls for sex, age at GPRD study cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site. The index date for controls was set as the date at which the follow-up time from cohort entry was the same as the case. The index date was chosen so as to give the control equal follow-up time to that of the case for ascertainment of use of gabapentin. Participants with a history of cancer prior to the index date were excluded as controls.	Incidence of bladder cancer defined as first time incident cancer diagnosis (READ/Oxford Medical Information System [OXMIS] codes) in the General Practice Research Database (GPRD) study cohort. Entry into the study cohort began January 1, 1993, or at the time of GPRD registration if after January 1, 1993. Follow-up ended December 31, 2008, or earlier if the respective cancer was diagnosed or if the participant left the GPRD for any reason including death. Participants with a prior history of cancer were excluded.	Cancer cases were risk set matched with up to 10 controls for sex, age at GPRD study cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site. The index date for controls was set as the date at which the follow-up time from cohort entry was the same as the case. The index date was chosen so as to give the control equal follow-up time to that of the case for ascertainment of use of gabapentin. Participants with a history of cancer prior to the index date were excluded as controls.	Incidence of other nervous system cancer defined as first time incident cancer diagnosis (READ/Oxford Medical Information System [OXMIS] codes) in the General Practice Research Database (GPRD) study cohort. Entry into the study cohort began January 1, 1993, or at the time of GPRD registration if after January 1, 1993. Follow-up ended December 31, 2008, or earlier if the respective cancer was diagnosed or if the participant left the GPRD for any reason including death. Participants with a prior history of cancer were excluded.	Cancer cases were risk set matched with up to 10 controls for sex, age at GPRD study cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site. The index date for controls was set as the date at which the follow-up time from cohort entry was the same as the case. The index date was chosen so as to give the control equal follow-up time to that of the case for ascertainment of use of gabapentin. Participants with a history of cancer prior to the index date were excluded as controls.	Incidence of pancreatic cancer defined as first time incident cancer diagnosis (READ/Oxford Medical Information System [OXMIS] codes) in the General Practice Research Database (GPRD) study cohort. Entry into the study cohort began January 1, 1993, or at the time of GPRD registration if after January 1, 1993. Follow-up ended December 31, 2008, or earlier if the respective cancer was diagnosed or if the participant left the GPRD for any reason including death. Participants with a prior history of cancer were excluded.	Cancer cases were risk set matched with up to 10 controls for sex, age at GPRD study cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site. The index date for controls was set as the date at which the follow-up time from cohort entry was the same as the case. The index date was chosen so as to give the control equal follow-up time to that of the case for ascertainment of use of gabapentin. Participants with a history of cancer prior to the index date were excluded as controls.	Incidence of renal cancer defined as first time incident cancer diagnosis (READ/Oxford Medical Information System [OXMIS] codes) in the General Practice Research Database (GPRD) study cohort. Entry into the study cohort began January 1, 1993, or at the time of GPRD registration if after January 1, 1993. Follow-up ended December 31, 2008, or earlier if the respective cancer was diagnosed or if the participant left the GPRD for any reason including death. Participants with a prior history of cancer were excluded.	Cancer cases were risk set matched with up to 10 controls for sex, age at GPRD study cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site. The index date for controls was set as the date at which the follow-up time from cohort entry was the same as the case. The index date was chosen so as to give the control equal follow-up time to that of the case for ascertainment of use of gabapentin. Participants with a history of cancer prior to the index date were excluded as controls.
Period Title: Overall Study
STARTED	179138	1710950	1877	17853	212	2067	10855	102836	300	2935	19564	188924	148	1396	4600	43559	38	380	2155	20382	1272	12167
COMPLETED	179138	1710950	1877	17853	212	2067	10855	102836	300	2935	19564	188924	148	1396	4600	43559	38	380	2155	20382	1272	12167
NOT COMPLETED	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0

Baseline Characteristics

Arm/Group Title	All-Cancer: Cases	All-Cancer: Controls	Stomach Cancer: Cases	Stomach Cancer: Controls	Anal Cancer: Cases	Anal Cancer: Controls	Lung Cancer: Cases	Lung Cancer: Controls	Bone/Joint Cancer: Cases	Bone/Joint Cancer: Controls	Breast Cancer: Cases	Breast Cancer: Controls	Penile Cancer: Cases	Penile Cancer: Controls	Bladder Cancer: Cases	Bladder Cancer: Controls	Other Nervous System Cancer: Cases	Other Nervous System Cancer: Controls	Pancreatic Cancer: Cases	Pancreatic Cancer: Controls	Renal Cancer: Cases	Renal Cancer: Controls	Total
Arm/Group Description	Incidence of all cancers defined as first time incident cancer diagnosis (READ/Oxford Medical Information System [OXMIS] codes) in the General Practice Research Database (GPRD) study cohort. Entry into the study cohort began January 1, 1993, or at the time of GPRD registration if after January 1, 1993. Follow-up ended December 31, 2008, or earlier if the respective cancer was diagnosed or if the participant left the GPRD for any reason including death. Participants with a prior history of cancer were excluded.	Cancer cases were risk set matched with up to 10 controls for sex, age at GPRD study cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site. The index date for controls was set as the date at which the follow-up time from cohort entry was the same as the case. The index date was chosen so as to give the control equal follow-up time to that of the case for ascertainment of use of gabapentin. Participants with a history of cancer prior to the index date were excluded as controls.	Incidence of stomach cancer defined as first time incident cancer diagnosis (READ/Oxford Medical Information System [OXMIS] codes) in the General Practice Research Database (GPRD) study cohort. Entry into the study cohort began January 1, 1993, or at the time of GPRD registration if after January 1, 1993. Follow-up ended December 31, 2008, or earlier if the respective cancer was diagnosed or if the participant left the GPRD for any reason including death. Participants with a prior history of cancer were excluded.	Cancer cases were risk set matched with up to 10 controls for sex, age at GPRD study cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site. The index date for controls was set as the date at which the follow-up time from cohort entry was the same as the case. The index date was chosen so as to give the control equal follow-up time to that of the case for ascertainment of use of gabapentin. Participants with a history of cancer prior to the index date were excluded as controls.	Incidence of anal cancer defined as first time incident cancer diagnosis (READ/Oxford Medical Information System [OXMIS] codes) in the General Practice Research Database (GPRD) study cohort. Entry into the study cohort began January 1, 1993, or at the time of GPRD registration if after January 1, 1993. Follow-up ended December 31, 2008, or earlier if the respective cancer was diagnosed or if the participant left the GPRD for any reason including death. Participants with a prior history of cancer were excluded.	Cancer cases were risk set matched with up to 10 controls for sex, age at GPRD study cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site. The index date for controls was set as the date at which the follow-up time from cohort entry was the same as the case. The index date was chosen so as to give the control equal follow-up time to that of the case for ascertainment of use of gabapentin. Participants with a history of cancer prior to the index date were excluded as controls.	Incidence of lung cancer defined as first time incident cancer diagnosis (READ/Oxford Medical Information System [OXMIS] codes) in the General Practice Research Database (GPRD) study cohort. Entry into the study cohort began January 1, 1993, or at the time of GPRD registration if after January 1, 1993. Follow-up ended December 31, 2008, or earlier if the respective cancer was diagnosed or if the participant left the GPRD for any reason including death. Participants with a prior history of cancer were excluded.	Cancer cases were risk set matched with up to 10 controls for sex, age at GPRD study cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site. The index date for controls was set as the date at which the follow-up time from cohort entry was the same as the case. The index date was chosen so as to give the control equal follow-up time to that of the case for ascertainment of use of gabapentin. Participants with a history of cancer prior to the index date were excluded as controls.	Incidence of bone/joint cancer defined as first time incident cancer diagnosis (READ/Oxford Medical Information System [OXMIS] codes) in the General Practice Research Database (GPRD) study cohort. Entry into the study cohort began January 1, 1993, or at the time of GPRD registration if after January 1, 1993. Follow-up ended December 31, 2008, or earlier if the respective cancer was diagnosed or if the participant left the GPRD for any reason including death. Participants with a prior history of cancer were excluded.	Cancer cases were risk set matched with up to 10 controls for sex, age at GPRD study cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site. The index date for controls was set as the date at which the follow-up time from cohort entry was the same as the case. The index date was chosen so as to give the control equal follow-up time to that of the case for ascertainment of use of gabapentin. Participants with a history of cancer prior to the index date were excluded as controls.	Incidence of breast cancer defined as first time incident cancer diagnosis (READ/Oxford Medical Information System [OXMIS] codes) in the General Practice Research Database (GPRD) study cohort. Entry into the study cohort began January 1, 1993, or at the time of GPRD registration if after January 1, 1993. Follow-up ended December 31, 2008, or earlier if the respective cancer was diagnosed or if the participant left the GPRD for any reason including death. Participants with a prior history of cancer were excluded.	Cancer cases were risk set matched with up to 10 controls for sex, age at GPRD study cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site. The index date for controls was set as the date at which the follow-up time from cohort entry was the same as the case. The index date was chosen so as to give the control equal follow-up time to that of the case for ascertainment of use of gabapentin. Participants with a history of cancer prior to the index date were excluded as controls.	Incidence of penile cancer defined as first time incident cancer diagnosis (READ/Oxford Medical Information System [OXMIS] codes) in the General Practice Research Database (GPRD) study cohort. Entry into the study cohort began January 1, 1993, or at the time of GPRD registration if after January 1, 1993. Follow-up ended December 31, 2008, or earlier if the respective cancer was diagnosed or if the participant left the GPRD for any reason including death. Participants with a prior history of cancer were excluded.	Cancer cases were risk set matched with up to 10 controls for sex, age at GPRD study cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site. The index date for controls was set as the date at which the follow-up time from cohort entry was the same as the case. The index date was chosen so as to give the control equal follow-up time to that of the case for ascertainment of use of gabapentin. Participants with a history of cancer prior to the index date were excluded as controls.	Incidence of bladder cancer defined as first time incident cancer diagnosis (READ/Oxford Medical Information System [OXMIS] codes) in the General Practice Research Database (GPRD) study cohort. Entry into the study cohort began January 1, 1993, or at the time of GPRD registration if after January 1, 1993. Follow-up ended December 31, 2008, or earlier if the respective cancer was diagnosed or if the participant left the GPRD for any reason including death. Participants with a prior history of cancer were excluded.	Cancer cases were risk set matched with up to 10 controls for sex, age at GPRD study cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site. The index date for controls was set as the date at which the follow-up time from cohort entry was the same as the case. The index date was chosen so as to give the control equal follow-up time to that of the case for ascertainment of use of gabapentin. Participants with a history of cancer prior to the index date were excluded as controls.	Incidence of other nervous system cancer defined as first time incident cancer diagnosis (READ/Oxford Medical Information System [OXMIS] codes) in the General Practice Research Database (GPRD) study cohort. Entry into the study cohort began January 1, 1993, or at the time of GPRD registration if after January 1, 1993. Follow-up ended December 31, 2008, or earlier if the respective cancer was diagnosed or if the participant left the GPRD for any reason including death. Participants with a prior history of cancer were excluded.	Cancer cases were risk set matched with up to 10 controls for sex, age at GPRD study cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site. The index date for controls was set as the date at which the follow-up time from cohort entry was the same as the case. The index date was chosen so as to give the control equal follow-up time to that of the case for ascertainment of use of gabapentin. Participants with a history of cancer prior to the index date were excluded as controls.	Incidence of pancreatic cancer defined as first time incident cancer diagnosis (READ/Oxford Medical Information System [OXMIS] codes) in the General Practice Research Database (GPRD) study cohort. Entry into the study cohort began January 1, 1993, or at the time of GPRD registration if after January 1, 1993. Follow-up ended December 31, 2008, or earlier if the respective cancer was diagnosed or if the participant left the GPRD for any reason including death. Participants with a prior history of cancer were excluded.	Cancer cases were risk set matched with up to 10 controls for sex, age at GPRD study cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site. The index date for controls was set as the date at which the follow-up time from cohort entry was the same as the case. The index date was chosen so as to give the control equal follow-up time to that of the case for ascertainment of use of gabapentin. Participants with a history of cancer prior to the index date were excluded as controls.	Incidence of renal cancer defined as first time incident cancer diagnosis (READ/Oxford Medical Information System [OXMIS] codes) in the General Practice Research Database (GPRD) study cohort. Entry into the study cohort began January 1, 1993, or at the time of GPRD registration if after January 1, 1993. Follow-up ended December 31, 2008, or earlier if the respective cancer was diagnosed or if the participant left the GPRD for any reason including death. Participants with a prior history of cancer were excluded.	Cancer cases were risk set matched with up to 10 controls for sex, age at GPRD study cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site. The index date for controls was set as the date at which the follow-up time from cohort entry was the same as the case. The index date was chosen so as to give the control equal follow-up time to that of the case for ascertainment of use of gabapentin. Participants with a history of cancer prior to the index date were excluded as controls.	Total of all reporting groups
Overall Participants	179138	1710950	1877	17853	212	2067	10855	102836	300	2935	19564	188924	148	1396	4600	43559	38	380	2155	20382	1272	12167	2323608
Age, Customized (participants) [Number]
< 40 years	10246 5.7%	101968 6%	16 0.9%	155 0.9%	12 5.7%	118 5.7%	36 0.3%	369 0.4%	120 40%	1173 40%	915 4.7%	9346 4.9%	9 6.1%	90 6.4%	56 1.2%	552 1.3%	5 13.2%	54 14.2%	11 0.5%	132 0.6%	31 2.4%	317 2.6%	125731 5.4%
40-49 years	13112 7.3%	132244 7.7%	48 2.6%	469 2.6%	18 8.5%	200 9.7%	294 2.7%	2973 2.9%	21 7%	254 8.7%	2827 14.5%	29167 15.4%	16 10.8%	154 11%	120 2.6%	1285 3%	8 21.1%	78 20.5%	65 3%	669 3.3%	99 7.8%	990 8.1%	185111 8%
50-59 years	28392 15.8%	279797 16.4%	135 7.2%	1388 7.8%	47 22.2%	467 22.6%	1256 11.6%	12388 12%	46 15.3%	422 14.4%	5319 27.2%	51266 27.1%	30 20.3%	300 21.5%	547 11.9%	5296 12.2%	5 13.2%	54 14.2%	261 12.1%	2564 12.6%	249 19.6%	2406 19.8%	392635 16.9%
60-69 years	43077 24%	415567 24.3%	395 21%	3850 21.6%	45 21.2%	464 22.4%	2944 27.1%	28526 27.7%	38 12.7%	360 12.3%	4665 23.8%	44769 23.7%	32 21.6%	320 22.9%	1127 24.5%	11021 25.3%	9 23.7%	78 20.5%	515 23.9%	5043 24.7%	362 28.5%	3498 28.7%	566705 24.4%
70-79 years	50072 28%	475887 27.8%	679 36.2%	6559 36.7%	50 23.6%	440 21.3%	4092 37.7%	38557 37.5%	51 17%	501 17.1%	3334 17%	32051 17%	44 29.7%	388 27.8%	1644 35.7%	15471 35.5%	4 10.5%	48 12.6%	716 33.2%	6813 33.4%	361 28.4%	3399 27.9%	641161 27.6%
80+ years	34239 19.1%	305487 17.9%	604 32.2%	5432 30.4%	40 18.9%	378 18.3%	2233 20.6%	20023 19.5%	24 8%	225 7.7%	2504 12.8%	22325 11.8%	17 11.5%	144 10.3%	1106 24%	9934 22.8%	7 18.4%	68 17.9%	587 27.2%	5161 25.3%	170 13.4%	1557 12.8%	412265 17.7%
Sex: Female, Male (Count of Participants)
Female	92099 51.4%	883331 51.6%	716 38.1%	6789 38%	133 62.7%	1294 62.6%	4259 39.2%	40345 39.2%	156 52%	1538 52.4%	19564 100%	188924 100%	0 0%	0 0%	1229 26.7%	11552 26.5%	14 36.8%	140 36.8%	1109 51.5%	10530 51.7%	507 39.9%	4820 39.6%	1269049 54.6%
Male	87039 48.6%	827619 48.4%	1161 61.9%	11064 62%	79 37.3%	773 37.4%	6596 60.8%	62491 60.8%	144 48%	1397 47.6%	0 0%	0 0%	148 100%	1396 100%	3371 73.3%	32007 73.5%	24 63.2%	240 63.2%	1046 48.5%	9852 48.3%	765 60.1%	7347 60.4%	1054559 45.4%
Number of participants with the indicated duration of follow-up from registration to index date (participants) [Number]
0-1 years	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	11846 6.3%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	11846 0.5%
2-3 years	12001 6.7%	99830 5.8%	115 6.1%	866 4.9%	14 6.6%	135 6.5%	644 5.9%	5014 4.9%	35 11.7%	331 11.3%	1361 7%	11846 6.3%	9 6.1%	73 5.2%	251 5.5%	2007 4.6%	2 5.3%	13 3.4%	118 5.5%	968 4.7%	91 7.2%	762 6.3%	136486 5.9%
4-5 years	12675 7.1%	108112 6.3%	123 6.6%	1048 5.9%	20 9.4%	164 7.9%	708 6.5%	5868 5.7%	35 11.7%	308 10.5%	1426 7.3%	12724 6.7%	15 10.1%	136 9.7%	283 6.2%	2403 5.5%	0 0%	0 0%	122 5.7%	1028 5%	72 5.7%	695 5.7%	147965 6.4%
6-7 years	13403 7.5%	119691 7%	157 8.4%	1321 7.4%	13 6.1%	121 5.9%	802 7.4%	6978 6.8%	29 9.7%	264 9%	1508 7.7%	14037 7.4%	12 8.1%	104 7.4%	304 6.6%	2694 6.2%	5 13.2%	41 10.8%	173 8%	1418 7%	95 7.5%	839 6.9%	164009 7.1%
8+ years	141059 78.7%	1383317 80.9%	1482 79%	14618 81.9%	165 77.8%	1647 79.7%	8681 80%	84976 82.6%	201 67%	2032 69.2%	15269 78%	150317 79.6%	112 75.7%	1083 77.6%	3762 81.8%	36455 83.7%	31 81.6%	326 85.8%	1742 80.8%	16968 83.2%	1014 79.7%	9871 81.1%	1875128 80.7%
Number of participants with the indicated duration of follow-up from cohort entry to index date (participants) [Number]
0-1 years	0 0%	0 0%	0 0%	3216 18%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	3216 0.1%
2-3 years	24693 13.8%	226195 13.2%	351 18.7%	3216 18%	32 15.1%	319 15.4%	1627 15%	14736 14.3%	48 16%	458 15.6%	2784 14.2%	26004 13.8%	24 16.2%	212 15.2%	592 12.9%	5351 12.3%	2 5.3%	20 5.3%	310 14.4%	2835 13.9%	186 14.6%	1704 14%	311699 13.4%
4-5 years	24123 13.5%	223084 13%	281 15%	2628 14.7%	32 15.1%	288 13.9%	1489 13.7%	13563 13.2%	48 16%	451 15.4%	2718 13.9%	25616 13.6%	28 18.9%	264 18.9%	609 13.2%	5502 12.6%	2 5.3%	20 5.3%	283 13.1%	2596 12.7%	149 11.7%	1388 11.4%	305162 13.1%
6-7 years	25608 14.3%	241586 14.1%	231 12.3%	2108 11.8%	26 12.3%	257 12.4%	1577 14.5%	14855 14.4%	49 16.3%	490 16.7%	2830 14.5%	27109 14.3%	17 11.5%	159 11.4%	618 13.4%	5824 13.4%	6 15.8%	60 15.8%	318 14.8%	2972 14.6%	165 13%	1564 12.9%	328429 14.1%
8+ years	104714 58.5%	1020085 59.6%	1014 54%	9901 55.5%	122 57.5%	1203 58.2%	6162 56.8%	59682 58%	155 51.7%	1536 52.3%	11232 57.4%	110195 58.3%	79 53.4%	761 54.5%	2781 60.5%	26882 61.7%	28 73.7%	280 73.7%	1244 57.7%	11979 58.8%	772 60.7%	7511 61.7%	1378318 59.3%
Number of participants with the indicated Body Mass Index (BMI) (participants) [Number]
BMI: < 18.5 kg/m^2	2393 1.3%	19861 1.2%	30 1.6%	190 1.1%	5 2.4%	30 1.5%	255 2.3%	1081 1.1%	7 2.3%	38 1.3%	267 1.4%	3059 1.6%	1 0.7%	7 0.5%	52 1.1%	380 0.9%	0 0%	4 1.1%	30 1.4%	232 1.1%	7 0.6%	128 1.1%	28057 1.2%
BMI: 18.5 to 24.99 kg/m^2	61913 34.6%	569726 33.3%	580 30.9%	5429 30.4%	82 38.7%	674 32.6%	3912 36%	31787 30.9%	72 24%	741 25.2%	7813 39.9%	75516 40%	42 28.4%	420 30.1%	1408 30.6%	13224 30.4%	11 28.9%	106 27.9%	662 30.7%	6313 31%	331 26%	3753 30.8%	784515 33.8%
BMI: 25 to 29.99 kg/m^2	54471 30.4%	508109 29.7%	579 30.8%	5483 30.7%	55 25.9%	581 28.1%	3070 28.3%	33035 32.1%	70 23.3%	580 19.8%	5054 25.8%	47989 25.4%	50 33.8%	461 33%	1591 34.6%	14390 33%	11 28.9%	111 29.2%	662 30.7%	6313 31%	447 35.1%	3958 32.5%	687070 29.6%
BMI: 30 kg/m^2 or greater	22152 12.4%	206232 12.1%	245 13.1%	2001 11.2%	21 9.9%	254 12.3%	1097 10.1%	12814 12.5%	25 8.3%	285 9.7%	2773 14.2%	25819 13.7%	22 14.9%	142 10.2%	587 12.8%	5117 11.7%	7 18.4%	37 9.7%	294 13.6%	2492 12.2%	245 19.3%	1626 13.4%	284287 12.2%
BMI: Missing	38209 21.3%	407022 23.8%	443 23.6%	4750 26.6%	49 23.1%	528 25.5%	2521 23.2%	24119 23.5%	126 42%	1291 44%	3657 18.7%	36541 19.3%	33 22.3%	366 26.2%	962 20.9%	10448 24%	9 23.7%	122 32.1%	507 23.5%	5027 24.7%	242 19%	2702 22.2%	539674 23.2%
Number of participants with the indicated smoking status (participants) [Number]
Current Smoker	38033 21.2%	306184 17.9%	375 20%	2902 16.3%	63 29.7%	379 18.3%	4806 44.3%	18856 18.3%	51 17%	476 16.2%	3572 18.3%	34818 18.4%	51 34.5%	297 21.3%	1310 28.5%	7502 17.2%	10 26.3%	71 18.7%	518 24%	3368 16.5%	309 24.3%	2324 19.1%	426275 18.3%
Ex-Smoker	41270 23%	354651 20.7%	531 28.3%	4134 23.2%	41 19.3%	380 18.4%	3594 33.1%	24290 23.6%	43 14.3%	385 13.1%	3156 16.1%	28210 14.9%	29 19.6%	334 23.9%	1307 28.4%	11070 25.4%	6 15.8%	77 20.3%	499 23.2%	4247 20.8%	319 25.1%	2768 22.8%	481341 20.7%
Never Smoked	85483 47.7%	854263 49.9%	810 43.2%	8540 47.8%	83 39.2%	1048 50.7%	1766 16.3%	48127 46.8%	133 44.3%	1255 42.8%	11396 58.2%	109463 57.9%	58 39.2%	587 42%	1636 35.6%	20018 46%	17 44.7%	176 46.3%	929 43.1%	10455 51.3%	570 44.8%	5875 48.3%	1162688 50%
Unknown	14352 8%	195852 11.4%	161 8.6%	2277 12.8%	25 11.8%	260 12.6%	689 6.3%	11563 11.2%	73 24.3%	819 27.9%	1440 7.4%	16433 8.7%	10 6.8%	178 12.8%	347 7.5%	4969 11.4%	5 13.2%	56 14.7%	209 9.7%	2312 11.3%	74 5.8%	1200 9.9%	253304 10.9%
Number of participants with the indicated alcohol consumption (per day) (participants) [Number]
0 units/day	36134 20.2%	340399 19.9%	422 22.5%	3817 21.4%	0 0%	0 0%	2412 22.2%	21471 20.9%	0 0%	0 0%	4585 23.4%	44823 23.7%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	454063 19.5%
1-2 units/day	86422 48.2%	801383 46.8%	839 44.7%	7659 42.9%	0 0%	0 0%	4661 42.9%	46769 45.5%	0 0%	0 0%	9947 50.8%	93964 49.7%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	1051644 45.3%
3-6 units/day	11929 6.7%	102201 6%	115 6.1%	1118 6.3%	0 0%	0 0%	868 8%	6843 6.7%	0 0%	0 0%	535 2.7%	4544 2.4%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	128153 5.5%
7+ units/day	1992 1.1%	15014 0.9%	18 1%	157 0.9%	0 0%	0 0%	202 1.9%	1132 1.1%	0 0%	0 0%	50 0.3%	399 0.2%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	18964 0.8%
Missing	42661 23.8%	451953 26.4%	483 25.7%	5102 28.6%	0 0%	0 0%	2712 25%	26621 25.9%	0 0%	0 0%	4447 22.7%	45194 23.9%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	579173 24.9%
Number of participants using indicated class of medications or having the indicated comorbidities (participants) [Number]
Acid suppressing drugs (>6 mo. before index date)	0 0%	0 0%	796 42.4%	5202 29.1%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	5998 0.3%
Back pain	87337 48.8%	765861 44.8%	924 49.2%	7779 43.6%	92 43.4%	924 44.7%	5626 51.8%	46417 45.1%	130 43.3%	1008 34.3%	9603 49.1%	91302 48.3%	59 39.9%	569 40.8%	2149 46.7%	18811 43.2%	23 60.5%	192 50.5%	1081 50.2%	9186 45.1%	678 53.3%	5473 45%	1055224 45.4%
Benign breast disease	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	894 4.6%	5508 2.9%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	6402 0.3%
Chronic obstructive pulmonary disease (COPD)	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	2532 23.3%	6639 6.5%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	9171 0.4%
Chronic pancreatitis	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	13 0.6%	8 0%	0 0%	0 0%	21 0%
Current estrogen (within 6 months of index date)	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	2916 14.9%	22576 11.9%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	25492 1.1%
Diabetes	15280 8.5%	130210 7.6%	208 11.1%	1592 8.9%	17 8%	139 6.7%	1086 10%	9448 9.2%	15 5%	155 5.3%	1162 5.9%	9899 5.2%	18 12.2%	104 7.4%	559 12.2%	4087 9.4%	3 7.9%	32 8.4%	241 11.2%	1415 6.9%	155 12.2%	968 8%	176793 7.6%
Diuretic use (at any time up to the index date)	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	554 43.6%	3815 31.4%	4369 0.2%
Epilepsy	3015 1.7%	24226 1.4%	32 1.7%	271 1.5%	8 3.8%	23 1.1%	214 2%	1577 1.5%	3 1%	33 1.1%	285 1.5%	2486 1.3%	0 0%	28 2%	78 1.7%	654 1.5%	3 7.9%	4 1.1%	35 1.6%	289 1.4%	20 1.6%	176 1.4%	33460 1.4%
Human immunodeficiency virus (HIV)	0 0%	0 0%	0 0%	0 0%	3 1.4%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	3 0%
Human papillomavirus (HPV) (genital)	0 0%	0 0%	0 0%	0 0%	15 7.1%	7 0.3%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	9 6.1%	12 0.9%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	43 0%
Hypertension	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	749 58.9%	5580 45.9%	6329 0.3%
Hysterectomy	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	3563 18.2%	34290 18.2%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	37853 1.6%
Neuropathic pain	38561 21.5%	338723 19.8%	382 20.4%	3485 19.5%	47 22.2%	406 19.6%	2485 22.9%	21202 20.6%	49 16.3%	387 13.2%	4055 20.7%	39494 20.9%	25 16.9%	229 16.4%	998 21.7%	8539 19.6%	15 39.5%	71 18.7%	474 22%	4195 20.6%	275 21.6%	2434 20%	466531 20.1%
Phimosis	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	54 36.5%	31 2.2%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	85 0%
Prior estrogen	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	6803 34.8%	60840 32.2%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	67643 2.9%
Upper gastro-intestinal disorder	0 0%	0 0%	442 23.5%	2860 16%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	0 0%	3302 0.1%

Outcome Measures

1. Primary Outcome

Title	Number of All-Cancer Cases and Matched Controls With the Indicated Exposure to Gabapentin
Description	Incidence of all cancers. Gabapentin Exposure Description: Without 2 year lag = Gabapentin prescription from cohort entry to index date. With 2 year lag = Gabapentin prescription from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer).
Time Frame	The case index date was the date of incident cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident All-cancer Patients	Matched Controls for Incident All-cancer Patients
Arm/Group Description	Patients with any type of incident cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident all-cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	179138	1710950
Ever (with 2 year lag)	804 0.4%	6507 0.4%
Never (with 2 year lag)	178334 99.6%	1704443 99.6%
Ever (without 2 year lag)	1729 1%	12470 0.7%
Never (without 2 year lag)	177409 99%	1698480 99.3%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0002
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.15
	Confidence Interval	(2-Sided) 95%; 1.07 to 1.24
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Ever (with 2 year lag)" and "Never (with 2 year lag)"

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0976
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.07
	Confidence Interval	(2-Sided) 95%; 0.99 to 1.15
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Ever (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy.

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	<0.0001
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.29
	Confidence Interval	(2-Sided) 95%; 1.23 to 1.36
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Ever (without 2 year lag)" and "Never (without 2 year lag)"

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	<0.0001
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.19
	Confidence Interval	(2-Sided) 95%; 1.13 to 1.26
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Ever (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy.

2. Primary Outcome

Title	Number of All-Cancer Cases and Matched Controls With the Indicated Number of Gabapentin Prescriptions
Description	Incidence of all cancers. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date. Tertile's with 2 year lag: Tertile 1 (1-2 prescriptions), Tertile 2 (3-7 prescriptions), Tertile 3 (8-298 prescriptions). Tertiles without 2 year lag: Tertile 1 (1-2 prescriptions), Tertile 2 (3-7 prescriptions), and Tertile 3 (8-388 prescriptions).
Time Frame	The case index date was the date of incident cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident All-cancer Patients	Matched Controls for Incident All-cancer Patients
Arm/Group Description	Patients with any type of incident cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident all-cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	179138	1710950
Never (with 2 year lag)	178334 99.6%	1704443 99.6%
Tertile 1 (with 2 year lag)	319 0.2%	2739 0.2%
Tertile 2 (with 2 year lag)	216 0.1%	1645 0.1%
Tertile 3 (with 2 year lag)	269 0.2%	2123 0.1%
Never (without 2 year lag)	177409 99%	1698480 99.3%
Tertile 1 (without 2 year lag)	781 0.4%	5285 0.3%
Tertile 2 (without 2 year lag)	413 0.2%	3033 0.2%
Tertile 3 (without 2 year lag)	535 0.3%	4152 0.2%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.1468
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.09
	Confidence Interval	(2-Sided) 95%; 0.97 to 1.23
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.8545
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.01
	Confidence Interval	(2-Sided) 95%; 0.90 to 1.14
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy.

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	<0.0001
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.37
	Confidence Interval	(2-Sided) 95%; 1.27 to 1.48
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	<0.0001
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.27
	Confidence Interval	(2-Sided) 95%; 1.18 to 1.37
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy.

Statistical Analysis 5

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0033
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.24
	Confidence Interval	(2-Sided) 95%; 1.07 to 1.43
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 6

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0474
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.16
	Confidence Interval	(2-Sided) 95%; 1.00 to 1.33
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy.

Statistical Analysis 7

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	<0.0001
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.29
	Confidence Interval	(2-Sided) 95%; 1.16 to 1.43
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 8

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0008
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.19
	Confidence Interval	(2-Sided) 95%; 1.08 to 1.32
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy.

Statistical Analysis 9

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0190
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.17
	Confidence Interval	(2-Sided) 95%; 1.03 to 1.33
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 10

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.3366
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.07
	Confidence Interval	(2-Sided) 95%; 0.94 to 1.21
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy.

Statistical Analysis 11

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0001
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.19
	Confidence Interval	(2-Sided) 95%; 1.09 to 1.31
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 12

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0464
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.10
	Confidence Interval	(2-Sided) 95%; 1.00 to 1.20
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy.

3. Primary Outcome

Title	Number of All-Cancer Cases and Matched Controls With the Indicated Duration of Exposure to Gabapentin
Description	Incidence of all cancers. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date. Tertile's with 2 year lag: Tertile 1 (0.01 - 1.38 months), Tertile 2 (1.39 - 5.56 months), and Tertile 3 (5.57 - 105.82 months). Tertile's without 2 year lag: Tertile 1 (0.01 - 1.38 months), Tertile 2 (1.39 - 5.72 months), and Tertile 3 (5.73 - 123.70 months).
Time Frame	The case index date was the date of incident cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident All-cancer Patients	Matched Controls for Incident All-cancer Patients
Arm/Group Description	Patients with any type of incident cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident all-cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	179138	1710950
Never (with 2 year lag)	178334 99.6%	1704443 99.6%
Tertile 1 (with 2 year lag)	258 0.1%	2175 0.1%
Tertile 2 (with 2 year lag)	269 0.2%	2163 0.1%
Tertile 3 (with 2 year lag)	277 0.2%	2169 0.1%
Never (without 2 year lag)	177409 99%	1698480 99.3%
Tertile 1 (without 2 year lag)	620 0.3%	4213 0.2%
Tertile 2 (without 2 year lag)	587 0.3%	4104 0.2%
Tertile 3 (without 2 year lag)	522 0.3%	4153 0.2%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.1194
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.11
	Confidence Interval	(2-Sided) 95%; 0.97 to 1.26
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.7004
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.03
	Confidence Interval	(2-Sided) 95%; 0.90 to 1.17
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy.

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	<0.0001
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.37
	Confidence Interval	(2-Sided) 95%; 1.26 to 1.49
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	<0.0001
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.26
	Confidence Interval	(2-Sided) 95%; 1.16 to 1.38
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy.

Statistical Analysis 5

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0134
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.17
	Confidence Interval	(2-Sided) 95%; 1.03 to 1.34
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 6

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.1592
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.10
	Confidence Interval	(2-Sided) 95%; 0.96 to 1.25
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy.

Statistical Analysis 7

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	<0.0001
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.35
	Confidence Interval	(2-Sided) 95%; 1.24 to 1.47
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 8

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	<0.0001
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.25
	Confidence Interval	(2-Sided) 95%; 1.14 to 1.36
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy.

Statistical Analysis 9

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0131
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.17
	Confidence Interval	(2-Sided) 95%; 1.03 to 1.33
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 10

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.2731
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.07
	Confidence Interval	(2-Sided) 95%; 0.95 to 1.22
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy.

Statistical Analysis 11

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0012
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.16
	Confidence Interval	(2-Sided) 95%; 1.06 to 1.28
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 12

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.1477
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.07
	Confidence Interval	(2-Sided) 95%; 0.98 to 1.17
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy.

4. Primary Outcome

Title	Number of All-Cancer Cases and Matched Controls With the Indicated Long Duration of Exposure to Gabapentin
Description	Incidence of all cancers. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date.
Time Frame	The case index date was the date of incident cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident All-cancer Patients	Matched Controls for Incident All-cancer Patients
Arm/Group Description	Patients with any type of incident cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident all-cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	179138	1710950
Never (with 2 year lag)	178334 99.6%	1704443 99.6%
> 1 year (with 2 year lag)	157 0.1%	1217 0.1%
> 2 year (with 2 year lag)	64 0%	463 0%
> 3 year (with 2 year lag)	25 0%	197 0%
Never (without 2 year lag)	177409 99%	1698480 99.3%
> 1 year (without 2 year lag)	301 0.2%	2499 0.1%
> 2 year (without 2 year lag)	159 0.1%	1207 0.1%
> 3 year (without 2 year lag)	76 0%	549 0%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0369
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.20
	Confidence Interval	(2-Sided) 95%; 1.01 to 1.41
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "> 1 year (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.3132
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.09
	Confidence Interval	(2-Sided) 95%; 0.92 to 1.29
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "> 1 year (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy.

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0762
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.12
	Confidence Interval	(2-Sided) 95%; 0.99 to 1.26
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "> 1 year (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.7243
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.02
	Confidence Interval	(2-Sided) 95%; 0.91 to 1.15
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "> 1 year (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy.

Statistical Analysis 5

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0775
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.27
	Confidence Interval	(2-Sided) 95%; 0.97 to 1.65
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "> 2 years (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 6

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.2649
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.16
	Confidence Interval	(2-Sided) 95%; 0.89 to 1.52
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "> 2 years (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy.

Statistical Analysis 7

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0257
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.21
	Confidence Interval	(2-Sided) 95%; 1.02 to 1.43
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "> 2 years (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 8

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.2540
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.10
	Confidence Interval	(2-Sided) 95%; 0.93 to 1.30
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "> 2 years (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy.

Statistical Analysis 9

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.5141
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.15
	Confidence Interval	(2-Sided) 95%; 0.76 to 1.75
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "> 3 years (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 10

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.8208
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.05
	Confidence Interval	(2-Sided) 95%; 0.69 to 1.60
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "> 3 years (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy.

Statistical Analysis 11

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0562
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.27
	Confidence Interval	(2-Sided) 95%; 0.99 to 1.62
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "> 3 years (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 12

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.2596
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.15
	Confidence Interval	(2-Sided) 95%; 0.90 to 1.47
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "> 3 years (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy.

5. Primary Outcome

Title	Number of All-Cancer Cases and Matched Controls With the Indicated Cumulative Dose of Gabapentin
Description	Incidence of all cancers. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date. Tertile's with 2 year lag: Tertile 1 (0.1 - 30.0 grams), Tertile 2 (30.1 - 189.0 grams), and Tertile 3 (189.1 - 9600.0 grams). Tertile's without 2 year lag: Tertile 1 (0.1 - 30.0 grams), Tertile 2 (30.1 - 189.0 grams), and Tertile 3 (189.1 - 11610.0 grams).
Time Frame	The case index date was the date of incident cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident All-cancer Patients	Matched Controls for Incident All-cancer Patients
Arm/Group Description	Patients with any type of incident cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident all-cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	179138	1710950
Never (with 2 year lag)	178334 99.6%	1704443 99.6%
Tertile 1 (with 2 year lag)	269 0.2%	2222 0.1%
Tertile 2 (with 2 year lag)	268 0.1%	2118 0.1%
Tertile 3 (with 2 year lag)	267 0.1%	2167 0.1%
Never (without 2 year lag)	177409 99%	1698480 99.3%
Tertile 1 (without 2 year lag)	640 0.4%	4438 0.3%
Tertile 2 (without 2 year lag)	585 0.3%	3882 0.2%
Tertile 3 (without 2 year lag)	504 0.3%	4150 0.2%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0709
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.12
	Confidence Interval	(2-Sided) 95%; 0.99 to 1.28
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.4996
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.05
	Confidence Interval	(2-Sided) 95%; 0.92 to 1.19
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy.

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	<0.0001
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.34
	Confidence Interval	(2-Sided) 95%; 1.23 to 1.45
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	<0.0001
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.24
	Confidence Interval	(2-Sided) 95%; 1.14 to 1.35
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy.

Statistical Analysis 5

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0062
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.20
	Confidence Interval	(2-Sided) 95%; 1.05 to 1.36
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 6

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.1144
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.11
	Confidence Interval	(2-Sided) 95%; 0.98 to 1.26
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy.

Statistical Analysis 7

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	<0.0001
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.42
	Confidence Interval	(2-Sided) 95%; 1.30 to 1.55
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 8

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	<0.0001
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.31
	Confidence Interval	(2-Sided) 95%; 1.20 to 1.43
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy.

Statistical Analysis 9

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0469
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.14
	Confidence Interval	(2-Sided) 95%; 1.00 to 1.29
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 10

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.5169
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.04
	Confidence Interval	(2-Sided) 95%; 0.92 to 1.19
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy.

Statistical Analysis 11

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0103
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.13
	Confidence Interval	(2-Sided) 95%; 1.03 to 1.24
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 12

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.4182
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.04
	Confidence Interval	(2-Sided) 95%; 0.95 to 1.14
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy.

6. Primary Outcome

Title	Number of Stomach Cancer Cases and Matched Controls With the Indicated Exposure to Gabapentin
Description	Incident stomach cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin prescription from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin prescription from cohort entry to index date.
Time Frame	The case index date was the date of incident stomach cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Stomach Cancer Patients	Matched Controls for Incident Stomach Cancer Patients
Arm/Group Description	Patients with incident stomach cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident stomach cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	1877	17853
Ever (with 2 year lag)	5 0%	70 0%
Never (with 2 year lag)	1872 1%	17783 1%
Ever (without 2 year lag)	13 0%	20 0%
Never (without 2 year lag)	1864 1%	17733 1%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.3597
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Unadjusted Odds Ratio
	Estimated Value	0.65
	Confidence Interval	(2-Sided) 95%; 0.26 to 1.62
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Ever (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.1845
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.54
	Confidence Interval	(2-Sided) 95%; 0.21 to 1.34
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Ever (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, upper GI disorder, acid-suppressing drugs.

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.9142
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.03
	Confidence Interval	(2-Sided) 95%; 0.58 to 1.84
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Ever (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.6524
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.87
	Confidence Interval	(2-Sided) 95%; 0.49 to 1.57
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Ever (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, upper GI disorder, acid-suppressing drugs.

7. Primary Outcome

Title	Number of Stomach Cancer Cases and Matched Controls With the Indicated Number of Gabapentin Prescriptions
Description	Incident stomach cancer. Gabapentin (Gaba.) Exposure Description: With 2 year (yr) lag=Gaba. exposure from cohort entry to 2 yr prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 yr lag=Gaba. exposure from cohort entry to index date. Tertiles (T) with 2 yr lag: T 1 (1-2 prescriptions [prescrip]), T 2 (3-7 prescrip), T 3 (8-298 prescrip). Tertiles without 2 yr lag: T 1 (1-2 prescrip), T 2 (3-7 prescrip), and T 3 (8-388 prescrip). Inestimable OR and 95% CI when no gaba.-exposed cancer cases or controls at the exposure level.
Time Frame	The case index date was the date of incident stomach cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Stomach Cancer Patients	Matched Controls for Incident Stomach Cancer Patients
Arm/Group Description	Patients with incident stomach cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident stomach cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	1877	17853
Never (with 2 year lag)	1872 1%	17783 1%
Tertile 1 (with 2 year lag)	0 0%	29 0%
Tertile 2 (with 2 year lag)	2 0%	10 0%
Tertile 3 (with 2 year lag)	3 0%	31 0%
Never (without 2 year lag)	1864 1%	17733 1%
Tertile 1 (without 2 year lag)	3 0%	55 0%
Tertile 2 (without 2 year lag)	7 0%	19 0%
Tertile 3 (without 2 year lag)	3 0%	46 0%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.2875
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.53
	Confidence Interval	(2-Sided) 95%; 0.17 to 1.70
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.2042
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.47
	Confidence Interval	(2-Sided) 95%; 0.15 to 1.51
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, upper GI disorder, acid-suppressing drugs.

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.4733
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.75
	Confidence Interval	(2-Sided) 95%; 0.38 to 8.06
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.7867
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.24
	Confidence Interval	(2-Sided) 95%; 0.26 to 5.85
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, upper GI disorder, acid-suppressing drugs.

Statistical Analysis 5

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0049
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	3.53
	Confidence Interval	(2-Sided) 95%; 1.47 to 8.50
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 6

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0218
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	2.87
	Confidence Interval	(2-Sided) 95%; 1.17 to 7.08
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, upper GI disorder, acid-suppressing drugs.

Statistical Analysis 7

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.8490
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.89
	Confidence Interval	(2-Sided) 95%; 0.27 to 2.92
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 8

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.6263
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.74
	Confidence Interval	(2-Sided) 95%; 0.22 to 2.46
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, upper GI disorder, acid-suppressing drugs.

Statistical Analysis 9

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.4122
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.61
	Confidence Interval	(2-Sided) 95%; 0.19 to 1.97
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 10

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.2642
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.51
	Confidence Interval	(2-Sided) 95%; 0.16 to 1.66
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, upper GI disorder, acid-suppressing drugs.

8. Primary Outcome

Title	Number of Stomach Cancer Cases and Matched Controls With the Indicated Duration of Exposure to Gabapentin
Description	Incident stomach cancer. Gabapentin (Gaba.) Exposure Description: With 2 year (yr) lag=Gaba. exposure from cohort entry to 2 yr prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 yr lag=Gaba. exposure from cohort entry to index date. Tertiles (T) with 2 yr lag: T 1 (0.01-1.38 months [mo.]), T 2 (1.3-5.56 mo.), and T 3 (5.57-105.82 mo.). Tertiles without 2 yr lag: T 1 (0.01-1.38 m.), T 2 (1.39-5.72 mo.), and T 3 (5.73-123.70 mo.). Inestimable OR and 95% CI when no gaba.-exposed cancer cases or controls at the exposure level.
Time Frame	The case index date was the date of incident stomach cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Stomach Cancer Patients	Matched Controls for Incident Stomach Cancer Patients
Arm/Group Description	Patients with incident stomach cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident stomach cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	1877	17853
Never (with 2 year lag)	1872 1%	17783 1%
Tertile 1 (with 2 year lag)	0 0%	24 0%
Tertile 2 (with 2 year lag)	2 0%	14 0%
Tertile 3 (with 2 year lag)	3 0%	32 0%
Never (without 2 year lag)	1864 1%	17733 1%
Tertile 1 (without 2 year lag)	4 0%	44 0%
Tertile 2 (without 2 year lag)	4 0%	30 0%
Tertile 3 (without 2 year lag)	5 0%	46 0%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.7891
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.87
	Confidence Interval	(2-Sided) 95%; 0.31 to 2.43
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.6216
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.77
	Confidence Interval	(2-Sided) 95%; 0.27 to 2.17
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, upper GI disorder, acid-suppressing drugs.

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.6893
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.35
	Confidence Interval	(2-Sided) 95%; 0.31 to 5.96
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.9273
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.07
	Confidence Interval	(2-Sided) 95%; 0.24 to 4.84
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, upper GI disorder, acid-suppressing drugs.

Statistical Analysis 5

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.5980
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.33
	Confidence Interval	(2-Sided) 95%; 0.47 to 3.78
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 6

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.8572
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.10
	Confidence Interval	(2-Sided) 95%; 0.38 to 3.18
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, upper GI disorder, acid-suppressing drugs.

Statistical Analysis 7

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.7807
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.84
	Confidence Interval	(2-Sided) 95%; 0.26 to 2.77
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 8

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.5579
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.70
	Confidence Interval	(2-Sided) 95%; 0.21 to 2.32
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, upper GI disorder, acid-suppressing drugs.

Statistical Analysis 9

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.9932
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.00
	Confidence Interval	(2-Sided) 95%; 0.40 to 2.53
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 10

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.6901
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.83
	Confidence Interval	(2-Sided) 95%; 0.32 to 2.11
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, upper GI disorder, acid-suppressing drugs.

9. Primary Outcome

Title	Number of Stomach Cancer Cases and Matched Controls With the Indicated Long Duration of Exposure to Gabapentin
Description	Incident stomach cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date.
Time Frame	The case index date was the date of incident stomach cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Stomach Cancer Patients	Matched Controls for Incident Stomach Cancer Patients
Arm/Group Description	Patients with incident stomach cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident stomach cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	1877	17853
Never (with 2 year lag)	1872 1%	17783 1%
> 1 year (with 2 year lag)	2 0%	15 0%
> 2 year (with 2 year lag)	1 0%	8 0%
> 3 year (with 2 year lag)	1 0%	4 0%
Never (without 2 year lag)	1864 1%	17733 1%
> 1 year (without 2 year lag)	3 0%	33 0%
> 2 year (without 2 year lag)	2 0%	19 0%
> 3 year (without 2 year lag)	1 0%	11 0%

10. Primary Outcome

Title	Number of Stomach Cancer Cases and Matched Controls With the Indicated Cumulative Dose of Gabapentin
Description	Incident stomach cancer. Gabapentin (Gaba.) Exposure Description: With 2 year (yr) lag=Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 yr lag=Gaba. exposure from cohort entry to index date. Tertiles (T) with 2 yr lag: T 1 (0.1-30.0 grams [g]), T 2 (30.1-189.0 g), and T 3 (189.1-9600.0 g). Tertiles without 2 yr lag: T 1 (0.1-30.0 g), T 2 (30.1-189.0 g), and T 3 (189.1-11610.0 g). Inestimable OR and 95% CI when no gaba.-exposed cancer cases or controls at the exposure level.
Time Frame	The case index date was the date of incident stomach cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Stomach Cancer Patients	Matched Controls for Incident Stomach Cancer Patients
Arm/Group Description	Patients with incident stomach cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident stomach cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	1877	17853
Never (with 2 year lag)	1872 1%	17783 1%
Tertile 1 (with 2 year lag)	0 0%	22 0%
Tertile 2 (with 2 year lag)	3 0%	21 0%
Tertile 3 (with 2 year lag)	2 0%	27 0%
Never (without 2 year lag)	1864 1%	17733 1%
Tertile 1 (without 2 year lag)	3 0%	46 0%
Tertile 2 (without 2 year lag)	6 0%	31 0%
Tertile 3 (without 2 year lag)	4 0%	43 0%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.4310
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.62
	Confidence Interval	(2-Sided) 95%; 0.19 to 2.01
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.3226
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.55
	Confidence Interval	(2-Sided) 95%; 0.17 to 1.79
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, upper GI disorder, acid-suppressing drugs.

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.6206
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.36
	Confidence Interval	(2-Sided) 95%; 0.40 to 4.59
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.8557
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.12
	Confidence Interval	(2-Sided) 95%; 0.33 to 3.84
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, upper GI disorder, acid-suppressing drugs.

Statistical Analysis 5

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.1498
	Comments
	Method	Unadjusted Odds Ratio
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.91
	Confidence Interval	(2-Sided) 95%; 0.79 to 4.62
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 6

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.2645
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.66
	Confidence Interval	(2-Sided) 95%; 0.68 to 4.07
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, upper GI disorder, acid-suppressing drugs.

Statistical Analysis 7

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.5728
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.66
	Confidence Interval	(2-Sided) 95%; 0.16 to 2.79
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 8

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.3727
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.52
	Confidence Interval	(2-Sided) 95%; 0.12 to 2.21
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, upper GI disorder, acid-suppressing drugs.

Statistical Analysis 9

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.7743
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.86
	Confidence Interval	(2-Sided) 95%; 0.31 to 2.40
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 10

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.4775
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.69
	Confidence Interval	(2-Sided) 95%; 0.24 to 1.94
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, upper GI disorder, acid-suppressing drugs.

11. Primary Outcome

Title	Number of Anal Cancer Cases and Matched Controls With the Indicated Exposure to Gabapentin
Description	Incident Anal cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin prescription from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin prescription from cohort entry to index date.
Time Frame	The case index date was the date of incident anal cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Anal Cancer Patients	Matched Controls for Incident Anal Cancer Patients
Arm/Group Description	Patients with incident anal cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident anal cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	212	2067
Ever (with 2 year lag)	4 0%	6 0%
Never (with 2 year lag)	208 0.1%	2061 0.1%
Ever (without 2 year lag)	5 0%	15 0%
Never (without 2 year lag)	207 0.1%	2052 0.1%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0032
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	7.33
	Confidence Interval	(2-Sided) 95%; 1.95 to 27.56
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Ever (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.1251
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	3.39
	Confidence Interval	(2-Sided) 95%; 0.71 to 16.17
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Ever (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, HIV, HPV (genital).

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0261
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	3.22
	Confidence Interval	(2-Sided) 95%; 1.15 to 9.01
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Ever (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.2090
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	2.16
	Confidence Interval	(2-Sided) 95%; 0.65 to 7.17
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Ever (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, HIV, HPV (genital).

12. Primary Outcome

Title	Number of Anal Cancer Cases and Matched Controls With the Indicated Number of Gabapentin Prescriptions
Description	Incident anal cancer. Gabapentin (Gaba.) Exposure Description: With 2 year (yr) lag=Gaba. exposure from cohort entry to 2 yr prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 yr lag=Gaba. exposure from cohort entry to index date. Tertiles (T) with 2 yr lag: T 1 (1-2 prescriptions [prescrip.]), T 2 (3-7 prescrip.), T 3 (8-298 prescrip.). Tertiles without 2 yr lag: T 1 (1-2 prescrip.), T 2 (3-7 prescrip.), and T 3 (8-388 prescrip.). Inestimable OR and 95% CI when no gaba.-exposed cancer cases or controls at the exposure level.
Time Frame	The case index date was the date of incident anal cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Anal Cancer Patients	Matched Controls for Incident Anal Cancer Patients
Arm/Group Description	Patients with incident anal cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident anal cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	212	2067
Never (with 2 year lag)	208 0.1%	2061 0.1%
Tertile 1 (with 2 year lag)	2 0%	1 0%
Tertile 2 (with 2 year lag)	1 0%	2 0%
Tertile 3 (with 2 year lag)	1 0%	3 0%
Never (without 2 year lag)	207 0.1%	2052 0.1%
Tertile 1 (without 2 year lag)	3 0%	7 0%
Tertile 2 (without 2 year lag)	0 0%	3 0%
Tertile 3 (without 2 year lag)	2 0%	5 0%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0144
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	20.00
	Confidence Interval	(2-Sided) 95%; 1.81 to 220.5
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0158
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	21.47
	Confidence Interval	(2-Sided) 95%; 1.78 to 258.8
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, HIV, HPV (genital).

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0524
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	3.86
	Confidence Interval	(2-Sided) 95%; 0.99 to 15.14
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0269
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	5.21
	Confidence Interval	(2-Sided) 95%; 1.21 to 22.51
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, HIV, HPV (genital).

Statistical Analysis 5

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.1888
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	5.00
	Confidence Interval	(2-Sided) 95%; 0.45 to 55.14
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 6

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.9990
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.00
	Confidence Interval	(2-Sided) 95%; 0.03 to 32.21
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, HIV, HPV (genital).

Statistical Analysis 7

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.2885
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	3.75
	Confidence Interval	(2-Sided) 95%; 0.33 to 43.08
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 8

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.9608
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.93
	Confidence Interval	(2-Sided) 95%; 0.06 to 13.87
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, HIV, HPV (genital).

Statistical Analysis 9

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0944
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	4.33
	Confidence Interval	(2-Sided) 95%; 0.78 to 24.07
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 10

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.9253
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.90
	Confidence Interval	(2-Sided) 95%; 0.10 to 7.82
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, HIV, HPV (genital).

13. Primary Outcome

Title	Number of Anal Cancer Cases and Matched Controls With the Indicated Duration of Exposure to Gabapentin
Description	Incident anal cancer. Gabapentin (Gaba.) Exposure Description: With 2 year (yr) lag=Gaba. exposure from cohort entry to 2 yr prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 yr lag=Gaba. exposure from cohort entry to index date. Tertiles (T) with 2 yr lag: T 1 (0.01-1.38 months [mo.]), T 2 (1.39-5.56 mo.), and T 3 (5.57 -105.82 mo.). Tertiles without 2 yr lag: T 1 (0.01-1.38 mo.), T 2 (1.39-5.72 mo.), and T 3 (5.73-123.70 mo.). Inestimable OR and 95% CI when no gaba.-exposed cancer cases or controls at the exposure level.
Time Frame	The case index date was the date of incident anal cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Anal Cancer Patients	Matched Controls for Incident Anal Cancer Patients
Arm/Group Description	Patients with incident anal cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident anal cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	212	2067
Never (with 2 year lag)	208 0.1%	2061 0.1%
Tertile 1 (with 2 year lag)	2 0%	1 0%
Tertile 2 (with 2 year lag)	1 0%	2 0%
Tertile 3 (with 2 year lag)	1 0%	3 0%
Never (without 2 year lag)	207 0.1%	2052 0.1%
Tertile 1 (without 2 year lag)	3 0%	5 0%
Tertile 2 (without 2 year lag)	0 0%	5 0%
Tertile 3 (without 2 year lag)	2 0%	5 0%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0144
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	20.00
	Confidence Interval	(2-Sided) 95%; 1.81 to 220.5
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0158
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	21.47
	Confidence Interval	(2-Sided) 95%; 1.78 to 258.8
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, HIV, HPV (genital).

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0141
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	6.00
	Confidence Interval	(2-Sided) 95%; 1.43 to 25.11
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0063
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	8.56
	Confidence Interval	(2-Sided) 95%; 1.83 to 40.02
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, HIV, HPV (genital).

Statistical Analysis 5

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.1888
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	5.00
	Confidence Interval	(2-Sided) 95%; 0.45 to 55.14
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 6

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.9990
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.00
	Confidence Interval	(2-Sided) 95%; 0.03 to 32.21
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, HIV, HPV (genital).

Statistical Analysis 7

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.2885
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	3.75
	Confidence Interval	(2-Sided) 95%; 0.33 to 43.08
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 8

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.9608
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.93
	Confidence Interval	(2-Sided) 95%; 0.06 to 13.87
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, HIV, HPV (genital).

Statistical Analysis 9

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0944
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	4.33
	Confidence Interval	(2-Sided) 95%; 0.78 to 24.07
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 10

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.9381
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.92
	Confidence Interval	() 95%; 0.11 to 7.97
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, HIV, HPV (genital).

14. Primary Outcome

Title	Number of Anal Cancer Cases and Matched Controls With the Indicated Long Duration of Exposure to Gabapentin
Description	Incident anal cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date.
Time Frame	The case index date was the date of incident anal cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Anal Cancer Patients	Matched Controls for Incident Anal Cancer Patients
Arm/Group Description	Patients with incident anal cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident anal cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	212	2067
Never (with 2 year lag)	208 0.1%	2061 0.1%
> 1 year (with 2 year lag)	0 0%	1 0%
> 2 year (with 2 year lag)	0 0%	1 0%
> 3 year (with 2 year lag)	0 0%	1 0%
Never (without 2 year lag)	207 0.1%	2052 0.1%
> 1 year (without 2 year lag)	2 0%	3 0%
> 2 year (without 2 year lag)	0 0%	1 0%
> 3 year (without 2 year lag)	0 0%	1 0%

15. Primary Outcome

Title	Number of Anal Cancer Cases and Matched Controls With the Indicated Cumulative Dose of Gabapentin
Description	Incident anal cancer. Gabapentin (Gaba.) Exposure Description: With 2 yr lag=Gaba. exposure from cohort entry to 2 yr prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 yr lag=Gaba. exposure from cohort entry to index date. Tertiles (T) with 2 yr lag: T 1 (0.1-30.0 grams [g]), T 2 (30.1-189.0 g), and T 3 (189.1-9600.0 g). Tertiles without 2 yr lag: T 1 (0.1-30.0 g), T 2 (30.1-189.0 g), and T 3 (189.1-11610.0 g). Inestimable OR and 95% CI when no gaba.-exposed cancer cases or controls at the exposure level.
Time Frame	The case index date was the date of incident anal cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Anal Cancer Patients	Matched Controls for Incident Anal Cancer Patients
Arm/Group Description	Patients with incident anal cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident anal cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	212	2067
Never (with 2 year lag)	208 0.1%	2061 0.1%
Tertile 1 (with 2 year lag)	2 0%	1 0%
Tertile 2 (with 2 year lag)	1 0%	3 0%
Tertile 3 (with 2 year lag)	1 0%	2 0%
Never (without 2 year lag)	207 0.1%	2052 0.1%
Tertile 1 (without 2 year lag)	3 0%	7 0%
Tertile 2 (without 2 year lag)	0 0%	3 0%
Tertile 3 (without 2 year lag)	2 0%	5 0%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0144
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	20.00
	Confidence Interval	(2-Sided) 95%; 1.81 to 220.5
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0157
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	21.55
	Confidence Interval	(2-Sided) 95%; 1.78 to 260.2
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, HIV, HPV (genital).

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0524
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	3.86
	Confidence Interval	(2-Sided) 95%; 0.99 to 15.14
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0333
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	4.80
	Confidence Interval	(2-Sided) 95%; 1.13 to 20.31
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, HIV, HPV (genital).

Statistical Analysis 5

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.2966
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	3.34
	Confidence Interval	(2-Sided) 95%; 0.35 to 32.06
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 6

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.7510
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.62
	Confidence Interval	(2-Sided) 95%; 0.03 to 12.31
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, HIV, HPV (genital).

Statistical Analysis 7

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.1868
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	6.71
	Confidence Interval	(2-Sided) 95%; 0.40 to 113.4
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 8

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.7558
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.66
	Confidence Interval	(2-Sided) 95%; 0.07 to 40.61
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, HIV, HPV (genital).

Statistical Analysis 9

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0944
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	4.33
	Confidence Interval	(2-Sided) 95%; 0.78 to 24.07
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 10

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.9244
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.90
	Confidence Interval	(2-Sided) 95%; 0.10 to 7.81
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, HIV, HPV (genital).

16. Primary Outcome

Title	Number of Lung Cancer Cases and Matched Controls With the Indicated Exposure to Gabapentin
Description	Incident lung cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin prescription from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin prescription from cohort entry to index date.
Time Frame	The case index date was the date of incident lung cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Lung Cancer Patients	Matched Controls for Incident Lung Cancer Patients
Arm/Group Description	Patients with incident lung cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident lung cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	10855	102836
Ever (with 2 year lag)	55 0%	402 0%
Never (with 2 year lag)	10800 6%	102434 6%
Ever (without 2 year lag)	155 0.1%	783 0%
Never (without 2 year lag)	10700 6%	102053 6%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0963
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.27
	Confidence Interval	(2-Sided) 95%; 0.96 to 1.70
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Ever (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.6710
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.94
	Confidence Interval	(2-Sided) 95%; 0.69 to 1.27
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Ever (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, COPD.

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	<0.0001
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.84
	Confidence Interval	(2-Sided) 95%; 1.54 to 2.20
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Ever (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	<0.0001
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.50
	Confidence Interval	(2-Sided) 95%; 1.24 to 1.81
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Ever (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, COPD.

17. Primary Outcome

Title	Number of Lung Cancer Cases and Matched Controls With the Indicated Number of Gabapentin Prescriptions
Description	Incident lung cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date. Tertile's with 2 year lag: Tertile 1 (1-2 prescriptions), Tertile 2 (3-7 prescriptions), Tertile 3 (8-298 prescriptions). Tertiles without 2 year lag: Tertile 1 (1-2 prescriptions), Tertile 2 (3-7 prescriptions), and Tertile 3 (8-388 prescriptions).
Time Frame	The case index date was the date of incident lung cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Lung Cancer Patients	Matched Controls for Incident Lung Cancer Patients
Arm/Group Description	Patients with incident lung cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident lung cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	10855	102836
Never (with 2 year lag)	10800 6%	102434 6%
Tertile 1 (with 2 year lag)	24 0%	160 0%
Tertile 2 (with 2 year lag)	13 0%	105 0%
Tertile 3 (with 2 year lag)	18 0%	137 0%
Never (without 2 year lag)	10700 6%	102053 6%
Tertile 1 (without 2 year lag)	81 0%	327 0%
Tertile 2 (without 2 year lag)	32 0%	186 0%
Tertile 3 (without 2 year lag)	42 0%	270 0%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.1184
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.41
	Confidence Interval	(2-Sided) 95%; 0.92 to 2.18
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.8016
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.94
	Confidence Interval	(2-Sided) 95%; 0.59 to 1.50
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, COPD.

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	<0.0001
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	2.27
	Confidence Interval	(2-Sided) 95%; 1.77 to 2.91
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	<0.0001
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.97
	Confidence Interval	(2-Sided) 95%; 1.51 to 2.58
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, COPD.

Statistical Analysis 5

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.6105
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.16
	Confidence Interval	(2-Sided) 95%; 0.65 to 2.08
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 6

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.6797
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.14
	Confidence Interval	(2-Sided) 95%; 0.62 to 2.09
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, COPD.

Statistical Analysis 7

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0103
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.64
	Confidence Interval	(2-Sided) 95%; 1.12 to 2.39
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 8

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.1629
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.34
	Confidence Interval	(2-Sided) 95%; 0.89 to 2.00
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, COPD.

Statistical Analysis 9

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.4713
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.20
	Confidence Interval	(2-Sided) 95%; 0.73 to 1.97
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 10

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.4505
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.82
	Confidence Interval	(2-Sided) 95%; 0.49 to 1.38
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, COPD.

Statistical Analysis 11

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0284
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.45
	Confidence Interval	(2-Sided) 95%; 1.04 to 2.01
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 12

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.6262
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.09
	Confidence Interval	(2-Sided) 95%; 0.77 to 1.54
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, COPD.

18. Primary Outcome

Title	Number of Lung Cancer Cases and Matched Controls With the Indicated Duration of Exposure to Gabapentin
Description	Incident lung cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date. Tertile's with 2 year lag: Tertile 1 (0.01 - 1.38 months), Tertile 2 (1.39 - 5.56 months), and Tertile 3 (5.57 - 105.82 months). Tertile's without 2 year lag: Tertile 1 (0.01 - 1.38 months), Tertile 2 (1.39 - 5.72 months), and Tertile 3 (5.73 - 123.70 months).
Time Frame	The case index date was the date of incident lung cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Lung Cancer Patients	Matched Controls for Incident Lung Cancer Patients
Arm/Group Description	Patients with incident lung cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident lung cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	10855	102836
Never (with 2 year lag)	10800 6%	102434 6%
Tertile 1 (with 2 year lag)	21 0%	131 0%
Tertile 2 (with 2 year lag)	15 0%	130 0%
Tertile 3 (with 2 year lag)	19 0%	141 0%
Never (without 2 year lag)	10700 6%	102053 6%
Tertile 1 (without 2 year lag)	67 0%	267 0%
Tertile 2 (without 2 year lag)	48 0%	247 0%
Tertile 3 (without 2 year lag)	40 0%	269 0%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0798
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.52
	Confidence Interval	(2-Sided) 95%; 0.95 to 2.41
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.8357
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.05
	Confidence Interval	(2-Sided) 95%; 0.64 to 1.73
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, COPD.

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	<0.0001
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	2.29
	Confidence Interval	(2-Sided) 95%; 1.75 to 3.01
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	<0.0001
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	2.05
	Confidence Interval	(2-Sided) 95%; 1.53 to 2.76
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, COPD.

Statistical Analysis 5

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.7794
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.08
	Confidence Interval	(2-Sided) 95%; 0.63 to 1.85
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 6

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.8075
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.93
	Confidence Interval	(2-Sided) 95%; 0.52 to 1.66
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, COPD.

Statistical Analysis 7

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0001
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.84
	Confidence Interval	(2-Sided) 95%; 1.35 to 2.52
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 8

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0161
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.51
	Confidence Interval	(2-Sided) 95%; 1.08 to 2.12
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, COPD.

Statistical Analysis 9

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.3972
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.23
	Confidence Interval	(2-Sided) 95%; 0.76 to 2.00
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 10

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.4950
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.84
	Confidence Interval	(2-Sided) 95%; 0.50 to 1.39
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, COPD.

Statistical Analysis 11

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0601
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.38
	Confidence Interval	(2-Sided) 95%; 0.99 to 1.94
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 12

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.9286
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.02
	Confidence Interval	(2-Sided) 95%; 0.71 to 1.45
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, COPD.

19. Primary Outcome

Title	Number of Lung Cancer Cases and Matched Controls With the Indicated Long Duration of Exposure to Gabapentin
Description	Incident lung cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date.
Time Frame	The case index date was the date of incident lung cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Lung Cancer Patients	Matched Controls for Incident Lung Cancer Patients
Arm/Group Description	Patients with incident lung cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident lung cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	10855	102836
Never (with 2 year lag)	10800 6%	102434 6%
> 1 year (with 2 year lag)	10 0%	75 0%
> 2 year (with 2 year lag)	4 0%	24 0%
> 3 year (with 2 year lag)	1 0%	6 0%
Never (without 2 year lag)	10700 6%	102053 6%
> 1 year (without 2 year lag)	21 0%	167 0%
> 2 year (without 2 year lag)	11 0%	77 0%
> 3 year (without 2 year lag)	4 0%	33 0%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.5394
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.23
	Confidence Interval	(2-Sided) 95%; 0.63 to 2.39
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "> 1 year (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.5070
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.79
	Confidence Interval	(2-Sided) 95%; 0.39 to 1.58
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "> 1 year (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, COPD.

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.5426
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.15
	Confidence Interval	(2-Sided) 95%; 0.73 to 1.83
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "> 1 year (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.5150
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.85
	Confidence Interval	(2-Sided) 95%; 0.52 to 1.38
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "> 1 year (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, COPD.

Statistical Analysis 5

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.4350
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.53
	Confidence Interval	(2-Sided) 95%; 0.53 to 4.42
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "> 2 years (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 6

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.8276
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.13
	Confidence Interval	(2-Sided) 95%; 0.37 to 3.43
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "> 2 years (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, COPD.

Statistical Analysis 7

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.4022
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.32
	Confidence Interval	(2-Sided) 95%; 0.69 to 2.50
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "> 2 years (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 8

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.8681
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.94
	Confidence Interval	(2-Sided) 95%; 0.48 to 1.86
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "> 2 years (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, COPD.

Statistical Analysis 9

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.6326
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.67
	Confidence Interval	(2-Sided) 95%; 0.20 to 13.86
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "> 3 years (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 10

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.5770
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.87
	Confidence Interval	(2-Sided) 95%; 0.21 to 16.86
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "> 3 years (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, COPD.

Statistical Analysis 11

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.8087
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.14
	Confidence Interval	(2-Sided) 95%; 0.40 to 3.21
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "> 3 years (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 12

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.8497
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.90
	Confidence Interval	(2-Sided) 95%; 0.31 to 2.64
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "> 3 years (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, COPD.

20. Primary Outcome

Title	Number of Lung Cancer Cases and Matched Controls With the Indicated Cumulative Dose of Gabapentin
Description	Incident lung cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date. Tertile's with 2 year lag: Tertile 1 (0.1 - 30.0 grams), Tertile 2 (30.1 - 189.0 grams), and Tertile 3 (189.1 - 9600.0 grams). Tertile's without 2 year lag: Tertile 1 (0.1 - 30.0 grams), Tertile 2 (30.1 - 189.0 grams), and Tertile 3 (189.1 - 11610.0 grams).
Time Frame	The case index date was the date of incident lung cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Lung Cancer Patients	Matched Controls for Incident Lung Cancer Patients
Arm/Group Description	Patients with incident lung cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident lung cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice
Measure Participants	10855	102836
Never (with 2 year lag)	10800 6%	102434 6%
Tertile 1 (with 2 year lag)	21 0%	139 0%
Tertile 2 (with 2 year lag)	16 0%	120 0%
Tertile 3 (with 2 year lag)	18 0%	143 0%
Never (without 2 year lag)	10700 6%	102053 6%
Tertile 1 (without 2 year lag)	62 0%	288 0%
Tertile 2 (without 2 year lag)	56 0%	224 0%
Tertile 3 (without 2 year lag)	37 0%	271 0%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.1528
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.40
	Confidence Interval	(2-Sided) 95%; 0.88 to 2.24
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.9625
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.01
	Confidence Interval	(2-Sided) 95%; 0.62 to 1.65
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, COPD.

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	<0.0001
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.96
	Confidence Interval	(2-Sided) 95%; 1.48 to 2.59
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0006
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.70
	Confidence Interval	(2-Sided) 95%; 1.25 to 2.29
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, COPD.

Statistical Analysis 5

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.3515
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.28
	Confidence Interval	(2-Sided) 95%; 0.76 to 2.16
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 6

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.9069
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.03
	Confidence Interval	(2-Sided) 95%; 0.59 to 1.82
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, COPD.

Statistical Analysis 7

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	<0.0001
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	2.39
	Confidence Interval	(2-Sided) 95%; 1.78 to 3.21
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 8

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	<0.0001
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	2.04
	Confidence Interval	(2-Sided) 95%; 1.48 to 2.81
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, COPD.

Statistical Analysis 9

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.5948
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.14
	Confidence Interval	(2-Sided) 95%; 0.70 to 1.88
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 10

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.3984
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.80
	Confidence Interval	(2-Sided) 95%; 0.47 to 1.35
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, COPD.

Statistical Analysis 11

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.1872
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.26
	Confidence Interval	(2-Sided) 95%; 0.89 to 1.79
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 12

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.6772
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.93
	Confidence Interval	(2-Sided) 95%; 0.64 to 1.34
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, COPD.

21. Primary Outcome

Title	Number of Bone/Joint Cancer Cases and Matched Controls With the Indicated Exposure to Gabapentin
Description	Incident bone/joint cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin prescription from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin prescription from cohort entry to index date. Inestimable OR and 95% CI when no gabapentin-exposed cancer cases or no gabapentin-exposed controls at the exposure level.
Time Frame	The case index date was the date of incident bone/joint cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as for the case.

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Bone/Joint Cancer Patients	Matched Controls for Incident Bone/Joint Cancer Patients
Arm/Group Description	Patients with incident bone/joint cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident bone/joint cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	300	2935
Ever (with 2 year lag)	0 0%	10 0%
Never (with 2 year lag)	300 0.2%	2925 0.2%
Ever (without 2 year lag)	1 0%	19 0%
Never (without 2 year lag)	299 0.2%	2916 0.2%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.4933
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.49
	Confidence Interval	(2-Sided) 95%; 0.07 to 3.72
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Ever (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.4147
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.43
	Confidence Interval	(2-Sided) 95%; 0.06 to 3.27
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Ever (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy.

22. Primary Outcome

Title	Number of Bone/Joint Cancer Cases and Matched Controls With the Indicated Number of Gabapentin Prescriptions
Description	Incident bone/joint cancer. Gabapentin (Gaba.) Exposure Description: With 2 year (yr) lag=Gaba. exposure from cohort entry to 2 yr prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 yr lag=Gaba. exposure from cohort entry to index date. Tertiles (T) with 2 yr lag: T 1 (1-2 prescriptions [prescrip.]), T 2 (3-7 prescrip.), T 3 (8-298 prescrip.). Tertiles without 2 yr lag: T 1 (1-2 prescrip.), T 2 (3-7 prescrip.), and T 3 (8-388 prescrip.). Inestimable OR and 95% CI when no gaba.-exposed cases or controls at the exposure level.
Time Frame	The case index date was the date of incident bone/joint cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as for the case.

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Bone/Joint Cancer Patients	Matched Controls for Incident Bone/Joint Cancer Patients
Arm/Group Description	Patients with incident bone/joint cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident bone/joint cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	300	2935
Never (with 2 year lag)	300 0.2%	2925 0.2%
Tertile 1 (with 2 year lag)	0 0%	4 0%
Tertile 2 (with 2 year lag)	0 0%	2 0%
Tertile 3 (with 2 year lag)	0 0%	4 0%
Never (without 2 year lag)	299 0.2%	2916 0.2%
Tertile 1 (without 2 year lag)	1 0%	7 0%
Tertile 2 (without 2 year lag)	0 0%	4 0%
Tertile 3 (without 2 year lag)	0 0%	8 0%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.8257
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.27
	Confidence Interval	(2-Sided) 95%; 0.15 to 10.38
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.9467
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.08
	Confidence Interval	(2-Sided) 95%; 0.13 to 9.02
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy.

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.9859
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.00
	Confidence Interval	(2-Sided) 95%; 0.00 to 0.00
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.9857
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.00
	Confidence Interval	(2-Sided) 95%; 0.00 to 0.00
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy.

Statistical Analysis 5

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.9860
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.00
	Confidence Interval	(2-Sided) 95%; 0.00 to 0.00
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 6

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.9858
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.00
	Confidence Interval	(2-Sided) 95%; 0.00 to 0.00
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy.

Statistical Analysis 7

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.9798
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.00
	Confidence Interval	(2-Sided) 95%; 0.00 to 0.00
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 8

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.9796
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.00
	Confidence Interval	(2-Sided) 95%; 0.00 to 0.00
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy.

23. Primary Outcome

Title	Number of Bone/Joint Cancer Cases and Matched Controls With the Indicated Duration of Exposure to Gabapentin
Description	Incident bone/joint cancer. Gabapentin (Gaba.) Exposure Description: With 2 year (yr) lag=Gaba. exposure from cohort entry to 2 yr prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 yr lag=Gaba. exposure from cohort entry to index date. Tertiles (T) with 2 yr lag: T 1 (0.01-1.38 months [mo.]), T 2 (1.39-5.56 mo.), and T 3 (5.57-105.82 mo.). Tertiles without 2 yr lag: T 1 (0.01-1.38 mo.), T 2 (1.39-5.72 mo.), and T 3 (5.73-123.70 mo.). Inestimable OR and 95% CI when no gaba.-exposed cases or controls at the exposure level.
Time Frame	The case index date was the date of incident bone/joint cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as for the case.

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Bone/Joint Cancer Patients	Matched Controls for Incident Bone/Joint Cancer Patients
Arm/Group Description	Patients with incident bone/joint cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident bone/joint cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	300	2935
Never (with 2 year lag)	300 0.2%	2925 0.2%
Tertile 1 (with 2 year lag)	0 0%	4 0%
Tertile 2 (with 2 year lag)	0 0%	1 0%
Tertile 3 (with 2 year lag)	0 0%	5 0%
Never (without 2 year lag)	299 0.2%	2916 0.2%
Tertile 1 (without 2 year lag)	1 0%	5 0%
Tertile 2 (without 2 year lag)	0 0%	5 0%
Tertile 3 (without 2 year lag)	0 0%	9 0%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.6252
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.71
	Confidence Interval	(2-Sided) 95%; 0.20 to 14.84
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.7208
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.49
	Confidence Interval	(2-Sided) 95%; 0.17 to 13.28
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy.

24. Primary Outcome

Title	Number of Bone/Joint Cancer Cases and Matched Controls With the Indicated Long Duration of Exposure to Gabapentin
Description	Incident bone/joint cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date.
Time Frame	The case index date was the date of incident bone/joint cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as for the case.

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Bone/Joint Cancer Patients	Matched Controls for Incident Bone/Joint Cancer Patients
Arm/Group Description	Patients with incident bone/joint cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident bone/joint cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	300	2935
Never (with 2 year lag)	300 0.2%	2925 0.2%
> 1 year (with 2 year lag)	0 0%	3 0%
> 2 year (with 2 year lag)	0 0%	0 0%
> 3 year (with 2 year lag)	0 0%	0 0%
Never (without 2 year lag)	299 0.2%	2916 0.2%
> 1 year (without 2 year lag)	0 0%	5 0%
> 2 year (without 2 year lag)	0 0%	1 0%
> 3 year (without 2 year lag)	0 0%	0 0%

25. Primary Outcome

Title	Number of Bone/Joint Cancer Cases and Matched Controls With the Indicated Cumulative Dose of Gabapentin
Description	Incident bone/joint cancer. Gabapentin (Gaba.) Exposure Description: With 2 year (yr) lag=Gaba. exposure from cohort entry to 2 yr prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 yr lag=Gaba. exposure from cohort entry to index date. Tertiles (T) with 2 yr lag: T 1 (0.1-30.0 grams [g]), T 2 (30.1-189.0 g), and T 3 (189.1-9600.0 g). Tertiles without 2 yr lag: T 1 (0.1-30.0 g), T 2 (30.1-189.0 g), and T 3 (189.1-11610.0 g). Inestimable OR and 95% CI when no gaba.-exposed cancer cases or controls at the exposure level.
Time Frame	The case index date was the date of incident bone/joint cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as for the case.

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Bone/Joint Cancer Patients	Matched Controls for Incident Bone/Joint Cancer Patients
Arm/Group Description	Patients with incident bone/joint cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident bone/joint cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	300	2935
Never (with 2 year lag)	300 0.2%	2925 0.2%
Tertile 1 (with 2 year lag)	0 0%	4 0%
Tertile 2 (with 2 year lag)	0 0%	6 0%
Tertile 3 (with 2 year lag)	0 0%	0 0%
Never (without 2 year lag)	299 0.2%	2916 0.2%
Tertile 1 (without 2 year lag)	1 0%	8 0%
Tertile 2 (without 2 year lag)	0 0%	3 0%
Tertile 3 (without 2 year lag)	0 0%	8 0%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.9141
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.12
	Confidence Interval	(2-Sided) 95%; 0.14 to 9.03
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.9613
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.95
	Confidence Interval	(2-Sided) 95%; 0.12 to 7.78
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy.

26. Primary Outcome

Title	Number of Breast Cancer Cases and Matched Controls With the Indicated Exposure to Gabapentin
Description	Incident breast cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin prescription from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin prescription from cohort entry to index date.
Time Frame	The case index date was the date of incident breast cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Breast Cancer Patients	Matched Controls for Incident Breast Cancer Patients
Arm/Group Description	Patients with incident breast cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident breast cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	19564	188924
Ever (with 2 year lag)	96 0.1%	725 0%
Never (with 2 year lag)	19468 10.9%	188199 11%
Ever (without 2 year lag)	174 0.1%	1392 0.1%
Never (without 2 year lag)	19390 10.8%	187532 11%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0323
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.27
	Confidence Interval	(2-Sided) 95%; 1.02 to 1.57
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Ever (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0682
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.22
	Confidence Interval	(2-Sided) 95%; 0.99 to 1.52
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Ever (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, hysterectomy, current estrogen, prior estrogen, benign breast disease.

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0326
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.19
	Confidence Interval	(2-Sided) 95%; 1.01 to 1.40
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Ever (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0699
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.16
	Confidence Interval	(2-Sided) 95%; 0.99 to 1.36
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Ever (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, hysterectomy, current estrogen, prior estrogen, benign breast disease.

27. Primary Outcome

Title	Number of Breast Cancer Cases and Matched Controls With the Indicated Number of Gabapentin Prescriptions
Description	Incident breast cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date. Tertile's with 2 year lag: Tertile 1 (1-2 prescriptions), Tertile 2 (3-7 prescriptions), Tertile 3 (8-298 prescriptions). Tertiles without 2 year lag: Tertile 1 (1-2 prescriptions), Tertile 2 (3-7 prescriptions), and Tertile 3 (8-388 prescriptions).
Time Frame	The case index date was the date of incident breast cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Cases	Controls
Arm/Group Description	Cases	Controls
Measure Participants	19564	188924
Never (with 2 year lag)	19468 10.9%	188199 11%
Tertile 1 (with 2 year lag)	38 0%	306 0%
Tertile 2 (with 2 year lag)	24 0%	182 0%
Tertile 3 (with 2 year lag)	34 0%	237 0%
Never (without 2 year lag)	19390 10.8%	187532 11%
Tertile 1 (without 2 year lag)	75 0%	577 0%
Tertile 2 (without 2 year lag)	32 0%	355 0%
Tertile 3 (without 2 year lag)	67 0%	460 0%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.2893
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.20
	Confidence Interval	(2-Sided) 95%; 0.86 to 1.69
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.4301
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.15
	Confidence Interval	(2-Sided) 95%; 0.82 to 1.61
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, hysterectomy, current estrogen, prior estrogen, benign breast disease.

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0806
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.24
	Confidence Interval	(2-Sided) 95%; 0.97 to 1.58
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.1226
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.21
	Confidence Interval	(2-Sided) 95%; 0.95 to 1.55
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, hysterectomy, current estrogen, prior estrogen, benign breast disease.

Statistical Analysis 5

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.3502
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.23
	Confidence Interval	(2-Sided) 95%; 0.80 to 1.89
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 6

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.4654
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.17
	Confidence Interval	(2-Sided) 95%; 0.76 to 1.81
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, hysterectomy, current estrogen, prior estrogen, benign breast disease.

Statistical Analysis 7

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.3987
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.86
	Confidence Interval	(2-Sided) 95%; 0.59 to 1.23
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 8

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.3349
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.84
	Confidence Interval	(2-Sided) 95%; 0.58 to 1.20
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, hysterectomy, current estrogen, prior estrogen, benign breast disease.

Statistical Analysis 9

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0827
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.38
	Confidence Interval	(2-Sided) 95%; 0.96 to 1.98
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 10

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0932
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.36
	Confidence Interval	(2-Sided) 95%; 0.95 to 1.96
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, hysterectomy, current estrogen, prior estrogen, benign breast disease.

Statistical Analysis 11

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0130
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.39
	Confidence Interval	(2-Sided) 95%; 1.07 to 1.80
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 12

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0241
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.35
	Confidence Interval	(2-Sided) 95%; 1.04 to 1.75
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, hysterectomy, current estrogen, prior estrogen, benign breast disease.

28. Primary Outcome

Title	Number of Breast Cancer Cases and Matched Controls With the Indicated Duration of Exposure to Gabapentin
Description	Incident breast cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date. Tertile's with 2 year lag: Tertile 1 (0.01 - 1.38 months), Tertile 2 (1.39 - 5.56 months), and Tertile 3 (5.57 - 105.82 months). Tertile's without 2 year lag: Tertile 1 (0.01 - 1.38 months), Tertile 2 (1.39 - 5.72 months), and Tertile 3 (5.73 - 123.70 months).
Time Frame	The case index date was the date of incident breast cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Breast Cancer Patients	Matched Controls for Incident Breast Cancer Patients
Arm/Group Description	Patients with incident breast cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident breast cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	19564	188924
Never (with 2 year lag)	19468 10.9%	188199 11%
Tertile 1 (with 2 year lag)	29 0%	245 0%
Tertile 2 (with 2 year lag)	32 0%	247 0%
Tertile 3 (with 2 year lag)	35 0%	233 0%
Never (without 2 year lag)	19390 10.8%	187532 11%
Tertile 1 (without 2 year lag)	58 0%	465 0%
Tertile 2 (without 2 year lag)	52 0%	473 0%
Tertile 3 (without 2 year lag)	64 0%	454 0%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.5043
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.14
	Confidence Interval	(2-Sided) 95%; 0.77 to 1.68
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.7129
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.08
	Confidence Interval	(2-Sided) 95%; 0.73 to 1.59
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, hysterectomy, current estrogen, prior estrogen, benign breast disease.

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.2055
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.19
	Confidence Interval	(2-Sided) 95%; 0.91 to 1.57
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.3005
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.16
	Confidence Interval	(2-Sided) 95%; 0.88 to 1.52
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, hysterectomy, current estrogen, prior estrogen, benign breast disease.

Statistical Analysis 5

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.2849
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.22
	Confidence Interval	(2-Sided) 95%; 0.84 to 1.77
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 6

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.3650
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.19
	Confidence Interval	(2-Sided) 95%; 0.82 to 1.72
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, hysterectomy, current estrogen, prior estrogen, benign breast disease.

Statistical Analysis 7

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.7559
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.05
	Confidence Interval	(2-Sided) 95%; 0.78 to 1.40
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 8

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.8520
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.03
	Confidence Interval	(2-Sided) 95%; 0.77 to 1.37
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, hysterectomy, current estrogen, prior estrogen, benign breast disease.

Statistical Analysis 9

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0459
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.44
	Confidence Interval	(2-Sided) 95%; 1.01 to 2.06
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 10

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0542
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.42
	Confidence Interval	(2-Sided) 95%; 0.99 to 2.04
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, hysterectomy, current estrogen, prior estrogen, benign breast disease.

Statistical Analysis 11

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0312
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.34
	Confidence Interval	(2-Sided) 95%; 1.03 to 1.74
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 12

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0508
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.30
	Confidence Interval	(2-Sided) 95%; 1.00 to 1.70
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, hysterectomy, current estrogen, prior estrogen, benign breast disease.

29. Primary Outcome

Title	Number of Breast Cancer Cases and Matched Controls With the Indicated Long Duration of Exposure to Gabapentin
Description	Incident breast cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date.
Time Frame	The case index date was the date of incident breast cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Breast Cancer Patients	Matched Controls for Incident Breast Cancer Patients
Arm/Group Description	Patients with incident breast cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident breast cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	19564	188924
Never (with 2 year lag)	19468 10.9%	188199 11%
> 1 year (with 2 year lag)	20 0%	131 0%
> 2 year (with 2 year lag)	9 0%	54 0%
> 3 year (with 2 year lag)	2 0%	27 0%
Never (without 2 year lag)	19390 10.8%	187532 11%
> 1 year (without 2 year lag)	46 0%	291 0%
> 2 year (without 2 year lag)	24 0%	133 0%
> 3 year (without 2 year lag)	10 0%	54 0%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0892
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.51
	Confidence Interval	(2-Sided) 95%; 0.94 to 2.42
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "> 1 year (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.1043
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.48
	Confidence Interval	(2-Sided) 95%; 0.92 to 2.39
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "> 1 year (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, hysterectomy, current estrogen, prior estrogen, benign breast disease.

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0101
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.51
	Confidence Interval	(2-Sided) 95%; 1.10 to 2.06
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "> 1 year (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0186
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.46
	Confidence Interval	(2-Sided) 95%; 1.07 to 2.00
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "> 1 year (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, hysterectomy, current estrogen, prior estrogen, benign breast disease.

Statistical Analysis 5

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.1458
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.69
	Confidence Interval	(2-Sided) 95%; 0.83 to 3.43
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "> 2 years (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 6

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.1859
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.62
	Confidence Interval	(2-Sided) 95%; 0.79 to 3.29
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "> 2 years (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, hysterectomy, current estrogen, prior estrogen, benign breast disease.

Statistical Analysis 7

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0122
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.75
	Confidence Interval	(2-Sided) 95%; 1.13 to 2.72
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "> 2 years (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 8

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0162
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.72
	Confidence Interval	(2-Sided) 95%; 1.10 to 2.67
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "> 2 years (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, hysterectomy, current estrogen, prior estrogen, benign breast disease.

Statistical Analysis 9

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.7573
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.80
	Confidence Interval	(2-Sided) 95%; 0.19 to 3.36
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "> 3 years (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 10

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.7088
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.76
	Confidence Interval	(2-Sided) 95%; 0.18 to 3.21
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "> 3 years (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, hysterectomy, current estrogen, prior estrogen, benign breast disease.

Statistical Analysis 11

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0787
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.84
	Confidence Interval	(2-Sided) 95%; 0.93 to 3.64
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "> 3 years (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 12

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0873
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.81
	Confidence Interval	(2-Sided) 95%; 0.92 to 3.58
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "> 3 years (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, hysterectomy, current estrogen, prior estrogen, benign breast disease.

30. Primary Outcome

Title	Number of Breast Cancer Cases and Matched Controls With the Indicated Cumulative Dose of Gabapentin
Description	Incident breast cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date. Tertile's with 2 year lag: Tertile 1 (0.1 - 30.0 grams), Tertile 2 (30.1 - 189.0 grams), and Tertile 3 (189.1 - 9600.0 grams). Tertile's without 2 year lag: Tertile 1 (0.1 - 30.0 grams), Tertile 2 (30.1 - 189.0 grams), and Tertile 3 (189.1 - 11610.0 grams).
Time Frame	The case index date was the date of incident breast cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Breast Cancer Patients	Matched Controls for Incident Breast Cancer Patients
Arm/Group Description	Patients with incident breast cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident breast cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	19564	188924
Never (with 2 year lag)	19468 10.9%	188199 11%
Tertile 1 (with 2 year lag)	31 0%	254 0%
Tertile 2 (with 2 year lag)	30 0%	225 0%
Tertile 3 (with 2 year lag)	35 0%	246 0%
Never (without 2 year lag)	19390 10.8%	187532 11%
Tertile 1 (without 2 year lag)	66 0%	493 0%
Tertile 2 (without 2 year lag)	44 0%	423 0%
Tertile 3 (without 2 year lag)	64 0%	476 0%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.4233
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.17
	Confidence Interval	(2-Sided) 95%; 0.80 to 1.70
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.5908
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.11
	Confidence Interval	(2-Sided) 95%; 0.76 to 1.62
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, hysterectomy, current estrogen, prior estrogen, benign breast disease.

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0540
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.29
	Confidence Interval	(2-Sided) 95%; 1.00 to 1.67
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0939
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.25
	Confidence Interval	(2-Sided) 95%; 0.96 to 1.62
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, hysterectomy, current estrogen, prior estrogen, benign breast disease.

Statistical Analysis 5

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.2140
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.27
	Confidence Interval	(2-Sided) 95%; 0.87 to 1.87
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 6

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.2940
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.23
	Confidence Interval	(2-Sided) 95%; 0.84 to 1.80
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, hysterectomy, current estrogen, prior estrogen, benign breast disease.

Statistical Analysis 7

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.8352
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.97
	Confidence Interval	(2-Sided) 95%; 0.71 to 1.32
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 8

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.7809
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.96
	Confidence Interval	(2-Sided) 95%; 0.70 to 1.31
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, hysterectomy, current estrogen, prior estrogen, benign breast disease.

Statistical Analysis 9

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0921
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.36
	Confidence Interval	(2-Sided) 95%; 0.95 to 1.94
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 10

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.1091
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.34
	Confidence Interval	(2-Sided) 95%; 0.94 to 1.92
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, hysterectomy, current estrogen, prior estrogen, benign breast disease.

Statistical Analysis 11

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0564
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.29
	Confidence Interval	(2-Sided) 95%; 0.99 to 1.68
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 12

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0939
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.25
	Confidence Interval	(2-Sided) 95%; 0.96 to 1.63
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, alcohol consumption, diabetes, neuropathic pain, back pain, epilepsy, hysterectomy, current estrogen, prior estrogen, benign breast disease.

31. Primary Outcome

Title	Number of Penile Cancer Cases and Matched Controls With the Indicated Exposure to Gabapentin
Description	Incident penile cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin prescription from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin prescription from cohort entry to index date.
Time Frame	The case index date was the date of incident penile cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Penile Cancer Patients	Matched Controls for Incident Penile Cancer Patients
Arm/Group Description	Patients with incident penile cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident penile cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	148	1396
Ever (with 2 year lag)	1 0%	2 0%
Never (with 2 year lag)	147 0.1%	1394 0.1%
Ever (without 2 year lag)	1 0%	6 0%
Never (without 2 year lag)	147 0.1%	1390 0.1%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.3082
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	3.58
	Confidence Interval	(2-Sided) 95%; 0.31 to 41.71
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Ever (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.3999
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	2.90
	Confidence Interval	(2-Sided) 95%; 0.24 to 34.69
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Ever (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, HIV, HPV (genital), phimosis/balanitis.

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.7509
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.42
	Confidence Interval	(2-Sided) 95%; 0.16 to 12.52
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Ever (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.6303
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.75
	Confidence Interval	(2-Sided) 95%; 0.18 to 17.07
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Ever (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, HIV, HPV (genital), phimosis/balanitis.

32. Primary Outcome

Title	Number of Penile Cancer Cases and Matched Controls With the Indicated Number of Gabapentin Prescriptions
Description	Incident penile cancer. Gabapentin (Gaba.) Exposure Description: With 2 year (yr) lag=Gaba. exposure from cohort entry to 2 yr prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 yr lag=Gaba. exposure from cohort entry to index date. Tertiles (T) with 2 yr lag: T 1 (1-2 prescriptions [prescrip.]), T 2 (3-7 prescrip.), T 3 (8-298 prescrip.). Tertiles without 2 yr lag: T 1 (1-2 prescrip.), T 2 (3-7 prescrip.), and T 3 (8-388 prescrip.). Inestimable OR/95% CI when no gaba.-exposed cancer cases or controls at the exposure level.
Time Frame	The case index date was the date of incident penile cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Penile Cancer Patients	Matched Controls for Incident Penile Cancer Patients
Arm/Group Description	Patients with incident penile cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident penile cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	148	1396
Never (with 2 year lag)	147 0.1%	1394 0.1%
Tertile 1 (with 2 year lag)	1 0%	1 0%
Tertile 2 (with 2 year lag)	0 0%	0 0%
Tertile 3 (with 2 year lag)	0 0%	1 0%
Never (without 2 year lag)	147 0.1%	1390 0.1%
Tertile 1 (without 2 year lag)	1 0%	1 0%
Tertile 2 (without 2 year lag)	0 0%	3 0%
Tertile 3 (without 2 year lag)	0 0%	2 0%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.2038
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	6.32
	Confidence Interval	(2-Sided) 95%; 0.37 to 108.8
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.2866
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	4.63
	Confidence Interval	(2-Sided) 95%; 0.28 to 77.51
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, HIV, HPV (genital), phimosis/balanitis.

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.2038
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	6.32
	Confidence Interval	(2-Sided) 95%; 0.37 to 108.8
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.1118
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	20.40
	Confidence Interval	(2-Sided) 95%; 0.50 to 839.0
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, HIV, HPV (genital), phimosis/balanitis.

33. Primary Outcome

Title	Number of Penile Cancer Cases and Matched Controls With the Indicated Duration of Exposure to Gabapentin
Description	Incident penile cancer. Gabapentin (Gaba.) Exposure Description: With 2 year (yr) lag=Gaba. exposure from cohort entry to 2 yr prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 yr lag=Gaba. exposure from cohort entry to index date. Tertiles (T) with 2 yr lag: T 1 (0.01-1.38 months [mo.]), T 2 (1.39-5.56 mo.), and T 3 (5.57-105.82 mo.). Tertiles without 2 yr lag: T 1 (0.01-1.38 mo.), T 2 (1.39-5.72 mo.), and T 3 (5.73-123.70 mo.). Inestimable OR and 95% CI when no gaba.-exposed cancer cases or controls at the exposure level.
Time Frame	The case index date was the date of incident penile cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Penile Cancer Patients	Matched Controls for Incident Penile Cancer Patients
Arm/Group Description	Patients with incident penile cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident penile cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	148	1396
Never (with 2 year lag)	147 0.1%	1394 0.1%
Tertile 1 (with 2 year lag)	1 0%	1 0%
Tertile 2 (with 2 year lag)	0 0%	0 0%
Tertile 3 (with 2 year lag)	0 0%	1 0%
Never (without 2 year lag)	147 0.1%	1390 0.1%
Tertile 1 (without 2 year lag)	1 0%	1 0%
Tertile 2 (without 2 year lag)	0 0%	3 0%
Tertile 3 (without 2 year lag)	0 0%	2 0%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.2038
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	6.32
	Confidence Interval	(2-Sided) 95%; 0.37 to 108.8
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.2866
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	4.63
	Confidence Interval	(2-Sided) 95%; 0.28 to 77.51
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, HIV, HPV (genital), phimosis/balanitis.

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.2038
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	6.32
	Confidence Interval	(2-Sided) 95%; 0.37 to 108.8
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.1118
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	20.40
	Confidence Interval	(2-Sided) 95%; 0.50 to 839.0
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, HIV, HPV (genital), phimosis/balanitis.

34. Primary Outcome

Title	Number of Penile Cancer Cases and Matched Controls With the Indicated Long Duration of Exposure to Gabapentin
Description	Incident penile cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date.
Time Frame	The case index date was the date of incident penile cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Penile Cancer Patients	Matched Controls for Incident Penile Cancer Patients
Arm/Group Description	Patients with incident penile cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident penile cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	148	1396
Never (with 2 year lag)	147 0.1%	1394 0.1%
> 1 year (with 2 year lag)	0 0%	1 0%
> 2 year (with 2 year lag)	0 0%	0 0%
> 3 year (with 2 year lag)	0 0%	0 0%
Never (without 2 year lag)	147 0.1%	1390 0.1%
> 1 year (without 2 year lag)	0 0%	1 0%
> 2 year (without 2 year lag)	0 0%	1 0%
> 3 year (without 2 year lag)	0 0%	1 0%

35. Primary Outcome

Title	Number of Penile Cancer Cases and Matched Controls With the Indicated Cumulative Dose of Gabapentin
Description	Incident penile cancer. Gabapentin (Gaba.) Exposure Description: With 2 year (yr) lag=Gaba. exposure from cohort entry to 2 yr prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 yr lag=Gaba. exposure from cohort entry to index date. Tertiles (T) with 2 yr lag: T 1 (0.1-30.0 grams [g]), T 2 (30.1-189.0 g), and T 3 (189.1-9600.0 g). Tertiles without 2 yr lag: T 1 (0.1-30.0 g), T 2 (30.1-189.0 g), and T 3 (189.1-11610.0 g). Inestimable OR and 95% CI when no gaba.-exposed cancer cases or controls at the exposure level.
Time Frame	The case index date was the date of incident penile cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Penile Cancer Patients	Matched Controls for Incident Penile Cancer Patients
Arm/Group Description	Patients with incident penile cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident penile cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	148	1396
Never (with 2 year lag)	147 0.1%	1394 0.1%
Tertile 1 (with 2 year lag)	1 0%	1 0%
Tertile 2 (with 2 year lag)	0 0%	0 0%
Tertile 3 (with 2 year lag)	0 0%	1 0%
Never (without 2 year lag)	147 0.1%	1390 0.1%
Tertile 1 (without 2 year lag)	1 0%	1 0%
Tertile 2 (without 2 year lag)	0 0%	2 0%
Tertile 3 (without 2 year lag)	0 0%	3 0%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.2038
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	6.32
	Confidence Interval	(2-Sided) 95%; 0.37 to 108.8
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.2866
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	4.63
	Confidence Interval	(2-Sided) 95%; 0.28 to 77.51
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, HIV, HPV (genital), phimosis/balanitis.

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.2038
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	6.32
	Confidence Interval	(2-Sided) 95%; 0.37 to 108.8
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.1118
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	20.40
	Confidence Interval	(2-Sided) 95%; 0.5 to 839.0
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, HIV, HPV (genital), phimosis/balanitis.

36. Primary Outcome

Title	Number of Bladder Cancer Cases and Matched Controls With the Indicated Exposure to Gabapentin
Description	Incident bladder cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin prescription from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin prescription from cohort entry to index date.
Time Frame	The case index date was the date of incident bladder cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Bladder Cancer Patients	Matched Controls for Incident Bladder Cancer Patients
Arm/Group Description	Patients with incident bladder cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident bladder cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	4600	43559
Ever (with 2 year lag)	15 0%	144 0%
Never (with 2 year lag)	4585 2.6%	43415 2.5%
Ever (without 2 year lag)	33 0%	299 0%
Never (without 2 year lag)	4567 2.5%	43260 2.5%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.9614
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.99
	Confidence Interval	(2-Sided) 95%; 0.58 to 1.69
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Ever (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.5179
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.84
	Confidence Interval	(2-Sided) 95%; 0.49 to 1.44
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Ever (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy.

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.8856
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.03
	Confidence Interval	(2-Sided) 95%; 0.71 to 1.48
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Ever (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.5078
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.88
	Confidence Interval	(2-Sided) 95%; 0.61 to 1.28
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Ever (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy.

37. Primary Outcome

Title	Number of Bladder Cancer Cases and Matched Controls With the Indicated Number of Gabapentin Prescriptions
Description	Incident bladder cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date. Tertile's with 2 year lag: Tertile 1 (1-2 prescriptions), Tertile 2 (3-7 prescriptions), Tertile 3 (8-298 prescriptions). Tertiles without 2 year lag: Tertile 1 (1-2 prescriptions), Tertile 2 (3-7 prescriptions), and Tertile 3 (8-388 prescriptions).
Time Frame	The case index date was the date of incident bladder cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Bladder Cancer Patients	Matched Controls for Incident Bladder Cancer Patients
Arm/Group Description	Patients with incident bladder cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident bladder cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	4600	43559
Never (with 2 year lag)	4585 2.6%	43415 2.5%
Tertile 1 (with 2 year lag)	6 0%	68 0%
Tertile 2 (with 2 year lag)	3 0%	35 0%
Tertile 3 (with 2 year lag)	6 0%	41 0%
Never (without 2 year lag)	4567 2.5%	43260 2.5%
Tertile 1 (without 2 year lag)	13 0%	138 0%
Tertile 2 (without 2 year lag)	10 0%	71 0%
Tertile 3 (without 2 year lag)	10 0%	90 0%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.6888
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.84
	Confidence Interval	(2-Sided) 95%; 0.36 to 1.95
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.5024
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.75
	Confidence Interval	(2-Sided) 95%; 0.32 to 1.75
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy.

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.6164
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.86
	Confidence Interval	(2-Sided) 95%; 0.49 to 1.53
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.4002
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.78
	Confidence Interval	(2-Sided) 95%; 0.44 to 1.39
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy.

Statistical Analysis 5

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.7329
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.81
	Confidence Interval	(2-Sided) 95%; 0.25 to 2.66
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 6

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.5731
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.71
	Confidence Interval	(2-Sided) 95%; 0.22 to 2.33
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy.

Statistical Analysis 7

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.3762
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.35
	Confidence Interval	(2-Sided) 95%; 0.69 to 2.63
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 8

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.5820
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.21
	Confidence Interval	(2-Sided) 95%; 0.62 to 2.36
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy.

Statistical Analysis 9

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.4782
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.37
	Confidence Interval	(2-Sided) 95%; 0.58 to 3.23
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 10

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.8998
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.06
	Confidence Interval	(2-Sided) 95%; 0.44 to 2.52
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy.

Statistical Analysis 11

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.9308
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.03
	Confidence Interval	(2-Sided) 95%; 0.53 to 1.99
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 12

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.5125
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.80
	Confidence Interval	(2-Sided) 95%; 0.41 to 1.56
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy.

38. Primary Outcome

Title	Number of Bladder Cancer Cases and Matched Controls With the Indicated Duration of Exposure to Gabapentin
Description	Incident bladder cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date. Tertile's with 2 year lag: Tertile 1 (0.01 - 1.38 months), Tertile 2 (1.39 - 5.56 months), and Tertile 3 (5.57 - 105.82 months). Tertile's without 2 year lag: Tertile 1 (0.01 - 1.38 months), Tertile 2 (1.39 - 5.72 months), and Tertile 3 (5.73 - 123.70 months).
Time Frame	The case index date was the date of incident bladder cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Bladder Cancer Patients	Matched Controls for Incident Bladder Cancer Patients
Arm/Group Description	Patients with incident bladder cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident bladder cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	4600	43559
Never (with 2 year lag)	4585 2.6%	43415 2.5%
Tertile 1 (with 2 year lag)	5 0%	51 0%
Tertile 2 (with 2 year lag)	4 0%	52 0%
Tertile 3 (with 2 year lag)	6 0%	41 0%
Never (without 2 year lag)	4567 2.5%	43260 2.5%
Tertile 1 (without 2 year lag)	12 0%	112 0%
Tertile 2 (without 2 year lag)	11 0%	100 0%
Tertile 3 (without 2 year lag)	10 0%	87 0%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.8655
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.92
	Confidence Interval	(2-Sided) 95%; 0.37 to 2.33
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.6482
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.80
	Confidence Interval	(2-Sided) 95%; 0.32 to 2.05
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy.

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.9003
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.96
	Confidence Interval	(2-Sided) 95%; 0.53 to 1.76
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.6000
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.85
	Confidence Interval	(2-Sided) 95%; 0.46 to 1.56
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy.

Statistical Analysis 5

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.5670
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.74
	Confidence Interval	(2-Sided) 95%; 0.27 to 2.06
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 6

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.4353
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.67
	Confidence Interval	(2-Sided) 95%; 0.24 to 1.85
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy.

Statistical Analysis 7

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.8922
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.04
	Confidence Interval	(2-Sided) 95%; 0.56 to 1.95
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 8

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.8591
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.94
	Confidence Interval	(2-Sided) 95%; 0.50 to 1.77
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy.

Statistical Analysis 9

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.4779
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.37
	Confidence Interval	(2-Sided) 95%; 0.58 to 3.23
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 10

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.8994
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.06
	Confidence Interval	(2-Sided) 95%; 0.44 to 2.52
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy.

Statistical Analysis 11

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.7800
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.10
	Confidence Interval	(2-Sided) 95%; 0.57 to 2.12
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 12

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.6577
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.86
	Confidence Interval	(2-Sided) 95%; 0.44 to 1.67
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy.

39. Primary Outcome

Title	Number of Bladder Cancer Cases and Matched Controls With the Indicated Long Duration of Exposure to Gabapentin
Description	Incident bladder cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date.
Time Frame	The case index date was the date of incident bladder cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Bladder Cancer Patients	Matched Controls for Incident Bladder Cancer Patients
Arm/Group Description	Patients with incident bladder cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident bladder cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	4600	43559
Never (with 2 year lag)	4585 2.6%	43415 2.5%
> 1 year (with 2 year lag)	2 0%	21 0%
> 2 year (with 2 year lag)	0 0%	7 0%
> 3 year (with 2 year lag)	0 0%	3 0%
Never (without 2 year lag)	4567 2.5%	43260 2.5%
> 1 year (without 2 year lag)	6 0%	52 0%
> 2 year (without 2 year lag)	2 0%	24 0%
> 3 year (without 2 year lag)	1 0%	10 0%

40. Primary Outcome

Title	Number of Bladder Cancer Cases and Matched Controls With the Indicated Cumulative Dose of Gabapentin
Description	Incident bladder cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date. Tertile's with 2 year lag: Tertile 1 (0.1 - 30.0 grams), Tertile 2 (30.1 - 189.0 grams), and Tertile 3 (189.1 - 9600.0 grams). Tertile's without 2 year lag: Tertile 1 (0.1 - 30.0 grams), Tertile 2 (30.1 - 189.0 grams), and Tertile 3 (189.1 - 11610.0 grams).
Time Frame	The case index date was the date of incident bladder cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Bladder Cancer Patients	Matched Controls for Incident Bladder Cancer Patients
Arm/Group Description	Patients with incident bladder cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident bladder cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	4600	43559
Never (with 2 year lag)	4585 2.6%	43415 2.5%
Tertile 1 (with 2 year lag)	6 0%	52 0%
Tertile 2 (with 2 year lag)	4 0%	51 0%
Tertile 3 (with 2 year lag)	5 0%	41 0%
Never (without 2 year lag)	4567 2.5%	43260 2.5%
Tertile 1 (without 2 year lag)	14 0%	112 0%
Tertile 2 (without 2 year lag)	10 0%	103 0%
Tertile 3 (without 2 year lag)	9 0%	84 0%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.8111
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.11
	Confidence Interval	(2-Sided) 95%; 0.48 to 2.59
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.9104
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.05
	Confidence Interval	(2-Sided) 95%; 0.45 to 2.46
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy.

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.5592
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.18
	Confidence Interval	(2-Sided) 95%; 0.68 to 2.07
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.8521
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.06
	Confidence Interval	(2-Sided) 95%; 0.60 to 1.85
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy.

Statistical Analysis 5

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.5529
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.73
	Confidence Interval	(2-Sided) 95%; 0.26 to 2.04
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 6

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.3275
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Adjusted Odds Ratio
	Estimated Value	0.60
	Confidence Interval	(2-Sided) 95%; 0.21 to 1.68
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy.

Statistical Analysis 7

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.7300
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.89
	Confidence Interval	(2-Sided) 95%; 0.46 to 1.72
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 8

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.4831
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.79
	Confidence Interval	(2-Sided) 95%; 0.41 to 1.53
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy.

Statistical Analysis 9

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.7693
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.15
	Confidence Interval	(2-Sided) 95%; 0.45 to 2.92
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 10

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.8189
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.90
	Confidence Interval	(2-Sided) 95%; 0.35 to 2.30
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy.

Statistical Analysis 11

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.9800
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.99
	Confidence Interval	(2-Sided) 95%; 0.50 to 1.98
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 12

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.4944
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.78
	Confidence Interval	(2-Sided) 95%; 0.39 to 1.58
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy.

41. Primary Outcome

Title	Number of Other Nervous System Cancer Cases and Matched Controls With the Indicated Exposure to Gabapentin
Description	Incident other nervous system cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin prescription from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin prescription from cohort entry to index date. Inestimable OR and 95% CI when no gabapentin-exposed cancer cases or no gabapentin-exposed controls at the exposure level.
Time Frame	The case index date was the date of incident other nervous system cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his cohort entry was the same as case.

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Other Nervous System Cancer Patients	Matched Controls for Incident ONS Cancer Patients
Arm/Group Description	Patients with incident other nervous system cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident other nervous system cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	38	380
Ever (with 2 year lag)	0 0%	0 0%
Never (with 2 year lag)	38 0%	380 0%
Ever (without 2 year lag)	2 0%	1 0%
Never (without 2 year lag)	36 0%	379 0%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0144
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	20.00
	Confidence Interval	(2-Sided) 95%; 1.81 to 220.5
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Ever (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0743
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	10.79
	Confidence Interval	(2-Sided) 95%; 0.79 to 147.0
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Ever (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy.

42. Primary Outcome

Title	Number of Other Nervous System (ONS) Cancer Cases and Matched Controls With the Indicated Number of Gabapentin Prescriptions
Description	Incident ONS cancer. Gabapentin (Gaba.) Exposure Description: With 2 year (yr) lag=Gaba. exposure from cohort entry to 2 yr prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 yr lag=Gaba. exposure from cohort entry to index date. Tertiles (T) with 2 yr lag: T 1 (1-2 prescriptions [prescrip.]), T 2 (3-7 prescrip.), T 3 (8-298 prescrip.). Tertiles without 2 yr lag: T 1 (1-2 prescrip.), T 2 (3-7 prescrip.), and T 3 (8-388 prescrip.). Inestimable OR and 95% CI when no gaba.-exposed cancer cases or controls at the exposure level.
Time Frame	The case index date was the date of incident other nervous system cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his cohort entry was same as for case.

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Other Nervous System Cancer Patients	Matched Controls for Incident ONS Cancer Patients
Arm/Group Description	Patients with incident other nervous system cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident other nervous system cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	38	380
Never (with 2 year lag)	38 0%	380 0%
Tertile 1 (with 2 year lag)	0 0%	0 0%
Tertile 2 (with 2 year lag)	0 0%	0 0%
Tertile 3 (with 2 year lag)	0 0%	0 0%
Never (without 2 year lag)	36 0%	379 0%
Tertile 1 (without 2 year lag)	0 0%	0 0%
Tertile 2 (without 2 year lag)	1 0%	1 0%
Tertile 3 (without 2 year lag)	1 0%	0 0%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.1035
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Unadjusted Odds Ratio
	Estimated Value	10.00
	Confidence Interval	(2-Sided) 95%; 0.63 to 159.9
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.3799
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	4.10
	Confidence Interval	(2-Sided) 95%; 0.18 to 95.82
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy.

43. Primary Outcome

Title	Number of Other Nervous System (ONS) Cancer Cases and Matched Controls With the Indicated Duration of Exposure to Gabapentin
Description	Incident ONS cancer. Gabapentin (Gaba.) Exposure Description: With 2 yr lag=Gaba. exposure from cohort entry to 2 yr prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 yr lag=Gaba. exposure from cohort entry to index date. Tertiles (T) with 2 yr lag: T 1 (0.01-1.38 months [mo.]), T 2 (1.39-5.56 mo.), and T 3 (5.57-105.82 mo.). Tertiles without 2 yr lag: T 1 (0.01-1.38 mo.), T 2 (1.39-5.72 mo.), and T 3 (5.73-123.70 mo.). Inestimable OR and 95% CI when no gaba.-exposed cancer cases or controls at the exposure level.
Time Frame	The case index date was the date of incident other nervous system cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was same as for case

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Other Nervous System Cancer Patients	Matched Controls for Incident ONS Cancer Patients
Arm/Group Description	Patients with incident other nervous system cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident other nervous system cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	38	380
Never (with 2 year lag)	38 0%	380 0%
Tertile 1 (with 2 year lag)	0 0%	0 0%
Tertile 2 (with 2 year lag)	0 0%	0 0%
Tertile 3 (with 2 year lag)	0 0%	0 0%
Never (without 2 year lag)	36 0%	379 0%
Tertile 1 (without 2 year lag)	0 0%	0 0%
Tertile 2 (without 2 year lag)	1 0%	1 0%
Tertile 3 (without 2 year lag)	1 0%	0 0%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.1035
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	10.00
	Confidence Interval	(2-Sided) 95%; 0.63 to 159.9
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.3799
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	4.10
	Confidence Interval	(2-Sided) 95%; 0.18 to 95.82
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy.

44. Primary Outcome

Title	Number of Other Nervous System Cancer Cases and Matched Controls With the Indicated Long Duration of Exposure to Gabapentin
Description	Incident other nervous system cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date.
Time Frame	The case index date was the date of incident other nervous system cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was same as for case

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Other Nervous System Cancer Patients	Matched Controls for Incident ONS Cancer Patients
Arm/Group Description	Patients with incident other nervous system cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident other nervous system cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	38	380
Never (with 2 year lag)	38 0%	380 0%
> 1 year (with 2 year lag)	0 0%	0 0%
> 2 year (with 2 year lag)	0 0%	0 0%
> 3 year (with 2 year lag)	0 0%	0 0%
Never (without 2 year lag)	36 0%	379 0%
> 1 year (without 2 year lag)	1 0%	0 0%
> 2 year (without 2 year lag)	0 0%	0 0%
> 3 year (without 2 year lag)	0 0%	0 0%

45. Primary Outcome

Title	Number of Other Nervous System (ONS) Cancer Cases and Matched Controls With the Indicated Cumulative Dose of Gabapentin
Description	Incident ONS cancer. Gabapentin (Gaba.) Exposure Description: With 2 year (yr) lag=Gaba. exposure from cohort entry to 2 yr prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 yr lag=Gaba. exposure from cohort entry to index date. Tertiles (T) with 2 yr lag: T 1 (0.1-30.0 grams [g]), T 2 (30.1-189.0 g), and T 3 (189.1-9600.0 g). Tertiles without 2 yr lag: T 1 (0.1-30.0 g), T 2 (30.1-189.0 g), and T 3 (189.1-11610.0 g). Inestimable OR and 95% CI when no gaba.-exposed cancer cases or controls at the exposure level.
Time Frame	The case index date was the date of incident other nervous system cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was same as for case

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Other Nervous System Cancer Patients	Matched Controls for Incident ONS Cancer Patients
Arm/Group Description	Patients with incident other nervous system cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident other nervous system cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	38	380
Never (with 2 year lag)	38 0%	380 0%
Tertile 1 (with 2 year lag)	0 0%	0 0%
Tertile 2 (with 2 year lag)	0 0%	0 0%
Tertile 3 (with 2 year lag)	0 0%	0 0%
Never (without 2 year lag)	36 0%	379 0%
Tertile 1 (without 2 year lag)	0 0%	0 0%
Tertile 2 (without 2 year lag)	1 0%	1 0%
Tertile 3 (without 2 year lag)	1 0%	0 0%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.1035
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	10.00
	Confidence Interval	(2-Sided) 95%; 0.63 to 159.9
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.3799
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	4.10
	Confidence Interval	(2-Sided) 95%; 0.18 to 95.82
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy.

46. Primary Outcome

Title	Number of Pancreatic Cancer Cases and Matched Controls With the Indicated Exposure to Gabapentin
Description	Incident pancreatic cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin prescription from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin prescription from cohort entry to index date.
Time Frame	The case index date was the date of incident pancreatic cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as for the case.

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Pancreatic Cancer Patients	Matched Controls for Incident Pancreatic Cancer Patients
Arm/Group Description	Patients with incident pancreatic cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident pancreatic cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	2155	20382
Ever (with 2 year lag)	20 0%	86 0%
Never (with 2 year lag)	2135 1.2%	20296 1.2%
Ever (without 2 year lag)	35 0%	160 0%
Never (without 2 year lag)	2120 1.2%	20222 1.2%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0066
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	2.01
	Confidence Interval	(2-Sided) 95%; 1.21 to 3.32
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Ever (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0494
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.68
	Confidence Interval	(2-Sided) 95%; 1.00 to 2.82
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Ever (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, chronic pancreatitis.

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0006
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.93
	Confidence Interval	(2-Sided) 95%; 1.33 to 2.82
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Ever (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0138
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.63
	Confidence Interval	(2-Sided) 95%; 1.11 to 2.41
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Ever (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, chronic pancreatitis.

47. Primary Outcome

Title	Number of Pancreatic Cancer Cases and Matched Controls With the Indicated Number of Gabapentin Prescriptions
Description	Incident pancreatic cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date. Tertile's with 2 year lag: Tertile 1 (1-2 prescriptions), Tertile 2 (3-7 prescriptions), Tertile 3 (8-298 prescriptions). Tertiles without 2 year lag: Tertile 1 (1-2 prescriptions), Tertile 2 (3-7 prescriptions), and Tertile 3 (8-388 prescriptions).
Time Frame	The case index date was the date of incident pancreatic cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as for the case.

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Pancreatic Cancer Patients	Matched Controls for Incident Pancreatic Cancer Patients
Arm/Group Description	Patients with incident pancreatic cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident pancreatic cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	2155	20382
Never (with 2 year lag)	2135 1.2%	20296 1.2%
Tertile 1 (with 2 year lag)	12 0%	34 0%
Tertile 2 (with 2 year lag)	2 0%	22 0%
Tertile 3 (with 2 year lag)	6 0%	30 0%
Never (without 2 year lag)	2120 1.2%	20222 1.2%
Tertile 1 (without 2 year lag)	17 0%	62 0%
Tertile 2 (without 2 year lag)	4 0%	30 0%
Tertile 3 (without 2 year lag)	14 0%	68 0%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0006
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	3.29
	Confidence Interval	(2-Sided) 95%; 1.66 to 6.53
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0020
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	3.01
	Confidence Interval	(2-Sided) 95%; 1.50 to 6.05
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, chronic pancreatitis.

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0013
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	2.49
	Confidence Interval	(2-Sided) 95%; 1.43 to 4.34
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0080
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	2.16
	Confidence Interval	(2-Sided) 95%; 1.22 to 3.80
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, chronic pancreatitis.

Statistical Analysis 5

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.8165
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.84
	Confidence Interval	(2-Sided) 95%; 0.20 to 3.61
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 6

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.6624
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.72
	Confidence Interval	(2-Sided) 95%; 0.17 to 3.12
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, chronic pancreatitis.

Statistical Analysis 7

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.6316
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.29
	Confidence Interval	(2-Sided) 95%; 0.45 to 3.68
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 8

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.9060
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.07
	Confidence Interval	(2-Sided) 95%; 0.35 to 3.22
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, chronic pancreatitis.

Statistical Analysis 9

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.3731
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.51
	Confidence Interval	(2-Sided) 95%; 0.61 to 3.74
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 10

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.8266
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.11
	Confidence Interval	(2-Sided) 95%; 0.43 to 2.87
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, chronic pancreatitis.

Statistical Analysis 11

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0712
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.73
	Confidence Interval	(2-Sided) 95%; 0.95 to 3.13
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 12

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.2674
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.41
	Confidence Interval	(2-Sided) 95%; 0.77 to 2.60
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, chronic pancreatitis.

48. Primary Outcome

Title	Number of Pancreatic Cancer Cases and Matched Controls With the Indicated Duration of Exposure to Gabapentin
Description	Incident pancreatic cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date. Tertile's with 2 year lag: Tertile 1 (0.01 - 1.38 months), Tertile 2 (1.39 - 5.56 months), and Tertile 3 (5.57 - 105.82 months). Tertile's without 2 year lag: Tertile 1 (0.01 - 1.38 months), Tertile 2 (1.39 - 5.72 months), and Tertile 3 (5.73 - 123.70 months).
Time Frame	The case index date was the date of incident pancreatic cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as for the case.

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Pancreatic Cancer Patients	Matched Controls for Incident Pancreatic Cancer Patients
Arm/Group Description	Patients with incident pancreatic cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident pancreatic cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	2155	20382
Never (with 2 year lag)	2135 1.2%	20296 1.2%
Tertile 1 (with 2 year lag)	11 0%	28 0%
Tertile 2 (with 2 year lag)	2 0%	27 0%
Tertile 3 (with 2 year lag)	7 0%	31 0%
Never (without 2 year lag)	2120 1.2%	20222 1.2%
Tertile 1 (without 2 year lag)	15 0%	53 0%
Tertile 2 (without 2 year lag)	5 0%	43 0%
Tertile 3 (without 2 year lag)	15 0%	64 0%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0005
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	3.61
	Confidence Interval	(2-Sided) 95%; 1.75 to 7.46
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0019
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	3.25
	Confidence Interval	(2-Sided) 95%; 1.55 to 6.83
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, chronic pancreatitis.

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0024
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	2.51
	Confidence Interval	(2-Sided) 95%; 1.39 to 4.55
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0115
	Comments
	Method	Wilcoxon (Mann-Whitney)
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	2.18
	Confidence Interval	(2-Sided) 95%; 1.19 to 4.00
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, chronic pancreatitis.

Statistical Analysis 5

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.6281
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.70
	Confidence Interval	(2-Sided) 95%; 0.17 to 2.96
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 6

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.5103
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.61
	Confidence Interval	(2-Sided) 95%; 0.14 to 2.62
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, chronic pancreatitis.

Statistical Analysis 7

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.7733
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Unadjusted Odds Ratio
	Estimated Value	1.15
	Confidence Interval	(2-Sided) 95%; 0.45 to 2.90
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 8

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.8370
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.10
	Confidence Interval	(2-Sided) 95%; 0.43 to 2.81
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, chronic pancreatitis.

Statistical Analysis 9

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.1980
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.75
	Confidence Interval	(2-Sided) 95%; 0.75 to 4.10
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 10

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.5391
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.32
	Confidence Interval	(2-Sided) 95%; 0.54 to 3.21
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, chronic pancreatitis.

Statistical Analysis 11

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0224
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.97
	Confidence Interval	(2-Sided) 95%; 1.10 to 3.53
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 12

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.1815
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.51
	Confidence Interval	(2-Sided) 95%; 0.82 to 2.77
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, chronic pancreatitis.

49. Primary Outcome

Title	Number of Pancreatic Cancer Cases and Matched Controls With the Indicated Long Duration of Exposure to Gabapentin
Description	Incident pancreatic cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date.
Time Frame	The case index date was the date of incident pancreatic cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as for the case

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Pancreatic Cancer Patients	Matched Controls for Incident Pancreatic Cancer Patients
Arm/Group Description	Patients with incident pancreatic cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident pancreatic cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	2155	20382
Never (with 2 year lag)	2135 1.2%	20296 1.2%
> 1 year (with 2 year lag)	5 0%	18 0%
> 2 year (with 2 year lag)	2 0%	6 0%
> 3 year (with 2 year lag)	1 0%	2 0%
Never (without 2 year lag)	2120 1.2%	20222 1.2%
> 1 year (without 2 year lag)	7 0%	46 0%
> 2 year (without 2 year lag)	2 0%	23 0%
> 3 year (without 2 year lag)	1 0%	6 0%

50. Primary Outcome

Title	Number of Pancreatic Cancer Cases and Matched Controls With the Indicated Cumulative Dose of Gabapentin
Description	Incident pancreatic cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date. Tertile's with 2 year lag: Tertile 1 (0.1 - 30.0 grams), Tertile 2 (30.1 - 189.0 grams), and Tertile 3 (189.1 - 9600.0 grams). Tertile's without 2 year lag: Tertile 1 (0.1 - 30.0 grams), Tertile 2 (30.1 - 189.0 grams), and Tertile 3 (189.1 - 11610.0 grams).
Time Frame	The case index date was the date of incident pancreatic cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as for the case.

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Pancreatic Cancer Patients	Matched Controls for Incident Pancreatic Cancer Patients
Arm/Group Description	Patients with incident pancreatic cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident pancreatic cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	2155	20382
Never (with 2 year lag)	2135 1.2%	20296 1.2%
Tertile 1 (with 2 year lag)	9 0%	33 0%
Tertile 2 (with 2 year lag)	6 0%	22 0%
Tertile 3 (with 2 year lag)	5 0%	31 0%
Never (without 2 year lag)	2120 1.2%	20222 1.2%
Tertile 1 (without 2 year lag)	15 0%	56 0%
Tertile 2 (without 2 year lag)	8 0%	45 0%
Tertile 3 (without 2 year lag)	12 0%	59 0%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0248
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	2.38
	Confidence Interval	(2-Sided) 95%; 1.12 to 5.09
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0520
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	2.16
	Confidence Interval	(2-Sided) 95%; 0.99 to 4.70
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, chronic pancreatitis.

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0046
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	2.33
	Confidence Interval	(2-Sided) 95%; 1.30 to 4.20
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0193
	Comments
	Method	Wilcoxon (Mann-Whitney)
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	2.05
	Confidence Interval	(2-Sided) 95%; 1.12 to 3.72
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, chronic pancreatitis.

Statistical Analysis 5

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0309
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	2.75
	Confidence Interval	(2-Sided) 95%; 1.10 to 6.88
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 6

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.1285
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	2.11
	Confidence Interval	(2-Sided) 95%; 0.81 to 5.51
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, chronic pancreatitis.

Statistical Analysis 7

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.1257
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.80
	Confidence Interval	(2-Sided) 95%; 0.85 to 3.83
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 8

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.2944
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.52
	Confidence Interval	(2-Sided) 95%; 0.70 to 3.31
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, chronic pancreatitis.

Statistical Analysis 9

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.7032
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.21
	Confidence Interval	(2-Sided) 95%; 0.45 to 3.21
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 10

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.9922
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.00
	Confidence Interval	(2-Sided) 95%; 0.37 to 2.69
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, chronic pancreatitis.

Statistical Analysis 11

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.1227
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.66
	Confidence Interval	(2-Sided) 95%; 0.87 to 3.16
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 12

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.3809
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.34
	Confidence Interval	(2-Sided) 95%; 0.69 to 2.61
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, chronic pancreatitis.

51. Primary Outcome

Title	Number of Renal Cancer Cases and Matched Controls With the Indicated Exposure to Gabapentin
Description	Incident renal cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin prescription from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin prescription from cohort entry to index date.
Time Frame	The case index date was the date of incident pancreatic cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as for the case.

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Renal Cancer Patients	Matched Controls for Incident Renal Cancer Patients
Arm/Group Description	Patients with incident renal cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident renal cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	1272	12167
Ever (with 2 year lag)	9 0%	53 0%
Never (with 2 year lag)	1263 0.7%	12114 0.7%
Ever (without 2 year lag)	15 0%	107 0%
Never (without 2 year lag)	1257 0.7%	12060 0.7%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.1757
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.64
	Confidence Interval	(2-Sided) 95%; 0.80 to 3.34
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Ever (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.4405
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.33
	Confidence Interval	(2-Sided) 95%; 0.64 to 2.75
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Ever (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, hypertension, diuretic use.

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.4420
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.24
	Confidence Interval	(2-Sided) 95%; 0.71 to 2.16
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Ever (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.9034
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.04
	Confidence Interval	(2-Sided) 95%; 0.59 to 1.82
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Ever (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, hypertension, diuretic use.

52. Primary Outcome

Title	Number of Renal Cancer Cases and Matched Controls With the Indicated Number of Gabapentin Prescriptions
Description	Incident renal cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date. Tertile's with 2 year lag: Tertile 1 (1-2 prescriptions), Tertile 2 (3-7 prescriptions), Tertile 3 (8-298 prescriptions). Tertiles without 2 year lag: Tertile 1 (1-2 prescriptions), Tertile 2 (3-7 prescriptions), and Tertile 3 (8-388 prescriptions).
Time Frame	The case index date was the date of incident renal cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Renal Cancer Patients	Matched Controls for Incident Renal Cancer Patients
Arm/Group Description	Patients with incident renal cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident renal cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	1272	12167
Never (with 2 year lag)	1263 0.7%	12114 0.7%
Tertile 1 (with 2 year lag)	4 0%	21 0%
Tertile 2 (with 2 year lag)	2 0%	15 0%
Tertile 3 (with 2 year lag)	3 0%	17 0%
Never (without 2 year lag)	1257 0.7%	12060 0.7%
Tertile 1 (without 2 year lag)	6 0%	46 0%
Tertile 2 (without 2 year lag)	3 0%	29 0%
Tertile 3 (without 2 year lag)	6 0%	32 0%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.2395
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.91
	Confidence Interval	(2-Sided) 95%; 0.65 to 5.60
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.3764
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.64
	Confidence Interval	(2-Sided) 95%; 0.55 to 4.93
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, hypertension, diuretic use.

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.7599
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.14
	Confidence Interval	(2-Sided) 95%; 0.48 to 2.71
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.9502
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.03
	Confidence Interval	(2-Sided) 95%; 0.43 to 2.46
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, hypertension, diuretic use.

Statistical Analysis 5

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.7164
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.31
	Confidence Interval	(2-Sided) 95%; 0.30 to 5.75
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 6

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.9345
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.06
	Confidence Interval	(2-Sided) 95%; 0.24 to 4.73
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, hypertension, diuretic use.

Statistical Analysis 7

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.8833
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.91
	Confidence Interval	(2-Sided) 95%; 0.28 to 3.02
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 8

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.6178
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.74
	Confidence Interval	(2-Sided) 95%; 0.22 to 2.46
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, hypertension, diuretic use.

Statistical Analysis 9

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.4556
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.60
	Confidence Interval	(2-Sided) 95%; 0.46 to 5.53
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 10

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.7478
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.23
	Confidence Interval	(2-Sided) 95%; 0.35 to 4.29
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, hypertension, diuretic use.

Statistical Analysis 11

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.2350
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.71
	Confidence Interval	(2-Sided) 95%; 0.71 to 4.13
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 12

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.5428
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.32
	Confidence Interval	(2-Sided) 95%; 0.54 to 3.24
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, hypertension, diuretic use.

53. Primary Outcome

Title	Number of Renal Cancer Cases and Matched Controls With the Indicated Duration of Exposure to Gabapentin
Description	Incident renal cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date. Tertile's with 2 year lag: Tertile 1 (0.01 - 1.38 months), Tertile 2 (1.39 - 5.56 months), and Tertile 3 (5.57 - 105.82 months). Tertile's without 2 year lag: Tertile 1 (0.01 - 1.38 months), Tertile 2 (1.39 - 5.72 months), and Tertile 3 (5.73 - 123.70 months).
Time Frame	The case index date was the date of incident renal cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Renal Cancer Patients	Matched Controls for Incident Renal Cancer Patients
Arm/Group Description	Patients with incident renal cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident renal cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	1272	12167
Never (with 2 year lag)	1263 0.7%	12114 0.7%
Tertile 1 (with 2 year lag)	3 0%	16 0%
Tertile 2 (with 2 year lag)	3 0%	22 0%
Tertile 3 (with 2 year lag)	3 0%	15 0%
Never (without 2 year lag)	1257 0.7%	12060 0.7%
Tertile 1 (without 2 year lag)	5 0%	32 0%
Tertile 2 (without 2 year lag)	4 0%	44 0%
Tertile 3 (without 2 year lag)	6 0%	31 0%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.3123
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.90
	Confidence Interval	(2-Sided) 95%; 0.55 to 6.59
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.3951
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.73
	Confidence Interval	(2-Sided) 95%; 0.49 to 6.15
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, hypertension, diuretic use.

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.4388
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.45
	Confidence Interval	(2-Sided) 95%; 0.56 to 3.75
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.5535
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.34
	Confidence Interval	(2-Sided) 95%; 0.51 to 3.47
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, hypertension, diuretic use.

Statistical Analysis 5

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.6332
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.34
	Confidence Interval	(2-Sided) 95%; 0.40 to 4.48
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 6

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.8743
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.10
	Confidence Interval	(2-Sided) 95%; 0.32 to 3.75
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, hypertension, diuretic use.

Statistical Analysis 7

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.6132
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.77
	Confidence Interval	(2-Sided) 95%; 0.27 to 2.15
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 8

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.3821
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Adjusted Odds Ratio
	Estimated Value	0.63
	Confidence Interval	(2-Sided) 95%; 0.22 to 1.78
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, hypertension, diuretic use.

Statistical Analysis 9

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.3591
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.80
	Confidence Interval	(2-Sided) 95%; 0.51 to 6.28
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 10

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.6899
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.29
	Confidence Interval	(2-Sided) 95%; 0.37 to 4.58
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, hypertension, diuretic use.

Statistical Analysis 11

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.2031
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.78
	Confidence Interval	(2-Sided) 95%; 0.73 to 4.30
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 12

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.4932
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.37
	Confidence Interval	(2-Sided) 95%; 0.56 to 3.37
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, hypertension, diuretic use.

54. Primary Outcome

Title	Number of Renal Cancer Cases and Matched Controls With the Indicated Long Duration of Exposure to Gabapentin
Description	Incident renal cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date.
Time Frame	The case index date was the date of incident renal cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Renal Cancer Patients	Matched Controls for Incident Renal Cancer Patients
Arm/Group Description	Patients with incident renal cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident renal cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	1272	12167
Never (with 2 year lag)	1263 0.7%	12114 0.7%
> 1 year (with 2 year lag)	2 0%	6 0%
> 2 year (with 2 year lag)	0 0%	2 0%
> 3 year (with 2 year lag)	0 0%	1 0%
Never (without 2 year lag)	1257 0.7%	12060 0.7%
> 1 year (without 2 year lag)	4 0%	16 0%
> 2 year (without 2 year lag)	4 0%	7 0%
> 3 year (without 2 year lag)	2 0%	3 0%

55. Primary Outcome

Title	Number of Renal Cancer Cases and Matched Controls With the Indicated Cumulative Dose of Gabapentin
Description	Incident renal cancer. Gabapentin Exposure Description: With 2 year lag = Gabapentin exposure from cohort entry to 2 years prior to index date (to control for prediagnostic prescribing for pain symptoms possibly related to cancer). Without 2 year lag = Gabapentin exposure from cohort entry to index date. Tertile's with 2 year lag: Tertile 1 (0.1 - 30.0 grams), Tertile 2 (30.1 - 189.0 grams), and Tertile 3 (189.1 - 9600.0 grams). Tertile's without 2 year lag: Tertile 1 (0.1 - 30.0 grams), Tertile 2 (30.1 - 189.0 grams), and Tertile 3 (189.1 - 11610.0 grams).
Time Frame	The case index date was the date of incident renal cancer diagnosis ascertained in the GPRD study cohort 1995-2008. The matched control index date was the date at which the follow-up time from his/her cohort entry was the same as that for the case.

Outcome Measure Data

Analysis Population Description

Cases & controls drawn from the GPRD study cohort. Entry into study cohort began January 1, 1993, or at time of GPRD registration if after Jan. 1, 1993. Follow-up ended December 31, 2008, or earlier if respective cancer was diagnosed or if participant left the GPRD for any reason including death. Participants with prior history of cancer excluded.

Arm/Group Title	Incident Renal Cancer Patients	Matched Controls for Incident Renal Cancer Patients
Arm/Group Description	Patients with incident renal cancer defined from READ/OXMIS codes in the years 1995-2008	Cancer-free control patients risk set matched with incident renal cancer patients for sex, age at cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site
Measure Participants	1272	12167
Never (with 2 year lag)	1263 0.7%	12114 0.7%
Tertile 1 (with 2 year lag)	4 0%	15 0%
Tertile 2 (with 2 year lag)	3 0%	19 0%
Tertile 3 (with 2 year lag)	2 0%	19 0%
Never (without 2 year lag)	1257 0.7%	12060 0.7%
Tertile 1 (without 2 year lag)	5 0%	32 0%
Tertile 2 (without 2 year lag)	6 0%	40 0%
Tertile 3 (without 2 year lag)	4 0%	35 0%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0843
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	2.64
	Confidence Interval	(2-Sided) 95%; 0.88 to 7.96
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0810
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	2.71
	Confidence Interval	(2-Sided) 95%; 0.88 to 8.28
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, hypertension, diuretic use.

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.4583
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.43
	Confidence Interval	(2-Sided) 95%; 0.55 to 3.70
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.4622
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.43
	Confidence Interval	(2-Sided) 95%; 0.55 to 3.73
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 1 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, hypertension, diuretic use.

Statistical Analysis 5

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.5512
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.45
	Confidence Interval	(2-Sided) 95%; 0.43 to 4.95
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 6

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.9277
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.06
	Confidence Interval	(2-Sided) 95%; 0.31 to 3.67
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, hypertension, diuretic use.

Statistical Analysis 7

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.6402
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.23
	Confidence Interval	(2-Sided) 95%; 0.51 to 2.95
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 8

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.9256
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.96
	Confidence Interval	(2-Sided) 95%; 0.40 to 2.32
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 2 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, hypertension, diuretic use.

Statistical Analysis 9

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.9514
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.05
	Confidence Interval	(2-Sided) 95%; 0.24 to 4.50
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)".

Statistical Analysis 10

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.7543
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.79
	Confidence Interval	(2-Sided) 95%; 0.18 to 3.44
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (with 2 year lag)" and "Never (with 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, hypertension, diuretic use.

Statistical Analysis 11

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.8781
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.08
	Confidence Interval	(2-Sided) 95%; 0.38 to 3.07
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)".

Statistical Analysis 12

Statistical Analysis Overview	Comparison Group Selection	Incident All-cancer Patients, Matched Controls for Incident All-cancer Patients
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.7429
	Comments
	Method	Conditional Logistic Regression
	Comments
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	0.84
	Confidence Interval	(2-Sided) 95%; 0.29 to 2.40
	Parameter Dispersion	Type: Value:
	Estimation Comments	Comparison: "Tertile 3 (without 2 year lag)" and "Never (without 2 year lag)". Adjusted for smoking, BMI, diabetes, neuropathic pain, back pain, epilepsy, hypertension, diuretic use.

Adverse Events

Adverse Event Reporting Description	This is a retrospective study of pre-existing medical record and/or health insurance claims data; all data are de-identified, and thus no assessments of Serious or Non-serious Adverse Events are possible.
Arm/Group Title	All-Cancer: Cases		All-Cancer: Controls		Stomach Cancer: Cases		Stomach Cancer: Controls		Anal Cancer: Cases		Anal Cancer: Controls		Lung Cancer: Cases		Lung Cancer: Controls		Bone/Joint Cancer: Cases		Bone/Joint Cancer: Controls		Breast Cancer: Cases		Breast Cancer: Controls		Penile Cancer: Cases		Penile Cancer: Controls		Bladder Cancer: Cases		Bladder Cancer: Controls		Other Nervous System Cancer: Cases		Other Nervous System Cancer: Controls		Pancreatic Cancer: Cases		Pancreatic Cancer: Controls		Renal Cancer: Cases		Renal Cancer: Controls
Arm/Group Description	Incidence of all cancers defined as first time incident cancer diagnosis (READ/Oxford Medical Information System [OXMIS] codes) in the General Practice Research Database (GPRD) study cohort. Entry into the study cohort began January 1, 1993, or at the time of GPRD registration if after January 1, 1993. Follow-up ended December 31, 2008, or earlier if the respective cancer was diagnosed or if the participant left the GPRD for any reason including death. Participants with a prior history of cancer were excluded.		Cancer cases were risk set matched with up to 10 controls for sex, age at GPRD study cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site. The index date for controls was set as the date at which the follow-up time from cohort entry was the same as the case. The index date was chosen so as to give the control equal follow-up time to that of the case for ascertainment of use of gabapentin. Participants with a history of cancer prior to the index date were excluded as controls.		Incidence of stomach cancer defined as first time incident cancer diagnosis (READ/Oxford Medical Information System [OXMIS] codes) in the General Practice Research Database (GPRD) study cohort. Entry into the study cohort began January 1, 1993, or at the time of GPRD registration if after January 1, 1993. Follow-up ended December 31, 2008, or earlier if the respective cancer was diagnosed or if the participant left the GPRD for any reason including death. Participants with a prior history of cancer were excluded.		Cancer cases were risk set matched with up to 10 controls for sex, age at GPRD study cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site. The index date for controls was set as the date at which the follow-up time from cohort entry was the same as the case. The index date was chosen so as to give the control equal follow-up time to that of the case for ascertainment of use of gabapentin. Participants with a history of cancer prior to the index date were excluded as controls.		Incidence of anal cancer defined as first time incident cancer diagnosis (READ/Oxford Medical Information System [OXMIS] codes) in the General Practice Research Database (GPRD) study cohort. Entry into the study cohort began January 1, 1993, or at the time of GPRD registration if after January 1, 1993. Follow-up ended December 31, 2008, or earlier if the respective cancer was diagnosed or if the participant left the GPRD for any reason including death. Participants with a prior history of cancer were excluded.		Cancer cases were risk set matched with up to 10 controls for sex, age at GPRD study cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site. The index date for controls was set as the date at which the follow-up time from cohort entry was the same as the case. The index date was chosen so as to give the control equal follow-up time to that of the case for ascertainment of use of gabapentin. Participants with a history of cancer prior to the index date were excluded as controls.		Incidence of lung cancer defined as first time incident cancer diagnosis (READ/Oxford Medical Information System [OXMIS] codes) in the General Practice Research Database (GPRD) study cohort. Entry into the study cohort began January 1, 1993, or at the time of GPRD registration if after January 1, 1993. Follow-up ended December 31, 2008, or earlier if the respective cancer was diagnosed or if the participant left the GPRD for any reason including death. Participants with a prior history of cancer were excluded.		Cancer cases were risk set matched with up to 10 controls for sex, age at GPRD study cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site. The index date for controls was set as the date at which the follow-up time from cohort entry was the same as the case. The index date was chosen so as to give the control equal follow-up time to that of the case for ascertainment of use of gabapentin. Participants with a history of cancer prior to the index date were excluded as controls.		Incidence of bone/joint cancer defined as first time incident cancer diagnosis (READ/Oxford Medical Information System [OXMIS] codes) in the General Practice Research Database (GPRD) study cohort. Entry into the study cohort began January 1, 1993, or at the time of GPRD registration if after January 1, 1993. Follow-up ended December 31, 2008, or earlier if the respective cancer was diagnosed or if the participant left the GPRD for any reason including death. Participants with a prior history of cancer were excluded.		Cancer cases were risk set matched with up to 10 controls for sex, age at GPRD study cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site. The index date for controls was set as the date at which the follow-up time from cohort entry was the same as the case. The index date was chosen so as to give the control equal follow-up time to that of the case for ascertainment of use of gabapentin. Participants with a history of cancer prior to the index date were excluded as controls.		Incidence of breast cancer defined as first time incident cancer diagnosis (READ/Oxford Medical Information System [OXMIS] codes) in the General Practice Research Database (GPRD) study cohort. Entry into the study cohort began January 1, 1993, or at the time of GPRD registration if after January 1, 1993. Follow-up ended December 31, 2008, or earlier if the respective cancer was diagnosed or if the participant left the GPRD for any reason including death. Participants with a prior history of cancer were excluded.		Cancer cases were risk set matched with up to 10 controls for sex, age at GPRD study cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site. The index date for controls was set as the date at which the follow-up time from cohort entry was the same as the case. The index date was chosen so as to give the control equal follow-up time to that of the case for ascertainment of use of gabapentin. Participants with a history of cancer prior to the index date were excluded as controls.		Incidence of penile cancer defined as first time incident cancer diagnosis (READ/Oxford Medical Information System [OXMIS] codes) in the General Practice Research Database (GPRD) study cohort. Entry into the study cohort began January 1, 1993, or at the time of GPRD registration if after January 1, 1993. Follow-up ended December 31, 2008, or earlier if the respective cancer was diagnosed or if the participant left the GPRD for any reason including death. Participants with a prior history of cancer were excluded.		Cancer cases were risk set matched with up to 10 controls for sex, age at GPRD study cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site. The index date for controls was set as the date at which the follow-up time from cohort entry was the same as the case. The index date was chosen so as to give the control equal follow-up time to that of the case for ascertainment of use of gabapentin. Participants with a history of cancer prior to the index date were excluded as controls.		Incidence of bladder cancer defined as first time incident cancer diagnosis (READ/Oxford Medical Information System [OXMIS] codes) in the General Practice Research Database (GPRD) study cohort. Entry into the study cohort began January 1, 1993, or at the time of GPRD registration if after January 1, 1993. Follow-up ended December 31, 2008, or earlier if the respective cancer was diagnosed or if the participant left the GPRD for any reason including death. Participants with a prior history of cancer were excluded.		Cancer cases were risk set matched with up to 10 controls for sex, age at GPRD study cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site. The index date for controls was set as the date at which the follow-up time from cohort entry was the same as the case. The index date was chosen so as to give the control equal follow-up time to that of the case for ascertainment of use of gabapentin. Participants with a history of cancer prior to the index date were excluded as controls.		Incidence of other nervous system cancer defined as first time incident cancer diagnosis (READ/Oxford Medical Information System [OXMIS] codes) in the General Practice Research Database (GPRD) study cohort. Entry into the study cohort began January 1, 1993, or at the time of GPRD registration if after January 1, 1993. Follow-up ended December 31, 2008, or earlier if the respective cancer was diagnosed or if the participant left the GPRD for any reason including death. Participants with a prior history of cancer were excluded.		Cancer cases were risk set matched with up to 10 controls for sex, age at GPRD study cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site. The index date for controls was set as the date at which the follow-up time from cohort entry was the same as the case. The index date was chosen so as to give the control equal follow-up time to that of the case for ascertainment of use of gabapentin. Participants with a history of cancer prior to the index date were excluded as controls.		Incidence of pancreatic cancer defined as first time incident cancer diagnosis (READ/Oxford Medical Information System [OXMIS] codes) in the General Practice Research Database (GPRD) study cohort. Entry into the study cohort began January 1, 1993, or at the time of GPRD registration if after January 1, 1993. Follow-up ended December 31, 2008, or earlier if the respective cancer was diagnosed or if the participant left the GPRD for any reason including death. Participants with a prior history of cancer were excluded.		Cancer cases were risk set matched with up to 10 controls for sex, age at GPRD study cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site. The index date for controls was set as the date at which the follow-up time from cohort entry was the same as the case. The index date was chosen so as to give the control equal follow-up time to that of the case for ascertainment of use of gabapentin. Participants with a history of cancer prior to the index date were excluded as controls.		Incidence of renal cancer defined as first time incident cancer diagnosis (READ/Oxford Medical Information System [OXMIS] codes) in the General Practice Research Database (GPRD) study cohort. Entry into the study cohort began January 1, 1993, or at the time of GPRD registration if after January 1, 1993. Follow-up ended December 31, 2008, or earlier if the respective cancer was diagnosed or if the participant left the GPRD for any reason including death. Participants with a prior history of cancer were excluded.		Cancer cases were risk set matched with up to 10 controls for sex, age at GPRD study cohort entry (within two years), calendar year of cohort entry (within one year), and General Practice site. The index date for controls was set as the date at which the follow-up time from cohort entry was the same as the case. The index date was chosen so as to give the control equal follow-up time to that of the case for ascertainment of use of gabapentin. Participants with a history of cancer prior to the index date were excluded as controls.
All Cause Mortality
	All-Cancer: Cases		All-Cancer: Controls		Stomach Cancer: Cases		Stomach Cancer: Controls		Anal Cancer: Cases		Anal Cancer: Controls		Lung Cancer: Cases		Lung Cancer: Controls		Bone/Joint Cancer: Cases		Bone/Joint Cancer: Controls		Breast Cancer: Cases		Breast Cancer: Controls		Penile Cancer: Cases		Penile Cancer: Controls		Bladder Cancer: Cases		Bladder Cancer: Controls		Other Nervous System Cancer: Cases		Other Nervous System Cancer: Controls		Pancreatic Cancer: Cases		Pancreatic Cancer: Controls		Renal Cancer: Cases		Renal Cancer: Controls
	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events
Total	/ (NaN)		/ (NaN)		/ (NaN)		/ (NaN)		/ (NaN)		/ (NaN)		/ (NaN)		/ (NaN)		/ (NaN)		/ (NaN)		/ (NaN)		/ (NaN)		/ (NaN)		/ (NaN)		/ (NaN)		/ (NaN)		/ (NaN)		/ (NaN)		/ (NaN)		/ (NaN)		/ (NaN)		/ (NaN)
Serious Adverse Events
	All-Cancer: Cases		All-Cancer: Controls		Stomach Cancer: Cases		Stomach Cancer: Controls		Anal Cancer: Cases		Anal Cancer: Controls		Lung Cancer: Cases		Lung Cancer: Controls		Bone/Joint Cancer: Cases		Bone/Joint Cancer: Controls		Breast Cancer: Cases		Breast Cancer: Controls		Penile Cancer: Cases		Penile Cancer: Controls		Bladder Cancer: Cases		Bladder Cancer: Controls		Other Nervous System Cancer: Cases		Other Nervous System Cancer: Controls		Pancreatic Cancer: Cases		Pancreatic Cancer: Controls		Renal Cancer: Cases		Renal Cancer: Controls
	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events
Total	0/0 (NaN)		0/0 (NaN)		0/0 (NaN)		0/0 (NaN)		0/0 (NaN)		0/0 (NaN)		0/0 (NaN)		0/0 (NaN)		0/0 (NaN)		0/0 (NaN)		0/0 (NaN)		0/0 (NaN)		0/0 (NaN)		0/0 (NaN)		0/0 (NaN)		0/0 (NaN)		0/0 (NaN)		0/0 (NaN)		0/0 (NaN)		0/0 (NaN)		0/0 (NaN)		0/0 (NaN)
Other (Not Including Serious) Adverse Events
	All-Cancer: Cases		All-Cancer: Controls		Stomach Cancer: Cases		Stomach Cancer: Controls		Anal Cancer: Cases		Anal Cancer: Controls		Lung Cancer: Cases		Lung Cancer: Controls		Bone/Joint Cancer: Cases		Bone/Joint Cancer: Controls		Breast Cancer: Cases		Breast Cancer: Controls		Penile Cancer: Cases		Penile Cancer: Controls		Bladder Cancer: Cases		Bladder Cancer: Controls		Other Nervous System Cancer: Cases		Other Nervous System Cancer: Controls		Pancreatic Cancer: Cases		Pancreatic Cancer: Controls		Renal Cancer: Cases		Renal Cancer: Controls
	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events
Total	0/0 (NaN)		0/0 (NaN)		0/0 (NaN)		0/0 (NaN)		0/0 (NaN)		0/0 (NaN)		0/0 (NaN)		0/0 (NaN)		0/0 (NaN)		0/0 (NaN)		0/0 (NaN)		0/0 (NaN)		0/0 (NaN)		0/0 (NaN)		0/0 (NaN)		0/0 (NaN)		0/0 (NaN)		0/0 (NaN)		0/0 (NaN)		0/0 (NaN)		0/0 (NaN)		0/0 (NaN)

Limitations/Caveats

[Not Specified]

More Information

Certain Agreements

Principal Investigators are NOT employed by the organization sponsoring the study.

GSK agreements may vary with individual investigators, but will not prohibit any investigator from publishing. GSK supports the publication of results from all centers of a multi-center trial but requests that reports based on single-site data not precede the primary publication of the entire clinical trial.