A Type 2 Diabetes Study of the Longer-Term Glycemic Effect of AVANDAMET vs. Metformin

• Sponsor: GlaxoSmithKline (Industry)
• Overall Status: Completed
• CT.gov ID: NCT00386100
• Collaborator: (none)
• Enrollment: 688
• Locations: 116
• Arms: 2
• Duration (Months): 35
• Patients Per Site: 5.9
• Patients Per Site Per Month: 0.2

Study Details

Study Description

Brief Summary

This study will evaluate the longer-term glycemic effect of two medicines approved for initial treatment of type 2 diabetes. The study consists of a 2 week screening period (2 study visits), followed by an 80 week double-blind treatment period (11 study visits). Also, a sub-study was included to look at changes in bone mineral density (BMD) at the lumbar spine.

Condition or Disease	Intervention/Treatment	Phase
Diabetes Mellitus, Type 2	Drug: Avandamet 6 mg/1500 mg (ttd); Drug: Avandamet 4 mg/1000 mg (ttd); Drug: Avandamet 2 mg/500 mg (ttd); Drug: Avandamet 8 mg/ 2000 mg (ttd); Drug: Metformin 500 mg (ttd); Drug: Metformin 1000 mg (ttd); Drug: Metformin 1500 mg (ttd); Drug: Metformin 2000 mg (ttd)	Phase 4

Detailed Description

This was a phase IV, randomized, double-blind, global, multi-centre study. The study consisted of a 2 week screening period followed by an 80 week double-blind treatment period. Subjects who met all eligibility requirements were randomized in a 1:1 ratio, stratified by country, gender (male and female) and pre-screening HbA1c (≤9% or>9) either to MET or AVM. When the substudy was added, a new randomization was created for the participating centers. Those subjects in the bone sub-study were stratified by country, gender (male, premenopausal female, and postmenopausal female), pre-screening HbA1c (i.e., ≤9%; >9%), and either to MET or AVM.

At randomization, Visit 3 (Week 0), subjects were initiated at Dose Level 1. Treatment with AVM was initiated at a dose of 4 mg/500 mg and titrated up to a maximum total daily dose of AVM 8 mg/2000 mg. Treatment with MET therapy was initiated at a dose of 500 mg and titrated up to a maximum daily dose of 2000mg.

Study Design

• Study Type: Interventional
• Actual Enrollment: 688 participants
• Allocation: Randomized
• Intervention Model: Parallel Assignment
• Masking: Quadruple (Participant, Care Provider, Investigator, Outcomes Assessor)
• Primary Purpose: Treatment
• Official Title: A Randomized, Parallel Group, Double-blind, Multi-center Study Comparing the Efficacy and Safety of AVANDAMET and Metformin After 80 Weeks of Treatment.
• Study Start Date: Oct 1, 2006
• Actual Primary Completion Date: Sep 1, 2009
• Actual Study Completion Date: Sep 1, 2009

Arms and Interventions

Arm	Intervention/Treatment
Placebo Comparator: Metformin MET began at a total daily dose of 500 mg and could be increased up to a maximum dose of MET 2000 mg. The dose level was to be increased unless a tolerability issue existed at the current dose level.	Drug: Metformin 500 mg (ttd); One placebo capsule will be taken in the AM with the morning meal. One 500 mg capsule will be taken in the PM with the evening meal.; Drug: Metformin 1000 mg (ttd); One 500 mg capsule will be taken in the AM with the morning meal. One 500 mg capsule will be taken in the PM with the evening meal.; Drug: Metformin 1500 mg (ttd); One 500 mg capsule will be taken in the AM with the morning meal. Two 500 mg capsules will be taken in the PM with the evening meal.; Drug: Metformin 2000 mg (ttd); Two 500 mg capsule will be taken in the AM with the morning meal. Two 500 mg capsule will be taken in the PM with the evening meal.
Active Comparator: Avandamet (Rosiglitazone maleate/metformin hydrochloride) AVM began at a total daily dose of 4 mg/500 mg and could be increased up to a maximum dose of AVM 8 mg/2000 mg	Drug: Avandamet 6 mg/1500 mg (ttd); One 2 mg/ 500 mg capsule will be taken in the AM with the morning meal Two 2 mg/ 500 mg capsules will be taken in the PM with the evening meal; Drug: Avandamet 4 mg/1000 mg (ttd); One 2 mg/500 mg capsule will be taken in the AM with the morning meal. One 2 mg/500 mg capsule will be taken in the PM with the evening meal.; Drug: Avandamet 2 mg/500 mg (ttd); one placebo capsule will be taken in the AM with the morning meal one 2 mg/ 500 mg capsule will be taken in the PM with the evening meal.; Drug: Avandamet 8 mg/ 2000 mg (ttd); Two 2 mg/ 500 mg capsules will be taken in the AM with the morning meal. Two 2 mg/ 500 mg capsules will be taken in the PM with the evening meal.

Outcome Measures

Primary Outcome Measures

1. Change From Baseline in HbA1c at Week 80 [Baseline and Week 80]

   Blood was taken for serum HbA1c measurements. Change from baseline was calculated as the Week 80 value minus the baseline value. Last observation carried forward (LOCF) was not used for this analysis.

Secondary Outcome Measures

1. Mean Change From Baseline in HbA1c at Week 80 [Baseline and Week 80]

   Blood was taken for serum Hb1AC measurements. Change from baseline was calculated as the Week 80 value minus the baseline value, with LOCF from Week 32 for withdrawn participants or missing values.

2. Number of Participants Achieving HbA1c <=6.5% and <7% at Week 80 [Week 80]

   Blood was taken for serum Hb1AC measurements. Hb1AC responders were described as participants having achieved Hb1AC <=6% and <7% at Week 80 with LOCF from Week 32.

3. Change in Fasting Plasma Glucose (FPG) From Baseline at Week 80 [Baseline and Week 80]

   Blood was taken for serum FPG measurements. Change from baseline was calculated as the Week 80 value minus the baseline value.

4. Change From Baseline in FPG at Week 80 [Baseline and Week 80]

   Blood was taken for serum FPG measurements. Change from baseline was calculated as the Week 80 value minus the baseline value with LOCF from Week 32 for withdrawn participants or missing values.

5. Number of Participants Achieving FPG <=6 mmol/L (110 mg/dL) and <=7 mmol/L (126 mg/dL) at Week 80 [Week 80]

   Blood was taken for serum FPG measurements. FPG responders were described as participants having achieved FPG <=6 mmol/L (110 mg/dL) and <7 mmol/L (126 mg/dL) Hb1AC at Week 80 with LOCF from Week 32.

6. Number of Participants Achieving Treatment Failure [Randomization to treatment failure (up to Week 80)]

   Treatment failure was defined as an HbA1c level >= 7% after Week 32 or withdrawal due to insufficient therapeutic effect (ITE) at any time.

7. Percent Change From Baseline in Total Cholesterol, Low-density Lipoprotein (LDL) Cholesterol, High-density Lipoprotein (HDL) Cholesterol, and Triglycerides at Week 80 [Baseline and Week 80]

   Blood was taken for measurement of total cholesterol, LDL cholesterol, HDL cholesterol, and triglycerides. Percent change from baseline at Week 80 was based on log transformed data. Geometric mean, GM; standard error, SE. n is the number of evaluable participants, which is the number of participants with a value at baseline and at the specified visit for the parameter of interest.

8. Percent Change From Baseline in Adiponectin at Week 80 (United States [US] and Mexico Subset of Participants ) [Baseline and Week 80]

   Blood was taken for measurement of adiponectin. Percent change from baseline at Week 80 was based on log transformed data. This outcome measure was analyzed for a subset of participants in the US and Mexico only.

9. Percent Change From Baseline in C-reactive Protein (CRP) at Week 80 (US and Mexico Subset of Participants) [Baseline and Week 80]

   Blood was taken for measurement of CRP. Percent change from baseline at Week 80 was based on log transformed data. This outcome measure was analyzed for a subset of participants in the US and Mexico only.

10. Percent Change in Free Fatty Acids (FFA) From Baseline at Week 80 (US and Mexico Subset of Participants). [Baseline and Week 80]

    Blood was taken for measurement of FFA. Percent change from baseline at Week 80 was based on log transformed data. This outcome measure was analyzed for a subset of participants in the US and Mexico only.

11. Change in Fasting Insulin From Baseline at Week 80 (US and Mexico Subset of Participants) [Baseline and Week 80]

    Blood was taken for fasting insulin measurements. Change from baseline was calculated as the Week 80 value minus the baseline value, with LOCF from Week 32 for withdrawn participants or missing values. This outcome measure was analyzed for a subset of participants in the US and Mexico only.

12. Change in C-peptide From Baseline at Week 80 (US and Mexico Subset of Participants) [Baseline and Week 80]

    Blood was taken for C-peptide measurements. Change from baseline was calculated as the Week 80 value minus the baseline value with LOCF from Week 32 for withdrawn participants or missing values. This outcome measure was analyzed for a subset of participants in the US and Mexico only.

13. Percent Change From Baseline in in HOMA-S and HOMA-B to Week 80 (US and Mexico Subset of Participants) [Baseline and Week 80]

    Blood was taken for measurement of homeostasis model assessment for insulin sensitivity (HOMA-S) and beta-cell function (HOMA-B). Percent change from baseline at Week 80 was based on log transformed data. This outcome measure was analyzed for a subset of participants in the US and Mexico only. GM, geometric mean; SE, standard error.

14. Slope of Delta-cell Function as Estimated by the Ratio deltaI/deltaG [Baseline and Week 80]

    The ratio Delta I/Delta G is calculated based on the oral glucose tolerance test (OGTT), where Delta I = (30 minute immunoreactive insulin minus 0 minute immunoreactive insulin) and Delta G = (30 minute plasma glucose minus 0 minute plasma glucose). The 0 minute values are fasting insulin and glucose; the 30 minute values are taken 30 minutes after the oral glucose challenge. This outcome measure was analyzed for a subset of participants in the US and Mexico only.

15. Number of Participants at Final Dose Level [Baseline to Week 80 or withdrawal]

16. Percent Change From Baseline in Lumbar Spine Bone Mass Density (BMD) at Weeks 20, 56, and 80 (Bone Sub-study Subset of Participants) [Baseline and Weeks 20, 56, and 80]

    BMD was measured by dual X-ray absorptiometry (DXA). The percent change from baseline in BMD at a given timepoint was defined at the participant level by the following formula: percent change = (BMD at given week minus BMD at baseline)/BMD at baseline x 100%. This outcome measure was analyzed for a subset of participants in the bone study only.

17. Percent Change From Baseline in Total Hip BMD at Weeks 20, 56, and 80 (Bone Sub-study Subset of Participants) [Baseline and Weeks 20, 56, and 80]

    BMD was measured by dual X-ray absorptiometry (DXA). The percent change from baseline in BMD at a given timepoint was defined at the participant level by the following formula: percent change = (BMD at given week minus BMD at baseline)/BMD at baseline x 100. This outcome measure was analyzed for a subset of participants in the bone study only.

18. Percent Change From Baseline in Trochanter BMD at Weeks 20, 56, and 80 (Bone Sub-study Subset of Participants) [Baseline and Weeks 20, 56, and 80]

    BMD was measured by dual X-ray absorptiometry (DXA). The percent change from baseline in BMD at a given timepoint was defined at the participant level by the following formula: percent change = (BMD at given week minus BMD at baseline)/BMD at baseline x 100. This outcome measure was analyzed for a subset of participants in the bone study only.

19. Percent Change From Baseline in Femoral Neck BMD at Weeks 20, 56, and 80 (Bone Sub-study Subset of Participants) [Baseline and Weeks 20, 56, and 80]

    BMD was measured by dual X-ray absorptiometry (DXA). The percent change from baseline in BMD at a given timepoint was defined at the participant level by the following formula: percent change = (BMD at given week minus BMD at baseline)/BMD at baseline x 100. This outcome measure was analyzed for a subset of participants in the bone study only.

20. Percent Change From Baseline in Distal Radius BMD at Weeks 20, 56, and 80 (Bone Sub-study Subset of Participants) [Baseline and Weeks 20, 56, and 80]

    BMD was measured by dual X-ray absorptiometry (DXA). The percent change from baseline in BMD at a given timepoint was defined at the participant level by the following formula: percent change = (BMD at given week minus BMD at baseline)/BMD at baseline x 100. This outcome measure was analyzed for a subset of participants in the bone study only.

21. Percent Change From Baseline in Total Body BMD at Weeks 20, 56, and 80 (Bone Sub-study Subset of Participants) [Baseline and Weeks 20, 56, and 80]

    BMD was measured by dual X-ray absorptiometry (DXA). The percent change from baseline in BMD at a given timepoint was defined at the participant level by the following formula: percent change = (BMD at given week minus BMD at baseline)/BMD at baseline x 100. This outcome measure was analyzed for a subset of participants in the bone study only.

22. Percent Change From Baseline in Serum Calcium at Weeks 12, 32, 56, and 80 [Baseline and Weeks 12, 32, 56, and 80]

    Blood was taken for measurement of serum calcium. Percent change from baseline was based on log transformed data. Geometric mean, GM; standard error, SE. This outcome measure was analyzed for a subset of participants in the bone study only. n is the number of evaluable participants, which is the number of participants with a value at baseline and at the specified visit for the parameter of interest.

23. Percent Change From Baseline in Intact Parathyroid Hormone at Week 80 [Baseline and Week 80]

    Blood was taken for measurement of intact parathyroid hormone. Percent change from baseline was based on log transformed data. Standard error, SE; Wk, Week; %, percent. This outcome measure was analyzed for a subset of participants in the bone study only. n is the number of evaluable participants, which is the number of participants with a value at baseline and at the specified visit for the parameter of interest.

24. Percent Change From Baseline in 25-hydroxy Vitamin D at Week 80 [Baseline and Week 80]

    Blood was taken for measurement of 25-hydroxy vitamin D. Percent change from baseline was based on log transformed data. Standard error, SE; Wk, Week; %, percent. This outcome measure was analyzed for a subset of participants in the bone study only. n is the number of evaluable participants, which is the number of participants with a value at baseline and at the specified visit for the parameter of interest.

25. Percent Change From Baseline in Estradiol at Weeks 20, 56, and 80 [Baseline and Weeks 20, 56, and 80]

    Blood was taken for measurement of estradiol. Percent change from baseline was based on log transformed data. Standard error, SE; Wk, Week; %, percent. This outcome measure was analyzed for a subset of female participants in the bone study only. n is the number of evaluable participants, which is the number of female participants with a value at baseline and at the specified visit for the parameter of interest.

26. Percent Change From Baseline in C-terminal Telopeptide (CTX) at Weeks 20, 56, and 80 [Baseline and Weeks 20, 56, and 80]

    Blood was taken for measurement of CTX. Percent change from baseline was based on log transformed data. Standard error, SE; Wk, Week; %, percent. This outcome measure was analyzed for a subset of participants in the bone study only. n is the number of evaluable participants, which is the number of participants with a value at baseline and at the specified visit for the parameter of interest.

27. Percent Change From Baseline in Procollagen Type-1 N-propeptide (P1NP) at Weeks 20, 56, and 80 [Baseline and Weeks 20, 56, and 80]

    Blood was taken for measurement of P1NP. Percent change from baseline was based on log transformed data. Standard error, SE; Wk, Week; %, percent. This outcome measure was analyzed for a subset of participants in the bone study only. n is the number of evaluable participants, which is the number of participants with a value at baseline and at the specified visit for the parameter of interest.

28. Percent Change From Baseline in Bone Alkaline Phosphatase (BSAP) at Weeks 20, 56, and 80 [Baseline and Weeks 20, 56, and 80]

    Blood was taken for measurement of BSAP. Percent change from baseline was based on log transformed data. Standard error, SE; Wk, Week; %, percent. This outcome measure was analyzed for a subset of participants in the bone study only. n is the number of evaluable participants, which is the number of participants with a value at baseline and at the specified visit for the parameter of interest.

Eligibility Criteria

Criteria

• Ages Eligible for Study: 18 Years to 75 Years
• Sexes Eligible for Study: All
• Accepts Healthy Volunteers: No

Inclusion Criteria:

• The subject provides written informed consent.

• The subject is male or female and 18 to 75 years of age at the time of pre-screening.

• The subject has an established clinical diagnosis of type 2 diabetes according to recommended guidelines (e.g., American Diabetes Association, International Diabetes Federation, World Health Organization, Canadian Diabetes Association, or American Association of Clinical Endocrinologists).

• The subject is currently treated with diet and exercise, and has not taken more than 2 weeks of an anti-diabetic monotherapy or insulin in the past 6 months.

• The subject has a BMI >25 kg/m2 at pre-screening.

• The subject has a Quest HbA1c 7.5% to 10.5% at pre-screening.

• The subject has a fasting capillary blood glucose 126 mg/dL (7mmol/L), as measured by the site staff at week 0.

• If the subject is a pre-menopausal female of child-bearing potential, she agrees to practice acceptable contraceptive measures (e.g. oral birth control pills, Norplant, Depo-Provera, an intrauterine device (IUD), a diaphragm with spermicide or a condom with spermicide, or abstinence) at least 1 month before screening, during the study, and for 30 days after the last dose of study medication is taken

• The subject is able and willing to perform self-monitoring of blood glucose as specified in this protocol.

Exclusion Criteria:

• The subject has taken an oral anti-diabetic monotherapy or insulin for more than 14 days in the past 6 months.

• The subject has presence of clinically significant renal or hepatic disease (serum creatinine 1.5 mg/dL (132.6 mol/L) for males and 1.4 mg/dL (123.8 mol/L) for females): ALT, AST, total bilirubin, or alkaline phosphatase >2.5 times the upper limit of the normal (ULN) reference range.

• The subject has anemia defined by hemoglobin concentration <11g/dL (110g/L) for males or <10g/dL (100g/L) for females.

• Presence of unstable or severe angina, coronary insufficiency or New York Heart Association (NYHA) class III-IV or any congestive heart failure requiring pharmacologic treatment.

• The subject has systolic blood pressure >160 mmHg or diastolic blood pressure >90 mmHg

• The subject has a chronic disease requiring intermittent or chronic treatment with oral, intravenous, or intra-articular corticosteroids (i.e., only use of topical, inhaled or nasal corticosteroids is permitted).

• The subject has acute or chronic metabolic acidosis or a history of diabetic ketoacidosis.

• The subject has a clinically significant abnormality which in the judgment of the investigator makes the subject unsuitable for inclusion in the study (e.g., physical examination, laboratory tests, or electrocardiogram, etc).

• The subject has used an investigational agent within 30 days or 5 half-lives (whichever was longer) prior to pre-screening.

• The subject is a female who is lactating, pregnant, or planned to become pregnant.

• The subject has a prior history of severe edema or a medically serious fluid related event (e.g., heart failure).

• The subject has a history of macular edema.

• The subject has significant hypersensitivity (e.g., difficulty swallowing, difficulty breathing, and tachycardia or skin reaction) to TZDs, biguanides, or compounds with similar chemical structures.

• The subject has a diagnosis of cancer (other than squamous, basal cell, or cervical cancer in-situ) in the past 3 years and is receiving treatment for cancer.

• The subject has a history or suspicion of drug abuse or alcohol abuse within the last 6 months.

• The subject is known to have severe lactose intolerance.

• The subject is not willing to comply with visits and procedures described in the protocol.

• The subject has a disease that may affect bone turnover including, but not limited to: Paget's disease, hypercalcemia, hypocalcemia, hyperparathyroidism, hyperthyroidism, osteomalacia, metastatic bone disease

• The subject has a weight of greater than 300 lbs (136.4 kg).

• The subject has received treatment with bisphosphonates (≥1 month cumulative treatment within the last 12 months) or fluoride (dose greater than 10mg/day within the previous 5 years).

Contacts and Locations

Locations

	Site	City	State	Country	Postal Code
1	GSK Investigational Site	Tuscaloosa	Alabama	United States	35406
2	GSK Investigational Site	Gilbert	Arizona	United States	85296
3	GSK Investigational Site	Glendale	Arizona	United States	85308
4	GSK Investigational Site	Phoenix	Arizona	United States	85020
5	GSK Investigational Site	Tucson	Arizona	United States	85712
6	GSK Investigational Site	Tucson	Arizona	United States	85745
7	GSK Investigational Site	Alhambra	California	United States	91801
8	GSK Investigational Site	Artesia	California	United States	90701
9	GSK Investigational Site	Greenbrae	California	United States	94904
10	GSK Investigational Site	Roseville	California	United States	95661
11	GSK Investigational Site	Sacramento	California	United States	95825
12	GSK Investigational Site	Wheat Ridge	Colorado	United States	80033
13	GSK Investigational Site	Hialeah	Florida	United States	33013
14	GSK Investigational Site	Ocala	Florida	United States	34471
15	GSK Investigational Site	Kahului	Hawaii	United States	96732
16	GSK Investigational Site	Peoria	Illinois	United States	61615
17	GSK Investigational Site	Avon	Indiana	United States	46123
18	GSK Investigational Site	Evansville	Indiana	United States	47710
19	GSK Investigational Site	Evansville	Indiana	United States	47712
20	GSK Investigational Site	Indianapolis	Indiana	United States	46254
21	GSK Investigational Site	Waterloo	Iowa	United States	50702
22	GSK Investigational Site	Slidell	Louisiana	United States	70461
23	GSK Investigational Site	Sunset	Louisiana	United States	70584
24	GSK Investigational Site	Elkridge	Maryland	United States	21075
25	GSK Investigational Site	Chaska	Minnesota	United States	55318
26	GSK Investigational Site	Minneapolis	Minnesota	United States	55407-3799
27	GSK Investigational Site	Excelsior Springs	Missouri	United States	64024
28	GSK Investigational Site	St. Louis	Missouri	United States	63110
29	GSK Investigational Site	St. Louis	Missouri	United States	63128
30	GSK Investigational Site	St. Peters	Missouri	United States	63376
31	GSK Investigational Site	Billings	Montana	United States	59102
32	GSK Investigational Site	Las Vegas	Nevada	United States	89016
33	GSK Investigational Site	Pahrump	Nevada	United States	89048
34	GSK Investigational Site	Hamilton	New Jersey	United States	08690
35	GSK Investigational Site	Albuquerque	New Mexico	United States	87102
36	GSK Investigational Site	East Syracuse	New York	United States	13057
37	GSK Investigational Site	Flushing	New York	United States	11355
38	GSK Investigational Site	Kingston	New York	United States	12401
39	GSK Investigational Site	Huntersville	North Carolina	United States	28078
40	GSK Investigational Site	Canal Fulton	Ohio	United States	44614
41	GSK Investigational Site	Canton	Ohio	United States	44718
42	GSK Investigational Site	Cleveland	Ohio	United States	44195
43	GSK Investigational Site	Columbus	Ohio	United States	43210
44	GSK Investigational Site	Kettering	Ohio	United States	45429
45	GSK Investigational Site	Mogadore	Ohio	United States	44260
46	GSK Investigational Site	Wandsworth	Ohio	United States	44281
47	GSK Investigational Site	Oregon City	Oregon	United States	97045
48	GSK Investigational Site	Beaver	Pennsylvania	United States	15009
49	GSK Investigational Site	Clairton	Pennsylvania	United States	15205
50	GSK Investigational Site	Coatsville	Pennsylvania	United States	19320
51	GSK Investigational Site	Erie	Pennsylvania	United States	16508
52	GSK Investigational Site	Sewickley	Pennsylvania	United States	15143
53	GSK Investigational Site	West Chester	Pennsylvania	United States	19382
54	GSK Investigational Site	Clinton	South Carolina	United States	29325
55	GSK Investigational Site	Columbia	South Carolina	United States	29201
56	GSK Investigational Site	Pelzer	South Carolina	United States	29669
57	GSK Investigational Site	Kingsport	Tennessee	United States	37660
58	GSK Investigational Site	Corpus Christi	Texas	United States	78404
59	GSK Investigational Site	Dallas	Texas	United States	75235
60	GSK Investigational Site	Georgetown	Texas	United States	78626
61	GSK Investigational Site	South Burlington	Vermont	United States	05403
62	GSK Investigational Site	Burke	Virginia	United States	22015
63	GSK Investigational Site	Manassas	Virginia	United States	20110
64	GSK Investigational Site	Salem	Virginia	United States	24153
65	GSK Investigational Site	Gig Harbor	Washington	United States	98335
66	GSK Investigational Site	Graham	Washington	United States	98338
67	GSK Investigational Site	Olympia	Washington	United States	98506
68	GSK Investigational Site	Tacoma	Washington	United States	98405
69	GSK Investigational Site	Vancouver	Washington	United States	98664
70	GSK Investigational Site	Wenatchee	Washington	United States	98801
71	GSK Investigational Site	Wauwatosa	Wisconsin	United States	53228
72	GSK Investigational Site	Buenos Aries	Buenos Aires	Argentina	C1425AWC
73	GSK Investigational Site	Capital Federal	Buenos Aires	Argentina	C1416DRJ
74	GSK Investigational Site	Ciudad Autonoma de Buenos Aires	Buenos Aires	Argentina	B1704ETD
75	GSK Investigational Site	Ciudad Autonoma de Buenos Aires	Buenos Aires	Argentina	C1155ADP
76	GSK Investigational Site	Cordoba	Córdova	Argentina	5000
77	GSK Investigational Site	Buenos Aires		Argentina	1425
78	GSK Investigational Site	Ciudad Autónoma de Buenos Aires		Argentina	C1117ABH
79	GSK Investigational Site	Mendoza		Argentina	5500
80	GSK Investigational Site	Fortaleza	Ceará	Brazil	60120-021
81	GSK Investigational Site	Goiânia	Goiás	Brazil	74110-010
82	GSK Investigational Site	Porto Alegre	Rio Grande Do Sul	Brazil	90035-170
83	GSK Investigational Site	Campinas	São Paulo	Brazil	13073-350
84	GSK Investigational Site	Brasília		Brazil	71625-009
85	GSK Investigational Site	São Paulo		Brazil	01323-001
86	GSK Investigational Site	Coquitlam	British Columbia	Canada	V3K 3P4
87	GSK Investigational Site	Bathurst	New Brunswick	Canada	E2A 4X7
88	GSK Investigational Site	Bay Roberts	Newfoundland and Labrador	Canada	A0A 1G0
89	GSK Investigational Site	St. John's	Newfoundland and Labrador	Canada	A1E 2C2
90	GSK Investigational Site	Brampton	Ontario	Canada	L6T 3T1
91	GSK Investigational Site	Smiths Falls	Ontario	Canada	K7A 4W8
92	GSK Investigational Site	Toronto	Ontario	Canada	M9W 4L6
93	GSK Investigational Site	Gatineau	Quebec	Canada	J8Y 6S8
94	GSK Investigational Site	Sherbrooke	Quebec	Canada	J1H 4J6
95	GSK Investigational Site	Gwangju		Korea, Republic of	501-757
96	GSK Investigational Site	Seoul		Korea, Republic of	110-749
97	GSK Investigational Site	Seoul		Korea, Republic of	139-872
98	GSK Investigational Site	Seoul		Korea, Republic of	152-703
99	GSK Investigational Site	Suwon, Kyonggi-do		Korea, Republic of	443-721
100	GSK Investigational Site	Uijeongbu-si, Kyonggi-do		Korea, Republic of	480-130
101	GSK Investigational Site	Tijuana	Baja California Norte	Mexico	22320
102	GSK Investigational Site	Pachuca	Hidalgo	Mexico	42039
103	GSK Investigational Site	Monterrey	Nuevo León	Mexico	64460
104	GSK Investigational Site	Durango		Mexico	34070
105	GSK Investigational Site	Karachi		Pakistan	74800
106	GSK Investigational Site	Lahore		Pakistan	54000
107	GSK Investigational Site	Cebu City		Philippines	6000
108	GSK Investigational Site	Manila		Philippines	1000
109	GSK Investigational Site	Manila		Philippines	1008
110	GSK Investigational Site	Marikina City		Philippines	1810
111	GSK Investigational Site	Quezon City		Philippines	1108
112	GSK Investigational Site	Changhua		Taiwan	500
113	GSK Investigational Site	Kaohsiung		Taiwan	833
114	GSK Investigational Site	Taichung		Taiwan	404
115	GSK Investigational Site	Taipei		Taiwan	114
116	GSK Investigational Site	Taoyuan Hsien		Taiwan	333

Sponsors and Collaborators

• GlaxoSmithKline

Investigators

• Study Director: GSK Clinical Trials, GlaxoSmithKline

Study Documents (Full-Text)

None provided.

More Information

Additional Information:

• Researchers can use this site to request access to anonymised patient level data and/or supporting documents from clinical studies to conduct further research.

Publications

• Borges JL, Bilezikian JP, Jones-Leone AR, Acusta AP, Ambery PD, Nino AJ, Grosse M, Fitzpatrick LA, Cobitz AR. A randomized, parallel group, double-blind, multicentre study comparing the efficacy and safety of Avandamet (rosiglitazone/metformin) and metformin on long-term glycaemic control and bone mineral density after 80 weeks of treatment in drug-naïve type 2 diabetes mellitus patients. Diabetes Obes Metab. 2011 Nov;13(11):1036-46. doi: 10.1111/j.1463-1326.2011.01461.x.

• Responsible Party: GlaxoSmithKline
• ClinicalTrials.gov Identifier: NCT00386100
• Other Study ID Numbers: AVT105913
• First Posted: Oct 11, 2006
• Last Update Posted: Nov 23, 2016
• Last Verified: Oct 1, 2016

• Individual Participant Data (IPD) Sharing Statement: Yes
• Plan to Share IPD: Yes
• Keywords provided by GlaxoSmithKline: Fasting Plasma Glucose; Dual energy X ray absorptiometry (DXA); Drug-naive; Type 2 diabetes mellitus; Bone Mineral Density; Hyperglycemia; HbA1c
• Additional relevant MeSH terms: Diabetes Mellitus; Diabetes Mellitus, Type 2; Metformin

Study Results

Participant Flow

Arm/Group Title	Metformin	Avandamet
Arm/Group Description	Beginning dose of Metformin (MET) 500 milligrams (mg) once a day. Dose could be increased up to a maximum dose of MET 2000 mg.	Beginning dose of Avandamet (AVM) 4 mg/500 mg once a day. Dose could be increased up to a maximum dose of 8 mg/2000 mg.
Period Title: Overall Study
STARTED	340	348
COMPLETED	186	217
NOT COMPLETED	154	131

Baseline Characteristics

Arm/Group Title	Metformin	Avandamet	Total
Arm/Group Description	Beginning dose of Metformin (MET) 500 milligrams (mg) once a day. Dose could be increased up to a maximum dose of MET 2000 mg.	Beginning dose of Avandamet (AVM) 4 mg/500 mg once a day. Dose could be increased up to a maximum dose of 8 mg/2000 mg.	Total of all reporting groups
Overall Participants	334	344	678
Age (years) [Mean (Standard Deviation) ]
Mean (Standard Deviation) [years]	50.7 (10.49)	51.5 (10.52)	51.1 (10.50)
Sex: Female, Male (Count of Participants)
Female	158 47.3%	160 46.5%	318 46.9%
Male	176 52.7%	184 53.5%	360 53.1%
Race/Ethnicity, Customized (participants) [Number]
African America/African	13 3.9%	17 4.9%	30 4.4%
American Indian or Alaskan Native	3 0.9%	7 2%	10 1.5%
Asian - Central/South Asian	23 6.9%	22 6.4%	45 6.6%
Asian - East Asian	44 13.2%	51 14.8%	95 14%
Asian - Japanese	0 0%	1 0.3%	1 0.1%
Asian - South East Asian	48 14.4%	48 14%	96 14.2%
White - Arabic/North African	4 1.2%	3 0.9%	7 1%
White - White/Caucasian/European	185 55.4%	183 53.2%	368 54.3%
Mixed Race	5 1.5%	5 1.5%	10 1.5%
Not Specified	9 2.7%	7 2%	16 2.4%
Duration of Diabetes (years) [Mean (Standard Deviation) ]
Mean (Standard Deviation) [years]	2.570 (3.2671)	2.293 (3.0774)	2.429 (3.1729)

Outcome Measures

1. Secondary Outcome

Title	Mean Change From Baseline in HbA1c at Week 80
Description	Blood was taken for serum Hb1AC measurements. Change from baseline was calculated as the Week 80 value minus the baseline value, with LOCF from Week 32 for withdrawn participants or missing values.
Time Frame	Baseline and Week 80

Outcome Measure Data

Analysis Population Description

Week 32 Evaluable Population with LOCF: a subset of the ITT Population with LOCF starting at Week 32. Only participants with assessment(s) at Week 32 or later were included in this population. Only evaluable participants, defined as participants with a value at baseline and at the specified visit for the parameter of interest, were analyzed.

Arm/Group Title	Metformin	Avandamet
Arm/Group Description	Beginning dose of Metformin (MET) 500 milligrams (mg) once a day. Dose could be increased up to a maximum dose of MET 2000 mg.	Beginning dose of Avandamet (AVM) 4 mg/500 mg once a day. Dose could be increased up to a maximum dose of 8 mg/2000 mg.
Measure Participants	258	265
Baseline	8.59 (0.898)	8.66 (0.954)
Change from Baseline	-1.42 (0.070)	-1.91 (0.069)

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	<0.0001
	Comments	No adjustment for multiple comparisons
	Method	ANCOVA
	Comments	ANCOVA with terms for treatment, region, gender, and baseline value with LOCF from Week 32 for withdrawn participants or missing values
Method of Estimation	Estimation Parameter	Mean Difference (Net)
	Estimated Value	-0.49
	Confidence Interval	(2-Sided) 95%; -0.669 to -0.305
	Parameter Dispersion	Type: Value:
	Estimation Comments	AVM mean change from baseline minus MET mean change from baseline based on ANCOVA model

2. Primary Outcome

Title	Change From Baseline in HbA1c at Week 80
Description	Blood was taken for serum HbA1c measurements. Change from baseline was calculated as the Week 80 value minus the baseline value. Last observation carried forward (LOCF) was not used for this analysis.
Time Frame	Baseline and Week 80

Outcome Measure Data

Analysis Population Description

Intent-to-Treat (ITT) Population: all participants who were randomized, received at least one dose of study medication, and had both a baseline and at least one on-therapy value. Only evaluable participants, defined as the number of participants with a baseline and at least one post baseline assessment, were analyzed.

Arm/Group Title	Metformin	Avandamet
Arm/Group Description	Beginning dose of Metformin (MET) 500 milligrams (mg) once a day. Dose could be increased up to a maximum dose of MET 2000 mg.	Beginning dose of Avandamet (AVM) 4 mg/500 mg once a day. Dose could be increased up to a maximum dose of 8 mg/2000 mg.
Measure Participants	294	301
Baseline	8.64 (0.053)	8.64 (0.055)
Change from Baseline	-1.36 (0.073)	-1.85 (0.070)

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	<0.0001
	Comments	No adjustment for multiple comparisons
	Method	Repeated measures analysis
	Comments	Repeated measures analysis with terms for baseline, region, treatment, gender, time, and treatment by time interaction
Method of Estimation	Estimation Parameter	Mean Difference (Net)
	Estimated Value	-0.50
	Confidence Interval	(2-Sided) 95%; -0.69 to -0.30
	Parameter Dispersion	Type: Value:
	Estimation Comments	AVM mean change from baseline minus MET mean change from baseline based on repeated measures analysis model

3. Secondary Outcome

Title	Number of Participants Achieving HbA1c <=6.5% and <7% at Week 80
Description	Blood was taken for serum Hb1AC measurements. Hb1AC responders were described as participants having achieved Hb1AC <=6% and <7% at Week 80 with LOCF from Week 32.
Time Frame	Week 80

Outcome Measure Data

Analysis Population Description

Week 32 Evaluable Population with LOCF. Only evaluable participants, defined as participants with a value at baseline and at the specified visit for the parameter of interest, were analyzed.

Arm/Group Title	Metformin	Avandamet
Arm/Group Description	Beginning dose of Metformin (MET) 500 milligrams (mg) once a day. Dose could be increased up to a maximum dose of MET 2000 mg.	Beginning dose of Avandamet (AVM) 4 mg/500 mg once a day. Dose could be increased up to a maximum dose of 8 mg/2000 mg.
Measure Participants	258	265
<=6.5%	97 29%	128 37.2%
<7%	133 39.8%	184 53.5%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0046
	Comments	Hb1AC <= 6.5%
	Method	Regression, Logistic
	Comments	Logistic reggression with terms for treatment, region, gender, and baseline Hb1AC with LOCF from Week 32.
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	1.18
	Confidence Interval	(2-Sided) 95%; 1.18 to 2.46
	Parameter Dispersion	Type: Value:
	Estimation Comments	Odds of having an HbA1c <= 6.5% at Week 80 on Avandamet compared to Metformin.

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	<0.0001
	Comments	Hb1AC < 7%
	Method	Regression, Logistic
	Comments	Logistic reggression with terms for treatment, region, gender, and baseline Hb1AC with LOCF from Week 32
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	2.59
	Confidence Interval	(2-Sided) 95%; 1.75 to 3.81
	Parameter Dispersion	Type: Value:
	Estimation Comments	Odds of having an HbA1c <7% at Week 80 on Avandamet compared to Metformin

4. Secondary Outcome

Title	Change in Fasting Plasma Glucose (FPG) From Baseline at Week 80
Description	Blood was taken for serum FPG measurements. Change from baseline was calculated as the Week 80 value minus the baseline value.
Time Frame	Baseline and Week 80

Outcome Measure Data

Analysis Population Description

ITT Population. Only evaluable participants, defined as the number of participants with a baseline and at least one post baseline assessment, were analyzed. Last observation carried forward (LOCF) was not used for this analysis.

Arm/Group Title	Metformin	Avandamet
Arm/Group Description	Beginning dose of Metformin (MET) 500 milligrams (mg) once a day. Dose could be increased up to a maximum dose of MET 2000 mg.	Beginning dose of Avandamet (AVM) 4 mg/500 mg once a day. Dose could be increased up to a maximum dose of 8 mg/2000 mg.
Measure Participants	319	332
Baseline	10.52 (0.184)	10.17 (0.162)
Change from Baseline	-2.25 (0.147)	-3.41 (0.139)

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	<0.0001
	Comments
	Method	Repeated measures analysis
	Comments	Terms for baseline, region, treatment, pre-screening Hb1Ac strata, gender, time, and treatment by time interaction
Method of Estimation	Estimation Parameter	Mean Difference (Net)
	Estimated Value	-1.15
	Confidence Interval	() 95%; -1.54 to -0.77
	Parameter Dispersion	Type: Value:
	Estimation Comments	AVM mean change from baseline minus MET mean change from baseline based on repeated measures analysis model

5. Secondary Outcome

Title	Change From Baseline in FPG at Week 80
Description	Blood was taken for serum FPG measurements. Change from baseline was calculated as the Week 80 value minus the baseline value with LOCF from Week 32 for withdrawn participants or missing values.
Time Frame	Baseline and Week 80

Outcome Measure Data

Analysis Population Description

Week 32 Evaluable Population with LOCF. Only evaluable participants, defined as participants with a value at baseline and at the specified visit for the parameter of interest, were analyzed.

Arm/Group Title	Metformin	Avandamet
Arm/Group Description	Beginning dose of Metformin (MET) 500 milligrams (mg) once a day. Dose could be increased up to a maximum dose of MET 2000 mg.	Beginning dose of Avandamet (AVM) 4 mg/500 mg once a day. Dose could be increased up to a maximum dose of 8 mg/2000 mg.
Measure Participants	258	265
Baseline	10.37 (3.268)	10.13 (2.927)
Change from Baseline	-2.53 (3.165)	-3.39 (2.908)

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	<0.001
	Comments	No adjustment for multiple comparisons
	Method	ANCOVA
	Comments	Terms for treatment, region, gender, pre-screening Hb1Ac strata, and baseline with LOCF from Week 32 for withdrawn participants or missing values
Method of Estimation	Estimation Parameter	Median Difference (Net)
	Estimated Value	-1.07
	Confidence Interval	(2-Sided) 95%; -1.425 to -0.710
	Parameter Dispersion	Type: Value:
	Estimation Comments	AVM mean change from baseline minus MET mean change from baseline based on ANCOVA model

6. Secondary Outcome

Title	Number of Participants Achieving FPG <=6 mmol/L (110 mg/dL) and <=7 mmol/L (126 mg/dL) at Week 80
Description	Blood was taken for serum FPG measurements. FPG responders were described as participants having achieved FPG <=6 mmol/L (110 mg/dL) and <7 mmol/L (126 mg/dL) Hb1AC at Week 80 with LOCF from Week 32.
Time Frame	Week 80

Outcome Measure Data

Analysis Population Description

Week 32 Evaluable Population with LOCF. Only evaluable participants, defined as participants with a value at baseline and at the specified visit for the parameter of interest, were analyzed.

Arm/Group Title	Metformin	Avandamet
Arm/Group Description	Beginning dose of Metformin (MET) 500 milligrams (mg) once a day. Dose could be increased up to a maximum dose of MET 2000 mg.	Beginning dose of Avandamet (AVM) 4 mg/500 mg once a day. Dose could be increased up to a maximum dose of 8 mg/2000 mg.
Measure Participants	258	265
FPG <=6.1 mmol/l	31 9.3%	83 24.1%
FPG <=7.0 mmol/l	55 16.5%	133 38.7%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	<0.0001
	Comments	FPG <=6.1 mmol/l
	Method	Regression, Logistic
	Comments	Terms for treatment, region, gender, pre-screening Hb1Ac strata, and baseline with LOCF from Week 32.
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	4.47
	Confidence Interval	(2-Sided) 95%; 2.94 to 6.81
	Parameter Dispersion	Type: Value:
	Estimation Comments	Odds of having an FPG <=6.1 mmol/l at Week 80 on Avandamet compared to Metformin

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	<0.0001
	Comments	FPG <=7 mmol/l
	Method	Regression, Logistic
	Comments	Terms for treatment, region, gender, pre-screening Hb1Ac, and baseline with LOCF from Week 32.
Method of Estimation	Estimation Parameter	Odds Ratio (OR)
	Estimated Value	3.33
	Confidence Interval	(2-Sided) 95%; 2.25 to 4.92
	Parameter Dispersion	Type: Value:
	Estimation Comments	Odds of having an FPG <=7 mmol/l at Week 80 on Avandamet compared to Metformin

7. Secondary Outcome

Title	Number of Participants Achieving Treatment Failure
Description	Treatment failure was defined as an HbA1c level >= 7% after Week 32 or withdrawal due to insufficient therapeutic effect (ITE) at any time.
Time Frame	Randomization to treatment failure (up to Week 80)

Outcome Measure Data

Analysis Population Description

ITT Population. Only evaluable participants, defined as the number of subjects with a baseline and at least one post baseline assessment, were analyzed. Last observation carried forward (LOCF) was not used for this analysis

Arm/Group Title	Metformin	Avandamet
Arm/Group Description	Beginning dose of Metformin (MET) 500 milligrams (mg) once a day. Dose could be increased up to a maximum dose of MET 2000 mg.	Beginning dose of Avandamet (AVM) 4 mg/500 mg once a day. Dose could be increased up to a maximum dose of 8 mg/2000 mg.
Measure Participants	319	332
Number [participants]	156 46.7%	114 33.1%

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	<0.0001
	Comments
	Method	Regression, Cox
	Comments	Cox proportional hazard regression with terms for treatment, region, baseline Hb1Ac strata, and gender
Method of Estimation	Estimation Parameter	Hazard Ratio (HR)
	Estimated Value	0.58
	Confidence Interval	(2-Sided) 95%; 0.45 to 0.74
	Parameter Dispersion	Type: Value:
	Estimation Comments

8. Secondary Outcome

Title	Percent Change From Baseline in Total Cholesterol, Low-density Lipoprotein (LDL) Cholesterol, High-density Lipoprotein (HDL) Cholesterol, and Triglycerides at Week 80
Description	Blood was taken for measurement of total cholesterol, LDL cholesterol, HDL cholesterol, and triglycerides. Percent change from baseline at Week 80 was based on log transformed data. Geometric mean, GM; standard error, SE. n is the number of evaluable participants, which is the number of participants with a value at baseline and at the specified visit for the parameter of interest.
Time Frame	Baseline and Week 80

Outcome Measure Data

Analysis Population Description

Week 32 evaluable population with LOCF from Week 32. n is the number of evaluable participants, which is defined as the number of participants with a value at baseline and at the specified visit for the parameter of interest.

Arm/Group Title	Metformin	Avandamet
Arm/Group Description	Beginning dose of Metformin (MET) 500 milligrams (mg) once a day. Dose could be increased up to a maximum dose of MET 2000 mg.	Beginning dose of Avandamet (AVM) 4 mg/500 mg once a day. Dose could be increased up to a maximum dose of 8 mg/2000 mg.
Measure Participants	258	265
% change + SE, Total cholesterol, n=258, 265	-1.615	4.218
% change, Total cholesterol, n=258, 265	-2.89	2.91
% change - SE, Total cholesterol, n=258, 265	-4.140	1.614
% change + SE, LDL cholesterol, n=248, 252	-4.933	3.304
% change, LDL cholesterol, n=248, 252	-7.11	0.98
% change - SE, LDL cholesterol, n=248, 252	-9.228	-1.286
% change + SE, HDL cholesterol, n=258, 265	6.761	9.483
% change, HDL cholesterol, n=258, 265	5.60	8.324
% change - SE, HDL cholesterol, n=258, 265	4.453	7.179
% change + SE, triglycerides, n=258, 265	-2.696	-2.042
% change, triglycerides, n=258, 265	-5.330	-4.627
% change - SE, triglycerides, n=258, 265	-7.893	-7.143

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0006
	Comments	Total cholesterol. No adjustment for multiple comparisons. Log transformed
	Method	ANCOVA
	Comments	ANCOVA with log (value) minus log (baseline [BL]) as dependent variable. Terms for treatment, region, gender, pre-screening HbA1c strata, and log (BL)
Method of Estimation	Estimation Parameter	Percent difference from metformin
	Estimated Value	5.97
	Confidence Interval	(2-Sided) 95%; 2.522 to 9.526
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.056
	Comments	LDL cholesterol. No adjustment for multiple comparisons. Log transformed
	Method	ANCOVA
	Comments	ANCOVA with log (value) minus log (baseline [BL]) as dependent variable. Terms for treatment, region, gender, pre-screening HbA1c strata, and log (BL)
Method of Estimation	Estimation Parameter	Percent difference from metformin
	Estimated Value	8.71
	Confidence Interval	(2-Sided) 95%; 2.486 to 15.304
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.072
	Comments	HDL cholesterol. No adjustment for multiple comparison. Log transformed.
	Method	ANCOVA
	Comments	ANCOVA with log (value) minus log (baseline [BL]) as dependent variable. Terms for treatment, region, gender, pre-screening HbA1c strata, and log (BL)
Method of Estimation	Estimation Parameter	Percent difference from metformin
	Estimated Value	2.580
	Confidence Interval	(2-Sided) 95%; -0.232 to 5.470
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.835
	Comments	Triglycerides. No adjustment for multiple comparisons. Log transformed.
	Method	ANCOVA
	Comments	ANCOVA with log (value) minus log (baseline [BL]) as dependent variable. Terms for treatment, region, gender, pre-screening HbA1c strata, and log (BL)
Method of Estimation	Estimation Parameter	Percent difference from metformin
	Estimated Value	0.743
	Confidence Interval	(2-Sided) 95%; -6.056 to 8.035
	Parameter Dispersion	Type: Value:
	Estimation Comments

9. Secondary Outcome

Title	Percent Change From Baseline in Adiponectin at Week 80 (United States [US] and Mexico Subset of Participants )
Description	Blood was taken for measurement of adiponectin. Percent change from baseline at Week 80 was based on log transformed data. This outcome measure was analyzed for a subset of participants in the US and Mexico only.
Time Frame	Baseline and Week 80

Outcome Measure Data

Analysis Population Description

Week 32 Evaluable Population with LOCF for US and Mexico subset. Only evaluable participants, defined as the number of participants with a value at baseline and at the specified visit for the parameter of interest, were analyzed.

Arm/Group Title	Metformin	Avandamet
Arm/Group Description	Beginning dose of Metformin (MET) 500 milligrams (mg) once a day. Dose could be increased up to a maximum dose of MET 2000 mg.	Beginning dose of Avandamet (AVM) 4 mg/500 mg once a day. Dose could be increased up to a maximum dose of 8 mg/2000 mg.
Measure Participants	91	96
Percent change from baseline + SE	18.58	139.28
Percent change from baseline	12.96	128.44
Percent change from baseline - SE	7.61	118.10

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	<0.0001
	Comments	No adjustment for multiple comparisons. Log transformed.
	Method	ANCOVA
	Comments	ANCOVA with log (value) minus log (baseline [BL]) as dependent variable. Terms for treatment, region, gender, pre-screening HbA1c strata, and log (BL)
Method of Estimation	Estimation Parameter	Percent difference from metformin
	Estimated Value	102.24
	Confidence Interval	(2-Sided) 95%; 79.23 to 128.19
	Parameter Dispersion	Type: Value:
	Estimation Comments

10. Secondary Outcome

Title	Percent Change From Baseline in C-reactive Protein (CRP) at Week 80 (US and Mexico Subset of Participants)
Description	Blood was taken for measurement of CRP. Percent change from baseline at Week 80 was based on log transformed data. This outcome measure was analyzed for a subset of participants in the US and Mexico only.
Time Frame	Baseline and Week 80

Outcome Measure Data

Analysis Population Description

Week 32 Evaluable Population with LOCF for US and Mexico subset. Only evaluable participants, defined as the number of participants with a value at baseline and at the specified visit for the parameter of interest, were analyzed.

Arm/Group Title	Metformin	Avandamet
Arm/Group Description	Beginning dose of Metformin (MET) 500 milligrams (mg) once a day. Dose could be increased up to a maximum dose of MET 2000 mg.	Beginning dose of Avandamet (AVM) 4 mg/500 mg once a day. Dose could be increased up to a maximum dose of 8 mg/2000 mg.
Measure Participants	92	95
Percent change from baseline + SE	-6.49	14.28
Percent change from baseline	-10.63	-17.96
Percent change from baseline - SE	-14.59	-21.49

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.1380
	Comments	No adjustment for multiple comparisons. Log transformed.
	Method	ANCOVA
	Comments	ANCOVA with log (value) minus log (baseline [BL]) as dependent variable. Terms for treatment, region, gender, pre-screening HbA1c strata, and log (BL)
Method of Estimation	Estimation Parameter	Percent difference from metformin
	Estimated Value	-8.20
	Confidence Interval	(2-Sided) 95%; -18.04 to 2.82
	Parameter Dispersion	Type: Value:
	Estimation Comments

11. Secondary Outcome

Title	Percent Change in Free Fatty Acids (FFA) From Baseline at Week 80 (US and Mexico Subset of Participants).
Description	Blood was taken for measurement of FFA. Percent change from baseline at Week 80 was based on log transformed data. This outcome measure was analyzed for a subset of participants in the US and Mexico only.
Time Frame	Baseline and Week 80

Outcome Measure Data

Analysis Population Description

Week 32 Evaluable Population with LOCF for US and Mexico subset. Only evaluable participants, defined as participants with a value at baseline and at the specified visit for the parameter of interest, were analyzed.

Arm/Group Title	Metformin	Avandamet
Arm/Group Description	Beginning dose of Metformin (MET) 500 milligrams (mg) once a day. Dose could be increased up to a maximum dose of MET 2000 mg.	Beginning dose of Avandamet (AVM) 4 mg/500 mg once a day. Dose could be increased up to a maximum dose of 8 mg/2000 mg.
Measure Participants	92	95
Percent change from baseline + SE	14.353	1.267
Percent change from baseline	9.40	-2.97
Percent change from baseline - SE	4.657	-7.028

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0342
	Comments	No adjustment for multiple comparisons. Log transformed.
	Method	ANCOVA
	Comments	ANCOVA with log (value) minus log (baseline [BL]) as dependent variable. Terms for treatment, region, gender, pre-screening HbA1c strata, and log (BL)
Method of Estimation	Estimation Parameter	Percent difference from metformin
	Estimated Value	-11.308
	Confidence Interval	(2-Sided) 95%; -20.617 to -0.899
	Parameter Dispersion	Type: Value:
	Estimation Comments

12. Secondary Outcome

Title	Change in Fasting Insulin From Baseline at Week 80 (US and Mexico Subset of Participants)
Description	Blood was taken for fasting insulin measurements. Change from baseline was calculated as the Week 80 value minus the baseline value, with LOCF from Week 32 for withdrawn participants or missing values. This outcome measure was analyzed for a subset of participants in the US and Mexico only.
Time Frame	Baseline and Week 80

Outcome Measure Data

Analysis Population Description

Week 32 Evaluable Population with LOCF for US and Mexico subset. Only evaluable participants, defined as participants with a value at baseline and at the specified visit for the parameter of interest, were analyzed.

Arm/Group Title	Metformin	Avandamet
Arm/Group Description	Beginning dose of Metformin (MET) 500 milligrams (mg) once a day. Dose could be increased up to a maximum dose of MET 2000 mg.	Beginning dose of Avandamet (AVM) 4 mg/500 mg once a day. Dose could be increased up to a maximum dose of 8 mg/2000 mg.
Measure Participants	92	97
Mean (Standard Error) [picomoles per Liter (pmol/l)]	-1.5 (9.18)	-35.2 (8.79)

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0042
	Comments	No adjustment for multiple comparisons.
	Method	ANCOVA
	Comments	ANCOVA with terms for treatment, region, gender, pre-screening HbA1c strata, and baseline.
Method of Estimation	Estimation Parameter	Mean Difference (Net)
	Estimated Value	-33.7
	Confidence Interval	(2-Sided) 95%; -56.666 to -0.99
	Parameter Dispersion	Type: Value:
	Estimation Comments	AVM mean change from baseline minus MET mean change from baseline based on ANCOVA model

13. Secondary Outcome

Title	Change in C-peptide From Baseline at Week 80 (US and Mexico Subset of Participants)
Description	Blood was taken for C-peptide measurements. Change from baseline was calculated as the Week 80 value minus the baseline value with LOCF from Week 32 for withdrawn participants or missing values. This outcome measure was analyzed for a subset of participants in the US and Mexico only.
Time Frame	Baseline and Week 80

Outcome Measure Data

Analysis Population Description

Week 32 Evaluable Population with LOCF for US and Mexico subset. Only evaluable participants, defined as participants with a value at baseline and at the specified visit for the parameter of interest, were analyzed.

Arm/Group Title	Metformin	Avandamet
Arm/Group Description	Beginning dose of Metformin (MET) 500 milligrams (mg) once a day. Dose could be increased up to a maximum dose of MET 2000 mg.	Beginning dose of Avandamet (AVM) 4 mg/500 mg once a day. Dose could be increased up to a maximum dose of 8 mg/2000 mg.
Measure Participants	92	95
Mean (Standard Error) [mmol/l]	-0.244 (0.0518)	-0.473 (0.0502)

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0006
	Comments	No adjustment for multiple comparisons.
	Method	ANCOVA
	Comments	ANCOVA with terms for treatment, region, gender, pre-screening HbA1c strata, and baseline
Method of Estimation	Estimation Parameter	Median Difference (Net)
	Estimated Value	-0.229
	Confidence Interval	(2-Sided) 95%; -0.359 to -0.099
	Parameter Dispersion	Type: Value:
	Estimation Comments	AVM mean change from baseline minus MET mean change from baseline based on ANCOVA model

14. Secondary Outcome

Title	Percent Change From Baseline in in HOMA-S and HOMA-B to Week 80 (US and Mexico Subset of Participants)
Description	Blood was taken for measurement of homeostasis model assessment for insulin sensitivity (HOMA-S) and beta-cell function (HOMA-B). Percent change from baseline at Week 80 was based on log transformed data. This outcome measure was analyzed for a subset of participants in the US and Mexico only. GM, geometric mean; SE, standard error.
Time Frame	Baseline and Week 80

Outcome Measure Data

Analysis Population Description

Week 32 Evaluable Population with LOCF for US and Mexico subset. Only evaluable participants, defined as participants with a value at baseline and at the specified visit for the parameter of interest, were analyzed.

Arm/Group Title	Metformin	Avandamet
Arm/Group Description	Beginning dose of Metformin (MET) 500 milligrams (mg) once a day. Dose could be increased up to a maximum dose of MET 2000 mg.	Beginning dose of Avandamet (AVM) 4 mg/500 mg once a day. Dose could be increased up to a maximum dose of 8 mg/2000 mg.
Measure Participants	91	96
Percent change HOMA-B + SE	88.25	92.32
Percent change HOMA-B	78.85	83.14
Percent change HOMA-B - SE	69.92	74.39
Percent change HOMA-S + SE	25.13	63.67
Percent change HOMA-S	18.65	55.58
Percent change HOMA-S - SE	12.49	47.91

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.7148
	Comments	HOMA-B. Log transformed.
	Method	ANCOVA
	Comments	ANCOVA with log (value) minus log (baseline [BL]) as dependent variable. Terms for treatment, region, gender, pre-screening HbA1c strata, and log (BL)
Method of Estimation	Estimation Parameter	Percent difference from metformin
	Estimated Value	2.40
	Confidence Interval	(2-Sided) 95%; -9.87 to 16.34
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	<0.001
	Comments	HOMA-S. Log transformed.
	Method	ANCOVA
	Comments	ANCOVA with log (value) minus log (baseline [BL]) as dependent variable. Terms for treatment, region, gender, pre-screening HbA1c strata, and log (BL)
Method of Estimation	Estimation Parameter	percent difference from metformin
	Estimated Value	31.14
	Confidence Interval	(2-Sided) 95%; 14.82 to 49.78
	Parameter Dispersion	Type: Value:
	Estimation Comments

15. Secondary Outcome

Title	Slope of Delta-cell Function as Estimated by the Ratio deltaI/deltaG
Description	The ratio Delta I/Delta G is calculated based on the oral glucose tolerance test (OGTT), where Delta I = (30 minute immunoreactive insulin minus 0 minute immunoreactive insulin) and Delta G = (30 minute plasma glucose minus 0 minute plasma glucose). The 0 minute values are fasting insulin and glucose; the 30 minute values are taken 30 minutes after the oral glucose challenge. This outcome measure was analyzed for a subset of participants in the US and Mexico only.
Time Frame	Baseline and Week 80

Outcome Measure Data

Analysis Population Description

Week 32 Evaluable Population with LOCF. Only evaluable participants, defined as participants from US and Mexico sites with a value at baseline and at the specified visit, were analyzed.

Arm/Group Title	Metformin	Avandamet
Arm/Group Description	Beginning dose of Metformin (MET) 500 milligrams (mg) once a day. Dose could be increased up to a maximum dose of MET 2000 mg.	Beginning dose of Avandamet (AVM) 4 mg/500 mg once a day. Dose could be increased up to a maximum dose of 8 mg/2000 mg.
Measure Participants	82	89
Mean (Standard Error) [ratio]	-86.23 (59.706)	-5.38 (55.229)

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.319
	Comments
	Method	Repeated measures analysis
	Comments	Repeated measures analysis with terms for baseline, region, treatment, gender, pre-screening Hb1AC, time, and treatment by time interaction
Method of Estimation	Estimation Parameter	Mean Difference (Net)
	Estimated Value	80.85
	Confidence Interval	(2-Sided) 95%; -79.035 to 240.744
	Parameter Dispersion	Type: Value:
	Estimation Comments

16. Secondary Outcome

Title	Number of Participants at Final Dose Level
Time Frame	Baseline to Week 80 or withdrawal

Outcome Measure Data

Analysis Population Description

Safety population. This population consisted of all participants who were randomized and received at least one dose of the study medication.

Arm/Group Title	Metformin	Avandamet
Arm/Group Description	Beginning dose of Metformin (MET) 500 milligrams (mg) once a day. Dose could be increased up to a maximum dose of MET 2000 mg.	Beginning dose of Avandamet (AVM) 4 mg/500 mg once a day. Dose could be increased up to a maximum dose of 8 mg/2000 mg.
Measure Participants	334	344
Dose level 1 AVM 4 mg/500 mg, MET 500 mg	17 5.1%	19 5.5%
Dose level 2 AVM 4 mg/1000 mg, MET 1000 mg	16 4.8%	18 5.2%
Dose level 3 AVM 6 mg/1500 mg, MET 1500 mg	27 8.1%	34 9.9%
Dose level 4 AVM 8 mg/2000 mg, MET 2000 mg	236 70.7%	242 70.3%

17. Secondary Outcome

Title	Percent Change From Baseline in Lumbar Spine Bone Mass Density (BMD) at Weeks 20, 56, and 80 (Bone Sub-study Subset of Participants)
Description	BMD was measured by dual X-ray absorptiometry (DXA). The percent change from baseline in BMD at a given timepoint was defined at the participant level by the following formula: percent change = (BMD at given week minus BMD at baseline)/BMD at baseline x 100%. This outcome measure was analyzed for a subset of participants in the bone study only.
Time Frame	Baseline and Weeks 20, 56, and 80

Outcome Measure Data

Analysis Population Description

Week 20 Evaluable for the Bone Sub-study Population. This is a subset of the bone sub-study with LOCF from Week 20. Only participants with BMD assessment(s) at Week 20 or later were included.

Arm/Group Title	Metformin	Avandamet
Arm/Group Description	Beginning dose of Metformin (MET) 500 milligrams (mg) once a day. Dose could be increased up to a maximum dose of MET 2000 mg.	Beginning dose of Avandamet (AVM) 4 mg/500 mg once a day. Dose could be increased up to a maximum dose of 8 mg/2000 mg.
Measure Participants	87	87
Overall population, Week 80, n=87, 87	-2.1 (0.54)	0.1 (0.54)
Overall population, Week 56, n=87, 87	-1.4 (0.45)	0.3 (0.44)
Overall population, Week 20, n=87, 87	-0.5 (0.41)	0.5 (0.40)
Male population, Week 80, n=38, 43	-0.7 (0.81)	1.0 (0.90)
Male population, Week 56, n=38, 43	-0.2 (0.61)	0.7 (0.68)
Male population, Week 20, n=38, 43	-0.1 (0.63)	-0.2 (0.71)
Female population, Week 80, n=49, 44	-2.7 (0.77)	0.2 (0.75)
Female population, Week 56, n=49, 44	-2.0 (0.70)	0.3 (0.68)
Female population, Week 20, n=49, 44	-0.9 (0.52)	1.2 (0.51)
Premenopausal population, Week 80, n=21, 14	-2.4 (1.01)	-0.7 (1.03)
Premenopausal population, Week 56, n=21, 14	-0.8 (0.88)	0.2 (0.90)
Premenopausal population, Week 20, n=21, 14	0.2 (0.75)	0.5 (0.77)
Postmenopausal population, Week 80, n=28, 30	-3.3 (1.22)	0 (1.16)
Postmenopausal population, Week 56, n=28, 30	-2.7 (1.09)	-0.1 (1.04)
Postmenopausal population, Week 20, n=28, 30	-1.4 (0.73)	-1.1 (0.69)

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0012
	Comments	Overall population, Week 80. No adjustment for multiple comparisons. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Mean Difference (Net)
	Estimated Value	-2.2
	Confidence Interval	(2-Sided) 95%; -3.5 to -0.9
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0031
	Comments	Overall population, Week 56. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Mean Difference (Net)
	Estimated Value	-1.6
	Confidence Interval	(2-Sided) 95%; -2.7 to -0.6
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0308
	Comments	Overall population, Week 20. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Mean Difference (Net)
	Estimated Value	-1.1
	Confidence Interval	(2-Sided) 95%; -2.0 to -0.1
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0954
	Comments	Male population, Week 80. No adjustment for multiple comparisons. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Mean Difference (Net)
	Estimated Value	-1.7
	Confidence Interval	(2-Sided) 95%; -3.7 to 0.3
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 5

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0015
	Comments	Female population, Week 80. No adjustment for multiple comparisons. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Mean Difference (Net)
	Estimated Value	-2.9
	Confidence Interval	(2-Sided) 95%; -4.7 to -1.2
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 6

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0045
	Comments	Female population, Week 56. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Mean Difference (Net)
	Estimated Value	-2.3
	Confidence Interval	(2-Sided) 95%; -3.9 to -0.7
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 7

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0005
	Comments	Female population, Week 20. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Mean Difference (Net)
	Estimated Value	-2.2
	Confidence Interval	(2-Sided) 95%; -3.3 to -1.0
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 8

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.2363
	Comments	Premenopausal female population, Week 80. No adjustment for multiple comparisons. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Mean Difference (Net)
	Estimated Value	-1.7
	Confidence Interval	(2-Sided) 95%; -4.6 to 1.2
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 9

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0161
	Comments	Postmenopausal female population, Week 80. No adjustment for multiple comparisons. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Mean Difference (Net)
	Estimated Value	-3.3
	Confidence Interval	(2-Sided) 95%; -5.9 to -0.6
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 10

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0338
	Comments	Postmenopausal female population, Week 56. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Mean Difference (Net)
	Estimated Value	-2.6
	Confidence Interval	(2-Sided) 95%; -4.9 to -0.2
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 11

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0020
	Comments	Postmenopausal female population, Week 20. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Median Difference (Net)
	Estimated Value	-2.6
	Confidence Interval	(2-Sided) 95%; -4.1 to -1.0
	Parameter Dispersion	Type: Value:
	Estimation Comments

18. Secondary Outcome

Title	Percent Change From Baseline in Total Hip BMD at Weeks 20, 56, and 80 (Bone Sub-study Subset of Participants)
Description	BMD was measured by dual X-ray absorptiometry (DXA). The percent change from baseline in BMD at a given timepoint was defined at the participant level by the following formula: percent change = (BMD at given week minus BMD at baseline)/BMD at baseline x 100. This outcome measure was analyzed for a subset of participants in the bone study only.
Time Frame	Baseline and Weeks 20, 56, and 80

Outcome Measure Data

Analysis Population Description

Week 20 Evaluable for the Bone Sub-study Population. This is a subset of the bone sub-study with LOCF from Week 20. Only participants with BMD assessment(s) at Week 20 or later were included.

Arm/Group Title	Metformin	Avandamet
Arm/Group Description	Beginning dose of Metformin (MET) 500 milligrams (mg) once a day. Dose could be increased up to a maximum dose of MET 2000 mg.	Beginning dose of Avandamet (AVM) 4 mg/500 mg once a day. Dose could be increased up to a maximum dose of 8 mg/2000 mg.
Measure Participants	87	87
Overall population, Week 80, n=87, 87	0.0 (0.35)	-1.5 (0.35)
Overall population, Week 56, n=87, 87	-0.2 (0.31)	-1.2 (0.32)
Overall population, Week 20, n=87, 87	0.2 (0.24)	-0.4 (0.24)
Male population, Week 80, n=38, 43	0.3 (0.46)	-1.0 (0.41)
Male population, Week 56, n=38, 43	0.2 (0.41)	-0.8 (0.37)
Male population, Week 20, n=38, 43	0.5 (0.35)	-0.3 (0.31)
Female population, Week 80, n=49, 44	0.0 (0.60)	-1.8 (0.61)
Female population, Week 56, n=49, 44	-0.2 (0.52)	-1.4 (0.54)
Female population, Week 20, n=49, 44	-0.3 (0.37)	-0.6 (0.38)
Premenopausal population, Week 80, n=21, 14	1.3 (0.71)	-0.9 (0.69)
Premenopausal population, Week 56, n=21, 14	0.8 (0.69)	-1.6 (0.67)
Premenopausal population, Week 20, n=21, 14	0.3 (0.48)	0.0 (0.47)
Postmenopausal population, Week 80, n=28, 30	-0.9 (0.97)	-2.3 (1.02)
Postmenopausal population, Week 56, n=28, 30	-1.1 (0.81)	-1.5 (0.85)
Postmenopausal population, Week 20, n=28, 30	-0.4 (0.57)	-0.6 (0.60)

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0005
	Comments	Overall population, Week 80. No adjustment for multiple comparisons. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Mean Difference (Net)
	Estimated Value	-1.5
	Confidence Interval	(2-Sided) 95%; -2.3 to -0.7
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0110
	Comments	Overall population, Week 56. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Median Difference (Net)
	Estimated Value	-1.0
	Confidence Interval	(2-Sided) 95%; -1.7 to -0.2
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0618
	Comments	Overall population, Week 20. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Mean Difference (Net)
	Estimated Value	-0.6
	Confidence Interval	(2-Sided) 95%; -1.1 to 0.0
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0110
	Comments	Male population, Week 80. No adjustment for multiple comparisons. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Mean Difference (Net)
	Estimated Value	-1.3
	Confidence Interval	(2-Sided) 95%; -2.3 to -0.3
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 5

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0272
	Comments	Male population, Week 56. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Mean Difference (Net)
	Estimated Value	-1.0
	Confidence Interval	(2-Sided) 95%; -1.9 to -0.1
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 6

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0337
	Comments	Male population, Week 20. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Mean Difference (Net)
	Estimated Value	-0.8
	Confidence Interval	(2-Sided) 95%; -1.6 to -0.1
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 7

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0152
	Comments	Female population, Week 80. No adjustment for multiple comparisons. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Median Difference (Net)
	Estimated Value	-1.7
	Confidence Interval	(2-Sided) 95%; -3.1 to -0.3
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 8

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0570
	Comments	Female population, Week 56. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Mean Difference (Net)
	Estimated Value	-1.2
	Confidence Interval	(2-Sided) 95%; -2.4 to 0.0
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 9

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0296
	Comments	Premenopausal female population, Week 80. No adjustment for multiple comparisons. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Median Difference (Net)
	Estimated Value	-2.2
	Confidence Interval	(2-Sided) 95%; -4.3 to -0.2
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 10

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0155
	Comments	Premenopausal female population, Week 56. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Mean Difference (Net)
	Estimated Value	-2.4
	Confidence Interval	(2-Sided) 95%; -4.4 to -0.5
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 11

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.5870
	Comments	Premenopausal female population, Week 20. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Median Difference (Net)
	Estimated Value	-0.4
	Confidence Interval	(2-Sided) 95%; -1.7 to 1.0
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 12

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.1854
	Comments	Postmenopausal female population, Week 80. No adjustment for multiple comparisons. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Mean Difference (Net)
	Estimated Value	-1.5
	Confidence Interval	(2-Sided) 95%; -3.6 to 0.7
	Parameter Dispersion	Type: Value:
	Estimation Comments

19. Secondary Outcome

Title	Percent Change From Baseline in Trochanter BMD at Weeks 20, 56, and 80 (Bone Sub-study Subset of Participants)
Description	BMD was measured by dual X-ray absorptiometry (DXA). The percent change from baseline in BMD at a given timepoint was defined at the participant level by the following formula: percent change = (BMD at given week minus BMD at baseline)/BMD at baseline x 100. This outcome measure was analyzed for a subset of participants in the bone study only.
Time Frame	Baseline and Weeks 20, 56, and 80

Outcome Measure Data

Analysis Population Description

Week 20 Evaluable for the Bone Sub-study Population. This is a subset of the bone sub-study with LOCF from Week 20. Only participants with BMD assessment(s) at Week 20 or later were included.

Arm/Group Title	Metformin	Avandamet
Arm/Group Description	Beginning dose of Metformin (MET) 500 milligrams (mg) once a day. Dose could be increased up to a maximum dose of MET 2000 mg.	Beginning dose of Avandamet (AVM) 4 mg/500 mg once a day. Dose could be increased up to a maximum dose of 8 mg/2000 mg.
Measure Participants	87	87
Overall population, Week 80, n=87, 87	-0.1 (0.54)	-2.1 (0.55)
Overall population, Week 56, n=87, 87	-0.2 (0.49)	-1.7 (0.49)
Overall population, Week 20, n=87, 87	0.1 (0.35)	-0.6 (0.34)
Male population, Week 80, n=38, 43	1.1 (0.69)	-0.4 (0.61)
Male population, Week 56, n=38, 43	0.9 (0.67)	-0.5 (0.60)
Male population, Week 20, n=38, 43	0.7 (0.55)	-0.2 (0.48)
Female population, Week 80, n=49, 44	-0.2 (0.93)	-2.6 (0.96)
Female population, Week 56, n=49, 44	-0.3 (0.80)	-2.1 (0.82)
Female population, Week 20, n=49, 44	-0.2 (0.52)	-0.8 (0.53)
Premenopausal population, Week 80, n=21, 14	1.7 (1.18)	-1.4 (1.12)
Premenopausal population, Week 56, n=21, 14	0.3 (0.94)	-2.6 (0.89)
Premenopausal population, Week 20, n=21, 14	0.0 (0.71)	0.0 (0.67)
Postmenopausal population, Week 80, n=28, 30	-1.1 (1.45)	-3.6 (1.53)
Postmenopausal population, Week 56, n=28, 30	-1.1 (1.25)	-2.2 (1.31)
Postmenopausal population, Week 20, n=28, 30	-0.3 (0.76)	-1.0 (0.80)

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0038
	Comments	Overall population, Week 80. No adjustment for multiple comparisons. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Median Difference (Net)
	Estimated Value	-1.9
	Confidence Interval	(2-Sided) 95%; -3.2 to -0.6
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0134
	Comments	Overall population, Week 56. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Mean Difference (Net)
	Estimated Value	-1.5
	Confidence Interval	(2-Sided) 95%; -2.7 to -0.3
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0967
	Comments	Overall population, Week 20. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Median Difference (Net)
	Estimated Value	-0.7
	Confidence Interval	(2-Sided) 95%; -1.5 to 0.1
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0512
	Comments	Male population, Week 80. No adjustment for multiple comparisons. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Mean Difference (Net)
	Estimated Value	-1.5
	Confidence Interval	(2-Sided) 95%; -3.0 to 0.0
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 5

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0330
	Comments	Female population, Week 80. No adjustment for multiple comparisons. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Mean Difference (Net)
	Estimated Value	-2.4
	Confidence Interval	(2-Sided) 95%; -4.6 to -0.2
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 6

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0547
	Comments	Female population, Week 56. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Mean Difference (Net)
	Estimated Value	-1.8
	Confidence Interval	(2-Sided) 95%; -3.7 to 0.0
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 7

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.0613
	Comments	Premenopausal female population, Week 80. No adjustment for multiple comparisons. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Mean Difference (Net)
	Estimated Value	-3.1
	Confidence Interval	(2-Sided) 95%; -6.3 to 0.2
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 8

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.1369
	Comments	Postmenopausal female population, Week 80. No adjustment for multiple comparisons. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Mean Difference (Net)
	Estimated Value	-2.5
	Confidence Interval	(2-Sided) 95%; -3.9 to 1.7
	Parameter Dispersion	Type: Value:
	Estimation Comments

20. Secondary Outcome

Title	Percent Change From Baseline in Femoral Neck BMD at Weeks 20, 56, and 80 (Bone Sub-study Subset of Participants)
Description	BMD was measured by dual X-ray absorptiometry (DXA). The percent change from baseline in BMD at a given timepoint was defined at the participant level by the following formula: percent change = (BMD at given week minus BMD at baseline)/BMD at baseline x 100. This outcome measure was analyzed for a subset of participants in the bone study only.
Time Frame	Baseline and Weeks 20, 56, and 80

Outcome Measure Data

Analysis Population Description

Week 20 Evaluable for the Bone Sub-study Population. This is a subset of the bone sub-study with LOCF from Week 20. Only participants with BMD assessment(s) at Week 20 or later were included.

Arm/Group Title	Metformin	Avandamet
Arm/Group Description	Beginning dose of Metformin (MET) 500 milligrams (mg) once a day. Dose could be increased up to a maximum dose of MET 2000 mg.	Beginning dose of Avandamet (AVM) 4 mg/500 mg once a day. Dose could be increased up to a maximum dose of 8 mg/2000 mg.
Measure Participants	87	87
Overall population, Week 80, n=87, 87	-0.7 (0.39)	-1.4 (0.39)
Overall population, Week 56, n=87, 87	-0.7 (0.39)	-0.8 (0.39)
Overall population, Week 20, n=87, 87	-0.1 (0.36)	-0.2 (0.36)
Male population, Week 80, n=38, 43	-1.4 (0.66)	-2.2 (0.59)
Male population, Week 56, n=38, 43	-0.7 (0.64)	-1.3 (0.57)
Male population, Week 20, n=38, 43	-0.4 (0.60)	-0.7 (0.53)
Female population, Week 80, n=49, 44	-0.3 (0.56)	-1.2 (0.57)
Female population, Week 56, n=49, 44	-0.6 (0.55)	-0.5 (0.56)
Female population, Week 20, n=49, 44	-0.4 (0.49)	-0.2 (0.50)
Premenopausal population, Week 80, n=21, 14	0.9 (0.73)	-0.3 (0.68)
Premenopausal population, Week 56, n=21, 14	-0.1 (0.96)	-0.4 (0.89)
Premenopausal population, Week 20, n=21, 14	0.9 (0.83)	-0.3 (0.77)
Postmenopausal population, Week 80, n=28, 30	-1.3 (0.87)	-2.1 (0.91)
Postmenopausal population, Week 56, n=28, 30	-1.2 (0.81)	-0.7 (0.85)
Postmenopausal population, Week 20, n=28, 30	-1.1 (0.71)	-0.4 (0.74)

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.1260
	Comments	Overall population, Week 80. No adjustment for multiple comparisons. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Median Difference (Net)
	Estimated Value	-0.7
	Confidence Interval	(2-Sided) 95%; -1.7 to 0.2
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.3117
	Comments	Male population, Week 80. No adjustment for multiple comparisons. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Median Difference (Net)
	Estimated Value	-0.7
	Confidence Interval	(2-Sided) 95%; -2.2 to 0.7
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.1776
	Comments	Female population, Week 80. No adjustment for multiple comparisons. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Mean Difference (Net)
	Estimated Value	-0.9
	Confidence Interval	(2-Sided) 95%; -2.2 to 0.4
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.2259
	Comments	Premenopausal female population, Week 80. No adjustment for multiple comparisons. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Median Difference (Net)
	Estimated Value	-1.2
	Confidence Interval	(2-Sided) 95%; -3.2 to 0.8
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 5

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.3840
	Comments	postmenopausal female population, Week 80. No adjustment for multiple comparisons. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Mean Difference (Net)
	Estimated Value	-0.9
	Confidence Interval	(2-Sided) 95%; -2.8 to 1.1
	Parameter Dispersion	Type: Value:
	Estimation Comments

21. Secondary Outcome

Title	Percent Change From Baseline in Distal Radius BMD at Weeks 20, 56, and 80 (Bone Sub-study Subset of Participants)
Description	BMD was measured by dual X-ray absorptiometry (DXA). The percent change from baseline in BMD at a given timepoint was defined at the participant level by the following formula: percent change = (BMD at given week minus BMD at baseline)/BMD at baseline x 100. This outcome measure was analyzed for a subset of participants in the bone study only.
Time Frame	Baseline and Weeks 20, 56, and 80

Outcome Measure Data

Analysis Population Description

Week 20 Evaluable for the Bone Sub-study Population. This is a subset of the bone sub-study with LOCF from Week 20. Only participants with BMD assessment(s) at Week 20 or later were included.

Arm/Group Title	Metformin	Avandamet
Arm/Group Description	Beginning dose of Metformin (MET) 500 milligrams (mg) once a day. Dose could be increased up to a maximum dose of MET 2000 mg.	Beginning dose of Avandamet (AVM) 4 mg/500 mg once a day. Dose could be increased up to a maximum dose of 8 mg/2000 mg.
Measure Participants	87	87
Overall population, Week 80, n=87, 87	0.3 (0.63)	-0.1 (0.66)
Overall population, Week 56, n=87, 87	0.5 (0.55)	0.5 (0.58)
Overall population, Week 20, n=87, 87	1.2 (0.58)	0.9 (0.61)
Male population, Week 80, n=38, 43	0.4 (0.84)	-0.6 (0.82)
Male population, Week 56, n=38, 43	0.7 (0.73)	0.4 (0.71)
Male population, Week 20, n=38, 43	1.5 (0.75)	0.2 (0.73)
Female population, Week 80, n=49, 44	0.1 (1.00)	0.1 (1.07)
Female population, Week 56, n=49, 44	0.6 (0.87)	0.7 (0.93)
Female population, Week 20, n=49, 44	0.6 (0.94)	1.1 (0.99)
Premenopausal population, Week 80, n=21, 14	1.1 (1.51)	-0.3 (1.44)
Premenopausal population, Week 56, n=21, 14	2.2 (1.37)	-0.6 (1.30)
Premenopausal population, Week 20, n=21, 14	2.6 (1.51)	0.7 (1.40)
Postmenopausal population, Week 80, n=28, 30	-0.5 (1.61)	-0.2 (1.76)
Postmenopausal population, Week 56, n=28, 30	-0.3 (1.34)	0.8 (1.46)
Postmenopausal population, Week 20, n=28, 30	0.5 (1.36)	1.3 (1.49)

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.6102
	Comments	Overall population, Week 80. No adjustment for multiple comparisons. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Mean Difference (Net)
	Estimated Value	-0.4
	Confidence Interval	(2-Sided) 95%; -1.9 to 1.1
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.3445
	Comments	Male population, Week 80. No adjustment for multiple comparisons. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Mean Difference (Net)
	Estimated Value	-0.9
	Confidence Interval	(2-Sided) 95%; -2.9 to 1.0
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.9735
	Comments	Female population, Week 80. No adjustment for multiple comparisons. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Mean Difference (Net)
	Estimated Value	0.0
	Confidence Interval	(2-Sided) 95%; -2.4 to 2.3
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.4861
	Comments	Premenopausal female population, Week 80. No adjustment for multiple comparisons. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Mean Difference (Net)
	Estimated Value	-1.4
	Confidence Interval	(2-Sided) 95%; -5.8 to 2.9
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 5

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.8770
	Comments	Postmenopausal female population, Week 80. No adjustment for multiple comparisons. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Mean Difference (Net)
	Estimated Value	0.3
	Confidence Interval	(2-Sided) 95%; -3.5 to 4.0
	Parameter Dispersion	Type: Value:
	Estimation Comments

22. Secondary Outcome

Title	Percent Change From Baseline in Total Body BMD at Weeks 20, 56, and 80 (Bone Sub-study Subset of Participants)
Description	BMD was measured by dual X-ray absorptiometry (DXA). The percent change from baseline in BMD at a given timepoint was defined at the participant level by the following formula: percent change = (BMD at given week minus BMD at baseline)/BMD at baseline x 100. This outcome measure was analyzed for a subset of participants in the bone study only.
Time Frame	Baseline and Weeks 20, 56, and 80

Outcome Measure Data

Analysis Population Description

Week 20 Evaluable for the Bone Sub-study Population. This is a subset of the bone sub-study with LOCF from Week 20. Only participants with BMD assessment(s) at Week 20 or later were included.

Arm/Group Title	Metformin	Avandamet
Arm/Group Description	Beginning dose of Metformin (MET) 500 milligrams (mg) once a day. Dose could be increased up to a maximum dose of MET 2000 mg.	Beginning dose of Avandamet (AVM) 4 mg/500 mg once a day. Dose could be increased up to a maximum dose of 8 mg/2000 mg.
Measure Participants	87	87
Overall population, Week 80, n=87, 87	1.1 (0.64)	1.1 (0.64)
Overall population, Week 56, n=87, 87	1.3 (0.63)	-0.1 (0.63)
Overall population, Week 20, n=87, 87	0.0 (0.37)	0.5 (0.37)
Male population, Week 80, n=38, 43	0.3 (0.43)	0.5 (0.38)
Male population, Week 56, n=38, 43	0.5 (0.41)	0.5 (0.37)
Male population, Week 20, n=38, 43	0.2 (0.43)	0.5 (0.38)
Female population, Week 80, n=49, 44	2.0 (1.00)	1.0 (1.01)
Female population, Week 56, n=49, 44	1.9 (1.03)	1.2 (1.04)
Female population, Week 20, n=49, 44	-0.2 (0.56)	0.5 (0.57)
Premenopausal population, Week 80, n=21, 14	5.3 (2.26)	-0.2 (2.73)
Premenopausal population, Week 56, n=21, 14	5.2 (2.42)	-1.0 (2.94)
Premenopausal population, Week 20, n=21, 14	-0.2 (1.38)	0.3 (1.66)
Postmenopausal population, Week 80, n=28, 30	0.2 (0.56)	-0.4 (0.67)
Postmenopausal population, Week 56, n=28, 30	-0.2 (0.49)	-0.6 (0.58)
Postmenopausal population, Week 20, n=28, 30	-0.2 (0.45)	0.1 (0.54)

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.7015
	Comments	Overall population, Week 80. No adjustment for multiple comparisons. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Median Difference (Net)
	Estimated Value	-0.2
	Confidence Interval	(2-Sided) 95%; -1.5 to 1.0
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.6767
	Comments	Male population, Week 80. No adjustment for multiple comparisons. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Mean Difference (Net)
	Estimated Value	0.2
	Confidence Interval	(2-Sided) 95%; -0.7 to 1.0
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.4199
	Comments	Female population, Week 80. No adjustment for multiple comparisons. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Mean Difference (Net)
	Estimated Value	-1.0
	Confidence Interval	(2-Sided) 95%; -3.4 to 1.5
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.2526
	Comments	Premenopausal female population, Week 80. No adjustment for multiple comparisons. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Median Difference (Net)
	Estimated Value	-5.5
	Confidence Interval	(2-Sided) 95%; -16.4 to 5.4
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 5

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.4153
	Comments	postmenopausal female population, Week 80. No adjustment for multiple comparisons. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Mean Difference (Net)
	Estimated Value	-0.5
	Confidence Interval	(2-Sided) 95%; -1.8 to 0.8
	Parameter Dispersion	Type: Value:
	Estimation Comments

23. Secondary Outcome

Title	Percent Change From Baseline in Serum Calcium at Weeks 12, 32, 56, and 80
Description	Blood was taken for measurement of serum calcium. Percent change from baseline was based on log transformed data. Geometric mean, GM; standard error, SE. This outcome measure was analyzed for a subset of participants in the bone study only. n is the number of evaluable participants, which is the number of participants with a value at baseline and at the specified visit for the parameter of interest.
Time Frame	Baseline and Weeks 12, 32, 56, and 80

Outcome Measure Data

Analysis Population Description

Week 20 Evaluable for the Bone Sub-study Population. This is a subset of the bone sub-study with LOCF from Week 20. Only participants with serum calcium measurements at Week 20 or later were included.

Arm/Group Title	Metformin	Avandamet
Arm/Group Description	Beginning dose of Metformin (MET) 500 milligrams (mg) once a day. Dose could be increased up to a maximum dose of MET 2000 mg.	Beginning dose of Avandamet (AVM) 4 mg/500 mg once a day. Dose could be increased up to a maximum dose of 8 mg/2000 mg.
Measure Participants	79	79
Wk80 % change + SE, overall n= 79, 79	0.929	1.104
Wk80 % change, overall n=79, 79	0.424	0.592
Wk80 % change - SE, overall n= 79, 79	-0.079	0.083
Wk56 % change + SE, overall n= 79, 79	1.098	1.313
Wk56 % change, overall n= 79, 79	0.527	0.735
Wk56 % change - SE, overall n 79, 79	-0.040	0.160
Wk32 % change + SE, overall n=78, 77	0.136	-0.735
Wk32 % change, overall n= 78, 77	-0.316	-1.204
Wk32 % change - SE, overall n 78, 79	-0.766	-1.607
Wk12 % change + SE, overall n=81, 78	0.568	-0.315
Wk12 % change, overall n= 81, 78	0.123	-0.767
Wk12 % change - SE, overall n 81, 78	-0.320	-1.216
Wk80 % change + SE, males n=34, 39	0.8777	1.551
Wk80 % change, males n= 34, 39	0.071	0.817
Wk80 % change - SE, males n 34, 39	-0.728	0.877
Wk56 % change + SE, males n=34, 39	1.304	1.092
Wk56 % change, males n= 34, 39	0.646	0.498
Wk56 % change - SE, males n 34, 39	-0.009	-0.094
Wk32 % change + SE, males n=33, 39	-0.538	-0.585
Wk32 % change, males n= 33, 39	-1.281	-1.254
Wk32 % change - SE, males n 33,39	-2.019	-1.919
Wk12 % change + SE, males n=37, 40	1.410	0.543
Wk12 % change, males n= 37, 40	0.656	-0.147
Wk12 % change - SE, males n 37, 40	-0.091	-0.833
Wk80 % change + SE, females n= 45, 40	1.508	1.049
Wk80 % change, females n=45, 40	0.735	0.279
Wk80 % change - SE, females n= 45, 40	-0.033	-0.484
Wk56 % change + SE, females n= 45, 40	1.692	1.884
Wk56 % change, females n=45, 40	0.720	0.911
Wk56 % change - SE, females n= 45, 40	-0.242	-0.053
Wk32 % change + SE, females n= 45, 38	0.637	-0.653
Wk32 % change, females n=45, 38	0.038	-1.293
Wk32 % change - SE, females n= 45, 38	-0.557	-1.930
Wk12 % change + SE, females n= 44, 38	0.715	-0.425
Wk12 % change, females n=44, 38	0.055	-1.081
Wk12 % change - SE, females n= 44, 38	-0.600	-1.733
Wk80 % change + SE, premenopausal females n=19, 14	1.429	0.704
Wk80 % change, premenopausal females n= 19, 14	0.531	-0.111
Wk80 % change - SE, premenopausal females n 19, 14	-0.359	-0.918
Wk56 % change + SE, premenopausal females n=19, 14	2.091	2.334
Wk56 % change, premenopausal females n= 19, 14	0.960	1.298
Wk56 % change - SE, premenopausal females n 19, 14	-0.159	0.273
Wk32 % change + SE, premenopausal females n=19, 13	1.074	-0.650
Wk32 % change, premenopausal females n= 19, 13	0.070	-1.679
Wk32 % change - SE, premenopausal females n 19, 13	-0.923	-2.698
Wk12 % change + SE, premenopausal females n=19, 13	0.420	-0.218
Wk12 % change, premenopausal females n= 19, 13	-0.657	-1.280
Wk12 % change - SE, premenopausal females n 19, 13	-1.722	-2.330
Wk80 % change+ SE, postmenopausal females n=26, 26	2.019	1.915
Wk80 % change, postmenopausal females n=26, 26	0.760	0.604
Wk80% change - SE, postmenopausal females n=26, 26	-0.484	-0.690
Wk56 % change+ SE, postmenopausal females n=26, 26	2.181	2.691
Wk56 % change, postmenopausal females n=26, 26	0.560	0.993
Wk56% change - SE, postmenopausal females n=26, 26	-1.035	-0.677
Wk32 % change+ SE, postmenopausal females n=26, 25	0.844	-0.535
Wk32 % change, postmenopausal females n=26, 25	0.056	-1.360
Wk32% change - SE, postmenopausal females n=26, 25	-0.726	-2.178
Wk12 % change+ SE, postmenopausal females n=25, 25	0.844	-1.163
Wk12 % change, postmenopausal females n=25, 25	-0.042	-2.074
Wk12% change - SE, postmenopausal females n=25, 25	-0.921	-2.976

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.7895
	Comments	Overall population, Week 80. Log transformed. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline.
Method of Estimation	Estimation Parameter	Percent difference from metformin
	Estimated Value	0.168
	Confidence Interval	(2-Sided) 95%; -1.066 to 1.417
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.4155
	Comments	Males, Week 80. Log transformed. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline.
Method of Estimation	Estimation Parameter	Percent difference from metformin
	Estimated Value	0.745
	Confidence Interval	(2-Sided) 95%; -1.064 to 2.587
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.6223
	Comments	Females, Week 80. Log transformed. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline.
Method of Estimation	Estimation Parameter	Percent difference from metformin
	Estimated Value	-0.452
	Confidence Interval	(2-Sided) 95%; -2.253 to 1.382
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.5908
	Comments	Premenopausal females, Week 80. Log transformed. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline.
Method of Estimation	Estimation Parameter	Percent difference from metformin
	Estimated Value	-0.638
	Confidence Interval	(2-Sided) 95%; -3.043 to 1.826
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 5

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.9154
	Comments	Postmenopausal females, Week 80. Log transformed. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline.
Method of Estimation	Estimation Parameter	Percent difference from metformin
	Estimated Value	-0.155
	Confidence Interval	(2-Sided) 95%; -3.037 to 2.814
	Parameter Dispersion	Type: Value:
	Estimation Comments

24. Secondary Outcome

Title	Percent Change From Baseline in Intact Parathyroid Hormone at Week 80
Description	Blood was taken for measurement of intact parathyroid hormone. Percent change from baseline was based on log transformed data. Standard error, SE; Wk, Week; %, percent. This outcome measure was analyzed for a subset of participants in the bone study only. n is the number of evaluable participants, which is the number of participants with a value at baseline and at the specified visit for the parameter of interest.
Time Frame	Baseline and Week 80

Outcome Measure Data

Analysis Population Description

Week 20 Evaluable for the Bone Sub-study Population. This is a subset of the bone sub-study with LOCF from Week 20. Only participants with intact parathyroid hormone measurements at Week 20 or later were included.

Arm/Group Title	Metformin	Avandamet
Arm/Group Description	Beginning dose of Metformin (MET) 500 milligrams (mg) once a day. Dose could be increased up to a maximum dose of MET 2000 mg.	Beginning dose of Avandamet (AVM) 4 mg/500 mg once a day. Dose could be increased up to a maximum dose of 8 mg/2000 mg.
Measure Participants	45	38
Wk80 % change + SE, overall n= 45, 38	-1.296	0.843
Wk80 % change, overall n=45, 38	-7.634	-6.279
Wk80 % change - SE, overall n= 45, 38	-13.565	-12.898
Wk80 % change + SE, males n=16, 21	-1.160	-2.619
Wk80 % change, males n= 16, 21	-9.238	-8.941
Wk80 % change - SE, males n=16, 21	-16.655	-14.852
Wk80 % change + SE, females n= 30, 22	4.662	11.166
Wk80 % change, females n=30, 22	-4.608	-1.759
Wk 80 % change - SE, females n= 30, 22	-13.056	-13.182
Wk80 % change + SE, premenopausal females n=13, 7	14.295	24.396
Wk80 % change, premenopausal females n= 13, 7	-7.686	-0.126
Wk80 % change - SE, premenopausal females n=13, 7	-25.440	-20.160
Wk80 % change+ SE, postmenopausal females n=16, 10	3.017	8.297
Wk80 % change, postmenopausal females n=16, 10	-12.109	-9.170
Wk80% change - SE, postmenopausal females n=16, 10	-25.014	-23.807

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.8682
	Comments	Overall population, Week 80. Log transformed.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Percent difference from metformin
	Estimated Value	1.467
	Confidence Interval	(2-Sided) 95%; -14.785 to 20.818
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.9740
	Comments	Males, Week 80. Log transformed.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline.
Method of Estimation	Estimation Parameter	Percent difference from metformin
	Estimated Value	0.328
	Confidence Interval	(2-Sided) 95%; -18.308 to 23.214
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.8378
	Comments	Females, Week 80. Log transformed.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline.
Method of Estimation	Estimation Parameter	Percent difference from metformin
	Estimated Value	2.986
	Confidence Interval	(2-Sided) 95%; -22.997 to 37.735
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.7889
	Comments	Pre-menopausal females, Week 80. Log transformed.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline.
Method of Estimation	Estimation Parameter	Percent difference from metformin
	Estimated Value	8.190
	Confidence Interval	(2-Sided) 95%; -43.839 to 108.422
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 5

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.8800
	Comments	Postmenopausal females, Week 80. Log transformed.
	Method	ANCOVA
	Comments	.Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline.
Method of Estimation	Estimation Parameter	Percent difference from metformin
	Estimated Value	3.344
	Confidence Interval	(2-Sided) 95%; -34.853 to 63.935
	Parameter Dispersion	Type: Value:
	Estimation Comments

25. Secondary Outcome

Title	Percent Change From Baseline in 25-hydroxy Vitamin D at Week 80
Description	Blood was taken for measurement of 25-hydroxy vitamin D. Percent change from baseline was based on log transformed data. Standard error, SE; Wk, Week; %, percent. This outcome measure was analyzed for a subset of participants in the bone study only. n is the number of evaluable participants, which is the number of participants with a value at baseline and at the specified visit for the parameter of interest.
Time Frame	Baseline and Week 80

Outcome Measure Data

Analysis Population Description

Week 20 Evaluable for the Bone Sub-study Population. This is a subset of the bone sub-study with LOCF from Week 20. Only participants with 25-hydroxy vitamin D measurements at Week 20 or later were included.

Arm/Group Title	Metformin	Avandamet
Arm/Group Description	Beginning dose of Metformin (MET) 500 milligrams (mg) once a day. Dose could be increased up to a maximum dose of MET 2000 mg.	Beginning dose of Avandamet (AVM) 4 mg/500 mg once a day. Dose could be increased up to a maximum dose of 8 mg/2000 mg.
Measure Participants	45	38
Wk80 % change + SE, overall n= 45, 38	-7.0053	-4.6305
Wk80 % change, overall n=45, 38	-13.9678	-12.9210
Wk 80 % change - SE, overall n= 45, 38	-20.4091	-20.4909
Wk80 % change + SE, males n=16, 21	8.9639	-4.2837
Wk80 % change, males n=16, 21	-4.4369	-14.1507
Wk80 % change - SE, males n=16, 21	-16.1896	-23.0005
Wk80 % change + SE, females n= 29, 17	-7.1080	4.7124
Wk80 % change, females n=29, 17	-16.4578	-9.0412
Wk 80 % change - SE, females n= 29, 17	-24.8664	-20.9882
Wk80 % change + SE, premenopausal females n=13, 7	-2.0833	0.3546
Wk80 % change, premenopausal females n= 13, 7	-15.3838	-16.2325
Wk80 % change - SE, premenopausal females n=13, 7	-26.8776	-30.0781
Wk80 % change+ SE, postmenopausal females n=16, 10	8.8368	17.3807
Wk80 % change, postmenopausal females n=16, 10	-7.1318	-2.7100
Wk80% change - SE, postmenopausal females n=16, 10	-20.7575	-19.3620

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.9069
	Comments	Overall population, Week 80. Log transformed.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline.
Method of Estimation	Estimation Parameter	Percent difference from metformin
	Estimated Value	1.2168
	Confidence Interval	(2-Sided) 95%; -17.6057 to 24.3392
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.5118
	Comments	Males, Week 80. Log transformed.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline.
Method of Estimation	Estimation Parameter	Percent difference from metformin
	Estimated Value	-10.1648
	Confidence Interval	(2-Sided) 95%; -35.6298 to 25.3742
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.5816
	Comments	Females, Week 80. Log transformed.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline.
Method of Estimation	Estimation Parameter	Percent difference from metformin
	Estimated Value	8.8777
	Confidence Interval	(2-Sided) 95%; -20.2636 to 48.6692
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.9587
	Comments	Pre-menopausal females, Week 80. Log transformed.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline.
Method of Estimation	Estimation Parameter	Percent difference from metformin
	Estimated Value	-1.0031
	Confidence Interval	(2-Sided) 95%; -35.9280 to 52.9590
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 5

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.8337
	Comments	Postmenopausal females, Week 80. Log transformed.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline.
Method of Estimation	Estimation Parameter	Percent difference from metformin
	Estimated Value	4.7614
	Confidence Interval	(2-Sided) 95%; -34.4741 to 67.4901
	Parameter Dispersion	Type: Value:
	Estimation Comments

26. Secondary Outcome

Title	Percent Change From Baseline in Estradiol at Weeks 20, 56, and 80
Description	Blood was taken for measurement of estradiol. Percent change from baseline was based on log transformed data. Standard error, SE; Wk, Week; %, percent. This outcome measure was analyzed for a subset of female participants in the bone study only. n is the number of evaluable participants, which is the number of female participants with a value at baseline and at the specified visit for the parameter of interest.
Time Frame	Baseline and Weeks 20, 56, and 80

Outcome Measure Data

Analysis Population Description

Week 20 Evaluable for the Bone Sub-study Population. This is a subset of the bone sub-study with LOCF from Week 20. Only participants with estradiol measurements at Week 20 or later were included.

Arm/Group Title	Metformin	Avandamet
Arm/Group Description	Beginning dose of Metformin (MET) 500 milligrams (mg) once a day. Dose could be increased up to a maximum dose of MET 2000 mg.	Beginning dose of Avandamet (AVM) 4 mg/500 mg once a day. Dose could be increased up to a maximum dose of 8 mg/2000 mg.
Measure Participants	30	22
Wk80 % change + SE, females n= 30, 22	-8.695	0.097
Wk80 % change, females n=30, 22	-21.360	-15.441
Wk 80 % change - SE, females n= 30, 22	-32.268	-28.567
Wk56 % change + SE, females n= 27, 21	12.015	10.268
Wk56 % change, females n=27, 21	-0.815	-2.439
Wk 56 % change - SE, females n= 27, 21	-12.175	-13.682
Wk20 % change + SE, females n= 36, 25	7.863	14.875
Wk20 % change, females n=36, 25	-5.097	-0.646
Wk 20 % change - SE, females n= 36, 25	-16.500	-14.875
Wk80 % change + SE, premenopausal females n=13, 9	25.386	67.328
Wk80 % change, premenopausal females n= 13, 9	-17.314	7.667
Wk80 % change - SE, premenopausal females n=13, 9	-45.472	-30.722
Wk56 % change + SE, premenopausal females n=12, 8	35.708	19.557
Wk56 % change, premenopausal females n= 12, 8	1.677	-11.431
Wk56 % change - SE, premenopausal females n=12, 8	-23.820	-34.386
Wk20 % change + SE, premenopausal females n=15, 10	5.738	8.866
Wk20 % change, premenopausal females n=15, 10	-3.721	-3.005
Wk20 % change - SE, premenopausal females n=15, 10	-12.334	-13.581
Wk80 % change+ SE, postmenopausal females n=17, 13	-27.303	-25.501
Wk80 % change, postmenopausal females n=17, 13	-33.045	-31.754
Wk80% change - SE, postmenopausal females n=17, 13	-38.334	-37.482
Wk56 % change+ SE, postmenopausal females n=15, 13	-25.758	-18.475
Wk56 % change, postmenopausal females n=15, 13	-32.766	-25.372
Wk56% change - SE, postmenopausal females n=15, 13	-39.113	-31.686
Wk20 % change+ SE, postmenopausal females n=21, 15	3.432	16.140
Wk20 % change, postmenopausal females n=21, 15	-12.870	-5.061
Wk20% change - SE, postmenopausal females n=21, 15	-26.603	-22.391

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.7041
	Comments	Females, Week 80. Log transformed. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline.
Method of Estimation	Estimation Parameter	Percent difference from metformin
	Estimated Value	7.527
	Confidence Interval	(2-Sided) 95%; -26.773 to 57.892
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.6403
	Comments	Pre-menopausal females, Week 80. Log transformed. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline.
Method of Estimation	Estimation Parameter	Percent difference from metformin
	Estimated Value	30.211
	Confidence Interval	(2-Sided) 95%; -61.586 to 341.371
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.8474
	Comments	Postmenopausal females, Week 80. Log transformed. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	.Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline.
Method of Estimation	Estimation Parameter	Percent difference from metformin
	Estimated Value	1.929
	Confidence Interval	(2-Sided) 95%; -17.015 to 25.198
	Parameter Dispersion	Type: Value:
	Estimation Comments

27. Secondary Outcome

Title	Percent Change From Baseline in C-terminal Telopeptide (CTX) at Weeks 20, 56, and 80
Description	Blood was taken for measurement of CTX. Percent change from baseline was based on log transformed data. Standard error, SE; Wk, Week; %, percent. This outcome measure was analyzed for a subset of participants in the bone study only. n is the number of evaluable participants, which is the number of participants with a value at baseline and at the specified visit for the parameter of interest.
Time Frame	Baseline and Weeks 20, 56, and 80

Outcome Measure Data

Analysis Population Description

Week 20 Evaluable for the Bone Sub-study Population. This is a subset of the bone sub-study with LOCF from Week 20. Only participants with CTX measurements at Week 20 or later were included.

Arm/Group Title	Metformin	Avandamet
Arm/Group Description	Beginning dose of Metformin (MET) 500 milligrams (mg) once a day. Dose could be increased up to a maximum dose of MET 2000 mg.	Beginning dose of Avandamet (AVM) 4 mg/500 mg once a day. Dose could be increased up to a maximum dose of 8 mg/2000 mg.
Measure Participants	69	62
Wk80 % change + SE, overall n=69, 62	-12.7	-7.5
Wk80 % change, overall n=69, 62	-18.7	-14.0
Wk80 % change - SE, overall n=79, 62	-24.2	-20.1
Wk56 % change + SE, overall n=68, 62	-5.3	3.7
Wk56 % change, overall n=68, 62	-11.4	-3.2
Wk56 % change - SE, overall n=68, 62	-17.1	-9.6
Wk20 % change + SE, overall n=61, 56	-0.8	-0.8
Wk20 % change, overall n=61, 56	-5.5	-5.9
Wk20 % change - SE, overall n=61, 56	-9.9	-10.7
Wk80 % change + SE, males n=28, 32	-11.0	-1.8
Wk80 % change, males n=28, 32	-20.7	-10.9
Wk80 % change - SE, males n=28, 32	-29.2	-19.1
Wk56 % change + SE, males n=28, 32	-11.5	6.5
Wk56 % change, males n=28, 32	-20.0	-2.2
Wk56 % change - SE, males n=28, 32	-27.7	-10.2
Wk20 % change + SE, males n=23, 29	4.8	13.2
Wk20 % change, males n=23, 29	-3.1	5.8
Wk20 % change - SE, males n=23, 29	-10.3	-1.1
Wk80 % change + SE, females n=41, 30	-12.9	-8.2
Wk80 % change, females n=41, 30	-20.1	-16.9
Wk80 % change - SE, females n=41, 30	-26.6	-24.8
Wk56 % change + SE, females n=40, 30	-8.5	1.2
Wk56 % change, females n=40, 30	-15.8	-8.0
Wk56 % change - SE, females n=40, 30	-22.5	-16.4
Wk20 % change + SE, females n=38, 27	-2.5	-8.3
Wk20 % change, females n=38, 27	-8.9	-15.3
Wk20 % change - SE, females n=38, 27	-15.0	-21.7
Wk80 % change + SE, premenopausal females n=18, 9	-18.6	10.9
Wk80 % change, premenopausal females n=18, 9	-27.8	-4.7
Wk80 % change - SE, premenopausal females n=18, 9	-36.0	-18.1
Wk56 % change + SE, premenopausal females n=18, 9	-25.5	12.6
Wk56 % change, premenopausal females n=18, 9	-31.8	0.8
Wk56 % change - SE, premenopausal females n=18, 9	-37.5	-9.8
Wk20 % change + SE, premenopausal females n=16, 9	-4.2	-7.8
Wk20 % change, premenopausal females n=16, 9	-14.0	-18.9
Wk20 % change - SE, premenopausal females n=16, 9	-22.8	-28.8
Wk80 % change+ SE, postmenopausal females n=23, 21	-11.5	-14.3
Wk80 % change, postmenopausal females n=23, 21	-21.0	-24.4
Wk80% change - SE, postmenopausal females n=23, 21	-29.5	-33.4
Wk56 % change+ SE, postmenopausal females n=22, 21	10.3	5.1
Wk56 % change, postmenopausal females n=22, 21	-2.1	-7.4
Wk56% change - SE, postmenopausal females n=22, 21	-13.0	-18.4
Wk20 % change+ SE, postmenopausal females n=22, 18	4.1	-3.2
Wk20 % change, postmenopausal females n=22, 18	-5.8	-13.6
Wk20% change - SE, postmenopausal females n=22, 18	-14.7	-22.8

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.4860
	Comments	Overall, Week 80. Log transformed. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Percent difference from metformin
	Estimated Value	5.7
	Confidence Interval	(2-Sided) 95%; -9.6 to 23.6
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.3791
	Comments	Males, Week 80. Log transformed. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Percent difference from metformin
	Estimated Value	12.3
	Confidence Interval	(2-Sided) 95%; -13.7 to 46.1
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.7065
	Comments	Females, Week 80. Log transformed. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Percent difference from metformin
	Estimated Value	4.0
	Confidence Interval	(2-Sided) 95%; -15.3 to 27.6
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.1381
	Comments	Pre-menopausal females, Week 80. Log transformed. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Percent difference from metformin
	Estimated Value	32.1
	Confidence Interval	(2-Sided) 95%; -9.5 to 92.7
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 5

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.7195
	Comments	Postmenopausal females, Week 80. Log transformed. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	.Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Percent difference from metformin
	Estimated Value	-4.3
	Confidence Interval	(2-Sided) 95%; -25.4 to 22.7
	Parameter Dispersion	Type: Value:
	Estimation Comments

28. Secondary Outcome

Title	Percent Change From Baseline in Procollagen Type-1 N-propeptide (P1NP) at Weeks 20, 56, and 80
Description	Blood was taken for measurement of P1NP. Percent change from baseline was based on log transformed data. Standard error, SE; Wk, Week; %, percent. This outcome measure was analyzed for a subset of participants in the bone study only. n is the number of evaluable participants, which is the number of participants with a value at baseline and at the specified visit for the parameter of interest.
Time Frame	Baseline and Weeks 20, 56, and 80

Outcome Measure Data

Analysis Population Description

Week 20 Evaluable for the Bone Sub-study Population. This is a subset of the bone sub-study with LOCF from Week 20. Only participants with P1NP measurements at Week 20 or later were included.

Arm/Group Title	Metformin	Avandamet
Arm/Group Description	Beginning dose of Metformin (MET) 500 milligrams (mg) once a day. Dose could be increased up to a maximum dose of MET 2000 mg.	Beginning dose of Avandamet (AVM) 4 mg/500 mg once a day. Dose could be increased up to a maximum dose of 8 mg/2000 mg.
Measure Participants	72	64
Wk80 % change + SE, overall n= 72, 64	-18.4	-20.7
Wk80 % change, overall n=72, 64	-22.0	-24.3
Wk80 % change - SE, overall n=72, 64	-25.5	-27.8
Wk56 % change + SE, overall n=72, 64	-23.5	-26.6
Wk56 % change, overall n=72, 64	-26.9	-30.0
Wk56 % change - SE, overall n=72, 64	-30.2	-33.3
Wk20 % change + SE, overall n=66, 59	-24.0	-31.6
Wk20 % change, overall n=66, 59	-27.0	-34.6
Wk20 % change - SE, overall n=66, 59	-30.0	-37.5
Wk80 % change + SE, males n=30, 32	-18.5	-18.2
Wk80 % change, males n=30, 32	-24.1	-23.4
Wk80 % change - SE, males n=30, 32	-29.4	-28.2
Wk56 % change + SE, males n=30, 32	-22.7	-23.0
Wk56 % change, males n=30, 32	-27.7	-27.6
Wk56 % change - SE, males n=30, 32	-32.4	-31.9
Wk20 % change + SE, males n=26, 29	-20.2	-24.7
Wk20 % change, males n=26, 29	-25.4	-29.5
Wk20 % change - SE, males n=26, 29	-30.3	-33.9
Wk80 % change + SE, females n=42, 32	-19.3	-19.2
Wk80 % change, females n=42, 32	-23.8	-24.4
Wk80 % change - SE, females n=42, 32	-28.2	-29.3
Wk56 % change + SE, females n=42, 32	-28.6	-27.0
Wk56 % change, females n=42, 32	-32.8	-32.0
Wk56 % change - SE, females n=42, 32	-36.9	-36.6
Wk20 % change + SE, females n=40, 30	-25.2	-31.6
Wk20 % change, females n=40, 30	-30.0	-36.6
Wk20 % change - SE, females n=40, 30	-34.5	-41.2
Wk80 % change + SE, premenopausal females n=19, 10	-16.9	-11.0
Wk80 % change, premenopausal females n=19, 10	-23.6	-20.1
Wk80 % change - SE, premenopausal females n=19, 10	-29.7	-28.3
Wk56 % change + SE, premenopausal females n=19, 10	-25.3	-23.4
Wk56 % change, premenopausal females n=19, 10	-31.0	-30.9
Wk56 % change - SE, premenopausal females n=19, 10	-36.2	-37.6
Wk20 % change + SE, premenopausal females n=17, 10	-24.6	-29.0
Wk20 % change, premenopausal females n=17, 10	-31.4	-36.5
Wk20 % change - SE, premenopausal females n=17, 10	-37.7	-43.1
Wk80 % change+ SE, postmenopausal females n=23, 22	-22.5	-27.6
Wk80 % change, postmenopausal females n=23, 22	-28.4	-33.8
Wk80% change - SE, postmenopausal females n=23, 22	-33.8	-39.4
Wk56 % change+ SE, postmenopausal females n=23, 22	-32.1	-35.2
Wk56 % change, postmenopausal females n=23, 22	-37.9	-41.4
Wk56% change - SE, postmenopausal females n=22, 22	-43.2	-47.0
Wk20 % change+ SE, postmenopausal females n=23, 20	-25.2	-34.0
Wk20 % change, postmenopausal females n=23, 20	-31.9	-40.9
Wk20% change - SE, postmenopausal females n=23, 20	-38.0	-47.1

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.5595
	Comments	Overall, Week 80. Log transformed. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Percent difference from metformin
	Estimated Value	-3.0
	Confidence Interval	(2-Sided) 95%; -12.3 to 7.4
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.9125
	Comments	Males, Week 80. Log transformed. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Percent difference from metformin
	Estimated Value	1.0
	Confidence Interval	(2-Sided) 95%; -15.3 to 20.4
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.9122
	Comments	Females, Week 80. Log transformed. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Percent difference from metformin
	Estimated Value	-0.8
	Confidence Interval	(2-Sided) 95%; -14.0 to 14.5
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.7435
	Comments	Pre-menopausal females, Week 80. Log transformed. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Percent difference from metformin
	Estimated Value	4.6
	Confidence Interval	(2-Sided) 95%; -21.2 to 38.7
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 5

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.3897
	Comments	Postmenopausal females, Week 80. Log transformed. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	.Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Percent difference from metformin
	Estimated Value	-7.5
	Confidence Interval	(2-Sided) 95%; -23.0 to 11.0
	Parameter Dispersion	Type: Value:
	Estimation Comments

29. Secondary Outcome

Title	Percent Change From Baseline in Bone Alkaline Phosphatase (BSAP) at Weeks 20, 56, and 80
Description	Blood was taken for measurement of BSAP. Percent change from baseline was based on log transformed data. Standard error, SE; Wk, Week; %, percent. This outcome measure was analyzed for a subset of participants in the bone study only. n is the number of evaluable participants, which is the number of participants with a value at baseline and at the specified visit for the parameter of interest.
Time Frame	Baseline and Weeks 20, 56, and 80

Outcome Measure Data

Analysis Population Description

Week 20 Evaluable for the Bone Sub-study Population. This is a subset of the bone sub-study with LOCF from Week 20. Only participants with BSAP measurements at Week 20 or later were included.

Arm/Group Title	Metformin	Avandamet
Arm/Group Description	Beginning dose of Metformin (MET) 500 milligrams (mg) once a day. Dose could be increased up to a maximum dose of MET 2000 mg.	Beginning dose of Avandamet (AVM) 4 mg/500 mg once a day. Dose could be increased up to a maximum dose of 8 mg/2000 mg.
Measure Participants	72	64
Wk80 % change + SE, overall n=72, 64	-21.19	-23.30
Wk80 % change, overall n=72, 64	-24.23	-26.37
Wk80 % change - SE, overall n=72, 64	-27.15	-29.31
Wk56 % change + SE, overall n=72, 64	-18.29	-24.39
Wk56 % change, overall n=72, 64	-21.52	-27.48
Wk56 % change - SE, overall n=72, 64	-24.61	-30.45
Wk20 % change + SE, overall n=66, 59	-24.16	-31.02
Wk20 % change, overall n=66, 59	-26.46	-33.31
Wk20 % change - SE, overall n=66, 59	-28.70	-35.53
Wk80 % change + SE, males n=30, 32	-15.44	-21.18
Wk80 % change, males n=30, 32	-20.35	-25.34
Wk80 % change - SE, males n=30, 32	-24.97	-29.27
Wk56 % change + SE, males n=30, 32	-12.22	-21.97
Wk56 % change, males n=30, 32	-17.05	-25.87
Wk56 % change - SE, males n=30, 32	-21.62	-29.57
Wk20 % change + SE, males n=26, 29	-16.27	-26.59
Wk20 % change, males n=26, 29	-20.50	-30.05
Wk20 % change - SE, males n=26, 29	-24.51	-33.35
Wk80 % change + SE, females n=42, 32	-24.30	-23.20
Wk80 % change, females n=42, 32	-28.11	-27.60
Wk80 % change - SE, females n=42, 32	-31.74	-31.74
Wk56 % change + SE, females n=42, 32	-22.54	-24.09
Wk56 % change, females n=42, 32	-26.75	-28.77
Wk56 % change - SE, females n=42, 32	-30.73	-33.17
Wk20 % change + SE, females n=40, 30	-26.82	-31.60
Wk20 % change, females n=40, 30	-30.23	-35.23
Wk20 % change - SE, females n=40, 30	-33.48	-38.66
Wk80 % change + SE, premenopausal females n=19, 10	-13.89	-23.69
Wk80 % change, premenopausal females n=19, 10	-19.28	-29.89
Wk80 % change - SE, premenopausal females n=19, 10	-24.33	-35.59
Wk56 % change + SE, premenopausal females n=19, 10	-12.43	-27.74
Wk56 % change, premenopausal females n=19, 10	-16.84	-32.47
Wk56 % change - SE, premenopausal females n=19, 10	-21.02	-36.89
Wk20 % change + SE, premenopausal females n=17, 10	-24.73	-32.69
Wk20 % change, premenopausal females n=17, 10	-29.57	-37.68
Wk20 % change - SE, premenopausal females n=17, 10	-34.10	-42.32
Wk80 % change+ SE, postmenopausal females n=23, 22	-29.10	-25.82
Wk80 % change, postmenopausal females n=23, 22	-34.39	-31.93
Wk80% change - SE, postmenopausal females n=23, 22	-39.29	-37.54
Wk56 % change+ SE, postmenopausal females n=23, 22	-27.51	-23.61
Wk56 % change, postmenopausal females n=23, 22	-34.04	-31.20
Wk56% change - SE, postmenopausal females n=23, 22	-39.99	-38.04
Wk20 % change+ SE, postmenopausal females n=23, 20	-28.14	-32.02
Wk20 % change, postmenopausal females n=23, 20	-32.70	-36.94
Wk20% change - SE, postmenopausal females n=23, 20	-36.97	-41.50

Statistical Analysis 1

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.5176
	Comments	Overall, Week 80. Log transformed. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Percent difference from metformin
	Estimated Value	-2.83
	Confidence Interval	(2-Sided) 95%; -10.97 to 6.06
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 2

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.3620
	Comments	Males, Week 80. Log transformed. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Percent difference from metformin
	Estimated Value	-6.26
	Confidence Interval	(2-Sided) 95%; -18.62 to 7.97
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 3

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.9115
	Comments	Females, Week 80. Log transformed. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Percent difference from metformin
	Estimated Value	0.72
	Confidence Interval	(2-Sided) 95%; -11.39 to 14.48
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 4

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.1956
	Comments	Pre-menopausal females, Week 80. Log transformed. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Percent difference from metformin
	Estimated Value	-13.51
	Confidence Interval	(2-Sided) 95%; -30.35 to 8.31
	Parameter Dispersion	Type: Value:
	Estimation Comments

Statistical Analysis 5

Statistical Analysis Overview	Comparison Group Selection	Metformin, Avandamet
	Comments
	Type of Statistical Test	Superiority or Other
	Comments
Statistical Test of Hypothesis	p-Value	0.6749
	Comments	Postmenopausal females, Week 80. Log transformed. If the p value was significant at Week 80 a comparison was done at Week 56; if it was significant at Weeks 80 and 56, a comparison was done at Week 20.
	Method	ANCOVA
	Comments	.Based on ANCOVA with terms for treatment, region, pre-screening HbA1c strata, sex, baseline
Method of Estimation	Estimation Parameter	Percent difference from metformin
	Estimated Value	3.75
	Confidence Interval	(2-Sided) 95%; -13.09 to 23.85
	Parameter Dispersion	Type: Value:
	Estimation Comments

Adverse Events

Time Frame	An on therapy adverse event (AE)/serious adverse event (SAE) was defined as an AE/SAE with onset on or after the start date of study medication but not later than two days after the last date of study medication.
Adverse Event Reporting Description	SAEs and AEs were collected in the Safety Population, comprised of all participants who were randomized and received at least one dose of the study medication.
Arm/Group Title	Metformin		Avandamet
Arm/Group Description	Beginning dose of Metformin (MET) 500 milligrams (mg) once a day. Dose could be increased up to a maximum dose of MET 2000 mg.		Beginning dose of Avandamet (AVM) 4 mg/500 mg once a day. Dose could be increased up to a maximum dose of 8 mg/2000 mg.
All Cause Mortality
	Metformin		Avandamet
	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events
Total	/ (NaN)		/ (NaN)
Serious Adverse Events
	Metformin		Avandamet
	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events
Total	27/334 (8.1%)		28/344 (8.1%)
Blood and lymphatic system disorders
Anaemia	0/334 (0%)		1/344 (0.3%)
Iron deficiency anaemia	1/334 (0.3%)		0/344 (0%)
Cardiac disorders
Coronary artery disease	1/334 (0.3%)		2/344 (0.6%)
Angina pectoris	0/334 (0%)		1/344 (0.3%)
Atrial fibrillation	0/334 (0%)		1/344 (0.3%)
Cardiac discomfort	0/334 (0%)		1/344 (0.3%)
Cardiac failure congestive	0/334 (0%)		1/344 (0.3%)
Myocardial infarction	1/334 (0.3%)		1/344 (0.3%)
Myocardial ischaemia	1/334 (0.3%)		0/344 (0%)
Supraventricular tachycardia	1/334 (0.3%)		0/344 (0%)
Eye disorders
Angle closure glaucoma	0/334 (0%)		1/344 (0.3%)
Hepatobiliary disorders
Cholelithiasis	1/334 (0.3%)		0/344 (0%)
Infections and infestations
Pneumonia	2/334 (0.6%)		2/344 (0.6%)
Appendicitis	1/334 (0.3%)		1/344 (0.3%)
Cellulitis	1/334 (0.3%)		2/344 (0.6%)
Gastroenteritis	2/334 (0.6%)		0/344 (0%)
Abdominal abscess	1/334 (0.3%)		0/344 (0%)
Arthritis infective	1/334 (0.3%)		0/344 (0%)
Infected skin ulcer	1/334 (0.3%)		0/344 (0%)
Localized infection	1/334 (0.3%)		0/344 (0%)
Staphylococcal infection	1/334 (0.3%)		0/344 (0%)
Subcutaneous abscess	1/334 (0.3%)		9/344 (2.6%)
Injury, poisoning and procedural complications
Ankle fracture	0/334 (0%)		2/344 (0.6%)
Lumbar vertebral fracture	1/334 (0.3%)		0/344 (0%)
Multiple injuries	1/334 (0.3%)		0/344 (0%)
Road traffic accident	1/334 (0.3%)		0/344 (0%)
Musculoskeletal and connective tissue disorders
Intrevertebral disc protusion	0/334 (0%)		2/344 (0.6%)
Back pain	0/334 (0%)		1/344 (0.3%)
Neoplasms benign, malignant and unspecified (incl cysts and polyps)
Colon cancer	1/334 (0.3%)		0/344 (0%)
Nasal sinus cancer	0/334 (0%)		1/344 (0.3%)
Uterine leiomyoma	0/334 (0%)		1/344 (0.3%)
Nervous system disorders
Brain stem infarction	1/334 (0.3%)		0/344 (0%)
Carotid artery occlusion	1/334 (0.3%)		0/344 (0%)
Cerebral infarction	1/334 (0.3%)		0/344 (0%)
Cerebrovascular accident	0/334 (0%)		1/344 (0.3%)
Cranial neuropathy	1/334 (0.3%)		0/344 (0%)
Haemorrhage itracranial	0/334 (0%)		1/344 (0.3%)
Ischaemic stroke	0/334 (0%)		1/344 (0.3%)
Syncope	0/334 (0%)		1/344 (0.3%)
Transient ischaemic attack	0/334 (0%)		1/344 (0.3%)
Pregnancy, puerperium and perinatal conditions
Abortion spontaneous	0/334 (0%)		1/344 (0.3%)
Renal and urinary disorders
Nephrolithiasis	2/334 (0.6%)		0/344 (0%)
Renal colic	1/334 (0.3%)		0/344 (0%)
Stress urinary incontinence	1/334 (0.3%)		0/344 (0%)
Reproductive system and breast disorders
Prostatomegaly	0/334 (0%)		1/344 (0.3%)
Respiratory, thoracic and mediastinal disorders
Asthma	0/334 (0%)		1/344 (0.3%)
Chronic obstructive pulmonary disease	0/334 (0%)		1/344 (0.3%)
Pleuritic pain	1/334 (0.3%)		0/344 (0%)
Pulmonary embolism	1/334 (0.3%)		0/344 (0%)
Surgical and medical procedures
Inguinal hernia repair	1/334 (0.3%)		0/344 (0%)
Vascular disorders
Hypertensive crisis	0/334 (0%)		1/344 (0.3%)
Varicose vein	0/334 (0%)		1/344 (0.3%)
Other (Not Including Serious) Adverse Events
	Metformin		Avandamet
	Affected / at Risk (%)	# Events	Affected / at Risk (%)	# Events
Total	159/334 (47.6%)		145/344 (42.2%)
Gastrointestinal disorders
Diarrhoea	63/334 (18.9%)		41/344 (11.9%)
Nausea	30/334 (9%)		31/344 (9%)
General disorders
Oedema peripheral	4/334 (1.2%)		16/344 (4.7%)
Infections and infestations
Upper respiratort tract infection	30/334 (9%)		22/344 (6.4%)
Nasopharyngitis	26/334 (7.8%)		18/344 (5.2%)
Urinary tract infection	20/334 (6%)		23/344 (6.7%)
Musculoskeletal and connective tissue disorders
Pain in extremity	22/334 (6.6%)		16/344 (4.7%)
Arthralgia	19/334 (5.7%)		25/344 (7.3%)
Back pain	19/334 (5.7%)		22/344 (6.4%)
Nervous system disorders
Headache	26/334 (7.8%)		19/344 (5.5%)
Dizziness	24/334 (7.2%)		19/344 (5.5%)

Limitations/Caveats

Ten participants who started the study and were randomized did not receive any study medication (6 participants in the metformin arm and 4 participants in the avandamet arm). These participants are not included in any of the analysis populations.

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.