The change from Baseline to Week 26 and Week 52 in HbA1c (the concentration of glucose bound to hemoglobin as a percent of the absolute maximum that can be bound).

Per-protocol set included all randomized patients who received at least 1 dose of double-blind study medication and who had no major protocol violations. Last observation carried forward (LOCF) imputation was utilized.

The change from Baseline in HbA1c (the concentration of glucose bound to hemoglobin as a percent of the absolute maximum that can be bound) during the study. Least Squares Means were from an Analysis of Covariance (ANCOVA) model with treatment, study schedule, and geographic region as class variables, and baseline metformin dose and baseline HbA1c as covariates.

Per-protocol set. Last observation carried forward (LOCF) imputation was utilized.

Clinical response at Weeks 26 and 52 was assessed by the percentage of participants with HbA1c less than or equal to 6.5%.

The full analysis set. Patients who did not complete the scheduled Week 26 or Week 52 visit were assessed based on their response at the time of discontinuation.

Clinical response at Weeks 26 and 52 was assessed by the percentage of participants with HbA1c less than or equal to 7%.

The full analysis set. Patients who did not complete the scheduled Week 26 or Week 52 visit were assessed based on their response at the time of discontinuation.

Clinical response at Weeks 26 and 52 was assessed by the percentage of participants with HbA1c less than or equal to 7.5%.

The full analysis set. Patients who did not complete the scheduled Week 26 or Week 52 visit were assessed based on their response at the time of discontinuation.

Clinical response at Weeks 26 and 52 was assessed by the percentage of participants with a decrease from Baseline in HbA1c of greater than or equal to 0.5%.

The full analysis set. Patients who did not complete the scheduled Week 26 or Week 52 visit were assessed based on their response at the time of discontinuation.

Clinical response at Weeks 26 and 52 was assessed by the percentage of participants with a decrease from Baseline in HbA1c of greater than or equal to 1.0%.

The full analysis set. Patients who did not complete the scheduled Week 26 or Week 52 visit were assessed based on their response at the time of discontinuation.

Clinical response at Weeks 26 and 52 was assessed by the percentage of participants with a decrease from Baseline in HbA1c of greater than or equal to 1.5%.

The full analysis set. Patients who did not complete the scheduled Week 26 or Week 52 visit were assessed based on their response at the time of discontinuation.

Clinical response at Weeks 26 and 52 was assessed by the percentage of participants with a decrease from Baseline in HbA1c of greater than or equal to 2.0%.

The full analysis set. Patients who did not complete the scheduled Week 26 or Week 52 visit were assessed based on their response at the time of discontinuation.

The change from Baseline in fasting plasma glucose (FPG) was assessed at Weeks 2, 4, 8, 12, 16, 20, 26, 34, 42 and 52. Least Squares Means were from an ANCOVA model with treatment, study schedule and geographic region as class variables, and baseline metformin dose and baseline FPG as covariates.

The full analysis set, which included all randomized patients who received at least 1 dose of double-blind study drug and where Baseline and at least 1 postbaseline value were available. Last observation carried forward (LOCF) imputation was utilized.

Marked Hyperglycemia is defined as fasting plasma glucose greater than or equal to 200 mg/dL (11.10 mmol/L).

Full Analysis Set including patients with at least one non-missing fasting plasma glucose result in each treatment group.

Rescue was defined as meeting 1 of the following criteria, confirmed by a 2nd sample drawn within 7 days after the first sample and analyzed by the central laboratory:

1. After more than 2 weeks of treatment but prior to the Week 4 Visit: A single fasting plasma glucose (FPG) ≥275 mg/dL;
2. From the Week 4 Visit but prior to the Week 8 Visit: A single FPG ≥250 mg/dL;
3. From the Week 8 Visit but prior to the Week 12 Visit: A single FPG ≥225 mg/dL;
4. From the Week 12 Visit through the End-of-Treatment Visit: HbA1c ≥8.5% AND ≤0.5% reduction in HbA1c as compared with the baseline HbA1c.

Full Analysis Set including patients with a postbaseline visit.

Proinsulin is a precursor to insulin, and was measured as an indicator of pancreatic function. The change from Baseline in fasting proinsulin was assessed at Weeks 4, 8, 12, 16, 20, 26, 34, 42 and 52. Least Squares Means were from an ANCOVA model with treatment, study schedule and geographic region as class variables, and baseline metformin dose and baseline fasting proinsulin as covariates.

The full analysis set where Baseline and at least 1 postbaseline value were available. Last observation carried forward (LOCF) imputation was utilized.

The change from Baseline in fasting insulin was assessed at Weeks 4, 8, 12, 16, 20, 26, 34, 42 and 52. Least Squares Means were from an ANCOVA model with treatment, study schedule and geographic region as class variables, and baseline metformin dose and baseline fasting insulin as covariates.

The full analysis set where Baseline and at least 1 postbaseline value were available. Last observation carried forward (LOCF) imputation was utilized.

The ratio of proinsulin to insulin was calculated as proinsulin (pmol/L) / insulin (μIU/mL) at weeks 4, 8, 12, 16, 20, 26, 34, 42 and 52 relative to the Baseline value. Least squares means were from an ANCOVA model with treatment, study schedule and geographic region as class variables, and baseline metformin dose and baseline fasting proinsulin/insulin ratio as covariates.

The full analysis set where Baseline and at least 1 postbaseline value were available. Last observation carried forward (LOCF) imputation was utilized.

C-peptide is a byproduct created when the hormone insulin is produced and is measured by a blood test. Change from Baseline was assessed at Weeks 4, 8, 12, 16, 20, 26, 34, 42 and 52. Least squares means are from an ANCOVA model with treatment, study schedule and geographic region as class variables, and baseline metformin dose and baseline fasting C-peptide as covariates.

The full analysis set where Baseline and at least 1 postbaseline value were available. Last observation carried forward (LOCF) imputation was utilized.

The Homeostasis Model Assessment of insulin resistance (HOMA IR) measures insulin resistance based on fasting glucose and insulin measurements:

HOMA IR = fasting plasma insulin (µIU/mL) * fasting plasma glucose (mmol/L) / 22.5

A higher number indicates a greater degree of insulin resistance. The change from Baseline in HOMA IR was assessed at Weeks 12, 26, 42 and 52. Least squares means are from an ANCOVA model with treatment, study schedule and geographic region as class variables, and baseline metformin dose and baseline HOMA insulin resistance as covariates.

The full analysis set where Baseline and at least 1 postbaseline value were available. Last observation carried forward (LOCF) imputation was utilized.

The Homeostasis Model Assessment (HOMA) estimates steady state beta cell function (%B) as a percentage of a normal reference population.

HOMA %B = 20 * insulin (µIU/mL) / fasting plasma glucose (mmol/L) - 3.5

The change from Baseline in the homeostasis model assessment of beta cell function was assessed at Weeks 12, 26, 42 and 52. Least squares means are from an ANCOVA model with treatment, study schedule and geographic region as class variables, and baseline metformin dose and baseline HOMA beta cell function as covariates.

The full analysis set where Baseline and at least 1 postbaseline value were available. Last observation carried forward (LOCF) imputation was utilized.

Change from Baseline in body weight was assessed at Weeks 4, 8, 12, 26, 42 and 52. Least squares means are from an ANCOVA model with treatment, study schedule and geographic region as class variables, and baseline metformin dose and baseline body weight as covariates.

The full analysis set where Baseline and at least 1 postbaseline value were available. Last observation carried forward (LOCF) imputation was utilized.

Change from Baseline in total cholesterol was assessed at Weeks 4, 8, 12, 16, 20, 26, 34, 42 and 52. Least squares means are from an ANCOVA model with treatment, study schedule and geographic region as class variables, and baseline metformin dose and baseline total cholesterol as covariates.

The full analysis set where Baseline and at least 1 postbaseline value were available. Last observation carried forward (LOCF) imputation was utilized.

Change from Baseline in high-density lipoprotein cholesterol (HDL-C) was assessed at Weeks 4, 8, 12, 16, 20, 26, 34, 42 and 52. Least squares means are from an ANCOVA model with treatment, study schedule and geographic region as class variables, and baseline metformin dose and baseline HDL cholesterol as covariates.

The full analysis set where Baseline and at least 1 postbaseline value were available. Last observation carried forward (LOCF) imputation was utilized.

Change from Baseline in low-density lipoprotein cholesterol (LDL-C) was assessed at Weeks 4, 8, 12, 16, 20, 26, 34, 42 and 52. Least squares means are from an ANCOVA model with treatment, study schedule and geographic region as class variables, and baseline metformin dose and baseline LDL cholesterol as covariates.

The full analysis set where Baseline and at least 1 postbaseline value were available. Last observation carried forward (LOCF) imputation was utilized.

Change from Baseline in triglycerides was assessed at Weeks 4, 8, 12, 16, 20, 26, 34, 42 and 52. Least squares means are from an ANCOVA model with treatment, study schedule and geographic region as class variables, and baseline metformin dose and baseline triglycerides as covariates.

The full analysis set where Baseline and at least 1 postbaseline value were available. Last observation carried forward (LOCF) imputation was utilized.

Change from Baseline in free fatty acids was assessed at Weeks 12, 26, 42 and 52. Least squares means are from an ANCOVA model with treatment, study schedule and geographic region as class variables, and baseline metformin dose and baseline free fatty acids as covariates.

The full analysis set where Baseline and at least 1 postbaseline value were available. Last observation carried forward (LOCF) imputation was utilized.

Change from Baseline in Apolipoprotein A1 was assessed at Weeks 12, 26, 42 and 52. Least squares means are from an ANCOVA model with treatment, study schedule and geographic region as class variables, and baseline metformin dose and baseline apolipoprotein A1 as covariates.

The full analysis set where Baseline and at least 1 postbaseline value were available. Last observation carried forward (LOCF) imputation was utilized.

Change from Baseline in Apolipoprotein A2 was assessed at Weeks 12, 26, 42 and 52. Least squares means are from an ANCOVA model with treatment, study schedule and geographic region as class variables, and baseline metformin dose and baseline apolipoprotein A2 as covariates.

The full analysis set where Baseline and at least 1 postbaseline value were available. Last observation carried forward (LOCF) imputation was utilized.

Change from Baseline in Apolipoprotein B was assessed at Weeks 12, 26, 42 and 52. Least squares means are from an ANCOVA model with treatment, study schedule and geographic region as class variables, and baseline metformin dose and baseline apolipoprotein B as covariates.

Full analysis set where Baseline and at least 1 postbaseline value were available. Last observation carried forward (LOCF) imputation was utilized.

Change from Baseline in Apolipoprotein C-III was assessed at Weeks 12, 26, 42 and 52. Least squares means are from an ANCOVA model with treatment, study schedule and geographic region as class variables, and baseline metformin dose and baseline apolipoprotein C-III as covariates.

Full analysis set where Baseline and at least 1 postbaseline value were available. Last observation carried forward (LOCF) imputation was utilized.

Change from Baseline in plasminogen activator inhibitor-1 (PAI-1) was assessed at Weeks 12, 26, 42 and 52. Least squares means are from an ANCOVA model with treatment, study schedule and geographic region as class variables, and baseline metformin dose and baseline PAI-1 as covariates.

The full analysis set where Baseline and at least 1 postbaseline value were available. Last observation carried forward (LOCF) imputation was utilized.

Change from Baseline in high-sensitivity C-Reactive Protein (hsCRP) was assessed at Weeks 12, 26, 42 and 52. Least squares means are from an ANCOVA model with treatment, study schedule and geographic region as class variables, and baseline metformin dose and baseline hsCRP as covariates.

The full analysis set where Baseline and at least 1 postbaseline value were available. Last observation carried forward (LOCF) imputation was utilized.

Change from Baseline in adiponectin was assessed at Weeks 12, 26, 42 and 52. Least squares means are from an ANCOVA model with treatment, study schedule and geographic region as class variables, and baseline metformin dose and baseline adiponectin as covariates.

The full analysis set where Baseline and at least 1 postbaseline value were available. Last observation carried forward (LOCF) imputation was utilized.

Nuclear Magnetic Resonance (NMR) lipid fractionation was used to assess the change from Baseline in total triglyceride levels at Weeks 12, 26, 42 and 52. Least squares means are from an ANCOVA model with treatment, study schedule and geographic region as class variables, and baseline metformin dose and baseline NMR triglycerides as covariates.

The full analysis set where Baseline and at least 1 postbaseline value were available. Last observation carried forward (LOCF) imputation was utilized.

The change from Baseline in levels of total VLDL/chylomicron particles and large VLDL/chylomicron particles was assessed by NMR lipid fractionation at Weeks 12, 26, 42 and 52.

Least squares means are from an ANCOVA model with treatment, study schedule and geographic region as class variables, and baseline metformin dose and baseline VLDL/chylomicron particles as covariates.

The full analysis set where Baseline and at least 1 postbaseline value were available. Last observation carried forward (LOCF) imputation was utilized.

The change from Baseline in levels of VLDL/chylomicron triglycerides was assessed by NMR lipid fractionation at Weeks 12, 26, 42 and 52.

Least squares means are from an ANCOVA model with treatment, study schedule and geographic region as class variables, and baseline metformin dose and baseline VLDL/chylomicron triglycerides as covariates.

The full analysis set where Baseline and at least 1 postbaseline value were available. Last observation carried forward (LOCF) imputation was utilized.

The change from Baseline in levels of medium VLDL particles and small VLDL particles was assessed by NMR fractionation at Weeks 12, 26, 42 and 52. Least squares means are from an ANCOVA model with treatment, study schedule and geographic region as class variables, and baseline metformin dose and baseline VLDL particles as covariates.

The full analysis set where Baseline and at least 1 postbaseline value were available. Last observation carried forward (LOCF) imputation was utilized.

Change from Baseline in mean VLDL particle size was assessed by NMR lipid fractionation at Weeks 12, 26, 42 and 52. Least squares means are from an ANCOVA model with treatment, study schedule and geographic region as class variables, and baseline metformin dose and baseline mean VLDL particle size as covariates.

The full analysis set where Baseline and at least 1 postbaseline value were available. Last observation carried forward (LOCF) imputation was utilized.

The change from Baseline in levels of IDL particles was assessed by NMR lipid fractionation at Weeks 12, 26, 42 and 52. Least squares means are from an ANCOVA model with treatment, study schedule and geographic region as class variables, and baseline metformin dose and baseline IDL particles as covariates.

The full analysis set where Baseline and at least 1 postbaseline value were available. Last observation carried forward (LOCF) imputation was utilized.

The change from Baseline in levels of total, large, medium-small, total small and very small LDL particles was assessed by NMR fractionation at Weeks 12, 26, 42 and 52. Least squares means are from an ANCOVA model with treatment, study schedule and geographic region as class variables, and baseline metformin dose and baseline LDL particles as covariates.

The full analysis set where Baseline and at least 1 postbaseline value were available. Last observation carried forward (LOCF) imputation was utilized.

Change from Baseline in mean LDL particle size was assessed by NMR lipid fractionation at Weeks 12, 26, 42 and 52. Least squares means are from an ANCOVA model with treatment, study schedule and geographic region as class variables, and baseline metformin dose and baseline mean LDL particle size as covariates.

The full analysis set where Baseline and at least 1 postbaseline value were available. Last observation carried forward (LOCF) imputation was utilized.

The change from Baseline in levels of total, large, medium and small HDL particles was assessed by NMR fractionation at Weeks 12, 26, 42 and 52. Least squares means are from an ANCOVA model with treatment, study schedule and geographic region as class variables, and baseline metformin dose and baseline HDL particles as covariates.

The full analysis set where Baseline and at least 1 postbaseline value were available. Last observation carried forward (LOCF) imputation was utilized.

Change from Baseline in mean HDL particle size was assessed by NMR lipid fractionation at Weeks 12, 26, 42 and 52. Least squares means are from an ANCOVA model with treatment, study schedule and geographic region as class variables, and baseline metformin dose and baseline mean HDL particle size as covariates.

The full analysis set where Baseline and at least 1 postbaseline value were available. Last observation carried forward (LOCF) imputation was utilized.