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

1. Trial at a Glance

This page covers the registry results for A Type 2 Diabetes Study of the Longer-Term Glycemic Effect of AVANDAMET vs. Metformin (NCT00386100), a randomized phase 4 study. The primary outcome is Change From Baseline in HbA1c at Week 80.

2. Clinical Question

3. Trial Design

This was a randomized, parallel-group, double-blind study with quadruple masking, registered as Phase 4. The official title describes a multicenter comparison of AVANDAMET and metformin after 80 weeks of treatment. The registry record reports enrollment of 688 participants and a completed status. The recorded study start is 2006-10, and primary completion is recorded as 2009-09.

The study included a 2 week screening period with 2 study visits, followed by an 80 week double-blind treatment period with 11 study visits. The parallel-group design means participants were assigned to a treatment group and followed in that group rather than crossing between treatments. The record does not report the allocation ratio, the randomization sequence method, or the procedures used to maintain masking. It identifies the study as randomized and quadruple masked, but does not specify which roles were masked.

AVANDAMET combined rosiglitazone maleate and metformin hydrochloride. Its described starting total daily dose was 4 mg/500 mg, with possible increases up to 8 mg/2000 mg. Metformin began at a total daily dose of 500 mg and could be increased up to 2000 mg; the dose was to increase unless there was a tolerability issue at the current level. The registry lists several dose levels for both study treatments, but does not give a complete titration schedule or describe how dose decisions were made at each visit.

The stated focus was longer-term glycemic effect. The study also included a bone mineral density sub-study examining changes at the lumbar spine. Other listed outcomes include fasting plasma glucose, achievement of specified HbA1c and glucose thresholds, treatment failure, blood lipids, and adiponectin in a United States and Mexico subset. The record lists 29 outcome measures and 100 analyses. It does not establish which analyses were prespecified as confirmatory beyond identifying the primary outcome.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: 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.

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

6. Primary Endpoint

The registry lists the primary outcome as Change From Baseline in HbA1c at Week 80, measured over Baseline and Week 80.

The primary endpoint was change from baseline in HbA1c at Week 80. HbA1c is a laboratory measure used to assess glycemic status over time, so comparing its change across randomized groups can address whether the groups followed different average glycemic trajectories during the treatment period. The registry record says blood was taken for serum HbA1c measurement and defines the change as the Week 80 value minus the baseline value.

The endpoint is described as a mean of a continuous measure, with standard error as the stated dispersion measure. A mean summarizes the average change among participants included in the analysis, while a standard error describes the statistical precision of an estimated mean. Neither gives the full range of individual participant responses. A confidence interval, if reported, would help show a range of values reasonably compatible with the estimated group difference; the primary outcome description does not provide such an interval.

The registry record labels the endpoint unit as “percent change,” while its description defines a subtraction of the Week 80 value from the baseline value. The record does not explain this wording further. It states that last observation carried forward was not used for this analysis. Group-specific estimates and the between-group contrast are not reported in the registry record, so the direction, magnitude, and precision of the primary comparison cannot be described from the endpoint information available.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does a mean change describe?

The primary outcome is a continuous measure summarized as a mean change from baseline. For each participant with usable measurements, change is defined in the registry record as the Week 80 HbA1c value minus the baseline value. A group mean then summarizes the average participant-level change among those included in the analysis. This summary is useful for comparing average change, but it does not show how variable individual changes were or whether all participants moved in the same direction.

The record identifies standard error as the dispersion measure. A standard error quantifies the estimated precision of a sample mean under the statistical model used. It is not the same as the spread of participant values: a standard deviation describes variation among observations, while a standard error describes uncertainty in an estimated mean. No primary outcome standard errors or group estimates are reported in the registry record, so their magnitudes cannot be interpreted.

How would a between-group estimate be read?

For a randomized comparison, the key efficacy quantity would ordinarily be an estimate of the difference in mean change between the AVANDAMET and metformin groups. The estimate would describe the direction and size of the contrast, and a confidence interval would indicate its precision. A confidence interval is a range constructed by a statistical procedure; it communicates uncertainty around an estimate rather than the range of individual participants’ outcomes. The registry record does not report a between-group estimate or confidence interval for the primary endpoint.

A p-value, if available, would describe how compatible the observed data are with a specified null hypothesis under the analysis assumptions. It would not measure the probability that a treatment is effective, nor would it establish clinical importance by itself. The primary endpoint’s p-value and the hypothesis-testing details are not reported in the registry record. There is therefore no basis in the posted primary outcome information for judging statistical evidence for a group difference.

What is known about the analysis population?

The record describes an Intent-to-Treat population as participants who were randomized, received at least one dose, and had both a baseline and at least one on-therapy value. This definition includes randomized participants meeting those conditions, but it is not a full randomized-population analysis because it excludes participants who do not meet the stated post-randomization measurement requirements. The endpoint description additionally refers to evaluable participants with baseline and at least one post-baseline assessment. The exact number analyzed in each group and the handling of participants with limited follow-up are not reported.

What does the missing-data statement establish?

The record explicitly says that last observation carried forward (LOCF) was not used for the primary analysis. LOCF is a method that replaces a missing later measurement with a participant’s earlier observed measurement. Avoiding LOCF does not by itself identify the alternative approach. The record does not state whether the primary analysis used only observed Week 80 measurements, a model using repeated measurements, another imputation method, or another approach. Those possibilities should not be inferred from the absence of LOCF.

Missing data matter because participants with observed endpoint values may differ from those whose values are missing. If missingness is related to outcomes or treatment experience, an analysis restricted to observed measurements may not represent all randomized participants. General statistical practice is to describe the amount and pattern of missingness and explain assumptions used to address it. Such details are not reported in the registry record for the primary endpoint.

How should multiple outcomes be considered?

The record lists a broad set of secondary outcomes and reports 29 outcome measures and 100 analyses. When many hypotheses are examined, the chance of obtaining at least one apparently notable result can increase if each is tested without accounting for multiplicity. Statistical plans can address this by defining a hierarchy, adjusting significance thresholds, or treating analyses as exploratory. The registry record does not report a multiplicity adjustment, a testing hierarchy, or which secondary analyses were confirmatory. No correction method should therefore be assumed.

Some listed outcomes are continuous changes, while others are counts meeting specified thresholds or treatment failure. These outcome types call for different summaries and statistical models. The primary endpoint is identified as a continuous mean, but the methods for the secondary outcomes are not reported in the registry record. In general, model choice should reflect the outcome’s scale, distribution, and study design; the actual methods used for every analysis cannot be established from the information recorded.

What can be inferred from the design?

Randomization is intended to create groups that are comparable on average at the point of assignment, and masking can reduce certain forms of bias in treatment and outcome assessment. These are design features, not guarantees against all bias. The registry identifies randomization, a parallel-group structure, and quadruple masking, but does not describe the allocation procedure or identify the masked roles. The analysis model, covariate adjustment, handling of protocol deviations, and any interim analyses are not reported in the registry record.

10. Multiplicity, Missing Data and Analysis Populations

The record describes a primary endpoint and a large collection of additional outcomes. It lists 29 outcome measures and 100 analyses, including continuous measures, threshold-based outcomes, treatment failure, and a bone mineral density sub-study. Testing multiple outcomes creates a multiplicity concern: without an appropriate plan, the chance of at least one false-positive finding can rise as more hypotheses are tested. Statistical approaches can control that risk or distinguish confirmatory from exploratory analyses, but the registry record does not report a multiplicity strategy or analysis hierarchy.

The primary analysis population is described as randomized participants who received at least one dose and had a baseline and at least one on-therapy value. The endpoint description also specifies evaluable participants with a baseline and at least one post-baseline assessment. The number of participants meeting these criteria in each group is not reported. The record says LOCF was not used for the primary analysis, but does not identify the replacement method, model, or assumptions used for missing Week 80 observations. It also does not report the pattern or amount of missing primary endpoint data.

These omissions limit interpretation. Complete-case results, model-based estimates, and imputed estimates can rely on different assumptions about participants without observed outcomes. The stated analysis-population definition and the statement about LOCF do not resolve those questions. Details about the full analysis set, protocol deviations, sensitivity analyses, and handling of incomplete follow-up are not reported in the registry record. The statistical plan for interim analyses is also not reported, so no interim procedure can be presumed.

11. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 5% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: subcutaneous abscess (2.6%), pneumonia (0.6%), gastroenteritis (0.6%), nephrolithiasis (0.6%), cellulitis (0.6%).

The serious adverse event figures are counts of participants at risk and participants affected. In the metformin group, 334 participants were at risk and 27 were affected; in the AVANDAMET group, 344 were at risk and 28 were affected. For other adverse events, the corresponding counts were 334 at risk and 159 affected in the metformin group, and 344 at risk and 145 affected in the AVANDAMET group. These are descriptive counts as recorded; they do not by themselves establish a difference between groups or its statistical uncertainty.

The record lists subcutaneous abscess at 2.6% among the top serious adverse event terms. Pneumonia, gastroenteritis, nephrolithiasis, cellulitis, and coronary artery disease are each listed at 0.6%. The registry entry does not define the denominator for these percentages in the information reported, so they should be read as listed rather than recalculated or interpreted as a group-specific risk.

Deaths are not reported for either group in the safety data. The listed information does not provide event timing, duration, recurrence, or a complete account of severity and relatedness. A participant count also does not necessarily represent the total number of events if individuals experienced more than one event; the record does not clarify this for the displayed counts. These figures should therefore be interpreted as a limited descriptive safety summary, not as a complete comparative safety analysis.

12. What the Result Does, and Does Not, Mean

The study addresses a clinically relevant statistical question: how HbA1c changed over a long treatment period in two randomized groups, one assigned to AVANDAMET and one assigned to metformin. The registry record gives the primary endpoint definition and the intended treatment period. It also records a randomized, parallel-group, quadruple-masked design. These details help identify the comparison the study was designed to make.

However, the primary endpoint results are not reported in the registry record. Without the group means, between-group estimate, and uncertainty, the observed comparative change cannot be characterized. The record does not support a statement that one group had a larger or smaller change, or that a difference was statistically distinguishable from chance. The absence of reported estimates is not evidence that the groups were identical, nor does it establish a difference.

The enrollment figure is not the same as the primary analysis sample size. The endpoint description restricts analysis to participants with specified baseline and follow-up measurements, but does not give the number contributing data by group. The missing-data method beyond the statement that LOCF was not used is also not reported. Those details are important because the estimate and its precision depend on the data included and the assumptions used for unavailable measurements.

The listed secondary outcomes and safety counts add context, but they do not fill the primary-outcome gap. The record does not report numerical results for the glycemic secondary outcomes in the information available, and it does not provide a full statistical comparison of adverse events. Counts alone do not show the timing or severity profile of events or quantify uncertainty in group differences.

Accordingly, the registry record can be used to understand the design, endpoint, analysis-population description, and some safety counts. It cannot establish the magnitude or precision of the primary HbA1c comparison. Any broader conclusion about comparative treatment effects would require results and methodological details not reported in the registry record.

13. Why This Trial Matters Statistically

This trial is statistically notable for combining a long treatment period with a randomized comparison of two regimens used in the study’s stated type 2 diabetes context. The primary endpoint was a continuous measure of HbA1c change from baseline at Week 80. A randomized parallel-group design is suited to comparing average outcomes between assigned groups, while masking may reduce some opportunities for treatment knowledge to influence care or outcome assessment. The registry record identifies these design features without giving their operational details.

The endpoint is interpretable in principle as an average within-group change and, most importantly for comparison, a between-group contrast. Precision matters alongside the point estimate: a small estimated difference with narrow uncertainty is not equivalent to the same estimate with wide uncertainty. The registry record does not provide the primary estimates or their uncertainty, so the central statistical comparison cannot be evaluated from the posted endpoint information.

The study also illustrates why analysis populations and missing-data methods need to be read alongside endpoint definitions. The reported population requires baseline and on-therapy data, and LOCF was not used. Neither fact reveals how many participants contributed or what assumptions governed incomplete Week 80 outcomes. Without those details, it is difficult to assess how closely the analyzed participants represent everyone randomized.

The record lists many outcome measures and analyses. That breadth makes prespecification and multiplicity management important: a clearly defined primary outcome can anchor interpretation, while a transparent plan can distinguish confirmatory questions from exploratory ones. The registry record identifies the primary endpoint but does not describe the multiplicity strategy. The statistical importance of this trial therefore lies not only in its design and long-term question, but also in the limits imposed by missing result and analysis details.

14. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

A major limitation of the registry information is that it does not report the primary HbA1c estimates. The stated endpoint is a Week 80 change from baseline, but no group-specific means, between-group contrast, standard errors, confidence intervals, or p-values are included. As a result, the primary comparison’s direction, size, and precision cannot be assessed.

The analysis-population description is not accompanied by the number of participants analyzed in each group. It requires randomization, at least one dose, baseline data, and at least one on-therapy value; the endpoint description also refers to evaluable participants with a post-baseline assessment. The relationship between enrollment and the analyzed population is therefore not fully documented. The record does not provide participant flow, reasons for missing measurements, or details about adherence and discontinuation.

Although the record states that LOCF was not used for the primary analysis, the method used instead is not reported. The amount and pattern of missing endpoint data, assumptions about missingness, and sensitivity analyses are also not described. These gaps prevent assessment of how robust the result might be to different assumptions about unavailable Week 80 values.

Other methods details are absent, including the primary model, any covariate adjustment, the randomization procedure, and a detailed masking description. The record lists many outcomes and analyses but does not state how multiplicity was handled. The listed safety information is limited: deaths are not reported, and details such as timing and event-relatedness are not given. The registry also mentions a lumbar spine bone mineral density sub-study without reporting its results. These limitations concern what is documented in the registry record and do not establish what may be available elsewhere.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: A Type 2 Diabetes Study of the Longer-Term Glycemic Effect of AVANDAMET vs. Metformin — NCT00386100.
• Registry record last updated 2016-11-23; status last verified 2016-10.
• Results first posted 2011-04-05.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

18. Record Summary

NCT00386100 is a completed Phase 4 randomized, parallel-group, quadruple-masked study of AVANDAMET and metformin in type 2 diabetes. The registry record reports enrollment of 688, a 2 week screening period, and an 80 week double-blind treatment period. The primary endpoint was change from baseline in HbA1c at Week 80, defined as the Week 80 value minus baseline; LOCF was not used for this analysis.

The primary endpoint estimates and their uncertainty are not reported in the registry record. The analysis-population description requires a baseline and follow-up measurement, but the number analyzed by group and the method for handling missing Week 80 data are not reported. Safety counts are listed for serious and other adverse events; deaths are not reported. The record also lists secondary glycemic and laboratory outcomes and a lumbar spine bone mineral density sub-study, but does not provide those results in the information available.