Bexagliflozin Efficacy and Safety Trial: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Bexagliflozin Efficacy and Safety Trial (NCT02558296), a randomized phase 3 study. The primary outcome is Change in HbA1c From Baseline to Week 24.

2. Clinical Question

The primary efficacy objective was to evaluate the change in HbA1c after 24 weeks of treatment. HbA1c is a continuous measure, and the posted arm results are least-squares means with 95% confidence intervals. The sponsor’s reported comparison is a net mean difference from a mixed-effects repeated-measures analysis.

The clinical question is specific to the endpoint and follow-up period recorded for the trial. It is not a question about cardiovascular outcomes: the registry summary says data from this study would be combined with other bexagliflozin studies in a meta-analysis of cardiovascular safety outcomes. The primary endpoint reported for this trial is HbA1c change, while the listed secondary outcomes include HbA1c among subjects prescribed insulin, body-weight change, and systolic blood-pressure change in subjects hypertensive at baseline.

The trial record does not report a threshold for a clinically important HbA1c difference, nor does it provide a basis for interpreting the endpoint as a direct measure of cardiovascular events. The posted estimates therefore describe the recorded HbA1c comparison, rather than establishing a broader clinical conclusion.

3. Trial Design

This was a Phase 3 randomized, quadruple-masked, parallel-group study comparing bexagliflozin with placebo. The registry lists an enrollment of 1700 and a completed status. The recorded start was 2015-10, and primary completion was 2018-08. Theracos is listed as the sponsor.

The active-treatment arm received bexagliflozin tablets, 20 mg once daily, and the comparator arm received placebo tablets once daily. The registry describes both interventions as drugs and states that each subject would receive the assigned tablet daily for the duration of the study. The primary endpoint was change in HbA1c from baseline to Week 24, measured as percentage of glycated hemoglobin.

Randomization and masking describe features of the study design, but the registry record does not report the randomization procedure, concealment details, or which specific roles were included in quadruple masking. It also does not report the duration of treatment beyond the stated once-daily schedule, or details about adherence and treatment discontinuation.

The registry’s enrollment figure describes the trial record; it should not be treated as the number contributing to every endpoint. For the primary endpoint, the stated population comprises subjects with a value at baseline and at Week 24. The posted arm counts for that endpoint are 1046.0 for bexagliflozin and 531.0 for placebo. A separate adverse-event summary lists 1132 at risk in the bexagliflozin group and 567 at risk in the placebo group. These figures correspond to different posted summaries, and the registry record does not explain all reasons for the differences.

The study’s stated purpose was to investigate HbA1c lowering in people with type 2 diabetes and increased cardiovascular risk. Cardiovascular safety outcomes were intended for combination with data from other bexagliflozin studies, rather than being the primary endpoint reported in this trial.

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.

Before assignment: Subjects ≥ 40 years old with inadequately controlled T2DM with HbA1c between 7.5% (7.0% since protocol version 8) and 11% on stable medications and elevated risk for CV adverse events were enrolled. All subjects must belong to 1 of 3 CV risk groups to be eligible. All eligible subjects took placebo during 13 ± 2 days run-in period.

Completed the Study Through Week 24

Reasons for not completing this period:

Completed the Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Number of subjects with a value at baseline and at Week 24.

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 in HbA1c From Baseline to Week 24, measured over 24 weeks.

The primary endpoint is change in HbA1c from baseline to Week 24. HbA1c is a continuous measure expressed in the registry as percentage of glycated hemoglobin. The endpoint compares each subject’s follow-up measurement with their own baseline measurement, then summarizes the change by treatment group.

The registry reports least-squares mean changes for the two arms. The bexagliflozin estimate is -0.85, with a 95% confidence interval from -0.9 to -0.8. The placebo estimate is -0.37, with a 95% confidence interval from -0.44 to -0.3. The sponsor’s reported net mean difference is -0.48, with a 95% confidence interval from -0.56 to -0.39.

A continuous endpoint retains information about the magnitude of change, rather than reducing each participant’s measurement to a category. HbA1c is used in this trial because the registered primary objective was to evaluate its change during treatment. The registry does not report a clinical-importance threshold for interpreting the size of the difference, so statistical estimates alone do not establish how important the magnitude is for patients.

The endpoint population is defined as subjects with measurements at baseline and Week 24. How missing measurements were handled in the sponsor’s model, beyond the stated mixed-effects repeated-measures method, is not reported in the registry record. The registry also does not report additional details about endpoint measurement procedures.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is continuous, and the registry reports least-squares means and a mean difference. These summaries describe average changes, while the confidence intervals indicate uncertainty around the estimated averages and comparison. The sponsor’s reported method is mixed-effects repeated measures. The registry record does not provide the full model specification.

What does a least-squares mean represent?

In general, a least-squares mean is a model-based estimate of a group mean. It can account for terms included in the statistical model, unlike a simple arithmetic average calculated directly from observed values. The exact covariates and other model terms used in this trial are not reported in the registry record. The posted estimates should therefore be read as the registry’s reported least-squares means, without assuming which adjustments were made.

The arm estimates are -0.85 for bexagliflozin and -0.37 for placebo. Each has its own 95% confidence interval. These arm-specific intervals describe uncertainty in each estimated mean; they should not be substituted for the interval on the between-group comparison. The sponsor reports the net mean difference separately, with its own 95% confidence interval from -0.56 to -0.39.

Why use repeated-measures analysis?

In general, repeated-measures methods are designed for outcomes measured on participants over time. A mixed-effects repeated-measures model can use the structure of repeated observations and account for correlation among measurements from the same participant. Such models can also accommodate unequal numbers of observed measurements under specified assumptions. The registry record does not state the full measurement schedule used in the primary analysis, the model’s covariance structure, or the assumptions used for missing observations. Those details should not be inferred from the method’s name alone.

Over-dispersion is a general concern when observed variability exceeds what a chosen model assumes. It can make model-based standard errors too small if the extra variability is not accounted for, leading to confidence intervals that are too narrow. Over-dispersion is especially discussed for count outcomes; the primary endpoint in this trial is continuous. The record does not report an over-dispersion assessment, and there is no basis to say that it was or was not a concern in this analysis.

How should the confidence interval be read?

In general, a confidence interval communicates precision for an estimate under the statistical model. A narrower interval indicates greater precision than a wider interval, but it does not rule out bias, model misspecification, or limitations in the measurement and analysis process. The confidence interval for the sponsor’s net difference is -0.56 to -0.39 at the 95% level. This interval concerns the average between-group estimate, not the range of plausible changes for individual participants.

The sponsor also reports p < 0.001 and classifies the analysis as superiority. A p-value is calculated relative to a statistical model and a specified null hypothesis. In general, it measures how incompatible the observed data are with that null model; it does not quantify the probability that the result is due to chance, and it does not establish clinical importance. The registry record does not report whether the p-value was adjusted for multiple testing.

What do the planning assumptions mean?

The registry describes sample-size calculations as a planning illustration computed from the observed effect, not from the trial’s own design. The stated assumptions are a two-sided significance level of 0.05, equal allocation between arms, a normal approximation, and an effect to detect set as a stated fraction of the observed effect. In general, such calculations illustrate the sample size associated with selected assumptions; they are not evidence about the trial’s actual design process and are never observed power. The registry record does not report the trial’s original sample-size justification or planning assumptions.

For non-time-to-event outcomes, the described planning quantities are participants per arm. The registry’s general note distinguishes these from time-to-event calculations, which give total events. This primary endpoint is continuous, and the planning illustration does not change the interpretation of the observed endpoint estimates.

What does randomization contribute?

In general, random assignment is intended to make treatment groups comparable on average at the start of a trial, including for characteristics that may not have been measured. Masking is intended to reduce the influence of expectations on care, measurement, or assessment. These design features can strengthen a comparison, but they do not eliminate every possible source of bias. The registry lists randomized allocation and quadruple masking; details of how allocation and masking were implemented are not reported in the registry record.

10. Analyses Reported by the Sponsor

The registry reports the sponsor’s own primary comparison as a mixed-effects repeated-measures analysis. It estimates a net mean difference of -0.48, with a 95% confidence interval from -0.56 to -0.39 and p < 0.001, and labels the result a superiority analysis. The posted arm-level least-squares means are -0.85 for bexagliflozin and -0.37 for placebo.

The registry record does not report a separate recalculated between-group estimate, so no second comparison can be set alongside the sponsor’s result. The arm-level estimates are descriptive summaries and do not replace the sponsor’s model-based contrast. No discrepancy is flagged between the sponsor analysis and a recalculation in the registry record. The sponsor’s own analysis is therefore the reported inferential result.

The sponsor’s estimate should still be interpreted within its stated endpoint and analysis population: change in HbA1c from baseline to Week 24 among subjects with values at both time points. The confidence interval reflects statistical precision under the reported method, not uncertainty about every limitation of the study. The registry does not report the model’s full specification, detailed missing-data handling, or whether multiplicity adjustments were used.

11. Multiplicity, Missing Data and Analysis Populations

The registry lists 11 outcome measures, 32 analyses, and 3 posted secondary outcomes. Multiple outcomes and analyses create the possibility that many statistical comparisons were considered. In general, the more comparisons made, the greater the chance of observing a small p-value even when there is no underlying difference for a particular comparison. Methods such as prespecifying a primary hypothesis or adjusting for multiplicity can address parts of this issue, but the registry record does not report a multiplicity strategy or identify how the analyses were prioritized beyond naming the primary endpoint.

The three listed secondary outcomes concern HbA1c change at Week 24 for subjects prescribed insulin, body-weight change from baseline to Week 48, and systolic blood-pressure change from baseline to Week 24 among subjects hypertensive at baseline. Their presence does not establish that each was confirmatory or that inferential testing was adjusted across them. The registry record does not report the detailed statistical plan, the testing hierarchy, or whether outcomes were added or changed after study initiation.

The primary endpoint population is described as subjects with a value at baseline and at Week 24. This is a definition based on observed measurements, not a full description of missing-data handling. The sponsor’s method is identified as mixed-effects repeated measures, but the assumptions used for unobserved outcomes and any sensitivity analyses are not reported in the registry record. The numbers and reasons for missing primary endpoint measurements are also not reported in the registry record.

The registry does not report an intention-to-treat definition, a per-protocol population, or other analysis-population rules beyond the stated endpoint population. It does not report how treatment discontinuation, deviations from assigned treatment, or use of additional medications affected inclusion in the primary analysis. Those omissions limit assessment of how closely the analysis reflects the randomized groups and how robust the result is to missing data.

These gaps do not invalidate the posted estimate, but they restrict what can be evaluated independently from the registry record. A reader can identify the reported comparison and its precision, while recognizing that the multiplicity and missing-data procedures remain insufficiently described.

12. Safety Results

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

Deaths

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: angina unstable (3.4%), acute myocardial infarction (2.7%), cardiac failure congestive (2.6%), myocardial infarction (2.3%), angina pectoris (1.9%).

The adverse-event summary lists serious adverse events affecting 373 subjects among 1132 at risk in the bexagliflozin group, and 208 subjects among 567 at risk in the placebo group. It also lists 739 subjects with other adverse events among 1132 at risk in the bexagliflozin group, and 386 among 567 at risk in the placebo group. These are posted counts; no rates or relative comparisons are calculated.

The summary lists 39 deaths in the bexagliflozin group and 26 in the placebo group. The registry record does not report causes of death, timing, adjudication details, or whether deaths were considered related to study treatment. Counts alone do not establish a causal relationship, and they do not account for differences in follow-up or exposure.

The listed serious-event terms include unstable angina at 3.4%, acute myocardial infarction at 2.7%, congestive cardiac failure at 2.6%, myocardial infarction at 2.3%, angina pectoris at 1.9%, and cardiac failure at 1.9%. The registry does not report the denominator or further coding details for these percentages, nor does it clarify whether terms may overlap for a subject.

These figures are descriptive safety summaries, not proof that an event was caused by a study drug. The registry record does not provide the detailed event-level information needed to evaluate timing, severity categories, relatedness assessments, or adjusted comparisons.

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

The posted primary endpoint estimates describe average HbA1c change from baseline to Week 24. Both arms have negative estimated changes, and the sponsor’s reported net mean difference is -0.48, with a 95% confidence interval from -0.56 to -0.39. The sponsor reports p < 0.001 and classifies the comparison as superiority. These results characterize the recorded statistical comparison; they do not, by themselves, establish how meaningful the magnitude is for patients.

The registry does not state a threshold for a clinically important difference in HbA1c. Statistical evidence against a no-difference model is not the same as evidence about the practical importance of an effect. The confidence interval provides information about precision under the analysis assumptions, but it does not show individual responses or resolve concerns about unreported model details, missing data, or multiplicity.

The primary endpoint is HbA1c, not cardiovascular events. The trial summary says that data from this study would be combined with other bexagliflozin studies in a meta-analysis of cardiovascular safety outcomes. The record does not report those combined results, so the HbA1c comparison should not be read as a conclusion about cardiovascular safety or cardiovascular benefit.

The safety summary records adverse-event and death counts in both groups, but it does not provide enough context for causal interpretation. It does not report the event-level timing, relatedness, or adjusted analysis needed to attribute events to treatment. Likewise, the endpoint results do not establish that each participant experienced a change similar to the group mean.

The record supports a focused statistical statement: the sponsor’s analysis estimated a more negative average HbA1c change in the bexagliflozin group than in the placebo group over the recorded primary endpoint period. The record does not support broader claims about clinical importance, individual benefit, or cardiovascular outcomes.

14. Why This Trial Matters Statistically

This trial is statistically notable for the way a continuous primary outcome is summarized alongside a model-based comparison. The arm-level least-squares means describe average HbA1c change, while the sponsor’s mixed-effects repeated-measures analysis supplies the between-group estimate and its confidence interval. Keeping those quantities distinct helps prevent an arm estimate from being mistaken for the treatment contrast.

The analysis also illustrates why a small p-value is only one part of interpretation. The sponsor reports p < 0.001, but the estimate’s magnitude, precision, endpoint definition, and model assumptions are all relevant. The registry does not report a clinical-importance threshold or the full model specification, so the statistical result has limits as a guide to practical importance.

The endpoint population is defined by having a value at baseline and at Week 24. Meanwhile, the safety summary uses separate at-risk counts, and the registry does not explain all differences between these populations. Recognizing that endpoint and safety summaries may use different populations is important when reading trial records; it avoids treating every posted count as if it described the same analysis set.

The record also distinguishes the trial’s primary HbA1c objective from the stated intention to combine data from other studies for cardiovascular safety outcomes. That distinction matters because an endpoint cannot answer questions outside its scope. The study’s results may contribute to a broader evidence base, but the registry record does not contain the planned meta-analysis results.

Finally, the planning assumptions are explicitly described as an illustration based on the observed effect, rather than as the trial’s own design calculation. This is an important distinction in statistical education: a retrospective planning exercise is not observed power and cannot reconstruct the original design rationale.

15. 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.

The registry record contains the primary endpoint estimates and the sponsor’s reported comparison, but does not provide the full analysis plan or model specification. Covariates, model structure, covariance assumptions, and details of repeated observations are not reported in the registry record. This limits independent assessment of how the least-squares means and contrast were estimated.

The primary endpoint population requires measurements at baseline and Week 24. The registry record does not report the number or reasons for missing endpoint measurements, how missing values were handled beyond identifying a mixed-effects repeated-measures method, or whether sensitivity analyses were performed. It also does not report an intention-to-treat definition or other population rules for the primary analysis.

The registry lists 11 outcome measures and 32 analyses, but a multiplicity plan, testing hierarchy, or adjustment procedure is not reported in the registry record. Without that information, the status of the reported p-value among the full set of analyses cannot be fully assessed. The record also does not report whether the listed outcomes or analyses were changed during the study.

The primary endpoint is a surrogate measure of glycemic change, and the registry does not report a threshold for clinical importance. Cardiovascular safety outcomes were intended for meta-analysis with other studies; those combined results are not reported in this record. The primary HbA1c finding therefore cannot answer cardiovascular outcome questions.

Safety figures are counts, and the record does not provide event timing, relatedness assessments, causes of death, or enough detail to evaluate differences in exposure and follow-up. The denominator for the percentages attached to the listed serious-event terms is not reported in the registry record. These omissions limit causal and comparative safety interpretation.

The registry does not report the randomization procedure, masking implementation details, adherence, treatment discontinuation, or the original sample-size justification. These limits do not negate the reported results, but they define what can be judged from the registry record alone.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Bexagliflozin Efficacy and Safety Trial — NCT02558296.
• Registry record last updated 2021-07-14; status last verified 2021-07.
• Results first posted 2021-07-07.
• 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.

19. Record Summary

The Bexagliflozin Efficacy and Safety Trial was a completed Phase 3 randomized, quadruple-masked, parallel-group study sponsored by Theracos. It enrolled 1700 and compared bexagliflozin tablets, 20 mg once daily, with placebo once daily in people with type 2 diabetes and increased cardiovascular risk.

The primary endpoint was change in HbA1c from baseline to Week 24. The posted least-squares mean changes were -0.85 for bexagliflozin and -0.37 for placebo. The sponsor’s mixed-effects repeated-measures analysis reported a net mean difference of -0.48, with a 95% confidence interval from -0.56 to -0.39 and p < 0.001.

The registry record describes a separate intention to combine study data with other bexagliflozin studies for cardiovascular safety outcomes. It does not report that combined analysis, a clinical-importance threshold for the primary endpoint, or full details of the statistical and missing-data procedures.