Study to Evaluate the Effects of a High-Fat Meal on Bexagliflozin in Healthy Subjects: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Study to Evaluate the Effects of a High-Fat Meal on Bexagliflozin in Healthy Subjects (NCT02820298), a randomized phase 1 study. The primary outcome is Cmax (Maximum Observed Plasma Concentration).

2. Clinical Question

The study focused on a pharmacokinetic question: whether taking bexagliflozin with a high-fat meal was associated with a different peak concentration in blood than taking it without food. It was not designed to assess a clinical outcome in people being treated for Type 2 Diabetes Mellitus. Although that condition appears in the registry record, the study title and summary describe healthy subjects.

The crossover arrangement means participants received both conditions, with the order randomized. One group received the fed condition in treatment period one and the fasted condition in period two; the other group received the reverse order. The record describes a 20 mg dose on day 1 and a second dose on day 8, with overnight fasting before dosing in both conditions.

The primary question is specifically about Cmax, the maximum observed plasma concentration. It does not by itself address other aspects of drug exposure, clinical benefit, or safety beyond the adverse-event counts posted for the study.

3. Trial Design

This was a Phase 1, open-label, randomized, two-period, two-treatment crossover study. The registry record lists the study as completed, with enrollment of 25 and Theracos as the sponsor. The recorded start date is 2016-06-20 and the primary completion date is 2016-08-29.

Participants received bexagliflozin under both fed and fasted conditions. In the first randomized sequence, the fed dose was taken on day 1 and the fasted dose on day 8. In the other sequence, the fasted dose came first and the fed dose came second. Each dose was 20 mg and followed an overnight fast; the distinction was whether food accompanied dosing. The record describes the fed condition as dosing with food, and the study purpose identifies the meal as high-fat.

Because the same participants could contribute measurements under both conditions, the comparison is within-subject in concept. A crossover can account for stable differences between participants when the analysis appropriately models the paired structure. The registry record does not report the analysis model or its terms, so the exact treatment of participant, period, or sequence effects cannot be established from the posted information.

The record identifies no masking. Participants and study personnel were therefore not described as blinded in the registry. The primary endpoint used a pharmacokinetic population: randomized subjects who received study drug and had sufficient bexagliflozin plasma measurements to derive at least one pharmacokinetic parameter following dosing. The number meeting that definition for the primary endpoint is not reported in the registry record as a single population count.

Blood samples were collected before dosing and at the scheduled post-dose times across the two treatment periods. The registry reports 5 outcome measures, but posted numerical results for other outcomes are not available in the record. The study design and population therefore support interpretation of the reported peak-concentration comparison, but not broader conclusions about unreported outcomes.

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: The PK Population included all randomized subjects who received study drug and who had sufficient plasma bexagliflozin measurements to derive at least one PK parameter following dosing.

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 Cmax (Maximum Observed Plasma Concentration), measured over 1, 2, 3, 4, 5, 6, 8, 10, 12, 16, 24, 36 and 48 h post-dose.

The primary endpoint was Cmax, defined as the maximum observed plasma concentration of bexagliflozin. Plasma concentrations were measured from blood samples collected before dosing and at 1, 2, 3, 4, 5, 6, 8, 10, 12, 16, 24, 36 and 48 h after dosing. Samples were collected during days 1-3 of treatment period one and days 8-10 of treatment period two.

Cmax is used in pharmacokinetic studies to describe the peak concentration reached in the measured sampling window. Comparing Cmax between fed and fasted conditions can characterize whether the observed peak differs with meal conditions. It is one part of a drug concentration profile: it does not describe the entire time course, total exposure, or a clinical outcome.

The posted arm summaries use geometric least-squares means, with geometric coefficients of variation as the reported measure of spread. A geometric summary is commonly used for concentration data when values are analyzed on a logarithmic scale, although the registry record does not state the specific model or transformation used. The primary result is a ratio of geometric least-squares means rather than a difference between arithmetic averages.

The endpoint is therefore suited to the study’s stated pharmacokinetic question, but its interpretation is bounded by the sampling schedule, the defined pharmacokinetic population, and the information reported in the registry.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does a crossover design compare?

In a crossover study, participants receive more than one study condition, usually in separate periods. Here, participants received bexagliflozin in both fed and fasted conditions, and the order was randomized. This structure allows a comparison that can use measurements from the same participants under both conditions. Within-person comparisons can be less affected by stable differences between participants than comparisons that assign separate groups to each condition.

A crossover design also requires attention to features that do not arise in the same way in a simple parallel-group comparison. Period effects can occur if measurements differ between the first and second treatment periods for reasons unrelated to the condition. Sequence effects can arise if the order matters. Carryover is a concern when effects from an earlier period persist into a later period. These issues are usually addressed through design choices and an analysis model. The registry record does not report whether or how period, sequence, or carryover were evaluated, so no particular adjustment or diagnostic procedure can be attributed to this study.

Why report geometric least-squares means?

Concentration measurements are often analyzed on a logarithmic scale because such data can be asymmetric and because changes may be more naturally expressed proportionally than as fixed absolute differences. A geometric mean summarizes values after that scale transformation and conversion back to the original scale. A least-squares mean is a model-based mean adjusted for the terms included in the statistical model. In general, the model matters: its specification determines what is adjusted for and how uncertainty is estimated. The registry identifies the result as a geometric least-squares mean but does not identify the model or its terms.

The fed and fasted summaries are 176.1 ng/mL and 133.7 ng/mL, respectively. Their reported comparison is a ratio of geometric least-squares means, 131.41. Ratios express the relative scale of two summaries, whereas a difference expresses their separation on the original measurement scale. The reported ratio should be interpreted as the registry’s comparative estimate, not as a recalculation from the arm summaries. The record does not provide enough model detail to reconstruct the analysis independently.

What does the confidence interval convey?

A confidence interval gives a range of values compatible with the observed data and the statistical procedure, under the procedure’s assumptions. A narrower interval indicates greater precision than a wider interval on the same scale; a wide interval signals that the estimate is less precise. The reported 90% confidence interval for the ratio extends from 117.03 to 147.56. It reflects uncertainty in the estimated comparison, not variation among individual concentrations and not a guarantee that the true value falls inside the interval.

Confidence intervals are useful because they show both the direction and the precision of an estimate. A point estimate alone can make a comparison appear more definitive than the data warrant. Interpretation still depends on the analysis method and on the target question. In this study, the record does not describe the model used to calculate the interval, so its construction cannot be verified from the registry entry.

How does equivalence differ from detecting a difference?

An equivalence analysis asks whether an estimated effect is sufficiently close to a prespecified reference value to fall within acceptable bounds. Those bounds must be chosen in advance and justified for the scientific question. Equivalence is not the same as failing to detect a difference: a lack of evidence for a difference does not establish that two conditions are sufficiently similar. The registry labels its comparison as equivalence, but the equivalence bounds and their rationale are not reported in the registry record. The label alone is not enough to judge whether the required criterion was met.

The posted comparison has a 90% confidence interval, but the record does not provide the decision rule or margins against which that interval was assessed. It also reports no p-value. The result can therefore be described numerically, while a definitive assessment of the equivalence procedure remains limited by the missing protocol and analysis details.

What does the coefficient of variation describe?

A coefficient of variation expresses dispersion relative to the scale of the measurement. The registry reports geometric coefficients of variation of 30.6 in the fed condition and 42.5 in the fasted condition. These values describe spread in the observed concentration summaries using a relative measure. They are not confidence intervals, and they do not identify the range of values for individual participants. The exact calculation and any model-based interpretation are not detailed in the registry record.

Relative dispersion can be useful for concentration data because variability may scale with the level of the measurement. In general, greater variability reduces precision unless offset by other features of the design or analysis. A crossover may help by comparing conditions within a participant, but the registry does not report the within-participant variability or the fitted model, so the reported group-level coefficients should not be mistaken for the uncertainty of the paired treatment comparison.

What can planning illustrations tell a reader?

Sample-size planning illustrations based on an observed effect are not the study’s original design justification and do not describe what the trial was powered to detect when planned. They are conditional calculations: their implications depend on assumptions such as the target effect, allocation, significance level, and statistical approximation. In general, planning values should be read as illustrations for a hypothetical future design, not as observed power or evidence that the current study had a particular probability of detecting an effect.

The registry record does not report the study’s own sample-size justification. It also does not report the planning assumptions used by the original investigators. Any planning interpretation must therefore remain distinct from the reported trial design and from the estimated pharmacokinetic comparison.

10. Analyses Reported by the Sponsor

The registry reports a ratio of geometric least-squares means of 131.41, with a 90% confidence interval from 117.03 to 147.56, and labels the analysis as equivalence. No statistical method, p-value, or equivalence bounds are reported alongside that result. The registry record therefore permits description of the sponsor’s posted estimate and interval, but does not provide enough information to reconstruct the underlying model or fully assess the equivalence decision.

No discrepancy between a sponsor result and a recalculated result is flagged in the registry record. The posted arm summaries and the reported ratio are different summaries of the data: the arm values are geometric least-squares means, while the comparison is a ratio. The reported ratio should not be replaced by a calculation from the arm summaries, particularly because the crossover structure and the model-based nature of the estimates matter to the comparison. The record does not identify a separate recalculated estimate for this result.

For interpretation, the sponsor’s posted estimate and confidence interval are the available inferential comparison. Their scope is the Cmax contrast between fed and fasted bexagliflozin dosing in the study’s pharmacokinetic population. Without the equivalence margins and analysis specification, readers cannot independently verify the equivalence claim from the registry entry alone.

11. Multiplicity, Missing Data and Analysis Populations

The registry lists 5 outcome measures, but it does not provide posted numerical results for secondary outcomes. It reports no secondary outcomes with posted results. For another endpoint, the record notes that values are missing or not estimable for a two-arm comparison. The available information therefore centers on the primary Cmax result; it does not support a complete review of all planned or measured outcomes.

When a study assesses multiple outcomes or makes multiple comparisons, the chance of at least one false-positive finding can increase if each result is interpreted separately without accounting for multiplicity. Statistical plans may specify a primary outcome, testing order, adjustment procedure, or another strategy for controlling this issue. Such a multiplicity plan is not reported in the registry record. There is also no basis to infer whether analyses were adjusted across outcomes or whether only selected outcomes were posted.

The registry defines the primary pharmacokinetic population as randomized subjects who received study drug and had sufficient plasma measurements to derive at least one pharmacokinetic parameter following dosing. This definition identifies who could contribute to the analysis, but the record does not report the full analysis population count for the primary endpoint or explain how missing samples and incomplete concentration profiles were handled.

The arm summaries list 23 participants for the fed condition and 24 for the fasted condition, compared with total enrollment of 25. These counts do not, on their own, explain participant flow, missing measurements, or the number contributing paired data to the crossover comparison. The record does not report reasons for missing data, imputation procedures, exclusions from the pharmacokinetic population, or sensitivity analyses addressing missingness.

In general, missing pharmacokinetic measurements can affect endpoint calculation if the missing samples include times relevant to identifying the observed maximum. The impact depends on the pattern of missingness and the analysis rules. Neither can be determined from the registry entry. Accordingly, the reported estimates should be read as results for the analysis represented in the record, with the population and missing-data details incompletely documented.

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 3% reporting threshold.

Most frequent non-serious adverse event terms

The serious adverse event figures are 0 among 23 participants in the fed condition and 0 among 24 participants in the fasted condition. Deaths are also reported as 0 in both conditions. These counts describe the posted safety record for the two dosing conditions; they do not establish that no adverse events of any kind occurred.

Other adverse events are reported for 5 of 23 participants in the fed condition and 1 of 24 participants in the fasted condition. The record does not provide the terms for these events or further detail about severity, timing, or attribution. It lists no top serious adverse event terms, consistent with the absence of serious events in the posted counts.

Safety counts in a small Phase 1 study are descriptive. A zero count means that no event meeting the serious-event definition was recorded among the participants represented in that condition; it does not rule out an event in a larger or different population. The registry record does not report a formal statistical comparison of adverse events between conditions. Because participants received both conditions in a crossover design, the counts should not be treated as if they necessarily came from independent groups. No conclusion about comparative safety can be drawn from these counts alone.

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

The result contributes a focused estimate of how the observed peak plasma concentration of bexagliflozin compared when dosing took place with a high-fat meal versus without food. The posted geometric least-squares mean summaries are 176.1 ng/mL for fed dosing and 133.7 ng/mL for fasted dosing. The reported ratio is 131.41, with a 90% confidence interval from 117.03 to 147.56. These values describe the registry’s primary pharmacokinetic comparison in healthy subjects.

The result does not establish that the two dosing conditions are equivalent. The registry labels the analysis as equivalence, but it does not report the equivalence bounds needed to assess that claim. Nor does the estimate alone establish a clinically important effect: the study reports a concentration endpoint, not a clinical outcome, and the registry does not state how a concentration difference should be interpreted for clinical use.

The result does not describe every aspect of bexagliflozin exposure. Cmax is a peak measure and does not, on its own, summarize total exposure or the complete concentration-time profile. Although samples were collected at multiple post-dose times, other endpoint results are not numerically posted in the registry record.

The study also does not directly answer how the comparison applies to people with Type 2 Diabetes Mellitus. The title and study summary describe healthy subjects, even though Type 2 Diabetes Mellitus appears in the record’s condition listing. Extrapolation beyond the studied population is not supported by this result alone.

Finally, the reported comparison depends on an analysis method that is not specified in the registry record. The result can be read as the posted estimate and interval for the fed-versus-fasted Cmax contrast, while recognizing that model details, equivalence criteria, participant disposition, and missing-data handling are not reported.

14. Why This Trial Matters Statistically

This trial is statistically informative because it uses a randomized crossover design to frame a within-person pharmacokinetic comparison. Rather than assigning one set of participants only to fed dosing and another only to fasted dosing, the design had participants receive both conditions in randomized order. In general, this structure can reduce the influence of stable differences between individuals when the analysis correctly accounts for the paired observations.

The case also illustrates why the label attached to an analysis is not a substitute for its decision criteria. The registry calls the primary comparison an equivalence analysis and reports a ratio with a 90% confidence interval. Without the prespecified equivalence margins, readers cannot determine from the record whether the interval met the study’s intended equivalence criterion. Reporting the estimate and interval is valuable, but full interpretation requires the criterion against which they were judged.

It also shows the distinction between descriptive summaries and an inferential comparison. The fed and fasted geometric least-squares means describe the endpoint summaries, while the ratio is the reported comparison. In a crossover study, a model-based contrast may account for pairing and design features that are not captured by simply comparing the two arm summaries.

Finally, the record demonstrates the limits of registry results for a complete statistical assessment. The primary outcome is reported, but the analysis method, multiplicity approach, missing-data handling, and equivalence margins are not. Those omissions do not erase the posted estimate; they define what can and cannot be concluded from it.

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 does not report the statistical model used for the primary comparison. In a crossover study, details such as handling of participant pairing, period, and treatment sequence can affect interpretation. The record also does not explain whether carryover was considered. These omissions limit independent assessment of the reported ratio and confidence interval.

Although the analysis is labeled as equivalence, the equivalence margins and their justification are not reported. The reader therefore cannot evaluate whether the confidence interval satisfied the intended criterion. No p-value is reported for the primary comparison, and the record gives no additional decision rule.

The primary population is defined by receipt of study drug and availability of sufficient plasma measurements to derive at least one pharmacokinetic parameter. The record does not provide a detailed participant flow or explain why endpoint counts differ from total enrollment. It also does not report how incomplete concentration profiles, missing samples, or exclusions were handled, or whether sensitivity analyses were performed.

The registry lists 5 outcome measures but provides no numerical secondary results. It does not report a multiplicity plan, so the handling of multiple outcomes or analyses cannot be assessed. The arm summaries include 23 participants in the fed condition and 24 in the fasted condition, but do not establish how many participants contributed complete paired observations to the reported comparison.

The study is described as involving healthy subjects, limiting the direct scope of the result. The condition listing includes Type 2 Diabetes Mellitus, but the registry title and summary identify healthy subjects. Finally, the posted safety information gives event counts without detailed terms for non-serious events, and the record does not report a formal comparison of safety outcomes.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Study to Evaluate the Effects of a High-Fat Meal on Bexagliflozin in Healthy Subjects — NCT02820298.
• Registry record last updated 2021-06-10; status last verified 2021-05.
• Results first posted 2021-04-28.
• 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

NCT02820298 is a completed Phase 1 randomized, open-label, two-period crossover study sponsored by Theracos. It enrolled 25 participants and examined bexagliflozin pharmacokinetics in healthy subjects under fed and fasted dosing conditions. Participants received a 20 mg dose in each condition, in randomized order.

The primary endpoint was Cmax, the maximum observed plasma concentration. The posted geometric least-squares means were 176.1 ng/mL in the fed condition and 133.7 ng/mL in the fasted condition. The reported ratio was 131.41, with a 90% confidence interval from 117.03 to 147.56; the analysis is labeled as equivalence.

Serious adverse events and deaths were reported as 0 in both conditions. Other adverse events were reported for 5 participants in the fed condition and 1 in the fasted condition. The registry does not report the equivalence margins, primary analysis method, multiplicity plan, or detailed missing-data handling.