Demonstration of Bioequivalence of Empagliflozin and Metformin Given in One Tablet Compared to the Intake of Single Tablets: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Demonstration of Bioequivalence of Empagliflozin and Metformin Given in One Tablet Compared to the Intake of Single Tablets (NCT01672788), a randomized phase 1 study. The primary outcome is Empa: Area Under the Curve 0 to Infinity (AUC0-∞).

2. Clinical Question

3. Trial Design

This was a Phase 1, randomized, open-label, single-dose, four-way crossover study in healthy volunteers. The stated enrollment was 36. A crossover design compares regimens through repeated treatment experiences within participants, rather than assigning each participant to only one treatment throughout. In general, this can reduce the influence of stable differences between people when comparing pharmacokinetic outcomes. The registry record identifies the design as crossover but does not report the treatment sequence details, the number or duration of periods, or any washout interval.

The study compared fixed-dose combination tablets with empagliflozin and metformin tablets administered together. Its stated objective was to establish bioequivalence after a high-fat, high-calorie meal. The posted primary endpoint concerns empagliflozin AUC0-∞. The record also lists AUC0-tz as a secondary outcome for empagliflozin and for metformin. Although multiple intervention entries appear in the registry record, the posted analyses present two empagliflozin comparisons, corresponding to Empa 12.5mg and Empa 5mg.

The primary endpoint analysis population is described as a pharmacokinetic set. It included participants meeting specified dosing, observation, and protocol-compliance conditions, but the full definition is not available in the registry record. The displayed description also begins a vomiting-related criterion and then ends before the rule is complete. The record does not report a detailed account of exclusions from the analysis population or how missing pharmacokinetic observations were handled.

Randomization describes how treatment assignment or treatment sequence was allocated; it does not by itself establish that the participants, investigators, or analysts were unaware of treatment. The study is explicitly described as open-label, with no masking. The registry record lists the study status as completed and gives the start as 2012-08 and primary completion as 2012-11. It does not report a sample-size justification, interim analyses, or additional design details beyond those stated.

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: This was an open-label, randomized, 4-way crossover trial. 36 patients were randomised to one of 4 treatment sequences and treated. Each treatment period consisted of a single dose of medication followed by 72 hours of pharmacokinetic sampling, with a washout of at least 7 days between treatment periods.

Treatment Period 1 (3 Days)

Washout Period 1 (7 Days)

Treatment Period 2 (3 Days)

Washout Period 2 (7 Days)

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Pharmacokinetic (PK) set: All subjects who took at least 1 dose of investigational treatment, provided at least 1 observation for at least 1 primary PK endpoint, did not have important protocol violations relevant to the evaluation of relative bioavailability, did not vomit at or before 2 times median tmax and did not use restricted medications

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 Empa: Area Under the Curve 0 to Infinity (AUC0-∞), measured over 1 hour (h) before drug administration and 20 minutes (min), 40min, 1h, 1.5h, 2h, 2.5h, 3h, 4h, 6h, 8h, 10h, 12h, 24h, 34h, 48h and 72h after drug administration.

The primary endpoint is empagliflozin area under the plasma concentration-time curve from zero extrapolated to infinity, abbreviated AUC0-∞. The area under a concentration-time curve represents drug exposure over time: it combines the measured plasma concentrations across the follow-up period and extrapolates the remaining area beyond the last relevant observation to infinity. The registry record describes measurements beginning 1 hour before administration and continuing at specified times through 72h after drug administration.

AUC is used in pharmacokinetic comparisons because it provides an overall measure of systemic exposure, rather than describing concentration at just one sampling time. In a bioequivalence comparison, exposure from a test formulation is compared with exposure from a reference formulation. The posted analysis expresses this comparison as a geometric mean ratio. A ratio near 100 indicates that the estimated geometric means are close on the scale used for the comparison; the uncertainty around each estimate is described by its confidence interval.

The record reports geometric coefficient of variation as the dispersion measure and notes that measured values show inter-individual variability, while the statistical analyses show intra-individual variability. These represent different sources of variation: differences between participants and variation within participants across treatment experiences. AUC0-tz for empagliflozin and metformin is listed as a secondary outcome, but the posted primary result is the empagliflozin AUC0-∞ comparison. The registry record does not report the equivalence margins or a formal conclusion tied to those margins.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry record identifies ANOVA as the method used for the two posted comparisons and reports each result as a geometric mean ratio. ANOVA, or analysis of variance, is a statistical framework for comparing treatment means while accounting for features specified in a model. In a crossover study, a suitable analysis can account for the fact that observations from the same participant are related. The exact model terms, transformations, and covariance structure used in this trial are not reported in the registry record, so the detailed implementation cannot be reconstructed from the method label alone.

What does a geometric mean ratio express?

A geometric mean summarizes values on a multiplicative scale, which is often useful for pharmacokinetic measurements that vary proportionally rather than by a constant additive amount. A geometric mean ratio compares the test regimen's geometric mean with the reference regimen's geometric mean. On the percentage-style scale used in the record, 100 corresponds to equal geometric means. A value above or below 100 indicates the direction of the estimated relative exposure, but the estimate should be interpreted together with its uncertainty and the prespecified decision rule.

The reported ratios are 101.31 for Empa 12.5mg and 100.30 for Empa 5mg. These are estimates for the respective test-reference comparisons, not percentages of participants and not measures of individual-level response. No new ratio or contrast is calculated from the arm summaries in this explanation. The registry record's reported analyses are the basis for the interpretation.

What does a 90% confidence interval mean?

A confidence interval gives a range of values compatible with the data and the analysis under the method's assumptions. Its width conveys precision: a narrower interval generally indicates more precise estimation, while a wider interval reflects greater uncertainty. A confidence interval is not a range containing a specified proportion of individual participants' measurements, and it does not mean that the true value has a probability equal to the confidence level of lying inside this particular calculated interval.

The posted intervals are 96.89 to 105.93 for Empa 12.5mg and 97.40 to 103.29 for Empa 5mg, each at the 90% confidence level. Both include 100, the ratio value for equal geometric means. Whether either interval falls within an acceptable equivalence range cannot be determined without the study's prespecified limits. Those limits are not reported in the registry record. This distinction matters because formal equivalence is assessed by the interval's relationship to defined boundaries, not simply by whether the point estimate is close to 100 or whether the interval includes 100.

Why use a crossover comparison?

In general, a crossover design allows each participant to receive more than one regimen, so comparisons can use within-participant information. This can make a treatment comparison less dependent on between-participant differences than a comparison of separate parallel groups. It also makes the analysis dependent on appropriate handling of the crossover structure, including treatment sequence and any period effects. The record identifies a four-way crossover but does not report the sequence allocation details, period terms in the model, or washout information. Those details should not be inferred from the design label.

Randomization supports unbiased assignment of treatment sequences when implemented as planned, but it does not remove every possible source of bias. The study was open-label, so treatment masking was not used. Pharmacokinetic concentration measurements may be more objective than some subjective outcomes, but the record does not describe laboratory blinding or the conduct of measurements. Such details are not reported in the registry record.

How should variability be read?

The primary endpoint description distinguishes inter-individual variability in the measured values from intra-individual variability in the statistical analyses. Inter-individual variability refers to differences among participants. Intra-individual variability refers to differences across a participant's treatment experiences. The reported geometric coefficient of variation summarizes relative dispersion on a multiplicative scale. It is not the confidence interval for the treatment ratio and should not be interpreted as uncertainty in the same way.

For pharmacokinetic data, variability can be multiplicative: values may spread by proportional rather than fixed amounts. General statistical methods for such data often analyze transformed values and then express treatment contrasts as ratios. The record identifies geometric means, geometric coefficients of variation, and ANOVA, but does not report the transformation or model specification. The general explanation does not establish which exact computational details were used.

What is not available for checking the analysis?

The registry record does not report the equivalence margins, the full statistical model, the treatment sequence details, or p-values for the posted comparisons. It also does not provide the complete primary pharmacokinetic analysis-population definition: the displayed description ends during a vomiting-related condition. The registry entries include two comparison results, but do not provide all endpoint estimates for the listed secondary outcomes. These omissions limit independent reconstruction of the analysis and the formal interpretation, even though the principal ratio estimates and confidence intervals are reported.

10. Analyses Reported by the Sponsor

The registry record posts the sponsor's ANOVA results as geometric mean ratios with 90% confidence intervals: 101.31 (96.89 to 105.93) for Empa 12.5mg and 100.30 (97.40 to 103.29) for Empa 5mg. The registry entries available for this analysis do not include a separate recalculated estimate or confidence interval for these comparisons. No discrepancy is flagged between the posted results and any recalculation, so there is no numerical conflict to resolve.

The sponsor's analysis is therefore the reported inferential analysis for these comparisons. Its point estimates are close to 100, and its intervals provide the posted measure of precision. The registry record does not state the prespecified equivalence limits, however, so the formal equivalence determination cannot be checked from these results alone. Nor does it provide p-values for the comparisons. The ratios and intervals should be reported as the sponsor's posted estimates, without extending them into a pass-or-fail conclusion that depends on information absent from the registry record.

11. Multiplicity, Missing Data and Analysis Populations

The registry record lists a primary empagliflozin endpoint, AUC0-∞, and secondary AUC0-tz outcomes for empagliflozin and metformin. It also reports two primary comparisons, one for Empa 12.5mg and one for Empa 5mg. The number of listed outcomes and comparisons creates more than one potential inferential question, but the registry record does not report a multiplicity-adjustment procedure or explain how the primary comparisons relate to any overall decision rule.

Multiplicity is a general statistical concern when several hypotheses or endpoints are assessed: without an appropriate plan, the chance of at least one apparently notable result can increase. Whether and how adjustment is needed depends on the study's goals and prespecified testing strategy. The existence of multiple comparisons alone does not establish that an adjustment was required or omitted. The registry record does not report a statistical plan that would allow this issue to be resolved.

Missing-data handling is also not reported in the registry record. The primary endpoint population description refers to participants who took at least one dose, contributed at least one observation for at least one primary pharmacokinetic endpoint, and had no important protocol violations relevant to relative bioavailability. The displayed definition is incomplete: it ends during a vomiting-related criterion. The full rule, the number excluded from each analysis, and the handling of missing concentrations or unevaluable endpoints are not reported.

The registry record reports enrollment of 36, while the displayed arm-level pharmacokinetic sample sizes are 34, 34, 35, and 33. These are summaries for the listed arms and should not be treated as a complete account of participant flow or as evidence of why any records were absent. The record does not provide the sequence-specific analysis counts or a detailed disposition of participants. It also does not report interim analyses, the sample-size justification, or a complete statistical analysis plan. Accordingly, the posted results support interpretation of the reported estimates but not a full audit of endpoint selection, multiplicity control, or missing-data procedures.

12. 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 non-serious adverse event terms

The safety summaries report zero serious adverse events in each of the four listed groups. The numbers at risk for serious adverse events are 35 for Empa 12.5mg Fixed-dose, 36 for Empa 12.5mg Free Dose, 34 for Empa 5mg Fixed-dose, and 35 for Empa 5mg Free Dose. These are group-specific figures and are not the same as the arm-level pharmacokinetic sample sizes.

The record also reports other adverse events affecting 14 participants in Empa 12.5mg Fixed-dose, 13 in Empa 12.5mg Free Dose, 14 in Empa 5mg Fixed-dose, and 9 in Empa 5mg Free Dose. These counts describe the posted safety summaries; they do not establish that every event was caused by treatment. The registry record does not list top serious adverse event terms. Death figures are not reported in the registry record.

Zero serious events in these groups is a description of the observed record, not proof that serious events cannot occur. Safety interpretation is limited by the study's healthy-volunteer setting, single-dose design, and the information available in the registry record. No broader safety conclusion follows from these counts alone.

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

The posted primary results estimate relative empagliflozin exposure for two fixed-dose combination comparisons against empagliflozin and metformin administered as single tablets together. The geometric mean ratios are 101.31 for Empa 12.5mg and 100.30 for Empa 5mg. Their 90% confidence intervals are 96.89 to 105.93 and 97.40 to 103.29, respectively. These values describe the estimated relative AUC0-∞ and its uncertainty in the trial's pharmacokinetic analysis.

The results do not, by themselves, establish a formal bioequivalence decision. That decision depends on prespecified acceptance limits, which are not reported in the registry record. The intervals include the equality value of 100, but whether their endpoints satisfy the applicable limits cannot be assessed without knowing those limits. The registry record does not report p-values for these analyses, and a p-value would not replace the missing equivalence criteria.

The findings are specific to the registered study conditions: healthy male and female volunteers, single-dose administration, and fed conditions after a high-fat, high-calorie meal. They do not directly describe repeated dosing, other populations, other meal conditions, or outcomes beyond the measured pharmacokinetic endpoints. The secondary AUC0-tz outcomes are listed, but their results are not reported in the registry entries available for this analysis.

Nor do the results establish clinical benefit, long-term safety, or comparative performance for outcomes other than drug exposure. The trial's stated question concerns bioequivalence, and the posted endpoint is pharmacokinetic. The results should therefore be read as estimates of exposure ratios with specified confidence intervals, not as evidence about treatment effects on clinical outcomes. Further detail on the exact statistical model, the complete analysis population, the equivalence limits, and missing-data procedures is not reported in the registry record.

14. Why This Trial Matters Statistically

This trial illustrates why the statistic used in a bioequivalence analysis differs from a simple comparison of two arm averages. The primary results are expressed as geometric mean ratios, which match the multiplicative nature of many pharmacokinetic comparisons. The confidence intervals make precision visible and permit assessment against prespecified equivalence limits when those limits are available.

The crossover design also makes the distinction between inter-individual and intra-individual variability important. The endpoint description explicitly notes that measured values show inter-individual variability, while the statistical analyses show intra-individual variability. These describe different sources of variation and should not be conflated. A reader who treats the arm-level dispersion measure as uncertainty around the ratio would be interpreting different quantities as though they were the same.

The two posted estimates are close to 100, but formal interpretation requires more than a point estimate. The intervals and acceptance limits must be considered together. Because the limits are not reported in the registry record, this example shows how a record can provide informative estimates yet leave the formal decision rule unverifiable. The same caution applies to the missing detail about the full analysis population and statistical model.

Finally, the registry lists both primary and secondary pharmacokinetic outcomes and two comparisons. That makes it useful to distinguish the outcome being interpreted from other listed outcomes and to ask what testing strategy was planned. The multiplicity approach is not reported. In this way, the record offers a practical example of the difference between reading a numerical result accurately and assessing whether the complete analysis plan can be independently evaluated.

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 main limitation for interpreting the primary result is that the registry record does not state the equivalence margins. Without the acceptance limits, the reported confidence intervals cannot be judged against the formal criterion for bioequivalence. The record also does not report p-values for the comparisons. These omissions prevent reconstruction of a complete decision statement from the posted numerical results.

Further details needed to audit the analysis are not reported in the registry record, including the full ANOVA model, transformation details, treatment sequence information, period handling, and washout interval. The record identifies a four-way crossover but does not give the operational details needed to assess those features. It also does not report a sample-size justification or interim analyses.

The primary pharmacokinetic analysis-population description is incomplete in the registry record. It identifies dosing, observation, and protocol-compliance conditions, then ends during a vomiting-related criterion. The full criterion, participant flow into each analysis, and handling of missing or unevaluable measurements are not reported. The posted arm-level sample sizes therefore should not be used to infer the reasons for differences in analysis counts.

The results are also limited to the stated setting: healthy volunteers, single-dose treatment, and administration after a high-fat, high-calorie meal. The record lists secondary AUC0-tz outcomes, but their estimates are not available in the registry record for this analysis. Safety summaries are limited to the reported event counts and do not provide serious-event terms or deaths. These are limits of what the registry record reports, rather than evidence that additional procedures or outcomes were absent.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Demonstration of Bioequivalence of Empagliflozin and Metformin Given in One Tablet Compared to the Intake of Single Tablets — NCT01672788.
• Registry record last updated 2015-08-31; status last verified 2015-08.
• Results first posted 2015-08-31.
• 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 registry record describes NCT01672788 as a completed Phase 1, randomized, open-label, single-dose, four-way crossover study in healthy male and female volunteers. Enrollment was 36. The study compared empagliflozin/metformin fixed-dose combination tablets with empagliflozin and metformin tablets taken together under fed conditions after a high-fat, high-calorie meal.

The primary endpoint was empagliflozin AUC0-∞. The posted ANOVA geometric mean ratios were 101.31 (90% confidence interval, 96.89 to 105.93) for Empa 12.5mg and 100.30 (90% confidence interval, 97.40 to 103.29) for Empa 5mg. The formal equivalence limits are not reported in the registry record, so the intervals cannot be evaluated against a stated acceptance range. The record also lists secondary AUC0-tz outcomes for empagliflozin and metformin, with results not reported in the registry record for this analysis.