This Study Tests Whether Taking the Medicines Empagliflozin, Linagliptin, and Metformin Together in 1 Pill is the Same as Taking Them in Separate Pills. The Study is Done in Healthy Men and Women and Measures the Amount of Each Medicine in the Blood: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for This Study Tests Whether Taking the Medicines Empagliflozin, Linagliptin, and Metformin Together in 1 Pill is the Same as Taking Them in Separate Pills. The Study is Done in Healthy Men and Women and Measures the Amount of Each Medicine in the Blood (NCT03259490), a randomized phase 1 study. The primary outcome is Area Under the Concentration-time Curve of the Analyte in Plasma Over the Time Interval From 0 to the Last Quantifiable Data Point (AUC0-tz) for Empagliflozin.

2. Clinical Question

3. Trial Design

This was a randomized, open-label, single-dose, two-period, two-sequence crossover study. In a crossover design, participants receive both treatments in different periods, with randomized sequence determining the order. This allows each participant's exposure under one formulation to be compared with that same participant's exposure under the other formulation, when the required measurements are available. The design description does not report the length of the periods or a washout interval; those details are not reported in the registry record.

The study compared a fixed-dose combination tablet of empagliflozin, linagliptin, and extended-release metformin with the free combination of the three medicines as separate tablets. Both were given as a single dose under fed conditions. The study was open label, meaning masking was not used. The stated purpose was to establish bioequivalence, and the primary endpoint assessed empagliflozin plasma exposure.

The registry record lists 30 participants enrolled. The primary pharmacokinetic analysis population was the PKS, defined in the endpoint description by treatment and availability of at least one eligible primary or secondary pharmacokinetic parameter. The posted arm results each report 29 participants. Because the analysis population definition allows a participant to contribute a parameter from one period even if the corresponding parameter from the other period is unavailable, the arm counts alone do not establish that every participant contributed a complete paired comparison.

The study is marked completed and is classified as Phase 1. Its population was healthy participants, not people selected for a clinical condition. Consequently, the primary question concerns relative drug exposure under the study conditions. The registry record does not report additional protocol details such as the randomization method, planned washout, or sample-size justification.

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 trial was a two-way crossover study. All subjects received the 2 treatments in randomised order.

Period 1

Reasons for not completing this period:

Period 2

Reasons for not completing this period:

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

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.

Population: Treated set (TS): This subject set included all subjects who were documented to have received one dose of study drug.

6. Primary Endpoint

The registry lists the primary outcome as Area Under the Concentration-time Curve of the Analyte in Plasma Over the Time Interval From 0 to the Last Quantifiable Data Point (AUC0-tz) for Empagliflozin, measured over Pharmacokinetic samples were collected at 1:30 hours: minutes pre dose and 0:30, 1:00, 1:30, 2:00, 2:30, 3:00, 4:00, 5:00, 6:00, 7:00, 8:00, 10:00, 12:00, 24:00, 34:00, 48:00 and 72:00 hours:minutes post dose.

The primary endpoint was empagliflozin AUC0-tz: the area under the plasma concentration-time curve from time zero to the last quantifiable data point. It summarizes measured plasma exposure across the sampled portion of the concentration-time profile. The unit was nanomoles*hours/ litres (nmol*h/L), and the endpoint was summarized using a geometric mean.

AUC is used in pharmacokinetic comparisons because it captures exposure across multiple sampling times, rather than relying on a single concentration measurement. In a formulation comparison, the relative AUC can indicate whether the amount of measured exposure is similar or different between formulations. It does not by itself describe every aspect of the concentration profile, such as the timing or height of a peak.

Samples were scheduled at 1:30 hours: minutes before dosing and at 0:30, 1:00, 1:30, 2:00, 2:30, 3:00, 4:00, 5:00, 6:00, 7:00, 8:00, 10:00, 12:00, 24:00, 34:00, 48:00 and 72:00 hours:minutes after dosing. The endpoint description states that a subject's concentration data or parameter could be considered non-evaluable in circumstances including early emesis, a predose concentration greater than 5% of that subject's Cmax, or missing samples at important phases of the pharmacokinetic disposition curve. These rules concern whether measurements could contribute to the endpoint; the registry record does not give counts for each exclusion reason.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary comparison used an analysis of variance (ANOVA) and reported an adjusted geometric mean ratio of test to reference. The ratio is expressed as a percentage, with 100% representing equal geometric means on the ratio scale. Values above 100% indicate a higher estimated geometric mean for the test relative to the reference; values below 100% indicate a lower one. This is a general interpretation of a ratio, not a statement about clinical benefit.

Why use a geometric mean?

Pharmacokinetic measurements are often analyzed on a logarithmic scale because exposure measures can be right-skewed and because treatment comparisons are naturally expressed as ratios. In general, a geometric mean is the exponentiated mean of log-transformed observations. A ratio of geometric means therefore describes relative exposure rather than an absolute difference in the endpoint's units. The registry reports geometric means for each treatment and an adjusted ratio from ANOVA, but does not give the full model specification or transformation details in the information.

What does the confidence interval convey?

A confidence interval gives a range of values compatible with the observed data and the analysis assumptions at its stated confidence level. The 95% interval for the adjusted ratio is 100.36% to 105.83%, around a point estimate of 103.06%. It conveys uncertainty in the estimate; it is not a range containing 95% of individual participants' exposure values, nor does it give the probability that the true ratio lies inside the interval. For formal bioequivalence assessment, the interval is ordinarily interpreted against prespecified acceptance limits. Such limits are not reported in the registry record, so no pass-or-fail conclusion against a particular range can be drawn from this information alone.

Why does the crossover structure matter?

In a crossover comparison, each participant can serve as their own control because both formulations are administered in different periods. This can account for stable differences between participants and is one reason the design is commonly used for pharmacokinetic comparisons. The analysis still needs to address the trial's period and sequence structure as appropriate. The registry identifies ANOVA but does not report the model terms, treatment sequence results, or period-specific findings. The crossover label alone does not establish which adjustments were included in the model.

The PKS definition also matters. It includes treated participants with at least one primary or secondary pharmacokinetic parameter that was not excluded. The definition explicitly allows a participant to contribute a parameter from one period even if the participant contributes only one period's value to the statistical assessment. Thus, the reported 29 participants in each arm should not automatically be read as 29 complete pairs. The record does not give the number of complete pairs for this endpoint.

How should the arm summaries be read?

The posted geometric means are 5656.0 for the fixed-dose combination and 5488.0 for the separate-tablet reference. The reported spread is 1.04 for each. The endpoint lists standard error as its dispersion measure, but the registry excerpt does not make clear how that spread is represented for these geometric-mean summaries. The adjusted ratio and its confidence interval are the more direct statistics for the stated treatment comparison. The separate summaries are descriptive and do not replace the model-based comparison.

The registry reports no p-value for the primary comparison. In a bioequivalence setting, a confidence interval assessed against a prespecified interval is often more informative than a test of whether the ratio differs from 100%. No specific equivalence interval or statistical decision rule is reported in the registry record, however. A missing p-value should not be interpreted as evidence for or against equivalence.

What do planning illustrations mean?

Planning illustrations based on an observed effect are not observed power and are not the trial's original sample-size justification. The stated illustration assumptions are a two-sided significance level of 0.05, equal allocation, and a normal approximation, with the target effect defined as a fraction of the observed effect. The actual numerical planning results and the trial's own sample-size justification are not reported in the registry record. Such illustrations are hypothetical planning quantities and should not be used to describe the study's achieved power.

Finally, statistical interpretation depends on more than the point estimate. The endpoint's evaluability rules, completeness of sampling, analysis population, model specification, confidence interval, and prespecified decision criteria all influence how a formulation comparison should be understood. Several of those details are not reported in the registry record. The posted estimate remains interpretable as a model-based relative exposure result, but the available information does not support reconstructing the full analysis or its decision framework.

10. Analyses Reported by the Sponsor

The registry reports the sponsor's primary comparison as an ANOVA-adjusted geometric mean ratio of 103.06%, with a 95% confidence interval from 100.36% to 105.83%. That model-based comparison is the relevant result for describing the sponsor's analysis. The two arm summaries, 5656.0 and 5488.0, are descriptive geometric means; a direct comparison of those summaries would not reproduce all adjustments in the ANOVA analysis.

The registry record does not report a p-value for this endpoint, and it does not give the model specification or a bioequivalence acceptance interval in the information. Therefore, the adjusted ratio and confidence interval can be reported, but the record does not support a further statement that a specified formal criterion was met. No conflicting sponsor and recalculated estimates are identified in the registry information for this comparison. The sponsor's reported estimate should not be replaced with an unadjusted calculation from the posted arm summaries.

11. Multiplicity, Missing Data and Analysis Populations

The record lists one primary endpoint: empagliflozin AUC0-tz. It also lists three secondary AUC outcomes, defined from time zero extrapolated to infinity for empagliflozin, metformin, and linagliptin. The registry reports nine outcome measures and nine analyses overall. These counts indicate that multiple outcomes or analyses were posted, but do not by themselves identify which analyses were confirmatory or how conclusions were prioritized.

The registry information does not report a multiplicity-control procedure, a hierarchy among endpoints, or whether secondary outcomes were considered exploratory. It also does not report whether any adjustment was used to account for multiple comparisons. Without those details, the primary endpoint should be distinguished from the secondary measures, and separate results should not be treated as if they necessarily had equal confirmatory status.

For missing or non-evaluable pharmacokinetic data, the primary endpoint description gives examples of exclusion conditions: emesis at or before two times the median tmax for the relevant treatment, a predose concentration greater than 5% of the participant's Cmax, or missing samples at important phases of the pharmacokinetic disposition curve. It defines the PKS as treated participants who posted at least one eligible primary or secondary parameter, and allows contribution of a value from only one period. The counts excluded for each reason, the extent of missing sampling, and any imputation approach are not reported in the registry record.

Accordingly, the arm counts of 29 do not establish that every participant had complete paired data for the primary comparison. The record does not report a separate missing-data sensitivity analysis, the number of complete pairs, or additional analysis populations for this endpoint. These omissions limit assessment of how robust the estimate is to exclusions or incomplete period data. They do not invalidate the reported result, but they leave uncertainty about the detailed handling of incomplete observations and the broader family of analyses.

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 non-serious adverse event terms

The registry reports zero serious adverse events among 29 participants in the fixed-dose combination group and zero among 29 participants in the separate-tablet reference group. It also reports zero deaths in each group. These are the serious-event and death counts recorded for the study; they do not establish that no event of any kind occurred.

For other adverse events, the record reports 11 affected participants in the test group and 6 in the reference group, each with 29 at risk. These are descriptive counts. The registry information does not give event-level detail in the serious-event term list, which is empty, and the counts alone do not establish whether an event was caused by a study treatment. No inferential comparison of these safety counts is reported.

This was a single-dose study in healthy participants. The safety figures should therefore be read in that limited context, rather than as evidence about longer exposure or other populations. Zero serious events in the posted counts is not proof that serious events cannot occur; it describes what was recorded among the participants at risk in this study.

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

The result is a comparison of empagliflozin exposure after two ways of taking the medicines: one fixed-dose combination tablet or the medicines in separate tablets. Under the study conditions—a single dose in healthy male and female participants after food—the adjusted estimate for empagliflozin AUC0-tz was 103.06% for test relative to reference, with a 95% confidence interval of 100.36% to 105.83%. This communicates both the estimated relative exposure and the statistical uncertainty around it.

The result does not, by itself, establish that the formulations satisfy a particular bioequivalence standard. That determination requires the relevant prespecified acceptance limits and decision rules, which are not reported in the registry record information. Nor does the result establish that the formulations are identical in every pharmacokinetic feature: AUC0-tz addresses exposure through the last quantifiable sample, not every feature of the concentration-time curve.

The trial also does not answer questions about repeated dosing, long-term outcomes, people with a clinical condition, or comparative clinical effectiveness. It was conducted in healthy participants and measured pharmacokinetic outcomes after a single dose. The result should not be extended beyond that population, dose context, and endpoint without evidence from other sources.

The safety counts are similarly limited. The registry records no serious adverse events or deaths in either group, and records other adverse events in both groups. Those observations describe this study's participants and duration; they are not evidence that events cannot occur in other settings. The absence of a reported p-value for the primary comparison and the absence of a reported bioequivalence margin also mean that the registry information does not support an independent formal pass-or-fail assessment.

14. Why This Trial Matters Statistically

This study illustrates why formulation comparisons are not usually interpreted by comparing two raw-looking group summaries alone. The central quantity is the test-to-reference ratio of geometric mean exposure, estimated through the reported ANOVA. The ratio provides a relative measure, while the confidence interval communicates the estimate's precision. Both pieces are necessary for understanding the primary pharmacokinetic result.

The crossover design is also statistically important. Because participants receive both formulations in different periods, comparisons can use within-participant information rather than relying only on differences between separate groups. At the same time, interpretation depends on which observations were evaluable and how the model accounted for the trial structure. The PKS definition allows a contribution from one period, so the posted arm counts do not establish that all analyzed participants contributed complete pairs.

The example further shows that a narrow-looking interval and an estimate near 100% are not, by themselves, a substitute for a formal equivalence rule. A decision needs prespecified limits and an analysis plan. Those details are not reported in the registry record information. The distinction between an estimate, its confidence interval, and the rule used to interpret them is central to sound statistical reading.

Finally, planning illustrations derived from an observed effect answer a different question from the original design justification. They are hypothetical sample-size calculations, not observed power. Keeping that distinction clear prevents a retrospective planning illustration from being mistaken for a feature of the completed trial's design or evidence about the reliability of its result.

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 provides the main design features and the primary comparison but does not provide the complete statistical analysis plan. In particular, the ANOVA model terms, transformation details, randomization procedure, period and sequence findings, and planned bioequivalence margins are not reported in the information. Those omissions limit independent assessment of the analysis and prevent a formal classification against an acceptance range.

The endpoint's evaluability criteria are described, but the record does not report how many observations were excluded under each rule, how many participants contributed complete paired data, or the extent of missing concentration samples. The PKS definition permits a contribution from only one period. The posted 29-participant counts for each treatment therefore should not be taken as confirmation of a complete within-participant dataset for all participants.

The study was open label, enrolled healthy participants, and used a single dose under fed conditions. Its findings are specific to that setting and the measured pharmacokinetic outcomes. The registry record does not provide results for repeated administration or a clinical outcome comparison. It also does not report the trial's sample-size justification or a multiplicity strategy for the nine posted outcome measures and analyses.

For safety, the registry reports event counts but does not provide a fuller account of event types in the serious-event term list or a formal comparison of the counts. The small study context and single-dose exposure also limit what can be inferred from the absence of recorded serious events. These limitations concern the scope and detail of the registry information; they do not change the numerical result that was posted for the primary endpoint.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: This Study Tests Whether Taking the Medicines Empagliflozin, Linagliptin, and Metformin Together in 1 Pill is the Same as Taking Them in Separate Pills. The Study is Done in Healthy Men and Women and Measures the Amount of Each Medicine in the Blood — NCT03259490.
• Registry record last updated 2020-03-05; status last verified 2020-02.
• Results first posted 2020-02-21.
• 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

NCT03259490 is a completed, randomized, open-label, Phase 1 crossover study sponsored by Boehringer Ingelheim. It enrolled 30 healthy male and female participants and compared a single fed dose of a fixed-dose empagliflozin/linagliptin/metformin extended-release tablet with the three medicines administered as separate tablets.

The primary endpoint was empagliflozin AUC0-tz. The registry reports an ANOVA-adjusted test-to-reference geometric mean ratio of 103.06%, with a 95% confidence interval from 100.36% to 105.83%. Each treatment arm reports 29 participants in the pharmacokinetic results. The record also lists three secondary AUC outcomes for empagliflozin, metformin, and linagliptin. It does not report a p-value for the primary result or a bioequivalence acceptance interval in the information.