Bioequivalence Study of Saxagliptin and Glucophage Combination Formulations in Healthy Subjects (B): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Bioequivalence Study of Saxagliptin and Glucophage Combination Formulations in Healthy Subjects (B) (NCT00897390), a randomized phase 1 study. The primary outcome is Saxagliptin Pharmacokinetic (PK) Parameter Area Under the Plasma Concentration-Time Curve From Time Zero Extrapolated to Infinite Time (AUC[INF]).

2. Clinical Question

3. Trial Design

The registry identifies this as a phase 1, randomized, unmasked crossover study, completed after starting in 2009-06, with primary completion in 2009-07. AstraZeneca is listed as sponsor. The enrollment was 24. The brief description states that the study aimed to demonstrate bioequivalence for a fixed-dose tablet containing saxagliptin 2.5 mg and metformin immediate release 1000 mg compared with the same drugs given as separate tablets, in fasted and fed states.

The four treatment conditions were the separate tablets while fasted, the fixed-dose tablet while fasted, the separate tablets with a standard meal, and the fixed-dose tablet with a standard meal. A crossover design means participants can receive more than one treatment condition, allowing within-participant comparisons and reducing the influence of stable differences between participants. The registry record does not report the treatment sequence, period allocation, washout interval, or whether every participant completed every condition. Those details are important for assessing potential period or carryover effects and cannot be inferred from the crossover label alone.

The primary pharmacokinetic outcome was saxagliptin AUC(INF), with measurements taken before dosing and at scheduled times through 48 hours in each period. The primary analysis population is described as treated participants with pharmacokinetic measures. Posted arm counts are 19 for each fasted condition, 21 for the separate-tablet fed condition, and 22 for the fixed-dose fed condition. These are analysis counts for the displayed conditions, not a basis for inferring the number of unique participants or the pattern of crossover completion.

The registry lists 20 outcome measures and 12 analyses. It also lists secondary pharmacokinetic outcomes, safety outcomes, and laboratory-abnormality outcomes. The record does not report the planned sample-size justification, the intended model for the primary analysis, or whether interim analyses were planned.

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: 58 participants were enrolled in the study; 34 participants were not dosed (23 no longer met study criteria, 4 withdrew consent, and 7 for other reasons).

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Treated participants with PK measures

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 Saxagliptin Pharmacokinetic (PK) Parameter Area Under the Plasma Concentration-Time Curve From Time Zero Extrapolated to Infinite Time (AUC[INF]), measured over pre-dose, post-dose at 15, 30, 45, 90 minutes, hours 1, 2, 3, 4, 6, 8, 12, 18, 24, 36 and 48 of each period.

The primary endpoint is saxagliptin AUC(INF), or the area under the plasma concentration-time curve from dosing extrapolated to infinite time. It is a pharmacokinetic measure of overall drug exposure rather than a measure of symptoms, disease control, or clinical events. The record says that single-dose parameters were derived from plasma concentration-versus-time data and identifies the endpoint unit as ng*h/mL.

AUC(INF) uses observed concentrations over the sampling period and an extrapolation beyond the final measured time. That extrapolated portion is an estimate, so the quality and timing of concentration measurements can matter to its precision. The posted measurement schedule extends from pre-dose through 48 hours in each period, with multiple early and later post-dose collections. The registry does not report the proportion of AUC(INF) that was extrapolated or the individual concentration profiles.

In a bioequivalence comparison, exposure after the combination formulation is assessed relative to exposure after the same medicines are taken separately. The registry reports geometric means and adjusted geometric-mean ratios, rather than an arithmetic difference in AUC. Ratios are natural for positive, often right-skewed exposure measures: a ratio near 1 indicates similar estimated exposure on the ratio scale. Whether a ratio and its confidence interval satisfy a formal bioequivalence criterion depends on prespecified margins and analysis rules, which are not reported in the registry record.

The primary endpoint concerns saxagliptin only. It does not by itself establish the corresponding exposure pattern for metformin or describe clinical effectiveness or longer-term safety.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

Bioequivalence analyses commonly compare exposure measures on a ratio scale. When an endpoint such as AUC is positive and may have a skewed distribution, analysis often uses logarithms and then expresses the contrast as a ratio of geometric means. A ratio of 1 represents equal estimated exposure on that scale; values on either side indicate relative differences in the estimated geometric means. This is a general explanation of the method, not confirmation of the exact model used in this trial. The registry labels its reported quantities as adjusted geometric means but does not report the adjustment method or model specification.

The registry gives 90% confidence intervals for both primary ratios. A confidence interval conveys statistical uncertainty around an estimated contrast, conditional on the data and analysis assumptions. In an equivalence setting, the key question is not simply whether a conventional test finds a difference. Rather, the interval is assessed against prespecified lower and upper equivalence limits. If the interval lies wholly inside those limits, that can support equivalence under the stated design and analysis. The margins, their justification, and the full decision rule are not reported in the registry record. The intervals should therefore be read as estimates and uncertainty ranges, not as a self-contained proof of a formal conclusion.

The crossover structure is relevant to the analysis. In general, crossover studies can compare treatments within participants, which can account for stable participant-level variation. Such studies also require attention to period effects, treatment sequence, and possible carryover. The registry confirms randomization and a crossover model, but the sequence, period structure, washout, and treatment of period or carryover effects are not reported in the registry record. It also does not report the exact statistical model or the assumptions used to obtain the adjusted ratios and confidence intervals.

The arm-level primary results use geometric means and full ranges, while the reported comparisons use adjusted geometric-mean ratios. Those quantities answer related but different questions: arm summaries describe the observed distribution, and the adjusted ratio describes the analysis contrast. A full range is not a confidence interval and does not quantify uncertainty in the estimated treatment contrast. The registry does not provide standard deviations for these primary summaries or participant-level data, so the adjusted comparison cannot be reconstructed from the displayed summaries alone.

A separate recalculation is available for the posted secondary T-Half summaries. It uses unadjusted mean differences from the arm-level means and standard deviations and should not be confused with the sponsor’s primary pharmacokinetic analysis. For these recalculations, the fasted fixed-dose comparison has a mean difference of -0.74 hours, with a confidence interval from -1.629 to 0.149 and an unadjusted p-value of 0.103. The comparison of the fasted separate-tablet arm with the fed separate-tablet arm has a mean difference of -0.57 hours, with a confidence interval from -1.479 to 0.339 and an unadjusted p-value of 0.219. These are descriptive, unadjusted analyses of the posted summary statistics, not reported sponsor analyses and not the primary endpoint analysis.

The planning illustrations based on the observed T-Half effects are not observed power and do not describe the original trial design. They use a two-sided significance level of 0.05, equal allocation, and a normal approximation, and consider different fractions of the observed effect. For the fasted fixed-dose comparison, planning for 100% of the observed effect gives 57 participants per arm for 80% power and 76 participants per arm for 90% power. Planning for 75% of the observed effect gives 100 and 134 participants per arm, respectively; planning for 50% gives 225 and 301. For the comparison between the fasted and fed separate-tablet conditions, planning for 100% gives 105 participants per arm for 80% power and 140 for 90%; planning for 75% gives 186 and 248; planning for 50% gives 417 and 558. These are hypothetical planning values, not evidence that the observed study had a particular power or that these were its design targets.

Over-dispersion is a general concern when variability exceeds what a simple model assumes: uncertainty can be understated if the analysis does not accommodate that variation. For continuous pharmacokinetic outcomes, the corresponding questions include the distributional assumptions, transformations, and handling of influential values. The registry does not report diagnostic checks or detailed model assumptions, so these cannot be assessed from the posted record.

10. Analyses Reported by the Sponsor

The sponsor’s reported primary analyses are the adjusted geometric-mean ratios with 90% confidence intervals. They compare the fixed-dose tablet with the co-administered tablets separately in fasted and fed conditions. The fasted ratio is 1.042, with a 90% confidence interval from 1.009 to 1.075; the fed ratio is 0.988, with a 90% confidence interval from 0.958 to 1.019. The registry does not report a p-value for either primary comparison or identify the analysis method used to produce these adjusted estimates.

The simple recalculations described for T-Half are not recalculations of the primary AUC(INF) endpoint. They use posted secondary summary statistics to form unadjusted mean differences, whereas the primary outcome is reported as an adjusted geometric-mean ratio. The scales, endpoints, and analyses differ, so the secondary recalculations cannot confirm or contradict the sponsor’s primary result. No discrepancy is identified between the posted sponsor analysis and the available recalculations, but that does not make the recalculations an independent reproduction of the primary analysis.

For interpretation of the primary endpoint, the sponsor’s own analysis takes precedence over unrelated unadjusted recalculations. Even so, the registry record does not report the equivalence margins, detailed model, or full decision rule. The posted estimates can be read as the sponsor-reported comparisons, while a complete independent assessment of the intended bioequivalence criterion remains limited by those omissions.

11. Multiplicity, Missing Data and Analysis Populations

The registry lists multiple outcomes and analyses, including pharmacokinetic endpoints, adverse events, and laboratory outcomes. It does not report whether a hierarchy of hypotheses was prespecified, how multiple comparisons were handled, or whether any secondary comparisons were intended to support formal claims. This means that reported secondary p-values, where present, should be treated cautiously unless their role in a prespecified testing strategy is known. Unadjusted testing across multiple outcomes can increase the chance of findings that appear noteworthy by chance alone.

Missing-data methods are also not reported in the registry record. The primary population is described as treated participants with pharmacokinetic measures, and the displayed group counts differ by condition. The record does not explain why all arm counts are not the same, how incomplete profiles were handled, or how missing concentration measurements affected AUC(INF) estimation. Nor does it report whether participants were excluded for protocol deviations or whether sensitivity analyses were conducted.

The record identifies a crossover design but does not describe the sequence assignment, period handling, carryover assessment, or analysis model. These details affect how the dependence between repeated measurements from the same participant is handled. The record also does not report the planned sample-size justification, equivalence margins, interim-analysis plan, or whether the presented analysis populations were fully prespecified. None of these omissions establishes that the conduct or analysis was inappropriate; they limit what can be verified from the registry record.

The primary ratios and intervals remain informative as posted estimates, but readers cannot fully assess the robustness of the formal bioequivalence decision without the missing analysis details. The secondary outcomes require additional caution because their multiplicity strategy and missing-data handling are not reported.

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 posted safety groups report 0 serious adverse events in each listed treatment condition and in the not-dosed group. For other adverse events, the reported numbers of affected participants are 3 for the fasted separate-tablet condition, 2 for the fasted fixed-dose condition, 2 for the fed separate-tablet condition, and 3 for the fed fixed-dose condition. The not-dosed group reports 0 affected participants for other adverse events. These are the registry’s group-level counts and should be read in the context of the listed risk populations for each group.

Zero reported serious adverse events in these groups means none were recorded in the posted safety results; it does not establish that serious events are impossible or rule out uncommon harms. This was a short, single-dose study in healthy subjects, and the registry does not provide detailed event descriptions in the serious-event terms. Death counts are not reported in the registry record. The safety results are descriptive and do not support broad conclusions about longer-term use, repeated dosing, or safety in people with diabetes.

The safety group counts should not be added together to infer a number of unique participants, particularly in a crossover design where participants may contribute under more than one condition. The listed risk counts also differ from some pharmacokinetic analysis counts, reflecting distinct posted outcome populations whose relationship is not fully explained in the registry record.

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

The primary result addresses a narrow pharmacokinetic question: the relative saxagliptin AUC(INF) after a fixed-dose combination compared with co-administered tablets, under fasted and fed conditions. The registry’s adjusted ratios and 90% confidence intervals show the estimates and their statistical uncertainty for each comparison. The two food conditions are separate comparisons, not a single pooled estimate.

These results do not by themselves establish equivalence under a formal regulatory criterion because the equivalence margins and full decision rule are not reported in the registry record. An interval near 1 may be consistent with similar estimated exposure, but formal equivalence depends on whether the whole interval falls within prespecified limits and on the adequacy of the underlying design and analysis. The record does not provide enough detail to verify that decision independently.

The endpoint is saxagliptin exposure, not metformin exposure, glucose control, symptoms, complications, or other clinical outcomes. The study involved healthy subjects and a single dose, so it does not directly describe treatment outcomes in people with Type 2 Diabetes Mellitus or effects with continued use. The registry’s condition listing does not change the described study population.

Nor do the posted safety counts establish the absence of risks. They describe events recorded in a limited study context; the study was not a comprehensive evaluation of uncommon or long-term adverse effects. The appropriate interpretation is therefore bounded: the posted data characterize single-dose pharmacokinetic comparisons and recorded short-term events under the study conditions, while broader clinical and safety questions remain outside what these results can answer.

14. Why This Trial Matters Statistically

This study is statistically interesting because a crossover bioequivalence question differs from a conventional parallel-group efficacy question. The treatment contrast is a ratio of exposure measures, and participant-level pairing can be informative when stable differences between people are substantial. At the same time, a crossover analysis depends on the sequence and period structure, and interpretation can be affected by incomplete periods or carryover. The registry identifies the design but does not report enough operational detail to evaluate those issues fully.

The primary results also illustrate why a ratio and its confidence interval should be read together. A point estimate alone does not express precision; the interval shows the range of uncertainty under the analysis. For an equivalence objective, simply observing a ratio near 1 or failing to show a conventional difference is not the complete decision rule. The margins and model matter, and those are not reported in the registry record.

The distinction between sponsor analyses and simple recalculations is important. Unadjusted secondary calculations from summary statistics can be educational, but they are not substitutes for the sponsor’s adjusted primary analysis. Nor do planning illustrations based on observed effects reveal the study’s actual design rationale or observed power. They demonstrate how a chosen effect assumption and target power can affect hypothetical sample-size planning, not what the original study was designed to detect.

Finally, the record shows how the interpretation of a small pharmacokinetic study depends on more than its headline estimate. Endpoint definition, food condition, population, model, missing data, and multiplicity all shape what can reasonably be concluded. Distinguishing these elements helps keep a narrowly framed exposure comparison from being mistaken for evidence about clinical benefit or broad safety.

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 leaves several trial-specific uncertainties. Although the study is labeled randomized and crossover, it does not report the treatment sequences, period structure, washout interval, or handling of period and carryover effects. These details would help assess whether the treatment comparisons adequately account for the way participants experienced the conditions. The statistical model used for the adjusted geometric-mean ratios is also not reported.

The record does not give the equivalence margins or explain how they were chosen. It therefore permits interpretation of the point estimates and confidence intervals but not a full check of whether the intended bioequivalence criterion was met. The primary summaries report full ranges rather than standard deviations, and participant-level data are not available in the registry record to reconstruct the adjusted analysis or assess influential observations.

Counts differ across displayed pharmacokinetic and safety groups, and the record does not fully explain the analysis-population differences or missing measurements. The proportion of AUC(INF) based on extrapolation is not reported. Detailed adverse-event descriptions, death counts, and a complete account of laboratory findings are also not available in the posted information summarized by these results.

Finally, the study evaluated single-dose pharmacokinetics in healthy subjects. Its scope is not equivalent to a study of clinical outcomes or longer-term safety in people with Type 2 Diabetes Mellitus. The record also does not report the sample-size justification, multiplicity strategy, or planned interim analyses. These limitations constrain independent assessment; they should not be taken as evidence that any particular unreported procedure was or was not conducted.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Bioequivalence Study of Saxagliptin and Glucophage Combination Formulations in Healthy Subjects (B) — NCT00897390.
• Registry record last updated 2015-05-21; status last verified 2015-05.
• Results first posted 2011-01-21.
• 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

The completed phase 1 trial was a randomized, unmasked crossover study sponsored by AstraZeneca. It enrolled 24 healthy subjects and compared a single dose of a saxagliptin/metformin immediate-release fixed-dose combination with co-administered separate tablets under fasted and fed conditions.

The primary endpoint was saxagliptin AUC(INF), analyzed in treated participants with pharmacokinetic measures. The registry reports adjusted geometric-mean ratios and 90% confidence intervals for the fasted and fed comparisons. The equivalence margins, detailed analysis model, crossover sequence, and missing-data procedures are not reported in the registry record, limiting independent assessment of a formal bioequivalence decision.

Secondary outcomes include additional pharmacokinetic, safety, and laboratory measures. Posted safety results report no serious adverse events in the listed groups, with other adverse events recorded in some groups. These single-dose findings describe the study conditions and do not establish clinical effectiveness or longer-term safety.