Bioequivalence Study of Individual Atazanavir and Cobicistat Compared With Atazanavir in Fixed-dose Combination With Cobicistat: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Bioequivalence Study of Individual Atazanavir and Cobicistat Compared With Atazanavir in Fixed-dose Combination With Cobicistat (NCT01837719), a randomized phase 1 study. The primary outcome is Maximum Observed Plasma Concentration (Cmax) of Atazanavir.

2. Clinical Question

The question is pharmacokinetic: it concerns the concentration of atazanavir in plasma after dosing, rather than a clinical outcome such as symptom improvement or disease progression. The study also examines whether the formulation comparison differs under fasted conditions and reports a comparison involving a high-fat meal. Those comparisons help describe the exposure patterns across the study’s treatment conditions, but they are not interchangeable with the primary light-meal comparison.

The registry describes a randomized, five-period crossover study in healthy subjects. In a crossover design, participants receive multiple treatments in different periods, allowing treatment conditions to be compared within the study population. The registry record does not report the treatment sequences or the precise details of the randomization process. It also does not report a clinical efficacy question. The results therefore address measured drug concentrations and selected safety outcomes in this study population, not the effects of treatment on HIV-1 illness.

3. Trial Design

The registry identifies this as a completed phase 1, randomized, open-label crossover study sponsored by Bristol-Myers Squibb. The official title describes a five-period study in healthy subjects. Enrollment was 64. The registry lists the study start as 2013-04 and primary completion as 2013-06. It does not report a participant flow account explaining how many participants completed each period or why any participant may have been excluded from a particular analysis.

The five treatments varied the formulation and meal condition. Treatment A was atazanavir 300 mg as a capsule plus cobicistat 150 mg as a tablet following a light meal. Treatment B was the fixed-dose combination at the same stated doses following a light meal. Treatment C used the separate agents in the fasted state, and Treatment D used the fixed-dose combination in the fasted state. Treatment E was the fixed-dose combination following a high-fat meal. The registry gives the scheduled dosing days as Day 1 or 8, Day 15 or 22, and Day 29 for these respective conditions.

Because participants received treatments across periods, the design can compare pharmacokinetic measurements under distinct formulation and meal conditions. The registry does not report the treatment sequence, washout duration, or other period-management details, so those features cannot be assessed from this record. Masking is listed as none. The primary endpoint is atazanavir Cmax; secondary outcomes include further pharmacokinetic measures and safety measures.

The numbers contributing to the reported Cmax summaries vary by treatment: 63 for Treatment A, 62 for Treatment B, 63 for Treatment C, 63 for Treatment D, and 30 for Treatment E. The registry describes the pharmacokinetic population as participants who received any study medication and had any available concentration-time data; only participants with adequate pharmacokinetic profiles were included in summary statistics and statistical analyses. The exact criteria for an adequate profile are not reported in the registry record.

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: A total of 149 patients were enrolled and 64 randomized to 1 of 8 treatment sequences (ABCDE, ABDCE, BACDE, BADCE, ABCD, ABDC, BACD, or BADC), with a 7-day washout period between treatments. Approximately 32 were to be discharged after Period 4, based on treatment sequence. Those remaining were to continue to and be discharged at end of Period 5.

Period 1: Days 1-7 (Treatment A or B)

Period 2: Days 8-14 (Treatment B or A)

Reasons for not completing this period:

Period 3: Days 15-21 (Treatment C or D)

Period 4: Days 22-28 (Treatment D or C)

Analysis population for the primary outcome, as defined in the registry: All participants who received any study medication and had any available concentration-time data. All available derived pharmacokinetic (PK) parameter values were included in the PK data set and reported, but only those with adequate PK profiles were included in the summary statistics and statistical analysis.

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: All participants who received at least 1 dose of study drug

6. Primary Endpoint

The registry lists the primary outcome as Maximum Observed Plasma Concentration (Cmax) of Atazanavir, measured over Days 1, 8, 15, 22, and 29 (predose and at 1, 2, 2.5, 3, 4, 5, 6, 8, 12, and 16 hours postdose); Days 2, 9, 16, 23, and 30 (24, 30, and 36 hours postdose); Days 3, 10, 17, 24, and 31 (48 hours postdose).

The primary endpoint is atazanavir maximum observed plasma concentration, abbreviated Cmax and measured in ng/mL. It is the highest measured plasma concentration identified from the concentration-versus-time profile after a dose. The registry describes blood sampling from predose through 48 hours after dosing on the specified study days, with multiple postdose sampling times. Cmax was derived from those concentration-time data.

Cmax is used in pharmacokinetic comparisons because it characterizes peak exposure after dosing. Comparing Cmax between formulations can indicate whether the observed peak concentration differs when atazanavir and cobicistat are posted as a fixed-dose combination rather than as separate agents. It is one aspect of drug exposure; it does not, by itself, describe the full concentration-time profile or establish a clinical effect.

The registry reports geometric means and geometric coefficients of variation for each treatment, then reports mixed-model geometric mean ratios for selected treatment comparisons. A geometric mean is appropriate for summarizing measurements that are analyzed on a multiplicative scale. The coefficient of variation describes relative spread around the geometric mean. The registry does not report the prespecified equivalence limits or the full statistical analysis plan for the primary endpoint, so the ratio and confidence interval should be read as the posted estimates rather than as a complete account of the decision criteria.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary result is reported as a geometric mean ratio with a 90% confidence interval. A ratio compares the scale of one treatment’s geometric mean with another’s. A ratio of 1 corresponds to equal geometric means on that ratio scale; a value above 1 places the first-named treatment’s estimate above the comparator, while a value below 1 places it below. For the light-meal comparison, the registry’s ordering is Treatment A followed by Treatment B, and the reported ratio is interpreted as the fixed-dose combination relative to separate agents. The confidence interval expresses uncertainty in the model-based ratio. It is not a range containing a stated percentage of individual participants, nor does it give the probability that a particular value is true.

The record identifies a mixed models analysis for the reported Cmax comparisons. Mixed models are commonly used for designs in which observations have a structured relationship, such as repeated measurements or treatment periods within participants. Such a model can account for features of the study design when those terms are specified and fitted. The registry record does not provide the model formula, its included effects, or details of how the covariance structure was handled. It is therefore possible to describe the method as reported, but not to reconstruct or independently assess its full specification from this record.

Geometric summaries and ratios are used because pharmacokinetic exposure measures are often analyzed on a multiplicative scale. In general, analyzing log-transformed measurements and then expressing contrasts as ratios can make interpretation more natural when relative differences are of interest. That is a general explanation; the registry does not provide every transformation or implementation detail. The geometric coefficient of variation is a measure of relative variability around the geometric mean. It should not be read as a confidence interval or as a direct measure of uncertainty in the treatment ratio.

The recalculated secondary comparisons use posted summary statistics to give simple, unadjusted mean differences for the half-life outcomes. For these comparisons, the reported difference is the first treatment’s mean minus the comparator’s mean as represented in the displayed summaries. The corresponding confidence intervals and p-values are labeled unadjusted. These calculations do not reproduce the sponsor’s mixed-model analysis, and they do not incorporate information that is not present in the summary statistics. In a crossover design, participant-level pairing and period structure may matter; the summary-statistic recalculations cannot recover those details. The sponsor’s own analysis takes precedence if a discrepancy is present, but the record identifies no discrepancy for these results.

A confidence interval for a mean difference gives a range of values compatible with the estimate under the calculation and its assumptions. If that interval includes zero, the data are compatible with no mean difference as well as with differences on either side of zero. A p-value measures how unusual the observed data, or more extreme data, would be under a specified null hypothesis and model. It is not the probability that the null hypothesis is true. The secondary p-values are unadjusted, so they do not account for making multiple comparisons or testing multiple outcomes.

Multiplicity matters because repeated testing increases the chance of obtaining at least one apparently notable result by chance, when each test is considered separately. Whether and how to adjust depends on the prespecified analysis strategy and the role of each endpoint. The registry record does not report a multiplicity plan, a hierarchy of hypotheses, or the decision rules used for the several Cmax comparisons. Accordingly, the reported confidence intervals and secondary p-values should be interpreted as the individual analyses recorded, not as evidence that a family-wise error rate was controlled.

Sample-size illustrations associated with the secondary recalculations are planning scenarios, not observed power. They use the observed effect or a stated fraction of it under the stated assumptions: two-sided significance level 0.05, equal allocation, and a normal approximation. Such illustrations show how estimated sample requirements change under those assumptions; they do not establish what power the completed study had, nor do they reveal the original design justification. The original sample-size justification is not reported in the registry record.

Missingness and analysis populations also affect interpretation. The registry says that available derived pharmacokinetic parameter values were included in the pharmacokinetic data set, while only adequate profiles contributed to summary statistics and statistical analyses. It does not give the number of missing concentration measurements, reasons for missingness, profile adequacy rules, or an analysis of how missing data could affect results. The analysis population therefore reflects the registry’s stated criteria, but the record does not allow a fuller assessment of selection or missing-data assumptions.

10. Analyses Reported by the Sponsor

The sponsor’s primary Cmax result is a mixed-model geometric mean ratio of 1.073 for the light-meal fixed-dose combination comparison, with a 90% confidence interval from 1.012 to 1.137. This is the analysis to use when describing the sponsor-reported primary comparison. The record gives no p-value for that result and does not state the equivalence bounds or full model specification.

The unadjusted recalculations are limited to the posted secondary half-life summaries. They calculate mean differences and related intervals from treatment-level summary statistics; they do not recreate the sponsor’s mixed-model method or its treatment of the crossover design. They therefore provide a separate, simplified view of those secondary outcomes, not an alternative primary Cmax result. Their unadjusted p-values should not be substituted for the sponsor’s primary analysis.

The registry reports no discrepancy for the posted results. Where a recalculation and a sponsor analysis differ, the sponsor’s own analysis takes precedence. In this record, no recalculated primary comparison is presented, so the primary interpretation rests on the sponsor-reported ratio and confidence interval. The original analysis plan and the reasons for selecting the reported model are not reported in the registry record.

11. Multiplicity, Missing Data and Analysis Populations

The registry reports a primary Cmax endpoint and multiple treatment comparisons, along with a range of secondary pharmacokinetic and safety outcomes. It does not report how the family of comparisons was defined, whether a hierarchy was prespecified, or whether any adjustment for multiplicity was applied. This matters because interpreting each comparison at its own confidence level or p-value does not account for the chance of at least one apparently unusual result among multiple analyses. The reported 90% confidence intervals should be read as the intervals attached to the individual mixed-model comparisons, not as evidence of a stated family-wide error control.

The record also does not report interim analyses or stopping rules. It does not state a sample-size justification, the targeted precision for Cmax, or the assumptions used to plan enrollment. The sample-size illustrations associated with the recalculated secondary outcomes are not the original trial design calculation and are not observed power. They are hypothetical planning estimates under stated assumptions, not evidence about the study’s actual power.

For the pharmacokinetic population, the registry says participants had to receive study medication and have available concentration-time data; only adequate pharmacokinetic profiles contributed to summary statistics and statistical analyses. It does not define adequacy, list participant-level missingness, state why profiles were excluded, or describe sensitivity analyses. The treatment-specific Cmax counts vary, and Treatment E has fewer observations than the other listed treatment groups. The record does not explain the reasons for these differences. Without that information, the impact of missing or excluded profiles cannot be assessed.

Safety risk counts also differ by treatment group, and the record does not explain the precise counting approach across crossover periods. It does not report details of how adverse events were attributed when participants received more than one treatment. These omissions limit interpretation of both missing data and safety comparisons.

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 registry reports zero serious adverse events in each listed treatment group. It records 1 participant with a non-serious adverse event in Treatment A and 3 in Treatment B, with zero in Treatments C, D, and E. The corresponding at-risk counts are 64 for Treatments A and B, 63 for Treatments C and D, and 31 for Treatment E. Death counts are not reported in the registry record.

These are group-level counts from a crossover study, not necessarily counts of distinct participants across the entire study. A participant could contribute to more than one treatment period, and the registry record does not explain how adverse events were assigned or deduplicated across periods. The summaries also do not provide event descriptions or timing for the non-serious events. The listed zero serious adverse event counts mean no such events were recorded in these group summaries; they do not establish that an event could not occur or provide a precise comparative safety estimate. This phase 1 pharmacokinetic study’s safety counts should be interpreted within that limited reporting context.

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

The posted primary result describes atazanavir peak plasma concentration for the fixed-dose combination relative to separate atazanavir and cobicistat after a light meal. The sponsor’s mixed-model estimate is a geometric mean ratio of 1.073, with a 90% confidence interval from 1.012 to 1.137. The estimate is above 1, and the interval is also above 1, indicating a higher estimated geometric mean peak concentration for the fixed-dose combination in that comparison. The interval conveys uncertainty about the magnitude of the relative difference.

This result does not, on its own, establish bioequivalence. Such a conclusion depends on prespecified acceptance limits and analysis rules, which are not reported in the registry record. Nor does the result establish a clinical benefit or harm: Cmax is a pharmacokinetic measure, not a patient-centered outcome. The study enrolled healthy subjects, and the record does not report clinical efficacy outcomes.

The additional reported comparisons involve different meal conditions and formulations. They should be interpreted according to their stated contrasts rather than combined into a single general claim about food or formulation. The registry does not report enough detail about the statistical plan to determine how those contrasts were prioritized or how multiplicity was handled.

The secondary half-life recalculations and adverse event counts provide limited context, not definitive evidence of no differences or of comparative safety. Unadjusted secondary analyses can be imprecise and do not address the full crossover model. Safety summaries with zero recorded serious adverse events are also limited by the study’s size, duration, and reporting detail. The record supports a description of the observed pharmacokinetic estimates and reported events, but not broader conclusions beyond those measurements and conditions.

14. Why This Trial Matters Statistically

This study is statistically notable because it uses a randomized crossover design to compare multiple formulations and meal conditions in a pharmacokinetic setting. Repeated treatment periods can make within-participant contrasts useful, since participants serve as their own reference across conditions. At the same time, valid interpretation depends on how the analysis handles period structure and within-participant correlation. The registry identifies a mixed models analysis for the Cmax comparisons but does not report the model specification, limiting independent assessment of those choices.

The primary estimate is reported on a ratio scale, which directly expresses relative peak exposure between formulations. Its 90% confidence interval gives readers information about precision, while the absence of reported equivalence bounds prevents a definitive assessment of whether the comparison meets a formal bioequivalence criterion. This distinction is important: a numerical ratio and confidence interval are evidence about an estimated contrast, but they do not reveal the decision rule on their own.

The record also illustrates why secondary outcomes and recalculations require care. Summary-statistic calculations can offer a transparent descriptive comparison, but they cannot reproduce participant-level crossover modeling. Multiple outcomes and contrasts create additional interpretive uncertainty when the multiplicity strategy is not reported. Finally, variation in the number of observations by treatment and limited reporting on missing profiles constrain assessment of precision and possible selection. These design and reporting features shape what can be concluded from the posted results.

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 record does not give treatment sequences, washout details, period effects, or the full mixed-model specification. Those details are especially relevant for a crossover study because treatment, period, and within-participant structure can influence estimated contrasts. Although the primary analysis method is named, the record does not provide enough information to reproduce it.

The registry does not report equivalence margins, a sample-size justification, the original precision target, or a detailed statistical analysis plan. It is therefore not possible to establish from the record whether the primary comparison met a predefined bioequivalence decision rule or how the set of comparisons was prioritized. It also does not report a multiplicity strategy, interim analyses, or stopping rules.

The number of participants represented in treatment summaries varies, particularly for the high-fat-meal condition. The registry does not explain these differences or define the criteria for an adequate pharmacokinetic profile. It does not report participant flow by period, missing concentration data, reasons for exclusion, or sensitivity analyses. Those gaps limit assessment of selection and missing-data effects.

The recalculated secondary results use posted summary statistics and are unadjusted; they do not reconstruct the crossover model. Safety results are reported as group counts without detailed event descriptions or an explanation of attribution across periods. Finally, the study population is described as healthy subjects, and clinical efficacy outcomes are not reported. The pharmacokinetic findings should therefore remain tied to the study’s measured endpoints and treatment conditions.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Bioequivalence Study of Individual Atazanavir and Cobicistat Compared With Atazanavir in Fixed-dose Combination With Cobicistat — NCT01837719.
• Registry record last updated 2014-08-29; status last verified 2014-08.
• Results first posted 2014-08-29.
• 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

NCT01837719 is a completed phase 1 randomized, open-label, five-period crossover study sponsored by Bristol-Myers Squibb. It enrolled 64 healthy subjects to compare atazanavir and cobicistat given separately with their fixed-dose combination under light-meal and fasted conditions, and to assess the fixed-dose combination after a high-fat meal.

The primary outcome was atazanavir Cmax. For the light-meal formulation comparison, the registry reports a mixed-model geometric mean ratio of 1.073, with a 90% confidence interval from 1.012 to 1.137. The record also reports other Cmax contrasts, secondary pharmacokinetic outcomes, and safety counts. It does not report the equivalence margins, full statistical plan, or detailed missing-data handling, which limits interpretation beyond the posted estimates.