Bioequivalence And Food Effect Study Comparing The Commercial Formulation Of Crizotinib To Its Clinical Study Formulations And Commercial Formulation With Or Without Food In Healthy Volunteers: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Bioequivalence And Food Effect Study Comparing The Commercial Formulation Of Crizotinib To Its Clinical Study Formulations And Commercial Formulation With Or Without Food In Healthy Volunteers (NCT01154218), a randomized phase 1 study. The primary outcome is Area Under the Curve From Time Zero to Extrapolated Infinite Time (AUC [0 - ∞]).

2. Clinical Question

3. Trial Design

This was a Phase 1, randomized, unmasked, crossover study in healthy volunteers. The registry reports enrollment of 36 and a completed status. Its stated purpose was to compare the Commercial Image Capsule (CIC) with the Immediate Release Tablet (IRT) and Powder in Capsule (PIC) formulations of crizotinib, and to compare CIC administration in fasted and fed states.

The record describes four treatments administered in one of four randomized orders: A-B-C-D, B-D-A-C, C-A-D-B, or D-C-B-A. The record does not identify which letter corresponds to each named treatment. It says participants were to receive all four treatments, making the planned design a within-participant comparison across formulations and food states. In a crossover design, each participant can contribute information under multiple conditions; comparisons can therefore account for variation between people. The registry does not report the period-specific allocation counts or the details of how order was assigned.

The primary outcome was AUC [0 - ∞], a pharmacokinetic measure calculated from plasma concentrations after dosing. Sampling was scheduled from pre-dose through 144 hours post-dose. The registry also lists secondary pharmacokinetic outcomes, including maximum concentration, time to maximum concentration, half-life, clearance, volume of distribution, and measures for a crizotinib metabolite.

The record gives no details about washout intervals, blinding procedures beyond masking being none, protocol deviations, or the statistical model used for the sponsor’s adjusted comparisons. Those aspects are not reported in the registry record. The crossover structure matters when interpreting group summaries: treatment-specific counts and measurements should not be treated as independent groups of people without accounting for the repeated treatment periods.

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.

First Intervention Period

Washout Period (at Least 14 Days)

Second Intervention Period

Washout Period (at Least 14 Days)

Analysis population for the primary outcome, as defined in the registry: Pharmacokinetic (PK) parameter analysis population included all randomized and treated participants who had at least 1 of the PK parameters of primary interest in at least 1 treatment period.

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 Area Under the Curve From Time Zero to Extrapolated Infinite Time (AUC [0 - ∞]), measured over 0 (pre-dose), 1, 2, 4, 5, 6, 8, 10, 12, 24, 36, 48, 72, 96 and 144 hours (hrs) post crizotinib dose.

The primary endpoint was area under the plasma concentration-time curve from time zero to extrapolated infinite time, AUC [0 - ∞], measured in ng*hr/mL. It represents overall plasma exposure after a dose, rather than the concentration at one particular sampling time. The registry defines it as the area from time zero to the last measured time plus the extrapolated area from that time to infinity. The listed sampling schedule began pre-dose and continued through 144 hours after crizotinib dosing.

Pharmacokinetic exposure is used in formulation comparisons because it characterizes how much drug is present in plasma over time. For a bioequivalence question, exposure is compared across formulations; for a food-effect question, exposure is compared between fed and fasted administration. These are statistical comparisons of measured pharmacokinetic quantities, not direct measures of clinical benefit.

The primary arm summaries are reported as geometric means, with standard deviations. The registry also reports adjusted-mean ratios and 90% confidence intervals for the three primary comparisons. A confidence interval describes the uncertainty around an estimated comparison under the analysis used. The registry does not report the prespecified equivalence margins, the adjustment model, or the sponsor’s detailed analysis plan. Without those details, the intervals can be described, but the registry record alone does not establish the decision criteria used for a formal bioequivalence conclusion.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does a crossover design contribute?

In a crossover study, participants receive more than one treatment condition over separate periods. That structure allows each participant’s measurements under different conditions to contribute to the comparison. Because measurements from the same person are related, a statistical analysis should account for that repeated-measures structure rather than treating all treatment-period observations as independent. The registry identifies the model as crossover and reports randomized treatment orders. It does not describe the fitted model, the treatment of period or sequence effects, or any washout procedures, so those details cannot be evaluated from the record.

The four listed orders distribute the sequence of treatments among participants. Randomizing order is intended to reduce systematic differences in which condition comes first or later. In general, crossover interpretation also depends on whether effects from an earlier period could carry into a later period. The record does not report how the study addressed that possibility. This is a design consideration, not evidence that carryover occurred.

What do the geometric means represent?

The primary arm summaries are geometric means. A geometric mean is commonly used for pharmacokinetic measurements when values are positive and may be asymmetric, because it summarizes values on a multiplicative scale. The reported standard deviations describe dispersion as recorded by the registry, but the record does not state whether those deviations were calculated on the original or transformed scale. The arm summaries are descriptive; they do not supply the full within-participant covariance information needed to reconstruct a crossover comparison.

How should the adjusted ratios and confidence intervals be read?

The primary analyses report ratios of adjusted means, together with 90% confidence intervals. A ratio compares estimated exposure between two treatment conditions on a relative scale. The point estimate is the central estimate from the reported analysis; the confidence interval communicates uncertainty around it. A narrower interval generally indicates greater precision, while a wider interval indicates more uncertainty. These are general explanations of the statistics, not additional results from the trial.

For bioequivalence, an estimate is generally evaluated against prespecified limits on the ratio and its confidence interval. The registry does not state the limits, the transformation used in analysis, or the model producing the adjusted means. Consequently, the reported intervals can be read as the sponsor’s estimates and uncertainty, but the record does not contain enough detail to independently verify the formal equivalence decision. The absence of a posted p-value does not make the confidence interval unusable; it means the record presents the comparison in interval form rather than as a reported p-value test.

Why can’t the arm summaries reproduce the crossover analysis?

A simple comparison of treatment-specific geometric means does not account for the fact that the same participant may contribute measurements in multiple periods. A crossover estimate can use paired information, and its precision depends on how measurements within the same person vary together. The record supplies neither those participant-level data nor the covariance structure. It also does not report the model used for adjusted means. As a result, the posted group summaries are useful for describing each condition, while the reported adjusted ratios and intervals are the relevant registered comparisons.

For the secondary endpoint of plasma decay half-life, the posted recalculations are explicitly unadjusted comparisons of summary statistics. They include mean differences, confidence intervals, and unadjusted p-values. Such recalculations do not replace a sponsor analysis that accounts for the trial design. No sponsor-reported analyses are listed for this secondary endpoint in the registry data. The recalculated values should therefore be treated as descriptive statistical checks, not as the study’s definitive inferential analysis.

What is over-dispersion, and is it relevant to these outcomes?

Over-dispersion is a general concern in count-data models when variability exceeds what a simple model assumes. These primary and secondary outcomes are continuous pharmacokinetic measures, not event counts, so over-dispersion is not the central issue for the reported AUC comparison. More relevant considerations include the distribution of pharmacokinetic measurements, within-participant correlation, treatment-period effects, and missing measurements. The registry does not report enough model detail to determine how those considerations were handled.

What can the planning illustrations say?

The posted sample-size illustrations for the half-life recalculations are based on the observed effect and state assumptions of a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. They show participants per arm for 80% or 90% power at different fractions of the observed effect. These are planning illustrations, not observed power and not evidence about the study’s original design. The registry does not report the trial’s own sample-size justification, nor does it establish that the original study was planned using these assumptions. Because the comparison is from a crossover trial while the illustration assumes equal allocation between two arms, the figures should not be treated as a reconstruction of the study design.

More broadly, statistical precision depends on the amount of information and on variability, not only on the number of measurements. In a crossover setting, the relationship between a participant’s measurements across periods also matters. The registry’s summaries do not supply enough information to rebuild that dependence or produce a design-matched power calculation. The planning figures therefore answer a limited hypothetical question about an unadjusted observed effect under stated assumptions.

10. Analyses Reported by the Sponsor

The sponsor-reported primary analyses are adjusted-mean ratios with 90% confidence intervals for the formulation and food-state comparisons. Those comparisons should be distinguished from the descriptive geometric means reported for each treatment. The adjusted estimates are the registry’s formal primary comparisons, while the arm summaries alone do not contain the within-participant information needed to recreate the crossover analysis.

No independently recalculated primary comparison is posted for AUC [0 - ∞]. The half-life comparisons do have simple unadjusted recalculations, but they concern a secondary endpoint and do not replicate the sponsor’s primary analysis. The registry does not report a sponsor analysis for those secondary comparisons. Accordingly, the half-life recalculations are a separate descriptive check and should not be treated as corroboration or contradiction of the adjusted primary results.

The record does not flag a discrepancy between the reported and recalculated analyses. It also does not provide the model specification, equivalence margins, or full statistical analysis plan needed to assess the sponsor analysis independently. The appropriate reading is therefore to report the sponsor’s adjusted primary estimates as posted, explain their uncertainty, and avoid drawing a formal equivalence conclusion from the registry record without the decision criteria.

11. Multiplicity, Missing Data and Analysis Populations

The record lists 11 outcome measures and 18 analyses, but does not report a multiplicity plan. Multiplicity concerns arise when several outcomes or comparisons are tested: interpreting each result as though it were the only test can overstate the strength of evidence. A prespecified hierarchy or adjustment can help manage that issue. The registry record does not state whether the primary comparisons were treated as a family, whether secondary outcomes were exploratory, or whether any adjustment was used.

The primary analysis population is defined as randomized and treated participants with at least one primary-interest pharmacokinetic parameter in at least one treatment period. Treatment-specific primary counts are 35 for IRT fasted, 35 for PIC fasted, 35 for CIC fasted, and 36 for CIC fed. The record does not explain the variation in those counts or identify the reasons for unavailable treatment-period measurements. It does not report missing-data handling, imputation, or sensitivity analyses.

In a crossover study, missingness in one period can affect the available paired comparisons even when a participant contributes information in another period. The record’s population definition permits inclusion when a participant has a primary parameter in at least one period, but it does not describe how incomplete treatment sequences contributed to each comparison. It is therefore not possible to assess from the registry record whether missingness was balanced across treatments or whether conclusions were robust to alternative handling.

The registry does not report interim analyses, a formal statistical analysis plan, protocol deviations, or the sample-size justification. It does not identify a prespecified equivalence margin or the decision rule for bioequivalence. These omissions are important because the primary confidence intervals are most informative when interpreted against the criteria set before the data were examined. Without that context, the intervals can be described precisely, but the registry record alone cannot establish the sponsor’s complete decision process.

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 0% reporting threshold.

Most frequent non-serious adverse event terms

The adverse-event record reports no serious adverse events attributed to treatment in the listed groups: the serious-event affected count is 0 for IRT fasted, PIC fasted, CIC fasted, and CIC fed. The registry lists 36 participants at risk for serious events in each treatment group. The top serious-event terms list is empty. These figures describe what is posted; they do not establish that no serious event occurred outside the recorded attribution or reporting definitions.

Other adverse events attributed to treatment are reported for 21 participants in the IRT fasted group, 24 in PIC fasted, 20 in CIC fasted, and 21 in CIC fed. Each group lists 36 at risk. The record does not provide event-level details in the registry record for these summaries, and the groups are treatment periods in a crossover study rather than separate parallel cohorts. Counts across treatment groups should not be added or interpreted as distinct participants, because a person could receive more than one treatment.

Death counts are not reported in the registry record. The adverse-event summaries describe observed reporting during this study; they do not establish comparative safety beyond the recorded follow-up and definitions.

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

The results describe pharmacokinetic exposure after crizotinib under different formulation and food-state conditions in healthy volunteers. The primary outcome is a measure of plasma exposure over time, and the registry reports adjusted ratios and 90% confidence intervals for the three comparisons. Those estimates communicate both the direction and the uncertainty of the comparisons.

They do not, by themselves, establish that two formulations are bioequivalent or that a meal has no meaningful effect. Such conclusions depend on prespecified criteria, including equivalence limits and the statistical model. Those details are not reported in the registry record. Nor do pharmacokinetic comparisons directly establish clinical effectiveness, long-term safety, or outcomes in people with a health condition. The population was healthy volunteers, and the study was a Phase 1 single-dose study.

The arm-level geometric means give a descriptive view of exposure by treatment. The adjusted ratios are the formal comparisons reported by the registry. Because the design is crossover, the same participants could contribute under multiple treatment conditions; the group-level summaries do not reveal the paired structure or reproduce the adjusted analysis. Interpretation should not treat the treatment groups as independent parallel arms.

The safety counts show no serious adverse events attributed to treatment in the listed groups, alongside recorded non-serious adverse events. These observations are limited to the study and reporting information in the registry. The counts do not establish the absence of rare or later events, and they should not be turned into treatment advice.

Overall, the posted results support a statistical description of the measured pharmacokinetic comparisons and their reported precision. They leave important questions about formal decision criteria, missing data, multiplicity, and model details unresolved in the registry record.

14. Why This Trial Matters Statistically

This trial is statistically informative because it uses a crossover design for comparisons of formulation and food state. When participants receive multiple conditions, their own measurements can help distinguish treatment-related variation from differences between people. That can be a useful design for pharmacokinetic questions, where measurements are taken repeatedly after dosing. It also makes the analysis more dependent on the within-participant structure than a simple comparison of independent group averages.

The registry illustrates why an endpoint summary and an inferential comparison are not interchangeable. Geometric means describe exposure under each condition, while adjusted ratios and confidence intervals describe the sponsor’s comparisons. The latter are the relevant reported estimates for the primary questions, but their full interpretation depends on the model and prespecified limits. Without those, readers can assess the reported values and precision but cannot fully reproduce the decision framework.

The record also highlights the distinction between bioequivalence questions and ordinary tests for a difference. A bioequivalence assessment generally asks whether uncertainty is sufficiently constrained within predefined limits, rather than whether an observed difference is statistically distinguishable from zero. The registry reports intervals but not the limits needed to evaluate that question. This is a key reason not to infer a formal conclusion from the point estimates alone.

Finally, the secondary summaries and safety counts demonstrate the importance of accounting for repeated treatment periods, multiple outcomes, and incomplete information. Clear reporting of analysis models, missing-data procedures, multiplicity handling, and decision criteria would make the statistical evidence more fully assessable. These are reporting limitations, not evidence that the analyses were necessarily inappropriate.

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.

Several trial-specific details needed for independent assessment are not reported in the registry record. The exact mapping between the randomized sequence letters and the four named treatments is not stated. The washout interval, period effects, sequence effects, and handling of possible carryover are also not described. These details matter to understanding a multi-period crossover analysis.

The sponsor’s adjusted-mean model is not specified, and the record does not provide the equivalence margins or formal decision rule. The primary results therefore cannot be independently interpreted against the study’s intended acceptance criteria. The registry also does not report a detailed statistical analysis plan, interim analyses, or a sample-size justification.

The pharmacokinetic analysis population is defined, and treatment-specific counts are posted, but the reasons for differing counts are not given. The record does not describe missing measurements, protocol deviations, or sensitivity analyses. Because a crossover comparison depends on measurements within the same participant, the impact of incomplete periods cannot be assessed from the summaries alone.

The record lists numerous outcomes and analyses but does not state whether secondary analyses were adjusted for multiplicity or considered exploratory. For half-life, the posted recalculations are unadjusted and do not reproduce a crossover model. The additional endpoint results are not all represented in the available posted summaries, limiting assessment of consistency across the full pharmacokinetic profile.

Finally, the study involved healthy volunteers and a single-dose pharmacokinetic question. The findings described in the registry should not be extended to other populations, repeated dosing, clinical outcomes, or long-term safety without evidence beyond this record.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Bioequivalence And Food Effect Study Comparing The Commercial Formulation Of Crizotinib To Its Clinical Study Formulations And Commercial Formulation With Or Without Food In Healthy Volunteers — NCT01154218.
• Registry record last updated 2011-10-24; status last verified 2011-10.
• Results first posted 2011-10-19.
• 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

NCT01154218 is a completed Phase 1 randomized, unmasked crossover study sponsored by Pfizer. It enrolled 36 healthy volunteers and compared crizotinib in the CIC, IRT, and PIC formulations, including CIC in fasted and fed states. The primary endpoint was AUC [0 - ∞], with treatment-specific geometric means and sponsor-reported adjusted-mean ratios with 90% confidence intervals for three comparisons.

The registry also lists secondary pharmacokinetic outcomes and adverse-event summaries. It reports no serious adverse events attributed to treatment in the listed groups, while other treatment-attributed adverse events were recorded. Important details—including the adjusted analysis model, equivalence margins, multiplicity strategy, and missing-data handling—are not reported in the registry record, limiting independent evaluation of formal conclusions.