A Crossover Study to Assess the Effects of Vorinostat (MK0683, SAHA) in Patients With Advanced Cancer (0683-070)(COMPLETED): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Crossover Study to Assess the Effects of Vorinostat (MK0683, SAHA) in Patients With Advanced Cancer (0683-070)(COMPLETED) (NCT00632931), a randomized phase 1 study. The primary outcome is Change From Baseline in QTcF at 0.5 Hours.

2. Clinical Question

3. Trial Design

This was a randomized, phase 1, partially blinded, placebo-controlled, two-period crossover study. It enrolled 25 participants with relapsed or refractory advanced cancer and is recorded as completed. The trial’s primary purpose was to assess the effect of a single dose of vorinostat on QTc interval, rather than to compare long-term clinical outcomes. The planned sequence structure assigned participants to Drug/Placebo or Placebo/Drug, allowing each sequence to include both treatments in different periods.

In a crossover design, treatment comparisons can use measurements collected from the same participants under different treatment conditions. This can reduce the influence of stable differences between participants, because each participant may contribute observations under both treatments. That general advantage depends on design and analysis assumptions, including how period effects and possible carryover are handled. The registry record does not report a washout duration, a carryover assessment, or the statistical method used to account for period or sequence effects. Those details should not be presumed from the crossover label alone.

The official title calls the study randomized, partially blind, placebo controlled, and two-period crossover. The design information separately records single masking. The identity of the masked parties and the practical masking procedures are not reported in the registry record. The comparator is described as unspecified placebo, so further details about its formulation or matching are also not reported.

The registry lists a start in 2007-07 and primary completion in 2008-05. The endpoint was QTcF change at 0.5 hours, with baseline and 0.5 hours named as its timeframe. The endpoint description says QTcF was determined at each time point from five replicate measurements. The reported primary endpoint is a continuous measure in milliseconds, summarized as a mean and confidence interval.

The primary endpoint population is described as all patients as treated in Part 1. Its accompanying text gives 22 patients with QTcF data from the vorinostat period and 23 from the placebo period, and mentions exclusions for a protocol violation and missing predose measurements. The arm summaries list 23 for vorinostat and 22 for placebo. Because those counts differ across registry descriptions, the precise relationship between the endpoint population counts and the arm-summary counts is not resolved by the record. The allocation sequence generation, sample-size justification, and detailed analysis plan 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: Enrolled patients were assigned to 1 of 2 dose sequences (vorinostat then placebo or placebo then vorinostat) in Part 1 of the study to determine the effect of vorinostat on the QTcF (Fridericia-corrected QT) interval. Following Part 1, active patients continued into Part 2, daily dosing of vorinostat.

Part 1, Period 1

Reasons for not completing this period:

Part 1, Washout

Part 1, Period 2

Post-Part 1 Washout

Analysis population for the primary outcome, as defined in the registry: All patients as treated population in Part 1 of the Study; 22 patients had QTcF data from the vorinostat period (One patient with protocol violation, and two patients without predose measurements were excluded) and 23 patients had QTcF data from the placebo period (one patient with protocol violation and one patient who discontinued were excluded)

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 Change From Baseline in QTcF at 0.5 Hours, measured over Baseline and 0.5 hours.

The primary endpoint is change from baseline in QTcF at 0.5 hours, expressed in milliseconds. The QT interval is a measurement on an electrocardiogram; QTcF denotes correction of the interval using Fridericia’s method. Correcting the interval is intended to account for the influence of heart rate on the measured QT interval, which helps make values collected under differing heart rates more comparable. This is general context about the measure, not a finding from this trial.

The registry says QTcF was determined from five replicate measurements at each time point. Replicate measurements can help characterize the value at a given time point, although the record does not explain how they were combined. The endpoint compares the value at 0.5 hours with the predose baseline value. Importantly, the registry describes the calculation as subtracting the QTcF at each time point from the baseline value. Thus, the stated direction is baseline minus the time-point value, and positive or negative changes should be interpreted according to that definition rather than an assumed convention.

The endpoint is used to assess an electrocardiographic interval under the study conditions. The posted arm summaries give mean changes and 95% confidence intervals. The sponsor’s comparison is described as a mean difference in final values, not explicitly as a contrast of the arm-level change scores. The record does not explain the precise relationship between those estimands or give the calculation details. The 0.5-hour measurement is the specified primary time point; other endpoint details and the rationale for selecting this time point are not reported in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does a mean describe?

A mean is a measure of the average value in the analyzed observations. Here, the reported outcome is a continuous measure of QTcF change in milliseconds. An arm mean condenses the observed changes into one summary, but it does not describe the full distribution of individual measurements. Participants may have values dispersed around the mean, and the registry does not provide a standard deviation or another measure of that participant-level spread for this endpoint.

The registry reports a mean for each treatment condition and a confidence interval around each mean. A confidence interval expresses statistical uncertainty in an estimate under the assumptions of the method used. It should not be interpreted as a range that contains a specified share of individual participant values. The arm confidence intervals are 95% intervals; the sponsor’s contrast uses a 90% interval. These interval levels are different, so their widths should not be compared as if they came from an identical uncertainty procedure.

What is the treatment contrast?

A treatment contrast compares outcomes under the study treatments. In a crossover study, the same participant may contribute data under both treatment conditions, which creates a within-participant relationship between observations. A method that accounts for this relationship can differ from a method that treats treatment groups as independent. The registry’s sponsor result is described as a mean difference in final values, but the method is not specified. It is therefore not possible to determine from the record whether the analysis accounted for the crossover structure, period, sequence, baseline, or within-participant correlation.

The distinction between an arm summary and a treatment contrast is important. An arm mean describes a treatment-specific average. A contrast is an estimate of how the treatment conditions differ under a particular definition and analysis. The record reports the contrast as 3.00, with a 90% confidence interval from -0.28 to 6.28, but does not state the detailed formula or model. Because the endpoint description defines change as baseline minus the time-point value, the direction of the arm-level change follows that convention. The sponsor contrast is labeled as final values, and its direction convention is not elaborated in the record.

How should confidence intervals be read?

In general, a confidence interval combines an estimate with information about its precision under a statistical procedure. A relatively narrow interval indicates more precision than a wider interval, while an interval spanning values on both sides of zero can leave the direction of a contrast uncertain. The sponsor’s interval extends from a small negative value to a positive value and includes zero. This is not proof that the treatments are identical; it indicates that the reported interval does not exclude zero under that analysis. The record does not report a p-value to accompany the sponsor comparison.

Arm-specific confidence intervals answer a different question from an interval for the between-treatment difference. Overlap or lack of overlap between separate arm intervals is not a substitute for the interval of the contrast, particularly in a crossover design where observations may be paired. The posted arm intervals are useful for describing uncertainty around each reported mean, while the sponsor interval is the direct reported comparison. The missing method description limits assessment of how that direct comparison was calculated.

What does the crossover structure require?

Randomized treatment order can help balance treatment sequence across participants. Crossover designs also require attention to period effects, treatment sequence, and possible carryover from an earlier period. General statistical practice may address these features in the design or analysis, but the registry does not report whether or how they were assessed. It would be inappropriate to assume a washout period or a particular model. The record also does not state whether the primary comparison used paired measurements or an alternative approach.

The registry gives a primary endpoint population description for Part 1 and labels it all patients as treated. It reports QTcF data counts by treatment period and describes exclusions, but the arm-summary counts do not match those counts. Neither a missing-data method nor an imputation approach is reported. Without those details, the effect of missing or excluded measurements on the estimates cannot be evaluated from the registry record.

What do the planning illustrations represent?

General sample-size illustrations based on an observed effect are planning scenarios, not evidence about what the study was designed to detect and not observed power. The stated illustration assumptions use 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. Such calculations depend on assumptions and should not be read as a reconstruction of the trial’s original design. The registry’s own sample-size justification and detailed design assumptions are not reported in the registry record. The illustration results and individual planning values are not reported in the record summarized for this analysis.

Multiplicity concerns arise when several outcomes, time points, or comparisons are assessed, because the chance of at least one apparently unusual result can change as the number of analyses grows. The record lists eight outcome measures and eight analyses, but reports zero posted secondary results. It does not describe a multiplicity adjustment or identify a hierarchy for secondary endpoints. No interim analysis plan is reported in the registry record.

What is not available for evaluating the method?

The method for the sponsor’s comparison, the handling of period and sequence, the approach to missing observations, and the basis for the confidence interval are not reported in the registry record. The record does not give a detailed protocol-level statistical plan, the sample-size justification, or the identity of the parties who were masked. These omissions do not invalidate the posted estimates, but they constrain independent assessment of their precision and reproducibility. Interpretation should remain anchored to the outcomes and analysis description that are actually recorded.

10. Analyses Reported by the Sponsor

The sponsor’s reported analysis gives a mean difference in final values of 3.00 milliseconds, with a 90% confidence interval from -0.28 to 6.28. The interval includes zero and extends to values on both sides of it. The registry does not report a p-value or the statistical method used for this comparison. Its label indicates final values, while the primary endpoint summaries are mean changes from baseline, so the two presentations should not automatically be treated as the same estimand.

The arm summaries report a mean change of 1.78 for vorinostat and -1.22 for placebo, each with a 95% confidence interval. The sponsor contrast is the registry’s own comparison and should be kept distinct from a simple comparison of the arm summaries. The record does not explain whether the sponsor analysis adjusted for baseline or accounted for period, sequence, and within-participant correlation in the crossover design.

No discrepancy flag is recorded for the posted analysis, but the absence of a reported method still limits independent evaluation. The sponsor’s analysis is therefore the primary reported inferential comparison; the arm-level summaries provide context rather than a substitute for its analysis. The record does not provide enough detail to reproduce the sponsor result or determine exactly how it relates to the reported changes.

11. Multiplicity, Missing Data and Analysis Populations

The registry lists eight outcome measures and eight analyses, while zero secondary results are posted. The primary result discussed is the QTcF change at 0.5 hours. Although the number of listed measures and analyses indicates that the registry contains more than the primary endpoint, the record does not give a complete account of which analyses were prespecified, which were exploratory, or how the outcomes relate to one another.

Multiplicity is a general concern when several statistical tests or estimates are considered. If many comparisons are examined, the chance of observing an apparently unusual result among them can increase. Statistical plans may address this through prespecified endpoint hierarchies or adjustments, but such procedures are not reported in the registry record. An interim analysis plan and any associated stopping criteria are also not reported. No inference about whether multiplicity was addressed should be made from the posted counts alone.

The primary endpoint population is described as all patients as treated in Part 1. Its description states that 22 patients had QTcF data from the vorinostat period and 23 from the placebo period, and mentions exclusions for a protocol violation and missing predose measurements. The arm summaries instead list 23 for vorinostat and 22 for placebo. The record does not reconcile these counts or explain the relationship between the population description and the arm summaries.

A missing-data strategy, imputation method, or sensitivity analysis is not reported in the registry record. The stated exclusion of participants without predose measurements indicates that baseline availability was relevant to the endpoint, but the record does not say how other incomplete observations were handled. It also does not describe whether the analysis used all available observations or required complete data across both crossover periods.

The record does not report a detailed analysis population definition beyond the all-patients-as-treated description, nor does it provide a protocol or statistical analysis plan that clarifies deviations from the intended approach. These gaps make it difficult to assess the effects of missingness, exclusions, or multiple analyses on the reported estimate. They should be treated as limitations of the registry description, not as evidence that a particular procedure was or was not used.

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 serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: disease progression (12.0%), abdominal pain (8.0%), vomiting (8.0%), pyrexia (8.0%), anemia (4.0%).

The safety record gives figures for all participants, with 25 at risk for serious adverse events and 11 affected. It also reports 25 at risk for other adverse events and 24 affected. These are descriptive counts in the registry record. They do not, on their own, establish that an event was caused by vorinostat, because the figures are not presented as a treatment-specific causal comparison.

The listed serious adverse event terms and percentages are Disease Progression at 12.0%, Abdominal Pain at 8.0%, Vomiting at 8.0%, Pyrexia at 8.0%, Anemia at 4.0%, and Ascites at 4.0%. These are the terms and percentages recorded; no additional event-level details, treatment-period attribution, or comparative risk estimates are given. The record does not report deaths in the available safety summary.

Serious adverse events and other adverse events are separate registry categories and should not be conflated. The figures describe events recorded among participants, not a complete causal assessment or a comparison between treatment periods. The registry does not provide enough detail to determine timing, recurrence, severity beyond the serious classification, or how events were adjudicated. These data are best read as a limited descriptive safety record, not as a basis for medical advice or a broad safety conclusion.

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

The results describe QTcF measurements in a small phase 1 crossover study of patients with relapsed or refractory advanced cancer. The posted arm means are 1.78 milliseconds for vorinostat and -1.22 milliseconds for placebo, with 95% confidence intervals of -1.53 to 5.09 and -4.53 to 2.08, respectively. The sponsor’s reported mean difference in final values is 3.00 milliseconds, with a 90% confidence interval from -0.28 to 6.28. These are estimates with uncertainty, not exact values known without error.

The sponsor interval includes zero, leaving uncertainty about the direction and size of the contrast under the reported analysis. This does not prove the absence of a difference, nor does it establish that a difference exists. The registry gives no p-value for this comparison, and its analysis method is not specified. The arm summaries and sponsor comparison have different labels, and the record does not explain precisely how the contrast relates to the arm-level changes.

The primary endpoint measures an electrocardiographic interval at a specified time point after a single dose. It does not measure a long-term clinical outcome. The study’s results cannot be extended automatically to other populations, doses, schedules, or settings. The registry does not report a clinical interpretation threshold for the endpoint, so the displayed statistical estimates alone do not establish the clinical importance of any change.

The design has a crossover structure, but the registry does not report details needed to assess handling of period effects, treatment sequence, or carryover. It also gives inconsistent treatment-specific counts between the endpoint population description and arm summaries. The missing statistical-method details and missing-data approach further constrain interpretation. These limitations affect how confidently the reported contrast can be evaluated, without changing the fact that the registry reports the estimates stated above.

The safety figures are descriptive and do not provide a causal treatment comparison. Overall, the registry records a set of short-term QTcF estimates and safety counts for this study. They do not, by themselves, support medical advice or a general claim about treatment benefit or harm.

14. Why This Trial Matters Statistically

This trial illustrates why statistical interpretation depends on the design as well as the numerical estimate. A crossover study can compare treatments within participants, but the analysis needs to reflect the sequence and period structure. The registry identifies a two-period crossover design yet does not report the sponsor’s statistical method or explain how the paired structure was handled. That omission makes the contrast harder to assess independently than the point estimate alone might suggest.

It also demonstrates the importance of distinguishing treatment-specific means from a between-treatment comparison. The arm summaries describe mean changes from baseline and give 95% confidence intervals. The sponsor’s reported result is a mean difference in final values with a 90% confidence interval. Those labels describe related but not necessarily identical quantities. A reader should avoid treating the intervals or estimates as directly interchangeable when the analysis details are not stated.

The endpoint definition adds another interpretive issue: the registry defines change as baseline minus the value at the time point. Making the direction explicit matters because positive and negative values depend on the subtraction convention. Statistical summaries can be misunderstood when the direction of an outcome is assumed rather than checked.

Finally, the registry’s descriptions of treatment-specific sample counts do not align: the primary population narrative gives 22 vorinostat-period and 23 placebo-period observations, while the arm summaries list 23 and 22, respectively. This highlights the value of checking population definitions and reporting consistency before drawing conclusions from confidence intervals. The study is a useful example of how estimates, design, endpoint definitions, and reporting completeness jointly shape statistical interpretation.

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 principal limitation is incomplete reporting of the analysis. The sponsor comparison is labeled as a mean difference in final values, but its method is not specified. The registry does not state whether the crossover structure, period, sequence, baseline, or within-participant correlation was accounted for. It also does not report a p-value or the assumptions used to construct the confidence interval.

The treatment-specific counts are inconsistent across the endpoint narrative and arm summaries. The endpoint description says 22 patients had QTcF data from the vorinostat period and 23 from the placebo period, while the arm summaries list 23 for vorinostat and 22 for placebo. The record does not explain this difference. The endpoint narrative mentions exclusions for a protocol violation and missing predose measurements, but the text does not provide a complete account of all missing or excluded observations.

The study enrolled 25 participants and was focused on a short-term electrocardiographic outcome at 0.5 hours. The results therefore have a limited scope. The registry does not report the sample-size justification, detailed statistical analysis plan, missing-data method, or any sensitivity analyses. Multiplicity procedures and interim analyses are not reported. The record also does not identify which parties were masked, describe placebo matching, or provide details about crossover period spacing and washout.

The arm summaries do not include a measure of participant-level spread beyond the confidence intervals around the means. The sponsor’s reported comparison and the arm-level change summaries are not fully reconciled, and the endpoint’s subtraction direction must be kept in view. These reporting gaps mean the registry estimates can be described, but their derivation cannot be independently reconstructed from the record. These limitations should temper interpretation without implying that unreported procedures were absent.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: A Crossover Study to Assess the Effects of Vorinostat (MK0683, SAHA) in Patients With Advanced Cancer (0683-070)(COMPLETED) — NCT00632931.
• Registry record last updated 2015-07-30; status last verified 2015-07.
• Results first posted 2010-02-15.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

19. Record Summary

The registry describes a completed randomized, phase 1, partially blinded, placebo-controlled, two-period crossover study of a single dose of vorinostat in patients with relapsed or refractory advanced cancer. It enrolled 25 participants and assessed QTcF change at 0.5 hours. The primary endpoint is reported as a continuous measure in milliseconds, with arm means and 95% confidence intervals.

The sponsor’s reported mean difference in final values is 3.00, with a 90% confidence interval from -0.28 to 6.28. The method and p-value are not reported in the registry record. Treatment-specific counts differ between the endpoint population description and the arm summaries. The record also does not report the detailed crossover analysis, missing-data method, sample-size justification, or multiplicity plan. Safety figures are descriptive and do not establish causality.