Sample Size Calculation for Thorough QT/QTc Study Considering Various Factors Related to Multiple Time Points

Abstract

The sample size requirement in a thorough QT/QTc study is discussed under a balanced parallel or crossover study design. First, we explore the impacts of various factors on the study power, including the mean effect profile across time and correlation among time points. Then we estimate the variability parameters needed based on multiple historical studies. Different baseline usage is illustrated to have a significant impact on the analysis variability in the parallel studies. Finally, the sample size calculations and recommendations are given for demonstrating a “negative” drug effect and the study assay sensitivity, respectively.

Key Words:

• False positive
• IUT test
• Multiple test correction
• Positive control
• Power

ACKNOWLEDGMENTS

The authors thank Frank Poitiers, David Hoffman, Fan Li, Gordon Sun, Ji Zhang, Chunpeng Fan, Catherine Ortemann-Renon, Denis Bonnet, and three reviewers for their helpful comments.

Notes

Note. SD within-sub = Sqrt(Var(subj∗period) + Var(residual)); partial correlation = Var(subj ∗ period)/Var(within-sub).

Note. SD(total) = Sqrt(Var(sub) +Var(residual)); Correlation = Var(subj)/Var(total). All ANCOVA or ANOVA models presented here are repeated measurement models. D-1 is the baseline day.

Note. SDwithin-sub = Sqrt(Var(subj∗period) + Var(residual)); partial correlation = Var(subj∗period)/ Var(within-sub).

Note. SD(total) = Sqrt(Var(sub) +Var(residual)); correlation = Var(subj)/Var(total).

Note. SD within-sub = Sqrt(Var(subj∗period) +Var(residual)); partial correlation = Var(subj∗period)/Var(within-sub).

Note. SD(total) = Sqrt(Var(sub) +Var(residual)); correlation = Var(subj)/Var(total).