Assessing the Relationship between Investigator Global Evaluation and Acne Lesion Counts *

Abstract

Efficacy evaluation for acne clinical trials is usually based on analysis of: (1) change (or relative change) in lesion counts from baseline and (2) the investigator global evaluation (IGE) at the final study endpoint, usually dichotomized to treatment success or failure. As lesion counts provide a quantitative assessment of the subject's acne condition and the IGE reflects the dermatologist's visual evaluation of the same condition, one would expect the two evaluations to be related. In this article we investigate the relationship between the two assessments by fitting logistic regression models to three data sets with different features. The results of the analysis show that final lesion counts explain very well the variability in the IGE treatment success. However, changes in lesion counts per se do not fully explain the variability in IGE treatment success, as baseline lesion counts are still a significant covariate in the model. Finally, inflammatory lesions consistently have a much greater impact than noninflammatory lesions on explaining the variability in the IGE treatment success. The findings of this study emphasize the importance of having an initial estimate of treatment effect before embarking on phase 3 trials. Such an estimate would be very useful in guiding the decision on whether to seek a lesion-specific or general acne indication, which, in turn, impacts the sample size required for establishing efficacy.