Midazolam is a common probe used to predict CYP3A activity, but multiple blood samples are necessary to determine midazolam's area under the concentration-time curve (AUC). As such, single sampling strategies have been examined. The purpose of this study was to assess the ability of single midazolam concentrations to predict midazolam AUC in the presence and absence of CYP3A modulation by Ginkgo biloba extract (GBE). Subjects received oral midazolam 8 mg before and after 28 days of GBE administration. Postdose blood samples were collected during both study periods and midazolam AUC determined. Linear regression was used to generate measures of predictive performance for each midazolam concentration. The geometric mean ratio (90% confidence intervals) of midazolam AUC(0-infinity) post-GBE/AUC(0-infinity) pre-GBE was 0.66 (0.49-0.84) (P = .03). Before and after GBE administration, optimal midazolam sampling times were identified at 3.5 to 5 hours and 2 to 3 hours, respectively. Single midazolam concentrations between 2 and 5 hours correctly predicted the reduction in midazolam AUC following GBE exposure, but confidence intervals were generally wide. Intersubject variability in CYP3A activity (either inherent or from drug administration) alters the prediction of optimal midazolam sampling times; therefore, midazolam AUC is preferred for assessing CYP3A activity in drug-drug interaction studies.