Large Intra-subject Variability in Caffeine Pharmacokinetics: Randomized Cross-over Study of Single Caffeine Product

Background Average bioequivalence has been criticized for not adequately addressing individual variations. Importance of subjects' blinding in bioequivalence studies has not been well studied. We explored the extent of intra-subject pharmacokinetic variability and effect of drug-ingestion unawareness in subjects taking single caffeine product. Methods A single-dose randomized cross-over design was used to compare pharmacokinetics of 200 mg caffeine, described as caffeine (overt) or as placebo (covert). Maximum concentration (C_max), C_max first time (T_max), area-under-the-concentration-time-curve, to last measured concentration (AUC_T), extrapolated to infinity (AUC_I), or to T_max of overt caffeine (AUC_Overttmax), and C_max/AUC_I were calculated blindly using standard non-compartmental method. Percentages of individual covert/overt ratios that are outside the ±25% range were determined. Covert-vs-overt effect on caffeine pharmacokinetics was evaluated by 90% confidence interval (CI) and 80.00-125.00% bioequivalence range. Results 32 healthy subjects (6% females, mean (SD) age 33.3 (7.2) year) participated in the study (28 analysed). Out of the 28 individual covert/overt ratios, 23% were outside the ±25% range for AUC_T, 30% for AUC_I, 20% for AUC_Overttmax, 30% for C_max, and 43% for T_max. There was no significant covert-vs-overt difference in any of the pharmacokinetic parameters studied. Further, the 90% CIs for AUC_T, AUC_I, C_max, AUC_Overttmax, and C_max/AUC_I were all within the 80.00-125.00% bioequivalence range with mean absolute deviation of covert/overt ratios of 3.31%, 6.29%, 1.43%, 1.87%, and 5.19%, respectively. Conclusions Large intra-subject variability in main caffeine pharmacokinetic parameters was noted when comparing an oral caffeine product to itself. Subjects' blinding may not be important in average bioequivalence studies.