INTRODUCTION: Systemic clearance of intravenous (IV) alfentanil (ALF) is an in vivo probe for hepatic cytochrome P450 (CYP) 3A activity, miosis is a surrogate for plasma ALF concentrations, and IV ALF miosis is a noninvasive probe for hepatic CYP3A. This investigation characterized the bioavailability and first-pass metabolism of oral ALF and tested the hypotheses that (1) first-pass ALF clearance reflects first-pass CYP3A activity, (2) miosis after oral ALF will reflect intestinal and hepatic CYP3A activity, and (3) miosis can approximate plasma concentration-based pharmacokinetic measures for IV and oral ALF as a noninvasive in vivo probe for hepatic and first-pass CYP3A activity and drug interactions. Results were compared with those for midazolam (MDZ), an alternative CYP3A probe. METHODS: Ten volunteers were studied by use of a randomized, 9-way, crossover design after administration of rifampin (INN, rifampicin) (hepatic and intestinal CYP3A induction), troleandomycin (TAO) (hepatic and intestinal CYP3A inhibition), grapefruit juice (selective intestine CYP3A inhibition), or nothing (control). For each condition, they received 1 mg IV MDZ and then 15 microg/kg IV ALF, as well as 3 mg oral MDZ and then oral ALF (23 or 60 microg/kg) on another day. Plasma concentrations were determined by liquid chromatography-mass spectrometry. Dark-adapted pupil diameters were measured coincident with blood sampling. ALF effect was analyzed similarly to concentration to yield an effect "clearance" (Dose/Area under the pupil diameter change versus time curve). RESULTS: Bioavailability (Foral), hepatic extraction (EH), and intestinal availability (FG) were 0.26 +/- 0.08, 0.52 +/- 0.09, and 0.56 +/- 0.20, respectively, for MDZ and 0.42 +/- 0.15, 0.28 +/- 0.09, and 0.56 +/- 0.18, respectively, for ALF. Oral clearance (CL/F) was 34.7 +/- 12.8 and 10.9 +/- 3.5 mL.kg -1.min -1 , respectively, for MDZ and ALF. After rifampin, TAO, and grapefruit juice, ALF F oral was 0.04 +/- 0.02 (P <.05, versus control), 0.99 +/- 0.18 (P <.05, versus control), and 0.62 +/- 0.18 (P <.05, versus control), respectively; E H was 0.69 +/- 0.14 (P < .05, versus control), 0.04 +/- 0.01 (P <.05, versus control), and 0.26 +/- 0.08, respectively; F G was 0.16 +/- 0.10 (P <.05, versus control), 1.0 +/- 0.2 (P <.05, versus control), and 0.85 +/- 0.30 (P <.05, versus control), respectively; CL/F was 339 +/- 233 (P <.05, versus control), 0.62 +/- 0.26 (P <.05, versus control), and 6.7 +/- 2.5 (P <.05, versus control), respectively, and effect clearance was 2.1 +/- 1.1 (P <.05, versus control), 0.087 +/- 0.056 (P <.05, versus control), and 0.54 +/- 0.30 (0.73 +/- 0.43 mg.mm -1.h -1 in controls), respectively. There were significant correlations between ALF and MDZ systemic clearances (r2= 0.92), EH (r2=0.93), and CL/F (r2= 0.97), as well as between oral ALF effect (miosis) clearance and oral clearance (r2=0.59). CONCLUSIONS: ALF and MDZ have similar intestinal extraction but low and intermediate hepatic extraction, respectively. Systemic and oral clearances of ALF are excellent in vivo probes for hepatic and first-pass CYP3A activities and drug interactions. Miosis was an acceptable surrogate for plasma ALF. ALF miosis may be a suitable noninvasive in vivo probe for both hepatic and first-pass CYP3A.