The ability of various human derived in vitro systems to predict various aspects of the in vivo metabolism and kinetics of almokalant have been investigated in a multicenter collaborative study. Although almokalant has been withdrawn from further clinical development, its metabolic and pharmacokinetic properties have been well characterized. Studies with precision-cut liver slices, primary hepatocyte cultures, and hepatic microsomal fractions fortified with UDP-glucuronic acid all suggested that almokalant is mainly glucuronidated to the stereoisomers M18a and M18b, which is in good agreement with the results in vivo. Both in vivo and in vitro studies indicate that the formation of M18b dominates over that of M18a, although the difference is more pronounced with the in vitro systems. Molecular modeling, cDNA-expressed enzyme analysis, correlation analysis, and inhibition studies did not clearly indicate which P450 enzymes catalyze the oxidative pathways, which may indicate a problem in identifying responsible enzymes for minor metabolic routes by in vitro methods. All of the in vitro systems underpredicted the metabolic clearance of almokalant, which has previously been reported to be a general problem for drugs that are cleared by P450-dependent metabolism. Although few studies on in vivo prediction of primarily glucuronidated drugs have appeared, in vitro models may consistently underpredict in vivo metabolic clearance. We conclude that in vitro systems, which monitor phase II metabolism, would be beneficial for prediction of the in vivo metabolism, although all of the candidate liver-derived systems studied here, within their intrinsic limitations, provided useful information for predicting metabolic routes and rates.