To demonstrate that with carbamazepine (CBZ), positive results of bioequivalence trials in healthy volunteers cannot be extrapolated to patients because metabolic enzyme induction fundamentally changes the pharmacokinetics of CBZ. Bioequivalence trials were simulated assuming normal and induced metabolic clearance. The relevant pharmacokinetic parameters were drawn from published studies. The Cmax ratio depends on the clearance. A generic product which is fully compliant with the regulatory requirements in healthy volunteers could be non-bioequivalent in patients with enhanced elimination. A statement of bioequivalence of CBZ formulations, based on a single-dose study performed in healthy subjects, may not hold for a target patient population in the steady state.