BACKGROUND All-trans-retinoic acid (all-trans RA) induces complete remission in most patients with acute promyelocytic leukemia (APL). However, continuous oral dosing results in progressive decline in plasma drug concentrations, which is associated with relapse and resistance to this retinoid. We speculated that the decline in drug levels, indicating acquired resistance, resulted partly from inducible cytochrome-P450 oxidative enzymes, which can catabolize all-trans RA. OBJECTIVE We studied the clinical pharmacology of all-trans RA in cancer patients to determine possible mechanisms of acquired resistance and evaluated the potential for reversal by ketoconazole, an inhibitor of cytochrome-P450 oxidative enzymes. METHODS Serial plasma samples were obtained from 54 patients with APL or advanced lung cancer after a single oral dose of all-trans RA (45 mg/m2). In the 34 patients with advanced lung cancer, all-trans RA (45 mg/m2) was administered twice daily for 4 weeks, and, on days 2, 28, and 29, serial plasma samples were again obtained after a single 45-mg/m2 dose. One hour prior to drug administration on days 2 and 29, a single oral dose (200-1200 mg) of ketoconazole was administered. Endogenous plasma concentrations of all-trans RA and 13-cis-retinoic acid were measured in a subset of these patients and in 11 with early-stage lung cancer. RESULTS The mean area under the curve for plasma drug concentration times time (AUC) for all-trans RA on day 1 varied substantially among patients. Compared with patients with APL, the 28 patients with advanced lung cancer who completed therapy demonstrated significantly lower AUC levels on day 1 (P = .06); a subgroup with levels less than 300 ng/mL per hour on day 1 had lower endogenous plasma all-trans RA concentrations than patients with APL or early-stage lung cancer or 14 normal subjects. Following continuous oral treatment, the mean day 28 AUC for all-trans RA was significantly lower than that on day 1 (213 ng/mL per hour versus 467 ng/mL per hour; P < .01), a decline significantly attenuated by ketoconazole, which increased the mean plasma all-trans RA AUC on day 29 to 375 ng/mL per hour (P < .01). CONCLUSIONS Reported variability for the pharmacokinetics of all-trans RA may result from disease-related or population-based differences in basal catabolic rates influenced by genetic or environmental factors. However, the pattern of inducible catabolism of all-trans RA is not disease specific. Ketoconazole attenuates this accelerated catabolism, suggesting that oxidation by cytochrome-P450 enzymes is an important pathway for both constitutive and induced pathways of all-trans RA metabolism.