A phase I clinical trial of 13-cis-retinoic acid (cis-RA) was undertaken to determine the maximally tolerated dose (MTD) and pharmacokinetics (PK) of cis-RA following bone marrow transplantation (BMT) in children with high-risk neuroblastoma. Mean peak serum levels of cis-RA in 31 pediatric patients ranged from 4.9 to 8.9 microM following doses of 100-200 mg/m2 per day, divided into two doses every 12 h administered orally. The PK of cis-RA obeyed a single-compartment model following first-order absorption in the majority of patients. A linear increase in the mean peak serum levels and area under the time-concentration curve (AUC) with increasing dose was observed. The average half-lives of absorption and elimination were 1.0 and 5.8 h, respectively. At the MTD of 160 mg/m2 per day, the mean cis-RA peak serum concentration was 7.2 +/- 5.3 microM. AUC values were not altered significantly during a 2-week course of treatment or over a long period of multiple courses. Levels of trans-retinoic acid, a metabolite of cis-RA, remained low but were similar on days 1 and 14, whereas the 4-oxo-13-cis-RA metabolite had increased in 64% of patients by day 14. Peak serum cis-RA concentrations correlated with clinical toxicity as grade 3 to 4 toxicity was seen in 44% of patient-courses (8/18) with peak serum levels > 10 microM, but only 13% (12/96) with peak serum levels < 10 microM. These results show that cis-RA given at 160 mg/m2 to children achieved serum concentrations known to be effective against neuroblastoma in vitro, and the PK for cis-RA differs from that reported for trans-retinoic acid in children.