Hexamethylene bisacetamide (HMBA, NSC 95580), a potent polar-planar differentiating agent in vitro, was studied in a phase I trial as a 10-day continuous infusion administered every 4 weeks. Since preclinical evidence had demonstrated that the duration of HMBA exposure was an important variable in the induction of differentiation, and HMBA steady-state concentrations (Css) achieved during 5-day infusions were minimally effective at inducing in vitro differentiation, this study was conducted to evaluate the feasibility of maintaining HMBA Css for 10 days similar to levels achieved and maintained for 5 days. Twelve patients received 17 evaluable courses at three dose levels, 12, 15.8, and 20 g/m2/day. Platelet toxicity limited further dose escalation. Mean nadir and percentage of decrement in platelet counts were 175,000/microliter and 66%, and 43,500/microliter and 83%, at 15.8 and 20 g/m2/day, respectively. In this 10-day study, the percentage of decrement of platelet counts was linearly related to mean HMBA Css and to the area under the plasma concentration versus time curve (AUC). However, when combined platelet and pharmacological data from both 5- and 10-day studies were analyzed, it was apparent that the duration of HMBA exposure was an additional significant variable in predicting the magnitude of thrombocytopenia. Renal and metabolic toxicities that precluded dose escalation in previous 5-day trials of HMBA were mild and insignificant in this study. Mean HMBA Css were 0.65 +/- 0.09 mmol/liter at 15.8 g/m2/day, and 0.99 +/- 0.22 mmol/liter at 20 g/m2/day. Depletion of intracellular polyamines in peripheral blood mononuclear leukocytes was noted during several courses that were associated with profound myelosuppression, but no clear relationships were apparent between the magnitude of polyamine changes and toxicity or between fluctuations in polyamines and the magnitude of mean HMBA Css values. Based on this study, the maximum tolerated and recommended phase II doses for HMBA administered on this schedule were 20 and 15.8 g/m2/d, respectively. The administration of HMBA by a ten-day infusion at the maximum tolerated dose resulted in the delivery of lower daily doses and lower HMBA Css than on the 5-day schedule. The major toxicities differed on these schedules with thrombocytopenia being most prominent on the 10-day schedule and metabolic and CNS toxicities on the 5-day schedule.(ABSTRACT TRUNCATED AT 400 WORDS)