Hexamethylene bisacetamide (HMBA, NSC 95580) has been demonstrated to be the most effective of the known and studied polar-planar compounds at inducing differentiation in a wide variety of leukemic and nonleukemic cell lines. Although HMBA demonstrated no antineoplastic activity in preclinical testing, it was selected for clinical development on the basis of its potent differentiating capabilities in vitro. In this phase I study, HMBA was administered as a continuous five-day infusion every 3 weeks to patients with advanced cancer. Twenty-three patients received 35 evaluable courses at doses that ranged from 4.8 to 33.6 g/m2/d. Dose-limiting toxicities included renal insufficiency, a hyperchloremic metabolic acidemia/acidosis, and CNS toxicities manifested by agitation and delirium, which progressed to coma in one patient who developed concomitant renal insufficiency. Moderate myelosuppression, mucositis, nausea, and vomiting were also observed. The pharmacokinetics of HMBA best fit a single compartmental model and disposition is primarily by renal elimination. Renal excretion of HMBA and of the primary metabolite, 6-acetoamidohexanoic acid, together account for the disposition of 66% to 93% (mean, 74%) of the infused drug. Based on this trial, the maximum tolerated and recommended phase II doses for HMBA administered on this schedule are 33.6 and 24 g/m2/d, respectively. However, since steady-state HMBA levels at these doses were in the range of 1 to 2 mmol/L, only approaching the lower limit demonstrated for in vitro differentiating effectiveness, and because of evidence suggesting that the exposure period is an important variable in the induction of differentiation, additional studies examining longer periods of infusion are warranted.