Etoposide is a widely used cytotoxic agent with a broad spectrum of activity in human malignancies. This agent has been incorporated into many transplant regimens although toxicity occurs because of its poor water solubility and toxic excipients. Etoposide phosphate, a water soluble prodrug of etoposide, has been studied at conventional dosages in man and shown to have advantages over the parent compound. We have extended our previous experience with this new agent to evaluate the levels needed in transplantation protocols. This phase I study of intravenous high-dose etoposide phosphate over 2 h on days 1 and 2 was designed to determine whether or not dose linearity between the amount of etoposide phosphate administered to patients and generation of etoposide in vivo as seen with conventional dosages of this agent would be present at transplant-dose levels. In addition, the toxicities of these dose levels with the short infusion schedule were defined. A conservative dose escalation scheme was chosen based upon prior knowledge of etoposide. Thirty-one patients (19 male, 12 female) with CALGB performance status 0-1 with a variety of solid tumors entered this study. The patients were treated with dose levels of etoposide phosphate given as the etoposide-equivalent doses of 250, 500, 750, 1000, 1200, 1400, and 1600 mg/m2/day in 250-400 ml of normal saline given as an intravenous infusion over 2 h on days 1 and 2 every 28 days. After the maximal tolerated dose level was determined on this schedule, additional patients received etoposide phosphate as a 4 h infusion on both days in an attempt to reduce toxicities. G-CSF (5 micrograms/kg/day) was administered subcutaneously to all patients from day 3 until the WBC > or = 10000/microliters. Nonhematologic toxicity was considered to be dose limiting. Serial plasma samples for pharmacokinetics were obtained from patients on day 1 of cycle 1. For the 2 h infusion, the maximum tolerated dose of etoposide phosphate was 1000 mg/m2/day x 2 with dose limiting mucositis. In the small number of patients studied, the maximum tolerated dose was reached for the 4 h infusion at 1400 mg/m2/day of drug, again due to mucositis. Other toxicities, despite the rapid infusion schedule, were modest with transient mild headache being most common. At the highest doses etoposide phosphate was efficiently and rapidly dephosphorylated to etoposide. Etoposide generated by dephosphorylation of etoposide phosphate had plasma disposition curves characteristic of etoposide administered parenterally. One partial response occurred in a patient with small cell lung cancer. Etoposide phosphate can be rapidly infused in modest fluid volumes at dosages required for transplantation protocols with minimal acute side-effects. On a 2 h schedule, mucositis becomes the dose limiting nonhematologic toxicity. Mucositis seems to correlate with peak dose levels of the drug rather than total drug administered. On a 4 h infusion schedule given sequentially for 2 days, the maximum tolerated dosage could be increased 40% compared to the 2 h schedule. The relative ease of administration and the rapid conversion of this prodrug into etoposide should make it useful in high-dose therapy settings.