The sodium salt of pyrazine-2-diazohydroxide (PZDH; NSC 361456) was identified as an active congener of the antitumor lead pyridine-2-diazotate with enhanced chemical stability under physiological conditions. In a phase I trial of PZDH administered as a single i.v. bolus injection, 19 patients with refractory solid tumors received 44 courses of therapy at dose levels ranging from 50 to 350 mg/m2. No objective responses to PZDH were noted. Myelosuppression characterized by prolonged, delayed onset leukopenia and thrombocytopenia was the dose limiting toxicity. A maximum tolerated dose of 350 mg/m2 was identified for this treatment schedule. Nonhematological toxicity was limited to severe nausea and vomiting, experienced by all patients treated at the lower doses, although reasonably well controlled when antiemetics were given prior to chemotherapy. The plasma pharmacokinetics of PZDH was evaluated following a single course of therapy in 16 patients. Drug levels were monitored using a specific capillary gas chromatographic assay with a 1-ng/ml lower limit of quantitation. In patients treated with doses greater than 50 mg/m2, the concentration of PZDH in plasma declined in a distinctly triexponential manner and remained above 1.5 ng/ml for at least 8 h. However, the initial decay phase, characterized by a harmonic mean half-life of 3.9 +/- 3.5 (SD) min (range, 2.2-6.3 min), was the primary determinant of drug disposition, as indicated by its 85.5-93.1% contribution to the area under the plasma concentration-time profiles from time zero to infinity. The harmonic mean terminal half-life increased with escalations in dose from 2.7 +/- 0.8 h (n = 2) at 100 mg/m2 to 8.5 +/- 3.0 h at 350 mg/m2 (n = 6). Total plasma drug clearance was very similar in patients treated with doses of 50-250 mg/m2, exhibiting a mean value of 42.5 +/- 7.8 liters/h/m2 (n = 10); however, it was significantly lower at the 350 mg/m2 dose level, 27.2 +/- 6.6 liters/h/m2 (n = 6; P < 0.002), denoting a departure from linear pharmacokinetic behavior. The rather low steady state apparent volume of distribution, which ranged from 6.0 +/- 1.5 (50 mg/m2, n = 2) to 12.7 +/- 8.0 (350 mg/m2, n = 6) liters/m2, was indicative of limited distribution of the drug into body tissue. The absence of objective antitumor effects should not discourage continued evaluation of PZDH against solid tumors selected for probable sensitivity as indicated by preclinical testing. A dose of 250 mg/m2 on a single i.v. bolus schedule is recommended for these phase II trials.