The effect of duration of induction treatment on complete response (CR) and event-free survival (EFS) (time to death or relapse) was evaluated in 162 "good-risk" germ cell tumor (GCT) patients treated from November 1982 to July 1986 in a randomized prospective trial with VAB-6 (cyclophosphamide + vinblastine + bleomycin + dactinomycin + cisplatin; 81 patients) versus etoposide + cisplatin (EP; 81 patients). Patients received three cycles of VAB-6 every 28 days or four cycles of EP every 21 days. Treatment cycle with either regimen was routinely postponed for 7 days for leukocytes less than 3000/mm3 or platelets less than 100,000/mm3 and then administered regardless of blood count. The number of treatment days was calculated for each patient from initial treatment day to final induction date plus 28 days for VAB-6 and plus 21 days for EP. The proportion of CR and the EFS for good-risk GCT patients treated with cisplatin-based chemotherapy were not influenced by a less than or equal to 7-day delay resulting from chemotherapy-induced myelosuppression. Short, planned delays in chemotherapy for good-risk GCT patients (less than or equal to 7 days per cycle) appear to be acceptable since they may prevent serious toxicity in this curable patient population. Delays of longer than 7 days are strongly discouraged except in extraordinary life-threatening circumstances.