Thirty-nine patients with extremity osteogenic sarcoma and no discernible metastases were treated with amputation and postoperative adjuvant high-dose methotrexate with leucovorin rescue. One half of the patients were also randomized to receive bacillus Calmette-Guérin by a multiple-puncture technique. Results have been analyzed with a minimum followup of 10 months and a median followup of 27 months. Actuarial analysis estimates that 38% of current protocol patients remain continuously free of disease for 24 months compared to only 17.4% of historical control patients (P = 0.029; one-sided generalized Kruskal-Wallis test). Bacillus Calmette-Guérin administered by a multiple-puncture technique had no effect on disease-free interval. Minor differences between current protocol and historical control patients with regard to race, age, histologic type, and site and size of primary tumors do not affect the difference in disease-free interval between these two patient groups. However, current patients had somewhat lower grade lesions and consideration of the patients with grade III and IV lesions only, lessens the difference between current and historical control patients (P = 0.11; one-sided generalized Kruskal-Wallis test). High-dose methotrexate was administered with virtually no morbidity and no deaths. The small differences observed in this study between protocol patients treated with surgery plus high-dose methotrexate and historical control patients treated with surgery alone point to the need for a prospective randomized study to establish the role of high-dose methotrexate in the adjuvant treatment of patients with osteogenic sarcoma.