We compared the clinical and microbiologic efficacy of two broad-spectrum combination antimicrobial regimens in the treatment of 148 patients with acute pelvic inflammatory disease. Patients were randomized to inpatient treatment with either cefoxitin and doxycycline (n = 75) or clindamycin and tobramycin (n = 73). These antibiotics were administered intravenously for at least 4 days, and up to 48 hours beyond defervescence. Patients were discharged on a regimen of oral doxycycline or clindamycin in accordance with the intravenous regimen to complete a total duration of therapy of 2 weeks. Neisseria gonorrhoeae (53%) and Chlamydia trachomatis (31%) were the microorganisms that were isolated most frequently from the genital tract of enrolled patients. At follow-up, N. gonorrhoeae was isolated in two patients, and C. trachomatis was isolated in none. The overall initial favorable response rate to combination antimicrobial therapy was 98.5% (130/132) in patients with uncomplicated pelvic inflammatory disease and 81% (13/16) in patients with pelvic inflammatory disease that was complicated by tuboovarian abscess. A greater than 70% decrease in abdominal tenderness score occurred in 89% of 111 patients within 6 weeks of hospital discharge. There were no significant differences between antibiotic treatment groups in any response categories or in toxicity. During the initial hospitalization, five patients (three with tuboovarian abscess; one with a pyosalpinx, and one with intractable acute and chronic pelvic inflammatory disease) required surgical intervention. These results support the recommendation to use broad-spectrum combination antimicrobial therapy for the treatment of acute pelvic inflammatory disease.