The best way of treating pelvic inflammatory disease (PID) is not known. The clinical response to two treatment regimens (penicillin plus metronidazole v doxycycline plus metronidazole) was studied in 33 patients with PID confirmed by laparoscopy and endometrial biopsy. The overall failure rate, according to the criteria used in this study was five of 11 (45%) women with chlamydial PID, none of six women with gonococcal PID, all of four women with chlamydial gonococcal PID, and three (25%) of 12 women with non-chlamydial non-gonococcal PID. The failure rate with penicillin plus metronidazole treatment was unacceptably high (53%), and significantly higher than that with doxycycline plus metronidazole (19%) (p = 0.038). In most cases the microbiological and histopathological evaluations identified a probable explanation for the poor response to the treatment regimen used.