Women with human immunodeficiency virus (HIV) infection who develop pelvic inflammatory disease (PID) are thought to have more severe illness and a greater need for surgical intervention. The current recommendation from the Centers for Disease Control and Prevention (CDC) is to administer parenteral antibiotics to all HIV-infected women with PID. Some recent studies have suggested an increased risk of nongonococcal, nonchlamydial PID among HIV-infected women. Bukusi and associates studied the effect of HIV infection on symptoms of PID, causal organisms and response to ambulatory therapy. Only one study has evaluated ambulatory treatment of HIV-negative women with PID.

A total of 162 women 18 to 40 years of age with lower abdominal pain for less than one month fitting the clinical criteria for PID underwent evaluation including endometrial biopsy and screening for HIV and other sexually transmitted infections. The women had delayed seeking medical attention for a median of seven days after first developing abdominal pain. Patients were treated with a single 500-mg oral dosage of ciprofloxacin and 10 days of oral metronidazole in a dosage of 500 mg and oral doxycycline in a dosage of 100 mg every 12 hours. Patients were reevaluated one week and one month after the initial visit. Repeat cervical cultures were obtained at the one-month follow-up visit. Of the 34 percent of subjects who were HIV-positive, 26 percent met World Health Organization criteria for symptomatic HIV, and 15 percent met the CDC definition for acquired immunodeficiency syndrome. Histologic criteria for endometritis were met in 63 (39 percent) of the 162 cases. Women with confirmed PID were more likely to be HIV positive than patients who did not have PID.

The HIV-positive patients with PID were found to have significantly more vaginal discharge and more severe abdominal pain. Neisseria gonorrhoeae, Chlamydia trachomatis, or both were detected more commonly among women with confirmed PID, and the white blood cell count was significantly higher. The degree of immunosuppression as measured by the CD4 T-cell count in HIV-positive women did not seem to affect the clinical severity of symptoms. Adnexal masses were rare overall, detected in 10 percent of HIV-positive women and 3 percent of HIV-negative women. Bacterial vaginosis was most commonly diagnosed in HIV-infected women. The HIV-infected women with the lowest CD4 T-cell counts had the highest rate of confirmed endometritis, usually bacterial vaginosis.

A total of 83 percent of the women with confirmed endometritis returned for follow-up at one week and at one month. At the one-week visit, all patients reported diminished lower abdominal pain. Seventy-five percent of women were considered cured at one week, and 83 percent were considered cured at one month. The cure rates did not differ according to HIV status. Women with PID who were infected with N. gonorrhoeae, C. trachomatis, or both did not achieve higher cure rates than patients infected with nongonococcal, nonchlamydial PID. No patient required hospitalization because of failed ambulatory therapy. Thirteen percent of women infected with N. gonorrhoeae, C. trachomatis, or both had persistent or recurrent gonococcal infections at follow-up, but none had chlamydial infection.

Results of this study demonstrate that histologically confirmed endometritis is more common among HIV-infected women than among HIV-negative women, and that HIV-infected women are more likely to have bacterial vaginosis than a gonococcal or chlamydial infection. The results also demonstrate that HIV-positive women achieved high cure rates with outpatient oral antimicrobial therapy. It was also shown that outpatient therapy was adequate regardless of HIV serostatus and degree of immunosuppression. None of these patients required surgical intervention or adjustments of antibiotic regimens.

The authors conclude that HIV-infected women with mild to moderate PID can safely be treated as outpatients although it is not clear if outpatient regimens would be adequate for severely immunocompromised women. A relationship between HIV-associated PID and bacterial vaginosis appears to exist.

editor's note: One of the interesting aspects of this study is that the median delay in seeking care for symptoms of lower abdominal pain was seven days (range 1 to 30 days). One of the CDC's recommendations for beginning early empiric therapy for suspected PID is based on the delay of care data demonstrating that even a few days' delay in treatment can result in long-term adverse sequelae. This study raises the question if an aggressive approach is necessary. Long-term follow-up on pregnancy and infertility rates in these women would be helpful in determining the optimal strategy.—b.a.