Objectives: A retrospective study to confirm the clinical impression of an increasing prevalence of granuloma inguinale (GI) in women, and to evaluate its association with pregnancy and HIV infection. Methods: Clinical records of all patients with a definitive diagnosis of GI attending the gynecology and antenatal clinics at King Edward VIII Hospital, Durban, South Africa, over a period of 36 months (January 1991–December 1993). Results: A total of 123 women were diagnosed with GI. The diagnosis was made by tissue smear alone in 21% ( n = 26), histology 43% ( n = 53) and by a combination of smear and histology in the rest. Forty-two percent ( n = 52) were pregnant. The only difference between pregnant and non-pregnant women were the presence of rectal and pelvic lesions in the latter. Sixty-nine percent ( n = 36) delivered vaginally while the remaining ( n = 16) were delivered by cesarean section. The indications for cesarean section were obstetric except for a patient in labor with extensive untreated vulval granuloma. In the majority (85%) GI had no influence on pregnancy outcome. There was no evidence of congenital GI in the neonates. Twenty-seven percent ( 30 113 ) had positive syphilis serology and 16% ( 18 110 ) had antibody to HIV. There were no differences in the clinical features and outcome of HIV positive and negative women. Conclusion: This study shows that GI is increasing in pregnancy in Durban, South Africa. Despite the concern that pregnancy promotes dissemination of GI, such an effect could not be established as the clinical response to treatment and outcome were similar in both pregnant and non-pregnant women. Infection with HIV also did not alter the clinical presentation and outcome of the disease in the patients studied.