Genital ulcer disease: accuracy of clinical diagnosis and strategies to improve control in Durban, South Africa.

OBJECTIVE: To investigate the accuracy of clinical diagnosis in genital ulcer disease (GUD); to devise management strategies for improving the control of GUD and thereby limit the spread of HIV-1 infection. DESIGN: Clinical and microbiological assessment of GUD in men and women. The index of suspicion, diagnostic accuracy, diagnostic efficiency and positive and negative predictive values of a clinical diagnosis were investigated. SETTING: City Health Sexually Transmitted Diseases Clinic, King Edward VIII Hospital, Durban, South Africa. PARTICIPANTS: 100 men and 100 women with genital ulcers. RESULTS: The accuracy of a clinical diagnosis was, in men: lymphogranuloma venereum (LGV) 66%, donovanosis 63%, chancroid 42%, genital herpes 39%, primary syphilis 32%, mixed infections 8%, and in women; secondary syphilis 94%, donovanosis 83%, genital herpes 60%, primary syphilis 58%, chancroid 57%, LGV 40%, mixed infections 14%. Overall, diagnostic efficiency was greater in women than in men. When compared with other causes of GUD, donovanosis ulcers bled to the touch and were larger and not usually associated with inguinal lymphadenopathy. In women, extensive vulval condylomata lata were readily differentiated from all other causes of GUD. CONCLUSION: A clinical diagnosis in genital ulceration was less accurate in men than in women. The diagnostic accuracies for donovanosis and secondary syphilis were relatively high but for most other conditions were low. Differences between clinical and laboratory diagnostic accuracies may reflect similarities between the clinical appearances of the various causes of GUD, the presence of mixed infections, atypical ulceration due to longstanding disease, and insensitive laboratory tests. In this community all large ulcers should be treated empirically for syphilis and donovanosis. Uncircumcised men with GUD are an important HIV core or "superspreader" group locally, and prevention strategies should include counselling and health education in the light of the inaccuracy of clinical diagnosis found in this study. The development of rapid accurate tests for GUD is urgently required.