Objective: To evaluate the usefulness of World Health Organisation (WHO's) clinical case-definition (CCD) for AIDS in a private hospital.
Design: A prospective study.
Setting: Shree Hindu Mandal Hospital, Dar es Salaam, Tanzania.
Subjects: A total of 601 patients (> 14 years) were studied from January 1995 to December 1997.
Methods: Using HIV test results as a reference standard, sensitivity, specificity, positive predictive values (PPV) and negative predictive values of signs and symptoms were calculated. Multiple logistic regression was used to determine a set of predictive symptoms and signs. Stepwise logistic regression modelling was used to choose the final model.
Results: The frequently occurring signs and symptoms among the 473 sero-positive patients were fever (226), oral candidiasis (167), weight loss (161), chronic cough (157), diarrhoea (100) and pulmonary tuberculosis in 69 cases. The presence of anorectal lesions and the rarity of pneumocystis carinii pneumonia in this study are important findings. Seven clinical characteristics predicted HIV infection. These included pulmonary tuberculosis (p=0.009), lymphadenopathy (p=0.007), diarrhoea (p=0.000), chronic cough (p=0.001), dermatitis (p=0.003), herpes zoster (p=0.01) and oral candidiasis (p=0.000).
Conclusions: A greater number of HIV positive patients presented with signs and symptoms different from those proposed by WHO's CCD were observed in this study. With environmental pathogens varying from one geographical region to another and new ones appearing, opportunistic disease cannot be constant in AIDS patients. Therefore, AIDS diagnosis based on clinical case definition alone without at least one positive HIV antibody test is inaccurate and no longer justified.