The prognostic value of the World Health Organisation staging system for HIV infection and disease in rural Uganda.

Abstract

To assess whether the WHO staging classification for HIV provides prognostically valuable and applicable information in rural Uganda. Data were obtained from a population-based cohort of 232 HIV-infected individuals. Clinical information was obtained using a detailed questionnaire and ascertained by physical examination. Participants were seen routinely every 3 months and when they were sick. A computer algorithm based on clinical history, examination and laboratory findings was used to stage HIV-positive participants at each routine visit. Kaplan-Meier survival estimates and the Cox proportional hazard model were used to assess the prognostic strength of the clinical and laboratory categories of the system. An attendance rate of 81% and 799 person-years of follow-up were achieved. Survival probability estimates at 6 years from being seen in clinical stages 1, 2, 3 and 4 were 63, 46, 24 and 6% respectively. When staging was revised to incorporate lymphocyte categories, the survival probabilities were 73, 62, 39 and 6% respectively. Unexplained prolonged fever and severe bacterial infection had survival probabilities closer to stage 2 conditions, mucocutaneous herpes simplex virus infection for more than 1 month and crytosporidiosis with diarrhoea for more than 1 month closer to stage 3 and oral candidiasis closer to stage 4 conditions. Even without the laboratory markers, the clinical category of the WHO staging system is useful for predicting survival in individuals with HIV disease. This is important for areas with limited access to laboratory markers. A simple rearrangement of a few clinical conditions could improve the prognostic significance of the WHO system.