Survival prognosis of HIV-infected patients.

Previous studies of survival after a diagnosis of acquired immunodeficiency syndrome (AIDS) have reported variation in temporal trends in association with age, gender, race, mode of transmission, lymphadenopathy, antiretroviral therapy, and presence of specific opportunistic infections at diagnosis. We used a logistic regression model to assess the effect of these factors while controlling for other potential predictors of time from initial CD4 cell count to death in 839 HIV-infected patients at a public hospital in Atlanta, Georgia. Our study found that a CD4 level of < 200 cells/microliters [odds ratio (OR) = 1.71; 95% confidence interval (CI) of 1.58, 1.85] and the presence of an AIDS-indicating condition (initial diagnosis OR = 2.50 and CI of 1.93, 3.24; diagnosis of a second AIDS condition OR = 3.02 and CI of 2.08, 4.40) are independently predictive of survival in HIV-infected persons. Furthermore, specific clinical manifestations of AIDS vary in their contribution to progression from diagnosis of AIDS to death. Therefore, changes in survival of AIDS patients must take into account changes over time in the relative frequency of specific AIDS-indicating diagnoses.