LONG-TERM SURVIVAL IN PATIENTS WITH ADVANCED IMMUNODEFICIENCY

Objective: To identify prognostic factors associated with survival tim e in HIV-infected patients with advanced immunodeficiency. Design: Pro spective cohort study. Participants: A total of 1284 HIV-infected pati ents with serial CD4 count measurements and at least one CD4 cell coun t less than or equal to 50 x 10(6)/l (CD4 less than or equal to 50). M ain outcome measure: Survival from initial CD4 cell count less than or equal to 50 x 10(6)/l. Results: The median survival from initial CD4 less than or equal to 50 x 10(6)/l was 17.1 months. The risk of death increased by 2% [95% confidence interval (CI), 1-3] for each year of a ge, by 10% (95% Cl, 3-16) for each 10 x 10(6)/l decrease in CD4 count, and by 14% (95% Cl, 9-18) for each 1 g/dl decrease in haemoglobin lev el. Compared to AIDS-free patients with CD4 less than or equal to 50 x 10(6) cells/l, the risk of dying was 1.5 fold (95% Cl, 1.2-1.9) that of patients who had an AIDS diagnosis for fewer than 3 months prior to CD4 less than or equal to 50, 1.8-fold for patients with an AIDS diag nosis for 4-11 months prior to CD4 less than or equal to 50, and twice that of patients with AIDS for greater than or equal to 12 months pri or to CD4 less than or equal to 50. The risk of dying for patients who se rate of CD4 cell decline was > 40 x 10(6)/l per 6 months was 1.7-fo ld (95% Cl, 1.3-2.3) that of patients with an average CD4 cell loss < 40 x 10(6)/l per 6 months, after adjusting for age, haemoglobin and du ration of AIDS prior to CD4 less than or equal to 50 x 10(6) cells/l. A prognostic score was developed from the final multivariate model, ba sed on age at CD4 less than or equal to 50, haemoglobin at CD4 less th an or equal to 50, duration of AIDS and rate of CD4 decline prior to C D4 less than or equal to 50. Conclusions: Routinely available clinical and laboratory data including haemoglobin level, rate of CD4 decline and duration of AIDS can be readily translated into a prognostic score and then used to predict the survival experience of an HIV-infected p atient with advanced immunodeficiency.