Zidovudine therapy, CD4+ and CD8+ counts are associated with a longer survival following AIDS onset

In 60 patients followed from the onset of acquired immunodeficiency syndrome (AIDS) to death, survival was determined by Cox Proportional Hazards Analysis in relationship to seven variables: time-dependent CD4+ and CD8+ peripheral lymphocyte counts, zidovudine treatment, cytomegalovirus (CMV) retinitis, time from AIDS onset, calendar year of AIDS onset (cohort effect), and age. Two significant prognostic variables were identified: zidovudine therapy and either CD4+ or CD8+ counts (the latter could not be distinguished due to concomitant high correlation). Treatment with zidovudine reduced the death rate by 75% compared to no treatment. When included in a proportional hazards regression with all covariates except for the other T lymphocyte count, every increase in CD4+ count of 10 cells was equivalent to a decline in the mortality rate by 13% (p = 0.046), and every increase in CD8+ count of 10 cells lowered the mortality by 1.4% (p = 0.0031). Patients treated with zidovudine and without CMV retinitis showed the slowest decline of both CD4+ and CD8+ counts. Both CD4+ and CD8+ levels are useful predictors of survival in patients with AIDS.