Objective: To assess the long-term effectiveness of zidovudine (AZT) in patients with acquired immunodeficiency syndrome (AIDS). This assessment has never been adequately done because controlled clinical trials were stopped early and survival comparisons were made with historical controls.
Design: Nonrandomized contemporary observational study of patients treated and not treated with zidovudine.
Setting: Twenty-three AIDS treatment centers throughout Italy that reported cases to the National Registry of AIDS Cases between July 1987 and March 1988.
Patients: One hundred fifty-nine zidovudine-treated and 112 untreated patients with AIDS, the majority of whom had acquired human immunodeficiency virus (HIV) infection through intravenous drug use.
Outcome measures: Median survival and 1- and 2-year survival for treated and untreated groups, as estimated using the Kaplan-Meier method. Cox proportional hazards regression analysis was also used to identify independent predictors of survival among the variables studied.
Results: Patients were similar with respect to CD4/CD8 ratio, age, sex, clinical and immunological status at diagnosis, and source of HIV infection. After 24 months, survival was 45.9% (95% confidence interval [CI], 36.1% to 55.7%) in the treated group and 20.5% (95% CI, 12.6% to 28.3%) in the untreated group, with median survival of 21.2 and 9.6 months, respectively.
Conclusions: Possible biases of this study include imperfect matching for clinical status and better overall medical care of treated patients. Nevertheless, we believe that the observed differences in survival were primarily due to zidovudine treatment.