RATE OF CD4 CELL DECLINE AND PREDICTION OF SURVIVAL IN ZIDOVUDINE-TREATED PATIENTS

Objective: To investigate the relationship between survival and the ra te of CD4 cell decline before and in the first year following initiati on of zidovudine (ZDV) therapy. Design: Retrospective observational st udy. Setting: Hospital-based HIV clinics within the Riverside District Health Authority in London. Patients: Patients (total, 1415) with AID S (n = 476), symptomatic (n = 687) or asymptomatic (n = 194) HIV-1 inf ection, or of unknown clinical status (n = 58), who first received ZDV between June 1986 and October 1991. Intervention: The majority of pat ients received ZDV at an initial dose of 200 mg every 6 h or 250 mg tw ice daily. The median duration of follow-up after receipt of ZDV was 1 7 months (range, 2-54 months). Main measurements: CD4 cell counts prio r to and following initiation of ZDV; rate of decline of log-transform ed CD4 cell count before ZDV therapy and during the first year of ther apy; survival. Results: As of 31 December 1991, 432 patients had died. Patients with the highest rate of log CD4 decline before initiation o f ZDV (less-than-or-equal-to -0.06 log cells per month) as well as in the first year of ZDV therapy (less-than-or-equal-to -0.08 log cells p er month) had a much poorer 3-year survival from initiation of ZDV (23 and 40.5%, respectively) compared with patients with no decline or an increase in their CD4 count before (39.0%) or after (72.3%) ZDV thera py. In a series of multivariate analyses, a high rate of log CD4 decli ne in the first year of ZDV therapy (less-than-or-equal-to -0.08 log c ells per month) was predictive of poor survival, after adjustment for age and clinical status at initiation of ZDV and most recent CD4 count . In contrast, rate of CD4 decline before ZDV, presence of an initial CD4 rise and the magnitude of change in the rate of CD4 decline follow ing ZDV were no longer significantly associated with outcome. Conclusi ons: In this retrospective study, the rate of CD4 decline in the first year of ZDV therapy, but not the occurrence of an initial CD4 rise wa s predictive of survival, suggesting that the early CD4 response may b e a poor measure of the impact of ZDV. Patients with a high rate of CD 4 decline despite ZDV therapy represent a subgroup of patients with a poor prognosis who might benefit from alternative or combination antir etroviral therapies.