Current issues of ACP Journal Club are published in Annals of Internal Medicine


Zidovudine prolonged progression to AIDS or death

ACP J Club. 1995 Jan-Feb;122:18. doi:10.7326/ACPJC-1995-122-1-018

Source Citation

Volberding PA, Lagakos SW, Grimes JM, et al. The duration of zidovudine benefit in persons with asymptomatic HIV infection. Prolonged evaluation of protocol of the AIDS Clinical Trials Group. JAMA. 1994 Aug 10;272: 437-42.



To determine the duration of the clinical benefit of zidovudine in patients with asymptomatic human immunodeficiency virus (HIV) infection.


Analysis of follow-up data from a randomized, double-blind, placebo-controlled trial (Protocol 019 of the AIDS [acquired immunodeficiency syndrome] Clinical Trials Group). Follow-up was 0 to 4.5 years (mean, 2.6 y).


University-based and university-affiliated AIDS research clinics.


1565 patients with asymptomatic HIV infection and entry CD4+ cell counts < 500/mL were recruited between 1987 and 1989, and 1548 patients began treatment. 1030 (67%) patients continued treatment with open-label zidovudine in 1989 and 683 (66%) of those received open-label zidovudine until study end. 383 (24%) patients were unavailable for follow-up.


Patients were assigned to either placebo (n = 493); zidovudine, 500 mg/d (n = 542); or zidovudine, 1500 mg/d (n = 530). In 1989, because of a 2-fold reduction in clinical progression to AIDS in the zidovudine groups compared with the placebo group, all patients were offered open-label zidovudine, 500 mg/d.

Main Outcome Measures

Time to progression to AIDS or death.

Main Results

During the follow-up period, 232 patients progressed to AIDS or died. In the follow-up analyses of 1) all patients according to original treatment group; 2) all patients in the original placebo group followed only until open-label zidovudine was offered; and 3) patients initially assigned to receive placebo, the time to progression to AIDS or death was longer for patients receiving zidovudine compared with patients not receiving zidovudine (for the pooled zidovudine groups [500 mg/d and 1500 mg/d], P = 0.008, P = 0.004, and P < 0.001 for the 3 analyses, respectively). In all 3 analyses, the risk for progression to AIDS increased or tended to increase with duration of zidovudine use (for the pooled zidovudine groups, P = 0.005, 0.08, and 0.04 for the 3 analyses, respectively). For the analysis of all patients with entry CD4+ counts ≤ 300/µL, no treatment difference was noted (P = 0.50). For patients on zidovudine with entry CD4+ counts > 300/mL, the risk for progression to AIDS was decreased (P = 0.008). No significant differences existed in survival between zidovudine and placebo groups.


Zidovudine was effective in delaying the progression to the acquired immunodeficiency syndrome or to death in patients with asymptomatic human immunodeficiency virus infection. The treatment effect decreased over time.

Source of funding: Burroughs Wellcome.

For article reprint: Dr. P.A. Volberding, AIDS Program, San Francisco General Hospital, Ward 84, 995 Potrero Avenue, San Francisco, CA 94110. FAX 415-476-9233.


The study by Volberding and colleagues is of a size, duration, and design that will be difficult to duplicate for other drugs. The study is valuable because it was placebo controlled, included a dosing regimen that is commonly used today (zidovudine, 500 mg/d), used a clinical end point (time to progression to AIDS or death) as its main outcome, and had a sufficient duration of follow-up (up to 4.5 y) to yield more clinical end points than any other North American-based clinical trial to date.

Useful messages from this study that can be applied to physician practices include the following: 1) Monotherapy with zidovudine is effective in delaying progression to AIDS and death in patients with asymptomatic HIV infection; 2) the duration of benefit of zidovudine monotherapy in delaying progression is time-limited (about 2 years in this study); and 3) the durability of the response is better in patients with higher rather than lower CD4+ cell counts. When death was analyzed as a single end point, no difference existed between groups.

If only zidovudine was available and its effect was time-limited, decisions on whether to start therapy early or late in the course of disease would depend on patient preferences, drug side effects, and cost. Today, 4 approved antiretroviral drugs are available. Choices are broader (sequential monotherapy with different drugs, combination therapy), and decisions are more difficult (When should a switch be made?, What surrogate marker can best predict when to switch?, and What is the basis for drug failure?).

Until ongoing trials answer these pressing questions, this study can give our patients an indication of the durability of the zidovudine effect and an idea of the effect of CD4+ counts on that durability.

Donald H. Batts, MD
Upjohn Laboratories Kalamazoo, Michigan