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


Changing from zidovudine to didanosine for HIV infection

ACP J Club. 1993 Jan-Feb;118:12. doi:10.7326/ACPJC-1993-118-1-012

Source Citation

Kahn JO, Lagakos SW, Richman DD, et al. A controlled trial comparing continued zidovudine with didanosine in human immunodeficiency virus infection. N Engl J Med. 1992 Aug 27;327:581-7.



To determine whether changing from zidovudine to didanosine reduces the risk for acquired immunodeficiency syndrome (AIDS)-related events or death in comparison with continued zidovudine treatment among persons with human immunodeficiency virus (HIV) infection.


2-year, randomized, double-blind, controlled trial of 3 regimens.


Multicenter American trial.


Eligibility criteria were HIV infection, tolerance of zidovudine for ≥16 weeks, CD4 count ≥ 60%. 913 persons (mainly male, white, and homosexual) were randomized; 30% had a previous AIDS event, 60% had AIDS-related complex, and 10% were asymptomatic. 93% received Pneumocystis carinii prophylaxis. 5% were lost to follow-up.


Patients were assigned to 750 mg/d or 500 mg/d of didanosine (2 sachets/d, equivalent to 400 mg in 4 tablets) or to 600 mg/d of zidovudine (100 mg every 4 h) and to placebo. Patients with serious symptoms could discontinue or reduce the study medication or change to the other drug with blinding maintained.

Main outcome measures

Progression to an AIDS-defining event (excluding Kaposi sarcoma and stage 1 AIDS dementia) or death for patients with AIDS-related complex or asymptomatic HIV infection, or a new, nonrecurrent AIDS-defining event or death for patients with AIDS. Events were confirmed by blinded review.

Main results

115 patients (37%) assigned to 750 mg of didanosine, 94 (32%) assigned to 500 mg of didanosine, and 125 (41%) assigned to zidovudine had a new, nonrecurrent AIDS-defining event or died (relative risk [RR], zidovudine vs. 750 mg of didanosine, 1.10, 95% CI 0.86 to 1.42; and RR, zidovudine vs. 500 mg of didanosine, 1.39, CI 1.06 to 1.82). Rates of withdrawal (57% overall) from zidovudine therapy were higher than from the 750-mg (P = 0.03) and 500-mg (P = 0.001) didanosine groups. Didanosine was associated with less severe hematologic toxicity, but the 750-mg dose was associated with more pancreatitis than was zidovudine. More improvements occurred in the number of CD4 cells and in p24 antigen levels in the didanosine groups as compared with the zidovudine group (P < 0.03). Effectiveness of didanosine was not related to duration of previous zidovudine therapy.


Changing treatment from zidovudine to 500 mg per day of didanosine slowed the rate of progression of HIV and AIDS-related events.

Source of funding: Not stated.

For article reprint: Dr. J.O. Kahn, AIDS Program, Ward 84, San Francisco General Hospital, 995 Potrero Avenue, San Francisco, CA 94110, USA. FAX 415-476-6953.


What is the best currently licensed anti-HIV drug? Zidovudine has the advantage of being first and remains the mainstay of treatment of HIV infection. Didanosine is, however, providing an alternative. The article by Kahn and colleagues shows that in patients who have been on zidovudine for at least 4 months, it may be better to switch to didanosine than to continue with zidovudine because new AIDS-defining events will be rarer and toxicity will be lesser.

What about patients who have not yet received antiviral treatment? At this time zidovudine remains the recommended drug because a direct comparison of zidovudine with didanosine in previously untreated patients has not been completed. We may speculate, however, that the difference in effectiveness between these 2 drugs as initial therapy will be small. The trials that are under way would otherwise have been terminated early after an interim analysis, and the results would already be known—this occurred in a comparison of dideoxycytidine with zidovudine when the inferiority of dideoxycytidine became apparent (Soo W. Personal communication).

What next? It is unlikely that further refinement of single-agent therapy will dramatically improve the prognosis of patients with HIV infection. The combination of dideoxycytidine (1) or didanosine and zidovudine does, however, increase CD4 counts more than single agents (Collier A. Personal communication). The next step is to translate these theoretical advantages into benefits that matter for patients: fewer opportunistic infections, improved quality of life, and longer survival. New drugs, those inhibiting protease in particular, are under investigation in small groups of patients. Their clinical value should be known within the next few years.

B. Hirschel, MD
Hôpital Cantonal Universitaire Geneva, Switzerland