Objectives:

To evaluate the association between time since initiation of pre-AIDS antiretroviral therapy [mainly with zidovudine (ZDV)] and AIDS-free survival in a cohort of HIV seroconverters, and to assess possible differences in this association and in the use of antiretroviral therapy by HIV exposure group.

Design:

Observational study of HIV-infected individuals, both those treated with antiretroviral therapy and those untreated, enrolled in an ongoing prospective cohort (median follow-up, 5.3 years).

Setting:

Sixteen HIV outpatient clinics throughout Italy.

Patients:

A total of 1078 individuals infected with HIV through injecting drug use or homo-/heterosexual activity, and with accurately estimated dates of seroconversion.

Main outcome measures and methods:

Kaplan–Meier estimates of the probability of receiving antiretroviral therapy before AIDS. Crude and adjusted relative hazards of AIDS and of death from AIDS using Cox regression models.

Results:

The cumulative incidence of beginning pre-AIDS antiretroviral therapy within 7 years of seroconversion was 49.2%. Injecting drug users (IDU) were less likely to undergo antiretroviral treatment before AIDS than homosexual men and heterosexual contacts. The adjusted relative hazard of developing AIDS for patients treated with ZDV (relative hazard adjusted for occurrence of acute HIV disease, pre-AIDS HIV-related diseases, CD4 count, and use of prophylaxis for Pneumocystis carinii pneumonia) was 0.57 within the first year of starting zidovudine and 0.92 after 1 year of therapy. Stratifying by HIV exposure category, the adjusted relative hazards of AIDS for individuals who started ZDV less and more than 1 year before AIDS were 0.74 and 0.99 among IDU, 0.31 and 0.89 among homosexual men, and 0.69 and 0.72 among heterosexuals, respectively. Similar results were obtained when using death from AIDS as an endpoint.

Conclusions:

IDU began pre-AIDS antiretroviral therapy significantly later than homosexual men and heterosexuals, even after adjusting for CD4 count. Results from this non-randomized study confirm that antiretroviral treatment has only a short-term clinical benefit. There was a stronger association between antiretroviral treatment and lower risk of AIDS in homosexual men than in IDU.