Sterne JA, Hernan M, Ledergerber B, Rickenbach M, Weber R, Tilling K, Egger M; International Conference on AIDS (15th : 2004 : Bangkok, Thailand).
Int Conf AIDS. 2004 Jul 11-16; 15: abstract no. C11843.
University of Bristol, Bristol, United Kingdom
Background: In countries where it is available, HAART has led to dramatic reductions of AIDS and death in HIV-1 infected individuals. However, in the absence of randomized trials (RCTs) comparing HAART with no therapy, the exact size of this effect remains unknown. Methods: We used data on 3,881 patients from the SHCS, who were not treated with HAART at registration. Follow up started after the later of 1 Jan 1996 or the date HIV-1 RNA was first measured, and ended on 1 Nov 2002. We used weighted Cox regression models (marginal structural models) to estimate the causal effect of HAART on the hazard of AIDS or death, allowing for "confounding by indication" (the fact that patients are treated because they are more immunosuppressed or show clinical signs of HIV infection, and such patients have worse prognosis). Results: In 15,768 patient-years of follow up, 500 patients progressed to AIDS or died. Patient characteristics such as lower CD4 count, decline in HIV-1 RNA, more advanced disease and previous treatment with monotherapy or dual therapy were associated with starting HAART. Overall, the hazard ratio (HR) for the effect of HAART on progression to AIDS or death was 0.23 (95% CI 0.15 to 0.35). The HR for HAART compared to dual therapy was 0.48 (0.30 to 0.79): consistent with the results of RCTs. In treatment naive patients the HR was 0.08 (0.04 to 0.16). This effect did not vary with gender, age or time period. In patients with transmission through intravenous drug use (IDU) the HR was 0.12 (0.04 to 0.31) compared to 0.06 (0.03 to 0.12) in non-IDU. The effect of HAART in treatment naive patients increased with time since starting therapy: HRs 0-6, 7-24 and >24 months after starting therapy were 0.17 (0.06 to 0.45), 0.12 (0.06 to 0.23) and 0.04 (0.02 to 0.10) respectively. Conclusions: Data from cohort studies can provide important information about the effects of HAART that is not available from RCTs. In treatment-naive patients HAART has reduced rates of progression to AIDS or death by between 84% and 96%.
Publication Types:
Keywords:
- Acquired Immunodeficiency Syndrome
- Antiretroviral Therapy, Highly Active
- CD4 Lymphocyte Count
- Cohort Studies
- Death
- Disease Progression
- HIV Infections
- HIV Seropositivity
- HIV-1
- Humans
- Proportional Hazards Models
- Substance Abuse, Intravenous
- drug therapy
- therapy
Other ID:
UI: 102277438
From Meeting Abstracts