Résumé :
|
[BDSP. Notice produite par INIST-CNRS F57tR0xi. Diffusion soumise à autorisation]. The effect of highly active antiretroviral therapy (HAART) on multiple acquired immunodeficiency syndrome (AIDS) - defining illnesses remains unclear. Between 1984 and 2005,573 male human immunodeficiency virus seroconverters in four US urban centers were followed for a median of 9.7 years. During follow-up, 345,113,50, and 65 men incurred 0,1,2, and>2 AIDS-defining illnesses, respectively. The authors extend the Cox proportional hazards model to determine whether the effect of HAART, as measured by calendar periods, persists beyond the first AIDS-defining illness. After adjustment for race and age at seroconversion, the hazards of a first through third AIDS-defining illness in the HAART calendar period (beyond July 1995) were 0.31 (95% confidence interval (Cl) : 0.21,0.46), 0.39 (95% Cl : 0.22,0.74), and 0.33 (95% Cl : 0.14,0.79), respectively, relative to the mono-therapy and combination therapy reference calendar period (January 1990-July 1995) and therefore did not attenuate with the number of prior AIDS-defining illnesses (p for homogeneity=0.83). After the authors averaged over multiple AIDS-defining illnesses, the hazard of an AIDS-defining illness in the HAART calendar period was 0.34 (95% Cl : 0.25,0.45) relative to the reference calendar period. HAART protects against initial and subsequent AIDS-defining illnesses, whose inclusion in analysis markedly increased the precision of the estimated hazard ratio.
|