Abstract :
Background
Highly active antiretroviral therapy (HAART) was introduced in 1997. We aimed to assess the continuing effect of this treatment on survival and progression to AIDS after HIV-1 seroconversion.
Methods
We used Cox models to estimate the effect of calendar year on time to AIDS and death in 22 cohorts of people from Europe, Australia, and Canada who had seroconverted. Retrospective and prospective data were used. We compared the effects of age at seroconversion, exposure category, sex, and presentation during acute HIV-1 infection pre-1997 (pre-HAART), in 1997-98 (limited use of HAART), and 1999-2001 (widespread use of HAART).
Findings
Of 7740 seroconverters, 2000 (26%) had died. Compared with pre-1997 data, the hazard ratio (HR) for death fell sharply to 0•47 [95% CI 0•39–0•56] in 1997, dropping further to 0•16 [0•12–0•22] in 2001. Correspondingly, the proportion of person-time on HAART increased from 22% in 1997 to 57% in 2001. By contrast with the pre-HAART era, injecting drug users had significantly higher mortality in 1999–2001 than did men infected through sex with men (HR 4•28 [2•86–6•41]). However, whereas pre-1997 the risk of AIDS was higher in those aged 45 years or older at seroconversion than in people who were 16–24 years (2•03 [1•67–2•47]), in 1999–2001 there was little evidence of a difference in risk by age (HR=1•17 [0•60–2•30]; interaction p=0•06). No such attenuation in the effect of age on survival was observed (p=0•63).
Interpretation
Predicted survival for people with HIV-1 has continued to increase, since the introduction of HAART; however, the importance of age and exposure category as determinants of progression seems to have changed.
