EFFECT OF GENDER, AGE, TRANSMISSION CATEGORY, AND ANTIRETROVIRAL THERAPY ON THE PROGRESSION OF HUMAN-IMMUNODEFICIENCY-VIRUS INFECTION USINGMULTISTATE MARKOV-MODELS

This article illustrates the use of time homogeneous Markov models wit h covariates to estimate the AIDS incubation period distribution from prevalent cohorts and to evaluate the effect of factors such as gender , age, human immunodeficiency virus (HN) transmission category, and an tiretroviral therapy on disease progression We applied this methodolog y to the analysis of data from a cohort of 3,027 patients enrolled fro m a hospital-based surveillance system of HIV infection in the Bordeau x University Hospital and four secondary public hospitals in southwest ern France. A total of 998 individuals (33%) progressed to AIDS during a median follow up period of 34 months. Based on a progressive three state Markov model, the estimated mean and median incubation periods w ere 9.1 years [95% confidence interval(CI) = 8.7-9.6] and 7.5 years (9 5% CI = 7.2-7.9), respectively. Our analyses showed a similar disease progression in men and women; we observed a more rapid progression for older subjects compared with younger ones and for homosexual men comp ared with heterosexuals, intravenous drug users, and transfusion recip ients, who had similar disease progression rates after adjusting for a ge. The use of antiretroviral therapy appeared to slow disease progres sion. Moreover, the results indicated that a combination therapy of zi dovudine with another antiretroviral drug may be more efficient than z idovudine monotherapy.