Incidence and Predictors of Antiretroviral Treatment Modification in HIV-Infected Adults: A Brazilian Historical Cohort from 2001 to 2010.

This study estimated the incidence of and time to first antiretroviral therapy (ART) modification. This longitudinal analysis comprised a sample of 236 patients from three HIV/AIDS referral centers in Belo Horizonte, Brazil-part of a major historical cohort. Inclusion criteria were as follows: having been treatment-naive patient ≥18 years old who initiated ART between 2001 and 2005 in these three referral centers. The main endpoint was time to first ART modification. Patients were followed up for five years, covering the period 2001-2010, during which time Pearson's chi-square test was performed to compare ART modification between groups. Kaplan-Meier inverse survival curves were employed to describe the probability of ART modification and Cox proportional hazard regression was used to estimate the adjusted hazard ratio (aHR) of ART modification. Among 247 patients from the major cohort, 236 were eligible. Median follow-up time was 37.2 months and the contribution in person-months was 7,615.4 months. A total of 108 (45.8%) patients had their ART regimen modified at least once (incidence rate: 1.42 per 100 person-months). Adverse drug reactions were the main reason for ART modification. Women (aHR = 1.62; = 0.022) and patients on protease inhibitor- (PI-) based regimens (aHR = 2.70; < 0.001) were at higher risk of ART modification.