Predictors and consequences of antiretroviral resistance among patients with late stage human immunodeficiency virus (HIV -1) disease in supervised treatment settings

Recent evidence suggests that positive treatment outcomes for patients with Human Immunodeficiency Virus-1 (HIV-1) disease may be compromised by selection of HIV-1 drug resistant mutants (HIVDRM). HIVDRMs have been associated with virologic failure for patients on highly active antiretroviral therapy (HAART), but this relationship is still nebulous. Whether HIVDRMs predict immunologic or clinical outcomes remains largely unexplored. Conversely, poor adherence is a major predictor of HIVDRM, but data from several studies suggests that mutations also appear among highly adherent patients. Among other factors, the potency and composition of a HAART regimen is thought to be both a cause and a consequence of drug resistance. The Genotypic sensitivity score (GSS), a measure of regimen potency, has been shown to predict treatment response in the context of HIVDRM, but the observational studies exploring this relationship did not control for adherence or for other possible confounders. Thus, understanding the causes and effects of HIVDRMs is provocative. The aims of this dissertation study are to: (1) identify independent predictors of virologic, immunologic and clinical antiretroviral (ARV) responses (e.g. GSS, sociodemographic markers, ARV exposure, HCV); (2) identify independent predictors of antiretroviral (ARV) resistance and viral evolution-weighted and unweighted measurement of amino acid changes in the reverse transcriptase and protease gene; and (3) systematically determine the key mutations contributing to treatment failure, and to viral evolution among treatment failures, among patients with late stage HIV disease in a supervised treatment setting. We hypothesized that higher GSS would predict more viral suppression, higher CD4 T cell count increase, lower occurrence of category c opportunistic conditions and mortality. We hypothesized further that among patients experiencing treatment failure, higher GSS would be associated with higher viral evolution rate, and that additional information on baseline viral diversity would explain more of the variability in treatment failures, and in viral evolution rate among treatment failures.;In the first study, we tabulated the GSS of the resistance testing informed regimen using the genotypic data from the Virtual phenotype report, based on the IAS-USA consensus guidelines, and examined whether GSS was predictive of viral suppression (HIV-1 RNA copies < 50 copies/ml), mean change in CD4 T-lymphocyte counts, occurrence of category C opportunistic conditions, and death. In the second study, GSS was explored as a predictor of resistance at the protein level - mean number of acquired amino acid substitutions, amino acid substitution rate, and evolutionary distance - PAM 250 score of amino acid substitutions with weighted changes from an evolutionary standpoint. We also investigated whether baseline measurements of drug resistance using full sequence data rather then selected mutations significantly improve prediction of HIVDRMs and viral evolution.;The results of this dissertation suggest: (1) higher GSS significantly increased the likelihood of viral suppression after controlling for therapy adherence levels and other covariates; (2) higher GSS was also associated to higher mean increase in CD4 T-lymphocyte counts, although this did not reach formal statistical significance; (3) higher GSS was not predictive of category C opportunistic conditions and death; (4) higher GSS was not significantly associated with a higher HIV-1 mutation rate among treatment failures; and (5) total number of baseline mutations and PAM 250 scored at baseline were predictive of PAM 250 score difference and viral mutation rate respectively. Baseline PAM 250 scores explained significantly more variability in the mean number of mutations and PAM 250 score differences. The implications of the findings and directions for future research are discussed.