Advisory Committee Chair
Donna K Arnett
Advisory Committee Members
Michael S Saag
Jeffrey S A Stringer
Sten H Vermund
Craig M Wilson
Document Type
Dissertation
Date of Award
2008
Degree Name by School
Doctor of Philosophy (PhD) School of Public Health
Abstract
This dissertation had three central goals. The first was to review the existing literature for reports of virologic failure among adults initiating non-nucleoside reverse transcriptase inhibitor-based antiretroviral therapy in resource-limited settings. The second was to describe the epidemiology of virologic failure over the first 6 months of therapy in Lusaka, Zambia. The third was to examine the performance of different approaches for diagnosing virologic failure. We performed a meta-analysis using data from previously published reports to create summary estimates for the prevalence of virologic failure and found that the lowest estimate of the 95% confidence interval was 10% or higher at all time points evaluated between 3 months and 2 years. We then examined data from one arm of a randomized trial of treatment monitoring strategies in Lusaka, Zambia and found the prevalence of virologic failure to be 6.7% (95% confidence interval, 4.9%-9.0%) at 6 months. We also found that poor adherence as determined from pharmacy attendance was associated with virologic failure (adjusted relative risk, 2.7; 95% confidence interval, 1.2-6.3). We created an algorithm to predict virologic failure using regression modeling techniques. We used adherence and baseline anemia to estimate the probability of failure and classified any patient with an estimated probability > 9% as a potential failure. The resulting predictive-score-based algorithm had a sensitivity of 24.4% (95% CI, 12.4%- 40.3%) and a specificity of 91.7% (95% CI, 89.2%-93.8%). We algorithm recommended iii by the World Health Organization had a sensitivity of 27.6% (95% CI, 12.7%-47.2%) and a specificity of 78.4% (95% CI, 74.4%-82.0%). The results of this dissertation research indicate that antiretroviral treatment failure is a significant concern as 10% of patients have circulating viremia at any given time point between 3 months and 2 years. Furthermore, we found that only adherence was significantly associated with virologic failure and were unable to produce an algorithm for the diagnosis of virologic failure that performed well at 6 months. Although cost and technical requirements prevent routine virologic monitoring from being widely used in resource-limited settings, these results argue that efforts should be made to address this problem.
Recommended Citation
Cantrell, Ronald Alexander, "Diagnosing Antiretroviral Treatment Failure In Resource-Limited Settings" (2008). All ETDs from UAB. 3676.
https://digitalcommons.library.uab.edu/etd-collection/3676