2006

Abstract:

Drug resistance is a major obstacle to the successful treatment of human immunodeficiency virus type 1 (HIV-1) infection. We have been using quantitative predictions of drug resistance to model the success of drug therapies in vivo. To estimate drug resistance levels, we used standard stepwise linear regression to construct drug resistance models for 7 protease inhibitors and 10 reverse transcriptase inhibitors using data obtained from the Stanford HIV drug resistance database.We evaluated these models by hold-one-out experiments and by tests on an independent dataset.

Interestingly, this simple model outperformed other publicly available genotypic interpretation algorithms, including decision tree, support vector machine, and four rules-based algorithms (HIVdb, VGI, ANRS and Rega) under both tests. To study how drug resistance influences therapy, we created a mathematical model for viral competition that includes CD4+ target cells, free virions, and different types of infected cells. In cases where drug resistance imposes a fitness cost, these models show that the trajectory of viral load change at the time therapy is initiated can be an important predictor of whether therapy will be successful: initiating treatment when viral load is decreasing can often reduce the risk of selecting for drug resistant mutants relative to treatments initiated when viral load is stable or increasing. These findings suggest new strategies for optimizing therapy regimens in salvage therapy patients and other patients with limited therapy options.