Management of antiretroviral therapy for HIV infection: Modelling when to change therapy

Objective: To evaluate four strategies for monitoring plasma HIV RNA levels and/or resistance genotypes to decide when to change antiretroviral therapy. The strategies include: (i) 1997 guidelines recommending a therapy switch when plasma RNA exceeds a threshold level; (ii) a viral load policy, using a fixed increase in viral load as the trigger; (iii) a genotype policy, requiring a smaller viral rebound than (ii) and detection of genotypic resistance before switching; and (iv) a proactive policy, switching drug regimens at a predetermined time if viral load has not rebounded. Design and setting: A Monte Carlo simulation tracks patients' viral loads and presence of opportunistic infection during therapy. The model uses clinical and virological data and statistical variation in patient parameters for the evaluation of therapeutic strategies. Main outcome measures: To determine which strategies minimize viral rebound detection delay while maintaining a low (prespecified) probability of switching therapy before rebound. Results: 1997 Guidelines and the viral load policy create lengthy delays in detection of rebound, particularly when patients are drug-naive and the detection limit of the viral load assay is 500 copies/ml. A detection limit of 20 copies/ml decreases this delay substantially. Genotyping achieves only minor additional delay reductions. Of the strategies tested, the proactive policy leads to the shortest delays. Conclusions: This model indicates that prolonged periods may be required for viral load to rebound to detectable levels following prolonged suppression. Proactive switching produces the best outcome in our model because it may reduce the duration of viral replication under pressure of a failing regimen before detection of viral rebound. This strategy should be evaluated in clinical trials.