Monitoring Decisions Depend on Risk for Virologic Failure — and on Cost

Antiretroviral therapy (ART) programs in developing countries face a critical question: For any given amount of available funds, which strategy saves the most lives — monitoring patients already on treatment or providing ART to more patients? Answering that question requires a determination of the best approach to monitoring. Three recently published papers provide empirical data to educate responses.

In Nigeria, Rawizza and colleagues completed the largest study to date on the use of WHO-recommended CD4-based criteria for monitoring virologic outcomes. A total of 9690 patients underwent CD4-cell–count and viral-load monitoring every 6 months after starting ART. During a median follow-up of nearly 3 years, 2097 patients (22%) experienced virologic failure (defined as 2 consecutive viral loads >1000 copies/mL after ≥6 months on ART). CD4-based criteria were found to be poor predictors of virologic failure, with a sensitivity of 58%, specificity of 75%, and positive predictive value of 39%.

In rural Uganda, Mermin, Kahn, and colleagues compared three monitoring strategies among 1116 adults initiating ART through a home-based program. Participants were randomized to undergo clinical monitoring alone (CM arm), clinical monitoring plus quarterly CD4-cell–count assessments (CD4 arm), or clinical monitoring plus both quarterly viral-load and CD4-cell–count assessments (VL arm). During a median follow-up of 3 years, 126 patients died (48% in the first 3 months of treatment), 148 individuals experienced a new AIDS-defining condition, and 61 (6%) had virologic failure (defined as 2 consecutive viral loads >500 copies/mL after 6 months of treatment).

In an intent-to-treat, multivariate analysis, the rate of new AIDS-defining conditions or death was significantly higher in the CM arm than in the VL or CD4 arms (hazard ratios, 1.8 and 1.5, respectively), with no significant difference between the VL and CD4 arms. Similar results were obtained when the analysis was restricted to all-cause mortality. Interestingly, in a per-protocol analysis that was restricted to patients who received ART for >90 days (and thus excluded the period with highest mortality, when CD4 or viral-load monitoring would be unlikely to affect outcome), mortality rates were identical in the CD4 and VL arms. A cost-effectiveness analysis of these data indicated that adding CD4-cell counts to routine clinical monitoring is not only more cost-effective than including viral-load testing, but also more cost-effective than ART itself.