THAB01: Monitoring of ART.

This was a high-yield and informative session at the cutting edge of ART monitoring research. It is highly recommended for viewers.

 THAB0101: CD4 and VL monitoring: research and development needs and policy implications, C. Gilks, Switzerland.

This was a broad overview of monitoring issues within the public health delivery of ART. Monitoring has multiple purposes: to improve ART outcomes, guide ART switching, detect toxicity, promote adherence, and prevent resistance. It must be locally accessible, and a cost-effective investment. He gave an overview of methods for identifying ART failures (clinical, CD4, or VL-based). He reviewed the topic of clinical vs. immunologic vs. virologic monitoring, highlighting data from DART, as well as several recent cost effectiveness modeling studies on whether it is “worth it” to add CD4, VL, or both types of monitoring to clinical care. He raised the idea of using VL as a “tie-breaker” to confirm ART failure, rather than a routine measurement. He lastly called for innovative point of care CD4 and VL tests, research to establish a threshold for virologic failure.

 THAB0102: Outcomes in ART treatment programmes with and without access to routine viral load monitoring: results from IeDEA Southern Africa, O. Keiser, Switzerland.

Keiser et al. presented a very large scale observational database analysis (from IeDEA) comparing mortality in 2 regions of southern Africa that do or do not have VL monitoring. Study was based on roughly n=80,000 patients from Malawi and Zimbabwe, and n=20,000 patients from South Africa. There was a 26% lower mortality where VL monitoring was used (S. Africa). Author argued that neither losses to follow-up, differences in baseline mortality, or patient characteristics account for the positive health impact of VL monitoring. She also presented data showing that sites that lack VL monitoring also lack the capacity to diagnose and treat many opportunistic infections, raising the possibility of possible unmeasured confounders of the relationship between VL monitoring and mortality.

 THAB0103: Dried blood spots for measurement of HIV-1 viral load (VL) in Argentina, M. Vasquez, Argentina.

Vasquez et al. presented a simple, elegant, and very informative study that compared DBS-based VL determination to a standard method (both methods performed in n=82 patients from a clinic in Argentina). The numbers of samples that showed VL>50 vs. VL<50 were similar between both methods, and the absolute VLs were strongly correlated by Spearman’s test and a Bland-Altman plot. Only 4 samples showed VL<50 by DBS that were >50 by standard VL (“false-negatives”); these all had a VL<3000 copies/cc. This study strongly supports DBS-based VL monitoring.

THAB0104: Evaluation of point-of-care CD4 and toxicity monitoring for resource limited ART clinic settings in Mozambique, I. Jani, Mozambique.

Jani et al. gave an excellent talk, starting by outlining a key monitoring problem: there are too many point of care technologies, many of which are inappropriate for resource limited settings. This study therefore assessed a list of 50 such technologies, and narrowed its focus to a short list of 18, based on feasibility, local availability, and price. These 18 systems were then quantitatively scored on 30 important criteria such as mobility, size, hand-held or not, heat/humidity stability, reagent costs, and others. The top-scoring CD4 system (by capillary blood) was the PIMA (by Alere), the top clinical chemistry system was the Reflotron (by Roche), and the top hemoglobin system was the HemoCue. These three then performed excellently in the field against gold standard methods, arguing that these may be valuable point of care tests.

 THAB0105: Determining acceptable properties for novel CD4 enumeration technologies from cross-sectional studies of candidate methods against flow cytometers, F. Noubar, USA.

Noubar et al. presented a statistically sophisticated yet very well-explained study asking the important question of how different frequencies of CD4 monitoring might affect the propensity to begin ART in naïve patients being observed over time. Data from the MACS cohort was used to model CD4 count over time via a large scale simulation of 50,000 CD4 trajectories, accounting for measurement error and intrinsic diurnal variation. The model predicts that with more frequent CD4 measurements, ART would get started slightly earlier and that bi-annual (every 6 months) measurements would be adequate (higher frequency not useful). Authors also state that novel technologies that have greater variability in CD4 measurements will lead to slightly earlier ART initiation times. A final key point made was that patients from all strata of CD4 should be recruited into studies evaluating novel measurement technologies.