When testing is not possible

Several solutions have been forwarded to address the problem of CD4 cell count cost and lack of laboratory facilities and/or equipment in resource-limited areas.

Researchers working with a cohort of nearly 1200 patients from the University of Alabama at Birmingham observational database designed a decision tree based on local circumstances to see if it could more reliably identify patients with fewer than 200 cells/mm³ than did the surrogate of a total lymphocyte count (TLC) of 1200 cells/ml (the WHO-recommended cut point for initiating ARV therapy).

Variables included TLC, haemoglobin, platelet count, gender, body mass index, and ARV treatment use in the past month. A similar decision tree, developed using local circumstances, was also tested in Zambia with data from a cohort of nearly 600 female patients.¹

Using a decision tree algorithm to identify low CD4 cell counts was more effective than using the TLC cut point of 1200 cells/ml in any single population. Further, in Zambia, a custom-designed decision tree was more effective than one developed using data from a different population. The use of inexpensive, easily obtained variables were relevant, regardless of whether the patient was receiving antiretroviral treatment or not. The variables TLC, haemoglobin, and platelet count were significant.

The authors suggest that since different areas have their own local laboratory baseline reference values (resulting from genetic, environmental, infectious, or nutritional factors), additional studies should be done to see whether local models could be developed for different populations and their use considered alongside the recommended surrogate of a TLC limit of 1200 cells/ml.

The DART study looked at clinical outcomes in the absence of routine laboratory monitoring by randomising over 3000 antiretroviral-naive patients in Uganda and Zimbabwe to receive clinically driven monitoring alone or both laboratory and clinical monitoring. Patients were followed for nearly five years.

Investigators found that clinical monitoring alone was sufficient for the first two years of treatment and that CD4 count monitoring could be deferred until after that period. If patients on the clinical monitoring arm experienced a WHO stage 4 clinical event, their ART regimen was changed. After five years of treatment, survival in the clinical monitoring arm was 87 versus 90% in the laboratory and clinical monitoring arm.^{2 3}