HIV viral load (VL) testing is the recommended method for monitoring the response of people living with HIV receiving antiretroviral therapy (ART). Availability of and access to standard plasma VL testing in low and middle-income countries (LMICs) is limited by the need to use fresh plasma. Good performing specimen collection methods for HIV VL, which are applicable to resource-constrained settings, are needed. We assessed the diagnostic performance of filtered dried plasma spot (FDPS), created using the newly developed, instrument-free VLPlasma® device, to identify treatment failure at a VL threshold of 1,000 copies/ml in fresh plasma. Performance was compared with the conventional dried blood spot (DBS). Venous blood samples from 201 people living with HIV, attending an infectious disease clinic in Malaysia, were collected and HIV VL was quantified using fresh plasma (reference standard), FDPS and DBS specimens. VL testing was done using the Roche Cobas AmpliPrep/Cobas TaqMan v2.0 assay. At 1,000 copies/ml threshold, the diagnostic performance of FDPS was superior (sensitivity=100% (95% CI=89.1-100%); specificity=100% (95% CI=97.8-100%) compared with DBS (sensitivity=100% (95% CI=89.4-100%); specificity=36.8% (95% CI=29.4-44.7%; p<0.001). A stronger correlation between FDPS VL and plasma VL was observed (r=0.94, p<0.001), compared with that of DBS VL and plasma VL (r=0.85, p<0.001). The mean difference in VL measures between FDPS and plasma (plasma VL - FDPS VL) was 0.127 log10 copies/ml (SD=0.32) compared to -0.95 log10 copies/ml (SD=0.84) for DBS. HIV VL measured using FDPS outperformed DBS in identifying treatment failure at 1,000 copies/ml threshold and compared well with plasma quantification of VL. FDPS can be an attractive alternative to fresh plasma for improving access to HIV VL monitoring among people living with HIV on ART in LMICs.

The authors would like to thank the patients, service providers and laboratory technicians at the Medical centre University of Malaya, Malaysia who made this study possible. The authors gratefully acknowledge the contribution to this work of the Victorian Operational Infrastructure Support Program received by the Burnet Institute. Support was provided by the National Health and Medical Research Council of Australia (NHMRC) through Project grant GNT1063725, and Career Development Fellowship to S Luchters.