ABSTRACT
While reporting human immunodeficiency virus (HIV) viral load (VL) using dried blood spot (DBS) in the BioMerieux NucliSENS platform, application of the hematocrit correction factor has been suggested. In this cross-sectional study from the National Microbiology Reference Laboratory of Zimbabwe, we assessed whether hematocrit correction (individual and/or mean) in DBS results improved the correlation with plasma VL and prediction of VL non-suppression (≥1000 copies per ml in plasma). Of 517 specimens during August-December 2018, 65(12.6%) had non-suppressed plasma VL results. The hematocrit correction factor ranged from 1.3 to 2.0 with a mean of 1.6, standard deviation (SD: 1.5, 1.7). The intraclass correlation (ICC) for mean (0.859, 95% CI: 0.834, 0.880) and individual (0.809, 95% CI: 0.777, 0.837) hematocrit corrected DBS results were not significantly different. The uncorrected DBS results had a significantly lower ICC (0.640, 95% CI: 0.586, 0.688) when compared to corrected DBS results. There were no significant differences in validity, predictive values, and areas under the receiver operating characteristics curves for all three DBS results when predicting VL non-suppression. To conclude, hematocrit correction of DBS VL results improved agreement with the plasma results but did not improve prediction of VL non-suppression. The results were not significantly different for individual and mean corrected results.