Higher specificity of nucleic acid sequence-based amplification isothermal technology than of real-time PCR for quantification of HIV-1 RNA on dried blood spots

Abstract

Dried blood spots (DBS) are widely proposed as a plasma surrogate for monitoring antiretroviral treatment efficacy based on the HIV-1 RNA level (viral load [VL]) in resource-limited settings. Interfering coamplification of cell-associated HIV-1 DNA during reverse transcription (RT)-PCR can be avoided by using nucleic acid sequence-based amplification (NASBA) technology, which is based on an RNA template and isothermic conditions. We analyzed VL values obtained with DBS and plasma samples by comparing isothermic NASBA (NucliSENS EasyQ HIV-1 V2.0; bioMérieux) with real-time RT-PCR (Cobas TaqMan HIV-1 V2.0; Roche). Samples from 197 HIV-1-infected patients were tested (non-B subtypes in 51% of the cases). Nucleic acid extractions were performed by use of NucliSENS EasyMAG (bioMérieux) and Cobas AmpliPrep (Roche) before the NASBA and RT-PCR quantifications, respectively. Both quantification assays have lower limits of detection of 20 (1.3) and 800 (2.9) log10 copies/ml (log) in plasma and DBS, respectively. The mean (DBS minus plasma) differences were -0.39 and -0.46 log, respectively, for RT-PCR and NASBA. RT-PCR on DBS identified virological failure in 122 of 126 patients (sensitivity, 97%) and viral suppression in 58 of 70 patients (specificity, 83%), yielding 12 false-positive results (median, 3.2 log). NASBA on DBS identified virological failure in 85 of 96 patients (sensitivity, 89%) and viral suppression in 95 of 97 patients (specificity, 98%) and yielded 2 false-positive results (3.0 log for both). Both technologies detected HIV-1 RNA in DBS at a threshold of 800 copies/ml. This higher specificity of NASBA technology could avoid overestimation of poor compliance or the emergence of resistance when monitoring antiretroviral efficacy with the DBS method.

FIG 1

(a) Correlation curve for VL on plasma and DBS using RT-PCR. (b) Bland-Altman representation of VL in plasma and DBS using RT-PCR.

FIG 2

(a) Correlation curve for VL on plasma and DBS using NASBA. (b) Bland-Altman representation of VL in plasma and DBS using NASBA.