Cold chain and centrifugation requirements of plasma have limited the access of remote and vulnerable populations to HIV viral load (VL) testing. Roche Molecular Diagnostics developed a sample collection device, the plasma separation card (PSC) where plasma is obtained from the addition of whole blood to the PSC, eliminating the need for centrifugation and cold chain. We sought to independently evaluate the analytical and clinical performance of the PSC as an alternative matrix for HIV-1 VL across multiple Roche testing platforms for WHO prequalification and for use in President’s Emergency Plan for AIDS Relief (PEPFAR) countries.
Performance of PSC dried plasma spots were compared to centrifuged plasma samples using HIV negative whole blood spiked with cultured virus or using the third HIV-1 International WHO Standard and remnant clinical samples submitted for HIV VL testing. Prepared dried plasma spots were tested within 56 days using the COBAS® AmpliPrep/COBAS® TaqMan® (CAP/CTM) HIV-1 Test, v2, and the cobas® HIV-1 test on the cobas® 4800 (c4800) and 6800/8800 (c6800/8800) platforms. The limit of detection (LOD) was calculated using PROBIT analysis. The precision of measurement, cross-contamination, and linearity for subtypes A, B, C, D and CRF02-AG were also determined. Bland-Altman and correlation analysis were applied to evaluate clinical performance of the PSC on the c4800 and the c6800/8800.
The LOD for PSC samples was calculated to be 579.3, 745.2, and 489.4 copies/mL on the CAP/CTM, c4800 and c6800/8800 respectively. No cross-contamination was detected on any platform among the 40 samples tested, alternating between 20 high positive and 20 negative samples. Standard deviation between runs for all three platforms was within 0.14 log10 copies/mL. Linearity assessment of HIV-1 subtypes A, B, C, D and CRF02-AG showed R2 values greater than 0.98 for all three platforms. Invalid rates using the PSC were below 5%. When compared to centrifuged plasma, the PSC had a demonstrated sensitivity of 96% and 95% and a specificity of 94% and 97% on the c4800 and c6800/8800, respectively. The average bias between PSC and plasma was less than 0.2 log10 copies/mL for the c4800 and c6800/8800.
These findings reveal the PSC as a suitable alternative to plasma HIV-1 VL for monitoring patients on antiretroviral treatments in remote settings and in low-and middle-income countries. This PSC evaluation contributed to WHO prequalification and PEPFAR approval.