Multi-site clinical evaluation of the Xpert® HIV-1 viral load assay

Abstract

BackgroundThe most important reason for measuring HIV-1 viral load (VL) is to monitor the effectiveness of antiretroviral therapy (ART), both for the initial therapeutic response and sustained responses. Maintaining low or undetectable HIV-1 VL levels can reduce both the risks of progression to AIDS and transmission of infection to others. ObjectivesTo evaluate the diagnostic accuracy of Xpert® HIV-1 Viral Load (VL) assay compared to the Abbott RealTime HIV-1 assay, including assessing specificity by testing plasma specimens from confirmed HIV-1 negative blood donors. Study designSubjects were enrolled from 4 participating sites, 2 in Europe and 2 in the USA. Fresh plasma samples were tested prospectively, while frozen plasma samples were collected prospectively, and tested retrospectively after selection of specimens to cover the assay’s quantification range (40cp/mL–10,000,000 cp/mL). Eligibility criteria included a clinician ordered HIV-1 VL test from a confirmed HIV-1 positive adult (≥18 years) with a known antiviral treatment status. Exclusion criteria included previous enrollment in this study or improper specimen collection. Human blood donor specimens determined to be HIV-1 negative by standard blood bank antibody and nucleic acid amplification methods were used to assess specificity. ResultsOf the 764 specimens collected, 752 were eligible for inclusion but 5 were not tested by the Xpert, leaving 747 specimens tested (28.2% from females and 71.8% from males). Valid results were obtained for 724/747 (96.9%) specimens tested using the Xpert HIV-1 VL assay. The Xpert HIV-1 VL assay detected or quantified 568/724 (78.5%) specimens, while the RealTime HIV-1 assay detected or quantified 559/724 (77.2%). Of the 724 specimens tested by both assays, 390 were quantified by both assays and showed strong correlation: r=0.9847, with an R2=0.9696. Bland-Altman analysis showed good agreement between the two assays (381/390; 97.7%) with a distribution within 0.5 log10 cp/mL centered around zero. Xpert yielded VLs for 393 (80%) of the 494 quantifiable samples by Abbott. VLs of those specimens quantified by one of the assays, and either detected but not quantified or not detected by the other assay were all <170cp/mL. Specificity of the Xpert assay was found to be 100% (109/109), 95% CI: 96.7–100.0. ConclusionVery good correlation was seen between the Xpert HIV-1 VL and Abbott RealTime HIV-1 assays, with added benefits for Xpert HIV-1 VL of: (1) lot-to-lot consistency traceable to WHO International Standard, (2) requiring both high and low level internal controls to be in range to have a valid result, (3) use of a single HIV-1 target for PCR and (4) faster turn-around-time for results, no need to wait to do batch testing of specimens. In summary, Xpert HIV-1 VL generated accurate VL results that if implemented could allow for actionable and timely treatment decisions during the same clinic visit. This scenario could reduce the loss to follow up often seen when these test results take days to weeks to become available to the clinician and patient.