Evaluation of the performance of a novel HIV‐1 viral load assay for HIV quantification in China

• Published in: HIV medicine 2023-07, Vol.24 (7), p.777-784
• Main Authors: Deng, Xizi, Li, Liya, Cai, Xiaoli, Lin, Yaqing, Lan, Yun
• Format: Article
• Language: English
• Subjects: Acetic acid; Agreements; Assaying; China; Confidence limits; Correlation coefficient; Correlation coefficients; Edetic acid; Hepatitis; Hepatitis B; Hepatitis C; HIV; HIV Infections - diagnosis; HIV Seropositivity; HIV-1 - genetics; HIV‐1 viral load assay; Human immunodeficiency virus; Humans; Linearity; Performance evaluation; Qualitative analysis; quantification; Reagents; RNA, Viral; Sensitivity and Specificity; Viral Load; Viruses
• ISSN: 1464-2662; 1468-1293; 1468-1293
• DOI: 10.1111/hiv.13473

Objectives Our objective was to assess the HIV‐1 quantification performance of the Livzon HIV‐1 viral load (VL) assay and the Roche Cobas HIV‐1 assay to evaluate an HIV‐1 VL testing reagent for application in China. Method We compared the Livzon and Roche Cobas HIV‐1 VL assays using ethylenediaminetetraacetic acid plasma samples collected between May 2021 and November 2021 from patients with HIV‐1 and healthy controls. We used Cohen's κ coefficient to measure agreement of qualitative values and Pearson's correlation coefficient (r) values and the coefficient of determination (R2) to determine the linear relationship between the two assays. We performed a Bland–Altman analysis to assess VL quantification agreement. Results In total, 11 plasma samples from patients with hepatitis B virus (HBV) or hepatitis C virus (HCV) and nine samples from healthy controls were undetectable on both assays. Overall agreement was seen in 419 of 500 specimens (91.40%), with a κ value of 0.59. Pearson's correlation coefficient between the two assays was 0.970. Using the Bland–Altman method, 95.14% (352/370) of paired VLs fell within the 95% confidence limits of agreement (−0.51 to 0.95 log10 copies/mL). Higher VLs had a better correlation and a smaller mean difference between the two assays. Pearson's correlation coefficient for the samples of subtype CRF01_AE, CRF07_BC, and CRF55_01B was 0.950, 0.935, and 0.952, respectively. Conclusion The Livzon HIV‐1 VL assay exhibits good precision and linearity and a high correlation with the Roche Cobas HIV‐1 assay. The Livzon HIV‐1 VL assay has salient advantages in terms of the lyophilized powder reagent, which gives the assay greater stability and sensitivity and can be readily used in low‐resource areas.