The accurate identification of recent HIV infection remains an important area of research to inform population-based HIV prevention and treatment intervention policies. Methods that use cross-sectional testing and biomarker information might be an alternative to longitudinal testing, as combinations of serological and molecular methods can potentially provide a means to identify recent HIV infections. The aim of this study was to develop a predictive scoring system which was based on combining the results of Schupbach’s Line assay algorithm 15.1 and the viral load (VL) to predict the risk of having recently acquired HIV infection.
New HIV diagnoses in Ireland from January to April 2016 (n=151) were included in the study. All samples were tested on the Sedia limiting antigen avidity assay (LAg). The normalised cut-off in the LAg assay is 1.5. Schupbach’s (2015) Algorithm 15.1 was applied to the Line assay results. This algorithm is derived from antibody reaction scores to HIV antigen bands on the LIA strip. Patients on antiretroviral treatment or previously diagnosed (>12 months) were excluded from the analysis (n=43). The Speigelhalter-Knill-Jones method was used to develop a predictive scoring system which is based on the LAg assay as a gold standard as this assay demonstrated 100% accuracy in identifying recency based on samples tested from the CEPHIA repository. Viral loads above the median of the cohort were used as an indicator of recency.
Patient demographics showed that the recent cohort was predominantly male (87.1%) and MSM/bisexual (78.9%). The LAg assay and LIA Algorithm 15.1 identified 32 cases (21.2%) and 19 cases (12.6%) respectively as recently acquired HIV. The median viral load of the cohort was 16,428 copies/ml (log4.2). The results of the regression model including both the LIA and the HIV VL to predict recency are shown in Table 1. Combining both Algorithm 15.1 and VL demonstrated that the observed risk in our cohort of being recent is 100%. Using standard laboratory assays, this predictive scoring system allows an individualised assessment of risk for prediction of recency.
We have developed a predictive scoring system which is based on combining LIA Algorithm 15.1 and the VL. This proof of concept analysis has demonstrated that data such as LIA and VL can be used to develop a multiassay algorithm to accurately predict an individual’s risk of recently acquired HIV infection.