Point of care technology: expediting diagnosis to improve clinical outcomes for HIV-infected children

Abstract

Only half the HIV-infected children in South Africa access antiretroviral treatment (ART), despite it being available for over 15 years. Translating infant diagnosis into early treatment and retention in care remains a key problem. This project aims to evaluate the new strategy of birth HIV polymerase chain reaction (PCR) testing in the context of point-of-care (POC) technology, thus enabling early treatment and linkage to care in the days after birth. The first paper of the thesis describes the advent of birth HIV PCR testing in a large maternity/delivery unit in Johannesburg, South Africa. High coverage (93%) and excellent uptake (>99%) were achieved in implementing the Roche COBAS® TaqMan® (Version 2.0) HIV-1 Qualitative central laboratory-based test (LABT) between June 2014 and December 2016. Overall, maternal HIV prevalence was 23%, and 1.6% of tested infants were found to be HIV-infected at birth. Of these, 16% were born to women newly diagnosed at delivery. The second paper evaluated the concurrent implementation of the Cepheid Xpert HIV-1 Qualitative POC test (POCT), run together with the LABT when staff capacity allowed. While only achieving 52% coverage, the POCT performed at sensitivity of 100% (95% CI 88·4–100) and specificity of 99·9% (95% CI 99·7–100) with a Cohen’s κ coefficient of 0·967 (95% CI 0·922–1·00). It enabled a higher proportion of result return (p<0.0001) than LABT alone (96% vs. 53%) in all infants and earlier initiation of ART (1 day of age vs. 6 days of age, p=0.0001) in infected infants. The third paper (diagnostic challenges) showed that 24% of initial non-negative birth tests (i.e. positive or indeterminate) were indeterminate, and in 5% of cases a significant diagnostic uncertainty remained despite a protracted course of testing. The final paper outlines the one-year outcomes of confirmed HIV-infected infants diagnosed at birth. The diagnosis enabled 98% of infants to commence ART. The 12-month mortality was 0·14 (95% CI 0·08-0·24) and retention in care of those initiated on site, excluding deaths, was 78%, with 71% achieving viral suppression to <400 copies/ml. Although diagnostic dilemmas are a key challenge and keeping the cohort in care requires significant effort, this study nevertheless demonstrates that diagnosis at birth serves as an important avenue to improve ART coverage in children. In conclusion, birth HIV testing together with POC testing, if implemented strategically, can effectively increase access to ART for HIV-infected infants.