Early infant HIV diagnosis and characterization of HIV drug resistance in Gauteng, South Africa

Abstract

Despite the high prevalence of the human immunodeficiency virus (HIV) in South African women of reproductive age, the South African (SA) Prevention of Mother-to-Child Transmission (PMTCT) programme has significantly reduced the incidence of new HIV infections in infants from >20% in 2004 to <2% and overall MTCT approximately >5%. The PMTCT programme, however, faces challenges in terms of early infant diagnosis (EID) because of HIV polymerase chain reaction (PCR) indeterminate results as well as HIV drug resistance (HIVDR) secondary to antiretroviral therapy (ART) exposure of mothers and infants. The National Health and Laboratory Services (NHLS) uses the Roche COBAS®AmpliPrep (CAP)/COBAS®TaqMan® (CTM) HIV-1 Qualitative Test (Roche Molecular Systems, Pleasanton, CA) (CAP/CTM) platform as part of EID. Recently, the Roche cobas® 6800/8800 System has been introduced to test HIV viral load and HIV DNA PCR for EID. The platform is already processing samples for HIV viral load; however, verification for HIV DNA PCR for EID with dried blood samples (DBS) is needed, especially for CAP/CTM HIV PCR indeterminate results (Cycle threshold [Ct]>33 with any relative fluorescent intensity [RFI] value or Ct≤33 and RFI <5 on the CAP/CTM, Ct>38 on the Roche cobas® 6800/8800 System). In addition, HIVDR in newly diagnosed infants significantly limits treatment options. Therefore, the current study verified the Roche cobas® 6800/8800 System against the CAP/CTM system for the detection of HIV in EID and determined the HIVDR prevalence and profiles in infants <6 months, and how this affects current SA PMTCT and EID guidelines. The study comprised 642 DBS samples (235 HIV PCR positive, 193 HIV PCR negative and 214 HIV PCR indeterminate) previously tested on the CAP/CTM assay. Overall, 99.6% (234/235) CAP/CTM HIV PCR positive samples remained positive, while 99.5% (192/193) HIV PCR negative samples remained negative with the Roche cobas® 6800/8800 System. The HIV PCR indeterminate results as detected by the CAP/CTM decreased from 100% (214/214) to 8.4% (18/214) with the Roche cobas® 6800/8800 System. The Roche cobas® 6800/8800 System had a specificity of 99.5% and a sensitivity of 99.6%, but this decreased to 96.3% and 90.8% when HIV PCR indeterminate results were included. The kappa value increased from 0.5, which signifies moderate agreement, to 0.9, which is excellent agreement, when RFI from the CAP/CTM was excluded for result determination. The overall agreement between the two assays, taking only cycle threshold values into account, was 93.8%. As for HIVDR, mutations were detected in 42.9% (24/56) of infants <6 months. The most common non-nucleoside reverse transcriptase inhibitor (NNRTI) mutation causing high-level resistance was K103N (21.4% [12/56]), followed by Y181C and the NRTI mutation, M184V, both in 8.9% (5/56) of infants. Also, major protease inhibitor (PI) mutations, M46L and V82A were detected in one case each (1.8%). In conclusion, the performance of the Roche cobas®6800/8800 System was comparable to the CAP/CTM; however, it detected fewer HIV PCR indeterminate results, thus potentially offering conclusive results in a larger proportion of infants. The detection of high levels of the NNRTI mutation, K103N, emphasises the need for constant surveillance since nevirapine is included as part of the SA PMTCT guidelines and the World Health Organization recommends that NNRTIs should be phased-out of as part of PMTCT once the resistance prevalence exceeds 10%.