Response to Zhao and Zhou: Diagnosis of HIV infection in breastfed infants of mothers on antiretroviral therapy.

Abstract

We thank Zhao and Zhou [1] for their thoughtful comments on our report of two infants born to women with HIV who had unusual trajectories of diagnostic tests. The two HIV-exposed infants were initially positive on nucleic acid amplification tests in the first month of life but had multiple negative tests during breastfeeding and presented with definitive evidence of HIV infection around the time of breastfeeding cessation. We hypothesize that constituents of breast milk, or processes induced by breastfeeding, most likely in combination with antiviral effects of maternal antiretroviral therapy, led to the undetectable test results that we observed in already infected infants until breastfeeding ended. We considered the alternative explanation of postnatal transmission proposed by Zhao and Zhou [1], but this explanation requires dismissing the initial positive test results on the early samples as transient infections or false-positive results. A rigorous investigation of 42 cases of suspected transient infection, failed to confirm a single one, despite the plausibility of this concept [2]. In one of the infants in our report, virus was detected in peripheral blood mononuclear cells (PBMCs) from three separate blood samples collected over the first month of life making laboratory errors an unlikely explanation. In this same case, CD4+ T cells from one of these early age time points were found to be positive for 2-LTR episomal HIV-1 DNA. Detection of 2-LTR episomal DNA in CD4+ T cells, although at low levels, provides evidence of viral replication as these episomal circles are a by-product thereof [3]. We acknowledge that early samples were not available for the second case making the alternative explanation of an initial false-positive result followed by postnatal infection a possible explanation of this second case. It should be noted that current birth testing and early treatment guidelines in South Africa would provide indefinite treatment to transient infections should they occur [4]. We would like to clarify the implications of our observations in the context of current guidance around antiretroviral use for women and their children to prevent vertical transmission of HIV. Maintenance of maternal therapeutic antiretroviral drug regimens over the full duration of breastfeeding and beyond is now considered standard of care for prevention of HIV transmission to children. This recommendation is consistent with clinical trials demonstrating exceptionally low rates of postnatal transmission if effective maternal antiretroviral drug treatment is in place [5]. Longer term infant prophylaxis is recommended only for infants whose mothers initiated treatment late and/or who are nonadherent with treatment. In both cases, maternal viral loads through the breastfeeding period were less than 1000 RNA copies/ml and in-country treatment guidelines were followed in terms of prophylactic regimens [6]. Because of concerns that infant prophylaxis may mask underlying HIV infections, some have encouraged repeat testing after prophylaxis has ended [7]. What distinguishes our cases is that the pattern of negative results was observed after discontinuation of antiretrovirals given to the infants but during continuation of maternal antiretroviral treatment and breastfeeding. Although the profile of the results that we observed is rare, we support repeat testing of HIV-exposed breastfed infants after complete cessation of all breastfeeding. The profile of HIV antibodies is of great interest, but it is unlikely that changes in standard anti-HIV IgG antibody titres over time in these infants will be helpful to resolve the timing of transmission. Infant antiretroviral prophylaxis. and on-going antiretroviral exposure via breast milk, like early treatment, may lead to undetectable antibodies [8]. The ability to detect HIV antibody would probably be at the time when these factors are no longer in place and viral levels rise and drive antibody production. Thus, a rise in antibodies, only once breastfeeding has stopped, might not necessarily mean the children were infected later. More sensitive assays differentiating antibody isotypes and antigen specificity may be more informative but will need comparison with the usual profile of antibodies in HIV-exposed uninfected children. The utility of HIV antibodies as markers of viral processes is well appreciated and this is likely to be a productive, albeit complex, line of research. Acknowledgements Conflicts of interest There are no conflicts of interest.