Why U=U does not cover breastfeeding
One of the last sessions at R4P2018 included six talks about U=U, one of which reviewed the evidence about breastfeeding. This is important because rather than just sexual transmission, U=U is sometimes wrongly interpreted as covering all ways that HIV can be transmitted.
Lena Serghides from University Health Network and University of Toronto answered the opening question of whether ART eliminates HIV transmission with a clear “no”. While it is easy to show that ART dramatically reduces the risk for all routes of transmission, it is only sexual transmission where the risk is proven to be effectively zero. 
This talk emphasised that cases of HIV transmission have already been reported when mothers are on ART with undetectable viral load, notably in the PROMISE trial. Latest results presented at AIDS 2018 reported that 2/8 cases when transmission occurred form breastfeeding were in women whose most recent viral load was <40 copies/mL. 
So although HIV treatment reduces the risk of transmission during breastfeeding, this risk is not zero. None of the studies looking at this risk have reported a zero result, with a meta analysis that drove WHO guidelines reported risk of 1.1% and 2.9% with 6- and 12-months breastfeeding in mothers on ART. 
Possible reasons for risk not becoming zero
Possible explanations for this residual risk include:
- High lifetime exposure volume compared to sexual fluid: 3-4 cups of milk a day is 150 litres over six months.
- Breast milk is likely to include 1 million immune cells/day and 150 million over six months.
- Breast milk induces immune activation and HIV replication is 10-fold higher in milk vs blood.
- Cell associated HIV DNA persists on ART and latently infected resting cells are transmitted (and then activated) in milk.
- Breast inflammation can activate HIV (mastitis, abscess, engorgement etc).
- Great immune vulnerability in infant GI tissue.
- Adherence difficulties post partum (erratic sleep, depression, support) are well documented in all settings, including high-income countries. 
Breastfeeding starts off with a much higher risk of transmission because the infant is exposed to far higher amounts of infectious fluid.
Context: low- vs high-income settings
Crucially, context for risk drives current WHO recommendations.
In low-income settings with limited access to clean water, formula feed and medical care the recommendation to breastfeed on ART means the positive benefits from breast milk in fighting malnutrition, diarrhoea and pneumonia outweigh the low risk of HIV transmission.
In high-income settings where the rates of infant mortality from background health risks are much lower, the recommendation to use formula feed has the best outcome in terms of infant health and the higher risk comes from HIV via breastfeeding.
This is an area where more data on transmission risk from mothers on ART should become increasingly available now that the WHO recommends both continued ART and breastfeeding in low income countries. Earlier reviews on risk showed heterogenicity of approaches with many studies not fully reporting timing and use of ART and importantly, maternal viral load data.
Prospectively collecting cases in large observational databases, similar to the approach in the PARTNER studies for sexual transmission, is not only ethical, but essential. This will need to included phylogenetic analyses to confirm that infections are linked, as shared breastfeeding is also reported in many settings.
Although a recent discussion paper from some of the doctors involved in the original Swiss Statement, suggested there might be equipoise on risk for women in high-income countries, this was published before the cases were presented from the PROMISE study. Phylogenetic analysis of these cases, if not already available, should also be prioritised, given their importance as crucial evidence on risk and safety of breastfeeding on effective ART.
Currently, National guidelines, including in the UK (and Switzerland), continue to recommend that the safest option for the baby in high-income countries is to use formula feed and not to breastfeed. 
At least two earlier papers – from 2005 and 2012 – report stronger links with cell-associated HIV DNA in breastmilk with increased risk of transmission to the infant. Although ART was not used throughout in these studies, the later 2012 seems to report transmissions at the lowest ranges of both HIV RNA and DNA levels. [7, 8]
However, no transmissions were reported in a more recent Tanzania cohort when approximately 200 breastfeeding mothers had viral load <1000 copies/mL on ART (75% were <100 copies/mL). The only two transmissions from breastfeeding were in cases of high maternal plasma viral load. 
- Serghides L. What does U=U mean for breastfeeding? R4P2018, 21-25 October 2018, Oral abstract R2.03.
- Flynn P et al. Association of maternal viral load and CD4 count with perinatal HIV-1 transmission risk during breastfeeding in the PROMISE postpartum component. AIDS 2018. Amsterdam. 23–27 July 2018. Poster abstract THPBE155.
- Bispo S et al. Postnatal HIV transmission in breastfed infants of HIV-infected women on ART: a systematic review and meta-analysis.J Int AIDS Soc. 2017 Feb 22;20(1):21251. doi: 10.7448/IAS.20.1.21251.
- Kahlert C et al. Is breastfeeding an equipoise option in effectively treated HIV-infected mothers in a high-income setting? Swiss Medical Weekly, 148:w14648, 2018.
- Huntington S et al. UK Collaborative HIV Cohort (UK CHIC) Study and the UK and Ireland National Study of HIV in pregnancy and childhood (NSHPC). The risk of viral rebound in the year after delivery in women remaining on antiretroviral therapy. AIDS. 2015;29(17):2269–78. doi: 10.1097/QAD.0000000000000826. Open access. https://journals.lww.com/aidsonline/fulltext/2012/10230/Adherence_to_antiretroviral_therapy_during_and.6.asp
- BHIVA. Management of HIV infection in pregnant women 2018 – Consultation version.
https://www.bhiva.org/guidelines https://www.bhiva.org/file/WrhwAPoKvRmeV/BHIVA-Pregnancy-guidelines-consultation-draft-final.pdf (PDF)
- Ndirangu J et al. (2012) Cell-free (RNA) and cell-associated (DNA) HIV-1 and postnatal transmission through breastfeeding. PLoS ONE 7(12): e51493. doi:10.1371/journal.pone.0051493
- Koulinska IN et al. Transmission of cell-free and cell-associated HIV-1 through breast-feeding. J Acquir Immune Defic Syndr 2006;41:93–99.
- Luoga E et al. Brief Report: No HIV transmission from virally suppressed mothers during breastfeeding in rural Tanzania. JAIDS Journal of Acquired Immune Deficiency Syndromes. 79(1):e17–e20. DOI: 10.1097/QAI.0000000000001758.