Two cases of HIV transmission from mother to infant during
the breastfeeding period when mothers had an undetectable viral load have been
reported by PROMISE, a large international study of the effectiveness
of antiretroviral treatment in preventing vertical HIV transmission.

The findings were presented in a poster at the 22nd
International AIDS Conference in Amsterdam in July.

Earlier this year an
international group of researchers called for more research
to determine if
HIV can be transmitted through breast milk even if the breastfeeding mother has
an undetectable viral load in blood. Swiss
doctors have argued that pregnant women with HIV should be informed of the
uncertain evidence about the risk of transmission
during breastfeeding, and
rather than being prohibited from breastfeeding while taking antiretroviral
drugs, should be supported to breastfeed safely through regular viral load
testing and education about factors that might increase the risk of
transmission, such as mastitis.

The PROMISE study was a large international study conducted
in 14 low- and middle-income countries, investigating the effectiveness of maternal
antiretroviral therapy in preventing HIV transmission and its impact on
maternal health. The study recruited women with CD4 cell counts above 350
cells/mm3 and randomised participants at three time-points: before delivery,
after delivery during the breastfeeding period (postpartum), or after
breastfeeding was discontinued.

In the postpartum randomisation, 2,431 mother-infant pairs
were randomised to either maternal antiretroviral therapy or infant prophylaxis
with nevirapine during the breastfeeding period, 6-14 days after delivery.

Maternal viral load was tested at study entry and at weeks
6, 14, 26 and 50 weeks postpartum. Infant specimens for nucleic acid testing
were collected at study entry, week 6 and every four weeks subsequently to
detect HIV DNA. (HIV diagnosis in infants must be established by testing for HIV DNA as infant antibodies do not appear until around 18 months of age). Infants were classified as infected with HIV if they had two
positive nucleic acid test results.

Seven infants in the antiretroviral therapy arm and seven
infants in the infant prophylaxis arm tested positive for HIV DNA in the primary analysis, which covered the period up to 56 days
after the end of the breastfeeding period or 18 months postpartum, whichever came first.

These data were reported previously in a 2018
publication of the study findings in the Journal
of Acquired Immune Deficiency Syndromes
. (A secondary analysis, which included all infections reported up to 24 months postpartum, reported eight infections in mother-infant pairs randomised to antiretroviral therapy.)

New data from the study, presented in Amsterdam, show that
two infants tested positive for HIV DNA either at the same time as their
mothers had an undetectable viral load or viral load 40 copies/ml, or
shortly afterwards.

In one case, an infant tested positive for HIV DNA at a week
14 postpartum visit while its mother had an undetectable viral load. However,
at the preceding baseline and week 6 visits, the mother had a detectable viral
load above 40 copies/ml but below 1000 copies/ml.

In the second case, an infant tested positive for HIV DNA at
around week 36 postpartum and the positive status was confirmed at another
visit shortly afterwards (unspecified interval) and again at week 50. The
infant’s mother had an undetectable viral load at weeks 14, 26 and at each of
the visits at which the infant tested positive for HIV DNA.

There are several possible explanations for these cases. One
possibility is that even when HIV is undetectable in blood, it may still be
transmitted in breast milk through cell-associated virus. The volume of breast
milk consumed during the breastfeeding period and the amount of potentially
infected cells in breast milk mean that the risk differs from sexual
transmission and may be much higher, despite undetectable viral load in blood.

Another possibility is that, in the first case, viral
suppression was too slow to prevent HIV transmission, and that infection had
taken place at some point after week 6 when viral load might still have been
detectable. If that were the case, it suggests that HIV transmission through
breast milk might occur when viral load measured in blood is below 1,000 copies/ml.

In the second case, poor adherence might explain
transmission, but it is hard to see how viral load could have rebounded after
week 26 but returned to undetectable levels by week 36.

Taken together, these cases of transmission suggest that
undetectable does not mean untransmittable in the case of breastfeeding.