Infant feeding and length of treatment: possible changes to UK pregnancy guidelines

While refraining from breastfeeding remains the best and safest option for woman with HIV in the UK, women taking effective anti-HIV treatment should be supported to breastfeed if they choose to do so, Gareth Tudor-Williams argued at the joint conference of the British HIV Association (BHIVA) and the British Association of Sexual Health and HIV (BASHH) on Thursday.

In addition, two studies were presented to the conference that could lead to alterations in treatment in pregnancy: one suggested that treatment needs to be started earlier in the pregnancy than is current practice and another found that a rupture of membranes (breaking of waters) long before delivery may no longer pose the risk it once did if the mother is taking effective antiretroviral therapy.

Infant feeding

Gareth Tudor-Williams outlined a proposed revision of current guidelines, which is open for consultation until May 21. He said the primary impetus for reconsidering this question had come from a number of studies conducted in sub-Saharan Africa, such as BAN and Mma Bana. He noted that patients in the UK sometimes asked clinicians about the implications of these studies.

The proposed guidelines make it clear that the safest infant feeding option is the avoidance of breastfeeding, and that this should be recommended to all mothers with HIV, regardless of the mother’s use of treatment or viral load.

Moreover, mothers should be supported to formula-feed their babies. This means that formula milk, sterilisers, bottles and other equipment should be freely available to mothers. Tudor-Williams noted that this was not always current practice in every part of the country, especially for refused asylum seekers.

However the guidelines “acknowledge that, in the UK, the risk of mother-to-child transmission from a woman who is on HAART and has a consistently undetectable HIV viral load is likely to be low but has not yet been quantified”. The draft text proposes that “if a woman is on effective HAART and chooses to exclusively breast-feed having carefully considered the aforementioned advice, she should be supported to do so as safely, and for as short a period, as possible”.

At present, breastfeeding is discouraged and could even be considered a child-protection issue. As a result, mothers may hide the fact that they breastfeed from healthcare providers and may not access support.

Safety issues would include the mother continuing to take antiretroviral therapy until a week after ending breastfeeding; monthly maternal viral load testing; monthly assessment of the infant (side-effects and HIV status); and avoidance of mixed feeding.

It is not recommended that babies should be offered antiretroviral prophylaxis throughout the breastfeeding period, as data are lacking on the safety of this approach.

Tudor-Williams stressed that this strategy would not be appropriate for a woman who was “in denial” and didn’t want to follow clinician’s advice. “It’s very much for a woman who has really thought about this, who has great desire to breast-feed, who really understands the need to stay on HAART and not to miss doses and to keep her viral load undetectable.”

When to start treatment in pregnancy

Annemiek de Ruiter, presenting on behalf of Philip Read, gave the results of a retrospective analysis of pregnancies, conducted to determine the optimum time to commence HIV therapy during pregnancy.

For pregnant women who do not need HIV treatment for their own health, current British guidelines say that a short course of treatment (START) should begin between weeks 20 and 28 of pregnancy. Moreover, vaginal delivery is supported by the guidelines as long as viral load is undetectable at the time of delivery. However, little is known about the likelihood of achieving an undetectable viral load by the time of delivery if treatment is started between weeks 20 and 28.

The study analysed data on 378 pregnancies collected since the year 2000 at five large HIV clinics in London and Brighton. A pregnancy was included if combination therapy was begun in pregnancy and viral load had been recorded with a test whose cut-off point for ‘undetectable’ was 50 copies/ml or lower.

In line with the current guidelines, most women began treatment between 20 and 26 weeks (median 23 weeks). Two-thirds planned to come off treatment after delivery, with the others needing to take treatment for their own health. Just over three quarters of women (77%) had a viral load below 50 copies/ml at the time of delivery.

The analysis demonstrated that the level of viral load at baseline (the time of starting treatment) was strongly correlated with the ability to achieve an undetectable viral load by the time of delivery.

• Baseline above 100,000 copies/ml: 37% achieved undetectable by the time of delivery.
• Baseline 50,000 – 100,000 copies/ml: 54% undetectable.
• Baseline 10,000 – 50,000 copies/ml: 73% undetectable.
• Baseline below 10,000 copies/ml: 91% undetectable.

For women in the lowest baseline viral load group, the date of beginning treatment did not affect their chances of achieving an undetectable viral load. However, for all other groups, starting treatment later was associated with less chances of success.

“There are many factors to consider when initiating HIV therapy in pregnant women,” de Ruiter said. “However in terms of achieving an undetectable viral load by delivery these data suggest that only women with a viral load of less than 10,000 should defer treatment to beyond 20 or 21 weeks. Women with a baseline viral load of over 100,000 should start HAART without delay... It may be feasible to delay START to 26 weeks if the viral load is less than 10,000, however this does not take into account that some babies may be delivered prematurely.”

Rupture of membranes

Pat Tookey presented national surveillance data, looking into whether prolonged rupture of membranes was associated with an increased risk of HIV transmission from mother to child.

As noted above, current UK guidelines suggest that a vaginal delivery is possible if a woman is taking triple combination therapy and her viral load is undetectable at the time of delivery.

However, data from the pre-HAART era suggested that prolonged rupture of membranes (“breaking of the waters”) was associated with an increased risk of HIV transmission. There has been concern that for women who planned to have a vaginal delivery, the occurrence of rupture of membranes was leading to a decision to have an emergency caesarean section. Over a quarter of HIV-positive women now have an emergency caesarean and the rates are continuing to rise.

The researchers wished to look into whether rupture of membranes was still associated with an increased risk of transmission when effective antiretroviral treatment was being taken.

Data on rupture of membranes has been collected in national surveillance since 2007. Information on 421 pregnancies was presented – in these cases, rupture of membranes occurred before delivery, its duration was known and the infection status of the child was known.

There was no difference in the infection rate. When rupture lasted less than six hours, the infection rate was 1.4%; when it lasted longer, the infection rate was 1.5%. Sub-analyses of women delivering vaginally, through emergency caesarean or according to viral load also showed no difference according to length of rupture of membranes.

In the handful of cases where an infection did occur following rupture of membranes, it is thought that in most cases infection took place in utero (before delivery), as HIV tests were positive in the baby on the day of delivery.

However, Tookey emphasised that the current data is insufficient to warrant a change in the current guidelines: continued monitoring of the outcome of pregnancies is essential.