'Real world' prevention of mother to baby HIV transmission studied in Kenya

It is one thing to show that a medicine works under closely supervised conditions in clinical trials, another to show that it works in 'real world' medical practice.

Two Kenyan studies, reported this week in the January 3 2003 edition of AIDS (now available online), highlight some of the differences that can emerge between approaches that, on the basis of previous studies, should have been equivalent. Both studies are, in fact, randomised clinical trials which are still likely to have given better results than routine practice.

One study explored the impact of rapid antibody tests on access to voluntary counselling and testing for pregnant women. It showed that when women had same-day access to their results, more positive women received their results than was the case if they had to come back for them a week later.

Among 1249 women who agreed to be tested, 627 were randomly assigned to rapid testing and 622 to conventional tests, which meant coming back for results at least seven days later. Those given the rapid tests were given the option of deferring the results, but few wanted to do this. Overall, 96% of the HIV positive women who received rapid tests learned of their HIV status, compared to 65% of those who received conventional tests (Malonza).

However, rates of uptake of preventive treatment were just as low, overall, in both groups. Only 15 out of the 79 seropositive women offered rapid testing and only 9 out of the 82 offered conventional testing actually took ARV treatment to prevent mother to child transmission. The main reason was because women had to accept referral to a different medical centre in order to receive treatment. They failed to take up this offer for reasons which seem to have included concerns over the attitude of their partner and immediate family, being seen by unfamiliar hospital staff, difficulties with transport, and a preference to give birth in a home village or town, where they could rely on support from their extended family.

Researchers and public health authorities in a number of countries have found that people attending VCT services prefer same-day results, for example, in South Africa, where the use of rapid tests supported by trained volunteer counsellors has been seen as vital for improving access and uptake of VCT, including among pregnant women (Campbell; Hausler; Van Dyk). The use of rapid tests has also been adopted where women are seen late in pregnancy, as the only practical way to increase access to preventive treatment, in countries such as Thailand (Teeraratkul) and Brazil (Veloso) as well as South Africa.

Clearly, however, women should have received more than tests alone at the centres where they were first seen. But what else should that have been?

The second study, carried out at that very referral centre - the Kenyatta National Hospital in Nairobi - compared two short-course preventive treatments, one based on nevirapine and the other on AZT. Indeed, those women who were referred from the trial of rapid testing compared to conventional testing were invited to join the second trial. A total number of 139 women were randomly assigned to one or the other treatment (Kiarie).

It emerged that the nevirapine (just two doses, one to the mother and one to the baby) was far more likely to be taken than the AZT (twice daily from 36 weeks and then every 3 hours during labour). Taking the AZT during labour seems to have been particularly difficult.

When adherence/compliance was defined as taking 80% of the prescribed doses, overall adherence was 87% to the nevirapine and just 41% to the AZT. Women whose partners were perceived as supporting them in taking the treatment and/or whose partners were aware of their own HIV status were far more likely to take their treatment correctly than women whose partners were perceived as ignorant of their status and/or unsupportive in their attitudes.

Adherence to nevirapine treatment was measured solely on what the women said, but adherence to AZT - dispensed 20 pills at a time - was also measured using an electronic pill-counting bottle and pill counts to check what women had said. (There was very close agreement between the results from the three methods, however.) In focus groups, some of the women reported sharing AZT pills with their partner - a problem that did not arise with the single dose provision of nevirapine.

The number of births in the study was not large enough to be sure of the different efficacy of the two regimens. However, there was a trend for the AZT to be more effective, especially in the minority of women who took all the doses as prescribed. Of 110 babies tested for HIV at 6 weeks, 17 were found to be HIV positive (15%). The transmission rate on the nevirapine arm (22%) was higher than on the AZT arm (9%). At least 40% of the babies in both groups were breast-fed.

The real take-home message from this study is that social factors - especially the perceived and actual support given to pregnant women with HIV by their partners and families - will be the real determinant of whether they can access treatment where it is available free of charge. This applies equally to treatment given to prevent HIV transmission and, when it becomes available, treatment to support their own and their partners' health.