What can we do to reduce mother to child transmission of HIV?

In rich countries antiretroviral treatment in pregnancy has been highly successful in reducing transmission of HIV from mother to child. James McIntyre and Glenda Gray examine whether the success of that treatment—and other strategies, such as changing infant feeding practices—can be repeated in poor countries, which lack both money and infrastructure

HIV infection and AIDS threaten to reverse the gains made in child survival through the more widespread use of childhood vaccines and improved management of diarrhoeal and acute respiratory infections. Although the use of antiretroviral prophylaxis has dramatically reduced mother to child transmission of HIV in the world's rich countries, the effects of these interventions on infant and child survival in poor countries remain undocumented. Our knowledge of the mechanisms and timing of transmission, associated risk factors, and successful methods to reduce the risk have improved over the past decade. Indeed we know how to reduce mother to child transmission of HIV in even the worst affected regions.¹,²

The World Health Organization has promoted a three pronged approach to reducing mother to child transmission: the prevention of (a) new infections in parents to be, (b) unwanted pregnancies in HIV infected women, and (c) transmission from an HIV infected mother to her infant. The focus to date has been on strategies to prevent transmission to the infant; much is known about these strategies, but the challenge remains in their implementation.

Summary points

• Over 600 000 infants worldwide are infected with HIV from their mothers each year

• Transmission rates are as high as 35% when there is no intervention and below 5% when antiretroviral treatment and appropriate care are available

• Antiretroviral treatment, as long or short course prophylaxis or for treatment of maternal condition, reduces the risk of mother to child transmission

• Breast feeding is a major contributing factor to mother to child transmission

• Implementing programmes to prevent mother to child transmission has been difficult and slow in poor countries

• Future research should focus on prevention of postpartum infection and operational issues

Methods

This article is based on a Medline search of original papers and reviews, unpublished material, presentations and abstracts from recent scientific meetings, information obtained from the internet, and personal communications with scientists and clinicians.

Preventing mother to child transmission

UNAIDS (the joint United Nations programme on HIV and AIDS) estimates that there are over 600 000 new infections from mother to child transmission annually,³ a number that is expected to increase rapidly as prevalence rises in South East Asia. This is in sharp contrast to the situation in the United States, where perinatally acquired HIV infection is a disappearing disease as more people gain access to appropriate treatment. Fewer than 200 cases occurred in the Unites States in the year 2000⁴; fewer children are infected there in one year than in one morning in the rest of the world. Elective caesarean section can reduce the risk of transmission by half but is not yet a realistic option for poor countries. Alternative approaches to reducing transmission, such as vitamin A supplementation and chlorhexidine cleansing of the birth canal, have not showed success. For poor countries therefore, the two most effective interventions are antiretroviral prophylaxis and modification of infant feeding practices.

Antiretroviral prophylaxis

Antiretroviral treatment during pregnancy led to a dramatic fall in perinatal infections in the United States and Europe. Within the first three years of routine use of zidovudine in pregnancy, transmission rates in the United States declined by between a half and two thirds.⁵ The transmission rate is now less than 5%.⁶ A similar pattern was seen in Europe, where the rate of transmission fell from 15.5% in 1994 to 2.6% after 1998.⁷

Key ongoing research

• Use of antiretroviral treatment to the baby only, as post-exposure prophylaxis of infection

• Use of antiretroviral treatment during the period of breast feeding to protect the infant from infection

• Potential protective effect of exclusive breast feeding

• Development and persistence of nevirapine resistance following single dose nevirapine treatment

• Low cost pasteurisation mechanisms for breast milk

• Effectiveness of targeted versus universal administration of nevirapine

• Operational research on service provision of mother to child prevention strategies

Zidovudine

In 1994 the landmark PACTG trial 076 showed that zidovudine was effective in reducing transmission.⁸ But it was obvious that this expensive regimen—antenatal oral zidovudine, intrapartum intravenous zidovudine, and six weeks of oral zidovudine in the babies—would not be feasible in most poor countries.

Several trials began to investigate the efficacy of short course zidovudine in late pregnancy. The first results, from Thailand, showed that four weeks of zidovudine given in late pregnancy produced a 50% reduction in transmission in a non-breastfeeding population.⁹ A second randomised trial of this regimen was conducted in 260 women in Côte d'Ivoire, where more than 95% of the infants in the trial were breast fed by their mothers. Zidovudine resulted in a 37% reduction in transmission by three months of age.¹⁰ Another trial, in Burkina Faso and Côte d'Ivoire, compared placebo with oral short course zidovudine in over 350 women. In this trial over 85% of infants were breast fed for longer than three months. Zidovudine reduced transmission by 38% (95% confidence interval 5% to 60%) at age 6 months,¹¹ and 30% at age 15 months.¹²

Zidovudine plus lamivudine

The PETRA study, in five sites in South Africa, Tanzania, and Uganda, investigated different regimens of a combination of zidovudine and lamivudine (also known as 3TC) in over 1700 women. The trial compared the effectiveness of three regimens of different length of zidovudine and lamivudine with placebo. In group 1 the treatment was given from 36 weeks' gestation and during labour and to mother and child for one week postpartum. In group 2 the treatment was given from the onset of labour and to mother and child for one week postpartum. In group 3 the treatment was give during labour only.

Seventy four per cent of the women started breast feeding. Early efficacy results at age 6 weeks showed that the risk of transmission was 5.9% in group 1, 8.9% in group 2, 14.2% in group 3, and 15.3% in the placebo group. However, by age 18 months, HIV infection rates had increased to 14.9%, 18.1%, 20.2%, and 22.2% respectively, with little benefit seen because of continued breast milk transmission.¹³

Nevirapine

Nevirapine is a fast acting, potent antiretroviral with a long half life. The HIVNET 012 trial in Uganda investigated the use of one 200 mg dose of nevirapine orally to women at the onset of labour and one dose of 2 mg/kg to the baby within 72 hours, compared with intrapartum zidovudine and one week of zidovudine treatment to the child. Almost all babies were breast fed. The transmission rate at age 14-16 weeks was 25.1% in the zidovudine group and 13.1% in the nevirapine group, a reduction of 47% (95% confidence interval 20% to 64%). Side effects were similar with the two regimens, which were both well tolerated.¹⁴

The HIVNET 012 nevirapine regimen has many advantages for poor countries in terms of feasibility, adherence, and cost. One disadvantage may be the development of resistance to the drug, even after single dose administration. A study in Uganda showed that 19% of women who received only one intrapartum dose of nevirapine developed nevirapine resistant virus.¹⁵ High viral load and low CD4 cell counts were associated with the development of resistant variants. Resistance is more likely to occur in certain HIV subtypes. Resistant mutations at six to eight weeks postpartum were detected more frequently in women with HIV subtype D than in women with subtype A (adjusted odd ratio 4.94; 95% confidence interval 1.21 to 20.22).¹⁶ No comparable results are available for subtype C, the most common subtype in the high prevalence areas of southern Africa, although studies are in progress to determine this. In children, resistant mutations were detected in 46% (11/24) of the infants who were infected by age 6-8 weeks and who could be evaluated.

The long term impact of the selection of nevirapine resistant virus on future treatment options and subsequent pregnancies is not known. The current international recommendation is that worries about resistance should not delay the implementation of nevirapine based programmes to reduce mother to child transmission.

Infant feeding

Breast feeding continues to be a major contributing factor to high transmission rates in Africa. The debate on infant feeding has been lively, not always driven by science, and confusing to the public and health professionals. Concern remains about the effect of type of infant feeding on HIV transmission, infant survival, and efficacy of antiretroviral interventions. Although breast feeding may be the major determinant for the difference in transmission rates between rich and poor countries, safe replacement feeding options are not always available in poor countries. International guidelines recommend that HIV positive women should be given the information to make an informed choice about the risks and benefits of breast feeding and replacement feeding and be supported in their choice.

Miotti et al in Malawi showed a 0.7% per month incidence of transmission from breast feeding at age 2-6 months and 0.3% incidence at age 12-24 months, equivalent to about 3% additional risk at 12-24 months.¹⁷ A randomised controlled trial comparing breast feeding and replacement feeding in Nairobi showed a cumulative probability of HIV infection at age 24 months of 36.7% in the breastfeeding arm and 20.5% in the formula feeding arm¹⁸; 44% of HIV infection in breastfed infants was attributable to breast milk. Although two year mortality was similar in the two groups, HIV-free survival was significantly lower in the breastfeeding arm (58% v 70%, P=0.02).

Coutsoudis et al proposed that the pattern of infant feeding influenced the rate of transmission.¹⁹ In their subanalysis of mother and infant pairs enrolled into a vitamin A supplementation study, transmission rates in exclusively breastfed and formula fed infants were similar at age 3-6 months, whereas those with mixed feeding had higher rates of infection. At age 15 months the rates were higher in the exclusive breastfeeding group than in the formula feeding group,²⁰ although the cumulative infection rate remained lower with exclusive than with mixed feeding. These findings require further investigation in a larger prospective trial.

Effect of HIV related maternal mortality on infant survival

Preventing mother to child transmission of HIV-1 may be only one of several interventions neeed to increase the likelihood that infants survive. Studies in Africa have shown a threefold to fourfold increased risk of death in children whose mothers have died.²¹ In areas of high HIV prevalence, HIV/AIDS has become a major cause of maternal mortality. HIV/AIDS was the leading cause of maternal mortality in the South African Confidential Enquiry into Maternal Deaths in 2000.²² A recent study from Kenya found that mortality among HIV-1 infected mothers was higher in women who breast fed than in women who fed their babies with formula (18 v 6 deaths; log rank test P=0.009).²³ In this study, maternal death was associated with an increased risk of subsequent infant death, and, even after controlling for HIV-1 infection in the infant, it was shown that infants of mothers who died had an eightfold increase in the likelihood of subsequent death. We urgently need to assess the association between breast feeding and maternal death in HIV-1 infected women and the reasons that infants die after the death of their mothers.

Priority research needs

• Alternative antiretroviral regimens

• Need for elective caesarean section when antiretroviral treatment is available

• Interventions to make breast feeding safer

• Interventions to make replacement feeding safer

• Use of vaccines to prevent mother to child transmission

• The influence of HIV subtype on the development of antiretroviral resistance

• Effective strategies to prevent new infections in HIV negative women identified in mother to child transmission programmes

• Development of single dose technologies for nevirapine syrup that can be used for deliveries outside the health sector

Moving towards implementation

Although trials have identified, and shown cost effectiveness of, suitable preventive regimens for poor countries, implementation has been slow in the worst affected areas in Africa owing to a combination of real and perceived barriers and divergent political agendas.

Pilot projects in nine African countries under the auspices of the United Nations agencies have provided important information on the problems of introducing these interventions. Over 80 000 women have been seen in these programmes, 19% of whom were HIV positive, and around a third of eligible women have been treated.²⁴ The Elizabeth Glaser Pediatric AIDS Foundation is supporting 70 sites worldwide through its “Call to Action” programme. But few African countries have made any commitment to nationwide implementation, with Botswana the notable exception. In South Africa, politics and science on this issue have diverged so much that it has ended up before the courts.

Major barriers to implementation are cost and infrastructure. Few places have antenatal care services sufficiently well established to introduce voluntary counselling and testing without some additional resources. Up to half of deliveries occur outside the health sector, making baby dosing difficult, and drug supply logistics are inadequate. Equally important is the abiding stigmatisation of HIV, which limits the uptake of testing, treatment, and infant feeding modifications. Concern about the implications of nevirapine resistance has caused some hesitation, as has the ongoing dilemma of infant feeding advice.

These considerations have led some experts to recommend the universal provision of nevirapine in high prevalence areas, without prior HIV testing. The “universalists” argue that the emergency situation of perinatal HIV transmission demands action, that more children could be protected in this way, that this is more cost effective, and that it will take years to put in place the necessary infrastructure to provide testing and counselling in all centres. In response, an “individualist” approach argues that women should have the right to know their HIV status and that programmes to reduce mother to child transmission could and should be the entry point for more appropriate care for the family. Universal administration may exacerbate problems with resistance, may cause low adherence rates, and may have a detrimental effect on prevention campaigns if communities mistake them as an indication that they are all infected. It would also make any modification of infant feeding impossible, rendering the programme ineffective in the long term.

While the challenges to implementation remain, and additional research issues exist, one thing is certain: effective options to prevent mother to child transmission of HIV do exist, and we have a responsibility to implement them as soon as possible.

Figure.

Effect of interventions on rates of mother to child transmission of HIV