We evaluated a two-tiered PMTCT strategy in which treatment was selected based on maternal medical status (as indicated by CD4 T cell count and WHO staging) in a West African population. HAART was prescribed to pregnant women with advanced HIV disease who met WHO eligibility criteria based on their own health status, whereas short-course various combinations of ARV regimens for PMTCT were given to pregnant women with less advanced HIV disease, who did not qualify for HAART. Three-quarters of the infants were breast-fed, for a median of 5.4 mo. Overall, the rate of peripartum HIV transmission was 2.2%, the cumulative rate of infant HIV infection at 12 mo was 5.7%, and the 12-mo HIV-free survival was 88.3%, without measurable differences between the two groups.
The rate of peripartum MTCT that we observed in women on HAART was very low (1.0%), similar to rates reported from industrialized countries [2
], despite a later identification of HIV infection during pregnancy, and a relatively short duration of HAART prior to delivery. In the Women and Infants Transmission Study in the US, HIV-1 transmission rate was 1.2% in infants whose mothers had received HAART but did not breast-feed [3
]. Similarly, the European Collaborative Study Group reported the decline of MTCT transmission rates from 15.5% in 1994, to 5.1% in 1997–1998, and down to 0.99% in 2001–2002, i.e., when HAART became more widely used [2
The maternal and PMTCT benefits of HAART initiated during pregnancy that have been described in the industrialized world may differ for populations from sub-Saharan Africa because of the wider spectrum of opportunistic diseases, the deeper level of immunodeficiency at which they occur [26
], and the poorer nutritional background and obstetrical conditions. In the DREAM cohort in Mozambique where formula feeding was systematic, pregnant women initiated a ZDV/NVP/3TC HAART combination at 25 wk of gestation, irrespective of clinical or immunological staging; the rate of HIV transmission from mother to infant was 1.2% at infant age of 1 mo [27
]. HAART was continued in this cohort up to 6 mo postdelivery for those not requiring therapy for their own health, while the others continued treatment indefinitely. We observed a comparable 1 mo of age rate of transmission with a more targeted prescription of HAART. As emphasized in the recently published WHO PMTCT guidelines, it is critically important that pregnant women who are eligible for HAART receive such treatment for the sake of their own health as well as for reducing MTCT [14
]. Women with advanced HIV disease are at the highest risks both for HIV-related morbidity and mortality and for transmitting HIV to their child [28
]. By decreasing the mother's plasma HIV VL and enhancing immune function, a targeted HAART both treats the mother's own HIV infection and reduces dramatically the likelihood of HIV transmission to the newborn. If HAART is used for PMTCT in women not requiring it for their own health, however, the potential risk to the woman's health of interrupting HAART after delivery and the potential risks of continuous HAART exposure to uninfected infants are unclear.
We observed low rates of postnatal transmission in both HAART-treated women and those receiving scARV for PMTCT only. These results were observed in conditions of short breast-feeding exposure with an overall median duration of 5.4 mo. We observed one case of postnatal HIV transmission among 52 breast-fed infants of HAART-treated mothers and 3/86 among the breast-fed infants of women on scARV regimens. There was no statistically significant difference between the two groups, although sample size was limited; this study was not designed to answer the question “does HAART substantially reduce postnatal transmission risk? While HAART might diminish postnatal MTCT because it suppresses cell-free HIV RNA in breast milk, other data counteract this hypothesis [29
]. Further studies are needed to document the long-term safety and efficacy of HAART for the prevention of breast-feeding–associated HIV transmission, including among children breast-fed for long periods.
The universal use of HAART in HIV-infected pregnant women irrespective of their clinical or immune status could have several advantages, as shown in resource-rich countries. It avoids using scARV prophylactic regimens, which though effective in reducing MTCT, do not benefit the women's health and are associated with some risk of development of HIV resistance in the mother and infant, particularly with sdNVP [19
]. However, in low-resource settings, many challenges hinder the universal use of HAART for all pregnant women, including limited access to ARVs, uncertain effects on maternal health from the short-term use of HAART with postpartum interruption, and toxicity concerns about NVP-containing HAART in women with higher CD4 T cell counts [18
]. It is therefore reassuring that the selective use of HAART by HIV-infected pregnant women with advanced disease, and scARV for PMTCT by pregnant women with less advanced disease yield equivalent results in terms of HIV-free survival in their infants. Furthermore, the use of a ZDV+3TC postpartum “tail” for 7 d reduces the risk of acquiring viral resistance in women receiving scARV for PMTCT including sdNVP [19
We observed a relatively low rate of severe adverse events in HAART-treated HIV-infected pregnant women. Adverse events that did occur were rapidly resolved with appropriate drug management, and unexpected toxicities were absent; together these observations suggest that HAART during pregnancy for women with advanced disease is reasonably safe. Of note, this study did not address the question of the use of NVP-based HAART in women with high CD4 T cell counts. The fact that 7.5% of the women developed toxicities requiring a change of ARV drugs highlights the importance of careful monitoring of pregnant women on HAART. Clinical monitoring can detect many adverse events even without laboratory results: in this study mucocutaneous rashes constituted more than two-thirds of the maternal adverse events, and one woman had severe anaemia with associated clinical symptoms. However, the diagnosis of severe liver toxicity requires frequent laboratory monitoring during pregnancy. Previous reports on adverse events in pregnant women in Mozambique [15
], Kenya [33
], and Thailand [34
] noted that rates of serious hepatic or cutaneous adverse events among pregnant women receiving NVP-containing HAART were comparable to those observed in nonpregnant individuals. Liver toxicity, as indicated by serial liver function tests, was reported in 6.3% of 606 women in the DREAM cohort who were systematically treated by HAART in Mozambique [35
]. In our study, we found a rate of 2% of liver toxicity in HAART-treated pregnant women with advanced disease.
We also found that the rate of low birth weight in infants born to mothers treated with HAART was significantly greater than among those exposed to scARV regimens for PMTCT. HAART exposure has been associated with prematurity in some series [16
] while other studies have found no association [37
]. Low birth weight might be related to the fact that the mothers had more advanced HIV disease, rather than being a direct effect of HAART itself. Further assessments of these infant outcomes are underway in our cohort [39
The main limitation of our study is the cohort sample size, with relatively small numbers per treatment group, thus affecting the precision of some MTCT rates and the power to fully perform statistical comparisons between groups. However, we were able to document that the upper confidence limit of the cumulative MTCT rate at 12 mo was less than 7% in the HAART-treated population where breast-feeding was commonly practiced for 5 mo, whereas natural risk may have exceeded 30% [10
]. Also this study is not a comparative study, but a description of two tiers with one group receiving HAART and another group a short-course of ARVs for PMTCT based on HIV disease staging. Thus, comparisons between the two tiers should be interpreted with caution. A more convincing group for the comparison would be a group of HIV-infected pregnant women eligible for ART who did not receive it. However, it would be unethical to perform such a study in a setting where HAART is available for women who meet eligibility criteria.
Our team has recently summarized available historical data on HIV-infected pregnant women eligible for ART in the same Abidjan population [40
]. We estimated the transmission rate at 6 wk at 26.2% in such women having received antepartum a short-course regimen of 4 wk of ZDV alone and practicing predominant breast-feeding. In a comparable group of ART-eligible women who had received 4 wk of ZDV and sdNVP, with short-term breast-feeding, this rate was 10.1%. At 6 mo the corresponding transmission rates were 41.6% and 15.9%, respectively. It is important to note that, in the current study, the infant HIV-free survival rates in the two groups was 11.2% (HAART tier) and 12.1% (scARV tier), not different between infants born from women with high or low risk of HIV transmission. It remains unclear, however, whether the benefits of maternal HAART were counterbalanced by their effect on infants such as early neonatal death in association with prematurity.
This study has several strengths. First, retention was high, with the final HIV status available in more than 86% of live births and only 2% of infants lost to follow-up among HAART-treated mothers. Second, the high level of adherence achieved is a likely explanation for the low rate of MTCT in the HAART-treated tier. Finally, we demonstrated that it is feasible to screen pregnant women for HIV, assess disease status by performing CD4 T cell counts, and quickly initiate HAART. Those who did not require HAART were treated with standard scART regimens for PMTCT, and the vast majority of women and their families were engaged in long-term HIV services. This outcome further demonstrates that PMTCT is an excellent entry point for women and families in HIV care and treatment.
In summary, a two-tiered approach to PMTCT, in which pregnant women with advanced HIV disease receive HAART, and those with less advanced disease receive scARV regimens for PMTCT, was safe during pregnancy and highly effective for the prevention of peripartum and postnatal transmission in an African population. While further investigations in other larger or pooled studies are needed to further explore the balance of benefits and risks related to HAART in breast-feeding mothers, our results provide evidence supporting the most recent WHO guidelines [14