High rates of HIV acquisition in pregnancy and post partum in Francistown, Botswana
Polly Clayden, HIV i-Base
In an oral presentation Lydia Lu from the CDC showed worrying findings from a study of HIV incidence during pregnancy and the first post-partum year (in which the majority of women breastfed) among women in Francistown, Botswana. 
Dr Lu explained that women are routinely tested for HIV in Botswana at a median of 22 weeks gestation. However, this strategy may fail to identify women with acute infection in the window period or infection acquired after testing. Maternal seroconversion during pregnancy or breastfeeding greatly increases the risk of mother to child transmission.
The study included women with a documented negative HIV test during pregnancy: 400 on maternity wards, and 400 attending immunisation clinics with infants age 9-15 months.
To calculate the number of women infected post testing and in turn paediatric infections, the study investigators assumed a total of 43,000 annual births in Botswana and a 32.4% HIV prevalence, giving 29,068 women whose first ante natal clinic (ANC) test is negative. The transmission rate for mothers receiving prevention of mother to child transmission (PMTCT) interventions in Botswana is currently 4.7%. Using data previously reported describing very high rates among women seroconverting during pregnancy and breastfeeding, receiving no intervention, they assumed rates of 73% and 36% respectively among post partum transmissions. 
Rapid testing and counselling were conducted in accordance with local guidelines, and HIV-positive women were referred for HIV care, PMTCT, and infant testing. Women tested on maternity wards (n=400) had a median interval of 17 weeks from their negative test and those attending immunisation clinics with available data (n=244) a median of 62 weeks (with infants of a median age of 11 months).
Dr Lu reported HIV incidence of 5/400, 1.3% (95% CI, 0.5-3.1%) among women tested on maternity wards and 7/244, 2.9% (95% CI 1.3-5.6) at immunisation clinics. The investigators calculated an overall HIV incidence of 1.8% at one year (see table 1).
Table 1. Calculation to estimate incidence at 1-year post partum
|Weeks||(62-17)||45 weeks||52 weeks|
They then estimated, of 43000 pregnant women in Botswana in 2007, 13932 (34.7%) were diagnosed during ANC and the remainder, 29068 (65.3%), assumed to be HIV negative. Of this group 378 (1.3%) would be infected during pregnancy and 450 (1.8%) while breastfeeding.
A transmission of 4.7% among 13932 women would result in 620 HIV-positive infants. A transmission rate of 73% among 378 women infected in pregnancy would result in 276 HIV-positive infants; and of 36% among 450 women infected 1-year post partum would result in 186. Therefore they estimated incident cases of maternal HIV to account for 462/1082 (43%) of infant infections.
The investigators concluded: In this mature and successful PMTCT programme, new and undetected maternal infections may be causing nearly half of infant infections. They offered a number of recommendations including: better prevention strategies; routine re-testing; identification of the most appropriate intervention in pregnancy (start HAART?) and the safest infant feeding for women with new infections while breastfeeding.
These data, while sobering, provide one of the most important messages from this conference and the investigators must be congratulated for such a clear analysis.
Previous reports from other settings have revealed transmissions that are most likely attributed to maternal serconversion. In a survey of mothers and infants at 6-month immunisation in KwaZulu-Natal, Rollllins at al reported a group of women (7.6%), having tested HIV-negative during the antenatal period but with HIV antibodies identified in the dried blood spots of their infants and 31.2% of these infants were infected. 
The investigators suggested that these women may have been in the window period at the time of their HIV tests or they may have been infected during pregnancy. And among the 54 infants born to undiagnosed women reported in the Perinatal Transmission Survey of HIV in England, at least 20% were born follllowing maternal seroconversion during pregnancy. 
The significant question is, of course, what to do about it? To which, as yet, there is simply no straightforward answer. The investigators suggest better prevention strategies, but coy discussions about husbands that we heard following the presentation are unlikely to be very effective. Some have suggested that this woulld be a potential useful role for PrEP.
Re-testing is an obvious answer, but when? Too early and the risk of seroconversion remains, too late and the efficacy of PMTCT interventions decrease and earlier in utero infection will be missed.
It is unclear what was meant by better infant feeding. Avoiding breastfeeding complletely for all women? But the same group last year showed scary mortality findings from a group of formula fed infants in the PMTCT programme in Francistown, when the water became contaminated. 
This issue provides a big obstacle to prevention of paediatric HIV.
- Lu L et al. HIV Incidence in Pregnancy and the First Post-partum Year and Implications for PMTCT Programs, Francistown, Botswana, 2008. 16th Conference on Retroviruses and Opportunistic Infections 2009. Montreal. Abstract 91.
- Humphrey JH et al. Mother to child transmission among Zimbabwean women who had their primary HIV infection during pregnancy or while breastfeeding. 16th International AIDS Conference. Toronto. 2006.
- Rollins N et al. HIV prevalence rates amongst 6 week old infants in South Africa: the case for universal screening at immunisation clinics. 16th International AIDS Conference, 2006, Toronto. Abstract THAC0104.
- Perinatal transmission of HIV in England 2002-2005. October 2007. Download from CHIVA website (last accessed February 2009)
- Creek T et al. Role of infant feeding and HIV in a severe outbreak of diarrhoea and malnutrition among young children, Botswana, 2006. 14th Conference on Retroviruses and Opportunistic Infections, 2007, Los Angeles. Abstract 770.