HIV prevalence rates amongst 6 week old infants in South Africa: the case for universal screening at immunisation clinics

Polly Clayden, HIV i-Base

Standard PMTCT surveillance methods have failed to assess the number of vertical infections being prevented in infants. Although rates of 11.9% with single dose nevirapine in the HIVNET 012 study are quoted liberally in discussion around the success of this strategy, field data can show considerably higher rates of transmission.

“We are spending millions of dollars and we don’t know how well we are doing” said Nigel Rollins of the University of KwaZulu-Natal (KZN) presenting data from a pilot surveillance study to determine mother to child transmission (MTCT) rates in KwaZulu-Natal.

In this programme, routine, anonymous, unlinked, HIV prevalence testing was performed on infants with consenting parents or guardians, attending 6-week immunisation clinics at seven primary health care clinics offering PMTCT services.

Dried blood spot (DBS) samples were collected on filter paper and tested for HIV antibodies (maternal) using a commercial ELISA. DBS samples of infants that were antibody positive were then tested for HIV RNA by PCR. The mothers were also asked about any previous pregnancies and whether the child was alive or dead.

Dr Rollins explained that trained lay people from local communities collected the DBS and conducted the interviews.

DBS samples were collected from 2,439 infants aged 4-8 weeks of age who were brought for their first DTP immunisation. The investigators reported HIV antibodies in 914 infants representing the maternal seroprevalence rate of 37.6% (CI: 35.7-39.6%); amongst mothers aged 20-29 years this was 46.9% (CI: 42.9% – 50.9%). This concurs with previously reported prevalence rates for KZN.

189 children born to these positive mothers tested HIV positive by PCR on the DBS (7.6%) giving a transmission rate of 20.8% (CI: 18.2 – 23.6%). Amongst mothers who reported that they were HIV-positive who had taken single dose nevirapine, transmission rate was 15.3%.

A group of 7.6% women reported having tested HIV-negative during the antenatal period but HIV antibodies were identified in the DBS of their infant and 31.2% of these infants were infected. Dr Rollins 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. High maternal viral load following seroconversion is likely to have caused this high transmission rate.

He concluded that screening of all infants at immunisation clinics using this method effective is feasible for monitoring the overall impact of PMTCT programmes.

He then went on to discuss factors that could contribute the poor performance of these programmes. Although there is a group of women for whom testing and nevirapine are effective for PMTCT, he reminded us that women may become infected in the antenatal period, or be in the window period during routine screening. Additionally a number of women have no access to services and therefore receive no intervention.

Child mortality rates – which had reduced considerably from 48 per 1000 before 1990 to 31 per 1000 between 1990 and 1994 – has increased dramatically in the last five years to 99 per 1000. Dr Rollins attributes this to HIV infection of the children and greater mortality among children whose mothers die of AIDS.


Outcomes in clinical trials can be expected to be better than in the field, but, nevertheless, these results are disappointing, as even by age 6 weeks the HIV infection rate among those infants exposed to NVP in labour was 15.3%.

The high rate of primary infection during pregnancy among these mothers is a major cause for concern. The 7.6% HIV positive women not identified in pregnancy is higher than has been reported for seroconversion rates elsewhere. One large study in progress in KZN found a rate of about 5% – reported in a conference abstract – and a lower rate was reported in an article looking at patients from Western Cape.

It is unclear what the uptake of testing was in pregnant women in this programme, and it could be that women who knew they were positive did not agree to be tested. The 15% transmission rate in women who took NVP is similar to the Kampala data presented at CROI, but higher than reported rates with NVP in urban Johannesburg settings (Bara and Coronation reported around 9%) and probably reflects the predominantly breastfeeding environment.

Although these data are depressing, they make a strong argument for opt-out testing in pregnancy, and repeat tests in late pregnancy or at delivery in high prevalence areas, and the implementation of the best possible regimen available, with an emphasis on antenatal and maternal care.

Dr Rollins made “the case for universal screening at immunisation clinics”. This would offer another opportunity for a woman to learn about her status and look after her own health as well as her child’s, and an opportunity to identify positive children and get them into early HIV care and follow up.

Single dose nevirapine still remains the mainstay of many PMTCT programmes, and represents the minimum intervention where other interventions are not available. But once again these data highlight a massive variation in this strategy’s efficacy. Even with this minimum intervention, there is room for improvement of services. The latest version of the WHO guidelines clearly advocate for the use of more complex interventions such as HAART, or short course AZT plus single dose nevirapine where possible. To a question about adding short course AZT to the single dose nevirapine, Dr Rollins argued that not only the regimen, but also the quality of the programmes and services need to be addressed and the health system supported.


Rollins N, Mzolo S, Little K et al. HIV prevalence rates amongst 6 week old infants in South Africa: the case for universal screening at immunization clinics. XVI International AIDS Conference, Toronto, Canada. 13-18 August 2006. Oral abstract THAC0104.

Links to other websites are current at date of posting but not maintained.