HIV surveillance at immunisation clinics reveals weakness in KwaZulu Natal's PMTCT programme

This article is more than 17 years old. Click here for more recent articles on this topic

The programme for the prevention of mother to child transmission (PMTCT) of HIV in KwaZulu Natal, South Africa is failing to reach even conservative HIV prevention targets according to a pilot surveillance project set up at immunisation clinics in the province.

The surveillance study detected a vertical transmission rate (VTR) of 20.8% in 6 week old infants of HIV-positive mothers, according to Dr. Nigel Rollins of the University of KwaZulu-Natal, who presented the findings at the recent Sixteenth International AIDS Conference in Toronto.

He compared these results to the findings in HIVNET 012, the study which evaluated the same single-dose nevirapine (sdNVP) regimen that South Africa’s PMTCT protocol utilises. In that study, only 11.9% of the infants aged six to eight weeks old whose mothers were randomised to sdNVP became infected — so the performance of the PMTCT programme is considerably below expectations.

Evaluating the effectiveness of PMTCT programmes

Although PMTCT trials in resource-constrained settings have demonstrated rates of transmission almost as low as those achieved in more industrialised countries, it is quite another thing to extend these benefits to the general populace outside of the clinical trial setting.



Immunisation is the process whereby a person is made immune or resistant to an infectious disease, typically by the administration of a vaccine. Vaccines stimulate the body’s own immune system to protect the person against subsequent infection or disease.


pilot study

Small-scale, preliminary study, conducted to evaluate feasibility, time, cost, adverse events, and improve upon the design of a future full-scale research project.



A detailed research plan that describes the aims and objectives of a clinical trial and how it will be conducted.


The period of time from conception up to birth.

window period

In HIV testing, the period of time after infection and before seroconversion during which markers of infection are still absent or too scarce to be detectable. All tests have a window period, the length of which depends on the marker of infection (HIV RNA, p24 antigen or HIV antibodies) and the specific test used. During the window period, a person can have a negative result on an HIV test despite having HIV.

Integrating effective PMTCT services into the health system involves making extensive changes and improvements to the existing mother-child healthcare infrastructure that go far beyond simply providing antiretroviral drugs to prevent transmission. Staff must be trained; routine HIV testing services for pregnant women, primary prevention for those who test negative and support services for those women who test positive including infant feeding support services — all have to be made widely available.

In addition, there needs to some way to monitor the uptake of those services and evaluate whether they are having their desired effect. Most studies have focused on individual aspects of the programme, such as the adequacy of counselling and so on. “But what we are really interested in preventing is death in children and the number of transmissions, said Dr. Rollins.

Existing methods fail to do this at the programmatic level. Standard PMTCT surveillance has usually involved HIV testing at the antenatal clinic — but most children are lost to follow-up shortly after delivery. Furthermore, there is no way to access women who have chosen not to participate in PMTCT programmes.

However, immunisation clinics are generally very well attended in most of southern Africa and could serve as a point of access to virtually all the infants born (not merely those who accessed PMTCT services).

The pilot programme in KZN

So in order to evaluate the effectiveness of the PMTCT programme in KwaZulu Natal, a pilot programme was set up to perform routine anonymous, unlinked, HIV prevalence testing on all infants aged between four to eight weeks old, attending 6-week immunisation clinics at seven primary health care clinics offering PMTCT services.

After obtaining the consent of the parent or legal guardian, dried blood spot (DBS) samples were collected from the infant’s heel, the eardrum or the thumb and screened for HIV antibodies. Antibody testing in infants is a fairly reliable way of determining HIV exposure (and maternal status) since babies carry their mother’s antibodies for several months after delivery and/or breastfeeding. Samples that were antibody-positive were then tested for HIV RNA by PCR.

In addition to testing the infants, the staff also asked all mothers, irrespective of the age of the child being immunised, about any pregnancy that they had had, and whether the child was alive or dead. This was done in order to determine child mortality rates over the previous fifteen year period.

Over a period of two and a half years, over 6500 such interviews were conducted and information was gathered on around 11,000 children (of various ages). Dried blood spots have been collected from 2,439 infants between the ages of 4-8 weeks who were brought in for their first DTP immunisation. About 11% of the mothers refused, knowing that the sample was going to be tested for HIV (even though the testing was unlinked and anonymous).


HIV antibodies were detected in 907 infants (37.6%, CI 35.7 - 39.6), which indicated the maternal seroprevalence rate — in mothers between 20-29 years old the rate was 46.9% (CI 42.9 – 50.9). “This reflects very closely the maternal prevalence rates in KZN,” said Dr. Rollins.

Of the exposed children, 189 were HIV-infected, (7.6% of the entire 6 week old population) indicating a vertical transmission rate (VTR) of 20.9% (CI 18.2 – 23.6%).

“Going back to the 012 trial, if every women had been identified, and everyone had received nevirapine and everyone of their children, then this number (the VTR) should have been 11.9%,” said Dr. Rollins.

In the women who reported that they had previously tested positive for HIV (virtually all of whom had received sdNVP), the VTR was between 15-16%.

However, a number of the women who claimed to be uninfected were actually HIV-positive, and among their infants the VTR was 31.2%. These women may have been in the window period before antibodies appear when they were previously tested, or they may have become infected during the antenatal period. Either way, they did not receive sdNVP, and high viral loads experienced during acute primary infection may have contributed to the high VTR.

Infant mortality rates have also been increasing dramatically in the last fifteen years. In children born before 1990, the infant mortality rate had been 48 per 1000 but between 1990-1994, a point at which things were improving, the infant mortality rate dropped to 31 per 1000. But in the fifteen years since the HIV epidemic, the infant mortality rate has been swinging up again, trebling in the last five years to 99 per 1000 births.

Implications for universal screening of infants at immunisation clinics

Although the study evaluated whether testing at immunisation clinics could be used for surveillance, it also raises the issue of whether this sort of testing should be done on a routine basis for improved provision of care.

“We generally provide screening services when we have something to offer, and today, we believe we have something to offer,” said Dr. Rollins. He noted that screening is routinely offered for rarer health issues in children, such as neural tube defects (which occurs in one of 10,000 infants), Down's syndrome (seen in one in 600-1000 infants), congenital hyperthyroidism (found one in 3000). “We are looking at a prevalence of 7% of all children at six weeks being HIV-infected. By providing this type of screening, it provides a second chance to the woman to learn her status as well as her children, and it picks up those who maybe seroconverted during pregnancy or those who had declined antenatal testing in the first place. It gives us a very immediate referral to HIV care services for both mother and child, including feeding support.”

In the discussion that followed, Dr. Cheri Luo from UNICEF agreed that the implications of these results need to be considered in policy recommendations. “I think we need to look at what threshold we need to have before making HIV screening routine at immunisation clinics. I’m not sure we’d make the same argument in China or India [which have a lower HIV prevalence among the general population] but it is critically important to start talking to that as a public health approach — and whether this would be cost beneficial for countries.”

Implications for the PMTCT programme in KZN

Dr. Rollins believes that this surveillance method is fairly simple to implement and should be rolled out to monitor the efficiency of PMTCT programmes throughout the province and potentially the country. It could also help provide projections of what sort of HIV treatment and care will be needed by children in the future.

However, it does not provide reasons for why the system is not performing well.

Dr. Rollins suggested a number of possible reasons: “There’s a group for whom testing and nevirapine are effective but there’s also a high risk group of women who may have become infected during pregnancy or were in the window period and were missed by routine screening. And there’s probably a significant proportion of women who had not accessed services and therefore any form of intervention.”

He feels that the increase in infant mortality rates may be largely due to HIV infection of the infants in and of itself, altered feeding care practices in mothers who are infected (formula fed infants have consistently poorer survival in most resource-constrained settings). Other studies also report poorer survival in children whose mothers die of HIV and AIDS.

“The implications are that the quality of the programme needs to be addressed and not just coverage,” Dr. Rollins said.

In the discussion session after the presentation, Dr. Harry Moultrie of Johannesburg asked whether Dr. Rollins had shown the South African Department of Health these data in order to encourage them to consider adding short-course AZT to the national PTMCT protocol. Other studies have shown that giving short-course AZT, in addition to sdNVP, and a short-course of post-natal AZT/3TC, result in lower rates of transmission than observed in HIVNET 012 — and these are now the PMTCT regimens recommended by WHO.

Dr. Rollins said that that he had presented this study with a smaller data set about a year ago to both the provincial and national officials, and that they had been very responsive to it.

However, the focus had been the quality of the programme and services being offered rather than the particular PMTCT protocol. The Department of Health in KZN is planning a series of interventions to improve the quality of the delivery of services delivered.

“What I think needs the most amount of work, is actually health systems intervention,” said Dr. Rollins.”How to help the facilities do the things that we already know, that’s actually the greater challenge. Its not just an issue of dual therapy. But the actual intervention is the health system’s support.”


Rollins H et al. HIV prevalence rates amongst 6 week old infants in South Africa: the case for universal screening at immunisation clinics. Sixteenth International AIDS Conference, Toronto, abstract THAC0104, 2006.