Nine years after the start of a national prevention of
mother-to-child transmission (PMTCT) programme, South Africa’s mother-to-child
HIV transmission (MTCT) rate is under 4%
at four to eight weeks after birth, according to South Africa’s first national
PMTCT impact evaluation. Results from this evaluation were presented at the Sixth
International AIDS Society conference (IAS 2011) in Rome.
The dramatic reduction is the result of the implementation of a comprehensive national programme to prevent transmission of HIV from mother to infant, through antenatal HIV testing and provision of antiretroviral prophylaxis or treatment for mothers and infants.
There is much documented evidence of the efficacy of PMTCT
interventions to reduce perinatal transmissions. However, evidence of the
impact and effectiveness of PMTCT programmes at a national level is lacking.
Most high-prevalence countries do not have ongoing
country-level monitoring of PMTCT. So the impact of PMTCT programmes on MTCT
rates at the population level is unknown; this is the case for South Africa
where PMTCT programmes began in 2002.
With international momentum for the elimination of
paediatric HIV by 2015, global targets have been set to reduce new paediatric
infections by 90% and population-level mother-to-child infection rates to under
5% at 18 months in breastfeeding populations.
Measuring the impact and effectiveness of PMTCT programmes
is critical for national programme planning and the eventual elimination of
MTCT. No standardised, internationally recognised methods for measuring
national mother-to-child transmission rates and PMTCT programme impact exist.
Data from clinical trials provide estimates of efficacy,
whereas data from programmes can provide estimates of effectiveness.
Transmission rates from clinical data and programme data may appear similar,
but it does not mean they can be generalised to all programmes.
One of the key measures of effectiveness is how many
complete the PMTCT cascade: from antenatal care, counselling and testing,
through starting timely treatment, to returning to test the infant and
follow-up care for the infant.
Africa, with approximately one million live
births annually, coverage of an infant’s first immunisation at six weeks of age
was close to 100% in 2010.
The median HIV prevalence rate among pregnant women is 30%,
ranging from 17% in the Northern and Western Capes
to over 40% in KwaZulu-Natal.
PMTCT is established in close to 100% of healthcare
facilities. Distribution of ART is
decentralised and nurses are qualified to start people on ART.
To determine national and provincial MTCT rates at four to
eight weeks after birth in 2009 and early 2010, and to identify factors
contributing to MTCT, the authors undertook a national cross-sectional
facility-based survey of 9915 infant-caregiver pairs at their first infant
Africa adopted the recommended World Health
Organization (WHO) PMTCT guidelines in 2010, this
evaluation looked at the effectiveness of the South African 2008 guidelines
(Option A) in use at the time.
Data were collected from June 2010 until November 2010, from
565 facilities in all nine provinces. Dried blood specimens (DBS) from the infants were tested for
Infants were considered HIV-exposed if born to women who had
reported their HIV status and /or their DBS test was antibody-positive; their
DBS were tested for HIV infection by DNA polymerase chain reaction (PCR). In
total, 30.3% (3003) HIV-exposed infants were identified among the
infant-caregivers pairs, of whom 98.5% (2958) had a PCR test result.
One-third (33.9%; 95% CI: 32.0-35.8) of the HIV-infected
mothers received triple-drug ART and 20% (95% CI: 18.2-21.8) reported exclusive
breastfeeding. The longer and more comprehensive the treatment, the lower were
the risks for transmission.
While exclusive breastfeeding and triple-drug antiretroviral
treatment (ART) were both protective factors, unplanned pregnancies (AOR=1.7,
95% CI: 0.9-2.9) and mixed feeding (AOR=1.6 95% CI: 1.0-2.5) were risk factors
associated with MTCT. Close to two-thirds (62%, 95% CI: 59.9-64.2) of the
pregnancies among HIV-infected women were unplanned.
Sixty-two per cent of mothers did no breastfeeding. The
presenter, Thu-Ha Dinh, cautioned that this may reduce MTCT but in the long run
would increase mortality rates.
The national MTCT rate at four to eight weeks calculated
according to population live births was 3.5% (95% CI: 2.9-4.1), while the proportion of infants nationwide that were exposed to the risk of HIV infection during pregnancy or around the time of delivery was 31.4% (95% CI: 30.1-32.6).
Rates varied in the provinces from a low MTCT rate in the Northern Cape (1.7%) and a corresponding exposure rate of
15.6%, to a high of 6% and corresponding exposure rate of 31.1% in the Free State. KwaZulu-Natal, with one
of the higher exposure rates of 43.9%, had a relatively low rate of MTCT at
2.8%. These reflect standards of care as well as coverage within the respective
MTCT rates did not differ according to sociodemographic
characteristics, medical care during pregnancy and childbirth, PMTCT knowledge
Dr Dinh noted that limitations of this model included the
potential for selection bias. The results are representative of a population
who attend primary healthcare facilities. Extremely sick children needing
emergency care were excluded from this analysis.
The South African perinatal MTCT rate of under 4% showed low
levels of transmission can be achieved, but also highlights the need for
follow-up throughout the breastfeeding period.
Given the high uptake of early infant diagnosis (92%) Dr
Dinh suggested that a DBS-HIV ELISA test be offered to all infants at routine
Contrary to previous understanding, delivery by caesarean
section and having a doctor at delivery were not better options than vaginal
delivery or having only a midwife or other healthcare worker in attendance.
Thu-Ha Dinh concluded reducing both unplanned pregnancies
among HIV-infection women and mixed feeding have the potential to further
reduce MTCT rates.