There is growing emphasis on testing HIV-exposed infants at birth to detect intrauterine infections. However concerns have been raised that implementation of birth testing may reduce uptake of routine early infant diagnostic testing (EID) at around 6 weeks of age. As there are no data on this question, we examined the association between birth PCR and subsequent routine EID testing in the Western Cape province of South Africa (SA), a setting where 6-week EID test coverage is estimated to reach >70% of all HIV-exposed infants.
Data on all HIV PCR tests conducted in the province (approximately 14 000 per year) were accessed from the SA National Health Laboratory Service. Infants with birth PCR (defined as testing in the first 7 days of life) and subsequent EID testing (8 to 182 days of age) were linked probabilistically using Finegrained Record Integration and Linkage Tool (Atlanta, USA); a range of sensitivity analyses were used to maximise linkage of birth tests to subsequent EID testing. Two periods of birth testing policies were compared: discretionary testing by clinicians (DT; Jan 2009- March 2014) and testing of all ‘high risk’ pregnancies based on duration of maternal ART, documented HIV viraemia, and related factors (HR; April 2014-June 2015).
Overall 3322 newborns received birth testing (80% within three days of life), comprising 3% of all HIV PCR tests conducted in infants. The number of tests increased >20-fold after the start of HR testing policies. Throughout most birth tests were undertaken in obstetric hospitals, though the proportion of birth tests from primary care facilities increased from 11% under DT to 33% under HR testing (Table). Birth PCR positivity rates decreased from 6% in 2009 to 1.6% in 2015 (p=0.001). Of children with negative birth PCR results, only 49% had any evidence of a follow-up EID test. This proportion decreased to 43% when restricted to a window around prescribed EID testing, but was stable over time with no difference in follow-up testing rates or mean age at retesting under DT vs HR testing policies (p=0.506 and 0.112, respectively).
Changing birth testing policies have led to dramatic increases in birth tests. However follow-up EID resting rates in infants testing PCR negative at birth are substantially lower than local estimates of 6-week EID coverage. These data suggest that implementation of birth testing will require particular care to avoid undermining postpartum EID services.