The effect of the Xpert HIV-1 Qual test on early infant diagnosis of HIV in Myanmar and Papua New Guinea: a pragmatic, cluster-randomised, stepped-wedge, open-label trial

Despite proven benefits for child health, coverage of early infant diagnosis of HIV remains suboptimal in many settings. We aimed to assess the effect of a point-of-care early infant diagnosis test on time-to-results communication for infants vertically exposed to HIV.This pragmatic, cluster-randomised, stepped-wedge, open-label trial assessed the effect of the Xpert HIV-1 Qual early infant diagnosis test (Cepheid) on time-to-results communication, compared with standard care laboratory-based testing of dried blood spots using PCR. Hospitals were the unit of randomisation for one-way crossover from control to intervention phase. Each site had between 1 month and 10 months of control phase before transitioning to the intervention, with a total of 33 hospital-months in the control phase and 45 hospital-months in the intervention phase. We enrolled infants vertically exposed to HIV at six public hospitals: four in Myanmar and two in Papua New Guinea. Infants had to have mothers with confirmed HIV infection, be younger than 28 days, and required HIV testing to be eligible for enrolment. Health-care facilities providing prevention of vertical transmission services were eligible for participation. The primary outcome was communication of early infant diagnosis results to the infant's caregiver by 3 months of age, assessed by intention to treat. This completed trial was registered with the Australian and New Zealand Clinical Trials Registry, 12616000734460.In Myanmar, recruitment took place between Oct 1, 2016, and June 30, 2018; in Papua New Guinea, recruitment was between Dec 1, 2016, and Aug 31, 2018. A total of 393 caregiver-infant pairs were enrolled in the study across both countries. Independent of study time, the Xpert test reduced time to early infant diagnosis results communication by 60%, compared with the standard of care (adjusted time ratio 0·40, 95% CI 0·29-0·53, p<0·0001). In the control phase, two (2%) of 102 study participants received an early infant diagnosis test result by 3 months of age compared with 214 (74%) of 291 in the intervention phase. No safety and adverse events were reported related to the diagnostic testing intervention.This study reinforces the importance of scaling up point-of-care early infant diagnosis testing in resource-constrained and low HIV-prevalence settings, typical of the UNICEF East Asia and Pacific region.National Health and Medical Research Council of Australia.