Background Many HIV-positive children in low-income and middle-income countries remain undiagnosed. Although HIV testing in children at health facilities is recommended by WHO, it is not well implemented. This systematic review and meta-analysis examines the case-finding benefit of HIV screening in children aged 0-5 years in low-income and middle-income countries. Methods We did this systematic review and meta-analysis in accordance with an a-priori protocol. We searched PubMed, MEDLINE, WHO Global Index Medicus, Web of Science, Medecins Sans Frontieres, Cochrane, Embase, CABS Abstracts, and LILACS databases for articles published between Jan 1, 2004, and April 30, 2016, that reported the quantitative prevalence of HIV detected through screening in four key contexts (paediatric inpatient settings, paediatric outpatient settings, nutrition centres, and expanded programme on immunisation centres) in paediatric populations in low-income and middle-income countries. Articles were identified and data were extracted in duplicate. The primary outcome was HIV prevalence, for which we used a DerSimonian-Laird random-effects meta-analysis to pool prevalence data and 95% CIs. We did stratified analyses according to geographical context and testing strategy. This study is registered with PROSPERO, number CRD42014014372. Findings Our search found 2996 studies, of which 26 met the inclusion criteria. Paediatric HIV prevalence across all settings was 15.6% (95% CI 11.8-19.5). HIV prevalence by setting was highest in paediatric inpatient settings (21.1%, 95% CI 14.9-27.3), followed by nutrition centres (13.1%, 95% CI 3.4-22.7), expanded programme on immunisation centres (3.3%, 95% CI 0-6.9), and paediatric outpatient settings (2.7%, 95% CI 0.3-5.2). Universal testing and testing triggered by symptoms had similar diagnostic yield in the inpatient setting (21.3%, 95% CI 11.6-31.0 in triggered testing vs 20.9%, 95% CI 13.5-28.3 in universal testing). Interpretation HIV testing in paediatric populations in low-income and middle-income countries outside the context of prevention of mother-to-child transmission programmes provides an important opportunity to identify HIV-positive children. For countries wishing to prioritise interventions, the highest diagnostic yields were obtained from inpatient wards and nutrition centres. Universal testing might be the preferred approach since it did not have a substantially lower diagnostic yield than triggered testing
