Attrition one year after starting antiretroviral therapy before and after the programmatic implementation of HIV “Treat All” in Sub-Saharan Africa: a systematic review and meta- analysis

Introduction: Evidence on the real-world effects of “Treat All” on attrition has not been summarised. We conducted a systematic review and meta-analysis to compare attrition 12 months after antiretroviral therapy (ART) initiation, before and after “Treat All” was implemented in Sub-Saharan Africa. We also describe predictors of attrition. Methods: We searched Embase, Google Scholar, PubMed, and Web of Science in July 2020 and created alerts up to the end of March 2022. We also searched for preprints and conference abstracts. Two co-authors screened and selected the articles. Risk of bias was assessed using the modified Newcastle-Ottawa Scale. We extracted and tabulated data on study characteristics, attrition 12 months after ART initiation, and predictors of attrition. We estimated a pooled estimate for attrition using random-effects meta-analysis. Results: Seven articles and one conference abstract (eight studies) out of 5782 screened records were included in the meta-analysis. The random-effects adjusted pooled risk ratio (RR) comparing attrition before and after “Treat All” 12 months after ART initiation was not significant [RR = 1.08 (95% Confidence interval (CI): 0.97–1.21)] with 82% heterogeneity (I 2 ). Being a pregnant or breastfeeding woman, starting ART with advanced HIV and within the same week were reported as risk factors for attrition both before and after “Treat All”. Conclusions: We found no significant difference in attrition before and after “Treat All” one year after ART initiation. While “Treat All” is being widely implemented, differentiated approaches to enhance retention should be prioritised for those subgroups at risk of attrition. PROSPERO Number CRD42020191582