Prevalence of HIV-Seropositivity and Associated Impact on Mortality among Injured Patients from Low-and Middle-Income Countries: A Systematic Review and Meta-Analysis.

Aluisio, Adam R, Rege, Soham, Stewart, Barclay T, Kinuthia, John, Levine, Adam C, Mello, Michael J, Farquhar, Carey

Although HIV and injury contribute substantially to disease burdens in low-and middle-income countries (LMIC), their intersection is poorly characterized. This systematic review assessed the prevalence and associated mortality impact of HIV-seropositivity among injured patients in LMIC. A systematic search of PubMed, EMBASE, Global Health, CINAHL, POPLINE and Cochrane databases through August 2016 was performed. Prospective and cross-sectional reports of injured patients from LMIC that evaluated HIV-serostatus were included. Two reviewers identified eligible records (kappa=0.83); quality was assessed using GRADE criteria. HIV-seroprevalence and mortality risks were summarized; pooled estimates were calculated using random-effects models with heterogeneity assessed. Of 472 retrieved records sixteen met inclusion. All reports were low/very low quality and derived from sub-Saharan Africa. HIV-serostatus was available for 3,994 patients. Individual report and pooled HIV-seroprevalence estimates were uniformly greater than temporally matched national statistics (range: 4.5-35.0%). Pooled reports from South Africa were three-fold greater than matched national prevalence (32.0%, 95% CI, 28.0-37.0%). Mortality data were available for 1,398 patients. Heterogeneity precluded pooled mortality analysis. Among individual reports, 66.7% demonstrated significantly increased relative risks (RR) of death; none found reduced risk of death among HIV-seropositive patients. Increased mortality risk among HIV-seropositive patients ranged from 1.86 (95% CI, 1.11-3.09) in Malawi to 10.7 (95% CI, 1.32-86.1) in South Africa. Available data indicate that HIV-seropositivity among the injured is high relative to national rates and may increase mortality, suggesting that integrated HIV-injury programming could be beneficial. Given the data limitations, further study of the HIV-injury intersection is crucially needed.