Weight Loss and Mortality in People Living with HIV: a Systematic Review and Meta-Analysis

Abstract This project was designed to evaluate whether people living with HIV (PLHIV) on HAART, hospitalized, with weight loss, have higher mortality than those hospitalized using HAART and without weight loss. It was a meta-analysis of prospective cohort studies published in English, Spanish or Portuguese. We searched MedLine, Embase, and LILACS and conducted the report according to MOOSE. We analyzed methodological quality and risk of bias by the Joanna Briggs Institute's Critical Assessment Tool for Cohort Studies, used the risk ratio (RR) to calculate the probability of hospitalized PLHIV who lost weight evolve to death, applied the random effect model, and created the funnel chart. Ten of the 670 studies identified were included for analysis, with a total of 1,637 PLHIV. We used the inverse test of variances estimated by the Mantel-Haenszel method, considering a 95% confidence interval, heterogeneity (I2), total effect size (Z), and a significance value of p < 0.05. The risk of hospitalized PLHIV who lost weight dying was 1.5 times greater than that of hospitalized PLHIV who did not lose weight (RR: 1.50, 95% CI: 1.03, 2.19, p = 0.04). We conclude that weight is still a determinant of mortality in PLHIV, even those on HAART. Prospero registration number: CRD42020191246