Abstract Context Some studies have surprisingly indicated that serum adiponectin level is positively related to mortality rate, thus casting doubts on its role as a therapeutic target for cardiovascular disease. Objective To summarize evidence about direction, strength, and modulators of this controversial association. Methods MEDLINE, Web of Science, CINHAL, Cochrane Library, and Scopus databases were searched from their inception dates through June 2018 for English-language prospective studies reporting the association between adiponectin and all-cause or cardiovascular mortality. Two investigators independently extracted data and assessed study quality using standard criteria following the Preferred Reporting Items for Systematic Reviews and Meta-analyses and The Newcastle-Ottawa Scale. Pooled hazard ratios (HRs) and 95% CIs were derived using fixed- or random-effects models when appropriate, and results were expressed to a 1-SD increment of adiponectin. Results We identified 55 studies (n = 61,676 subjects) with all-cause mortality data and 28 (n = 43,979 subjects) studies with cardiovascular mortality data. Pooled HRs were 1.24 (1.17-1.31) and 1.28 (1.19-1.37) for all-cause and cardiovascular mortality, respectively. Similar results were obtained for high-molecular-weight adiponectin. When meta-analyses were restricted to studies reporting data on natriuretic peptides, reductions of 43% and 28% on a log scale of these respective associations were observed after adjusting for natriuretic peptides. Conclusions Our results point strongly to a paradoxical association between high adiponectin levels and increased mortality rate, which is partly modulated by natriuretic peptides.