The Adiponectin-Mortality Paradox—A Systematic Review and Meta-analysis

Abstract

Surprisingly, several, though not all, studies have shown a positive relationship between serum adiponectin (ADPN) and mortality rate. In order to deeply address direction and strength of this paradoxical association, we carried out a systematic review and meta-analysis (MA) of prospective studies present in MEDLINE, Cochrane Library and Scopus by December, 1st 2017, with ADPN being the exposure and all-cause (AC) and cardiovascular (CV) mortality the outcomes. We performed random effect MA to pool individual hazard ratios (HRs; 95% CIs). To address the role of Natriuretic Peptides (NPs), known to increase both ADPN expression and mortality rate, a MA focused only on studies reporting HRs before and after NPs adjustment was also carried out. We assessed study heterogeneity by Q statistic and then performed meta-regressions considering male percentage, age, BMI, eGFR and clinical setting (general population, type 2 diabetes, CV disease, dialysis) as study level covariates. We identified 56 (n=64,775 subjects) and 30 (n=48,439) studies for AC and CV mortality, respectively. Heterogeneity across studies was observed for both outcomes (Q-test p<0.01). HRs, expressed as 1 SD increment of total ADPN, were 1.24 (1.17-1.30) and 1.28 (1.20-1.37) for AC and CV mortality, respectively. MA restricted to the 12 studies (n=19,010) reporting NPs data showed HRs for AC mortality equal to 1.24 (1.16-1.34) and 1.13 (1.06-1.21) without and with NPs adjustment, respectively (p<0.between HRs, when their correlation was >0.1). For CV mortality (6 studies, n=15,252), HRs were 1.21 (1.10-1.33) and 1.15 (1.06-1.24) without and with NPs adjustment, respectively (p<0.between HRs, when their correlation was >0.8). Male percentage, age, BMI, eGFR and clinical setting did not explain the observed heterogeneity among studies. Our results confirm the existence of a paradoxical association between high ADPN levels and increased mortality rate. Such paradox seems to be only partly explained by NPs levels. Disclosure M. Scarale: None. A. Fontana: None. V. Trischitta: None. M. Copetti: None. C. Menzaghi: None.