Abstract P111: Untargeted Metabolomics Study And Bidirectional Mendelian Randomization Analyses Identified Novel Metabolites Mediating The Effect Of Current Smoking On Body Mass Index
Ruiyuan Zhang,Tanika N Kelly,Lydia A Bazzano,Wei Chen,Jiang He,Shengxu Li,Changwei Li
DOI: https://doi.org/10.1161/circ.143.suppl_1.p111
IF: 37.8
2021-05-25
Circulation
Abstract:Introduction: Many studies identified that smokers had lower body mass index (BMI) than non-smokers, but the underlying mechanisms are not well understood. Hypothesis: An untargeted metabolomics study will identify novel mechanisms underlying smoking associated weight loss. Methods: We performed cross-sectional analyses among all 248 current smokers and 637 never smokers participated in the Bogalusa Heart Study 2016 survey. After stringent quality control and normalization, 1,202 plasma metabolites were assessed for mediation effects on smoking-BMI association after controlling for age, gender, race, education, drinking and physical activity. The collective mediation effects of metabolites were estimated by comparing the smoking-BMI associations before and after controlling for all significant metabolites after Bonferroni correction. Bidirectional Mendelian randomization analyses were then performed on a subset of participants (n=654) to delineate causal directions among smoking, BMI, and significant metabolites. We constructed multi-locus genetic risk scores (GRS) for smoking and BMI based on findings from published GWASs with a sample size of more than 250,000. Multivariable linear regression controlling for the same covariates as in the mediation analyses was used to estimate associations between smoking- and BMI-GRSs and the significant metabolites. Results: Compared to never smokers, current smokers had a 3.31 kg/m 2 (95% confidence interval [CI]: 2.15-4.47) lower BMI after adjusting for all covariables. A total of 40 metabolites significantly mediated and explained 90.7% of the smoking-BMI association. Seven of the significant metabolites were associated with smoking GRS but not BMI GRS, suggesting a true mediation role in the smoking-BMI association. The seven metabolites, including homoarginine (mediation beta=-0.49 kg/m 2 , 95% CI: -0.78 to -0.26), 1-oleoyl-GPE (18:1) (mediation beta=-0.44 kg/m 2 , 95% CI: -0.72 to -0.22), 1-linoleoyl-GPE (18:2)* (mediation beta=-0.59 kg/m 2 , 95% CI: -0.90 to -0.32), beta-cryptoxanthin (mediation beta=0.53 kg/m 2 , 95% CI: 0.27 to 0.83), 1-stearoyl-2-arachidonoyl-GPE (18:0/20:4) (mediation beta=0.39 kg/m 2 , 95% CI: 0.18 to 0.65), sphingomyelin (d18:2/21:0, d16:2/23:0)* (mediation beta=-0.54 kg/m 2 , 95% CI: -0.84 to -0.28), and sphingomyelin (d17:2/16:0, d18:2/15:0)* (mediation beta=-0.92 kg/m 2 , 95% CI: -1.30 to -0.59), mediated 36.3% of the association between current smoking and BMI. We also identified 6 metabolites significantly associated with BMI GRS but not with smoking GRS, and 2 significantly associated with both smoking- and BMI-GRSs. Conclusions: We identified 7 plasma metabolites that mediated a large proportion of the association between smoking and BMI.
cardiac & cardiovascular systems,peripheral vascular disease