Temporal Relationship Between Obesity and Cardiometabolic Risk Markers: an Integrated Two-Sample Bidirectional Mendelian Randomization Study and Cross-Lagged Panel Analysis
Lu Wang,Qiong Wu,Sitong Liu,Xuhui Zhang,Yimin Zhu
DOI: https://doi.org/10.2139/ssrn.4720752
2024-01-01
Abstract:Background: Although there is a significant correlation between obesity and cardiometabolic risk factors, it remains unclear whether these associations indicate causal relationships in either direction. This study aims to investigate the potential temporal relationship between obesity and markers of cardiometabolic risk. Methods: A two-sample bidirectional Mendelian randomization (MR) analysis was conducted using the inverse-variance weighted method and other supplementary MR approaches. These approaches were based on a summary of genome-wide association studies, with the aim of evaluating the temporal relationship between obesity and cardiometabolic risk markers, including glucose, systolic blood pressure (SBP), diastolic blood pressure (DBP), triglyceride (TG), and high-density lipoprotein cholesterol (HDL-C). Additionally, a cross-lagged panel analysis was conducted in the China Health and Retirement Longitudinal Study cohort (n = 3890), measuring body mass index (BMI), fasting plasma glucose (FPG), SBP, DBP, TG, and HDL-C twice at two-time points. Findings: The results of the MR analysis indicated a potential causal association between BMI and increased levels of glucose (b =0.148, 95%CI: 0.128–0.168, P<0.001), SBP (b =0.138, 95%CI: 0.115–0.161, P<0.001), DBP (b =0.221, 95%CI: 0.197–0.245, P<0.001), and TG (b =0.225, 95%CI: 0.193–0.258, P<0.001), along with a decreased risk of HDL-C (b=-0.294, 95%CI: -0.332 – -0.257, P<0.001). Conversely, in the reverse direction, the results suggested a negative correlation between HDL-C and BMI (b =-0.029, 95%CI: -0.043 – -0.014, P<0.001), whilst no significant causal effects of other cardiometabolic risk markers on BMI were identified (all P>0.05). Sensitivity analyses confirmed the robustness of these findings. The outcomes of the cross-lagged panel analysis aligned with the MR results. After adjusting for potential confounders, the panel coefficients (b1) from baseline BMI to follow-up FPG, SBP, DBP, and TG were significantly greater than the panel coefficients (b2) from baseline FPG, SBP, DBP, and TG to follow-up BMI (b1=0.098 versus b2=0.001, Pdifference<0.001 for FPG; b1=0.092 versus b2=0.014, Pdifference<0.05 for SBP; b1=0.079 versus b2=0.030, Pdifference<0.05 for DBP; b1=0.130 versus b2=0.041. Pdifference<0.001 for TG). Additionally, there was no significant difference between the panel coefficient from baseline BMI to follow-up HDL-C and the panel coefficient from baseline HDL-C to follow-up BMI (b1=-0.075 versus b2=-0.049, Pdifference=0.250). Interpretation: Our findings indicate a potential unidirectional causal association from BMI to glucose, blood pressure, and TG. Furthermore, a bidirectional relationship was observed between BMI and HDL-C. Funding: This study was partly funded by National Key Research and Development Program of China (2017YFC0907004). We sincerely thank the compilers of the GWAS summary dataset and the participants and investigators of CHARLS for data collection and management of data resources. We also would like to thank the grants from National Key Research and Development Program of China (2017YFC0907004).Declaration of Interest: The authors declare that they have no competing interests.