Type 2 Diabetes and Adiposity Induce Different Lipid Profile Disorders: A Mendelian Randomization Analysis

Type 2 Diabetes and Adiposity Induce Different Lipid Profile Disorders: A Mendelian Randomization... Abstract Context Type 2 diabetes and obesity often coexist, so it is difficult to judge whether diabetes or obesity induce certain types of hyperlipidemia due to mutual confounds and reverse causation. We used Mendelian randomization analyses to explore the causal relationships of diabetes and adiposity with lipid profiles. Design, Setting, and Main Outcome Measures From 23 sites in East China, 9798 participants were enrolled during 2014 to 2016. We calculated two weighted genetic risk scores as instrumental variables for type 2 diabetes and body mass index (BMI). These scores were used to measure the causal relationships of diabetes and BMI with lipid profiles that included total cholesterol, high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), and triglycerides (TGs). Results The causal regression coefficients (βIV) of genetically determined diabetes for the total cholesterol, LDL-C, and log10TG were 0.130 [95% confidence interval (CI): 0.020, 0.240; P = 0.014], 0.125 (96% CI: 0.041, 0.209; P = 0.001), and 0.019 (95% CI: –0.001, 0.039; P = 0.055), respectively. The βIV for HDL-C was –0.008 (95% CI: –0.032. 0.016), which was not significant (P = 0.699). The causal regression coefficients of a genetically determined 10 kg/m2 increase in BMI for HDL-C and log10TG were –0.409 (96% CI: –0.698, –0.120; P = 0.004) and 0.227 (95% CI: 0.039, 0.415; P = 0.026), respectively. The βIVs for TGs and LDL-C were not significant. Conclusions This study has provided evidence for the biologically plausible causal effects of diabetes and adiposity by BMI on different elements of the lipid profile using Mendelian randomization analyses. Atherosclerotic cardiovascular disease (CVD) is the leading cause of death worldwide (1). The incidence and prevalence of CVD are continuously increasing in parallel with the increase in dyslipidemia, especially in low- and middle-income countries such as China (1). According to the Coronary Heart Disease Policy Model-China, total cholesterol (TC) increases of 0.58 mmol/L (22.4 mg/dL) and 0.55 mmol/L (21.6 mg/dL) in Chinese men and women, respectively, from 2010 to 2030 would lead to approximately 960 million new CVD events in total, which represents the largest increase in CVD events among all modeled risk factors, including diabetes and body mass index (BMI) (2). Many well-designed epidemiological studies have demonstrated the associations of diabetes and obesity with lipid profile (3–5). Previous studies have reported that the most common pattern of dyslipidemia in patients with type 2 diabetes includes elevated triglyceride (TG) levels and decreased high-density lipoprotein cholesterol (HDL-C) levels (6). The concentration of low-density lipoprotein cholesterol (LDL-C) exhibits no significant increase (6). In obesity, hypercholesterolemia, hypertriglyceridemia, and high LDL-C and low HDL-C levels are very common (5). However, diabetes and obesity have often coexisted in the samples of previous studies, so it is difficult to judge whether diabetes or obesity induces certain types of hyperlipidemia due to mutual confounds, reverse causation, and other issues with conventional observational studies that ultimately lead to imprecision in the estimation of both the directions and magnitudes of the effects (7). An alternative to classical observational studies are Mendelian randomization (MR) studies. MR uses genetic variants associated with an intermediate phenotype (in the current study, diabetes and BMI) as instrumental variables (IVs) in nonexperimental data to make causal inferences about the effect of an exposure on an outcome (Fig. 1) (8). Because genetic variants are assumed to be randomly distributed within a population and exist before the outcome, the IV is regarded as independent of the confounders that affect the exposure-outcome relationship and guide definite causal direction (9). In this study, if type 2 diabetes and BMI causally induce certain types of hyperlipidemia, then genetic variants associated with type 2 diabetes and BMI could be expected to affect certain types of hyperlipidemia. Figure 1. View largeDownload slide Study design and tested associations. BMI_GRS, body mass index genetic risk score; DM_GRS, diabetes mellitus genetic risk score. Figure 1. View largeDownload slide Study design and tested associations. BMI_GRS, body mass index genetic risk score; DM_GRS, diabetes mellitus genetic risk score. Based on the large community-based sample of Chinese participants from the Survey on Prevalence in East China for metabolic diseases and risk factors (SPECT)-China study (www.chictr.org.cn; registration no. ChiCTR-ECS-14005052), we performed MR analyses to explore the causal associations of diabetes and BMI with lipid profiles, including TC, LDL-C, HDL-C, and TG. Diabetes mellitus genetic risk scores (DM_GRSs) and BMI genetic risk scores (BMI_GRSs) were constructed to represent genetic susceptibility, and we evaluated the causal links of genetically determined diabetes and genetically determined BMI with lipid profiles. Methods Study design and participants Figure 1 presents the study design. We analyzed the observed associations between the measured variables (residual confounds and reverse causation) and the genetic associations between genotypes and the measured variables (no residual confounds or reverse causation). First, we assessed the association between type 2 diabetes mellitus–related genetic variants and present diabetes (Fig. 1A) and the relation between BMI-related genetic variants and measured BMI (Fig. 1A). Second, we measured the effects of genetic variants associated with type 2 diabetes and BMI on the changes in lipid profiles, including TC, LDL-C, HDL-C, and TG (Fig. 1C). Third, we performed observational assessments of the relation between present diabetes and lipid profile and between BMI and lipid profile (Fig. 1B). The data were taken from the ongoing SPECT-China study, which is a cross-sectional investigation of the prevalence of metabolic diseases and risk factors in East China. Recruitment and enrollment have previously been described in detail (10–12). Chinese citizens ≥18 years old who had lived in their current area for ≥6 months were included. From 2014 to 2016, 12,666 subjects who were 18 to 93 years old from 23 sites in the Shanghai, Zhejiang, Jiangsu, Anhui, and Jiangxi provinces were recruited into the SPECT-China study. Among these participants, genotype information was available for 10,516 (83.0%). We excluded participants with missing information on more than two single-nucleotide polymorphism (SNP) genotypes (n = 297), missing lipid profile values (n = 3), and missing measured BMI (n = 242) and those who were using lipid-lowering drugs (n = 176). In total, 9,798 participants were involved in the final analyses. The study protocol was approved by the Ethics Committee of Shanghai Ninth People’s Hospital, Shanghai JiaoTong University School of Medicine. The study protocol conformed to the ethical guidelines of the 1975 Declaration of Helsinki as reflected in the a priori approval by the appropriate institutional review committee. Informed consent was obtained from all participants included in the study. Procedures A single-assessment protocol was adopted by trained staff during the interviews and the collection of biological specimens at each study site. Blood samples were obtained between 7:00 am and 10:00 am after fasting for at least 8 hours. The blood was refrigerated immediately after phlebotomy. After 2 to 4 hours, the blood was centrifuged, and the serum was aliquoted and frozen in a central laboratory. Glycated hemoglobin (HbA1c) was measured by high-performance liquid chromatography (MQ-2000PT; Medconn, Shanghai, China). Fasting plasma glucose (FPG), TG, TC, HDL-C, and LDL-C were measured with a Beckman Coulter AU680 (Brea, CA). BMI was calculated as the weight in kilograms divided by the height in meters squared. Body weight, height, and blood pressure were measured using standard methods as described previously (13). Diabetes was defined by the following criteria (1): a self-reported previous diagnosis by a health care professional (13), (2) an FPG level of 7.0 mmol/L or higher, or (3) an HbA1c concentration of 6.5% or more. Hypertension was assessed as a systolic blood pressure ≥140 mm Hg, a diastolic blood pressure ≥90 mm Hg, or a self-reported previous diagnosis of hypertension by a physician. The 16 study sites were categorized into high and low economic status according to the gross domestic product per capita of the whole nation in 2013 (10). DNA was extracted from white bloods cells using a blood genomic DNA extraction kit (DP603; TIANGEN BIOTECH CO, LTD, Beijing, China) on an automated nucleic acid extraction instrument (YOSE-S32; TIANGEN BIOTECH CO, LTD, Beijing, China). Specific assays were designed using Geneious Pro (version 4.8.3; www.geneious.com/, Auckland, New Zealand). Mass determination was performed with JUNO, and data acquisition was performed using Fluidigm SNP Genotyping Analysis version 4.1.3 software (Fluidigm Corporation, San Francisco, CA) (14). The researchers were masked to the data regarding present diabetes, BMI, and phenotype. The call rates of all SNPs were higher than 98%. For the selection of the diabetes- and BMI-related SNPs, we mainly considered the SNPs that have been identified in genome-wide association studies or replicated in Asians and then successfully used in two Chinese MR studies (15, 16). For diabetes, we a priori selected 18 SNPs that were identified and validated in meta-analyses including genome-wide association studies in East Asians (17, 18). A previous Chinese MR study (15) included 34 SNPs. To minimize the pleiotropic effects, we further selected the 18 SNPs that exhibited no significant associations with BMI or systolic and diastolic blood pressure in that study (15). These SNPs were TCF2 (HNF1B) rs4430796, GCC1/PAX4 rs6467136, MAEA rs6815464, GLIS3 rs7041847, C2CD4A/C2CD4B rs7172432, TCF7L2 rs7903146, ZFAND3 rs9470794, CDKN2A/B rs10811661, CDC123/CAMK1D rs12779790, PROX1 rs340874, SRR rs391300, PSMD6 rs831571, JAZF1 rs864745, TP53INP1 rs896854, SPRY2 rs1359790, PPARG rs1801282, KCNQ1 rs2237892, and IRS1 rs2943641. For BMI, 14 SNPs from a previous MR study (16) that were selected from genome-wide association studies in East Asians (19, 20) were adopted. These SNPs included TFAP2B rs4715210, MAP2K5 rs4776970, ADCY3-DNAJC27 rs6545814, MC4R rs6567160, CDKAL1 rs9356744, GNPDA2 rs10938397, SMC5-KLF9 rs11142387, GIPR-QPCTL rs11671664, GP2 rs12597579, FTO rs17817449, BDNF rs6265, PCSK1 rs261967, SEC16B rs574367, and PAX6 rs652722. There were no linkage disequilibrium relationships (r2 = 0) among the previous diabetes and BMI SNPs (15, 16). The full list of each SNP is presented in Supplemental Table 1. Statistical analysis The data analyses were performed using IBM SPSS Statistics version 22 (IBM Corporation, Armonk, NY). All analyses were two sided. A P value <0.05 indicated significance. Continuous variables are expressed as the mean ± the standard deviation (SD), and categorical variables are expressed as percentages (%). Serum TG was logarithmically transformed before the analysis. The P values for trends of the characteristics according to the genetic risk score (GRS) quintiles were calculated with one-way analysis of variance and χ2 tests. The additive genetic model for each SNP (coded as 0-2) was used to construct the DM_GRS and BMI_GRS. The scores were defined as the sum of the number of BMI-increasing and diabetes risk-increasing alleles at each locus multiplied by the respective β coefficient as reported from the meta-analysis. Next, the total scores were divided by the average effect sizes of these SNPs. Those who were missing one or two SNPs were imputed by the regression method using covariates. The missing rates of one or two diabetes- and BMI-related SNPs were 16.4% and 15.5%, respectively. There were also 82 missing blood pressure values, which were not imputed. For our analyses of the association between DM_GRS and present diabetes and that between BMI_GRS and measured BMI (Fig. 1A), we used logistic and linear regression analyses. The data were adjusted for age, sex, BMI (only in the DM_GRS model), diabetes (only in the BMI_GRS model), hypertension, and economic status. To assess the reliabilities and strengths of the GRSs as instruments, the F statistic was adopted (F = [R2 × (n–2)]/[1–R2]) (21). An F value greater than 10 indicates that the instrument is sufficiently strong as an IV, and the R2 is the proportion of variation in the respective phenotype that is explained by the genotype (22). We used regression analyses to calculate the β coefficients for the associations of present diabetes and BMI with lipid profile, including TC, LDL-C, HDL-C, and TG (Fig. 1B). Then, we also investigated the associations of the DM_GRS and BMI_GRS with lipid profile (Fig. 1C) by linear regressions that were adjusted for age, sex, BMI (only in the DM_GRS model), diabetes (only in the BMI_GRS model), hypertension, and economic status. The association of each individual SNP with lipid profile was analyzed by linear regression analysis with an unadjusted model. Next, we calculated the IV estimates of the genetically determined β coefficients with the Wald-type estimator (7, 22). For abnormal lipid profile trends due to present diabetes and increased BMI, the computational formulae were βIV(DM-lipid) = βDM_GRS-lipid/βDM_GRS-DM and βIV(BMI-lipid) = βBMI_GRS-lipid/βBMI_GRS-BMI, respectively. The standard errors and confidence intervals (CIs) for the IV estimators were estimated by the δ method. The formulae are: SEIV=abs(βIV)(SEGRS_exposureβGRS_exposure)2+(SEGRS_eoutcomeβGRS_outcome)2 95% CIIV=βIV±1.96×SEIV According to a previous study, a 10 kg/m2 BMI increment was used because this increment roughly reflects a shift from normal weight to obesity. Thus, to increase the parity between the observational and genetic estimates, the effect of the BMI_GRS was extrapolated to correspond to a 10 kg/m2 increase in BMI for the genetic analyses. A GRS as an IV should meet three assumptions (23). First, the GRS should be associated with the exposure. This assumption is confirmed in the aforementioned Fig. 1A. Second, the GRS should be independent of the confounding factors. Thus, we also measured the associations of each individual SNP and GRS with BMI, blood pressure, FPG, HbA1c, and age. Third, the two GRSs should have effects on lipid profile that are mediated solely through diabetes and BMI (i.e., there are no pleiotropic effects) (23). However, this is only an idealized view and could only be verified to a limited extent. We additionally adjusted the HbA1c or BMI in the previous models to observe the changes in the associations of the DM_GRS or BMI_GRS with the lipid profiles. In the sensitivity analysis, because the BMI_GRS exhibited significant associations with age and hypertension, in consideration of pleiotropy, we constructed a BMI_GRS12SNP that excluded the most obvious loci that may have been involved in the regulation of other traits (i.e., FTO rs17817449 and MC4R rs6567160) (24). Results Associations of SNPs with exposure and outcome We tested the associations of the diabetes- and BMI-related SNPs with exposure (present diabetes and measured BMI) and outcome (lipid profile). The unstandardized coefficients (95% CIs) and odds ratios (ORs) are presented in Tables 1 and 2 and Supplemental Figs. 1 and 2. Among the 18 diabetes-related SNPs, seven SNPs located at the GLIS3, CDKN2A/B, CDC123/CAMK1D, TP53INP1, PPARG, KCNQ1, and IRS1 loci were significantly associated with present diabetes (Supplemental Fig. 1). Among the 14 BMI-related SNPs, eight SNPs located at the MAP2K5, ADCY3-DNAJC27, MC4R, CDKAL1, GIPR-QPCTL, FTO, BDNF, and SEC16B loci were significantly associated with BMI (Supplemental Fig. 2). Table 1. Association of the 18 Diabetes-Related SNPs With Lipid Profile Gene SNP TC LDL-C HDL-C log10TG TCF2 (HNF1B) rs4430796 0.047 (0.009, 0.085) 0.031 (0.004, 0.059) −0.002 (−0.013, 0.009) 0.003 (−0.005, 0.011) GCC1/PAX4 rs6467136 0.077 (0.028, 0.125) 0.042 (0.008, 0.077) 0.014 (0.001, 0.028) 0.003 (−0.007, 0.014) MAEA rs6815464 −0.016 (−0.052, 0.019) −0.009 (−0.035, 0.016) 0.007 (−0.003, 0.017) 0.001 (−0.006, 0.009) GLIS3 rs7041847 0.031 (−0.003, 0.066) 0.027 (0.002, 0.051) 0.005 (−0.005, 0.014) 0.001 (−0.006, 0.008) C2CD4A/C2CD4B rs7172432 0.001 (−0.034, 0.036) 0.010 (−0.015, 0.035) −0.002 (−0.012, 0.008) −0.002 (−0.010, 0.005) TCF7L2 rs7903146 0.036 (−0.056, 0.129) 0.043 (−0.024, 0.109) 0.006 (−0.020, 0.032) −0.006 (−0.025, 0.014) ZFAND3 rs9470794 −0.017 (−0.054, 0.020) −0.002 (−0.029, 0.024) −0.007 (−0.018, 0.004) 0.004 (−0.003, 0.012) CDKN2A/B rs10811661 0.018 (−0.016, 0.053) 0.017 (−0.007, 0.042) 0.004 (−0.006, 0.014) 0.004 (−0.003, 0.011) CDC123/CAMK1D rs12779790 0.050 (0.004, 0.096) 0.017 (−0.016, 0.050) 0.003 (−0.010, 0.016) 0.003 (−0.007, 0.012) PROX1 rs340874 0.008 (−0.028, 0.043) 0.005 (−0.021, 0.030) 0.008 (−0.002, 0.018) 0.002 (−0.005, 0.010) SRR rs391300 0.004 (−0.033, 0.042) 0.011 (−0.015, 0.038) −0.001 (−0.011, 0.010) 0.000 (−0.008, 0.008) PSMD6 rs831571 0.038 (0.001, 0.074) 0.038 (0.012, 0.064) 0.003 (−0.007, 0.014) 0.003 (−0.004, 0.011) JAZF1 rs864745 0.043 (0.003, 0.084) 0.032 (0.003, 0.061) −0.001 (−0.013, 0.010) 0.000 (−0.004, 0.011) TP53INP1 rs896854 0.016 (−0.022, 0.053) 0.006 (−0.021, 0.033) 0.003 (−0.008, 0.014) 0.004 (−0.004, 0.012) SPRY2 rs1359790 −0.011 (−0.050, 0.028) −0.014 (−0.042, 0.014) −0.005 (−0.016, 0.006) −0.007 (−0.016, 0.001) PPARG rs1801282 0.013 (−0.061, 0.087) 0.011 (−0.042, 0.064) 0.014 (−0.007, 0.035) 0.015 (0.000, 0.031) KCNQ1 rs2237892 0.014 (−0.023, 0.051) 0.014 (−0.012, 0.040) −0.002 (−0.013, 0.008) 0.001 (−0.007, 0.009) IRS1 rs2943641 0.012 (−0.057, 0.081) −0.011 (−0.061, 0.038) −0.019 (−0.039, 0.001) 0.017 (0.003, 0.032) Gene SNP TC LDL-C HDL-C log10TG TCF2 (HNF1B) rs4430796 0.047 (0.009, 0.085) 0.031 (0.004, 0.059) −0.002 (−0.013, 0.009) 0.003 (−0.005, 0.011) GCC1/PAX4 rs6467136 0.077 (0.028, 0.125) 0.042 (0.008, 0.077) 0.014 (0.001, 0.028) 0.003 (−0.007, 0.014) MAEA rs6815464 −0.016 (−0.052, 0.019) −0.009 (−0.035, 0.016) 0.007 (−0.003, 0.017) 0.001 (−0.006, 0.009) GLIS3 rs7041847 0.031 (−0.003, 0.066) 0.027 (0.002, 0.051) 0.005 (−0.005, 0.014) 0.001 (−0.006, 0.008) C2CD4A/C2CD4B rs7172432 0.001 (−0.034, 0.036) 0.010 (−0.015, 0.035) −0.002 (−0.012, 0.008) −0.002 (−0.010, 0.005) TCF7L2 rs7903146 0.036 (−0.056, 0.129) 0.043 (−0.024, 0.109) 0.006 (−0.020, 0.032) −0.006 (−0.025, 0.014) ZFAND3 rs9470794 −0.017 (−0.054, 0.020) −0.002 (−0.029, 0.024) −0.007 (−0.018, 0.004) 0.004 (−0.003, 0.012) CDKN2A/B rs10811661 0.018 (−0.016, 0.053) 0.017 (−0.007, 0.042) 0.004 (−0.006, 0.014) 0.004 (−0.003, 0.011) CDC123/CAMK1D rs12779790 0.050 (0.004, 0.096) 0.017 (−0.016, 0.050) 0.003 (−0.010, 0.016) 0.003 (−0.007, 0.012) PROX1 rs340874 0.008 (−0.028, 0.043) 0.005 (−0.021, 0.030) 0.008 (−0.002, 0.018) 0.002 (−0.005, 0.010) SRR rs391300 0.004 (−0.033, 0.042) 0.011 (−0.015, 0.038) −0.001 (−0.011, 0.010) 0.000 (−0.008, 0.008) PSMD6 rs831571 0.038 (0.001, 0.074) 0.038 (0.012, 0.064) 0.003 (−0.007, 0.014) 0.003 (−0.004, 0.011) JAZF1 rs864745 0.043 (0.003, 0.084) 0.032 (0.003, 0.061) −0.001 (−0.013, 0.010) 0.000 (−0.004, 0.011) TP53INP1 rs896854 0.016 (−0.022, 0.053) 0.006 (−0.021, 0.033) 0.003 (−0.008, 0.014) 0.004 (−0.004, 0.012) SPRY2 rs1359790 −0.011 (−0.050, 0.028) −0.014 (−0.042, 0.014) −0.005 (−0.016, 0.006) −0.007 (−0.016, 0.001) PPARG rs1801282 0.013 (−0.061, 0.087) 0.011 (−0.042, 0.064) 0.014 (−0.007, 0.035) 0.015 (0.000, 0.031) KCNQ1 rs2237892 0.014 (−0.023, 0.051) 0.014 (−0.012, 0.040) −0.002 (−0.013, 0.008) 0.001 (−0.007, 0.009) IRS1 rs2943641 0.012 (−0.057, 0.081) −0.011 (−0.061, 0.038) −0.019 (−0.039, 0.001) 0.017 (0.003, 0.032) The data are presented as unstandardized coefficients and 95% CIs using linear regression analyses. The model was unadjusted. View Large Table 1. Association of the 18 Diabetes-Related SNPs With Lipid Profile Gene SNP TC LDL-C HDL-C log10TG TCF2 (HNF1B) rs4430796 0.047 (0.009, 0.085) 0.031 (0.004, 0.059) −0.002 (−0.013, 0.009) 0.003 (−0.005, 0.011) GCC1/PAX4 rs6467136 0.077 (0.028, 0.125) 0.042 (0.008, 0.077) 0.014 (0.001, 0.028) 0.003 (−0.007, 0.014) MAEA rs6815464 −0.016 (−0.052, 0.019) −0.009 (−0.035, 0.016) 0.007 (−0.003, 0.017) 0.001 (−0.006, 0.009) GLIS3 rs7041847 0.031 (−0.003, 0.066) 0.027 (0.002, 0.051) 0.005 (−0.005, 0.014) 0.001 (−0.006, 0.008) C2CD4A/C2CD4B rs7172432 0.001 (−0.034, 0.036) 0.010 (−0.015, 0.035) −0.002 (−0.012, 0.008) −0.002 (−0.010, 0.005) TCF7L2 rs7903146 0.036 (−0.056, 0.129) 0.043 (−0.024, 0.109) 0.006 (−0.020, 0.032) −0.006 (−0.025, 0.014) ZFAND3 rs9470794 −0.017 (−0.054, 0.020) −0.002 (−0.029, 0.024) −0.007 (−0.018, 0.004) 0.004 (−0.003, 0.012) CDKN2A/B rs10811661 0.018 (−0.016, 0.053) 0.017 (−0.007, 0.042) 0.004 (−0.006, 0.014) 0.004 (−0.003, 0.011) CDC123/CAMK1D rs12779790 0.050 (0.004, 0.096) 0.017 (−0.016, 0.050) 0.003 (−0.010, 0.016) 0.003 (−0.007, 0.012) PROX1 rs340874 0.008 (−0.028, 0.043) 0.005 (−0.021, 0.030) 0.008 (−0.002, 0.018) 0.002 (−0.005, 0.010) SRR rs391300 0.004 (−0.033, 0.042) 0.011 (−0.015, 0.038) −0.001 (−0.011, 0.010) 0.000 (−0.008, 0.008) PSMD6 rs831571 0.038 (0.001, 0.074) 0.038 (0.012, 0.064) 0.003 (−0.007, 0.014) 0.003 (−0.004, 0.011) JAZF1 rs864745 0.043 (0.003, 0.084) 0.032 (0.003, 0.061) −0.001 (−0.013, 0.010) 0.000 (−0.004, 0.011) TP53INP1 rs896854 0.016 (−0.022, 0.053) 0.006 (−0.021, 0.033) 0.003 (−0.008, 0.014) 0.004 (−0.004, 0.012) SPRY2 rs1359790 −0.011 (−0.050, 0.028) −0.014 (−0.042, 0.014) −0.005 (−0.016, 0.006) −0.007 (−0.016, 0.001) PPARG rs1801282 0.013 (−0.061, 0.087) 0.011 (−0.042, 0.064) 0.014 (−0.007, 0.035) 0.015 (0.000, 0.031) KCNQ1 rs2237892 0.014 (−0.023, 0.051) 0.014 (−0.012, 0.040) −0.002 (−0.013, 0.008) 0.001 (−0.007, 0.009) IRS1 rs2943641 0.012 (−0.057, 0.081) −0.011 (−0.061, 0.038) −0.019 (−0.039, 0.001) 0.017 (0.003, 0.032) Gene SNP TC LDL-C HDL-C log10TG TCF2 (HNF1B) rs4430796 0.047 (0.009, 0.085) 0.031 (0.004, 0.059) −0.002 (−0.013, 0.009) 0.003 (−0.005, 0.011) GCC1/PAX4 rs6467136 0.077 (0.028, 0.125) 0.042 (0.008, 0.077) 0.014 (0.001, 0.028) 0.003 (−0.007, 0.014) MAEA rs6815464 −0.016 (−0.052, 0.019) −0.009 (−0.035, 0.016) 0.007 (−0.003, 0.017) 0.001 (−0.006, 0.009) GLIS3 rs7041847 0.031 (−0.003, 0.066) 0.027 (0.002, 0.051) 0.005 (−0.005, 0.014) 0.001 (−0.006, 0.008) C2CD4A/C2CD4B rs7172432 0.001 (−0.034, 0.036) 0.010 (−0.015, 0.035) −0.002 (−0.012, 0.008) −0.002 (−0.010, 0.005) TCF7L2 rs7903146 0.036 (−0.056, 0.129) 0.043 (−0.024, 0.109) 0.006 (−0.020, 0.032) −0.006 (−0.025, 0.014) ZFAND3 rs9470794 −0.017 (−0.054, 0.020) −0.002 (−0.029, 0.024) −0.007 (−0.018, 0.004) 0.004 (−0.003, 0.012) CDKN2A/B rs10811661 0.018 (−0.016, 0.053) 0.017 (−0.007, 0.042) 0.004 (−0.006, 0.014) 0.004 (−0.003, 0.011) CDC123/CAMK1D rs12779790 0.050 (0.004, 0.096) 0.017 (−0.016, 0.050) 0.003 (−0.010, 0.016) 0.003 (−0.007, 0.012) PROX1 rs340874 0.008 (−0.028, 0.043) 0.005 (−0.021, 0.030) 0.008 (−0.002, 0.018) 0.002 (−0.005, 0.010) SRR rs391300 0.004 (−0.033, 0.042) 0.011 (−0.015, 0.038) −0.001 (−0.011, 0.010) 0.000 (−0.008, 0.008) PSMD6 rs831571 0.038 (0.001, 0.074) 0.038 (0.012, 0.064) 0.003 (−0.007, 0.014) 0.003 (−0.004, 0.011) JAZF1 rs864745 0.043 (0.003, 0.084) 0.032 (0.003, 0.061) −0.001 (−0.013, 0.010) 0.000 (−0.004, 0.011) TP53INP1 rs896854 0.016 (−0.022, 0.053) 0.006 (−0.021, 0.033) 0.003 (−0.008, 0.014) 0.004 (−0.004, 0.012) SPRY2 rs1359790 −0.011 (−0.050, 0.028) −0.014 (−0.042, 0.014) −0.005 (−0.016, 0.006) −0.007 (−0.016, 0.001) PPARG rs1801282 0.013 (−0.061, 0.087) 0.011 (−0.042, 0.064) 0.014 (−0.007, 0.035) 0.015 (0.000, 0.031) KCNQ1 rs2237892 0.014 (−0.023, 0.051) 0.014 (−0.012, 0.040) −0.002 (−0.013, 0.008) 0.001 (−0.007, 0.009) IRS1 rs2943641 0.012 (−0.057, 0.081) −0.011 (−0.061, 0.038) −0.019 (−0.039, 0.001) 0.017 (0.003, 0.032) The data are presented as unstandardized coefficients and 95% CIs using linear regression analyses. The model was unadjusted. View Large Table 2. Association of the 14 BMI-Related SNPs With Lipid Profile Gene SNP TC LDL-C HDL-C log10TG TFAP2B rs4715210 0.002 (−0.044, 0.049) 0.015 (−0.019, 0.049) −0.002 (−0.015, 0.012) 0.006 (−0.004, 0.016) MAP2K5 rs4776970 −0.004 (−0.044, 0.036) 0.001 (−0.028, 0.030) −0.015 (−0.027, −0.004) 0.005 (−0.003, 0.014) ADCY3-DNAJC27 rs6545814 −0.007 (−0.042, 0.027) 0.019 (−0.006, 0.045) −0.007 (−0.017, 0.003) −0.002 (−0.010, 0.005) MC4R rs6567160 0.019 (−0.023, 0.060) 0.025 (−0.005, 0.055) −0.014 (−0.026, –0.002) 0.008 (−0.001, 0.017) CDKAL1 rs9356744 −0.019 (−0.053, 0.015) −0.020 (−0.045, 0.005) 0.001 (−0.008, 0.011) −0.001 (−0.008, 0.006) GNPDA2 rs10938397 0.024 (−0.012, 0.060) 0.012 (−0.014, 0.038) 0.007 (−0.004, 0.017) −0.002 (−0.010, 0.005) SMC5-KLF9 rs11142387 −0.012 (−0.048, 0.025) −0.002 (−0.029, 0.025) 0.001 (−0.009, 0.012) 0.005 (−0.003, 0.012) GIPR-QPCTL rs11671664 0.001 (−0.032, 0.035) −0.004 (−0.028, 0.021) 0.002 (−0.007, 0.012) 0.003 (−0.004, 0.010) GP2 rs12597579 0.014 (−0.023, 0.052) 0.008 (−0.019, 0.035) −0.004 (−0.015, 0.007) 0.006 (−0.002, 0.014) FTO rs17817449 0.033 (−0.020, 0.086) 0.029 (−0.009, 0.068) −0.015 (−0.031, 0.000) 0.017 (0.005, 0.028) BDNF rs6265 −0.015 (−0.048, 0.019) 0.008 (−0.016, 0.033) −0.012 (−0.022, −0.002) −0.003 (−0.010, 0.004) PCSK1 rs261967 −0.020 (−0.054, 0.014) −0.003 (−0.028, 0.022) −0.009 (−0.019, 0.001) 0.005 (−0.003, 0.012) SEC16B rs574367 0.027 (−0.016, 0.07) 0.006 (−0.025, 0.037) −0.013 (−0.025, 0.000) 0.008 (−0.002, 0.017) PAX6 rs652722 0.002 (−0.033, 0.037) 0.002 (−0.023, 0.028) 0.000 (−0.010, 0.010) 0.000 (−0.008, 0.007) Gene SNP TC LDL-C HDL-C log10TG TFAP2B rs4715210 0.002 (−0.044, 0.049) 0.015 (−0.019, 0.049) −0.002 (−0.015, 0.012) 0.006 (−0.004, 0.016) MAP2K5 rs4776970 −0.004 (−0.044, 0.036) 0.001 (−0.028, 0.030) −0.015 (−0.027, −0.004) 0.005 (−0.003, 0.014) ADCY3-DNAJC27 rs6545814 −0.007 (−0.042, 0.027) 0.019 (−0.006, 0.045) −0.007 (−0.017, 0.003) −0.002 (−0.010, 0.005) MC4R rs6567160 0.019 (−0.023, 0.060) 0.025 (−0.005, 0.055) −0.014 (−0.026, –0.002) 0.008 (−0.001, 0.017) CDKAL1 rs9356744 −0.019 (−0.053, 0.015) −0.020 (−0.045, 0.005) 0.001 (−0.008, 0.011) −0.001 (−0.008, 0.006) GNPDA2 rs10938397 0.024 (−0.012, 0.060) 0.012 (−0.014, 0.038) 0.007 (−0.004, 0.017) −0.002 (−0.010, 0.005) SMC5-KLF9 rs11142387 −0.012 (−0.048, 0.025) −0.002 (−0.029, 0.025) 0.001 (−0.009, 0.012) 0.005 (−0.003, 0.012) GIPR-QPCTL rs11671664 0.001 (−0.032, 0.035) −0.004 (−0.028, 0.021) 0.002 (−0.007, 0.012) 0.003 (−0.004, 0.010) GP2 rs12597579 0.014 (−0.023, 0.052) 0.008 (−0.019, 0.035) −0.004 (−0.015, 0.007) 0.006 (−0.002, 0.014) FTO rs17817449 0.033 (−0.020, 0.086) 0.029 (−0.009, 0.068) −0.015 (−0.031, 0.000) 0.017 (0.005, 0.028) BDNF rs6265 −0.015 (−0.048, 0.019) 0.008 (−0.016, 0.033) −0.012 (−0.022, −0.002) −0.003 (−0.010, 0.004) PCSK1 rs261967 −0.020 (−0.054, 0.014) −0.003 (−0.028, 0.022) −0.009 (−0.019, 0.001) 0.005 (−0.003, 0.012) SEC16B rs574367 0.027 (−0.016, 0.07) 0.006 (−0.025, 0.037) −0.013 (−0.025, 0.000) 0.008 (−0.002, 0.017) PAX6 rs652722 0.002 (−0.033, 0.037) 0.002 (−0.023, 0.028) 0.000 (−0.010, 0.010) 0.000 (−0.008, 0.007) The data are presented as unstandardized coefficients and 95% CIs using linear regression analyses. The model was unadjusted. View Large Table 2. Association of the 14 BMI-Related SNPs With Lipid Profile Gene SNP TC LDL-C HDL-C log10TG TFAP2B rs4715210 0.002 (−0.044, 0.049) 0.015 (−0.019, 0.049) −0.002 (−0.015, 0.012) 0.006 (−0.004, 0.016) MAP2K5 rs4776970 −0.004 (−0.044, 0.036) 0.001 (−0.028, 0.030) −0.015 (−0.027, −0.004) 0.005 (−0.003, 0.014) ADCY3-DNAJC27 rs6545814 −0.007 (−0.042, 0.027) 0.019 (−0.006, 0.045) −0.007 (−0.017, 0.003) −0.002 (−0.010, 0.005) MC4R rs6567160 0.019 (−0.023, 0.060) 0.025 (−0.005, 0.055) −0.014 (−0.026, –0.002) 0.008 (−0.001, 0.017) CDKAL1 rs9356744 −0.019 (−0.053, 0.015) −0.020 (−0.045, 0.005) 0.001 (−0.008, 0.011) −0.001 (−0.008, 0.006) GNPDA2 rs10938397 0.024 (−0.012, 0.060) 0.012 (−0.014, 0.038) 0.007 (−0.004, 0.017) −0.002 (−0.010, 0.005) SMC5-KLF9 rs11142387 −0.012 (−0.048, 0.025) −0.002 (−0.029, 0.025) 0.001 (−0.009, 0.012) 0.005 (−0.003, 0.012) GIPR-QPCTL rs11671664 0.001 (−0.032, 0.035) −0.004 (−0.028, 0.021) 0.002 (−0.007, 0.012) 0.003 (−0.004, 0.010) GP2 rs12597579 0.014 (−0.023, 0.052) 0.008 (−0.019, 0.035) −0.004 (−0.015, 0.007) 0.006 (−0.002, 0.014) FTO rs17817449 0.033 (−0.020, 0.086) 0.029 (−0.009, 0.068) −0.015 (−0.031, 0.000) 0.017 (0.005, 0.028) BDNF rs6265 −0.015 (−0.048, 0.019) 0.008 (−0.016, 0.033) −0.012 (−0.022, −0.002) −0.003 (−0.010, 0.004) PCSK1 rs261967 −0.020 (−0.054, 0.014) −0.003 (−0.028, 0.022) −0.009 (−0.019, 0.001) 0.005 (−0.003, 0.012) SEC16B rs574367 0.027 (−0.016, 0.07) 0.006 (−0.025, 0.037) −0.013 (−0.025, 0.000) 0.008 (−0.002, 0.017) PAX6 rs652722 0.002 (−0.033, 0.037) 0.002 (−0.023, 0.028) 0.000 (−0.010, 0.010) 0.000 (−0.008, 0.007) Gene SNP TC LDL-C HDL-C log10TG TFAP2B rs4715210 0.002 (−0.044, 0.049) 0.015 (−0.019, 0.049) −0.002 (−0.015, 0.012) 0.006 (−0.004, 0.016) MAP2K5 rs4776970 −0.004 (−0.044, 0.036) 0.001 (−0.028, 0.030) −0.015 (−0.027, −0.004) 0.005 (−0.003, 0.014) ADCY3-DNAJC27 rs6545814 −0.007 (−0.042, 0.027) 0.019 (−0.006, 0.045) −0.007 (−0.017, 0.003) −0.002 (−0.010, 0.005) MC4R rs6567160 0.019 (−0.023, 0.060) 0.025 (−0.005, 0.055) −0.014 (−0.026, –0.002) 0.008 (−0.001, 0.017) CDKAL1 rs9356744 −0.019 (−0.053, 0.015) −0.020 (−0.045, 0.005) 0.001 (−0.008, 0.011) −0.001 (−0.008, 0.006) GNPDA2 rs10938397 0.024 (−0.012, 0.060) 0.012 (−0.014, 0.038) 0.007 (−0.004, 0.017) −0.002 (−0.010, 0.005) SMC5-KLF9 rs11142387 −0.012 (−0.048, 0.025) −0.002 (−0.029, 0.025) 0.001 (−0.009, 0.012) 0.005 (−0.003, 0.012) GIPR-QPCTL rs11671664 0.001 (−0.032, 0.035) −0.004 (−0.028, 0.021) 0.002 (−0.007, 0.012) 0.003 (−0.004, 0.010) GP2 rs12597579 0.014 (−0.023, 0.052) 0.008 (−0.019, 0.035) −0.004 (−0.015, 0.007) 0.006 (−0.002, 0.014) FTO rs17817449 0.033 (−0.020, 0.086) 0.029 (−0.009, 0.068) −0.015 (−0.031, 0.000) 0.017 (0.005, 0.028) BDNF rs6265 −0.015 (−0.048, 0.019) 0.008 (−0.016, 0.033) −0.012 (−0.022, −0.002) −0.003 (−0.010, 0.004) PCSK1 rs261967 −0.020 (−0.054, 0.014) −0.003 (−0.028, 0.022) −0.009 (−0.019, 0.001) 0.005 (−0.003, 0.012) SEC16B rs574367 0.027 (−0.016, 0.07) 0.006 (−0.025, 0.037) −0.013 (−0.025, 0.000) 0.008 (−0.002, 0.017) PAX6 rs652722 0.002 (−0.033, 0.037) 0.002 (−0.023, 0.028) 0.000 (−0.010, 0.010) 0.000 (−0.008, 0.007) The data are presented as unstandardized coefficients and 95% CIs using linear regression analyses. The model was unadjusted. View Large Most of the diabetes-related SNPs did not exhibit significant associations with lipid profile; the exceptions were rs4430796, rs6467136, rs12779790, rs831571, rs864745, rs7041847, and rs2943641 (Table 1). Most of the BMI-related SNPs also did not exhibit significant relationships with lipid profile; the exceptions were rs17817449, rs4776970, rs6567160, rs6265, and rs574367 (Table 2). However, additional adjustments for HbA1c and BMI greatly attenuated the associations of most of these exceptive SNPs with lipid profile. Pleiotropic effects of SNPs We measured the potential associations of the SNPs with BMI, systolic and diastolic blood pressure, FPG, HbA1c, and age using an additive model (Supplemental Figs. 3 and 4). Three of the diabetes-related SNPs exhibited an association with at least one trait at the nominal P value ≤0.05. Two SNPs (rs4430796 and rs391300) were associated with systolic and diastolic blood pressure, and two SNPs (rs4430796 and rs896854) were associated with age. Seven of the BMI-related SNPs exhibited a significant association with at least one trait. Seven SNPs were associated with blood pressure (rs4776970, rs12597579, rs17817449, rs261967, rs4715210, rs6567160, and rs574367), two SNPs were associated with FPG (rs4715210 and rs6545814), and two SNPs were associated with age (rs4715210 and rs17817449). Study characteristics based on the DM_GRS and BMI_GRS As expected, with increases in the DM_GRS, the FPG, HbA1c, and prevalence of diabetes significantly increased. The BMI_GRS was also significantly associated with an increase in BMI (all P < 0.001; Table 3). Table 3. Characteristics of the Study Participants According to the DM_GRS and the BMI_GRS (n = 9798) Characteristic Q1 Q2 Q3 Q4 Q5 P for Trend DM_GRS ≤18.013 18.014–19.740 19.741–21.312 21.313–23.011 ≥23.012  n 1959 1961 1958 1961 1959  Age, y 54.3 (13.0) 55.0 (12.9) 55.4 (12.5) 54.3 (13.0) 54.4 (12.9) 0.669  Men, % 40.3 40.1 39.0 39.9 40.0 0.850  BMI, kg/m2 24.7 (3.5) 24.8 (3.7) 24.6 (3.6) 24.6 (3.7) 24.5 (3.5) 0.176  TC, mmol/L 5.19 (1.20) 5.15 (1.00) 5.25 (1.04) 5.23 (1.08) 5.24 (1.21) 0.010  TGs, mmol/L 1.65 (1.51) 1.63 (1.22) 1.67 (1.44) 1.67 (1.54) 1.68 (1.55) 0.220  HDL-C, mmol/L 1.40 (0.32) 1.39 (0.32) 1.40 (0.33) 1.41 (0.32) 1.40 (0.31) 0.726  LDL-C, mmol/L 3.17 (0.79) 3.18 (0.78) 3.25 (0.82) 3.23 (0.81) 3.22 (0.82) 0.010  FPG, mmol/L 5.5 (1.39) 5.6 (1.35) 5.7 (1.53) 5.7 (1.54) 5.8 (1.65) 3.9E-7  HbA1c, % 5.5 (0.9) 5.6 (0.9) 5.6 (1.0) 5.6 (1.0) 5.7 (1.1) 3.4E-8  Diabetes, % 11.7 13.7 14.2 15.5 18.3 2.9E-8  Hypertension, % 48.3 49.7 48.3 47.0 46.2 0.052  Economic status, low %/high % 32.7/67.3 34.6/65.4 33.7/66.3 35.1/64.9 35.6/64.4 0.062 BMI_GRS ≤8.186 8.187–9.404 9.405–10.621 10.622–12.007 ≥12.008  n 2358 1857 1855 1849 1879  Age, y 54.8 (12.9) 54.8 (13.0) 55.1 (12.6) 54.6 (12.9) 54.0 (13.0) 0.036  Men, % 40.5 39.1 39.3 41.3 38.9 0.702  BMI, kg/m2 24.4 (3.6) 24.5 (3.5) 24.7 (3.7) 24.8 (3.5) 25.0 (3.6) 3.5E-9  TC, mmol/L 5.19 (1.07) 5.21 (1.10) 5.23 (1.06) 5.25 (1.24) 5.18 (1.07) 0.906  TGs, mmol/L 1.64 (1.57) 1.65 (1.47) 1.63 (1.23) 1.71 (1.46) 1.68 (1.49) 0.019  HDL-C, mmol/L 1.41 (0.33) 1.41 (0.32) 1.41 (0.32) 1.39 (0.32) 1.38 (0.31) 0.005  LDL-C, mmol/L 3.19 (0.81) 3.20 (0.78) 3.23 (0.81) 3.23 (0.81) 3.21 (0.81) 0.197  FPG, mmol/L 5.7 (1.6) 5.6 (1.5) 5.6 (1.4) 5.6 (1.6) 5.7 (1.5) 0.953  HbA1c, % 5.6 (1.0) 5.6 (1.0) 5.6 (1.0) 5.6 (1.0) 5.7 (1.0) 0.367  Diabetes, % 14.6 14.3 15.6 13.8 15.1 0.864  Hypertension, % 46.9 46.6 47.3 48.9 49.9 0.022  Economic status, low %/high % 34.1/65.9 32.7/67.3 35.0/65.0 35.6/64.4 34.4/65.6 0.313 Characteristic Q1 Q2 Q3 Q4 Q5 P for Trend DM_GRS ≤18.013 18.014–19.740 19.741–21.312 21.313–23.011 ≥23.012  n 1959 1961 1958 1961 1959  Age, y 54.3 (13.0) 55.0 (12.9) 55.4 (12.5) 54.3 (13.0) 54.4 (12.9) 0.669  Men, % 40.3 40.1 39.0 39.9 40.0 0.850  BMI, kg/m2 24.7 (3.5) 24.8 (3.7) 24.6 (3.6) 24.6 (3.7) 24.5 (3.5) 0.176  TC, mmol/L 5.19 (1.20) 5.15 (1.00) 5.25 (1.04) 5.23 (1.08) 5.24 (1.21) 0.010  TGs, mmol/L 1.65 (1.51) 1.63 (1.22) 1.67 (1.44) 1.67 (1.54) 1.68 (1.55) 0.220  HDL-C, mmol/L 1.40 (0.32) 1.39 (0.32) 1.40 (0.33) 1.41 (0.32) 1.40 (0.31) 0.726  LDL-C, mmol/L 3.17 (0.79) 3.18 (0.78) 3.25 (0.82) 3.23 (0.81) 3.22 (0.82) 0.010  FPG, mmol/L 5.5 (1.39) 5.6 (1.35) 5.7 (1.53) 5.7 (1.54) 5.8 (1.65) 3.9E-7  HbA1c, % 5.5 (0.9) 5.6 (0.9) 5.6 (1.0) 5.6 (1.0) 5.7 (1.1) 3.4E-8  Diabetes, % 11.7 13.7 14.2 15.5 18.3 2.9E-8  Hypertension, % 48.3 49.7 48.3 47.0 46.2 0.052  Economic status, low %/high % 32.7/67.3 34.6/65.4 33.7/66.3 35.1/64.9 35.6/64.4 0.062 BMI_GRS ≤8.186 8.187–9.404 9.405–10.621 10.622–12.007 ≥12.008  n 2358 1857 1855 1849 1879  Age, y 54.8 (12.9) 54.8 (13.0) 55.1 (12.6) 54.6 (12.9) 54.0 (13.0) 0.036  Men, % 40.5 39.1 39.3 41.3 38.9 0.702  BMI, kg/m2 24.4 (3.6) 24.5 (3.5) 24.7 (3.7) 24.8 (3.5) 25.0 (3.6) 3.5E-9  TC, mmol/L 5.19 (1.07) 5.21 (1.10) 5.23 (1.06) 5.25 (1.24) 5.18 (1.07) 0.906  TGs, mmol/L 1.64 (1.57) 1.65 (1.47) 1.63 (1.23) 1.71 (1.46) 1.68 (1.49) 0.019  HDL-C, mmol/L 1.41 (0.33) 1.41 (0.32) 1.41 (0.32) 1.39 (0.32) 1.38 (0.31) 0.005  LDL-C, mmol/L 3.19 (0.81) 3.20 (0.78) 3.23 (0.81) 3.23 (0.81) 3.21 (0.81) 0.197  FPG, mmol/L 5.7 (1.6) 5.6 (1.5) 5.6 (1.4) 5.6 (1.6) 5.7 (1.5) 0.953  HbA1c, % 5.6 (1.0) 5.6 (1.0) 5.6 (1.0) 5.6 (1.0) 5.7 (1.0) 0.367  Diabetes, % 14.6 14.3 15.6 13.8 15.1 0.864  Hypertension, % 46.9 46.6 47.3 48.9 49.9 0.022  Economic status, low %/high % 34.1/65.9 32.7/67.3 35.0/65.0 35.6/64.4 34.4/65.6 0.313 The data are summarized as the means (SD) for continuous variables and as numerical proportions for categorical variables. The P values for the trends were as calculated with analysis of variance and χ2 tests. View Large Table 3. Characteristics of the Study Participants According to the DM_GRS and the BMI_GRS (n = 9798) Characteristic Q1 Q2 Q3 Q4 Q5 P for Trend DM_GRS ≤18.013 18.014–19.740 19.741–21.312 21.313–23.011 ≥23.012  n 1959 1961 1958 1961 1959  Age, y 54.3 (13.0) 55.0 (12.9) 55.4 (12.5) 54.3 (13.0) 54.4 (12.9) 0.669  Men, % 40.3 40.1 39.0 39.9 40.0 0.850  BMI, kg/m2 24.7 (3.5) 24.8 (3.7) 24.6 (3.6) 24.6 (3.7) 24.5 (3.5) 0.176  TC, mmol/L 5.19 (1.20) 5.15 (1.00) 5.25 (1.04) 5.23 (1.08) 5.24 (1.21) 0.010  TGs, mmol/L 1.65 (1.51) 1.63 (1.22) 1.67 (1.44) 1.67 (1.54) 1.68 (1.55) 0.220  HDL-C, mmol/L 1.40 (0.32) 1.39 (0.32) 1.40 (0.33) 1.41 (0.32) 1.40 (0.31) 0.726  LDL-C, mmol/L 3.17 (0.79) 3.18 (0.78) 3.25 (0.82) 3.23 (0.81) 3.22 (0.82) 0.010  FPG, mmol/L 5.5 (1.39) 5.6 (1.35) 5.7 (1.53) 5.7 (1.54) 5.8 (1.65) 3.9E-7  HbA1c, % 5.5 (0.9) 5.6 (0.9) 5.6 (1.0) 5.6 (1.0) 5.7 (1.1) 3.4E-8  Diabetes, % 11.7 13.7 14.2 15.5 18.3 2.9E-8  Hypertension, % 48.3 49.7 48.3 47.0 46.2 0.052  Economic status, low %/high % 32.7/67.3 34.6/65.4 33.7/66.3 35.1/64.9 35.6/64.4 0.062 BMI_GRS ≤8.186 8.187–9.404 9.405–10.621 10.622–12.007 ≥12.008  n 2358 1857 1855 1849 1879  Age, y 54.8 (12.9) 54.8 (13.0) 55.1 (12.6) 54.6 (12.9) 54.0 (13.0) 0.036  Men, % 40.5 39.1 39.3 41.3 38.9 0.702  BMI, kg/m2 24.4 (3.6) 24.5 (3.5) 24.7 (3.7) 24.8 (3.5) 25.0 (3.6) 3.5E-9  TC, mmol/L 5.19 (1.07) 5.21 (1.10) 5.23 (1.06) 5.25 (1.24) 5.18 (1.07) 0.906  TGs, mmol/L 1.64 (1.57) 1.65 (1.47) 1.63 (1.23) 1.71 (1.46) 1.68 (1.49) 0.019  HDL-C, mmol/L 1.41 (0.33) 1.41 (0.32) 1.41 (0.32) 1.39 (0.32) 1.38 (0.31) 0.005  LDL-C, mmol/L 3.19 (0.81) 3.20 (0.78) 3.23 (0.81) 3.23 (0.81) 3.21 (0.81) 0.197  FPG, mmol/L 5.7 (1.6) 5.6 (1.5) 5.6 (1.4) 5.6 (1.6) 5.7 (1.5) 0.953  HbA1c, % 5.6 (1.0) 5.6 (1.0) 5.6 (1.0) 5.6 (1.0) 5.7 (1.0) 0.367  Diabetes, % 14.6 14.3 15.6 13.8 15.1 0.864  Hypertension, % 46.9 46.6 47.3 48.9 49.9 0.022  Economic status, low %/high % 34.1/65.9 32.7/67.3 35.0/65.0 35.6/64.4 34.4/65.6 0.313 Characteristic Q1 Q2 Q3 Q4 Q5 P for Trend DM_GRS ≤18.013 18.014–19.740 19.741–21.312 21.313–23.011 ≥23.012  n 1959 1961 1958 1961 1959  Age, y 54.3 (13.0) 55.0 (12.9) 55.4 (12.5) 54.3 (13.0) 54.4 (12.9) 0.669  Men, % 40.3 40.1 39.0 39.9 40.0 0.850  BMI, kg/m2 24.7 (3.5) 24.8 (3.7) 24.6 (3.6) 24.6 (3.7) 24.5 (3.5) 0.176  TC, mmol/L 5.19 (1.20) 5.15 (1.00) 5.25 (1.04) 5.23 (1.08) 5.24 (1.21) 0.010  TGs, mmol/L 1.65 (1.51) 1.63 (1.22) 1.67 (1.44) 1.67 (1.54) 1.68 (1.55) 0.220  HDL-C, mmol/L 1.40 (0.32) 1.39 (0.32) 1.40 (0.33) 1.41 (0.32) 1.40 (0.31) 0.726  LDL-C, mmol/L 3.17 (0.79) 3.18 (0.78) 3.25 (0.82) 3.23 (0.81) 3.22 (0.82) 0.010  FPG, mmol/L 5.5 (1.39) 5.6 (1.35) 5.7 (1.53) 5.7 (1.54) 5.8 (1.65) 3.9E-7  HbA1c, % 5.5 (0.9) 5.6 (0.9) 5.6 (1.0) 5.6 (1.0) 5.7 (1.1) 3.4E-8  Diabetes, % 11.7 13.7 14.2 15.5 18.3 2.9E-8  Hypertension, % 48.3 49.7 48.3 47.0 46.2 0.052  Economic status, low %/high % 32.7/67.3 34.6/65.4 33.7/66.3 35.1/64.9 35.6/64.4 0.062 BMI_GRS ≤8.186 8.187–9.404 9.405–10.621 10.622–12.007 ≥12.008  n 2358 1857 1855 1849 1879  Age, y 54.8 (12.9) 54.8 (13.0) 55.1 (12.6) 54.6 (12.9) 54.0 (13.0) 0.036  Men, % 40.5 39.1 39.3 41.3 38.9 0.702  BMI, kg/m2 24.4 (3.6) 24.5 (3.5) 24.7 (3.7) 24.8 (3.5) 25.0 (3.6) 3.5E-9  TC, mmol/L 5.19 (1.07) 5.21 (1.10) 5.23 (1.06) 5.25 (1.24) 5.18 (1.07) 0.906  TGs, mmol/L 1.64 (1.57) 1.65 (1.47) 1.63 (1.23) 1.71 (1.46) 1.68 (1.49) 0.019  HDL-C, mmol/L 1.41 (0.33) 1.41 (0.32) 1.41 (0.32) 1.39 (0.32) 1.38 (0.31) 0.005  LDL-C, mmol/L 3.19 (0.81) 3.20 (0.78) 3.23 (0.81) 3.23 (0.81) 3.21 (0.81) 0.197  FPG, mmol/L 5.7 (1.6) 5.6 (1.5) 5.6 (1.4) 5.6 (1.6) 5.7 (1.5) 0.953  HbA1c, % 5.6 (1.0) 5.6 (1.0) 5.6 (1.0) 5.6 (1.0) 5.7 (1.0) 0.367  Diabetes, % 14.6 14.3 15.6 13.8 15.1 0.864  Hypertension, % 46.9 46.6 47.3 48.9 49.9 0.022  Economic status, low %/high % 34.1/65.9 32.7/67.3 35.0/65.0 35.6/64.4 34.4/65.6 0.313 The data are summarized as the means (SD) for continuous variables and as numerical proportions for categorical variables. The P values for the trends were as calculated with analysis of variance and χ2 tests. View Large The ORZX of the observed association between DM_GRS (per 1 SD increase) and diabetes was 1.231 (95% CI: 1.159, 1.307) with an F statistic of 69. The βZX regression coefficient between BMI_GRS (per 1 SD increase) and BMI was 0.215 (95% CI: 0.147, 0.284) with an F statistic of 39. The F statistics greater than 10 commonly suggest these two GRSs are sufficiently strong as IVs. Regarding the pleiotropic effects, the DM_GRS quintiles were not significantly associated with BMI or hypertension, and each 1 SD increment of the DM_GRS also did not exhibit a significant association with BMI or hypertension in the adjusted model. The BMI_GRSs quintiles were associated with age and hypertension (Table 3). Therefore, in the sensitivity analyses, we constructed a BMI_GRS12SNP that excluded FTO rs17817449 and MC4R rs6567160, which were the two most obvious loci that had pleiotropic effects. Association of diabetes and DM_GRS with lipid profile Next, we analyzed the associations of present diabetes and DM_GRS with lipid profile including TC, HDL-C, LDL-C, and TG (Table 4). In this cross-sectional study, present diabetes (βXY) was significantly associated with log10TG (β = 0.078; 95% CI: 0.066, 0.091) and HDL-C (β = –0.067; 95% CI: –0.085, –0.049) but not with TC or LDL-C. However, the DM_GRS (βZY) was significantly related to TC (β = 0.027; 95% CI: 0.005, 0.048) and LDL-C (β = 0.026; 95% CI: 0.011, 0.042) and marginally related to log10TG (β = 0.004; 95% CI: 0.000, 0.009) but not to HDL-C. Table 4. Causal Coefficients From the MR Analysis for the Associations of Type 2 Diabetes and BMI With Lipid Profile TC LDL-C HDL-C log10TG DM  βXY (present diabetes) 0.019 (–0.043, 0.082) 0.029 (–0.016, 0.073) −0.067 (–0.085, –0.049) 0.078 (0.066, 0.091)  βZY (per SD increase in DM_GRS) 0.027 (0.005, 0.048) 0.026 (0.011, 0.042) −0.001 (–0.007, 0.005) 0.004 (0.000, 0.009)  βIV (present diabetes) 0.130 (0.020, 0.240) 0.125 (0.041, 0.209) −0.005 (–0.033, 0.023) 0.019 (–0.001, 0.039) BMI  βXY (per 10 kg/m2 increase in BMI) 0.281 (0.219, 0.343) 0.328 (0.284, 0.373) −0.212 (–0.230, –0.195) 0.169 (0.157, 0.182)  βZY (per SD increase in BMI_GRS) −0.002 (–0.024, 0.019) 0.008 (–0.008, 0.023) −0.009 (–0.015, –0.003) 0.005 (0.001, 0.010)  βIV (per 10 kg/m2 increase in BMI) −0.091 (–1.071, 0.889) 0.364 (–0.365, 1.083) −0.409 (–0.698, –0.120) 0.227 (0.039, 0.415) TC LDL-C HDL-C log10TG DM  βXY (present diabetes) 0.019 (–0.043, 0.082) 0.029 (–0.016, 0.073) −0.067 (–0.085, –0.049) 0.078 (0.066, 0.091)  βZY (per SD increase in DM_GRS) 0.027 (0.005, 0.048) 0.026 (0.011, 0.042) −0.001 (–0.007, 0.005) 0.004 (0.000, 0.009)  βIV (present diabetes) 0.130 (0.020, 0.240) 0.125 (0.041, 0.209) −0.005 (–0.033, 0.023) 0.019 (–0.001, 0.039) BMI  βXY (per 10 kg/m2 increase in BMI) 0.281 (0.219, 0.343) 0.328 (0.284, 0.373) −0.212 (–0.230, –0.195) 0.169 (0.157, 0.182)  βZY (per SD increase in BMI_GRS) −0.002 (–0.024, 0.019) 0.008 (–0.008, 0.023) −0.009 (–0.015, –0.003) 0.005 (0.001, 0.010)  βIV (per 10 kg/m2 increase in BMI) −0.091 (–1.071, 0.889) 0.364 (–0.365, 1.083) −0.409 (–0.698, –0.120) 0.227 (0.039, 0.415) The data are presented as regression coefficients (β) and 95% CIs. In this MR framework, the IV estimator is βIV = βZY/βZX, which equals the causal coefficients of diabetes and BMI (per 10 kg/m2 increase) for changes in lipid parameters. The data were adjusted for age, sex, BMI (only in the the BMI_GRS model), hypertension, and economic status. Abbreviation: DM, diabetes mellitus. View Large Table 4. Causal Coefficients From the MR Analysis for the Associations of Type 2 Diabetes and BMI With Lipid Profile TC LDL-C HDL-C log10TG DM  βXY (present diabetes) 0.019 (–0.043, 0.082) 0.029 (–0.016, 0.073) −0.067 (–0.085, –0.049) 0.078 (0.066, 0.091)  βZY (per SD increase in DM_GRS) 0.027 (0.005, 0.048) 0.026 (0.011, 0.042) −0.001 (–0.007, 0.005) 0.004 (0.000, 0.009)  βIV (present diabetes) 0.130 (0.020, 0.240) 0.125 (0.041, 0.209) −0.005 (–0.033, 0.023) 0.019 (–0.001, 0.039) BMI  βXY (per 10 kg/m2 increase in BMI) 0.281 (0.219, 0.343) 0.328 (0.284, 0.373) −0.212 (–0.230, –0.195) 0.169 (0.157, 0.182)  βZY (per SD increase in BMI_GRS) −0.002 (–0.024, 0.019) 0.008 (–0.008, 0.023) −0.009 (–0.015, –0.003) 0.005 (0.001, 0.010)  βIV (per 10 kg/m2 increase in BMI) −0.091 (–1.071, 0.889) 0.364 (–0.365, 1.083) −0.409 (–0.698, –0.120) 0.227 (0.039, 0.415) TC LDL-C HDL-C log10TG DM  βXY (present diabetes) 0.019 (–0.043, 0.082) 0.029 (–0.016, 0.073) −0.067 (–0.085, –0.049) 0.078 (0.066, 0.091)  βZY (per SD increase in DM_GRS) 0.027 (0.005, 0.048) 0.026 (0.011, 0.042) −0.001 (–0.007, 0.005) 0.004 (0.000, 0.009)  βIV (present diabetes) 0.130 (0.020, 0.240) 0.125 (0.041, 0.209) −0.005 (–0.033, 0.023) 0.019 (–0.001, 0.039) BMI  βXY (per 10 kg/m2 increase in BMI) 0.281 (0.219, 0.343) 0.328 (0.284, 0.373) −0.212 (–0.230, –0.195) 0.169 (0.157, 0.182)  βZY (per SD increase in BMI_GRS) −0.002 (–0.024, 0.019) 0.008 (–0.008, 0.023) −0.009 (–0.015, –0.003) 0.005 (0.001, 0.010)  βIV (per 10 kg/m2 increase in BMI) −0.091 (–1.071, 0.889) 0.364 (–0.365, 1.083) −0.409 (–0.698, –0.120) 0.227 (0.039, 0.415) The data are presented as regression coefficients (β) and 95% CIs. In this MR framework, the IV estimator is βIV = βZY/βZX, which equals the causal coefficients of diabetes and BMI (per 10 kg/m2 increase) for changes in lipid parameters. The data were adjusted for age, sex, BMI (only in the the BMI_GRS model), hypertension, and economic status. Abbreviation: DM, diabetes mellitus. View Large In the IV analysis, the causal regression coefficients (βIV) of genetically determined diabetes for TC, LDL-C, and log10TG were 0.130 (95% CI: 0.020, 0.240; P = 0.014), 0.125 (96% CI: 0.041, 0.209; P = 0.001), and 0.019 (95% CI: –0.001, 0.039; P = 0.055), respectively. The βIV for HDL-C was –0.005 (95% CI: –0.033, 0.023), which was not significant (P = 0.699). Association of BMI and the BMI_GRS with lipid profile The associations of measured BMI and the BMI_GRS with lipid profile are presented in Table 4. A 10 kg/m2 increase in measured BMI (βXY) was significantly associated with TC (β = 0.281; 95% CI: 0.219, 0.343), LDL-C (β = 0.328: 95% CI: 0.284, 0.373), HDL-C (β = –0.212; 95% CI: –0.230, –0.195), and log10TG (β = 0.169; 95% CI: 0.157, 0.182). However, the BMI_GRS (βZY) was only significantly associated with HDL-C (β = –0.009; 95% CI: –0.015, –0.003) and log10TG (β = 0.005; 95% CI: 0.001, 0.010). In the IV analysis, the causal regression coefficients (βIV) of a genetically determined 10 kg/m2 higher BMI for HDL-C and log10TG were –0.409 (96% CI: –0.698, –0.120; P = 0.004) and 0.227 (95% CI: 0.039, 0.415; P = 0.026), respectively. The βIV for TC and LDL-C was not significant. Sensitivity analysis Considering pleiotropy, we constructed a GRS that excluded FTO rs17817449 and MC4R rs6567160, which were no longer significantly associated with age (P = 0.155) or hypertension (P = 0.091). A genetically determined 10 kg/m2 greater BMI was still causally related to HDL (β = –0.361; 95% CI: –0.675, –0.047; P = 0.024) and marginally related to log10TG (β = 0.227; 95% CI: –0.001, 0.455; P = 0.051). Three assumptions for IV in MR First, the GRSs were strongly associated with exposure with F statistics that were considerably greater than 10, which indicated that these GRSs were adequate IVs. Second, a GRS should not be associated with any confounder of the exposure-outcome association. In our study, the BMI_GRSs quintiles were associated with age and hypertension; therefore, the BMI_GRS12SNP was constructed, and the BMI determined by the BMI_GRS12SNP was still causally associated with HDL. Third, a GRS should be independent of the outcome, except possibly through its association with the exposure. This assumption means that the only causal route from the genetic variants to the outcome should be through the exposure (i.e., there should be no other routes between the DM_GRS and BMI_GRS and hyperlipidemia) (23). However, this assumption is only an idealized view and quite untestable because it is possible that a causal association is due to other diabetes characteristics as a direct consequence of being diabetic (15), and biologic pathways are complex and interconnected. Additional adjustments for HbA1c or BMI could greatly attenuate most of the associations of DM_GRS and BMI_GRS with lipid profile, which verified the third assumption to a limited extent. Discussion In this cross-sectional survey that included almost 10,000 community-dwelling Chinese adults, we used an MR design to examine the causal associations between adiposity, as assessed by an elevated BMI, and lipid profile and between diabetes and lipid profile. The current study is, to our knowledge, the first MR study to measure the effect of diabetes on elements of the lipid profile. This is an old topic, but the MR design seems to reveal results that differ from conventional opinions. We demonstrated evidence for causal relationships (1) between diabetes and higher TC and LDL-C but not lower HDL-C and (2) between obesity and lower HDL-C and higher TG but not higher TC and LDL-C. Our findings do not fully accord with those of conventional observational studies. However, the setting of our study was different: We assessed the effects of lifelong exposure to diabetes risks and high BMI on hyperlipidemia. Many previous studies have demonstrated that hyperlipidemia is more prevalent in individuals with diabetes and obesity than it is in their nondiabetic and nonobese counterparts (25). The features of dyslipidemia in patients with diabetes have been reported to be high TG levels and decreased HDL-C levels, but the concentration of LDL-C can be in the normal range, and small dense LDL-C can be increased (6). However, our study found that type 2 diabetes may induce a significant elevation of LDL-C but not a decrease of HDL-C. This finding further indicates the important role of diabetes in the formation of atherosclerosis because LDL-C level is recommended for use as the primary lipid analysis for screening, risk estimation, diagnosis, and management (25). Interestingly, in our study, the observed associations of diabetes with lipid profile were significant for HDL-C and TG but not for TC or LDL-C, which might reflect the flaws of conventional observational studies. Additional independent MR studies of this association should be performed. In obesity, elevations of TC, LDL-C, and TG levels and decreased HDL-C levels are very common (5). There are two additional MR studies that have measured the associations of BMI with lipid profile (7, 26). We replicated the results of Fall et al. (7) from individuals of European descent. Using the FTO locus as the IV, these authors also observed an unfavorable effect of BMI on decreased HDL-C levels and increased TG levels and not on LDL-C or TC (7). However, another study that used 14 BMI-associated SNPs reported that a 1 kg/m2 increase in BMI reduced HDL-C by –0.02 mmol/L (95% CI: –0.03, –0.01) and, interestingly, LDL-C by –0.04 mmol/L (95% CI: –0.07, –0.01) in individuals of European descent (26). We found that the IV estimates for the effects of BMI on HDL-C and TG were greater than those calculated from standard linear regressions. As suggested in previous studies (7), possible explanations include the following: the cross-sectional nature of a study that could result in reverse causation and confounds, and the notion that the lifelong effects of SNPs on adiposity are not entirely captured by a single BMI measurement (27). CVD is the main cause of death in patients with type 2 diabetes and obesity (28, 29). Hyperlipidemia may be one important pathway linking diabetes and obesity to cardiovascular outcomes. Obesity could lead to dyslipidemia, which further induces and aggravates CVD (25, 30). Furthermore, a sustainable weight loss of 5 to 8 kg results in a mean LDL-C reduction of 5 mg/dL and a mean HDL-C increase of 2 to 3 mg/dL (30). Additionally, weight loss is also expected to decrease TG (30). However, an MR study did not identify evidence of a causal relationship between BMI and coronary heart disease (OR: 1.01; 95% CI: 0.94, 1.08) (26), and this issue requires further evidence. Although type 2 diabetes is regarded as a risk factor for CVD, the effect of intensive glucose-lowering therapies on the prevention of CVD is controversial. Some studies have found no benefit of intensive glucose-lowering therapy (31, 32), whereas one meta-analysis of randomized controlled trials indicated that intensive glycemic control resulted in a 15% reduction in coronary heart disease events (33). Moreover, a very recent MR study revealed a causal association between diabetes and increased arterial stiffness (OR: 1.24; 95% CI: 1.06, 1.47) in a Chinese population, which indicates that long-term treatment of diabetes may be beneficial to the prevention of the development of arterial stiffening (15). Our study suggests that diabetes could induce an elevation of LDL-C, which is the core element of the lipid profile in the development of atherosclerosis. Thus, we suspect that long-term treatment of diabetes may have benefits for atherosclerosis prevention that are partly mediated through LDL-C reduction. The strengths of our study include the well-defined community setting and the use of MR to delineate the genetically determined relations between diabetes, BMI, and lipid profile. Second, the IVs we used for the MR were strong (F > 10). Last, our participants were all of Asian descent, which eliminated population admixture as a potential drawback. This study also has some limitations. All of the participants were of Asian origin. Therefore, our findings may not be applicable to other ethnicities. Second, the lipid profiles were measured only once at baseline. Hence, we were not able to control for intraindividual variability. Third, the self-reported diagnoses of hypertension and diabetes may not have been completely accurate; however, in an Asian population study, the self-reported medical histories were generally accurate, including self-reported hypertension and diabetes (34). Finally, we built our GRSs based only on common variants that are considered to represent limited diabetes and BMI heritability. We were unable to assess the potential contribution of rare variants. It was also difficult to use the GRSs to completely discriminate the effects of diabetes and obesity from the effects of other risk factors on changes in lipid profile because of the possible pleiotropic effects of SNPs, especially regarding the BMI_GRS. In conclusion, this analysis has provided evidence for the biologically plausible causal relationships of genetically determined type 2 diabetes with higher TC and LDL-C and genetically determined obesity with lower HDL-C and higher TG in a Chinese population. Diabetes and adiposity seem to cause different types of hyperlipidemia. Our study suggests that early diagnosis and long-term treatment of type 2 diabetes may be beneficial for LDL-C, which is the lipid element with the most strongly confirmed involvement in atherosclerosis development, and thus lower LDL-C levels could aid the prevention of CVD. Independent MR studies are needed to validate our findings in other ethnicities. Abbreviations: Abbreviations: BMI body mass index BMI_GRS body mass index genetic risk score CI confidence interval CVD cardiovascular disease DM_GRS diabetes mellitus genetic risk score FPG fasting plasma glucose GRS genetic risk score HbA1c glycated hemoglobin HDL-C high-density lipoprotein cholesterol IV instrumental variable LDL-C low-density lipoprotein cholesterol MR Mendelian randomization OR odds ratio SD standard deviation SNP single-nucleotide polymorphism SPECT Survey on Prevalence in East China for metabolic diseases and risk factors TC total cholesterol TG triglyceride Acknowledgments We thank thank Xiaojin Wang and Bingshun Wang from the Department of Biostatistics and Shanghai Jiaotong University School of Medicine for data processing, Weiping Tu, Bin Li, and Ling Hu for helping to organize this investigation, and all team members and participants in the SPECT-China study. Financial Support: This work was supported by the National Natural Science Foundation of China (81570726 to Y.L. and 81600609 to N.W.), the Shanghai JiaoTong University School of Medicine (2014 to Y.L.), the Science and Technology Commission of Shanghai Municipality (16411971200 to Y.L. and 16410723200 to N.W.), the Fourth Round of the Three-Year Public Health Action Plan of Shanghai by the Shanghai Municipal Commission of Health and Family Planning (15GWZK0202 to Y.L. and 20164Y0079 to N.W.), the Municipal Human Resources Development Program for Outstanding Young Talents in Medical and Health Sciences in Shanghai (2017YQ053 to N.W.), and the Clinical Research Plan of Shanghai Hospital Development Center (16CR3076B to Y.L.). Author Contributions: Y.L., L.G., and D.L. designed the study; N.W., J.C., Z.N., Y.C., B.H., Q.L., C.C., L.Z., and F.X conducted the research; and N.W., J.C., and Z.N. analyzed the data and wrote the manuscript. N.W., J.C., and Z.N. contributed equally. The final manuscript was read and approved by all authors. Disclosure Summary: The authors have nothing to disclose. References 1. Gersh BJ , Sliwa K , Mayosi BM , Yusuf S . Novel therapeutic concepts: the epidemic of cardiovascular disease in the developing world: global implications . Eur Heart J . 2010 ; 31 ( 6 ): 642 – 648 . Google Scholar CrossRef Search ADS PubMed 2. Moran A , Gu D , Zhao D , Coxson P , Wang YC , Chen CS , Liu J , Cheng J , Bibbins-Domingo K , Shen YM , He J , Goldman L . Future cardiovascular disease in china: markov model and risk factor scenario projections from the coronary heart disease policy model-china . Circ Cardiovasc Qual Outcomes . 2010 ; 3 ( 3 ): 243 – 252 . Google Scholar CrossRef Search ADS PubMed 3. Mooradian AD . Dyslipidemia in type 2 diabetes mellitus . Nat Clin Pract Endocrinol Metab . 2009 ; 5 ( 3 ): 150 – 159 . Google Scholar CrossRef Search ADS PubMed 4. Kannel WB . Lipids, diabetes, and coronary heart disease: insights from the Framingham Study . Am Heart J . 1985 ; 110 ( 5 ): 1100 – 1107 . Google Scholar CrossRef Search ADS PubMed 5. Yang W , Xiao J , Yang Z , Ji L , Jia W , Weng J , Lu J , Shan Z , Liu J , Tian H , Ji Q , Zhu D , Ge J , Lin L , Chen L , Guo X , Zhao Z , Li Q , Zhou Z , Shan G , He J ; China National Diabetes and Metabolic Disorders Study Investigators . Serum lipids and lipoproteins in Chinese men and women . Circulation . 2012 ; 125 ( 18 ): 2212 – 2221 . Google Scholar CrossRef Search ADS PubMed 6. Haffner SM ; American Diabetes Association . Dyslipidemia management in adults with diabetes . Diabetes Care . 2004 ; 27 ( Suppl 1 ): S68 – S71 . Google Scholar CrossRef Search ADS PubMed 7. Fall T , Hägg S , Mägi R , Ploner A , Fischer K , Horikoshi M , Sarin AP , Thorleifsson G , Ladenvall C , Kals M , Kuningas M , Draisma HH , Ried JS , van Zuydam NR , Huikari V , Mangino M , Sonestedt E , Benyamin B , Nelson CP , Rivera NV , Kristiansson K , Shen HY , Havulinna AS , Dehghan A , Donnelly LA , Kaakinen M , Nuotio ML , Robertson N , de Bruijn RF , Ikram MA , Amin N , Balmforth AJ , Braund PS , Doney AS , Döring A , Elliott P , Esko T , Franco OH , Gretarsdottir S , Hartikainen AL , Heikkilä K , Herzig KH , Holm H , Hottenga JJ , Hyppönen E , Illig T , Isaacs A , Isomaa B , Karssen LC , Kettunen J , Koenig W , Kuulasmaa K , Laatikainen T , Laitinen J , Lindgren C , Lyssenko V , Läärä E , Rayner NW , Männistö S , Pouta A , Rathmann W , Rivadeneira F , Ruokonen A , Savolainen MJ , Sijbrands EJ , Small KS , Smit JH , Steinthorsdottir V , Syvänen AC , Taanila A , Tobin MD , Uitterlinden AG , Willems SM , Willemsen G , Witteman J , Perola M , Evans A , Ferrières J , Virtamo J , Kee F , Tregouet DA , Arveiler D , Amouyel P , Ferrario MM , Brambilla P , Hall AS , Heath AC , Madden PA , Martin NG , Montgomery GW , Whitfield JB , Jula A , Knekt P , Oostra B , van Duijn CM , Penninx BW , Smith GD , Kaprio J , Samani NJ , Gieger C , Peters A , Wichmann HE , Boomsma DI , de Geus EJ , Tuomi T , Power C , Hammond CJ , Spector TD , Lind L , Orho-Melander M , Palmer CN , Morris AD , Groop L , Järvelin MR , Salomaa V , Vartiainen E , Hofman A , Ripatti S , Metspalu A , Thorsteinsdottir U , Stefansson K , Pedersen NL , McCarthy MI , Ingelsson E , Prokopenko I ; European Network for Genetic and Genomic Epidemiology (ENGAGE) consortium . The role of adiposity in cardiometabolic traits: a Mendelian randomization analysis . PLoS Med . 2013 ; 10 ( 6 ): e1001474 . Google Scholar CrossRef Search ADS PubMed 8. Smith GD , Ebrahim S . ‘Mendelian randomization’: can genetic epidemiology contribute to understanding environmental determinants of disease ? Int J Epidemiol . 2003 ; 32 ( 1 ): 1 – 22 . Google Scholar CrossRef Search ADS PubMed 9. Didelez V , Sheehan N . Mendelian randomization as an instrumental variable approach to causal inference . Stat Methods Med Res . 2007 ; 16 ( 4 ): 309 – 330 . Google Scholar CrossRef Search ADS PubMed 10. Wang N , Wang X , Han B , Li Q , Chen Y , Zhu C , Chen Y , Xia F , Cang Z , Zhu C , Lu M , Meng Y , Chen C , Lin D , Wang B , Jensen MD , Lu Y . Is exposure to famine in childhood and economic development in adulthood associated with diabetes ? J Clin Endocrinol Metab . 2015 ; 100 ( 12 ): 4514 – 4523 . Google Scholar CrossRef Search ADS PubMed 11. Wang N , Wang X , Li Q , Han B , Chen Y , Zhu C , Chen Y , Lin D , Wang B , Jensen MD , Lu Y . The famine exposure in early life and metabolic syndrome in adulthood . Clin Nutr . 2017 ; 36 ( 1 ): 253 – 259 . Google Scholar CrossRef Search ADS PubMed 12. Wang N , Zhang K , Han B , Li Q , Chen Y , Zhu C , Chen Y , Xia F , Zhai H , Jiang B , Shen Z , Lu Y . Follicle stimulating hormone, its novel association with sex hormone binding globulin in men and postmenopausal women . Endocrine . 2017 ; 56 ( 3 ): 649 – 657 . Google Scholar CrossRef Search ADS PubMed 13. Xu Y , Wang L , He J , Bi Y , Li M , Wang T , Wang L , Jiang Y , Dai M , Lu J , Xu M , Li Y , Hu N , Li J , Mi S , Chen CS , Li G , Mu Y , Zhao J , Kong L , Chen J , Lai S , Wang W , Zhao W , Ning G ; 2010 China Noncommunicable Disease Surveillance Group . Prevalence and control of diabetes in Chinese adults . JAMA . 2013 ; 310 ( 9 ): 948 – 959 . Google Scholar CrossRef Search ADS PubMed 14. Wang N , Chen C , Zhao L , Chen Y , Han B , Xia F , Cheng J , Li Q , Lu Y . Vitamin D and nonalcoholic fatty liver disease: bi-directional Mendelian randomization analysis . EBioMedicine . 2018 ; 28 : 187 – 193 . Google Scholar CrossRef Search ADS PubMed 15. Xu M , Huang Y , Xie L , Peng K , Ding L , Lin L , Wang P , Hao M , Chen Y , Sun Y , Qi L , Wang W , Ning G , Bi Y . Diabetes and risk of arterial stiffness: a Mendelian randomization analysis . Diabetes . 2016 ; 65 ( 6 ): 1731 – 1740 . Google Scholar CrossRef Search ADS PubMed 16. Huang Y , Xu M , Xie L , Wang T , Huang X , Lv X , Chen Y , Ding L , Lin L , Wang W , Bi Y , Sun Y , Zhang Y , Ning G . Obesity and peripheral arterial disease: a Mendelian randomization analysis . Atherosclerosis . 2016 ; 247 : 218 – 224 . Google Scholar CrossRef Search ADS PubMed 17. Cho YS , Lee JY , Park KS , Nho CW . Genetics of type 2 diabetes in East Asian populations . Curr Diab Rep . 2012 ; 12 ( 6 ): 686 – 696 . Google Scholar CrossRef Search ADS PubMed 18. Kato N . Insights into the genetic basis of type 2 diabetes . J Diabetes Investig . 2013 ; 4 ( 3 ): 233 – 244 . Google Scholar CrossRef Search ADS PubMed 19. Okada Y , Kubo M , Ohmiya H , Takahashi A , Kumasaka N , Hosono N , Maeda S , Wen W , Dorajoo R , Go MJ , Zheng W , Kato N , Wu JY , Lu Q , Tsunoda T , Yamamoto K , Nakamura Y , Kamatani N , Tanaka T ; GIANT consortium . Common variants at CDKAL1 and KLF9 are associated with body mass index in east Asian populations . Nat Genet . 2012 ; 44 ( 3 ): 302 – 306 . Google Scholar CrossRef Search ADS PubMed 20. Wen W , Cho YS , Zheng W , Dorajoo R , Kato N , Qi L , Chen CH , Delahanty RJ , Okada Y , Tabara Y , Gu D , Zhu D , Haiman CA , Mo Z , Gao YT , Saw SM , Go MJ , Takeuchi F , Chang LC , Kokubo Y , Liang J , Hao M , Le Marchand L , Zhang Y , Hu Y , Wong TY , Long J , Han BG , Kubo M , Yamamoto K , Su MH , Miki T , Henderson BE , Song H , Tan A , He J , Ng DP , Cai Q , Tsunoda T , Tsai FJ , Iwai N , Chen GK , Shi J , Xu J , Sim X , Xiang YB , Maeda S , Ong RT , Li C , Nakamura Y , Aung T , Kamatani N , Liu JJ , Lu W , Yokota M , Seielstad M , Fann CS , Wu JY , Lee JY , Hu FB , Tanaka T , Tai ES , Shu XO ; Genetic Investigation of ANthropometric Traits (GIANT) Consortium . Meta-analysis identifies common variants associated with body mass index in east Asians . Nat Genet . 2012 ; 44 ( 3 ): 307 – 311 . Google Scholar CrossRef Search ADS PubMed 21. Vimaleswaran KS , Berry DJ , Lu C , Tikkanen E , Pilz S , Hiraki LT , Cooper JD , Dastani Z , Li R , Houston DK , Wood AR , Michaëlsson K , Vandenput L , Zgaga L , Yerges-Armstrong LM , McCarthy MI , Dupuis J , Kaakinen M , Kleber ME , Jameson K , Arden N , Raitakari O , Viikari J , Lohman KK , Ferrucci L , Melhus H , Ingelsson E , Byberg L , Lind L , Lorentzon M , Salomaa V , Campbell H , Dunlop M , Mitchell BD , Herzig KH , Pouta A , Hartikainen AL , Streeten EA , Theodoratou E , Jula A , Wareham NJ , Ohlsson C , Frayling TM , Kritchevsky SB , Spector TD , Richards JB , Lehtimäki T , Ouwehand WH , Kraft P , Cooper C , März W , Power C , Loos RJ , Wang TJ , Järvelin MR , Whittaker JC , Hingorani AD , Hyppönen E ; Genetic Investigation of Anthropometric Traits-GIANT Consortium . Causal relationship between obesity and vitamin D status: bi-directional Mendelian randomization analysis of multiple cohorts . PLoS Med . 2013 ; 10 ( 2 ): e1001383 . Google Scholar CrossRef Search ADS PubMed 22. Lawlor DA , Harbord RM , Sterne JA , Timpson N , Davey Smith G . Mendelian randomization: using genes as instruments for making causal inferences in epidemiology . Stat Med . 2008 ; 27 ( 8 ): 1133 – 1163 . Google Scholar CrossRef Search ADS PubMed 23. Sheehan NA , Didelez V , Burton PR , Tobin MD . Mendelian randomisation and causal inference in observational epidemiology . PLoS Med . 2008 ; 5 ( 8 ): e177 . Google Scholar CrossRef Search ADS PubMed 24. Kraja AT , Chasman DI , North KE , Reiner AP , Yanek LR , Kilpeläinen TO , Smith JA , Dehghan A , Dupuis J , Johnson AD , Feitosa MF , Tekola-Ayele F , Chu AY , Nolte IM , Dastani Z , Morris A , Pendergrass SA , Sun YV , Ritchie MD , Vaez A , Lin H , Ligthart S , Marullo L , Rohde R , Shao Y , Ziegler MA , Im HK , Schnabel RB , Jørgensen T , Jørgensen ME , Hansen T , Pedersen O , Stolk RP , Snieder H , Hofman A , Uitterlinden AG , Franco OH , Ikram MA , Richards JB , Rotimi C , Wilson JG , Lange L , Ganesh SK , Nalls M , Rasmussen-Torvik LJ , Pankow JS , Coresh J , Tang W , Linda Kao WH , Boerwinkle E , Morrison AC , Ridker PM , Becker DM , Rotter JI , Kardia SL , Loos RJ , Larson MG , Hsu YH , Province MA , Tracy R , Voight BF , Vaidya D , O’Donnell CJ , Benjamin EJ , Alizadeh BZ , Prokopenko I , Meigs JB , Borecki IB ; Cross Consortia Pleiotropy Group ; Cohorts for Heart and ; Aging Research in Genetic Epidemiology ; Genetic Investigation of Anthropometric Traits Consortium ; Global Lipids Genetics Consortium ; Meta-Analyses of Glucose ; Insulin-related traits Consortium ; Global BPgen Consortium ; ADIPOGen Consortium ; Women’s Genome Health Study ; Howard University Family Study . Pleiotropic genes for metabolic syndrome and inflammation . Mol Genet Metab . 2014 ; 112 ( 4 ): 317 – 338 . Google Scholar CrossRef Search ADS PubMed 25. Catapano AL , Graham I , De Backer G , Wiklund O , Chapman MJ , Drexel H , Hoes AW , Jennings CS , Landmesser U , Pedersen TR , Reiner Ž , Riccardi G , Taskinen MR , Tokgozoglu L , Verschuren WMM , Vlachopoulos C , Wood DA , Zamorano JL , Cooney MT ; ESC Scientific Document Group . 2016 ESC/EAS guidelines for the management of dyslipidaemias . Eur Heart J . 2016 ; 37 ( 39 ): 2999 – 3058 . Google Scholar CrossRef Search ADS PubMed 26. Holmes MV , Lange LA , Palmer T , Lanktree MB , North KE , Almoguera B , Buxbaum S , Chandrupatla HR , Elbers CC , Guo Y , Hoogeveen RC , Li J , Li YR , Swerdlow DI , Cushman M , Price TS , Curtis SP , Fornage M , Hakonarson H , Patel SR , Redline S , Siscovick DS , Tsai MY , Wilson JG , van der Schouw YT , FitzGerald GA , Hingorani AD , Casas JP , de Bakker PI , Rich SS , Schadt EE , Asselbergs FW , Reiner AP , Keating BJ . Causal effects of body mass index on cardiometabolic traits and events: a Mendelian randomization analysis . Am J Hum Genet . 2014 ; 94 ( 2 ): 198 – 208 . Google Scholar CrossRef Search ADS PubMed 27. Meyre D . Is FTO a type 2 diabetes susceptibility gene ? Diabetologia . 2012 ; 55 ( 4 ): 873 – 876 . Google Scholar CrossRef Search ADS PubMed 28. Rydén L , Standl E , Bartnik M , Van den Berghe G , Betteridge J , de Boer MJ , Cosentino F , Jönsson B , Laakso M , Malmberg K , Priori S , Ostergren J , Tuomilehto J , Thrainsdottir I , Vanhorebeek I , Stramba-Badiale M , Lindgren P , Qiao Q , Priori SG , Blanc JJ , Budaj A , Camm J , Dean V , Deckers J , Dickstein K , Lekakis J , McGregor K , Metra M , Morais J , Osterspey A , Tamargo J , Zamorano JL , Deckers JW , Bertrand M , Charbonnel B , Erdmann E , Ferrannini E , Flyvbjerg A , Gohlke H , Juanatey JR , Graham I , Monteiro PF , Parhofer K , Pyörälä K , Raz I , Schernthaner G , Volpe M , Wood D ; Task Force on Diabetes and Cardiovascular Diseases of the European Society of Cardiology (ESC) ; European Association for the Study of Diabetes (EASD) . Guidelines on diabetes, pre-diabetes, and cardiovascular diseases: executive summary . Eur Heart J . 2007 ; 28 ( 1 ): 88 – 136 . Google Scholar PubMed 29. Ortega FB , Lavie CJ , Blair SN . Obesity and cardiovascular disease . Circ Res . 2016 ; 118 ( 11 ): 1752 – 1770 . Google Scholar CrossRef Search ADS PubMed 30. Jensen MD , Ryan DH , Apovian CM , Ard JD , Comuzzie AG , Donato KA , Hu FB , Hubbard VS , Jakicic JM , Kushner RF , Loria CM , Millen BE , Nonas CA , Pi-Sunyer FX , Stevens J , Stevens VJ , Wadden TA , Wolfe BM , Yanovski SZ , Jordan HS , Kendall KA , Lux LJ , Mentor-Marcel R , Morgan LC , Trisolini MG , Wnek J , Anderson JL , Halperin JL , Albert NM , Bozkurt B , Brindis RG , Curtis LH , DeMets D , Hochman JS , Kovacs RJ , Ohman EM , Pressler SJ , Sellke FW , Shen WK , Smith SC Jr , Tomaselli GF ; American College of Cardiology/American Heart Association Task Force on Practice Guidelines ; Obesity Society . 2013 AHA/ACC/TOS guideline for the management of overweight and obesity in adults: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines and The Obesity Society . Circulation . 2014 ; 129 ( 25 , Suppl 2 ) S102 – S138 . Google Scholar CrossRef Search ADS PubMed 31. Patel A , MacMahon S , Chalmers J , Neal B , Billot L , Woodward M , Marre M , Cooper M , Glasziou P , Grobbee D , Hamet P , Harrap S , Heller S , Liu L , Mancia G , Mogensen CE , Pan C , Poulter N , Rodgers A , Williams B , Bompoint S , de Galan BE , Joshi R , Travert F , Travert F ; ADVANCE Collaborative Group . Intensive blood glucose control and vascular outcomes in patients with type 2 diabetes . N Engl J Med . 2008 ; 358 ( 24 ): 2560 – 2572 . Google Scholar CrossRef Search ADS PubMed 32. Gerstein HC , Miller ME , Byington RP , Goff DC Jr , Bigger JT , Buse JB , Cushman WC , Genuth S , Ismail-Beigi F , Grimm RH Jr , Probstfield JL , Simons-Morton DG , Friedewald WT ; Action to Control Cardiovascular Risk in Diabetes Study Group . Effects of intensive glucose lowering in type 2 diabetes . N Engl J Med . 2008 ; 358 ( 24 ): 2545 – 2559 . Google Scholar CrossRef Search ADS PubMed 33. Ray KK , Seshasai SR , Wijesuriya S , Sivakumaran R , Nethercott S , Preiss D , Erqou S , Sattar N . Effect of intensive control of glucose on cardiovascular outcomes and death in patients with diabetes mellitus: a meta-analysis of randomised controlled trials . Lancet . 2009 ; 373 ( 9677 ): 1765 – 1772 . Google Scholar CrossRef Search ADS PubMed 34. Wada K , Yatsuya H , Ouyang P , Otsuka R , Mitsuhashi H , Takefuji S , Matsushita K , Sugiura K , Hotta Y , Toyoshima H , Tamakoshi K . Self-reported medical history was generally accurate among Japanese workplace population . J Clin Epidemiol . 2009 ; 62 ( 3 ): 306 – 313 . Google Scholar CrossRef Search ADS PubMed Copyright © 2018 Endocrine Society http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Clinical Endocrinology and Metabolism Oxford University Press

Type 2 Diabetes and Adiposity Induce Different Lipid Profile Disorders: A Mendelian Randomization Analysis

Loading next page...
 
/lp/ou_press/type-2-diabetes-and-adiposity-induce-different-lipid-profile-disorders-YCUiz1a3Dv
Publisher
Oxford University Press
Copyright
Copyright © 2018 Endocrine Society
ISSN
0021-972X
eISSN
1945-7197
D.O.I.
10.1210/jc.2017-02789
Publisher site
See Article on Publisher Site

Abstract

Abstract Context Type 2 diabetes and obesity often coexist, so it is difficult to judge whether diabetes or obesity induce certain types of hyperlipidemia due to mutual confounds and reverse causation. We used Mendelian randomization analyses to explore the causal relationships of diabetes and adiposity with lipid profiles. Design, Setting, and Main Outcome Measures From 23 sites in East China, 9798 participants were enrolled during 2014 to 2016. We calculated two weighted genetic risk scores as instrumental variables for type 2 diabetes and body mass index (BMI). These scores were used to measure the causal relationships of diabetes and BMI with lipid profiles that included total cholesterol, high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), and triglycerides (TGs). Results The causal regression coefficients (βIV) of genetically determined diabetes for the total cholesterol, LDL-C, and log10TG were 0.130 [95% confidence interval (CI): 0.020, 0.240; P = 0.014], 0.125 (96% CI: 0.041, 0.209; P = 0.001), and 0.019 (95% CI: –0.001, 0.039; P = 0.055), respectively. The βIV for HDL-C was –0.008 (95% CI: –0.032. 0.016), which was not significant (P = 0.699). The causal regression coefficients of a genetically determined 10 kg/m2 increase in BMI for HDL-C and log10TG were –0.409 (96% CI: –0.698, –0.120; P = 0.004) and 0.227 (95% CI: 0.039, 0.415; P = 0.026), respectively. The βIVs for TGs and LDL-C were not significant. Conclusions This study has provided evidence for the biologically plausible causal effects of diabetes and adiposity by BMI on different elements of the lipid profile using Mendelian randomization analyses. Atherosclerotic cardiovascular disease (CVD) is the leading cause of death worldwide (1). The incidence and prevalence of CVD are continuously increasing in parallel with the increase in dyslipidemia, especially in low- and middle-income countries such as China (1). According to the Coronary Heart Disease Policy Model-China, total cholesterol (TC) increases of 0.58 mmol/L (22.4 mg/dL) and 0.55 mmol/L (21.6 mg/dL) in Chinese men and women, respectively, from 2010 to 2030 would lead to approximately 960 million new CVD events in total, which represents the largest increase in CVD events among all modeled risk factors, including diabetes and body mass index (BMI) (2). Many well-designed epidemiological studies have demonstrated the associations of diabetes and obesity with lipid profile (3–5). Previous studies have reported that the most common pattern of dyslipidemia in patients with type 2 diabetes includes elevated triglyceride (TG) levels and decreased high-density lipoprotein cholesterol (HDL-C) levels (6). The concentration of low-density lipoprotein cholesterol (LDL-C) exhibits no significant increase (6). In obesity, hypercholesterolemia, hypertriglyceridemia, and high LDL-C and low HDL-C levels are very common (5). However, diabetes and obesity have often coexisted in the samples of previous studies, so it is difficult to judge whether diabetes or obesity induces certain types of hyperlipidemia due to mutual confounds, reverse causation, and other issues with conventional observational studies that ultimately lead to imprecision in the estimation of both the directions and magnitudes of the effects (7). An alternative to classical observational studies are Mendelian randomization (MR) studies. MR uses genetic variants associated with an intermediate phenotype (in the current study, diabetes and BMI) as instrumental variables (IVs) in nonexperimental data to make causal inferences about the effect of an exposure on an outcome (Fig. 1) (8). Because genetic variants are assumed to be randomly distributed within a population and exist before the outcome, the IV is regarded as independent of the confounders that affect the exposure-outcome relationship and guide definite causal direction (9). In this study, if type 2 diabetes and BMI causally induce certain types of hyperlipidemia, then genetic variants associated with type 2 diabetes and BMI could be expected to affect certain types of hyperlipidemia. Figure 1. View largeDownload slide Study design and tested associations. BMI_GRS, body mass index genetic risk score; DM_GRS, diabetes mellitus genetic risk score. Figure 1. View largeDownload slide Study design and tested associations. BMI_GRS, body mass index genetic risk score; DM_GRS, diabetes mellitus genetic risk score. Based on the large community-based sample of Chinese participants from the Survey on Prevalence in East China for metabolic diseases and risk factors (SPECT)-China study (www.chictr.org.cn; registration no. ChiCTR-ECS-14005052), we performed MR analyses to explore the causal associations of diabetes and BMI with lipid profiles, including TC, LDL-C, HDL-C, and TG. Diabetes mellitus genetic risk scores (DM_GRSs) and BMI genetic risk scores (BMI_GRSs) were constructed to represent genetic susceptibility, and we evaluated the causal links of genetically determined diabetes and genetically determined BMI with lipid profiles. Methods Study design and participants Figure 1 presents the study design. We analyzed the observed associations between the measured variables (residual confounds and reverse causation) and the genetic associations between genotypes and the measured variables (no residual confounds or reverse causation). First, we assessed the association between type 2 diabetes mellitus–related genetic variants and present diabetes (Fig. 1A) and the relation between BMI-related genetic variants and measured BMI (Fig. 1A). Second, we measured the effects of genetic variants associated with type 2 diabetes and BMI on the changes in lipid profiles, including TC, LDL-C, HDL-C, and TG (Fig. 1C). Third, we performed observational assessments of the relation between present diabetes and lipid profile and between BMI and lipid profile (Fig. 1B). The data were taken from the ongoing SPECT-China study, which is a cross-sectional investigation of the prevalence of metabolic diseases and risk factors in East China. Recruitment and enrollment have previously been described in detail (10–12). Chinese citizens ≥18 years old who had lived in their current area for ≥6 months were included. From 2014 to 2016, 12,666 subjects who were 18 to 93 years old from 23 sites in the Shanghai, Zhejiang, Jiangsu, Anhui, and Jiangxi provinces were recruited into the SPECT-China study. Among these participants, genotype information was available for 10,516 (83.0%). We excluded participants with missing information on more than two single-nucleotide polymorphism (SNP) genotypes (n = 297), missing lipid profile values (n = 3), and missing measured BMI (n = 242) and those who were using lipid-lowering drugs (n = 176). In total, 9,798 participants were involved in the final analyses. The study protocol was approved by the Ethics Committee of Shanghai Ninth People’s Hospital, Shanghai JiaoTong University School of Medicine. The study protocol conformed to the ethical guidelines of the 1975 Declaration of Helsinki as reflected in the a priori approval by the appropriate institutional review committee. Informed consent was obtained from all participants included in the study. Procedures A single-assessment protocol was adopted by trained staff during the interviews and the collection of biological specimens at each study site. Blood samples were obtained between 7:00 am and 10:00 am after fasting for at least 8 hours. The blood was refrigerated immediately after phlebotomy. After 2 to 4 hours, the blood was centrifuged, and the serum was aliquoted and frozen in a central laboratory. Glycated hemoglobin (HbA1c) was measured by high-performance liquid chromatography (MQ-2000PT; Medconn, Shanghai, China). Fasting plasma glucose (FPG), TG, TC, HDL-C, and LDL-C were measured with a Beckman Coulter AU680 (Brea, CA). BMI was calculated as the weight in kilograms divided by the height in meters squared. Body weight, height, and blood pressure were measured using standard methods as described previously (13). Diabetes was defined by the following criteria (1): a self-reported previous diagnosis by a health care professional (13), (2) an FPG level of 7.0 mmol/L or higher, or (3) an HbA1c concentration of 6.5% or more. Hypertension was assessed as a systolic blood pressure ≥140 mm Hg, a diastolic blood pressure ≥90 mm Hg, or a self-reported previous diagnosis of hypertension by a physician. The 16 study sites were categorized into high and low economic status according to the gross domestic product per capita of the whole nation in 2013 (10). DNA was extracted from white bloods cells using a blood genomic DNA extraction kit (DP603; TIANGEN BIOTECH CO, LTD, Beijing, China) on an automated nucleic acid extraction instrument (YOSE-S32; TIANGEN BIOTECH CO, LTD, Beijing, China). Specific assays were designed using Geneious Pro (version 4.8.3; www.geneious.com/, Auckland, New Zealand). Mass determination was performed with JUNO, and data acquisition was performed using Fluidigm SNP Genotyping Analysis version 4.1.3 software (Fluidigm Corporation, San Francisco, CA) (14). The researchers were masked to the data regarding present diabetes, BMI, and phenotype. The call rates of all SNPs were higher than 98%. For the selection of the diabetes- and BMI-related SNPs, we mainly considered the SNPs that have been identified in genome-wide association studies or replicated in Asians and then successfully used in two Chinese MR studies (15, 16). For diabetes, we a priori selected 18 SNPs that were identified and validated in meta-analyses including genome-wide association studies in East Asians (17, 18). A previous Chinese MR study (15) included 34 SNPs. To minimize the pleiotropic effects, we further selected the 18 SNPs that exhibited no significant associations with BMI or systolic and diastolic blood pressure in that study (15). These SNPs were TCF2 (HNF1B) rs4430796, GCC1/PAX4 rs6467136, MAEA rs6815464, GLIS3 rs7041847, C2CD4A/C2CD4B rs7172432, TCF7L2 rs7903146, ZFAND3 rs9470794, CDKN2A/B rs10811661, CDC123/CAMK1D rs12779790, PROX1 rs340874, SRR rs391300, PSMD6 rs831571, JAZF1 rs864745, TP53INP1 rs896854, SPRY2 rs1359790, PPARG rs1801282, KCNQ1 rs2237892, and IRS1 rs2943641. For BMI, 14 SNPs from a previous MR study (16) that were selected from genome-wide association studies in East Asians (19, 20) were adopted. These SNPs included TFAP2B rs4715210, MAP2K5 rs4776970, ADCY3-DNAJC27 rs6545814, MC4R rs6567160, CDKAL1 rs9356744, GNPDA2 rs10938397, SMC5-KLF9 rs11142387, GIPR-QPCTL rs11671664, GP2 rs12597579, FTO rs17817449, BDNF rs6265, PCSK1 rs261967, SEC16B rs574367, and PAX6 rs652722. There were no linkage disequilibrium relationships (r2 = 0) among the previous diabetes and BMI SNPs (15, 16). The full list of each SNP is presented in Supplemental Table 1. Statistical analysis The data analyses were performed using IBM SPSS Statistics version 22 (IBM Corporation, Armonk, NY). All analyses were two sided. A P value <0.05 indicated significance. Continuous variables are expressed as the mean ± the standard deviation (SD), and categorical variables are expressed as percentages (%). Serum TG was logarithmically transformed before the analysis. The P values for trends of the characteristics according to the genetic risk score (GRS) quintiles were calculated with one-way analysis of variance and χ2 tests. The additive genetic model for each SNP (coded as 0-2) was used to construct the DM_GRS and BMI_GRS. The scores were defined as the sum of the number of BMI-increasing and diabetes risk-increasing alleles at each locus multiplied by the respective β coefficient as reported from the meta-analysis. Next, the total scores were divided by the average effect sizes of these SNPs. Those who were missing one or two SNPs were imputed by the regression method using covariates. The missing rates of one or two diabetes- and BMI-related SNPs were 16.4% and 15.5%, respectively. There were also 82 missing blood pressure values, which were not imputed. For our analyses of the association between DM_GRS and present diabetes and that between BMI_GRS and measured BMI (Fig. 1A), we used logistic and linear regression analyses. The data were adjusted for age, sex, BMI (only in the DM_GRS model), diabetes (only in the BMI_GRS model), hypertension, and economic status. To assess the reliabilities and strengths of the GRSs as instruments, the F statistic was adopted (F = [R2 × (n–2)]/[1–R2]) (21). An F value greater than 10 indicates that the instrument is sufficiently strong as an IV, and the R2 is the proportion of variation in the respective phenotype that is explained by the genotype (22). We used regression analyses to calculate the β coefficients for the associations of present diabetes and BMI with lipid profile, including TC, LDL-C, HDL-C, and TG (Fig. 1B). Then, we also investigated the associations of the DM_GRS and BMI_GRS with lipid profile (Fig. 1C) by linear regressions that were adjusted for age, sex, BMI (only in the DM_GRS model), diabetes (only in the BMI_GRS model), hypertension, and economic status. The association of each individual SNP with lipid profile was analyzed by linear regression analysis with an unadjusted model. Next, we calculated the IV estimates of the genetically determined β coefficients with the Wald-type estimator (7, 22). For abnormal lipid profile trends due to present diabetes and increased BMI, the computational formulae were βIV(DM-lipid) = βDM_GRS-lipid/βDM_GRS-DM and βIV(BMI-lipid) = βBMI_GRS-lipid/βBMI_GRS-BMI, respectively. The standard errors and confidence intervals (CIs) for the IV estimators were estimated by the δ method. The formulae are: SEIV=abs(βIV)(SEGRS_exposureβGRS_exposure)2+(SEGRS_eoutcomeβGRS_outcome)2 95% CIIV=βIV±1.96×SEIV According to a previous study, a 10 kg/m2 BMI increment was used because this increment roughly reflects a shift from normal weight to obesity. Thus, to increase the parity between the observational and genetic estimates, the effect of the BMI_GRS was extrapolated to correspond to a 10 kg/m2 increase in BMI for the genetic analyses. A GRS as an IV should meet three assumptions (23). First, the GRS should be associated with the exposure. This assumption is confirmed in the aforementioned Fig. 1A. Second, the GRS should be independent of the confounding factors. Thus, we also measured the associations of each individual SNP and GRS with BMI, blood pressure, FPG, HbA1c, and age. Third, the two GRSs should have effects on lipid profile that are mediated solely through diabetes and BMI (i.e., there are no pleiotropic effects) (23). However, this is only an idealized view and could only be verified to a limited extent. We additionally adjusted the HbA1c or BMI in the previous models to observe the changes in the associations of the DM_GRS or BMI_GRS with the lipid profiles. In the sensitivity analysis, because the BMI_GRS exhibited significant associations with age and hypertension, in consideration of pleiotropy, we constructed a BMI_GRS12SNP that excluded the most obvious loci that may have been involved in the regulation of other traits (i.e., FTO rs17817449 and MC4R rs6567160) (24). Results Associations of SNPs with exposure and outcome We tested the associations of the diabetes- and BMI-related SNPs with exposure (present diabetes and measured BMI) and outcome (lipid profile). The unstandardized coefficients (95% CIs) and odds ratios (ORs) are presented in Tables 1 and 2 and Supplemental Figs. 1 and 2. Among the 18 diabetes-related SNPs, seven SNPs located at the GLIS3, CDKN2A/B, CDC123/CAMK1D, TP53INP1, PPARG, KCNQ1, and IRS1 loci were significantly associated with present diabetes (Supplemental Fig. 1). Among the 14 BMI-related SNPs, eight SNPs located at the MAP2K5, ADCY3-DNAJC27, MC4R, CDKAL1, GIPR-QPCTL, FTO, BDNF, and SEC16B loci were significantly associated with BMI (Supplemental Fig. 2). Table 1. Association of the 18 Diabetes-Related SNPs With Lipid Profile Gene SNP TC LDL-C HDL-C log10TG TCF2 (HNF1B) rs4430796 0.047 (0.009, 0.085) 0.031 (0.004, 0.059) −0.002 (−0.013, 0.009) 0.003 (−0.005, 0.011) GCC1/PAX4 rs6467136 0.077 (0.028, 0.125) 0.042 (0.008, 0.077) 0.014 (0.001, 0.028) 0.003 (−0.007, 0.014) MAEA rs6815464 −0.016 (−0.052, 0.019) −0.009 (−0.035, 0.016) 0.007 (−0.003, 0.017) 0.001 (−0.006, 0.009) GLIS3 rs7041847 0.031 (−0.003, 0.066) 0.027 (0.002, 0.051) 0.005 (−0.005, 0.014) 0.001 (−0.006, 0.008) C2CD4A/C2CD4B rs7172432 0.001 (−0.034, 0.036) 0.010 (−0.015, 0.035) −0.002 (−0.012, 0.008) −0.002 (−0.010, 0.005) TCF7L2 rs7903146 0.036 (−0.056, 0.129) 0.043 (−0.024, 0.109) 0.006 (−0.020, 0.032) −0.006 (−0.025, 0.014) ZFAND3 rs9470794 −0.017 (−0.054, 0.020) −0.002 (−0.029, 0.024) −0.007 (−0.018, 0.004) 0.004 (−0.003, 0.012) CDKN2A/B rs10811661 0.018 (−0.016, 0.053) 0.017 (−0.007, 0.042) 0.004 (−0.006, 0.014) 0.004 (−0.003, 0.011) CDC123/CAMK1D rs12779790 0.050 (0.004, 0.096) 0.017 (−0.016, 0.050) 0.003 (−0.010, 0.016) 0.003 (−0.007, 0.012) PROX1 rs340874 0.008 (−0.028, 0.043) 0.005 (−0.021, 0.030) 0.008 (−0.002, 0.018) 0.002 (−0.005, 0.010) SRR rs391300 0.004 (−0.033, 0.042) 0.011 (−0.015, 0.038) −0.001 (−0.011, 0.010) 0.000 (−0.008, 0.008) PSMD6 rs831571 0.038 (0.001, 0.074) 0.038 (0.012, 0.064) 0.003 (−0.007, 0.014) 0.003 (−0.004, 0.011) JAZF1 rs864745 0.043 (0.003, 0.084) 0.032 (0.003, 0.061) −0.001 (−0.013, 0.010) 0.000 (−0.004, 0.011) TP53INP1 rs896854 0.016 (−0.022, 0.053) 0.006 (−0.021, 0.033) 0.003 (−0.008, 0.014) 0.004 (−0.004, 0.012) SPRY2 rs1359790 −0.011 (−0.050, 0.028) −0.014 (−0.042, 0.014) −0.005 (−0.016, 0.006) −0.007 (−0.016, 0.001) PPARG rs1801282 0.013 (−0.061, 0.087) 0.011 (−0.042, 0.064) 0.014 (−0.007, 0.035) 0.015 (0.000, 0.031) KCNQ1 rs2237892 0.014 (−0.023, 0.051) 0.014 (−0.012, 0.040) −0.002 (−0.013, 0.008) 0.001 (−0.007, 0.009) IRS1 rs2943641 0.012 (−0.057, 0.081) −0.011 (−0.061, 0.038) −0.019 (−0.039, 0.001) 0.017 (0.003, 0.032) Gene SNP TC LDL-C HDL-C log10TG TCF2 (HNF1B) rs4430796 0.047 (0.009, 0.085) 0.031 (0.004, 0.059) −0.002 (−0.013, 0.009) 0.003 (−0.005, 0.011) GCC1/PAX4 rs6467136 0.077 (0.028, 0.125) 0.042 (0.008, 0.077) 0.014 (0.001, 0.028) 0.003 (−0.007, 0.014) MAEA rs6815464 −0.016 (−0.052, 0.019) −0.009 (−0.035, 0.016) 0.007 (−0.003, 0.017) 0.001 (−0.006, 0.009) GLIS3 rs7041847 0.031 (−0.003, 0.066) 0.027 (0.002, 0.051) 0.005 (−0.005, 0.014) 0.001 (−0.006, 0.008) C2CD4A/C2CD4B rs7172432 0.001 (−0.034, 0.036) 0.010 (−0.015, 0.035) −0.002 (−0.012, 0.008) −0.002 (−0.010, 0.005) TCF7L2 rs7903146 0.036 (−0.056, 0.129) 0.043 (−0.024, 0.109) 0.006 (−0.020, 0.032) −0.006 (−0.025, 0.014) ZFAND3 rs9470794 −0.017 (−0.054, 0.020) −0.002 (−0.029, 0.024) −0.007 (−0.018, 0.004) 0.004 (−0.003, 0.012) CDKN2A/B rs10811661 0.018 (−0.016, 0.053) 0.017 (−0.007, 0.042) 0.004 (−0.006, 0.014) 0.004 (−0.003, 0.011) CDC123/CAMK1D rs12779790 0.050 (0.004, 0.096) 0.017 (−0.016, 0.050) 0.003 (−0.010, 0.016) 0.003 (−0.007, 0.012) PROX1 rs340874 0.008 (−0.028, 0.043) 0.005 (−0.021, 0.030) 0.008 (−0.002, 0.018) 0.002 (−0.005, 0.010) SRR rs391300 0.004 (−0.033, 0.042) 0.011 (−0.015, 0.038) −0.001 (−0.011, 0.010) 0.000 (−0.008, 0.008) PSMD6 rs831571 0.038 (0.001, 0.074) 0.038 (0.012, 0.064) 0.003 (−0.007, 0.014) 0.003 (−0.004, 0.011) JAZF1 rs864745 0.043 (0.003, 0.084) 0.032 (0.003, 0.061) −0.001 (−0.013, 0.010) 0.000 (−0.004, 0.011) TP53INP1 rs896854 0.016 (−0.022, 0.053) 0.006 (−0.021, 0.033) 0.003 (−0.008, 0.014) 0.004 (−0.004, 0.012) SPRY2 rs1359790 −0.011 (−0.050, 0.028) −0.014 (−0.042, 0.014) −0.005 (−0.016, 0.006) −0.007 (−0.016, 0.001) PPARG rs1801282 0.013 (−0.061, 0.087) 0.011 (−0.042, 0.064) 0.014 (−0.007, 0.035) 0.015 (0.000, 0.031) KCNQ1 rs2237892 0.014 (−0.023, 0.051) 0.014 (−0.012, 0.040) −0.002 (−0.013, 0.008) 0.001 (−0.007, 0.009) IRS1 rs2943641 0.012 (−0.057, 0.081) −0.011 (−0.061, 0.038) −0.019 (−0.039, 0.001) 0.017 (0.003, 0.032) The data are presented as unstandardized coefficients and 95% CIs using linear regression analyses. The model was unadjusted. View Large Table 1. Association of the 18 Diabetes-Related SNPs With Lipid Profile Gene SNP TC LDL-C HDL-C log10TG TCF2 (HNF1B) rs4430796 0.047 (0.009, 0.085) 0.031 (0.004, 0.059) −0.002 (−0.013, 0.009) 0.003 (−0.005, 0.011) GCC1/PAX4 rs6467136 0.077 (0.028, 0.125) 0.042 (0.008, 0.077) 0.014 (0.001, 0.028) 0.003 (−0.007, 0.014) MAEA rs6815464 −0.016 (−0.052, 0.019) −0.009 (−0.035, 0.016) 0.007 (−0.003, 0.017) 0.001 (−0.006, 0.009) GLIS3 rs7041847 0.031 (−0.003, 0.066) 0.027 (0.002, 0.051) 0.005 (−0.005, 0.014) 0.001 (−0.006, 0.008) C2CD4A/C2CD4B rs7172432 0.001 (−0.034, 0.036) 0.010 (−0.015, 0.035) −0.002 (−0.012, 0.008) −0.002 (−0.010, 0.005) TCF7L2 rs7903146 0.036 (−0.056, 0.129) 0.043 (−0.024, 0.109) 0.006 (−0.020, 0.032) −0.006 (−0.025, 0.014) ZFAND3 rs9470794 −0.017 (−0.054, 0.020) −0.002 (−0.029, 0.024) −0.007 (−0.018, 0.004) 0.004 (−0.003, 0.012) CDKN2A/B rs10811661 0.018 (−0.016, 0.053) 0.017 (−0.007, 0.042) 0.004 (−0.006, 0.014) 0.004 (−0.003, 0.011) CDC123/CAMK1D rs12779790 0.050 (0.004, 0.096) 0.017 (−0.016, 0.050) 0.003 (−0.010, 0.016) 0.003 (−0.007, 0.012) PROX1 rs340874 0.008 (−0.028, 0.043) 0.005 (−0.021, 0.030) 0.008 (−0.002, 0.018) 0.002 (−0.005, 0.010) SRR rs391300 0.004 (−0.033, 0.042) 0.011 (−0.015, 0.038) −0.001 (−0.011, 0.010) 0.000 (−0.008, 0.008) PSMD6 rs831571 0.038 (0.001, 0.074) 0.038 (0.012, 0.064) 0.003 (−0.007, 0.014) 0.003 (−0.004, 0.011) JAZF1 rs864745 0.043 (0.003, 0.084) 0.032 (0.003, 0.061) −0.001 (−0.013, 0.010) 0.000 (−0.004, 0.011) TP53INP1 rs896854 0.016 (−0.022, 0.053) 0.006 (−0.021, 0.033) 0.003 (−0.008, 0.014) 0.004 (−0.004, 0.012) SPRY2 rs1359790 −0.011 (−0.050, 0.028) −0.014 (−0.042, 0.014) −0.005 (−0.016, 0.006) −0.007 (−0.016, 0.001) PPARG rs1801282 0.013 (−0.061, 0.087) 0.011 (−0.042, 0.064) 0.014 (−0.007, 0.035) 0.015 (0.000, 0.031) KCNQ1 rs2237892 0.014 (−0.023, 0.051) 0.014 (−0.012, 0.040) −0.002 (−0.013, 0.008) 0.001 (−0.007, 0.009) IRS1 rs2943641 0.012 (−0.057, 0.081) −0.011 (−0.061, 0.038) −0.019 (−0.039, 0.001) 0.017 (0.003, 0.032) Gene SNP TC LDL-C HDL-C log10TG TCF2 (HNF1B) rs4430796 0.047 (0.009, 0.085) 0.031 (0.004, 0.059) −0.002 (−0.013, 0.009) 0.003 (−0.005, 0.011) GCC1/PAX4 rs6467136 0.077 (0.028, 0.125) 0.042 (0.008, 0.077) 0.014 (0.001, 0.028) 0.003 (−0.007, 0.014) MAEA rs6815464 −0.016 (−0.052, 0.019) −0.009 (−0.035, 0.016) 0.007 (−0.003, 0.017) 0.001 (−0.006, 0.009) GLIS3 rs7041847 0.031 (−0.003, 0.066) 0.027 (0.002, 0.051) 0.005 (−0.005, 0.014) 0.001 (−0.006, 0.008) C2CD4A/C2CD4B rs7172432 0.001 (−0.034, 0.036) 0.010 (−0.015, 0.035) −0.002 (−0.012, 0.008) −0.002 (−0.010, 0.005) TCF7L2 rs7903146 0.036 (−0.056, 0.129) 0.043 (−0.024, 0.109) 0.006 (−0.020, 0.032) −0.006 (−0.025, 0.014) ZFAND3 rs9470794 −0.017 (−0.054, 0.020) −0.002 (−0.029, 0.024) −0.007 (−0.018, 0.004) 0.004 (−0.003, 0.012) CDKN2A/B rs10811661 0.018 (−0.016, 0.053) 0.017 (−0.007, 0.042) 0.004 (−0.006, 0.014) 0.004 (−0.003, 0.011) CDC123/CAMK1D rs12779790 0.050 (0.004, 0.096) 0.017 (−0.016, 0.050) 0.003 (−0.010, 0.016) 0.003 (−0.007, 0.012) PROX1 rs340874 0.008 (−0.028, 0.043) 0.005 (−0.021, 0.030) 0.008 (−0.002, 0.018) 0.002 (−0.005, 0.010) SRR rs391300 0.004 (−0.033, 0.042) 0.011 (−0.015, 0.038) −0.001 (−0.011, 0.010) 0.000 (−0.008, 0.008) PSMD6 rs831571 0.038 (0.001, 0.074) 0.038 (0.012, 0.064) 0.003 (−0.007, 0.014) 0.003 (−0.004, 0.011) JAZF1 rs864745 0.043 (0.003, 0.084) 0.032 (0.003, 0.061) −0.001 (−0.013, 0.010) 0.000 (−0.004, 0.011) TP53INP1 rs896854 0.016 (−0.022, 0.053) 0.006 (−0.021, 0.033) 0.003 (−0.008, 0.014) 0.004 (−0.004, 0.012) SPRY2 rs1359790 −0.011 (−0.050, 0.028) −0.014 (−0.042, 0.014) −0.005 (−0.016, 0.006) −0.007 (−0.016, 0.001) PPARG rs1801282 0.013 (−0.061, 0.087) 0.011 (−0.042, 0.064) 0.014 (−0.007, 0.035) 0.015 (0.000, 0.031) KCNQ1 rs2237892 0.014 (−0.023, 0.051) 0.014 (−0.012, 0.040) −0.002 (−0.013, 0.008) 0.001 (−0.007, 0.009) IRS1 rs2943641 0.012 (−0.057, 0.081) −0.011 (−0.061, 0.038) −0.019 (−0.039, 0.001) 0.017 (0.003, 0.032) The data are presented as unstandardized coefficients and 95% CIs using linear regression analyses. The model was unadjusted. View Large Table 2. Association of the 14 BMI-Related SNPs With Lipid Profile Gene SNP TC LDL-C HDL-C log10TG TFAP2B rs4715210 0.002 (−0.044, 0.049) 0.015 (−0.019, 0.049) −0.002 (−0.015, 0.012) 0.006 (−0.004, 0.016) MAP2K5 rs4776970 −0.004 (−0.044, 0.036) 0.001 (−0.028, 0.030) −0.015 (−0.027, −0.004) 0.005 (−0.003, 0.014) ADCY3-DNAJC27 rs6545814 −0.007 (−0.042, 0.027) 0.019 (−0.006, 0.045) −0.007 (−0.017, 0.003) −0.002 (−0.010, 0.005) MC4R rs6567160 0.019 (−0.023, 0.060) 0.025 (−0.005, 0.055) −0.014 (−0.026, –0.002) 0.008 (−0.001, 0.017) CDKAL1 rs9356744 −0.019 (−0.053, 0.015) −0.020 (−0.045, 0.005) 0.001 (−0.008, 0.011) −0.001 (−0.008, 0.006) GNPDA2 rs10938397 0.024 (−0.012, 0.060) 0.012 (−0.014, 0.038) 0.007 (−0.004, 0.017) −0.002 (−0.010, 0.005) SMC5-KLF9 rs11142387 −0.012 (−0.048, 0.025) −0.002 (−0.029, 0.025) 0.001 (−0.009, 0.012) 0.005 (−0.003, 0.012) GIPR-QPCTL rs11671664 0.001 (−0.032, 0.035) −0.004 (−0.028, 0.021) 0.002 (−0.007, 0.012) 0.003 (−0.004, 0.010) GP2 rs12597579 0.014 (−0.023, 0.052) 0.008 (−0.019, 0.035) −0.004 (−0.015, 0.007) 0.006 (−0.002, 0.014) FTO rs17817449 0.033 (−0.020, 0.086) 0.029 (−0.009, 0.068) −0.015 (−0.031, 0.000) 0.017 (0.005, 0.028) BDNF rs6265 −0.015 (−0.048, 0.019) 0.008 (−0.016, 0.033) −0.012 (−0.022, −0.002) −0.003 (−0.010, 0.004) PCSK1 rs261967 −0.020 (−0.054, 0.014) −0.003 (−0.028, 0.022) −0.009 (−0.019, 0.001) 0.005 (−0.003, 0.012) SEC16B rs574367 0.027 (−0.016, 0.07) 0.006 (−0.025, 0.037) −0.013 (−0.025, 0.000) 0.008 (−0.002, 0.017) PAX6 rs652722 0.002 (−0.033, 0.037) 0.002 (−0.023, 0.028) 0.000 (−0.010, 0.010) 0.000 (−0.008, 0.007) Gene SNP TC LDL-C HDL-C log10TG TFAP2B rs4715210 0.002 (−0.044, 0.049) 0.015 (−0.019, 0.049) −0.002 (−0.015, 0.012) 0.006 (−0.004, 0.016) MAP2K5 rs4776970 −0.004 (−0.044, 0.036) 0.001 (−0.028, 0.030) −0.015 (−0.027, −0.004) 0.005 (−0.003, 0.014) ADCY3-DNAJC27 rs6545814 −0.007 (−0.042, 0.027) 0.019 (−0.006, 0.045) −0.007 (−0.017, 0.003) −0.002 (−0.010, 0.005) MC4R rs6567160 0.019 (−0.023, 0.060) 0.025 (−0.005, 0.055) −0.014 (−0.026, –0.002) 0.008 (−0.001, 0.017) CDKAL1 rs9356744 −0.019 (−0.053, 0.015) −0.020 (−0.045, 0.005) 0.001 (−0.008, 0.011) −0.001 (−0.008, 0.006) GNPDA2 rs10938397 0.024 (−0.012, 0.060) 0.012 (−0.014, 0.038) 0.007 (−0.004, 0.017) −0.002 (−0.010, 0.005) SMC5-KLF9 rs11142387 −0.012 (−0.048, 0.025) −0.002 (−0.029, 0.025) 0.001 (−0.009, 0.012) 0.005 (−0.003, 0.012) GIPR-QPCTL rs11671664 0.001 (−0.032, 0.035) −0.004 (−0.028, 0.021) 0.002 (−0.007, 0.012) 0.003 (−0.004, 0.010) GP2 rs12597579 0.014 (−0.023, 0.052) 0.008 (−0.019, 0.035) −0.004 (−0.015, 0.007) 0.006 (−0.002, 0.014) FTO rs17817449 0.033 (−0.020, 0.086) 0.029 (−0.009, 0.068) −0.015 (−0.031, 0.000) 0.017 (0.005, 0.028) BDNF rs6265 −0.015 (−0.048, 0.019) 0.008 (−0.016, 0.033) −0.012 (−0.022, −0.002) −0.003 (−0.010, 0.004) PCSK1 rs261967 −0.020 (−0.054, 0.014) −0.003 (−0.028, 0.022) −0.009 (−0.019, 0.001) 0.005 (−0.003, 0.012) SEC16B rs574367 0.027 (−0.016, 0.07) 0.006 (−0.025, 0.037) −0.013 (−0.025, 0.000) 0.008 (−0.002, 0.017) PAX6 rs652722 0.002 (−0.033, 0.037) 0.002 (−0.023, 0.028) 0.000 (−0.010, 0.010) 0.000 (−0.008, 0.007) The data are presented as unstandardized coefficients and 95% CIs using linear regression analyses. The model was unadjusted. View Large Table 2. Association of the 14 BMI-Related SNPs With Lipid Profile Gene SNP TC LDL-C HDL-C log10TG TFAP2B rs4715210 0.002 (−0.044, 0.049) 0.015 (−0.019, 0.049) −0.002 (−0.015, 0.012) 0.006 (−0.004, 0.016) MAP2K5 rs4776970 −0.004 (−0.044, 0.036) 0.001 (−0.028, 0.030) −0.015 (−0.027, −0.004) 0.005 (−0.003, 0.014) ADCY3-DNAJC27 rs6545814 −0.007 (−0.042, 0.027) 0.019 (−0.006, 0.045) −0.007 (−0.017, 0.003) −0.002 (−0.010, 0.005) MC4R rs6567160 0.019 (−0.023, 0.060) 0.025 (−0.005, 0.055) −0.014 (−0.026, –0.002) 0.008 (−0.001, 0.017) CDKAL1 rs9356744 −0.019 (−0.053, 0.015) −0.020 (−0.045, 0.005) 0.001 (−0.008, 0.011) −0.001 (−0.008, 0.006) GNPDA2 rs10938397 0.024 (−0.012, 0.060) 0.012 (−0.014, 0.038) 0.007 (−0.004, 0.017) −0.002 (−0.010, 0.005) SMC5-KLF9 rs11142387 −0.012 (−0.048, 0.025) −0.002 (−0.029, 0.025) 0.001 (−0.009, 0.012) 0.005 (−0.003, 0.012) GIPR-QPCTL rs11671664 0.001 (−0.032, 0.035) −0.004 (−0.028, 0.021) 0.002 (−0.007, 0.012) 0.003 (−0.004, 0.010) GP2 rs12597579 0.014 (−0.023, 0.052) 0.008 (−0.019, 0.035) −0.004 (−0.015, 0.007) 0.006 (−0.002, 0.014) FTO rs17817449 0.033 (−0.020, 0.086) 0.029 (−0.009, 0.068) −0.015 (−0.031, 0.000) 0.017 (0.005, 0.028) BDNF rs6265 −0.015 (−0.048, 0.019) 0.008 (−0.016, 0.033) −0.012 (−0.022, −0.002) −0.003 (−0.010, 0.004) PCSK1 rs261967 −0.020 (−0.054, 0.014) −0.003 (−0.028, 0.022) −0.009 (−0.019, 0.001) 0.005 (−0.003, 0.012) SEC16B rs574367 0.027 (−0.016, 0.07) 0.006 (−0.025, 0.037) −0.013 (−0.025, 0.000) 0.008 (−0.002, 0.017) PAX6 rs652722 0.002 (−0.033, 0.037) 0.002 (−0.023, 0.028) 0.000 (−0.010, 0.010) 0.000 (−0.008, 0.007) Gene SNP TC LDL-C HDL-C log10TG TFAP2B rs4715210 0.002 (−0.044, 0.049) 0.015 (−0.019, 0.049) −0.002 (−0.015, 0.012) 0.006 (−0.004, 0.016) MAP2K5 rs4776970 −0.004 (−0.044, 0.036) 0.001 (−0.028, 0.030) −0.015 (−0.027, −0.004) 0.005 (−0.003, 0.014) ADCY3-DNAJC27 rs6545814 −0.007 (−0.042, 0.027) 0.019 (−0.006, 0.045) −0.007 (−0.017, 0.003) −0.002 (−0.010, 0.005) MC4R rs6567160 0.019 (−0.023, 0.060) 0.025 (−0.005, 0.055) −0.014 (−0.026, –0.002) 0.008 (−0.001, 0.017) CDKAL1 rs9356744 −0.019 (−0.053, 0.015) −0.020 (−0.045, 0.005) 0.001 (−0.008, 0.011) −0.001 (−0.008, 0.006) GNPDA2 rs10938397 0.024 (−0.012, 0.060) 0.012 (−0.014, 0.038) 0.007 (−0.004, 0.017) −0.002 (−0.010, 0.005) SMC5-KLF9 rs11142387 −0.012 (−0.048, 0.025) −0.002 (−0.029, 0.025) 0.001 (−0.009, 0.012) 0.005 (−0.003, 0.012) GIPR-QPCTL rs11671664 0.001 (−0.032, 0.035) −0.004 (−0.028, 0.021) 0.002 (−0.007, 0.012) 0.003 (−0.004, 0.010) GP2 rs12597579 0.014 (−0.023, 0.052) 0.008 (−0.019, 0.035) −0.004 (−0.015, 0.007) 0.006 (−0.002, 0.014) FTO rs17817449 0.033 (−0.020, 0.086) 0.029 (−0.009, 0.068) −0.015 (−0.031, 0.000) 0.017 (0.005, 0.028) BDNF rs6265 −0.015 (−0.048, 0.019) 0.008 (−0.016, 0.033) −0.012 (−0.022, −0.002) −0.003 (−0.010, 0.004) PCSK1 rs261967 −0.020 (−0.054, 0.014) −0.003 (−0.028, 0.022) −0.009 (−0.019, 0.001) 0.005 (−0.003, 0.012) SEC16B rs574367 0.027 (−0.016, 0.07) 0.006 (−0.025, 0.037) −0.013 (−0.025, 0.000) 0.008 (−0.002, 0.017) PAX6 rs652722 0.002 (−0.033, 0.037) 0.002 (−0.023, 0.028) 0.000 (−0.010, 0.010) 0.000 (−0.008, 0.007) The data are presented as unstandardized coefficients and 95% CIs using linear regression analyses. The model was unadjusted. View Large Most of the diabetes-related SNPs did not exhibit significant associations with lipid profile; the exceptions were rs4430796, rs6467136, rs12779790, rs831571, rs864745, rs7041847, and rs2943641 (Table 1). Most of the BMI-related SNPs also did not exhibit significant relationships with lipid profile; the exceptions were rs17817449, rs4776970, rs6567160, rs6265, and rs574367 (Table 2). However, additional adjustments for HbA1c and BMI greatly attenuated the associations of most of these exceptive SNPs with lipid profile. Pleiotropic effects of SNPs We measured the potential associations of the SNPs with BMI, systolic and diastolic blood pressure, FPG, HbA1c, and age using an additive model (Supplemental Figs. 3 and 4). Three of the diabetes-related SNPs exhibited an association with at least one trait at the nominal P value ≤0.05. Two SNPs (rs4430796 and rs391300) were associated with systolic and diastolic blood pressure, and two SNPs (rs4430796 and rs896854) were associated with age. Seven of the BMI-related SNPs exhibited a significant association with at least one trait. Seven SNPs were associated with blood pressure (rs4776970, rs12597579, rs17817449, rs261967, rs4715210, rs6567160, and rs574367), two SNPs were associated with FPG (rs4715210 and rs6545814), and two SNPs were associated with age (rs4715210 and rs17817449). Study characteristics based on the DM_GRS and BMI_GRS As expected, with increases in the DM_GRS, the FPG, HbA1c, and prevalence of diabetes significantly increased. The BMI_GRS was also significantly associated with an increase in BMI (all P < 0.001; Table 3). Table 3. Characteristics of the Study Participants According to the DM_GRS and the BMI_GRS (n = 9798) Characteristic Q1 Q2 Q3 Q4 Q5 P for Trend DM_GRS ≤18.013 18.014–19.740 19.741–21.312 21.313–23.011 ≥23.012  n 1959 1961 1958 1961 1959  Age, y 54.3 (13.0) 55.0 (12.9) 55.4 (12.5) 54.3 (13.0) 54.4 (12.9) 0.669  Men, % 40.3 40.1 39.0 39.9 40.0 0.850  BMI, kg/m2 24.7 (3.5) 24.8 (3.7) 24.6 (3.6) 24.6 (3.7) 24.5 (3.5) 0.176  TC, mmol/L 5.19 (1.20) 5.15 (1.00) 5.25 (1.04) 5.23 (1.08) 5.24 (1.21) 0.010  TGs, mmol/L 1.65 (1.51) 1.63 (1.22) 1.67 (1.44) 1.67 (1.54) 1.68 (1.55) 0.220  HDL-C, mmol/L 1.40 (0.32) 1.39 (0.32) 1.40 (0.33) 1.41 (0.32) 1.40 (0.31) 0.726  LDL-C, mmol/L 3.17 (0.79) 3.18 (0.78) 3.25 (0.82) 3.23 (0.81) 3.22 (0.82) 0.010  FPG, mmol/L 5.5 (1.39) 5.6 (1.35) 5.7 (1.53) 5.7 (1.54) 5.8 (1.65) 3.9E-7  HbA1c, % 5.5 (0.9) 5.6 (0.9) 5.6 (1.0) 5.6 (1.0) 5.7 (1.1) 3.4E-8  Diabetes, % 11.7 13.7 14.2 15.5 18.3 2.9E-8  Hypertension, % 48.3 49.7 48.3 47.0 46.2 0.052  Economic status, low %/high % 32.7/67.3 34.6/65.4 33.7/66.3 35.1/64.9 35.6/64.4 0.062 BMI_GRS ≤8.186 8.187–9.404 9.405–10.621 10.622–12.007 ≥12.008  n 2358 1857 1855 1849 1879  Age, y 54.8 (12.9) 54.8 (13.0) 55.1 (12.6) 54.6 (12.9) 54.0 (13.0) 0.036  Men, % 40.5 39.1 39.3 41.3 38.9 0.702  BMI, kg/m2 24.4 (3.6) 24.5 (3.5) 24.7 (3.7) 24.8 (3.5) 25.0 (3.6) 3.5E-9  TC, mmol/L 5.19 (1.07) 5.21 (1.10) 5.23 (1.06) 5.25 (1.24) 5.18 (1.07) 0.906  TGs, mmol/L 1.64 (1.57) 1.65 (1.47) 1.63 (1.23) 1.71 (1.46) 1.68 (1.49) 0.019  HDL-C, mmol/L 1.41 (0.33) 1.41 (0.32) 1.41 (0.32) 1.39 (0.32) 1.38 (0.31) 0.005  LDL-C, mmol/L 3.19 (0.81) 3.20 (0.78) 3.23 (0.81) 3.23 (0.81) 3.21 (0.81) 0.197  FPG, mmol/L 5.7 (1.6) 5.6 (1.5) 5.6 (1.4) 5.6 (1.6) 5.7 (1.5) 0.953  HbA1c, % 5.6 (1.0) 5.6 (1.0) 5.6 (1.0) 5.6 (1.0) 5.7 (1.0) 0.367  Diabetes, % 14.6 14.3 15.6 13.8 15.1 0.864  Hypertension, % 46.9 46.6 47.3 48.9 49.9 0.022  Economic status, low %/high % 34.1/65.9 32.7/67.3 35.0/65.0 35.6/64.4 34.4/65.6 0.313 Characteristic Q1 Q2 Q3 Q4 Q5 P for Trend DM_GRS ≤18.013 18.014–19.740 19.741–21.312 21.313–23.011 ≥23.012  n 1959 1961 1958 1961 1959  Age, y 54.3 (13.0) 55.0 (12.9) 55.4 (12.5) 54.3 (13.0) 54.4 (12.9) 0.669  Men, % 40.3 40.1 39.0 39.9 40.0 0.850  BMI, kg/m2 24.7 (3.5) 24.8 (3.7) 24.6 (3.6) 24.6 (3.7) 24.5 (3.5) 0.176  TC, mmol/L 5.19 (1.20) 5.15 (1.00) 5.25 (1.04) 5.23 (1.08) 5.24 (1.21) 0.010  TGs, mmol/L 1.65 (1.51) 1.63 (1.22) 1.67 (1.44) 1.67 (1.54) 1.68 (1.55) 0.220  HDL-C, mmol/L 1.40 (0.32) 1.39 (0.32) 1.40 (0.33) 1.41 (0.32) 1.40 (0.31) 0.726  LDL-C, mmol/L 3.17 (0.79) 3.18 (0.78) 3.25 (0.82) 3.23 (0.81) 3.22 (0.82) 0.010  FPG, mmol/L 5.5 (1.39) 5.6 (1.35) 5.7 (1.53) 5.7 (1.54) 5.8 (1.65) 3.9E-7  HbA1c, % 5.5 (0.9) 5.6 (0.9) 5.6 (1.0) 5.6 (1.0) 5.7 (1.1) 3.4E-8  Diabetes, % 11.7 13.7 14.2 15.5 18.3 2.9E-8  Hypertension, % 48.3 49.7 48.3 47.0 46.2 0.052  Economic status, low %/high % 32.7/67.3 34.6/65.4 33.7/66.3 35.1/64.9 35.6/64.4 0.062 BMI_GRS ≤8.186 8.187–9.404 9.405–10.621 10.622–12.007 ≥12.008  n 2358 1857 1855 1849 1879  Age, y 54.8 (12.9) 54.8 (13.0) 55.1 (12.6) 54.6 (12.9) 54.0 (13.0) 0.036  Men, % 40.5 39.1 39.3 41.3 38.9 0.702  BMI, kg/m2 24.4 (3.6) 24.5 (3.5) 24.7 (3.7) 24.8 (3.5) 25.0 (3.6) 3.5E-9  TC, mmol/L 5.19 (1.07) 5.21 (1.10) 5.23 (1.06) 5.25 (1.24) 5.18 (1.07) 0.906  TGs, mmol/L 1.64 (1.57) 1.65 (1.47) 1.63 (1.23) 1.71 (1.46) 1.68 (1.49) 0.019  HDL-C, mmol/L 1.41 (0.33) 1.41 (0.32) 1.41 (0.32) 1.39 (0.32) 1.38 (0.31) 0.005  LDL-C, mmol/L 3.19 (0.81) 3.20 (0.78) 3.23 (0.81) 3.23 (0.81) 3.21 (0.81) 0.197  FPG, mmol/L 5.7 (1.6) 5.6 (1.5) 5.6 (1.4) 5.6 (1.6) 5.7 (1.5) 0.953  HbA1c, % 5.6 (1.0) 5.6 (1.0) 5.6 (1.0) 5.6 (1.0) 5.7 (1.0) 0.367  Diabetes, % 14.6 14.3 15.6 13.8 15.1 0.864  Hypertension, % 46.9 46.6 47.3 48.9 49.9 0.022  Economic status, low %/high % 34.1/65.9 32.7/67.3 35.0/65.0 35.6/64.4 34.4/65.6 0.313 The data are summarized as the means (SD) for continuous variables and as numerical proportions for categorical variables. The P values for the trends were as calculated with analysis of variance and χ2 tests. View Large Table 3. Characteristics of the Study Participants According to the DM_GRS and the BMI_GRS (n = 9798) Characteristic Q1 Q2 Q3 Q4 Q5 P for Trend DM_GRS ≤18.013 18.014–19.740 19.741–21.312 21.313–23.011 ≥23.012  n 1959 1961 1958 1961 1959  Age, y 54.3 (13.0) 55.0 (12.9) 55.4 (12.5) 54.3 (13.0) 54.4 (12.9) 0.669  Men, % 40.3 40.1 39.0 39.9 40.0 0.850  BMI, kg/m2 24.7 (3.5) 24.8 (3.7) 24.6 (3.6) 24.6 (3.7) 24.5 (3.5) 0.176  TC, mmol/L 5.19 (1.20) 5.15 (1.00) 5.25 (1.04) 5.23 (1.08) 5.24 (1.21) 0.010  TGs, mmol/L 1.65 (1.51) 1.63 (1.22) 1.67 (1.44) 1.67 (1.54) 1.68 (1.55) 0.220  HDL-C, mmol/L 1.40 (0.32) 1.39 (0.32) 1.40 (0.33) 1.41 (0.32) 1.40 (0.31) 0.726  LDL-C, mmol/L 3.17 (0.79) 3.18 (0.78) 3.25 (0.82) 3.23 (0.81) 3.22 (0.82) 0.010  FPG, mmol/L 5.5 (1.39) 5.6 (1.35) 5.7 (1.53) 5.7 (1.54) 5.8 (1.65) 3.9E-7  HbA1c, % 5.5 (0.9) 5.6 (0.9) 5.6 (1.0) 5.6 (1.0) 5.7 (1.1) 3.4E-8  Diabetes, % 11.7 13.7 14.2 15.5 18.3 2.9E-8  Hypertension, % 48.3 49.7 48.3 47.0 46.2 0.052  Economic status, low %/high % 32.7/67.3 34.6/65.4 33.7/66.3 35.1/64.9 35.6/64.4 0.062 BMI_GRS ≤8.186 8.187–9.404 9.405–10.621 10.622–12.007 ≥12.008  n 2358 1857 1855 1849 1879  Age, y 54.8 (12.9) 54.8 (13.0) 55.1 (12.6) 54.6 (12.9) 54.0 (13.0) 0.036  Men, % 40.5 39.1 39.3 41.3 38.9 0.702  BMI, kg/m2 24.4 (3.6) 24.5 (3.5) 24.7 (3.7) 24.8 (3.5) 25.0 (3.6) 3.5E-9  TC, mmol/L 5.19 (1.07) 5.21 (1.10) 5.23 (1.06) 5.25 (1.24) 5.18 (1.07) 0.906  TGs, mmol/L 1.64 (1.57) 1.65 (1.47) 1.63 (1.23) 1.71 (1.46) 1.68 (1.49) 0.019  HDL-C, mmol/L 1.41 (0.33) 1.41 (0.32) 1.41 (0.32) 1.39 (0.32) 1.38 (0.31) 0.005  LDL-C, mmol/L 3.19 (0.81) 3.20 (0.78) 3.23 (0.81) 3.23 (0.81) 3.21 (0.81) 0.197  FPG, mmol/L 5.7 (1.6) 5.6 (1.5) 5.6 (1.4) 5.6 (1.6) 5.7 (1.5) 0.953  HbA1c, % 5.6 (1.0) 5.6 (1.0) 5.6 (1.0) 5.6 (1.0) 5.7 (1.0) 0.367  Diabetes, % 14.6 14.3 15.6 13.8 15.1 0.864  Hypertension, % 46.9 46.6 47.3 48.9 49.9 0.022  Economic status, low %/high % 34.1/65.9 32.7/67.3 35.0/65.0 35.6/64.4 34.4/65.6 0.313 Characteristic Q1 Q2 Q3 Q4 Q5 P for Trend DM_GRS ≤18.013 18.014–19.740 19.741–21.312 21.313–23.011 ≥23.012  n 1959 1961 1958 1961 1959  Age, y 54.3 (13.0) 55.0 (12.9) 55.4 (12.5) 54.3 (13.0) 54.4 (12.9) 0.669  Men, % 40.3 40.1 39.0 39.9 40.0 0.850  BMI, kg/m2 24.7 (3.5) 24.8 (3.7) 24.6 (3.6) 24.6 (3.7) 24.5 (3.5) 0.176  TC, mmol/L 5.19 (1.20) 5.15 (1.00) 5.25 (1.04) 5.23 (1.08) 5.24 (1.21) 0.010  TGs, mmol/L 1.65 (1.51) 1.63 (1.22) 1.67 (1.44) 1.67 (1.54) 1.68 (1.55) 0.220  HDL-C, mmol/L 1.40 (0.32) 1.39 (0.32) 1.40 (0.33) 1.41 (0.32) 1.40 (0.31) 0.726  LDL-C, mmol/L 3.17 (0.79) 3.18 (0.78) 3.25 (0.82) 3.23 (0.81) 3.22 (0.82) 0.010  FPG, mmol/L 5.5 (1.39) 5.6 (1.35) 5.7 (1.53) 5.7 (1.54) 5.8 (1.65) 3.9E-7  HbA1c, % 5.5 (0.9) 5.6 (0.9) 5.6 (1.0) 5.6 (1.0) 5.7 (1.1) 3.4E-8  Diabetes, % 11.7 13.7 14.2 15.5 18.3 2.9E-8  Hypertension, % 48.3 49.7 48.3 47.0 46.2 0.052  Economic status, low %/high % 32.7/67.3 34.6/65.4 33.7/66.3 35.1/64.9 35.6/64.4 0.062 BMI_GRS ≤8.186 8.187–9.404 9.405–10.621 10.622–12.007 ≥12.008  n 2358 1857 1855 1849 1879  Age, y 54.8 (12.9) 54.8 (13.0) 55.1 (12.6) 54.6 (12.9) 54.0 (13.0) 0.036  Men, % 40.5 39.1 39.3 41.3 38.9 0.702  BMI, kg/m2 24.4 (3.6) 24.5 (3.5) 24.7 (3.7) 24.8 (3.5) 25.0 (3.6) 3.5E-9  TC, mmol/L 5.19 (1.07) 5.21 (1.10) 5.23 (1.06) 5.25 (1.24) 5.18 (1.07) 0.906  TGs, mmol/L 1.64 (1.57) 1.65 (1.47) 1.63 (1.23) 1.71 (1.46) 1.68 (1.49) 0.019  HDL-C, mmol/L 1.41 (0.33) 1.41 (0.32) 1.41 (0.32) 1.39 (0.32) 1.38 (0.31) 0.005  LDL-C, mmol/L 3.19 (0.81) 3.20 (0.78) 3.23 (0.81) 3.23 (0.81) 3.21 (0.81) 0.197  FPG, mmol/L 5.7 (1.6) 5.6 (1.5) 5.6 (1.4) 5.6 (1.6) 5.7 (1.5) 0.953  HbA1c, % 5.6 (1.0) 5.6 (1.0) 5.6 (1.0) 5.6 (1.0) 5.7 (1.0) 0.367  Diabetes, % 14.6 14.3 15.6 13.8 15.1 0.864  Hypertension, % 46.9 46.6 47.3 48.9 49.9 0.022  Economic status, low %/high % 34.1/65.9 32.7/67.3 35.0/65.0 35.6/64.4 34.4/65.6 0.313 The data are summarized as the means (SD) for continuous variables and as numerical proportions for categorical variables. The P values for the trends were as calculated with analysis of variance and χ2 tests. View Large The ORZX of the observed association between DM_GRS (per 1 SD increase) and diabetes was 1.231 (95% CI: 1.159, 1.307) with an F statistic of 69. The βZX regression coefficient between BMI_GRS (per 1 SD increase) and BMI was 0.215 (95% CI: 0.147, 0.284) with an F statistic of 39. The F statistics greater than 10 commonly suggest these two GRSs are sufficiently strong as IVs. Regarding the pleiotropic effects, the DM_GRS quintiles were not significantly associated with BMI or hypertension, and each 1 SD increment of the DM_GRS also did not exhibit a significant association with BMI or hypertension in the adjusted model. The BMI_GRSs quintiles were associated with age and hypertension (Table 3). Therefore, in the sensitivity analyses, we constructed a BMI_GRS12SNP that excluded FTO rs17817449 and MC4R rs6567160, which were the two most obvious loci that had pleiotropic effects. Association of diabetes and DM_GRS with lipid profile Next, we analyzed the associations of present diabetes and DM_GRS with lipid profile including TC, HDL-C, LDL-C, and TG (Table 4). In this cross-sectional study, present diabetes (βXY) was significantly associated with log10TG (β = 0.078; 95% CI: 0.066, 0.091) and HDL-C (β = –0.067; 95% CI: –0.085, –0.049) but not with TC or LDL-C. However, the DM_GRS (βZY) was significantly related to TC (β = 0.027; 95% CI: 0.005, 0.048) and LDL-C (β = 0.026; 95% CI: 0.011, 0.042) and marginally related to log10TG (β = 0.004; 95% CI: 0.000, 0.009) but not to HDL-C. Table 4. Causal Coefficients From the MR Analysis for the Associations of Type 2 Diabetes and BMI With Lipid Profile TC LDL-C HDL-C log10TG DM  βXY (present diabetes) 0.019 (–0.043, 0.082) 0.029 (–0.016, 0.073) −0.067 (–0.085, –0.049) 0.078 (0.066, 0.091)  βZY (per SD increase in DM_GRS) 0.027 (0.005, 0.048) 0.026 (0.011, 0.042) −0.001 (–0.007, 0.005) 0.004 (0.000, 0.009)  βIV (present diabetes) 0.130 (0.020, 0.240) 0.125 (0.041, 0.209) −0.005 (–0.033, 0.023) 0.019 (–0.001, 0.039) BMI  βXY (per 10 kg/m2 increase in BMI) 0.281 (0.219, 0.343) 0.328 (0.284, 0.373) −0.212 (–0.230, –0.195) 0.169 (0.157, 0.182)  βZY (per SD increase in BMI_GRS) −0.002 (–0.024, 0.019) 0.008 (–0.008, 0.023) −0.009 (–0.015, –0.003) 0.005 (0.001, 0.010)  βIV (per 10 kg/m2 increase in BMI) −0.091 (–1.071, 0.889) 0.364 (–0.365, 1.083) −0.409 (–0.698, –0.120) 0.227 (0.039, 0.415) TC LDL-C HDL-C log10TG DM  βXY (present diabetes) 0.019 (–0.043, 0.082) 0.029 (–0.016, 0.073) −0.067 (–0.085, –0.049) 0.078 (0.066, 0.091)  βZY (per SD increase in DM_GRS) 0.027 (0.005, 0.048) 0.026 (0.011, 0.042) −0.001 (–0.007, 0.005) 0.004 (0.000, 0.009)  βIV (present diabetes) 0.130 (0.020, 0.240) 0.125 (0.041, 0.209) −0.005 (–0.033, 0.023) 0.019 (–0.001, 0.039) BMI  βXY (per 10 kg/m2 increase in BMI) 0.281 (0.219, 0.343) 0.328 (0.284, 0.373) −0.212 (–0.230, –0.195) 0.169 (0.157, 0.182)  βZY (per SD increase in BMI_GRS) −0.002 (–0.024, 0.019) 0.008 (–0.008, 0.023) −0.009 (–0.015, –0.003) 0.005 (0.001, 0.010)  βIV (per 10 kg/m2 increase in BMI) −0.091 (–1.071, 0.889) 0.364 (–0.365, 1.083) −0.409 (–0.698, –0.120) 0.227 (0.039, 0.415) The data are presented as regression coefficients (β) and 95% CIs. In this MR framework, the IV estimator is βIV = βZY/βZX, which equals the causal coefficients of diabetes and BMI (per 10 kg/m2 increase) for changes in lipid parameters. The data were adjusted for age, sex, BMI (only in the the BMI_GRS model), hypertension, and economic status. Abbreviation: DM, diabetes mellitus. View Large Table 4. Causal Coefficients From the MR Analysis for the Associations of Type 2 Diabetes and BMI With Lipid Profile TC LDL-C HDL-C log10TG DM  βXY (present diabetes) 0.019 (–0.043, 0.082) 0.029 (–0.016, 0.073) −0.067 (–0.085, –0.049) 0.078 (0.066, 0.091)  βZY (per SD increase in DM_GRS) 0.027 (0.005, 0.048) 0.026 (0.011, 0.042) −0.001 (–0.007, 0.005) 0.004 (0.000, 0.009)  βIV (present diabetes) 0.130 (0.020, 0.240) 0.125 (0.041, 0.209) −0.005 (–0.033, 0.023) 0.019 (–0.001, 0.039) BMI  βXY (per 10 kg/m2 increase in BMI) 0.281 (0.219, 0.343) 0.328 (0.284, 0.373) −0.212 (–0.230, –0.195) 0.169 (0.157, 0.182)  βZY (per SD increase in BMI_GRS) −0.002 (–0.024, 0.019) 0.008 (–0.008, 0.023) −0.009 (–0.015, –0.003) 0.005 (0.001, 0.010)  βIV (per 10 kg/m2 increase in BMI) −0.091 (–1.071, 0.889) 0.364 (–0.365, 1.083) −0.409 (–0.698, –0.120) 0.227 (0.039, 0.415) TC LDL-C HDL-C log10TG DM  βXY (present diabetes) 0.019 (–0.043, 0.082) 0.029 (–0.016, 0.073) −0.067 (–0.085, –0.049) 0.078 (0.066, 0.091)  βZY (per SD increase in DM_GRS) 0.027 (0.005, 0.048) 0.026 (0.011, 0.042) −0.001 (–0.007, 0.005) 0.004 (0.000, 0.009)  βIV (present diabetes) 0.130 (0.020, 0.240) 0.125 (0.041, 0.209) −0.005 (–0.033, 0.023) 0.019 (–0.001, 0.039) BMI  βXY (per 10 kg/m2 increase in BMI) 0.281 (0.219, 0.343) 0.328 (0.284, 0.373) −0.212 (–0.230, –0.195) 0.169 (0.157, 0.182)  βZY (per SD increase in BMI_GRS) −0.002 (–0.024, 0.019) 0.008 (–0.008, 0.023) −0.009 (–0.015, –0.003) 0.005 (0.001, 0.010)  βIV (per 10 kg/m2 increase in BMI) −0.091 (–1.071, 0.889) 0.364 (–0.365, 1.083) −0.409 (–0.698, –0.120) 0.227 (0.039, 0.415) The data are presented as regression coefficients (β) and 95% CIs. In this MR framework, the IV estimator is βIV = βZY/βZX, which equals the causal coefficients of diabetes and BMI (per 10 kg/m2 increase) for changes in lipid parameters. The data were adjusted for age, sex, BMI (only in the the BMI_GRS model), hypertension, and economic status. Abbreviation: DM, diabetes mellitus. View Large In the IV analysis, the causal regression coefficients (βIV) of genetically determined diabetes for TC, LDL-C, and log10TG were 0.130 (95% CI: 0.020, 0.240; P = 0.014), 0.125 (96% CI: 0.041, 0.209; P = 0.001), and 0.019 (95% CI: –0.001, 0.039; P = 0.055), respectively. The βIV for HDL-C was –0.005 (95% CI: –0.033, 0.023), which was not significant (P = 0.699). Association of BMI and the BMI_GRS with lipid profile The associations of measured BMI and the BMI_GRS with lipid profile are presented in Table 4. A 10 kg/m2 increase in measured BMI (βXY) was significantly associated with TC (β = 0.281; 95% CI: 0.219, 0.343), LDL-C (β = 0.328: 95% CI: 0.284, 0.373), HDL-C (β = –0.212; 95% CI: –0.230, –0.195), and log10TG (β = 0.169; 95% CI: 0.157, 0.182). However, the BMI_GRS (βZY) was only significantly associated with HDL-C (β = –0.009; 95% CI: –0.015, –0.003) and log10TG (β = 0.005; 95% CI: 0.001, 0.010). In the IV analysis, the causal regression coefficients (βIV) of a genetically determined 10 kg/m2 higher BMI for HDL-C and log10TG were –0.409 (96% CI: –0.698, –0.120; P = 0.004) and 0.227 (95% CI: 0.039, 0.415; P = 0.026), respectively. The βIV for TC and LDL-C was not significant. Sensitivity analysis Considering pleiotropy, we constructed a GRS that excluded FTO rs17817449 and MC4R rs6567160, which were no longer significantly associated with age (P = 0.155) or hypertension (P = 0.091). A genetically determined 10 kg/m2 greater BMI was still causally related to HDL (β = –0.361; 95% CI: –0.675, –0.047; P = 0.024) and marginally related to log10TG (β = 0.227; 95% CI: –0.001, 0.455; P = 0.051). Three assumptions for IV in MR First, the GRSs were strongly associated with exposure with F statistics that were considerably greater than 10, which indicated that these GRSs were adequate IVs. Second, a GRS should not be associated with any confounder of the exposure-outcome association. In our study, the BMI_GRSs quintiles were associated with age and hypertension; therefore, the BMI_GRS12SNP was constructed, and the BMI determined by the BMI_GRS12SNP was still causally associated with HDL. Third, a GRS should be independent of the outcome, except possibly through its association with the exposure. This assumption means that the only causal route from the genetic variants to the outcome should be through the exposure (i.e., there should be no other routes between the DM_GRS and BMI_GRS and hyperlipidemia) (23). However, this assumption is only an idealized view and quite untestable because it is possible that a causal association is due to other diabetes characteristics as a direct consequence of being diabetic (15), and biologic pathways are complex and interconnected. Additional adjustments for HbA1c or BMI could greatly attenuate most of the associations of DM_GRS and BMI_GRS with lipid profile, which verified the third assumption to a limited extent. Discussion In this cross-sectional survey that included almost 10,000 community-dwelling Chinese adults, we used an MR design to examine the causal associations between adiposity, as assessed by an elevated BMI, and lipid profile and between diabetes and lipid profile. The current study is, to our knowledge, the first MR study to measure the effect of diabetes on elements of the lipid profile. This is an old topic, but the MR design seems to reveal results that differ from conventional opinions. We demonstrated evidence for causal relationships (1) between diabetes and higher TC and LDL-C but not lower HDL-C and (2) between obesity and lower HDL-C and higher TG but not higher TC and LDL-C. Our findings do not fully accord with those of conventional observational studies. However, the setting of our study was different: We assessed the effects of lifelong exposure to diabetes risks and high BMI on hyperlipidemia. Many previous studies have demonstrated that hyperlipidemia is more prevalent in individuals with diabetes and obesity than it is in their nondiabetic and nonobese counterparts (25). The features of dyslipidemia in patients with diabetes have been reported to be high TG levels and decreased HDL-C levels, but the concentration of LDL-C can be in the normal range, and small dense LDL-C can be increased (6). However, our study found that type 2 diabetes may induce a significant elevation of LDL-C but not a decrease of HDL-C. This finding further indicates the important role of diabetes in the formation of atherosclerosis because LDL-C level is recommended for use as the primary lipid analysis for screening, risk estimation, diagnosis, and management (25). Interestingly, in our study, the observed associations of diabetes with lipid profile were significant for HDL-C and TG but not for TC or LDL-C, which might reflect the flaws of conventional observational studies. Additional independent MR studies of this association should be performed. In obesity, elevations of TC, LDL-C, and TG levels and decreased HDL-C levels are very common (5). There are two additional MR studies that have measured the associations of BMI with lipid profile (7, 26). We replicated the results of Fall et al. (7) from individuals of European descent. Using the FTO locus as the IV, these authors also observed an unfavorable effect of BMI on decreased HDL-C levels and increased TG levels and not on LDL-C or TC (7). However, another study that used 14 BMI-associated SNPs reported that a 1 kg/m2 increase in BMI reduced HDL-C by –0.02 mmol/L (95% CI: –0.03, –0.01) and, interestingly, LDL-C by –0.04 mmol/L (95% CI: –0.07, –0.01) in individuals of European descent (26). We found that the IV estimates for the effects of BMI on HDL-C and TG were greater than those calculated from standard linear regressions. As suggested in previous studies (7), possible explanations include the following: the cross-sectional nature of a study that could result in reverse causation and confounds, and the notion that the lifelong effects of SNPs on adiposity are not entirely captured by a single BMI measurement (27). CVD is the main cause of death in patients with type 2 diabetes and obesity (28, 29). Hyperlipidemia may be one important pathway linking diabetes and obesity to cardiovascular outcomes. Obesity could lead to dyslipidemia, which further induces and aggravates CVD (25, 30). Furthermore, a sustainable weight loss of 5 to 8 kg results in a mean LDL-C reduction of 5 mg/dL and a mean HDL-C increase of 2 to 3 mg/dL (30). Additionally, weight loss is also expected to decrease TG (30). However, an MR study did not identify evidence of a causal relationship between BMI and coronary heart disease (OR: 1.01; 95% CI: 0.94, 1.08) (26), and this issue requires further evidence. Although type 2 diabetes is regarded as a risk factor for CVD, the effect of intensive glucose-lowering therapies on the prevention of CVD is controversial. Some studies have found no benefit of intensive glucose-lowering therapy (31, 32), whereas one meta-analysis of randomized controlled trials indicated that intensive glycemic control resulted in a 15% reduction in coronary heart disease events (33). Moreover, a very recent MR study revealed a causal association between diabetes and increased arterial stiffness (OR: 1.24; 95% CI: 1.06, 1.47) in a Chinese population, which indicates that long-term treatment of diabetes may be beneficial to the prevention of the development of arterial stiffening (15). Our study suggests that diabetes could induce an elevation of LDL-C, which is the core element of the lipid profile in the development of atherosclerosis. Thus, we suspect that long-term treatment of diabetes may have benefits for atherosclerosis prevention that are partly mediated through LDL-C reduction. The strengths of our study include the well-defined community setting and the use of MR to delineate the genetically determined relations between diabetes, BMI, and lipid profile. Second, the IVs we used for the MR were strong (F > 10). Last, our participants were all of Asian descent, which eliminated population admixture as a potential drawback. This study also has some limitations. All of the participants were of Asian origin. Therefore, our findings may not be applicable to other ethnicities. Second, the lipid profiles were measured only once at baseline. Hence, we were not able to control for intraindividual variability. Third, the self-reported diagnoses of hypertension and diabetes may not have been completely accurate; however, in an Asian population study, the self-reported medical histories were generally accurate, including self-reported hypertension and diabetes (34). Finally, we built our GRSs based only on common variants that are considered to represent limited diabetes and BMI heritability. We were unable to assess the potential contribution of rare variants. It was also difficult to use the GRSs to completely discriminate the effects of diabetes and obesity from the effects of other risk factors on changes in lipid profile because of the possible pleiotropic effects of SNPs, especially regarding the BMI_GRS. In conclusion, this analysis has provided evidence for the biologically plausible causal relationships of genetically determined type 2 diabetes with higher TC and LDL-C and genetically determined obesity with lower HDL-C and higher TG in a Chinese population. Diabetes and adiposity seem to cause different types of hyperlipidemia. Our study suggests that early diagnosis and long-term treatment of type 2 diabetes may be beneficial for LDL-C, which is the lipid element with the most strongly confirmed involvement in atherosclerosis development, and thus lower LDL-C levels could aid the prevention of CVD. Independent MR studies are needed to validate our findings in other ethnicities. Abbreviations: Abbreviations: BMI body mass index BMI_GRS body mass index genetic risk score CI confidence interval CVD cardiovascular disease DM_GRS diabetes mellitus genetic risk score FPG fasting plasma glucose GRS genetic risk score HbA1c glycated hemoglobin HDL-C high-density lipoprotein cholesterol IV instrumental variable LDL-C low-density lipoprotein cholesterol MR Mendelian randomization OR odds ratio SD standard deviation SNP single-nucleotide polymorphism SPECT Survey on Prevalence in East China for metabolic diseases and risk factors TC total cholesterol TG triglyceride Acknowledgments We thank thank Xiaojin Wang and Bingshun Wang from the Department of Biostatistics and Shanghai Jiaotong University School of Medicine for data processing, Weiping Tu, Bin Li, and Ling Hu for helping to organize this investigation, and all team members and participants in the SPECT-China study. Financial Support: This work was supported by the National Natural Science Foundation of China (81570726 to Y.L. and 81600609 to N.W.), the Shanghai JiaoTong University School of Medicine (2014 to Y.L.), the Science and Technology Commission of Shanghai Municipality (16411971200 to Y.L. and 16410723200 to N.W.), the Fourth Round of the Three-Year Public Health Action Plan of Shanghai by the Shanghai Municipal Commission of Health and Family Planning (15GWZK0202 to Y.L. and 20164Y0079 to N.W.), the Municipal Human Resources Development Program for Outstanding Young Talents in Medical and Health Sciences in Shanghai (2017YQ053 to N.W.), and the Clinical Research Plan of Shanghai Hospital Development Center (16CR3076B to Y.L.). Author Contributions: Y.L., L.G., and D.L. designed the study; N.W., J.C., Z.N., Y.C., B.H., Q.L., C.C., L.Z., and F.X conducted the research; and N.W., J.C., and Z.N. analyzed the data and wrote the manuscript. N.W., J.C., and Z.N. contributed equally. The final manuscript was read and approved by all authors. Disclosure Summary: The authors have nothing to disclose. References 1. Gersh BJ , Sliwa K , Mayosi BM , Yusuf S . Novel therapeutic concepts: the epidemic of cardiovascular disease in the developing world: global implications . Eur Heart J . 2010 ; 31 ( 6 ): 642 – 648 . Google Scholar CrossRef Search ADS PubMed 2. Moran A , Gu D , Zhao D , Coxson P , Wang YC , Chen CS , Liu J , Cheng J , Bibbins-Domingo K , Shen YM , He J , Goldman L . Future cardiovascular disease in china: markov model and risk factor scenario projections from the coronary heart disease policy model-china . Circ Cardiovasc Qual Outcomes . 2010 ; 3 ( 3 ): 243 – 252 . Google Scholar CrossRef Search ADS PubMed 3. Mooradian AD . Dyslipidemia in type 2 diabetes mellitus . Nat Clin Pract Endocrinol Metab . 2009 ; 5 ( 3 ): 150 – 159 . Google Scholar CrossRef Search ADS PubMed 4. Kannel WB . Lipids, diabetes, and coronary heart disease: insights from the Framingham Study . Am Heart J . 1985 ; 110 ( 5 ): 1100 – 1107 . Google Scholar CrossRef Search ADS PubMed 5. Yang W , Xiao J , Yang Z , Ji L , Jia W , Weng J , Lu J , Shan Z , Liu J , Tian H , Ji Q , Zhu D , Ge J , Lin L , Chen L , Guo X , Zhao Z , Li Q , Zhou Z , Shan G , He J ; China National Diabetes and Metabolic Disorders Study Investigators . Serum lipids and lipoproteins in Chinese men and women . Circulation . 2012 ; 125 ( 18 ): 2212 – 2221 . Google Scholar CrossRef Search ADS PubMed 6. Haffner SM ; American Diabetes Association . Dyslipidemia management in adults with diabetes . Diabetes Care . 2004 ; 27 ( Suppl 1 ): S68 – S71 . Google Scholar CrossRef Search ADS PubMed 7. Fall T , Hägg S , Mägi R , Ploner A , Fischer K , Horikoshi M , Sarin AP , Thorleifsson G , Ladenvall C , Kals M , Kuningas M , Draisma HH , Ried JS , van Zuydam NR , Huikari V , Mangino M , Sonestedt E , Benyamin B , Nelson CP , Rivera NV , Kristiansson K , Shen HY , Havulinna AS , Dehghan A , Donnelly LA , Kaakinen M , Nuotio ML , Robertson N , de Bruijn RF , Ikram MA , Amin N , Balmforth AJ , Braund PS , Doney AS , Döring A , Elliott P , Esko T , Franco OH , Gretarsdottir S , Hartikainen AL , Heikkilä K , Herzig KH , Holm H , Hottenga JJ , Hyppönen E , Illig T , Isaacs A , Isomaa B , Karssen LC , Kettunen J , Koenig W , Kuulasmaa K , Laatikainen T , Laitinen J , Lindgren C , Lyssenko V , Läärä E , Rayner NW , Männistö S , Pouta A , Rathmann W , Rivadeneira F , Ruokonen A , Savolainen MJ , Sijbrands EJ , Small KS , Smit JH , Steinthorsdottir V , Syvänen AC , Taanila A , Tobin MD , Uitterlinden AG , Willems SM , Willemsen G , Witteman J , Perola M , Evans A , Ferrières J , Virtamo J , Kee F , Tregouet DA , Arveiler D , Amouyel P , Ferrario MM , Brambilla P , Hall AS , Heath AC , Madden PA , Martin NG , Montgomery GW , Whitfield JB , Jula A , Knekt P , Oostra B , van Duijn CM , Penninx BW , Smith GD , Kaprio J , Samani NJ , Gieger C , Peters A , Wichmann HE , Boomsma DI , de Geus EJ , Tuomi T , Power C , Hammond CJ , Spector TD , Lind L , Orho-Melander M , Palmer CN , Morris AD , Groop L , Järvelin MR , Salomaa V , Vartiainen E , Hofman A , Ripatti S , Metspalu A , Thorsteinsdottir U , Stefansson K , Pedersen NL , McCarthy MI , Ingelsson E , Prokopenko I ; European Network for Genetic and Genomic Epidemiology (ENGAGE) consortium . The role of adiposity in cardiometabolic traits: a Mendelian randomization analysis . PLoS Med . 2013 ; 10 ( 6 ): e1001474 . Google Scholar CrossRef Search ADS PubMed 8. Smith GD , Ebrahim S . ‘Mendelian randomization’: can genetic epidemiology contribute to understanding environmental determinants of disease ? Int J Epidemiol . 2003 ; 32 ( 1 ): 1 – 22 . Google Scholar CrossRef Search ADS PubMed 9. Didelez V , Sheehan N . Mendelian randomization as an instrumental variable approach to causal inference . Stat Methods Med Res . 2007 ; 16 ( 4 ): 309 – 330 . Google Scholar CrossRef Search ADS PubMed 10. Wang N , Wang X , Han B , Li Q , Chen Y , Zhu C , Chen Y , Xia F , Cang Z , Zhu C , Lu M , Meng Y , Chen C , Lin D , Wang B , Jensen MD , Lu Y . Is exposure to famine in childhood and economic development in adulthood associated with diabetes ? J Clin Endocrinol Metab . 2015 ; 100 ( 12 ): 4514 – 4523 . Google Scholar CrossRef Search ADS PubMed 11. Wang N , Wang X , Li Q , Han B , Chen Y , Zhu C , Chen Y , Lin D , Wang B , Jensen MD , Lu Y . The famine exposure in early life and metabolic syndrome in adulthood . Clin Nutr . 2017 ; 36 ( 1 ): 253 – 259 . Google Scholar CrossRef Search ADS PubMed 12. Wang N , Zhang K , Han B , Li Q , Chen Y , Zhu C , Chen Y , Xia F , Zhai H , Jiang B , Shen Z , Lu Y . Follicle stimulating hormone, its novel association with sex hormone binding globulin in men and postmenopausal women . Endocrine . 2017 ; 56 ( 3 ): 649 – 657 . Google Scholar CrossRef Search ADS PubMed 13. Xu Y , Wang L , He J , Bi Y , Li M , Wang T , Wang L , Jiang Y , Dai M , Lu J , Xu M , Li Y , Hu N , Li J , Mi S , Chen CS , Li G , Mu Y , Zhao J , Kong L , Chen J , Lai S , Wang W , Zhao W , Ning G ; 2010 China Noncommunicable Disease Surveillance Group . Prevalence and control of diabetes in Chinese adults . JAMA . 2013 ; 310 ( 9 ): 948 – 959 . Google Scholar CrossRef Search ADS PubMed 14. Wang N , Chen C , Zhao L , Chen Y , Han B , Xia F , Cheng J , Li Q , Lu Y . Vitamin D and nonalcoholic fatty liver disease: bi-directional Mendelian randomization analysis . EBioMedicine . 2018 ; 28 : 187 – 193 . Google Scholar CrossRef Search ADS PubMed 15. Xu M , Huang Y , Xie L , Peng K , Ding L , Lin L , Wang P , Hao M , Chen Y , Sun Y , Qi L , Wang W , Ning G , Bi Y . Diabetes and risk of arterial stiffness: a Mendelian randomization analysis . Diabetes . 2016 ; 65 ( 6 ): 1731 – 1740 . Google Scholar CrossRef Search ADS PubMed 16. Huang Y , Xu M , Xie L , Wang T , Huang X , Lv X , Chen Y , Ding L , Lin L , Wang W , Bi Y , Sun Y , Zhang Y , Ning G . Obesity and peripheral arterial disease: a Mendelian randomization analysis . Atherosclerosis . 2016 ; 247 : 218 – 224 . Google Scholar CrossRef Search ADS PubMed 17. Cho YS , Lee JY , Park KS , Nho CW . Genetics of type 2 diabetes in East Asian populations . Curr Diab Rep . 2012 ; 12 ( 6 ): 686 – 696 . Google Scholar CrossRef Search ADS PubMed 18. Kato N . Insights into the genetic basis of type 2 diabetes . J Diabetes Investig . 2013 ; 4 ( 3 ): 233 – 244 . Google Scholar CrossRef Search ADS PubMed 19. Okada Y , Kubo M , Ohmiya H , Takahashi A , Kumasaka N , Hosono N , Maeda S , Wen W , Dorajoo R , Go MJ , Zheng W , Kato N , Wu JY , Lu Q , Tsunoda T , Yamamoto K , Nakamura Y , Kamatani N , Tanaka T ; GIANT consortium . Common variants at CDKAL1 and KLF9 are associated with body mass index in east Asian populations . Nat Genet . 2012 ; 44 ( 3 ): 302 – 306 . Google Scholar CrossRef Search ADS PubMed 20. Wen W , Cho YS , Zheng W , Dorajoo R , Kato N , Qi L , Chen CH , Delahanty RJ , Okada Y , Tabara Y , Gu D , Zhu D , Haiman CA , Mo Z , Gao YT , Saw SM , Go MJ , Takeuchi F , Chang LC , Kokubo Y , Liang J , Hao M , Le Marchand L , Zhang Y , Hu Y , Wong TY , Long J , Han BG , Kubo M , Yamamoto K , Su MH , Miki T , Henderson BE , Song H , Tan A , He J , Ng DP , Cai Q , Tsunoda T , Tsai FJ , Iwai N , Chen GK , Shi J , Xu J , Sim X , Xiang YB , Maeda S , Ong RT , Li C , Nakamura Y , Aung T , Kamatani N , Liu JJ , Lu W , Yokota M , Seielstad M , Fann CS , Wu JY , Lee JY , Hu FB , Tanaka T , Tai ES , Shu XO ; Genetic Investigation of ANthropometric Traits (GIANT) Consortium . Meta-analysis identifies common variants associated with body mass index in east Asians . Nat Genet . 2012 ; 44 ( 3 ): 307 – 311 . Google Scholar CrossRef Search ADS PubMed 21. Vimaleswaran KS , Berry DJ , Lu C , Tikkanen E , Pilz S , Hiraki LT , Cooper JD , Dastani Z , Li R , Houston DK , Wood AR , Michaëlsson K , Vandenput L , Zgaga L , Yerges-Armstrong LM , McCarthy MI , Dupuis J , Kaakinen M , Kleber ME , Jameson K , Arden N , Raitakari O , Viikari J , Lohman KK , Ferrucci L , Melhus H , Ingelsson E , Byberg L , Lind L , Lorentzon M , Salomaa V , Campbell H , Dunlop M , Mitchell BD , Herzig KH , Pouta A , Hartikainen AL , Streeten EA , Theodoratou E , Jula A , Wareham NJ , Ohlsson C , Frayling TM , Kritchevsky SB , Spector TD , Richards JB , Lehtimäki T , Ouwehand WH , Kraft P , Cooper C , März W , Power C , Loos RJ , Wang TJ , Järvelin MR , Whittaker JC , Hingorani AD , Hyppönen E ; Genetic Investigation of Anthropometric Traits-GIANT Consortium . Causal relationship between obesity and vitamin D status: bi-directional Mendelian randomization analysis of multiple cohorts . PLoS Med . 2013 ; 10 ( 2 ): e1001383 . Google Scholar CrossRef Search ADS PubMed 22. Lawlor DA , Harbord RM , Sterne JA , Timpson N , Davey Smith G . Mendelian randomization: using genes as instruments for making causal inferences in epidemiology . Stat Med . 2008 ; 27 ( 8 ): 1133 – 1163 . Google Scholar CrossRef Search ADS PubMed 23. Sheehan NA , Didelez V , Burton PR , Tobin MD . Mendelian randomisation and causal inference in observational epidemiology . PLoS Med . 2008 ; 5 ( 8 ): e177 . Google Scholar CrossRef Search ADS PubMed 24. Kraja AT , Chasman DI , North KE , Reiner AP , Yanek LR , Kilpeläinen TO , Smith JA , Dehghan A , Dupuis J , Johnson AD , Feitosa MF , Tekola-Ayele F , Chu AY , Nolte IM , Dastani Z , Morris A , Pendergrass SA , Sun YV , Ritchie MD , Vaez A , Lin H , Ligthart S , Marullo L , Rohde R , Shao Y , Ziegler MA , Im HK , Schnabel RB , Jørgensen T , Jørgensen ME , Hansen T , Pedersen O , Stolk RP , Snieder H , Hofman A , Uitterlinden AG , Franco OH , Ikram MA , Richards JB , Rotimi C , Wilson JG , Lange L , Ganesh SK , Nalls M , Rasmussen-Torvik LJ , Pankow JS , Coresh J , Tang W , Linda Kao WH , Boerwinkle E , Morrison AC , Ridker PM , Becker DM , Rotter JI , Kardia SL , Loos RJ , Larson MG , Hsu YH , Province MA , Tracy R , Voight BF , Vaidya D , O’Donnell CJ , Benjamin EJ , Alizadeh BZ , Prokopenko I , Meigs JB , Borecki IB ; Cross Consortia Pleiotropy Group ; Cohorts for Heart and ; Aging Research in Genetic Epidemiology ; Genetic Investigation of Anthropometric Traits Consortium ; Global Lipids Genetics Consortium ; Meta-Analyses of Glucose ; Insulin-related traits Consortium ; Global BPgen Consortium ; ADIPOGen Consortium ; Women’s Genome Health Study ; Howard University Family Study . Pleiotropic genes for metabolic syndrome and inflammation . Mol Genet Metab . 2014 ; 112 ( 4 ): 317 – 338 . Google Scholar CrossRef Search ADS PubMed 25. Catapano AL , Graham I , De Backer G , Wiklund O , Chapman MJ , Drexel H , Hoes AW , Jennings CS , Landmesser U , Pedersen TR , Reiner Ž , Riccardi G , Taskinen MR , Tokgozoglu L , Verschuren WMM , Vlachopoulos C , Wood DA , Zamorano JL , Cooney MT ; ESC Scientific Document Group . 2016 ESC/EAS guidelines for the management of dyslipidaemias . Eur Heart J . 2016 ; 37 ( 39 ): 2999 – 3058 . Google Scholar CrossRef Search ADS PubMed 26. Holmes MV , Lange LA , Palmer T , Lanktree MB , North KE , Almoguera B , Buxbaum S , Chandrupatla HR , Elbers CC , Guo Y , Hoogeveen RC , Li J , Li YR , Swerdlow DI , Cushman M , Price TS , Curtis SP , Fornage M , Hakonarson H , Patel SR , Redline S , Siscovick DS , Tsai MY , Wilson JG , van der Schouw YT , FitzGerald GA , Hingorani AD , Casas JP , de Bakker PI , Rich SS , Schadt EE , Asselbergs FW , Reiner AP , Keating BJ . Causal effects of body mass index on cardiometabolic traits and events: a Mendelian randomization analysis . Am J Hum Genet . 2014 ; 94 ( 2 ): 198 – 208 . Google Scholar CrossRef Search ADS PubMed 27. Meyre D . Is FTO a type 2 diabetes susceptibility gene ? Diabetologia . 2012 ; 55 ( 4 ): 873 – 876 . Google Scholar CrossRef Search ADS PubMed 28. Rydén L , Standl E , Bartnik M , Van den Berghe G , Betteridge J , de Boer MJ , Cosentino F , Jönsson B , Laakso M , Malmberg K , Priori S , Ostergren J , Tuomilehto J , Thrainsdottir I , Vanhorebeek I , Stramba-Badiale M , Lindgren P , Qiao Q , Priori SG , Blanc JJ , Budaj A , Camm J , Dean V , Deckers J , Dickstein K , Lekakis J , McGregor K , Metra M , Morais J , Osterspey A , Tamargo J , Zamorano JL , Deckers JW , Bertrand M , Charbonnel B , Erdmann E , Ferrannini E , Flyvbjerg A , Gohlke H , Juanatey JR , Graham I , Monteiro PF , Parhofer K , Pyörälä K , Raz I , Schernthaner G , Volpe M , Wood D ; Task Force on Diabetes and Cardiovascular Diseases of the European Society of Cardiology (ESC) ; European Association for the Study of Diabetes (EASD) . Guidelines on diabetes, pre-diabetes, and cardiovascular diseases: executive summary . Eur Heart J . 2007 ; 28 ( 1 ): 88 – 136 . Google Scholar PubMed 29. Ortega FB , Lavie CJ , Blair SN . Obesity and cardiovascular disease . Circ Res . 2016 ; 118 ( 11 ): 1752 – 1770 . Google Scholar CrossRef Search ADS PubMed 30. Jensen MD , Ryan DH , Apovian CM , Ard JD , Comuzzie AG , Donato KA , Hu FB , Hubbard VS , Jakicic JM , Kushner RF , Loria CM , Millen BE , Nonas CA , Pi-Sunyer FX , Stevens J , Stevens VJ , Wadden TA , Wolfe BM , Yanovski SZ , Jordan HS , Kendall KA , Lux LJ , Mentor-Marcel R , Morgan LC , Trisolini MG , Wnek J , Anderson JL , Halperin JL , Albert NM , Bozkurt B , Brindis RG , Curtis LH , DeMets D , Hochman JS , Kovacs RJ , Ohman EM , Pressler SJ , Sellke FW , Shen WK , Smith SC Jr , Tomaselli GF ; American College of Cardiology/American Heart Association Task Force on Practice Guidelines ; Obesity Society . 2013 AHA/ACC/TOS guideline for the management of overweight and obesity in adults: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines and The Obesity Society . Circulation . 2014 ; 129 ( 25 , Suppl 2 ) S102 – S138 . Google Scholar CrossRef Search ADS PubMed 31. Patel A , MacMahon S , Chalmers J , Neal B , Billot L , Woodward M , Marre M , Cooper M , Glasziou P , Grobbee D , Hamet P , Harrap S , Heller S , Liu L , Mancia G , Mogensen CE , Pan C , Poulter N , Rodgers A , Williams B , Bompoint S , de Galan BE , Joshi R , Travert F , Travert F ; ADVANCE Collaborative Group . Intensive blood glucose control and vascular outcomes in patients with type 2 diabetes . N Engl J Med . 2008 ; 358 ( 24 ): 2560 – 2572 . Google Scholar CrossRef Search ADS PubMed 32. Gerstein HC , Miller ME , Byington RP , Goff DC Jr , Bigger JT , Buse JB , Cushman WC , Genuth S , Ismail-Beigi F , Grimm RH Jr , Probstfield JL , Simons-Morton DG , Friedewald WT ; Action to Control Cardiovascular Risk in Diabetes Study Group . Effects of intensive glucose lowering in type 2 diabetes . N Engl J Med . 2008 ; 358 ( 24 ): 2545 – 2559 . Google Scholar CrossRef Search ADS PubMed 33. Ray KK , Seshasai SR , Wijesuriya S , Sivakumaran R , Nethercott S , Preiss D , Erqou S , Sattar N . Effect of intensive control of glucose on cardiovascular outcomes and death in patients with diabetes mellitus: a meta-analysis of randomised controlled trials . Lancet . 2009 ; 373 ( 9677 ): 1765 – 1772 . Google Scholar CrossRef Search ADS PubMed 34. Wada K , Yatsuya H , Ouyang P , Otsuka R , Mitsuhashi H , Takefuji S , Matsushita K , Sugiura K , Hotta Y , Toyoshima H , Tamakoshi K . Self-reported medical history was generally accurate among Japanese workplace population . J Clin Epidemiol . 2009 ; 62 ( 3 ): 306 – 313 . Google Scholar CrossRef Search ADS PubMed Copyright © 2018 Endocrine Society

Journal

Journal of Clinical Endocrinology and MetabolismOxford University Press

Published: Mar 1, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off