Atrial Fibrillation Screening And Risk Factors In Elderly Community Residents:A Mendelian Randomization Study

Background:Atrial fibrillation（AF）,one of the most common cardiac arrhythmias,is associated with a five-fold increased risk of stroke.The lifetime risk of AF in China is approximately 20%and increases with age,and the disease burden of AF in the elderly population is a major public health problem.In patients with unrecognized and untreated AF,the progression from paroxysmal to persistent or permanent AF may be faster.Two-sample Mendelian randomization（TSMR）study was based on the data of genome-wide association studies（GWAS）.The genetic variations associated to the exposure factor was used as the instrumental variables（IVs）to assess the potential causal effect between the exposure and the increased risk of AF.TSMR study can avoid the issues of confounding and reserve causation.Therefore,in this study,the older residents were conducted with the AF screening.The case-control study and TSMR study were used to find the risk factors associated with the increased risk of AF.Objectives:In this study,the elder people aged ≥50 years old in Nanjing were conducted with the AF screening to understand the detection rate of AF in this area.Secondly,an epidemiological case-control study was used to explore the potential risk factors of AF.Finally,a TSMR study was carried out based on public GWAS data to explore whether there is a causal relationship between uric acid（UA）and AF.Methods:1.A pilot study on AF screening was conducted among 948 retired people aged 50 and above in a community in Pukou district,Nanjing from June to July 2017.Subsequently,from March to April 2019,3842 participants living in Qixia district were recruited during the health examination and the epidemiological data were collected.The detection rate of AF in this community was calculated,and then the standardized detection rate of AF was calculated using the data of the sixth national population census in 2010.2.The second part was a case-control study.The cases were included in two parts:AF cases newly screened in the community,and the cases newly diagnosed in the Cardiovascular Department of Jiangsu Provincial People’s Hospital.The control group was selected from population without AF in the AF screening.The groups of case and control were matched at a ratio of 1:1 by propensity score matching approach by gender and age.The DAGitty v3.0 was used to draw directed acyclic graphs to detect the potential confounders.Multiple logistic regression models were performed to explore the potential risk factors.The receiver operating characteristic（ROC）curve analysis was used to evaluate the value of the level of UA in predicting AF.Moreover,the dose-response relationship between UA and AF risk was analyzed using Restricted Cubic Splines analysis.3.The datasets from the serum UA Global Urate Genetics Consortium（GUGC）GWAS and the 2018 AF HRC GWAS were used for the TSMR analysis.The method of Wald Ratio(β_XY=β_ZY/β_ZX)was used to evaluate the effect of UA on the risk of AF using each SNP.Moreover,the methods of inverse variance weighted（IVW）,weighted median（WM）,MR-Egger regression analysis and the genetic risk score（GRS）were performed to infer the causal relationships between the UA and AF risks using multiple SNPs.Cochran’s Q test was used to estimate heterogeneities among SNPs.The method of MR pleiotropy residual sum and outlier（MR-PRESSO）was used to identify outlying SNP that was potentially horizontally pleiotropic.Results:1.A total of 4 790 people participated in the health examination,excluding 384 participants lacking ECG data,and finally 4 406 subjects were involved in the AF screening,with an effective rate of 92.0%.Among these participants,49 were detected with AF,with a detection rate of 1.11%,including 2.06%in males and 0.62%in females（P<0.05）.The detection rate of AF was 0.12%among the 50～59 years old group,0.88%in the 60～69 years old group,2.77%in the 70～79 years old group,and 6.90%in the more than 80 years old group,which were showed a significant increased trend along with the age（P<0.05）.Standardized by the age and gender based on the data from the sixth national census in 2010,the standardized detection rate of AF among the people more than 50 years old was 1.21%,of which males were 1.35%,females were 0.94%.2.Multivariate logistic regression showed that after adjusting for age and gender in Model 1,individuals with hyperuricemia,hyperlipidemia,stroke,and hyperthyroidism had an increased risk of AF ranged from 64%to 484%（P<0.05）.For all the serum biomarkers,individuals with low levels of LMR,total cholesterol（TC）,HDL-C,LDL-C,and high levels of PLR,UA,and urea nitrogen were significantly associated with the increased risk of AF（P<0.05）.In Model 2,after adjusted with the factors of age,gender,smoking,alcohol consumption,hypertension,diabetes,hyperlipidemia,coronary artery disease,and stroke,the subjects with the former factors except for the stroke had the significantly increased risk of AF（P<0.05）.Both in Model1 and Model 2,the subjects with the higher-level of UA（>313μmol/L,median of the control group）had 128%（OR=2.28,95%Cl:1.91-2.72）and 116%（OR=2.16,95%Cl:1.80-2.59）increased risk of AF,respectively.The diagnostic value of UA on AF risk was relatively high,with an area under the ROC curve of 0.62（95%CI:0.60-0.64）.Moreover,a significant dose-response relationship between UA level and AF risk was found in the restricted cubic spline analysis.3.A total of 24 SNPs significantly associated with the UA(P<5×10^-8)were used as the IVs in the TSMR analysis.When the single SNP of rs478607（β=-0.677,95%Cl:-1.081～-0.272）,rs7953704（β=-0.645,95%Cl:-1.125～-0.165）,rs17786744（β=-0.514,95%Cl:-1.007～-0.020）and rs6598541（β=1.072,95%Cl:0.739～1.405）was used as IV,the causal relationship between UA and AF risk was found significantly in the TSMR analysis（P<0.05）.After MR-PRESSO test,two SNPs（rs478607,rs6598541）were found to be outliers.Using the methods of IVW,WM,MR-Egger and GRS calculated with the multiple SNPs as IVs for MR analysis,the results of group 1(N_SNPs=24),group 2(N_SNPs=22,remove outliers),group 3(N_SNPs=4,single SNP is meaningful)were not found the significant causal association between UA level and the risk of AF.After Cochran’s Q test and MR-egger intercept test,group 2 were not found with the heterogeneity and pleiotropy（P>0.05）.Conclusions:1.Among people more than 50 years old in Nanjing communities,the crude detection rate of AF was 1.11%（2.06%for males,0.62%for females）,and the standardized detection rate of AF was 1.21%（1.35%for males,0.94%for females）.2.Multivariate logistic regression analysis showed that combined with hyperlipidemia,hyperthyroidism,high levels of PLR,UA,urea nitrogen,low levels of LMR were the potential risk factors for the increased risk of AF.There was a significant dose-response relationship between the level of serum UA and the risk of AF.3.The TSMR analysis found that there was a causal relationship between UA and AF when the single SNP of rs478607,rs7953704,rs17786744 and rs6598541 was used as IV.Based on the limitations of TSMR design,the effect was needed the further demonstration.