Mendelian randomization is a special type of instrumental variable analysis. Its application in the medical field increases in popularity because of its obvious advantages and the rapid development of genomics. This article aimed to introduce the basic concepts, principles, common methods, and limitations of Mendelian randomization. It is expected to provide guidance for researchers to conduct Mendelian randomization studies.
Mendelian randomization (MR) studies use genetic variants as instrumental variables to explore the effects of exposures on health outcomes. STROBE-MR (strengthening the reporting of observational studies in epidemiology using Mendelian randomization) assists authors in reporting their MR studies clearly and transparently, and helpfully to improve the quality of MR. This paper interpreted the STROBE-MR, aiming to help Chinese scholars better understand, disseminate, and apply it.
Objective To investigate the causal relationship between resistin and multiple myeloma (MM). Methods A two-sample Mendelian randomization analysis was conducted using genetic variants (SNPs) associated with resistin as instrumental variables and MM genome-wide association study (GWAS) data as the outcome event. Five analysis methods, including inverse-variance weighted (IVW), MR-Egger, weighted median, weighted model, and simple model were used to assess the causal impact of resistin on the risk of MM. Results None of the five analysis methods showed a causal relationship between resistin and multiple myeloma (P>0.05). Sensitivity analysis indicated consistent and robust results, with no evidence of horizontal pleiotropy, heterogeneity, outliers, or individual SNPs influencing the findings. Conclusion This Mendelian randomization study provides no support for a causal relationship between resistin and the risk of multiple myeloma.
ObjectiveTo analyze the causal relationship between obstructive sleep apnea (OSA) with its typical symptoms (daytime sleepiness and snoring) and cardiovascular diseases (hypertension, coronary heart disease, myocardial infarction, heart failure) by using Mendelian randomization. MethodsWe used the instrumental variables (IV) in the FINNGen database and the UK Biobank to perform two-sample Mendelian randomization (TSMR) analysis. The results of random-effects inverse variance weighting method (IVW) were the main results. MR-Egger method was used for pleiotropic analysis and sensitivity analysis was performed by the leave-one-out method to verify the reliability of the data. ResultsOSA could lead to hypertension (IVW β=0.043, 95%CI 0.012 to 0.074, P=0.006) and heart failure (IVW β=0.234, 95%CI 0.015 to 0.452, P=0.036). Daytime sleepiness also had a pathogenic effect on heart failure (IVW β=1.139, 95%CI 0.271 to 2.006, P=0.010). There was no causal association between OSA and CHD or MI, snoring and the four CVDs. There was no causal association between daytime sleepiness and hypertension, CHD or MI.ConclusionOSA and daytime sleepiness have pathogenic effects on hypertension and heart failure, with heart failure being the most affected.
Objective To assess any potential associations between lung cancer and gut microbiota. Methods Mendelian randomization (MR) analysis was carried out by utilizing summary data from genome-wide association studies (GWAS) of the gut microbiota and lung cancer. The gut microbiota served as an exposure. Instrumental ariables (IVs) were identified from the GWAS of 18340 participants. The GWAS study of lung cancer from Europe served as an outcome, including 29 266 lung cancer patients and 56450 controls. We used the inverse-variance weighted (IVW) method as the primary analysis. Sensitivity analysis was used to test the reliability of MR analysis results. Results IVW results showed that Genus Parabacteroides (OR=1.258, 95%CI 1.034 to 1.531, P=0.022) and Phylum Bacteroidetes (OR=1.192, 95%CI 1.001 to 1.419, P=0.048) had a positive causal association with lung cancer, and there was a negative causal association between family Bifidobacteriaceae (OR=0.845, 95%CI 0.721 to 0.989, P=0.037) and order Bifidobacteriales (OR=0.865, 95%CI 0.721 to 0.989, P=0.037) with lung cancer. Sensitivity analysis showed no evidence of reverse causality, pleiotropy, and heterogeneity. Conclusion This study demonstrates that Genus Parabacteroides and Phylum Bacteroidetes are related to an increased risk of lung cancer, family Bifidobacteriaceae and order Bifidobacteriales can reduce the risk of lung cancer. Our thorough investigations provide evidence in favor of a potential causal relationship between a number of gut microbiota-taxa and lung cancer. To demonstrate how gut microbiota influences the development of lung cancer, further research is necessary.
ObjectiveTo investigate the causal association between metabolic syndrome (MetS) components and osteoarthritis of the knee (KOA) by using Mendelian randomization analysis. MethodsThe genome-wide association study database (GWAS) was mined, in which the exposure factors were MetS components, namely waist circumference (WC) level, triglyceride (TG) level, high-density lipoprotein cholesterol (HDL-C) level, hypertension (HBP), and type 2 diabetes (T2DM), and the outcome factor was KOA. Mendelian randomization analysis was performed using regression models of inverse-variance weighted (IVW), MR-Egger, Simple Mode, Weighted Median, and Weighted Mode methods. ResultsIVW showed a causal relationship between WC level and KOA with a positive correlation (OR=3.088, 95%CI 2.574 to 3.704, P<0.01), and HDL-C level had a causal relationship with KOA with a negative correlation (OR=0.877, 95%CI 0.779 to 0.989, P<0.05). IVW did not show a causal relationship between TG levels, HBP, and T2DM with KOA (P>0.05). The results of the ME-Egger intercept test were not multiplicative (P>0.05), indicating that Mendelian randomization was a valid method for causal inference in this study. ConclusionCentral obesity and low HDL-C disorder are independent risk factors for KOA. The causal relationship between TG level, HBP, and T2DM with KOA is still uncertain.
ObjectiveTo identify causal effects and potential mechanisms of oxidative stress (OS) genes in lung cancer. MethodsOS-related genes were extracted from the GeneCards database. Integration analysis of genome-wide association study (GWAS) data for lung cancer with gene expression and DNA methylation quantitative trait loci (eQTL and mQTL) in blood was performed using the summary data-based Mendelian randomization (SMR) approach to determine the causal relationship between OS genes and lung cancer risk. Colocalization analysis of OS gene QTLs and lung cancer risk loci was performed to gain insight into the potential regulatory mechanisms of lung cancer risk. ResultsA potential causal relationship between OS-related genes and lung cancer was identified by SMR analysis. AGER expression level was found to be associated with lung cancer risk [OR=1.944, 95%CI (1.431, 2.640), P<0.001], and ATF6B expression level was associated with lung cancer risk [OR=1.508, 95%CI (1.287, 1.767), P<0.001]. Meanwhile, ATF6B methylation level was associated with lung cancer risk. ConclusionOS-related genes are associated with lung cancer, which may be a potential site of action for anti-cancer drugs.
Objective To analyze the causal relationship between educational attainment and the risk of systemic lupus erythematosus (SLE). Methods Based on the data from publicly available genome-wide association studies, we employed single nucleotide polymorphisms (SNPs) strongly associated with educational attainment as instrumental variables. Two-sample Mendelian randomization analysis was used to investigate the causal relationship between educational attainment and SLE. The primary analysis method used was the inverse variance weighted with multiplicative random effects. Validation methods included inverse variance weighted with fixed effects and MR-Egger methods. Additionally, sensitivity analysis was conducted using the leave-one-out approach. Results Finally, 433 SNPs were included. The inverse variance weighted with multiplicative random effects analysis indicated no causal effect of educational attainment on the risk of SLE [odds ratio =1.111, 95% confidence interval (0.813, 1.518), P=0.509]. Similarly, the other two methods did not find any evidence of a causal relationship (P>0.05); however, significant heterogeneity was observed. The MR-Egger regression analysis provided no evidence of horizontal pleiotropy among the included instrumental variables (P>0.05). The leave-one-out approach did not identify any individual SNP that had a significant impact on the overall effect estimate. ConclusionIn conclusion, this study does not support a causal effect of educational attainment on the risk of SLE.
Objective To investigate the association between neuroticism and gastroesophageal reflux disease (GERD) using Mendelian randomization (MR). Methods Exposure and outcome data were downloaded from the IEU database (https://gwas.mrcieu.ac.uk/), containing summary statistics from genome-wide association studies (GWAS) for neuroticism (n=374 323) and gastroesophageal reflux disease (n=602 604). Using the weighted median (WM), MR-Egger, inverse variance weighted (IVW), weighted mode and simple mode methods for Mendelian randomization analysis. Odds ratio (OR) values were used to assess the causal relationship, while sensitivity analysis was used to ensure the accuracy of the results. ResultsNeuroticism (OR=1.229, 95%CI 1.186-1.274, P<0.001) was associated with an increased risk of GERD. Meanwhile, gastroesophageal reflux disease (OR=1.786, 95%CI 1.623-1.965, P<0.001) was also associated with increased risk of neuroticism. Conclusion The study finds a bidirectional causal relationship between neuroticism and gastroesophageal reflux disease.
Objective Mendelian randomization (MR) was used to analyze the potential relationship between blood pressure and proliferative diabetic retinopathy (PDR). MethodsTwo-sample MR analysis was performed using summary statistics from genome-wide association studies. Systolic blood pressure (SBP) and diastolic blood pressure (DBP) were selected as the exposure, PDR as the outcome. The instrumental variable of SBP and DBP came from the publicly available data of the the UK Medical Research Council Comprehensive Epidemiology Unit and Neale Laboratory; the outcome data (8 681 cases in the case group, 204 208 cases in the control group, European population) are from the FinnGen database. Inverse variance weighting (IVW) and weighted median (WM) were used to analyze the potential relationships between SBP, DBP and PDR. ResultsMR analysis showed that IVW [SBP: odds ratio (OR)=1.36, 95% confidence interval (CI) 1.17-1.57, P=4.22E-05; DBP: OR=1.29, 95%CI 1.11-1.51, P=8.6E-04], WM (SBP: OR=1.33, 95%CI 1.07-1.66, P=0.009; DBP: OR=1.28, 95%CI=1.03-1.59, P=0.002). The results showed that elevated SBP and DBP increased the risk of PDR. ConclusionBlood pressure (SBP, DBP) change is positively correlated with the risk of PDR.