Re: Estimation of bias in nongenetic observational studies using "Mendelian triangulation" by Bautista et al

A Note on Modelling Bidirectional Feedback Loops in Mendelian Randomization Studies

Structural equation models (SEMs) involving feedback loops may offer advantages over standard instrumental variables estimators in terms of modelling causal effects in the presence of bidirectional relationships. In the following note, we show that in the case of a single "exposure" and "outcome" variable, modelling relationships using a SEM with a simple bidirectional linear feedback loop offers no advantage over traditional instrumental variables estimators in terms of consistency (i.e. both approaches yield consistent estimates of the causal effect, provided that causal estimates are obtained in both directions). In the case of finite samples, traditional IV estimators and SEM exhibited similar power across many of the conditions we examined, although which method performed best depended on the residual correlation between variables and the strength of the instruments. In particular, the power of SEM was insensitive to the residual correlation between variables, whereas the power of the Wald estimator/2SLS improved (deteriorated) relative to SEM as the magnitude of the residual correlation increased (decreased) assuming a positive causal effect of the exposure on the outcome. The power of SEM improved relative to the Wald estimator/2SLS as the instruments explained more residual variance in the "outcome" variable.

Genetic determinants of complement activation in the general population

Complement is a fundamental innate immune response component. Its alterations are associated with severe systemic diseases. To illuminate the complement's genetic underpinnings, we conduct genome-wide association studies of the functional activity of the classical (CP), lectin (LP), and alternative (AP) complement pathways in the Cooperative Health Research in South Tyrol study (n = 4,990). We identify seven loci, encompassing 13 independent, pathway-specific variants located in or near complement genes (CFHR4, C7, C2, MBL2) and non-complement genes (PDE3A, TNXB, ABO), explaining up to 74% of complement pathways' genetic heritability and implicating long-range haplotypes associated with LP at MBL2. Two-sample Mendelian randomization analyses, supported by transcriptome- and proteome-wide colocalization, confirm known causal pathways, establish within-complement feedback loops, and implicate causality of ABO on LP and of CFHR2 and C7 on AP. LP causally influences collectin-11 and KAAG1 levels and the risk of mouth ulcers. These results build a comprehensive resource to investigate the role of complement in human health.

Cannabis use and atherosclerotic cardiovascular disease: a Mendelian randomization study

BACKGROUND

Association between cannabis use and development of atherosclerotic cardiovascular disease (ASCVD) is inconsistent and challenging to interpret, given existing study limitations.

METHODS

Sixty five independent single-nucleotide polymorphisms (SNPs), obtained from a genome-wide association study on lifetime cannabis use, were employed as genetic instruments to estimate the effects of genetically indexed cannabis use on risk of coronary artery disease (CAD) and acute ischemic stroke (IS) using a two-sample Mendelian randomization (MR) approach. Summary statistics on CAD (CARDIoGRAMplusC4D; 60,801 cases and 123,504 controls) and IS (MEGASTROKE; 34,217 cases and 406,111 controls) were obtained separately. A comprehensive review of the observational literature on cannabis use and CAD or IS was also performed and contrasted with MR results.

RESULTS

There was no causal effect of cannabis use on the risk of CAD (odds ratio (OR) per ever-users vs. never-users 0.93; 95% confidence interval (CI), 0.83 to 1.03) or IS (OR 1.05; 95%CI, 0.93 to 1.19). Sensitivity analyses yielded similar results, and no heterogeneity and directional pleiotropy was observed. Our meta-analysis of observational studies showed no significant association between ever use of cannabis with risk of CAD (k = 6 studies; ORpooled = 1.23, 95%CI 0.78 to 1.69), nor with IS (k = 6 studies; ORpooled = 1.22, 95%CI 0.95 to 1.50).

CONCLUSION

Using a genetic approach approximating a clinical trial does not provide evidence consistent with a causal effect of genetic predisposition to cannabis use on CAD or IS development. Further studies are needed to replicate our findinds, an to investigate more precisely the risk of ASCVD in relation to the quantity, type, route of administration, or the age at exposure to cannabis.

Role of DNA methylation in the relationship between glioma risk factors and glioma incidence: a two-step Mendelian randomization study

Genetic evidence suggests glioma risk is altered by leukocyte telomere length, allergic disease (asthma, hay fever or eczema), alcohol consumption, childhood obesity, low-density lipoprotein cholesterol (LDLc) and triglyceride levels. DNA methylation (DNAm) variation influences many of these glioma-related traits and is an established feature of glioma. Yet the causal relationship between DNAm variation with both glioma incidence and glioma risk factors is unknown. We applied a two-step Mendelian randomization (MR) approach and several sensitivity analyses (including colocalization and Steiger filtering) to assess the association of DNAm with glioma risk factors and glioma incidence. We used data from a recently published catalogue of germline genetic variants robustly associated with DNAm variation in blood (32,851 participants) and data from a genome-wide association study of glioma risk (12,488 cases and 18,169 controls, sub-divided into 6191 glioblastoma cases and 6305 non-glioblastoma cases). MR evidence indicated that DNAm at 3 CpG sites (cg01561092, cg05926943, cg01584448) in one genomic region (HEATR3) had a putative association with glioma and glioblastoma risk (False discovery rate [FDR] < 0.05). Steiger filtering provided evidence against reverse causation. Colocalization presented evidence against genetic confounding and suggested that differential DNAm at the 3 CpG sites and glioma were driven by the same genetic variant. MR provided little evidence to suggest that DNAm acts as a mediator on the causal pathway between risk factors previously examined and glioma onset. To our knowledge, this is the first study to use MR to appraise the causal link of DNAm with glioma risk factors and glioma onset. Subsequent analyses are required to improve the robustness of our results and rule out horizontal pleiotropy.

Body shape and risk of glaucoma: A Mendelian randomization

Background

Body size (BS) is one of the risk factors for the development of many clinical diseases, but the relationship between BS and glaucoma is controversial. Herein, we try to use Mendelian randomization (MR) method to study BS causal association with glaucoma risk from the genetic level.

Methods

The Body Size was determined through anthropometric traits (ATs), such as body mass index (BMI), waist-to-hip ratio adjusted by body mass index (WHRadjBMI), waist-to-hip ratio (WHR), and waist circumference (WC). Association of single nucleotide polymorphisms (SNPs) with each AT and glaucoma were determined individually from the aggregated data of the Genetic Investigation of Anthropometric Traits (GIANT) consortium and the FinnGen study summary data (8,591 cases with glaucoma and 210,201 controls). To explore the role of BS and glaucoma, a two-sample MR analysis was performed on genome-wide association study (GWAS) data. Besides, three MR methods [inverse variance weighted (IVW), Weighted median, and MR-Egger regression] were used to get the whole causal estimate for multiple instrumental SNPs.

Results

BMI (OR = 1.20; 95% CI = 1.02-1.41; P = 0.03) and WC (OR = 1.32; 95% CI =1.04-1.69; P = 0.03) were associated with a risk of glaucoma. Besides, genetically predicted WHRadjBMI (OR = 1.10; 95% CI = 0.88-1.35; P = 0.43) and WHR (OR = 1.22; 95% CI = 0.93-1,572; P = 0.14) were not associated with glaucoma. No heterogeneity and directional pleiotropy were detected.

Conclusion

The data of this study revealed that increased BMI and WC are potential risk factors for glaucoma, and WHRadjBMI and WHR are not associated with the occurrence of glaucoma.

Mendelian Randomization Analysis of the Association of SOCS3 Methylation with Abdominal Obesity

This study was conducted to evaluate the potential causality association of SOCS3 methylation with abdominal obesity using Mendelian randomization. A case-control study, including 1064 participants, was carried out on Chinese subjects aged 18 to 79. MethylTargetTM was used to detect the methylation level for each CpG site of SOCS3, and SNPscan^® was applied to measure the single-nucleotide polymorphism (SNP) genotyping. The logistic regression was used to assess the relationship of SOCS3 methylation level and SNP genotyping with abdominal obesity. Three types of Mendelian randomization methods were implemented to examine the potential causality between SOCS3 methylation and obesity based on the SNP of SOCS3 as instrumental variables. SOCS3 methylation levels were inversely associated with abdominal obesity in five CpG sites (effect estimates ranged from 0.786 (Chr17:76356054) to 0.851 (Chr17:76356084)), and demonstrated positively association in 18 CpG sites (effect estimates ranged from 1.243 (Chr17:76354990) to 1.325 (Chr17:76355061)). The causal relationship between SOCS3 methylation and abdominal obesity was found using the maximum-likelihood method and Mendelian randomization method of penalized inverse variance weighted (MR-IVW), and the β values (95% CI) were 5.342 (0.215, 10.469) and 4.911 (0.259, 9.564), respectively. The causality was found between the SOCS3 methylation level and abdominal obesity in the Chinese population.

Differential Effects of Genetically Determined Cholesterol Efflux Capacity on Coronary Artery Disease and Ischemic Stroke

Background

Observational studies indicated that cholesterol efflux capacity (CEC) of high-density lipoprotein (HDL) is inversely associated with cardiovascular events, independently of the HDL cholesterol concentration. The aim of the study is to examine the casual relevance of CEC for coronary artery disease (CAD) and myocardial infarction (MI), and compare it with that for ischemic stroke and its subtypes using a Mendelian randomization approach.

Methods

We performed a 2-sample Mendelian randomization to estimate the casual relationship of CEC with the risk of CAD, MI, and ischemic stroke. A CEC-related genetic variant (rs141622900) and other five genetic variants were used as the instrumental variables. Association of genetic variants with CAD were estimated in a GWAS involving 60,801 CAD cases and 123,504 controls. They were then compared with the associations of these variants with ischemic stroke and its subtypes (large vessel, small vessel, and cardioembolic) involving 40,585 ischemic stroke cases and 406,111 controls.

Results

Using the SNP of rs141622900 as the instrument, a 1-SD increase in CEC was associated with 45% lower risk for CAD (odds ratio [OR] 0.55, 95% confidence interval [CI] 0.44–0.69, p < 0.001) and 33% lower risk for MI (odds ratio [OR] 0.67, 95% CI 0.52–0.87, p = 0.002). By contrast, the causal effect of CEC was much weaker for ischemic stroke (odds ratio [OR] 0.79, 95% CI 0.64–0.97, p = 0.02; p for heterogeneity = 0.03) and, in particular, for cardioembolic stroke (p for heterogeneity = 0.006) when compared with that for CAD. Results using five genetic variants as the instrument also indicated consistently weaker effects on ischemic stroke than on CAD.

Conclusion

Genetic predicted higher CEC may be associated with decreased risk of CAD. However, the casual association of CEC with ischemic stroke and specific subtypes would need to be validated in further Mendelian randomization studies.

Exploration of errors in variance caused by using the first-order approximation in Mendelian randomization

Mendelian randomization (MR) uses genetic variation as a natural experiment to investigate the causal effects of modifiable risk factors (exposures) on outcomes. Two-sample Mendelian randomization (2SMR) is widely used to measure causal effects between exposures and outcomes via genome-wide association studies. 2SMR can increase statistical power by utilizing summary statistics from large consortia such as the UK Biobank. However, the first-order term approximation of standard error is commonly used when applying 2SMR. This approximation can underestimate the variance of causal effects in MR, which can lead to an increased false-positive rate. An alternative is to use the second-order approximation of the standard error, which can considerably correct for the deviation of the first-order approximation. In this study, we simulated MR to show the degree to which the first-order approximation underestimates the variance. We show that depending on the specific situation, the first-order approximation can underestimate the variance almost by half when compared to the true variance, whereas the second-order approximation is robust and accurate.

Exploring the causal effect of maternal pregnancy adiposity on offspring adiposity: Mendelian randomisation using polygenic risk scores

BACKGROUND

Greater maternal adiposity before or during pregnancy is associated with greater offspring adiposity throughout childhood, but the extent to which this is due to causal intrauterine or periconceptional mechanisms remains unclear. Here, we use Mendelian randomisation (MR) with polygenic risk scores (PRS) to investigate whether associations between maternal pre-/early pregnancy body mass index (BMI) and offspring adiposity from birth to adolescence are causal.

METHODS

We undertook confounder adjusted multivariable (MV) regression and MR using mother-offspring pairs from two UK cohorts: Avon Longitudinal Study of Parents and Children (ALSPAC) and Born in Bradford (BiB). In ALSPAC and BiB, the outcomes were birthweight (BW; N = 9339) and BMI at age 1 and 4 years (N = 8659 to 7575). In ALSPAC only we investigated BMI at 10 and 15 years (N = 4476 to 4112) and dual-energy X-ray absorptiometry (DXA) determined fat mass index (FMI) from age 10-18 years (N = 2659 to 3855). We compared MR results from several PRS, calculated from maternal non-transmitted alleles at between 29 and 80,939 single nucleotide polymorphisms (SNPs).

RESULTS

MV and MR consistently showed a positive association between maternal BMI and BW, supporting a moderate causal effect. For adiposity at most older ages, although MV estimates indicated a strong positive association, MR estimates did not support a causal effect. For the PRS with few SNPs, MR estimates were statistically consistent with the null, but had wide confidence intervals so were often also statistically consistent with the MV estimates. In contrast, the largest PRS yielded MR estimates with narrower confidence intervals, providing strong evidence that the true causal effect on adolescent adiposity is smaller than the MV estimates (Pdifference = 0.001 for 15-year BMI). This suggests that the MV estimates are affected by residual confounding, therefore do not provide an accurate indication of the causal effect size.

CONCLUSIONS

Our results suggest that higher maternal pre-/early-pregnancy BMI is not a key driver of higher adiposity in the next generation. Thus, they support interventions that target the whole population for reducing overweight and obesity, rather than a specific focus on women of reproductive age.

Mendelian Randomization: Concepts and Scope

Mendelian randomization (MR) is a method of studying the causal effects of modifiable exposures (i.e., potential risk factors) on health, social, and economic outcomes using genetic variants associated with the specific exposures of interest. MR provides a more robust understanding of the influence of these exposures on outcomes because germline genetic variants are randomly inherited from parents to offspring and, as a result, should not be related to potential confounding factors that influence exposure-outcome associations. The genetic variant can therefore be used as a tool to link the proposed risk factor and outcome, and to estimate this effect with less confounding and bias than conventional epidemiological approaches. We describe the scope of MR, highlighting the range of applications being made possible as genetic data sets and resources become larger and more freely available. We outline the MR approach in detail, covering concepts, assumptions, and estimation methods. We cover some common misconceptions, provide strategies for overcoming violation of assumptions, and discuss future prospects for extending the clinical applicability, methodological innovations, robustness, and generalizability of MR findings.

Mendelian randomization analysis of the causal impact of body mass index and waist-hip ratio on rates of hospital admission

We analyze how measures of adiposity - body mass index (BMI) and waist hip ratio (WHR) - causally influence rates of hospital admission. Conventional analyses of this relationship are susceptible to omitted variable bias from variables that jointly influence both hospital admission and adipose status. We implement a novel quasi-Poisson instrumental variable model in a Mendelian randomization framework, identifying causal effects from random perturbations to germline genetic variation. We estimate the individual and joint effects of BMI, WHR, and WHR adjusted for BMI. We also implement multivariable instrumental variable methods in which the causal effect of one exposure is estimated conditionally on the causal effect of another exposure. Data on 310,471 participants and over 550,000 inpatient admissions in the UK Biobank were used to perform one-sample and two-sample Mendelian randomization analyses. The results supported a causal role of adiposity on hospital admissions, with consistency across all estimates and sensitivity analyses. Point estimates were generally larger than estimates from comparable observational specifications. We observed an attenuation of the BMI effect when adjusting for WHR in the multivariable Mendelian randomization analyses, suggesting that an adverse fat distribution, rather than a higher BMI itself, may drive the relationship between adiposity and risk of hospital admission.

Roles of C-reactive protein polymorphisms and life event changes on cognitive function in bipolar patients receiving valproate

Introduction

Patients with bipolar disorder (BD) exhibit an inflamed condition that is associated with metabolic disturbance and cognitive impairment. Whether inflammation, represented by C-reactive protein (CRP), is causally associated with BD and influences treatment outcome has not been established.

Methods

We examined whether CRP is a causal factor for the risk of BD in drug-naïve, depressed BD patients and investigated whether polymorphisms in CRP and life event changes influence cognitive function in BD patients receiving valproate (VPA) treatment.

Results

Our results showed that BD patients had significantly higher CRP levels and worse cognitive function than the controls, while the frequencies of CRP single nucleotide polymorphisms in BD patients and in controls were not different. In addition, the life event scale score was higher for BD patients than for controls. Furthermore, the genotypes of CRP polymorphisms and the interactions between polymorphisms of CRP and life event scale score had a significant influence on cognitive performance in BD patients after 12 weeks of VPA treatment.

Conclusion

Our study demonstrated the clinical utility of the application of functional genetics in clarifying the interactions among CRP, life event stress, and BD and suggested the important roles of CRP gene–environment interactions in developing treatment strategies for BD.

Estimating the influence of body mass index (BMI) on mortality using offspring BMI as an instrumental variable

OBJECTIVE

High body mass index (BMI) is an important predictor of mortality but estimating underlying causality is hampered by confounding and pre-existing disease. Here, we use information from the offspring to approximate parental BMIs, with an aim to avoid biased estimation of mortality risk caused by reverse causality.

METHODS

The analyses were based on information on 9674 offspring-mother and 9096 offspring-father pairs obtained from the 1958 British birth cohort. Parental BMI-mortality associations were analysed using conventional methods and using offspring BMI as a proxy, or instrument, for their parents' BMI.

RESULTS

In the conventional analysis, associations between parental BMI and all-cause mortality were U-shaped (P_curvature < 0.001), while offspring BMI had linear associations with parental mortality (P_trend < 0.001, P_curvature > 0.46). Curvature was particularly pronounced for mortality from respiratory diseases and from lung cancer. Instrumental variable analyses suggested a positive association between BMI and mortality from all causes [mothers: HR per SD of BMI 1.43 (95% CI 1.21-1.69), fathers: HR 1.17 (1.00-1.36)] and from coronary heart disease [mothers: HR 1.65 (1.15-2.36), fathers: HR 1.51 (1.17-1.97)]. These were larger than HR from the equivalent conventional analyses, despite some attenuation by adjustment for social indicators and smoking.

CONCLUSIONS

Analyses using offspring BMI as a proxy for parental BMI suggest that the apparent adverse consequences of low BMI are considerably overestimated and adverse consequences of overweight are underestimated in conventional epidemiological studies.

Target Discovery for Drug Development Using Mendelian Randomization

Making drug development more efficient by identifying promising drug targets can contribute to resource savings. Identifying promising drug targets using human genetic approaches can remove barriers related to translation. In addition, genetic information can be used to identify potentially causal relationships between a drug target and disease. Mendelian randomization (MR) is a class of approaches used to identify causal associations between pairs of genetically predicted traits using data from human genetic studies. MR can be used to prioritize candidate drug targets by predicting disease outcomes and adverse events that could result from the manipulation of a drug target. The theory behind MR is reviewed, including a discussion of MR assumptions, different MR analytical methods, tests for violations of assumptions, and MR methods that can be robust to some violations of MR assumptions. A protocol to perform two-sample MR (2SMR) with summary genome-wide association study (GWAS) results is described. An example of 2SMR examining the causal relationship between low-density lipoprotein (LDL) and coronary artery disease (CAD) is provided as an illustration of the protocol.

Associations between plasma fatty acid concentrations and schizophrenia: a two-sample Mendelian randomisation study

BACKGROUND

Although studies suggest that concentrations of omega-3 and omega-6 fatty acids are lower in individuals with schizophrenia, evidence for beneficial effects of fatty acid supplementation is scarce. Therefore, in this study, we aimed to determine whether omega-3 and omega-6 fatty acid concentrations are causally related to schizophrenia.

METHODS

We did a two-sample Mendelian randomisation study, using deidentified summary-level data that were publicly available. Exposure-outcome relationships were evaluated using the inverse variance weighted two-sample Mendelian randomisation method using results from genome-wide association studies (GWASs) of fatty acid concentrations and schizophrenia. GWAS results were available for European (fatty acids) and European and Asian (schizophrenia) ancestry samples. Overall age and gender information were not calculable from the summary-level GWAS results. Weighted median, weighted mode, and Mendelian randomisation Egger regression methods were used as sensitivity analyses. To address underlying mechanisms, further analyses were done using single instruments within the FADS gene cluster and ELOVL2 gene locus. FADS gene cluster and ELOVL2 gene causal effects on schizophrenia were calculated by dividing the single nucleotide polymorphism (SNP)-schizophrenia effect estimate by the SNP-fatty acid effect estimate with standard errors derived using the first term from a delta method expansion for the ratio estimate. Multivariable Mendelian randomisation was used to estimate direct effects of omega-3 fatty acids on schizophrenia, independent of omega-6 fatty acids, lipoproteins (ie, HDL and LDL), and triglycerides.

FINDINGS

Mendelian randomisation analyses indicated that long-chain omega-3 and long-chain omega-6 fatty acid concentrations were associated with a lower risk of schizophrenia (eg, inverse variance weighted odds ratio [OR] 0·83 [95% CI 0·75-0·92] for docosahexaenoic acid). By contrast, there was weak evidence that short-chain omega-3 and short-chain omega-6 fatty acids were associated with an increased risk of schizophrenia (eg, inverse variance weighted OR 1·07 [95% CI 0·98-1·18] for α-linolenic acid). Effects were consistent across the sensitivity analyses and the FADS single-SNP analyses, suggesting that long-chain omega-3 and long-chain omega-6 fatty acid concentrations were associated with lower risk of schizophrenia (eg, OR 0·74 [95% CI 0·58-0·96] for docosahexaenoic acid) whereas short-chain omega-3 and short-chain omega-6 fatty acid concentrations were associated with an increased risk of schizophrenia (eg, OR 1·08 [95% CI 1·02-1·15] for α-linolenic acid). By contrast, estimates from the ELOVL2 single-SNP analyses were more imprecise and compatible with both risk-increasing and protective effects for each of the fatty acid measures. Multivariable Mendelian randomisation indicated that the protective effect of docosahexaenoic acid on schizophrenia persisted after conditioning on other lipids, although evidence was slightly weaker (multivariable inverse variance weighted OR 0·84 [95% CI 0·71-1·01]).

INTERPRETATION

Our results are compatible with the protective effects of long-chain omega-3 and long-chain omega-6 fatty acids on schizophrenia, suggesting that people with schizophrenia might have difficulty converting short-chain polyunsaturated fatty acids to long-chain polyunsaturated fatty acids. Further studies are required to determine whether long-chain polyunsaturated fatty acid supplementation or diet enrichment might help prevent onset of schizophrenia.

FUNDING

National Institute for Health Research Biomedical Research Centre at University Hospitals Bristol and Weston NHS Foundation Trust and the University of Bristol.

Instrumental variable analysis using offspring BMI in childhood as an indicator of parental BMI in relation to mortality

Childhood BMI shows associations with adult mortality, but these may be influenced by effects of ill health in childhood on BMI and later mortality. To avoid this, we used offspring childhood BMI as an instrumental variable (IV) for own BMI in relation to mortality and compared it with conventional associations of own childhood BMI and own mortality. We included 36,097 parent-offspring pairs with measured heights and weights from the Copenhagen School Health Records Register and register-based information on death. Hazard ratios (HR) were estimated using adjusted Cox regression models. For all-cause mortality, per zBMI at age 7 the conventional HR = 1.07 (95%CI: 1.04-1.09) in women and 1.02 (95%CI: 0.92-1.14) in men, whereas the IV HR = 1.23 (95%CI: 1.15-1.32) in women and 1.05 (95%CI: 0.94-1.17) in men. Per zBMI at age 13, the conventional HR = 1.11 (95%CI: 1.08-1.15) in women and 1.03 (95%CI: 0.99-1.06) in men, whereas the IV HR = 1.30 (95%CI: 1.19-1.42) in women and 1.15 (95%CI: 1.04-1.29) in men. Only conventional models showed indications of J-shaped associations. Our IV analyses suggest that there is a causal relationship between BMI and mortality that is positive at both high and low BMI values.

Cannabis use does not impact on type 2 diabetes: A two-sample Mendelian randomization study

Cannabis has effects on the insulin/glucose metabolism. As the use of cannabis and the prevalence of type 2 diabetes increase worldwide, it is important to examine the effect of cannabis on the risk of diabetes. We conducted a Mendelian randomization (MR) study by using 19 single-nucleotide polymorphisms (SNPs) as instrumental variables for lifetime cannabis use and 14 SNPs to instrument cannabis use disorder and linking these to type 2 diabetes risk using genome-wide association study data (lifetime cannabis use [N = 184,765]; cannabis use disorder [2387 cases/48,985 controls], type 2 diabetes [74,124 cases/824,006 controls]). The MR analysis suggested no effect of lifetime cannabis use (inverse-variance weighted odds ratio [95% confidence interval] = 1.00 [0.93-1.09], P value = 0.935) and cannabis use disorder (OR = 1.03 [0.99-1.08]) on type 2 diabetes. Sensitivity analysis to assess potential pleiotropy led to no substantive change in the estimates. This study adds to the evidence base that cannabis use does not play a causal role in type 2 diabetes.

Shared Genetic Background between Parkinson's Disease and Schizophrenia: A Two-Sample Mendelian Randomization Study

Background and objectives: Parkinson's disease (PD) and schizophrenia often share symptomatology. Psychotic symptoms are prevalent in patients with PD, and similar motor symptoms with extrapyramidal signs are frequently observed in antipsychotic-naïve patients with schizophrenia as well as premorbid families. However, few studies have examined the relationship between PD and schizophrenia. We performed this study to evaluate whether genetic variants which increase PD risk influence the risk of developing schizophrenia, and vice versa. Materials and Methods: Two-sample Mendelian randomization (TSMR) with summary statistics from large-scale genome-wide association studies (GWAS) was applied. Summary statistics were extracted for these instruments from GWAS of PD and schizophrenia; Results: We found an increase in the risk of schizophrenia per one-standard deviation (SD) increase in the genetically-predicted PD risk (inverse-variance weighted method, odds ratio = 1.10; 95% confidence interval, 1.05-1.15; p = 3.49 × 10-5). The association was consistent in sensitivity analyses, including multiple TSMR methods, analysis after removing outlier variants with potential pleiotropic effects, and analysis after applying multiple GWAS subthresholds. No relationships were evident between PD and smoking or other psychiatric disorders, including attention deficit hyperactivity disorder, autism spectrum disorder, bipolar affective disorder, major depressive disorder, Alzheimer's disease, or alcohol dependence. However, we did not find a reverse relationship; genetic variants increasing schizophrenia risk did not alter the risk of PD; Conclusions: Overall, our findings suggest that increased genetic risk of PD can be associated with increased risk of schizophrenia. This association supports the intrinsic nature of the psychotic symptom in PD rather than medication or environmental effects. Future studies for possible comorbidities and shared genetic structure between the two diseases are warranted.

A novel Mendelian randomization method with binary risk factor and outcome

BACKGROUND

Mendelian randomization (MR) applies instrumental variable (IV) methods to observational data using a genetic variant as an IV. Several Monte-Carlo studies investigate the performance of MR methods with binary outcomes, but few consider them in conjunction with binary risk factors.

OBJECTIVE

To develop a novel MR estimator for scenarios with a binary risk factor and outcome; and compare to existing MR estimators via simulations and real data analysis.

METHODS

A bivariate Bernoulli distribution is adapted to the IV setting. Empirical bias and asymptotic coverage probabilities are estimated via simulations. The proposed method is compared to the Wald method, two-stage predictor substitution (2SPS), two-stage residual inclusion (2SRI), and the generalized method of moments (GMM). An analysis is performed using existing data from the CLEAR study to estimate the potential causal effect of smoking on rheumatoid arthritis risk in African Americans.

RESULTS

Bias was low for the proposed method and comparable to 2SPS. The Wald method was often biased towards the null. Coverage was adequate for the proposed method, 2SPS, and 2SRI. Coverage for the Wald and GMM methods was poor in several scenarios. The causal effect of ever smoking on rheumatoid arthritis risk was not statistically significant using a variety of genetic instruments.

CONCLUSIONS

Simulations suggest the proposed MR method is sound with binary risk factors and outcomes, and comparable to 2SPS and 2SRI in terms of bias. The proposed method also provides more natural framework for hypothesis testing compared to 2SPS or 2SRI, which require ad-hoc variance adjustments.

Mendelian Randomization With Refined Instrumental Variables From Genetic Score Improves Accuracy and Reduces Bias

Mendelian randomization (MR) can estimate the causal effect for a risk factor on a complex disease using genetic variants as instrument variables (IVs). A variety of generalized MR methods have been proposed to integrate results arising from multiple IVs in order to increase power. One of the methods constructs the genetic score (GS) by a linear combination of the multiple IVs using the multiple regression model, which was applied in medical researches broadly. However, GS-based MR requires individual-level data, which greatly limit its application in clinical research. We propose an alternative method called Mendelian Randomization with Refined Instrumental Variable from Genetic Score (MR-RIVER) to construct a genetic IV by integrating multiple genetic variants based on summarized results, rather than individual data. Compared with inverse-variance weighted (IVW) and generalized summary-data-based Mendelian randomization (GSMR), MR-RIVER maintained the type I error, while possessing more statistical power than the competing methods. MR-RIVER also presented smaller biases and mean squared errors, compared to the IVW and GSMR. We further applied the proposed method to estimate the effects of blood metabolites on educational attainment, by integrating results from several publicly available resources. MR-RIVER provided robust results under different LD prune criteria and identified three metabolites associated with years of schooling and additional 15 metabolites with indirect mediation effects through butyrylcarnitine. MR-RIVER, which extends score-based MR to summarized results in lieu of individual data and incorporates multiple correlated IVs, provided a more accurate and powerful means for the discovery of novel risk factors.

Lipid and metabolic syndrome traits in coronary artery disease: a Mendelian randomization study

Mendelian randomization (MR) of lipid traits in CAD has provided evidence for causal associations of LDL-C and TGs in CAD, but many lipid trait genetic variants have pleiotropic effects on other cardiovascular risk factors that may bias MR associations. The goal of this study was to evaluate pleiotropic effects of lipid trait genetic variants and to account for these effects in MR of lipid traits in CAD. We performed multivariable MR using inverse variance-weighted and MR-Egger methods in large (n ≥ 300,000) GWAS datasets. We found that 30% of lipid trait genetic variants have effects on metabolic syndrome traits, including BMI, T2D, and systolic blood pressure (SBP). Nonetheless, in multivariable MR analysis, LDL-C, HDL-C, TGs, BMI, T2D, and SBP are independently associated with CAD, and each of these associations is robust to adjustment for directional pleiotropy. MR at loci linked to direct effects on HDL-C and TGs suggests locus- and mechanism-specific causal effects of these factors on CAD.

Exploring the Relationship Between Psychiatric Traits and the Risk of Mouth Ulcers Using Bi-Directional Mendelian Randomization

Background

Although the association between mouth ulcers and psychiatric traits has been reported by observational studies, their causal relationship remains unclear. Mendelian randomization (MR), powered by large-scale genome-wide association studies (GWAS), provides an opportunity to clarify the causality between mouth ulcers and psychiatric traits.

Methods

We collected summary statistics of mouth ulcers (sample size n = 461,106) and 10 psychiatric traits from the largest publicly available GWAS on Europeans, including anxiety disorder (n = 83,566), attention deficit/hyperactivity disorder (n = 53,293), autism spectrum disorder (n = 46,350), bipolar disorder (n = 51,710), insomnia (n = 1,331,010), major depressive disorder (n = 480,359), mood instability (n = 363,705), neuroticism (n = 168,105), schizophrenia (n = 105,318), and subjective wellbeing (n = 388,538). We applied three two-sample bi-directional MR analysis methods, namely the Inverse Variance Weighted (IVW) method, the MR pleiotropy residual sum and outlier (MR-PRESSO) method, and the weighted median method, to assess the causal relationship between each psychiatric trait and mouth ulcers.

Results

We found significant effects of autism spectrum disorder, insomnia, major depressive disorder, and subjective wellbeing on mouth ulcers, with the corresponding odds ratio (OR) from the IVW method being 1.160 [95% confidence interval (CI): 1.066–1.261, P = 5.39 × 10^–4], 1.092 (1.062–1.122, P = 3.37 × 10^–10), 1.234 (1.134–1.342, P = 1.03 × 10^–6), and 0.703 (0.571–0.865, P = 8.97 × 10^–4), respectively. We also observed suggestive evidence for mood instability to cause mouth ulcers [IVW, OR = 1.662 (1.059–2.609), P = 0.027]. These results were robust to weak instrument bias and heterogeneity. We found no evidence on causal effects between other psychiatric traits and mouth ulcers, in either direction.

Conclusion

Our findings suggest a protective effect of subjective wellbeing and risk effects of autism spectrum disorder, insomnia, major depressive disorder, and mood instability on mouth ulcers. These results clarify the causal relationship between psychiatric traits and the development of mouth ulcers.

Sex-Specific Genetically Predicted Iron Status in relation to 12 Vascular Diseases: A Mendelian Randomization Study in the UK Biobank

Background

Iron overload has been implicated in the pathogenesis of varicose veins (VVs). However, the association of serum iron status with other vascular diseases (VDs) is not well understood, which might be a potential target for VD prevention. This study was aimed at investigating the causal associations between iron status and VDs using the Mendelian randomization (MR) method.

Methods

A two-sample MR was designed to investigate whether iron status was associated with VDs, based on iron data from a published genome-wide association study meta-analysis of 48,972 subjects of European descent and VD data obtained from the UK Biobank, including 361,194 British subjects (167,020 males and 194,174 females). We further explored whether there was sex difference in the associations between genetically predicted iron status and VDs.

Results

The results demonstrated that iron status had a significant causal effect on VVs of lower extremities (P < 0.001) and a potential effect on coronary atherosclerosis (P < 0.05 for serum iron, ferritin, and transferrin saturation, respectively), but not on other VDs. Furthermore, higher iron status exerted a detrimental effect on VVs of lower extremities in both genders (P < 0.05) and a protective effect on male patients with coronary atherosclerosis (P < 0.05 for serum iron, ferritin, and transferrin saturation, respectively).

Conclusions

This MR study provides robust evidence that higher iron status increases the risk of VVs of lower extremities, whereas it reduces the incidence of coronary atherosclerosis in the male population, which indicates that iron has divergent effects on vascular pathology.

The Boulder Workshop Question Box

The International Statistical Genetics Workshop (commonly referred to as the "Boulder Workshop") has been held annually in Boulder, Colorado almost every year since 1990. A staple feature of each workshop has been the presence of a "question box" (either a physical box or an online virtual one) where workshop participants are given the opportunity of asking questions to the faculty. In this manuscript, we have compiled a list of ten "classic" questions that have appeared in one form or another across multiple workshops and our attempts at answering them.

Investigating causal relationships between Body Mass Index and risk of atopic dermatitis: a Mendelian randomization analysis

Population studies suggest that atopic dermatitis (AD) is associated with an increased risk of obesity, however a causal relationship between these two conditions remains to be established. We therefore use Mendelian randomization (MR) to evaluate whether obesity and AD are causally interlinked. We used summary statistics extracted from genome wide association studies of Body Mass Index (BMI) and AD. MR analysis was performed in both directions to establish the direction of causality between BMI and AD. We find that genetically determined increase in adiposity is associated with increased risk of AD (odds ratio of AD 1.08 [95% CI 1.01 to 1.14; p = 0.015] per unit increase in BMI). Conversely, genetically determined increased risk of AD is not associated with a higher BMI (change in BMI attributable to AD based on genetic information: 0.00; 95% CI − 0.02 to 0.02; p = 0.862). There was no evidence for confounding of these genetic analyses by horizontal pleiotropy. Our results indicate that the association of AD with obesity is likely to reflect a causal role for adiposity in the development of AD. Our findings enhance understanding of the etiology of AD, and the basis for experimental studies to evaluate the mechanistic pathways by which adiposity promotes AD.

A Mendelian randomization study of the causal association between anxiety phenotypes and schizophrenia

Schizophrenia shows a genetic correlation with both anxiety disorder and neuroticism, a trait strongly associated with anxiety. However, genetic correlations do not discern causality from genetic confounding. We therefore aimed to investigate whether anxiety-related phenotypes lie on the causal pathway to schizophrenia using Mendelian randomization (MR). Four MR methods, each with different assumptions regarding instrument validity, were used to investigate casual associations of anxiety and neuroticism related phenotypes on schizophrenia, and vice versa: inverse variance weighted (IVW), weighted median, weighted mode, and, when appropriate, MR Egger regression. MR provided evidence of a causal effect of neuroticism on schizophrenia (IVW odds ratio [OR]: 1.33, 95% confidence interval [CI]: 1.12-1.59), but only weak evidence of a causal effect of anxiety on schizophrenia (IVW OR: 1.10, 95% CI: 1.01-1.19). There was also evidence of a causal association from schizophrenia liability to anxiety disorder (IVW OR: 1.28, 95% CI: 1.18-1.39) and worry (IVW beta: 0.05, 95% CI: 0.03-0.07), but effect estimates from schizophrenia to neuroticism were inconsistent in the main analysis. The evidence of neuroticism increasing schizophrenia risk provided by our results supports future efforts to evaluate neuroticism- or anxiety-based therapies to prevent onset of psychotic disorders.

A comparison of robust Mendelian randomization methods using summary data

The number of Mendelian randomization (MR) analyses including large numbers of genetic variants is rapidly increasing. This is due to the proliferation of genome-wide association studies, and the desire to obtain more precise estimates of causal effects. Since it is unlikely that all genetic variants will be valid instrumental variables, several robust methods have been proposed. We compare nine robust methods for MR based on summary data that can be implemented using standard statistical software. Methods were compared in three ways: by reviewing their theoretical properties, in an extensive simulation study, and in an empirical example. In the simulation study, the best method, judged by mean squared error was the contamination mixture method. This method had well-controlled Type 1 error rates with up to 50% invalid instruments across a range of scenarios. Other methods performed well according to different metrics. Outlier-robust methods had the narrowest confidence intervals in the empirical example. With isolated exceptions, all methods performed badly when over 50% of the variants were invalid instruments. Our recommendation for investigators is to perform a variety of robust methods that operate in different ways and rely on different assumptions for valid inferences to assess the reliability of MR analyses.

Mendelian randomization analysis does not support causal associations of birth weight with hypertension risk and blood pressure in adulthood

Epidemiology studies suggested that low birthweight was associated with a higher risk of hypertension in later life. However, little is known about the causality of such associations. In our study, we evaluated the causal association of low birthweight with adulthood hypertension following a standard analytic protocol using the study-level data of 183,433 participants from 60 studies (CHARGE-BIG consortium), as well as that with blood pressure using publicly available summary-level genome-wide association data from EGG consortium of 153,781 participants, ICBP consortium and UK Biobank cohort together of 757,601 participants. We used seven SNPs as the instrumental variable in the study-level analysis and 47 SNPs in the summary-level analysis. In the study-level analyses, decreased birthweight was associated with a higher risk of hypertension in adults (the odds ratio per 1 standard deviation (SD) lower birthweight, 1.22; 95% CI 1.16 to 1.28), while no association was found between genetically instrumented birthweight and hypertension risk (instrumental odds ratio for causal effect per 1 SD lower birthweight, 0.97; 95% CI 0.68 to 1.41). Such results were consistent with that from the summary-level analyses, where the genetically determined low birthweight was not associated with blood pressure measurements either. One SD lower genetically determined birthweight was not associated with systolic blood pressure (β = - 0.76, 95% CI - 2.45 to 1.08 mmHg), 0.06 mmHg lower diastolic blood pressure (β = - 0.06, 95% CI - 0.93 to 0.87 mmHg), or pulse pressure (β = - 0.65, 95% CI - 1.38 to 0.69 mmHg, all p > 0.05). Our findings suggest that the inverse association of birthweight with hypertension risk from observational studies was not supported by large Mendelian randomization analyses.

Smoking, DNA Methylation, and Lung Function: a Mendelian Randomization Analysis to Investigate Causal Pathways

Whether smoking-associated DNA methylation has a causal effect on lung function has not been thoroughly evaluated. We first investigated the causal effects of 474 smoking-associated CpGs on forced expiratory volume in 1 s (FEV1) in UK Biobank (n = 321,047) by using two-sample Mendelian randomization (MR) and then replicated this investigation in the SpiroMeta Consortium (n = 79,055). Second, we used two-step MR to investigate whether DNA methylation mediates the effect of smoking on FEV1. Lastly, we evaluated the presence of horizontal pleiotropy and assessed whether there is any evidence for shared causal genetic variants between lung function, DNA methylation, and gene expression by using a multiple-trait colocalization ("moloc") framework. We found evidence of a possible causal effect for DNA methylation on FEV1 at 18 CpGs (p < 1.2 × 10-4). Replication analysis supported a causal effect at three CpGs (cg21201401 [LIME1 and ZGPAT], cg19758448 [PGAP3], and cg12616487 [EML3 and AHNAK] [p < 0.0028]). DNA methylation did not clearly mediate the effect of smoking on FEV1, although DNA methylation at some sites might influence lung function via effects on smoking. By using "moloc", we found evidence of shared causal variants between lung function, gene expression, and DNA methylation. These findings highlight potential therapeutic targets for improving lung function and possibly smoking cessation, although larger, tissue-specific datasets are required to confirm these results.

Coffee Consumption and Risk of Stroke: A Mendelian Randomization Study

OBJECTIVE

Observational epidemiological studies have reported a relationship between coffee intake and risk of stroke. However, evidence for this association is inconsistent, and it remains uncertain whether the association is causal or due to confounding or reverse causality. To clarify this relationship, we adopted a Mendelian randomization (MR) approach to evaluate the effects of coffee consumption on the risk of stroke and its subtypes.

METHODS

A meta-analysis of genome-wide association studies (GWASs) including 91,462 coffee consumers was used to identify instruments for coffee consumption. Summary-level data for stroke, intracerebral hemorrhage, ischemic stroke (IS), and IS subtypes were obtained from GWAS meta-analyses conducted by the MEGASTROKE consortium. MR analyses were performed using the inverse-variance-weighted, weighted-median, MR-PRESSO (Pleiotropy RESidual Sum and Outlier) test and MR-Egger regression. Sensitivity analyses were further performed using alternative instruments to test the robustness of our findings.

RESULTS

Genetically predicted coffee consumption (high vs infrequent/no) was not associated with risk of stroke. Similarly, among coffee consumers, MR analysis did not indicate causal associations between coffee consumption (cups/day) and risk of stroke. However, in the subgroup analysis, we found weak suggestive evidence for a potential protective effect of coffee consumption on risk of small vessel (SV)-IS, although the association did not reach statistical significance after correction for multiple comparisons.

INTERPRETATION

This study suggests that coffee consumption is not causally associated with risk of stroke or its subtypes. Further studies are warranted to elucidate the possible association between coffee intake and risk of SV-IS, as well as its potential underlying mechanisms. ANN NEUROL 2020;87:525-532.

A robust and efficient method for Mendelian randomization with hundreds of genetic variants

Mendelian randomization (MR) is an epidemiological technique that uses genetic variants to distinguish correlation from causation in observational data. The reliability of a MR investigation depends on the validity of the genetic variants as instrumental variables (IVs). We develop the contamination mixture method, a method for MR with two modalities. First, it identifies groups of genetic variants with similar causal estimates, which may represent distinct mechanisms by which the risk factor influences the outcome. Second, it performs MR robustly and efficiently in the presence of invalid IVs. Compared to other robust methods, it has the lowest mean squared error across a range of realistic scenarios. The method identifies 11 variants associated with increased high-density lipoprotein-cholesterol, decreased triglyceride levels, and decreased coronary heart disease risk that have the same directions of associations with various blood cell traits, suggesting a shared mechanism linking lipids and coronary heart disease risk mediated via platelet aggregation.

Inflammatory biomarkers and risk of ischemic stroke and subtypes: A 2-sample Mendelian randomization study

Objective: Chronic inflammation is considered as playing an important role in the pathophysiology of atherosclerosis, but the exact contributing inflammatory pathway on stroke is not clear. We aimed to examine the causal association of inflammatory biomarkers, such as interleukin-1 receptor antagonist (IL-1Ra), soluble interleukin-6 receptor (sIL-6R) and C-reactive protein (CRP), with the risk of ischemic stroke and its subtypes.Methods: Two-sample mendelian randomization analyses were performed using IL-1Ra, sIL-6R and CRP related genetic variants as instrumental variables. Summary-level data on ischemic stroke and its subtypes were obtained from the largest GWAS meta-analysis on stroke to date - the Multiancestry Genome-wide Association Study of Stroke (MEGASTROKE) consortium. Associations of IL-1Ra with stroke or its subtypes were estimated using inverse-variance weighted (IVW) method with SNPs rs6743376 and rs1542176 as instruments. Wald ratio method with SNP rs2228145 as the instrument was used for sIL-6R and IVW, MR-Egger, simple and weighted median approaches with 4- or 18-SNPs instruments were used for CRP.Results: Genetically elevated ln(IL-1Ra), ln(sIL-6R) and ln(CRP) levels were not causally associated with ischemic stroke (OR = 1.00, 95% CI: 0.97-1.04, p = 0.80; OR = 0.93, 95% CI: 0.87-0.99, p = 0.03; OR = 1.01, 95% CI: 0.94-1.09, p = 0.78). No significant association was observed between ln(IL-1Ra), ln(sIL-6R) and ln(CRP) level and ischemic stroke subtypes.Conclusions: Our study did not find convincing evidence to support that inflammatory biomarkers like IL-1Ra, sIL-6R and CRP are causally associated with the risk of ischemic stroke or its subtypes.

Genetic Association of Plasma Homocysteine Levels with Gastric Cancer Risk: A Two-Sample Mendelian Randomization Study

BACKGROUND

The association of plasma homocysteine level (PHL) with gastric cancer risk was reported in observational studies. However, the causality is challenging due to confounding factors and the lack of evidence from well-designed cohort studies. Herein, we performed a two-sample Mendelian randomization (MR) analysis to investigate whether PHL is causally related to gastric cancer risk.

METHODS

We performed the MR analysis based on the results from genome-wide association studies consisting of 2,631 patients with gastric cancer and 4,373 controls. An externally weighted genetic risk score (wGRS) was constructed with 15 SNPs with well-established associations with PHL. We utilized logistic regression model to estimate associations of PHL-related SNPs and wGRS with gastric cancer risk in total population and in strata by sex, age, and study site, in addition to a series of sensitivity analyses.

RESULTS

High genetically predicted PHL was associated with an increased gastric cancer risk (per SD increase in the wGRS: OR = 1.07; 95% confidence interval, 1.01-1.12; P = 0.011), which was consistent in sensitivity analyses. Subgroup analyses provided evidence of a stronger association with gastric cancer risk in women than in men. MR-Egger and weighted median regression suggested that potentially unknown pleiotropic effects were not biasing the association between PHL and gastric cancer risk.

CONCLUSIONS

These results revealed that genetically predicted high PHL was associated with an increased gastric cancer risk, suggesting that high PHL may have a causal role in the etiology of gastric cancer.

IMPACT

These findings provide causal inference for PHL on gastric cancer risk, suggesting a causal role of high PHL in the etiology of gastric cancer.

Body Shape and Alzheimer's Disease: A Mendelian Randomization Analysis

Obesity has been reported to be related to memory impairment and decline in cognitive function, possibly further leading to the development of Alzheimer’s disease (AD). However, observational studies revealed both negative and positive associations between body shape (BS) and AD, thereby making it difficult to confirm causality due to residual confounds and reverse causation. Thus, using genome-wide association study summary data, two-sample Mendelian randomization (MR) analyses were applied to identify whether there exists a causal association between BS and AD. BS was measured using anthropometric traits (ATs) in this study, including body mass index (BMI), waist-to-hip ratio (WHR), waist-to-hip ratio adjusted by body mass index (WHRadjBMI), and waist circumference (WC). The associations of single nucleotide polymorphisms (SNP) with each AT and AD were obtained separately from aggregated data from the Genetic Investigation of Anthropometric Traits (GIANT) consortium and International Genomics of Alzheimer’s Project (IGAP) summary data (17,008 cases with AD and 37,154 controls). An inverse-variance weighted method was applied to obtain the overall causal estimate for multiple instrumental SNPs. The odds ratio (OR) [95% confidence interval (CI)] for AD risk per 1-SD difference in BMI was 1.04 (0.88, 1.23), in WHR was 1.01 (0.77, 1.33), in WHRadjBMI was 1.12 (0.89, 1.41), and in WC was 1.02 (0.82, 1.27). Furthermore, simulation analyses of survivor bias indicated the overall causal effect of BMI on risk of AD was not biased. In conclusion, the evidence from MR analyses showed no casual effect of BS on AD risk, which is inconsistent with the results from previous observational studies. The biological mechanism underlying the findings warrants further study.

Appraising the causal relevance of DNA methylation for risk of lung cancer

BACKGROUND

DNA methylation changes in peripheral blood have recently been identified in relation to lung cancer risk. Some of these changes have been suggested to mediate part of the effect of smoking on lung cancer. However, limitations with conventional mediation analyses mean that the causal nature of these methylation changes has yet to be fully elucidated.

METHODS

We first performed a meta-analysis of four epigenome-wide association studies (EWAS) of lung cancer (918 cases, 918 controls). Next, we conducted a two-sample Mendelian randomization analysis, using genetic instruments for methylation at CpG sites identified in the EWAS meta-analysis, and 29 863 cases and 55 586 controls from the TRICL-ILCCO lung cancer consortium, to appraise the possible causal role of methylation at these sites on lung cancer.

RESULTS

Sixteen CpG sites were identified from the EWAS meta-analysis [false discovery rate (FDR) < 0.05], for 14 of which we could identify genetic instruments. Mendelian randomization provided little evidence that DNA methylation in peripheral blood at the 14 CpG sites plays a causal role in lung cancer development (FDR > 0.05), including for cg05575921-AHRR where methylation is strongly associated with both smoke exposure and lung cancer risk.

CONCLUSIONS

The results contrast with previous observational and mediation analysis, which have made strong claims regarding the causal role of DNA methylation. Thus, previous suggestions of a mediating role of methylation at sites identified in peripheral blood, such as cg05575921-AHRR, could be unfounded. However, this study does not preclude the possibility that differential DNA methylation at other sites is causally involved in lung cancer development, especially within lung tissue.

Causal relationships between obesity and the leading causes of death in women and men

Obesity traits are causally implicated with risk of cardiometabolic diseases. It remains unclear whether there are similar causal effects of obesity traits on other non-communicable diseases. Also, it is largely unexplored whether there are any sex-specific differences in the causal effects of obesity traits on cardiometabolic diseases and other leading causes of death. We constructed sex-specific genetic risk scores (GRS) for three obesity traits; body mass index (BMI), waist-hip ratio (WHR), and WHR adjusted for BMI, including 565, 324, and 337 genetic variants, respectively. These GRSs were then used as instrumental variables to assess associations between the obesity traits and leading causes of mortality in the UK Biobank using Mendelian randomization. We also investigated associations with potential mediators, including smoking, glycemic and blood pressure traits. Sex-differences were subsequently assessed by Cochran's Q-test (Phet). A Mendelian randomization analysis of 228,466 women and 195,041 men showed that obesity causes coronary artery disease, stroke (particularly ischemic), chronic obstructive pulmonary disease, lung cancer, type 2 and 1 diabetes mellitus, non-alcoholic fatty liver disease, chronic liver disease, and acute and chronic renal failure. Higher BMI led to higher risk of type 2 diabetes in women than in men (Phet = 1.4×10-5). Waist-hip-ratio led to a higher risk of chronic obstructive pulmonary disease (Phet = 3.7×10-6) and higher risk of chronic renal failure (Phet = 1.0×10-4) in men than women. Obesity traits have an etiological role in the majority of the leading global causes of death. Sex differences exist in the effects of obesity traits on risk of type 2 diabetes, chronic obstructive pulmonary disease, and renal failure, which may have downstream implications for public health.

No Association Between Vitamin D Status and Risk of Barrett's Esophagus or Esophageal Adenocarcinoma: A Mendelian Randomization Study

BACKGROUND & AIMS

Epidemiology studies of circulating concentrations of 25 hydroxy vitamin D (25(OH)D) and risk of esophageal adenocarcinoma (EAC) have produced conflicting results. We conducted a Mendelian randomization study to determine the associations between circulating concentrations of 25(OH)D and risks of EAC and its precursor, Barrett's esophagus (BE).

METHODS

We conducted a Mendelian randomization study using a 2-sample (summary data) approach. Six single-nucleotide polymorphisms (SNPs; rs3755967, rs10741657, rs12785878, rs10745742, rs8018720, and rs17216707) associated with circulating concentrations of 25(OH)D were used as instrumental variables. We collected data from 6167 patients with BE, 4112 patients with EAC, and 17,159 individuals without BE or EAC (controls) participating in the Barrett's and Esophageal Adenocarcinoma Consortium, as well as studies from Bonn, Germany, and Cambridge and Oxford, United Kingdom. Analyses were performed separately for BE and EAC.

RESULTS

Overall, we found no evidence for an association between genetically estimated 25(OH)D concentration and risk of BE or EAC. The odds ratio per 20 nmol/L increase in genetically estimated 25(OH)D concentration for BE risk estimated by combining the individual SNP association using inverse variance weighting was 1.21 (95% CI, 0.77-1.92; P = .41). The odds ratio for EAC risk, estimated by combining the individual SNP association using inverse variance weighting, was 0.68 (95% CI, 0.39-1.19; P = .18).

CONCLUSIONS

In a Mendelian randomization study, we found that low genetically estimated 25(OH)D concentrations were not associated with risk of BE or EAC.

GWAS for urinary sodium and potassium excretion highlights pathways shared with cardiovascular traits

Urinary sodium and potassium excretion are associated with blood pressure (BP) and cardiovascular disease (CVD). The exact biological link between these traits is yet to be elucidated. Here, we identify 50 loci for sodium and 13 for potassium excretion in a large-scale genome-wide association study (GWAS) on urinary sodium and potassium excretion using data from 446,237 individuals of European descent from the UK Biobank study. We extensively interrogate the results using multiple analyses such as Mendelian randomization, functional assessment, co localization, genetic risk score, and pathway analyses. We identify a shared genetic component between urinary sodium and potassium expression and cardiovascular traits. Ingenuity pathway analysis shows that urinary sodium and potassium excretion loci are over-represented in behavioural response to stimuli. Our study highlights pathways that are shared between urinary sodium and potassium excretion and cardiovascular traits.

Effects of Genetically Determined Iron Status on Risk of Venous Thromboembolism and Carotid Atherosclerotic Disease: A Mendelian Randomization Study

Background Systemic iron status has been implicated in atherosclerosis and thrombosis. The aim of this study was to investigate the effect of genetically determined iron status on carotid intima-media thickness, carotid plaque, and venous thromboembolism using Mendelian randomization. Methods and Results Genetic instrumental variables for iron status were selected from a genome-wide meta-analysis of 48 972 subjects. Genetic association estimates for carotid intima-media thickness and carotid plaque were obtained using data from 71 128 and 48 434 participants, respectively, and estimates for venous thromboembolism were obtained using data from a study incorporating 7507 cases and 52 632 controls. Conventional 2-sample summary data Mendelian randomization was performed for the main analysis. Higher genetically determined iron status was associated with increased risk of venous thromboembolism. Odds ratios per SD increase in biomarker levels were 1.37 (95% CI 1.14-1.66) for serum iron, 1.25 (1.09-1.43) for transferrin saturation, 1.92 (1.28-2.88) for ferritin, and 0.76 (0.63-0.92) for serum transferrin (with higher transferrin levels representing lower iron status). In contrast, higher iron status was associated with lower risk of carotid plaque. Corresponding odds ratios were 0.85 (0.73-0.99) for serum iron and 0.89 (0.80-1.00) for transferrin saturation, with concordant trends for serum transferrin and ferritin that did not reach statistical significance. There was no Mendelian randomization evidence of an effect of iron status on carotid intima-media thickness. Conclusions These findings support previous work to suggest that higher genetically determined iron status is protective against some forms of atherosclerotic disease but increases the risk of thrombosis related to stasis of blood.

Associations of mortality with own blood pressure using son's blood pressure as an instrumental variable

High systolic blood pressure (SBP) causes cardiovascular disease (CVD) and is associated with mortality from other causes, but conventional multivariably-adjusted results may be confounded. Here we used a son’s SBP (>1 million Swedish men) as an instrumental variable for parental SBP and examined associations with parents’ cause-specific mortality, avoiding reverse causation. The hazard ratio for CVD mortality per SD (10.80 mmHg) of SBP was 1.49 (95% CI: 1.43, 1.56); SBP was positively associated with coronary heart disease and stroke. SBP was also associated positively with all-cause, diabetes and kidney cancer mortality, and negatively with external causes. Negative associations with respiratory-related mortality were probably confounded by smoking. Hazard ratios for other causes were imprecise or null. Diastolic blood pressure gave similar results to SBP. CVD hazard ratios were intermediate between those from conventional multivariable studies and Mendelian randomization and stronger than those from clinical trials, approximately consistent with an effect of exposure duration on effect sizes. Plots of parental mortality against offspring SBP were approximately linear, supporting calls for lower SBP targets. Results suggest that conventional multivariable analyses of mortality and SBP are not substantially confounded by reverse causation and confirm positive effects of SBP on all-cause, CVD and diabetes mortality.

Improving the accuracy of two-sample summary-data Mendelian randomization: moving beyond the NOME assumption

BACKGROUND

Two-sample summary-data Mendelian randomization (MR) incorporating multiple genetic variants within a meta-analysis framework is a popular technique for assessing causality in epidemiology. If all genetic variants satisfy the instrumental variable (IV) and necessary modelling assumptions, then their individual ratio estimates of causal effect should be homogeneous. Observed heterogeneity signals that one or more of these assumptions could have been violated.

METHODS

Causal estimation and heterogeneity assessment in MR require an approximation for the variance, or equivalently the inverse-variance weight, of each ratio estimate. We show that the most popular 'first-order' weights can lead to an inflation in the chances of detecting heterogeneity when in fact it is not present. Conversely, ostensibly more accurate 'second-order' weights can dramatically increase the chances of failing to detect heterogeneity when it is truly present. We derive modified weights to mitigate both of these adverse effects.

RESULTS

Using Monte Carlo simulations, we show that the modified weights outperform first- and second-order weights in terms of heterogeneity quantification. Modified weights are also shown to remove the phenomenon of regression dilution bias in MR estimates obtained from weak instruments, unlike those obtained using first- and second-order weights. However, with small numbers of weak instruments, this comes at the cost of a reduction in estimate precision and power to detect a causal effect compared with first-order weighting. Moreover, first-order weights always furnish unbiased estimates and preserve the type I error rate under the causal null. We illustrate the utility of the new method using data from a recent two-sample summary-data MR analysis to assess the causal role of systolic blood pressure on coronary heart disease risk.

CONCLUSIONS

We propose the use of modified weights within two-sample summary-data MR studies for accurately quantifying heterogeneity and detecting outliers in the presence of weak instruments. Modified weights also have an important role to play in terms of causal estimation (in tandem with first-order weights) but further research is required to understand their strengths and weaknesses in specific settings.

Using Mendelian randomisation to assess causality in observational studies

OBJECTIVE

Mendelian randomisation (MR) is a technique that aims to assess causal effects of exposures on disease outcomes. The paper aims to present the main assumptions that underlie MR, the statistical methods used to estimate causal effects and how to account for potential violations of the key assumptions.

METHODS

We discuss the key assumptions that should be satisfied in an MR setting. We list the statistical methodologies used in two-sample MR when summary data are available to estimate causal effects (ie, Wald ratio estimator, inverse-variance weighted and maximum likelihood method) and identify/adjust for potential violations of MR assumptions (ie, MR-Egger regression and weighted Median approach). We also present statistical methods and graphical tools used to evaluate the presence of heterogeneity.

RESULTS

We use as an illustrative example of a published two-sample MR study, investigating the causal association of body mass index with three psychiatric disorders (ie, bipolar disorder, schizophrenia and major depressive disorder). We highlight the importance of assessing the results of all available methods rather than each method alone. We also demonstrate the impact of heterogeneity in the estimation of the causal effects.

CONCLUSIONS

MR is a useful tool to assess causality of risk factors in medical research. Assessment of the key assumptions underlying MR is crucial for a valid interpretation of the results.

Height and Body Mass Index as Modifiers of Breast Cancer Risk in BRCA1/2 Mutation Carriers: A Mendelian Randomization Study

BACKGROUND

BRCA1/2 mutations confer high lifetime risk of breast cancer, although other factors may modify this risk. Whether height or body mass index (BMI) modifies breast cancer risk in BRCA1/2 mutation carriers remains unclear.

METHODS

We used Mendelian randomization approaches to evaluate the association of height and BMI on breast cancer risk, using data from the Consortium of Investigators of Modifiers of BRCA1/2 with 14 676 BRCA1 and 7912 BRCA2 mutation carriers, including 11 451 cases of breast cancer. We created a height genetic score using 586 height-associated variants and a BMI genetic score using 93 BMI-associated variants. We examined both observed and genetically determined height and BMI with breast cancer risk using weighted Cox models. All statistical tests were two-sided.

RESULTS

Observed height was positively associated with breast cancer risk (HR = 1.09 per 10 cm increase, 95% confidence interval [CI] = 1.0 to 1.17; P = 1.17). Height genetic score was positively associated with breast cancer, although this was not statistically significant (per 10 cm increase in genetically predicted height, HR = 1.04, 95% CI = 0.93 to 1.17; P = .47). Observed BMI was inversely associated with breast cancer risk (per 5 kg/m2 increase, HR = 0.94, 95% CI = 0.90 to 0.98; P = .007). BMI genetic score was also inversely associated with breast cancer risk (per 5 kg/m2 increase in genetically predicted BMI, HR = 0.87, 95% CI = 0.76 to 0.98; P = .02). BMI was primarily associated with premenopausal breast cancer.

CONCLUSION

Height is associated with overall breast cancer and BMI is associated with premenopausal breast cancer in BRCA1/2 mutation carriers. Incorporating height and BMI, particularly genetic score, into risk assessment may improve cancer management.

Using Openly Accessible Resources to Strengthen Causal Inference in Epigenetic Epidemiology of Neurodevelopment and Mental Health

The recent focus on the role of epigenetic mechanisms in mental health has led to several studies examining the association of epigenetic processes with psychiatric conditions and neurodevelopmental traits. Some studies suggest that epigenetic changes might be causal in the development of the psychiatric condition under investigation. However, other scenarios are possible, e.g., statistical confounding or reverse causation, making it particularly challenging to derive conclusions on causality. In the present review, we examine the evidence from human population studies for a possible role of epigenetic mechanisms in neurodevelopment and mental health and discuss methodological approaches on how to strengthen causal inference, including the need for replication, (quasi-)experimental approaches and Mendelian randomization. We signpost openly accessible resources (e.g., "MR-Base" "EWAS catalog" as well as tissue-specific methylation and gene expression databases) to aid the application of these approaches.

Genetic association and transcriptome integration identify contributing genes and tissues at cystic fibrosis modifier loci

Cystic Fibrosis (CF) exhibits morbidity in several organs, including progressive lung disease in all patients and intestinal obstruction at birth (meconium ileus) in ~15%. Individuals with the same causal CFTR mutations show variable disease presentation which is partly attributed to modifier genes. With >6,500 participants from the International CF Gene Modifier Consortium, genome-wide association investigation identified a new modifier locus for meconium ileus encompassing ATP12A on chromosome 13 (min p = 3.83x10(-10)); replicated loci encompassing SLC6A14 on chromosome X and SLC26A9 on chromosome 1, (min p<2.2x10(-16), 2.81x10(-11), respectively); and replicated a suggestive locus on chromosome 7 near PRSS1 (min p = 2.55x10(-7)). PRSS1 is exclusively expressed in the exocrine pancreas and was previously associated with non-CF pancreatitis with functional characterization demonstrating impact on PRSS1 gene expression. We thus asked whether the other meconium ileus modifier loci impact gene expression and in which organ. We developed and applied a colocalization framework called the Simple Sum (SS) that integrates regulatory and genetic association information, and also contrasts colocalization evidence across tissues or genes. The associated modifier loci colocalized with expression quantitative trait loci (eQTLs) for ATP12A (p = 3.35x10(-8)), SLC6A14 (p = 1.12x10(-10)) and SLC26A9 (p = 4.48x10(-5)) in the pancreas, even though meconium ileus manifests in the intestine. The meconium ileus susceptibility locus on chromosome X appeared shifted in location from a previously identified locus for CF lung disease severity. Using the SS we integrated the lung disease association locus with eQTLs from nasal epithelia of 63 CF participants and demonstrated evidence of colocalization with airway-specific regulation of SLC6A14 (p = 2.3x10(-4)). Cystic Fibrosis is realizing the promise of personalized medicine, and identification of the contributing organ and understanding of tissue specificity for a gene modifier is essential for the next phase of personalizing therapeutic strategies.

Circulating antioxidants and Alzheimer disease prevention: a Mendelian randomization study

Background

Higher circulating antioxidant concentrations are associated with a lower risk of late-onset Alzheimer disease (AD) in observational studies, suggesting that diet-sourced antioxidants may be modifiable targets for reducing disease risk. However, observational evidence is prone to substantial biases that limit causal inference, including residual confounding and reverse causation.

Objectives

In order to infer whether long-term circulating antioxidant exposure plays a role in AD etiology, we tested the hypothesis that AD risk would be lower in individuals with lifelong, genetically predicted increases in concentrations of 4 circulating antioxidants that are modifiable by diet.

Methods

Two-sample Mendelian randomization analyses were conducted. First, published genetic association studies were used to identify single-nucleotide polymorphisms (SNPs) that determine variation in circulating ascorbate (vitamin C), β-carotene, retinol (vitamin A), and urate. Second, for each set of SNP data, statistics for genotype associations with AD risk were extracted from data of a genome-wide association study of late-onset AD cases and controls (n = 17,008 and 37,154, respectively). Ratio-of-coefficients and inverse-variance-weighted meta-analyses were the primary methods used to assess the 4 sets of SNP-exposure and SNP-AD associations. Additional analyses assessed the potential impact of bias from pleiotropy on estimates.

Results

The models suggested that genetically determined differences in circulating ascorbate, retinol, and urate are not associated with differences in AD risk. All estimates were close to the null, with all ORs for AD ≥1 per unit increase in antioxidant exposure (ranging from 1.00 for ascorbate to 1.05 for retinol). There was little evidence to imply that pleiotropy had biased results.

Conclusions

Our findings suggest that higher exposure to ascorbate, β-carotene, retinol, or urate does not lower the risk of AD. Replication Mendelian randomization studies could assess this further, providing larger AD case-control samples and, ideally, using additional variants to instrument each exposure.

Negative effect of vitamin D on kidney function: a Mendelian randomization study

Background

The kidney plays a central role in the regulation of vitamin D metabolism. It is not clear, however, whether vitamin D influences kidney function. Previous studies have reported conflicting results, which may have been influenced by reverse causation and residual confounding. We conducted a Mendelian randomization (MR) study to obtain unconfounded estimates of the association between genetically instrumented vitamin D metabolites and estimated glomerular filtration rate (eGFR) as well as the urinary albumin:creatinine ratio (UACR).

Methods

We performed a two-sample MR study based on three single nucleotide variants associated with 25(OH)D levels: rs2282679, rs10741657 and rs12785878, related to the genes GC, CYP2R1 and DHCR7, respectively. Estimates of the allele-dependent effects on serum 25(OH)D and eGFR/UACR were obtained from summary statistics of published genome-wide association meta-analyses. Additionally, we performed a one-sample MR analysis for both 25(OH)D and 1,25(OH)2 D using individual-level data from six cohorts.

Results

The combined MR estimate supported a negative causal effect of log transformed 25(OH)D on log transformed eGFR (β = -0.013, P = 0.003). The analysis of individual-level data confirmed the main findings and also revealed a significant association of 1,25(OH)2 D on eGFR (β = -0.094, P = 0.008). These results show that a 10% increase in serum 25(OH)D levels causes a 0.3% decrease in eGFR. There was no effect of 25(OH)D on UACR (β = 0.032, P = 0.265).

Conclusion

Our study suggests that circulating vitamin D metabolite levels are negatively associated with eGFR. Further studies are needed to elucidate the underlying mechanisms.

BMI and Mortality in UK Biobank: Revised Estimates Using Mendelian Randomization

OBJECTIVE

The aim of this study was to obtain estimates of the causal relationship between BMI and mortality.

METHODS

Mendelian randomization (MR) with BMI-associated genotypic variation was used to test the causal effect of BMI on all-cause and cause-specific mortality in UK Biobank participants of White British ancestry.

RESULTS

MR analyses supported a causal association between higher BMI and greater risk of all-cause mortality (hazard ratio [HR] per 1 kg/m2 : 1.03; 95% CI: 0.99-1.07) and mortality from cardiovascular diseases (HR: 1.10; 95% CI: 1.01-1.19), specifically coronary heart disease (HR: 1.12; 95% CI: 1.00-1.25) and those excluding coronary heart disease/stroke/aortic aneurysm (HR: 1.24; 95% CI: 1.03-1.48), stomach cancer (HR: 1.18; 95% CI: 0.87-1.62), and esophageal cancer (HR: 1.22; 95% CI: 0.98-1.53), and a decreased risk of lung cancer mortality (HR: 0.96; 95% CI: 0.85-1.08). Sex stratification supported the causal role of higher BMI increasing bladder cancer mortality risk (males) but decreasing respiratory disease mortality risk (males). The J-shaped observational association between BMI and mortality was visible with MR analyses, but the BMI at which mortality was minimized was lower and the association was flatter over a larger BMI range.

CONCLUSIONS

Results support a causal role of higher BMI in increasing the risk of all-cause mortality and mortality from several specific causes.

Genetically high plasma vitamin C and urate: a Mendelian randomization study in 106 147 individuals from the general population

Objective

Gout is the most common form of inflammatory arthritis and is caused by hyperuricaemia. Some studies have found a reduction in plasma urate with vitamin C supplementation. We tested the hypothesis that high plasma vitamin C is causally associated with low plasma urate and low risk of hyperuricaemia, using a Mendelian randomization approach.

Methods

We measured plasma urate and genotyped for the SLC23A1 rs33972313 vitamin C variant in 106 147 individuals from the Copenhagen General Population Study, of which 24 099 had hyperuricaemia. We measured plasma vitamin C in 9234 individuals and genotyped for the SLC2A9 rs7442295 urate variant in 102 345 individuals.

Results

Each 10 µmol/l higher plasma vitamin C was associated with a -2.3(95%CI: -0.69 to -3.9) µmol/l lower plasma urate after multivariable adjustments. The SLC23A1 rs33972313 GG genotype was associated with a 9% (5.6%, 11.9%) higher plasma vitamin C compared with AA and AG combined but was not associated with plasma urate (P = 0.31). Likewise, for each 10 µmol/l higher plasma vitamin C the odds ratios for hyperuricaemia were 0.92 (0.86, 0.98) observationally after multivariable adjustments, but 1.01 (0.84, 1.23) genetically.

Conclusion

High plasma vitamin C was associated with low plasma urate and with low risk of hyperuricaemia. However, the SLC23A1 genetic variant causing lifelong high plasma vitamin C was not associated with plasma urate levels or with risk of hyperuricaemia. Thus, our data do not support a causal relationship between high plasma vitamin C and low plasma urate.

Vitamin D and Inflammatory Bowel Disease: Mendelian Randomization Analyses in the Copenhagen Studies and UK Biobank

CONTEXT

Vitamin D may be a modifiable risk factor for inflammatory bowel disease (IBD).

OBJECTIVES

We tested the hypothesis that plasma 25-hydroxyvitamin D levels are causally associated with risk of Crohn disease (CD) and ulcerative colitis (UC).

DESIGN, SETTING, PATIENTS, AND INTERVENTIONS

We used a Mendelian randomization design to study 120,013 individuals from the Copenhagen City Heart Study, the Copenhagen General Population Study, and a Copenhagen-based cohort of patients with IBD. Of these, 35,558 individuals had plasma 25-hydroxyvitamin D measurements available, and 115,110 were genotyped for rs7944926 and rs11234027 in DHCR7 and rs10741657 and rs12794714 in CYP2R1, all variants associated with plasma 25-hydroxyvitamin D levels. We identified 653 cases of CD and 1265 cases of UC, of which 58 and 113, respectively, had 25-hydroxyvitamin D measurements available. We also included genetic data from 408,455 individuals from the UK Biobank, including 1707 CD cases and 3147 UC cases.

MAIN OUTCOME MEASURE

Hazard ratios for higher plasma 25-hydroxyvitamin D levels.

RESULTS

The multivariable-adjusted hazard ratios for 10 nmol/L higher 25-hydroxyvitamin D level were 1.04 (95% CI: 0.93 to 1.16) for CD and 1.13 (95% CI: 1.06 to 1.21) for UC. A combined 25-hydroxyvitamin D allele score was associated with a 1.4-nmol/L increase in plasma 25-hydroxyvitamin D level and hazard ratios of 0.98 (95% CI: 0.94 to 1.03) for CD and 1.01 (95% CI: 0.97 to 1.05) for UC. Combining genetic data from the Copenhagen studies and the UK Biobank, genetically determined vitamin D did not appear to influence the risk of CD or UC.

CONCLUSIONS

Our results do not support a major role for vitamin D deficiency in the development of IBD.