Mendelian randomization: Using genes as instruments for making causal inferences in epidemiology

Vitamin D status, filaggrin genotype, and cardiovascular risk factors: a Mendelian randomization approach

Background

Vitamin D deficiency is associated with increased cardiovascular disease risk in observational studies. Whether these associations are causal is not clear. Loss-of-function mutations in the filaggrin gene result in up to 10% higher serum vitamin D concentrations, supposedly due to a decreased UV-protection of the keratinocytes. We used a Mendelian randomization approach to estimate the causal effect of vitamin D status on serum lipids, blood pressure, body mass index, waist circumference, and the metabolic syndrome.

Methods

Three population based studies were included, Monica10 (2,656 individuals aged 40–71 years), Inter99 (6,784 individuals aged 30–60 years), and Health2006 (3,471 individuals aged 18–69 years) conducted in 1993–94, 1999–2001, and 2006–2008, respectively. Participants were genotyped for the two most common filaggrin gene mutations in European descendants R501X and 2282del4, in all three studies and further for the R2447X mutation in the Inter99 and Health2006 studies. Filaggrin genotype was used as instrumental variable for vitamin D status. Baseline measurements of serum 25-hydroxyvitamin D were performed in all three studies.

Results

Instrumental variable analyses showed a 23.8% (95% confidence interval, CI 3.0, 48.6) higher HDL cholesterol level and a 30.5% (95% CI: 0.8, 51.3) lower serum level of triglycerides per doubling of vitamin D. These associations were, however, not statistically significant when applying the Bonferroni adjusted significance level. The remaining lipids showed non-significant changes in a favorable direction. Doubling of vitamin D gave a non-significantly lower odds ratio = 0.26 (95% CI: 0.06, 1.17) of the metabolic syndrome. There were no statistically significant causal effects of vitamin D status on blood pressure, body mass index, or waist circumference.

Conclusion

Our results support a causal effect of higher vitamin D status on a more favorable lipid profile, although more studies in other populations are needed to confirm our results.

Short communication: genetic variations of SLC2A9 in relation to Parkinson's disease

Background

Epidemiological studies showed that higher plasma urate was associated with lower risk for Parkinson’s disease (PD) and slower disease progression. Recent genome-wide association studies (GWAS) consistently showed that several single nucleotide polymorphisms (SNPs) in the solute carrier family 2 member 9 gene (SLC2A9 ) were associated with plasma urate concentration and the risk of gout.

Methods

We conducted a case–control study to examine twelve tag SNPs of the SLC2A9 gene in relation to PD among 788 cases and 911 controls of European ancestry. Odds ratios (OR) and 95% confidence intervals (CI) were derived from logistic regression models, adjusting for age, sex, smoking and caffeine consumption.

Results

These SNPs were all in linkage disequilibrium (R² > 0.7). None of them were associated with PD risk. Among women, however, there was a suggestion that the presence of the minor allele of one SNP (rs7442295) was related to a small increase in PD risk [OR (95% CI) = 1.48 (1.01-2.16)].

Conclusion

This study provides little support for genetic variations of SLC2A9 and PD risk.

Association of plasminogen activator inhibitor-type 1 (-675 4G/5G) polymorphism with pre-eclampsia: systematic review

BACKGROUND AND AIMS

Excessive generation of plasminogen activator inhibitor-type 1 (PAI-1) is implicated in the pathogenesis of pre-eclampsia and related conditions. The PAI-1 (-675 4G/5G) promoter polymorphism (rs1799889) affects transcriptional activity and is a putative genetic risk factor for pre-eclampsia. The aim of this study was identify, appraise and synthesise the available evidence for the association of the PAI-1 (-675 4G/5G) polymorphism with pre-eclampsia.

METHODS

Systematic review and random effects meta-analysis of genetic association studies.

RESULTS

We found 12 eligible genetic association studies in which a total of 1511 women with pre-eclampsia, eclampsia or HELLP syndrome and 3492 controls participated. The studies were generally small (median number of cases 102, range 24 to 403) and underpowered to detect plausible association sizes. Meta-analysis of all of the studies detected statistically significant gene-disease associations in the recessive [pooled odds ratio 1.28 (95% confidence interval 1.09, 1.50); population attributable risk 7.7%] and dominant [pooled odds ratio 1.21 (95% confidence interval 1.01, 1.44); population attributable risk 13.7%] models. We did not find evidence of statistical heterogeneity, funnel plot asymmetry or small study bias.

CONCLUSIONS

These data suggest that the fibrinolytic pathway regulated by the PAI-1 gene may contribute to the pathogenesis of pre-eclampsia and related conditions. This association, if confirmed in larger genetic association studies, may inform research efforts to develop novel interventions or help to prioritise therapeutic targets that merit evaluation in randomised clinical trials.

Genetic regulation of vitamin D levels

Vitamin D plays several roles in the body, influencing bone health as well as serum calcium and phosphate levels. Further, vitamin D may modify immune function, cell proliferation, differentiation, and apoptosis. Vitamin D deficiency has been associated with numerous health outcomes, including bone disease, cancer, autoimmune disease, infectious disease, type 1 and type 2 diabetes, hypertension, and heart disease, although it is unclear whether or not these associations are causal. Various twin and family studies have demonstrated moderate to high heritability for circulating vitamin D levels. Accordingly, many studies have investigated the genetic determinants of this hormone. Recent advances in the methodology of large-scale genetic association studies, including coordinated international collaboration, have identified associations of CG, DHCR1, CYP2R1, VDR, and CYP24A1 with serum levels of vitamin D. Here, we review the genetic determinants of vitamin D levels by focusing on new findings arising from candidate gene and genomewide association studies.

Cohort Profile: the Avon Longitudinal Study of Parents and Children: ALSPAC mothers cohort

Summary The Avon Longitudinal Study of Children and Parents (ALSPAC) was established to understand how genetic and environmental characteristics influence health and development in parents and children. All pregnant women resident in a defined area in the South West of England, with an expected date of delivery between 1st April 1991 and 31st December 1992, were eligible and 13761 women (contributing 13867 pregnancies) were recruited. These women have been followed over the last 19-22 years and have completed up to 20 questionnaires, have had detailed data abstracted from their medical records and have information on any cancer diagnoses and deaths through record linkage. A follow-up assessment was completed 17-18 years postnatal at which anthropometry, blood pressure, fat, lean and bone mass and carotid intima media thickness were assessed, and a fasting blood sample taken. The second follow-up clinic, which additionally measures cognitive function, physical capability, physical activity (with accelerometer) and wrist bone architecture, is underway and two further assessments with similar measurements will take place over the next 5 years. There is a detailed biobank that includes DNA, with genome-wide data available on >10000, stored serum and plasma taken repeatedly since pregnancy and other samples; a wide range of data on completed biospecimen assays are available. Details of how to access these data are provided in this cohort profile.

Causal relationship between obesity and vitamin D status: bi-directional Mendelian randomization analysis of multiple cohorts

BACKGROUND

Obesity is associated with vitamin D deficiency, and both are areas of active public health concern. We explored the causality and direction of the relationship between body mass index (BMI) and 25-hydroxyvitamin D [25(OH)D] using genetic markers as instrumental variables (IVs) in bi-directional Mendelian randomization (MR) analysis.

METHODS AND FINDINGS

We used information from 21 adult cohorts (up to 42,024 participants) with 12 BMI-related SNPs (combined in an allelic score) to produce an instrument for BMI and four SNPs associated with 25(OH)D (combined in two allelic scores, separately for genes encoding its synthesis or metabolism) as an instrument for vitamin D. Regression estimates for the IVs (allele scores) were generated within-study and pooled by meta-analysis to generate summary effects. Associations between vitamin D scores and BMI were confirmed in the Genetic Investigation of Anthropometric Traits (GIANT) consortium (n = 123,864). Each 1 kg/m(2) higher BMI was associated with 1.15% lower 25(OH)D (p = 6.52×10⁻²⁷). The BMI allele score was associated both with BMI (p = 6.30×10⁻⁶²) and 25(OH)D (-0.06% [95% CI -0.10 to -0.02], p = 0.004) in the cohorts that underwent meta-analysis. The two vitamin D allele scores were strongly associated with 25(OH)D (p≤8.07×10⁻⁵⁷ for both scores) but not with BMI (synthesis score, p = 0.88; metabolism score, p = 0.08) in the meta-analysis. A 10% higher genetically instrumented BMI was associated with 4.2% lower 25(OH)D concentrations (IV ratio: -4.2 [95% CI -7.1 to -1.3], p = 0.005). No association was seen for genetically instrumented 25(OH)D with BMI, a finding that was confirmed using data from the GIANT consortium (p≥0.57 for both vitamin D scores).

CONCLUSIONS

On the basis of a bi-directional genetic approach that limits confounding, our study suggests that a higher BMI leads to lower 25(OH)D, while any effects of lower 25(OH)D increasing BMI are likely to be small. Population level interventions to reduce BMI are expected to decrease the prevalence of vitamin D deficiency.

Low-density lipoprotein cholesterol and risk of gallstone disease: a Mendelian randomization study and meta-analyses

BACKGROUND & AIMS

Drugs which reduce plasma low-density lipoprotein cholesterol (LDL-C) may protect against gallstone disease. Whether plasma levels of LDL-C per se predict risk of gallstone disease remains unclear. We tested the hypothesis that elevated LDL-C is a causal risk factor for symptomatic gallstone disease.

METHODS

We used a Mendelian randomization approach and genotyped 63,051 individuals from a prospective cohort study of the general Danish population, including 3323 subjects with symptomatic gallstones. We selected eight genetic variants in APOE, APOB, LDLR, and PCSK9 affecting LDL-C. Furthermore, studies of APOE rs429358/rs7412 (defining ε2/ε3/ε4 alleles; 12 studies) and APOB rs693 (eight studies) were included in meta-analyses.

RESULTS

The observational hazard ratio (HR) for symptomatic gallstone disease for the fifth versus first quintile of LDL-C was 0.94 (95% confidence interval: 0.76-1.17), despite a corresponding 134% increase in LDL-C. Furthermore, although individual genetic variants in APOE, APOB, LDLR, and PCSK9 associated with stepwise increases/decreases in LDL-C of up to +59% compared with non-carriers (p <0.001), none predicted the risk of symptomatic gallstone disease. Combining all variants into 10 genotypes, carriers of 9 versus ⩽3 LDL-C increasing alleles associated with 41% increased LDL-C (p <0.001), but predicted a HR for symptomatic gallstone disease of 1.09 (0.70-1.69). Finally, in meta-analyses, random effects odds ratios for gallstone disease were 0.91 (0.78-1.06) for carriers of APOE ε4 versus non-carriers, and 1.25 (0.95-1.63) for APOB rs693 CT+TT versus CC.

CONCLUSIONS

Results from the observational study, genetic studies, and meta-analyses suggest that elevated plasma levels of LDL-C are not causally associated with increased risk of symptomatic gallstone disease.

Child height, health and human capital: Evidence using genetic markers

Height has long been recognized as being associated with better outcomes: the question is whether this association is causal. We use children's genetic variants as instrumental variables to deal with possible unobserved confounders and examine the effect of child/adolescent height on a wide range of outcomes: academic performance, IQ, self-esteem, depression symptoms and behavioral problems. OLS findings show that taller children have higher IQ, perform better in school, and are less likely to have behavioral problems. The IV results differ: taller girls (but not boys) have better cognitive performance and, in contrast to the OLS, greater height appears to increase behavioral problems.

Genetic instrumental variable studies of effects of prenatal risk factors

Identifying the effects of maternal risk factors during pregnancy on infant and child health is an area of tremendous research interest. However, policymakers are primarily interested in unraveling the causal effects of prenatal risk factors, not their associations with child health, which may be confounded by several unobserved factors. In this article, we evaluate the utility of genetic variants in three genes that have unequivocal evidence of being related to three major risk factors-CHRNA3 for smoking, ADH1B for alcohol use, and FTO for obesity-as instrumental variables for identifying the causal effects of such factors during pregnancy. Using two independent datasets, we find that these variants are overall predictive of the risk factors and are not systematically related to observed confounders, suggesting that they may be useful instruments. We also find some suggestive evidence that genetic effects are stronger during than before pregnancy. We provide an empirical example illustrating the use of these genetic variants as instruments to evaluate the effects of risk factors on birth weight. Finally, we offer suggestions for researchers contemplating the use of these variants as instruments.

Genetically elevated non-fasting triglycerides and calculated remnant cholesterol as causal risk factors for myocardial infarction

AIMS

Elevated non-fasting triglycerides mark elevated levels of remnant cholesterol. Using a Mendelian randomization approach, we tested whether genetically increased remnant cholesterol in hypertriglyceridaemia due to genetic variation in the apolipoprotein A5 gene (APOA5) associates with an increased risk of myocardial infarction (MI).

METHODS AND RESULTS

We resequenced the core promoter and coding regions of APOA5 in individuals with the lowest 1% (n = 95) and highest 2% (n = 190) triglyceride levels in the Copenhagen City Heart Study (CCHS, n = 10 391). Genetic variants which differed in frequency between the two extreme triglyceride groups (c.-1131T > C, S19W, and c.*31C > T; P-value: 0.06 to <0.001), thus suggesting an effect on triglyceride levels, were genotyped in the Copenhagen General Population Study (CGPS), the CCHS, and the Copenhagen Ischemic Heart Disease Study (CIHDS), comprising a total of 5705 MI cases and 54 408 controls. Genotype combinations of these common variants associated with increases in non-fasting triglycerides and calculated remnant cholesterol of, respectively, up to 68% (1.10 mmol/L) and 56% (0.40 mmol/L) (P < 0.001), and with a corresponding odds ratio for MI of 1.87 (95% confidence interval: 1.25-2.81). Using APOA5 genotypes in instrumental variable analysis, the observational hazard ratio for a doubling in non-fasting triglycerides was 1.57 (1.32-2.68) compared with a causal genetic odds ratio of 1.94 (1.40-1.85) (P for comparison = 0.28). For calculated remnant cholesterol, the corresponding values were 1.67(1.38-2.02) observational and 2.23(1.48-3.35) causal (P for comparison = 0.21).

CONCLUSION

These data are consistent with a causal association between elevated levels of remnant cholesterol in hypertriglyceridaemia and an increased risk of MI. Limitations include that remnants were not measured directly, and that APOA5 genetic variants may influence other lipoprotein parameters.

Is aldehyde dehydrogenase 2 a credible genetic instrument for alcohol use in Mendelian randomization analysis in Southern Chinese men?

BACKGROUND

Mendelian randomization studies provide a means of assessing causal relations without interventions, but require valid genetic instruments. We assessed the credibility of aldehyde dehydrogenase 2 (ALDH2) as a genetic instrument for alcohol use in Southern Chinese men.

METHODS

We genotyped the single nucleotide polymorphism rs671 of ALDH2 in 4867 men from the Guangzhou Biobank Cohort Study. We used linear regression to assess the strength of the association of ALDH2 variants with alcohol use, whether ALDH2 variants were independently associated with socio-economic position or other potential confounders and whether associations of ALDH2 variants with cardiovascular risk factors (systolic and diastolic blood pressure, HDL- and LDL-cholesterol, fasting glucose), triglycerides, body mass index, self reported cardiovascular disease, self-reported ischaemic heart disease, cognitive function (delayed 10-word recall and Mini Mental State Examination score) and liver function (alanine transaminase and aspartate transaminase) were fully mediated by alcohol use.

RESULTS

The minor allele frequency (A) of ALDH2 was 0.29. The F statistic for ALDH2 variants was 75.0, suggesting that substantial weak instrument bias is unlikely. ALDH2 variants were not associated with socio-economic position, smoking or physical activity. ALDH2 variants were only associated with diastolic blood pressure and HDL-cholesterol, but these genetic associations with blood pressure and HDL-cholesterol were attenuated after adjusting for alcohol use, suggesting the apparent genetic associations were possibly mediated by alcohol use.

CONCLUSIONS

ALDH2 variants are a credible genetic instrument for Mendelian randomization studies of alcohol use and many attributes of health in Southern Chinese men.

Genetic impact on uric acid concentration and hyperuricemia in the Japanese population

AIM

Using general Japanese populations, we performed a replication study of genetic loci previously identified in European-descent populations as being associated with uric acid and gout. The relative contribution of non-genetic and genetic factors to the variances in serum uric acid concentration was then evaluated.

METHODS

Seven single nucleotide polymorphisms (SNPs) were genotyped from 7 candidate loci robustly confirmed in Europeans. Genotyping was performed in up to 17,226 individuals, from which 237 hyperuricemia cases and 3,218 controls were chosen for a case-control study. For 6 SNPs showing a replication of uric acid association in 17,076 general population samples, we further tested the associations with other metabolic traits (n≤5,745) and with type 2 diabetes (931 cases and 1404 controls) and coronary artery disease (806 cases and 1337 controls).

RESULTS

Significant uric acid associations (one-tailed p<0.05) were replicated for 6 loci in Japanese. The strongest association was detected at SLC22A12 rs505802 for uric acid (p=2.4×10(-50)) and ABCG2 rs2231142 for hyperuricemia (p3.6×10(-10)). The combined genetic effect could explain some proportion of inter-individual variation in uric acid (R(2)=0.03) and was more or less comparable to the effect of well-recognized risk factors -BMI (R(2)=0.04) and alcohol intake (R(2)=0.01). The tested SNPs were not significantly associated with cardiovascular risk traits except for GCKR rs780094.

CONCLUSION

Our results confirm that 6 common uric acid variant loci are reproducible in Japanese. Further investigation is warranted to efficiently use the knowledge about genetic factors in combination with modifiable risk factors when we decide an individual's treatment strategy for hyperuricemia.

Vitamin B-12 status during pregnancy and child's IQ at age 8: a Mendelian randomization study in the Avon longitudinal study of parents and children

Vitamin B-12 is essential for the development and maintenance of a healthy nervous system. Brain development occurs primarily in utero and early infancy, but the role of maternal vitamin B-12 status during pregnancy on offspring cognitive function is unclear. In this study we assessed the effect of vitamin B-12 status in well-nourished pregnant women on the cognitive ability of their offspring in a UK birth cohort (ALSPAC). We then examined the association of SNPs in maternal genes FUT2 (rs492602) and TCN2 (rs1801198, rs9606756) that are related to plasma vitamin B-12, with offspring IQ. Observationally, there was a positive association between maternal vitamin B-12 intake and child’s IQ that was markedly attenuated after adjustment for potential confounders (mean difference in offspring IQ score per doubling of maternal B-12 intake, before adjustment: 2.0 (95% CI 1.3, 2.8); after adjustment: 0.7 (95% CI −0.04, 1.4)). Maternal FUT2 was weakly associated with offspring IQ: mean difference in IQ per allele was 0.9 (95% CI 0.1, 1.6). The expected effect of maternal vitamin B-12 on offspring IQ, given the relationships between SNPs and vitamin B-12, and SNPs and IQ was consistent with the observational result. Our findings suggest that maternal vitamin B-12 may not have an important effect on offspring cognitive ability. However, further examination of this issue is warranted.

Mendelian randomization: how genetics is pushing the boundaries of epidemiology to identify new causes of heart disease

The past 10 years have seen a remarkable revolution in the genetics of cardiovascular (CV) disease. Although much work remains to bring these discoveries to the bedside, genetics has opened up remarkable possibilities in understanding the causes of CV disease through a relatively novel study design known as "Mendelian randomization." Akin to a randomized trial, Mendelian randomization is a genetic study design that takes advantage of the "randomization" of genetic information at birth to evaluate a potential causal relationship between a genetically determined biomarker and an outcome. By providing evidence for causal relationships, Mendelian randomization can improve our understanding of fundamental mechanisms in human disease, potentially accelerate the identification of bona fide drug targets, and ultimately improve the care of patients with CV disease. This review describes the concept and design of Mendelian randomization genetic studies, discusses their strengths and weaknesses, and presents recent examples of Mendelian randomization studies in the CV literature that have helped clarify the causal role of selected biomarkers in CV medicine.

Interpreting noncoding genetic variation in complex traits and human disease

Association studies provide genome-wide information about the genetic basis of complex disease, but medical research has primarily focused on protein-coding variants, due to the difficulty of interpreting non-coding mutations. This picture has changed with advances in the systematic annotation of functional non-coding elements. Evolutionary conservation, functional genomics, chromatin state, sequence motifs, and molecular quantitative trait loci all provide complementary information about non-coding function. These functional maps can help prioritize variants on risk haplotypes, filter mutations encountered in the clinic, and perform systems-level analyses to reveal processes underlying disease associations. Advances in predictive modeling can enable dataset integration to reveal pathways shared across loci and alleles, and richer regulatory models can guide the search for epistatic interactions. Lastly, new massively parallel reporter experiments can systematically validate regulatory predictions. Ultimately, advances in regulatory and systems genomics can help unleash the value of whole-genome sequencing for personalized genomic risk assessment, diagnosis, and treatment.

Maternal adiposity--a determinant of perinatal and offspring outcomes?

Experimental and animal data suggest that maternal obesity during pregnancy adversely affects offspring health in the short-term and the long-term. Whether these effects occur in humans and influence population health is less clear. This Review explores evidence from intervention studies and observational studies that have used designs (such as family-based comparisons and Mendelian randomization) that might help improve understanding of the causal effects of maternal obesity in humans. Collectively, human studies provide evidence that maternal overweight and obesity is causally related to pregnancy complications, increased offspring weight and adiposity at birth, and the difficulties associated with delivery of large-for-gestational-age infants. The underlying mechanisms for these effects probably involve maternal and fetal dysregulation of glucose, insulin, lipid and amino acid metabolism. Some evidence exists that extreme maternal obesity (BMI ≥40 kg/m(2)) is causally related to a long-term increase in offspring adiposity, but further exploration of this relationship is needed. High gestational weight gain may result in a long-term increase in offspring adiposity if women are already overweight or have obesity at the start of pregnancy. To date, little high-quality human evidence exists that any of these effects are mediated by epigenetic mechanisms, but approaches to appropriately test this possibility are being developed.

A case-control study on the effect of Apolipoprotein E genotypes on gastric cancer risk and progression

Background

Apolipoprotein E (ApoE) is a multifunctional protein playing both a key role in the metabolism of cholesterol and triglycerides, and in tissue repair and inflammation. The ApoE gene (19q13.2) has three major isoforms encoded by ε2, ε3 and ε4 alleles with the ε4 allele associated with hypercholesterolemia and the ε2 allele with the opposite effect. An inverse relationship between cholesterol levels and gastric cancer (GC) has been previously reported, although the relationship between apoE genotypes and GC has not been explored so far.

Methods

One hundred and fifty-six gastric cancer cases and 444 hospital controls were genotyped for apoE polymorphism (ε2, ε3, ε4 alleles). The relationship between GC and putative risk factors was measured using the adjusted odds ratios (ORs) and their 95% confidence intervals (CIs) from logistic regression analysis. A gene-environment interaction analysis was performed. The effect of the apoE genotypes on survival from GC was explored by a Kaplan–Meier analysis and Cox proportional hazard regression model.

Results

Subjects carrying at least one apoE ε2 allele have a significant 60% decrease of GC risk (OR=0.40, 95% CI: 0.19 – 0.84) compared with ε3 homozygotes. No significant interaction emerged between the ε4 or ε2 allele and environmental exposures, nor ε2 or ε4 alleles affected the median survival times, even after correcting for age, gender and stadium.

Conclusions

Our study reports for the first time a protective effect of the ε2 allele against GC, that might be partly attributed to the higher antioxidant properties of ε2 compared with the ε3 or ε4 alleles. Given the study’s sample size, further studies are required to confirm our findings.

Severity of bias of a simple estimator of the causal odds ratio in Mendelian randomization studies

Mendelian randomization studies estimate causal effects using genetic variants as instruments. Instrumental variable methods are straightforward for linear models, but epidemiologists often use odds ratios to quantify effects. Also, odds ratios are often the quantities reported in meta-analyses. Many applications of Mendelian randomization dichotomize genotype and estimate the population causal log odds ratio for unit increase in exposure by dividing the genotype-disease log odds ratio by the difference in mean exposure between genotypes. This 'Wald-type' estimator is biased even in large samples, but whether the magnitude of bias is of practical importance is unclear. We study the large-sample bias of this estimator in a simple model with a continuous normally distributed exposure, a single unobserved confounder that is not an effect modifier, and interpretable parameters. We focus on parameter values that reflect scenarios in which we apply Mendelian randomization, including realistic values for the degree of confounding and strength of the causal effect. We evaluate this estimator and the causal odds ratio using numerical integration and obtain approximate analytic expressions to check results and gain insight. A small simulation study examines finite sample bias and mild violations of the normality assumption. For our simple data-generating model, we find that the Wald estimator is asymptotically biased with a bias of around 10% in fairly typical Mendelian randomization scenarios but which can be larger in more extreme situations. Recently developed methods such as structural mean models require fewer untestable assumptions and we recommend their use when the individual-level data they require are available. The Wald-type estimator may retain a role as an approximate method for meta-analysis based on summary data.

Effect of long-term exposure to lower low-density lipoprotein cholesterol beginning early in life on the risk of coronary heart disease: a Mendelian randomization analysis

OBJECTIVES

The purpose of this study was to estimate the effect of long-term exposure to lower plasma low-density lipoprotein cholesterol (LDL-C) on the risk of coronary heart disease (CHD).

BACKGROUND

LDL-C is causally related to the risk of CHD. However, the association between long-term exposure to lower LDL-C beginning early in life and the risk of CHD has not been reliably quantified.

METHODS

We conducted a series of meta-analyses to estimate the effect of long-term exposure to lower LDL-C on the risk of CHD mediated by 9 polymorphisms in 6 different genes. We then combined these Mendelian randomization studies in a meta-analysis to obtain a more precise estimate of the effect of long-term exposure to lower LDL-C and compared it with the clinical benefit associated with the same magnitude of LDL-C reduction during treatment with a statin.

RESULTS

All 9 polymorphisms were associated with a highly consistent reduction in the risk of CHD per unit lower LDL-C, with no evidence of heterogeneity of effect (I(2) = 0.0%). In a meta-analysis combining nonoverlapping data from 312,321 participants, naturally random allocation to long-term exposure to lower LDL-C was associated with a 54.5% (95% confidence interval: 48.8% to 59.5%) reduction in the risk of CHD for each mmol/l (38.7 mg/dl) lower LDL-C. This represents a 3-fold greater reduction in the risk of CHD per unit lower LDL-C than that observed during treatment with a statin started later in life (p = 8.43 × 10(-19)).

CONCLUSIONS

Prolonged exposure to lower LDL-C beginning early in life is associated with a substantially greater reduction in the risk of CHD than the current practice of lowering LDL-C beginning later in life.

Traits of ADHD and autism in girls with a twin brother: a Mendelian randomization study

It has been hypothesized that prenatal exposure to testosterone may be associated with traits of attention-deficit/hyperactivity disorder (ADHD) or autism spectrum disorder (ASD). We conducted a population-based study of dizygotic female twins to elucidate this hypothesis, assuming that the sex of the co-twin influences the level of prenatal exposure to testosterone. We invited parents of 24,552 3- to 15-year-old twins to answer questionnaires on traits of ADHD and ASD. We analysed the data using a proportional odds model with sex of the co-twin as an instrumental variable for prenatal exposure to testosterone of female twins. We received responses for 6,339 girls from dizygotic twin pairs. Odds ratios for male versus female co-twin were 0.71 (95 % confidence interval 0.61-0.81) for ADHD traits and 0.74 (0.66-0.83) for ASD traits, indicating that a twin brother reduces traits of ADHD and ASD in females. In conclusion, we found that female twins with a twin brother scored significantly lower in parent-reported traits of ADHD and ASD than those with a twin sister. The reason for this may be parental reporting bias, or confounding by unmeasured variables, or a causal effect of an intrauterine environment modified by the sex of the co-twin in the opposite direction of what we expected.

Pancreatic beta-cell function and type 2 diabetes risk: quantify the causal effect using a Mendelian randomization approach based on meta-analyses

The objective of the study is to quantify the causal effect of β-cell function on type 2 diabetes by minimizing residual confounding and reverse causation. We employed a Mendelian randomization (MR) approach using TCF7L2 variant rs7903146 as an instrument for lifelong levels of β-cell function. We first conducted two sets of meta-analyses to quantify the association of the TCF7L2 variant with the risk of type 2 diabetes among 55 436 cases and 106 020 controls from 66 studies by calculating pooled odds ratio (OR) and to quantify the associations with multiple direct or indirect measures of β-cell function among 35 052 non-diabetic individuals from 31 studies by calculating pooled mean difference. We further applied the method of MR to obtain the causal estimates for the effect of β-cell function on type 2 diabetes risk based on findings from the meta-analyses. The OR [95% confidence interval (CI)] was 0.87 (0.81-0.93) for each five unit increment in homeostasis model assessment of insulin secretion (HOMA-%B) (P = 3.0 × 10(-5)). In addition, for measures based on intravenous glucose tolerance test, ORs (95% CI) associated with type 2 diabetes risk were 0.24 (0.08-0.74) (P = 0.01) and 0.14 (0.04-0.48) (P = 0.002) for per 1 standard deviation increment in insulin sensitivity index and disposition index, respectively. Findings from the present study lend support to a causal role of pancreatic β-cell function itself in the etiology of type 2 diabetes.

Mendelian randomization suggests non-causal associations of testosterone with cardiometabolic risk factors and mortality

Prospective studies showed that low serum testosterone concentrations are associated with various cardiometabolic risk factors and mortality. However, the causal nature of these associations is controversial. We studied 1 882 men aged 20-79 years with serum testosterone concentrations and genotyping data from the longitudinal population-based Study of Health in Pomerania. Testosterone concentrations were cross-sectionally associated with cardiometabolic risk factors, including anthropometric, lipid, blood pressure and glycaemic parameters; and prospectively with all-cause mortality (277 deaths, 14.7%) during the 10-year follow-up. To overcome problems of residual confounding, reverse causation, or regression dilution bias in the investigated testosterone-outcome associations, we used two-stage least square regression models with previously identified polymorphisms at the SHBG gene (rs12150660) and X chromosome (rs5934505) as multiple genetic instruments in an instrumental variable (IV) approach, also known as Mendelian randomization. In standard regression analyses, testosterone was robustly associated with a wide range of cardiometabolic risk factors. In subsequent IV analyses, no such significant associations were observed. Similarly, prospective analyses showed a consistent association of low testosterone concentrations with increased all-cause mortality risk, which was not apparent in subsequent IV analyses. The present Mendelian randomization analyses did not detect any evidence for causal associations of testosterone concentrations with cardiometabolic risk factors and mortality, suggesting that previously reported associations might largely result from residual confounding or reverse causation. Although testosterone assessment might improve risk prediction, implementation of testosterone replacement therapy requires further evidence of a direct effect on cardiometabolic outcomes from double-blinded randomized controlled trials and large-scale Mendelian randomization meta-analyses.

Nonfasting glucose, ischemic heart disease, and myocardial infarction: a Mendelian randomization study

OBJECTIVES

The purpose of this study was to test whether elevated nonfasting glucose levels associate with and cause ischemic heart disease (IHD) and myocardial infarction (MI).

BACKGROUND

Elevated fasting plasma glucose levels associate with increased risk of IHD, but whether this is also true for nonfasting levels and whether this is a causal relationship is unknown.

METHODS

Using a Mendelian randomization approach, we studied 80,522 persons from Copenhagen, Denmark. Of those, IHD developed in 14,155, and MI developed in 6,257. Subjects were genotyped for variants in GCK (rs4607517), G6PC2 (rs560887), ADCY5 (rs11708067), DGKB (rs2191349), and ADRA2A (rs10885122) associated with elevated fasting glucose levels in genome-wide association studies.

RESULTS

Risk of IHD and MI increased stepwise with increasing nonfasting glucose levels. The hazard ratio for IHD in subjects with nonfasting glucose levels ≥11 mmol/l (≥198 mg/dl) versus <5 mmol/l (<90 mg/dl) was 6.9 (95% confidence interval [CI]: 4.2 to 11.2) adjusted for age and sex, and 2.3 (95% CI: 1.3 to 4.2) adjusted multifactorially; corresponding values for MI were 9.2 (95% CI: 4.6 to 18.2) and 4.8 (95% CI: 2.1 to 11.2). Increasing number of glucose-increasing alleles was associated with increasing nonfasting glucose levels and with increased risk of IHD and MI. The estimated causal odds ratio for IHD and MI by instrumental variable analysis for a 1-mmol/l (18-mg/dl) increase in nonfasting glucose levels due to genotypes combined were 1.25 (95% CI: 1.03 to 1.52) and 1.69 (95% CI: 1.28 to 2.23), and the corresponding observed hazard ratio for IHD and MI by Cox regression was 1.18 (95% CI: 1.15 to 1.22) and 1.09 (95% CI: 1.07 to 1.11), respectively.

CONCLUSIONS

Like common nonfasting glucose elevation, plasma glucose-increasing polymorphisms associate with increased risk of IHD and MI. These data are compatible with a causal association.

Plasma HDL cholesterol and risk of myocardial infarction: a mendelian randomisation study

BACKGROUND

High plasma HDL cholesterol is associated with reduced risk of myocardial infarction, but whether this association is causal is unclear. Exploiting the fact that genotypes are randomly assigned at meiosis, are independent of non-genetic confounding, and are unmodified by disease processes, mendelian randomisation can be used to test the hypothesis that the association of a plasma biomarker with disease is causal.

METHODS

We performed two mendelian randomisation analyses. First, we used as an instrument a single nucleotide polymorphism (SNP) in the endothelial lipase gene (LIPG Asn396Ser) and tested this SNP in 20 studies (20,913 myocardial infarction cases, 95,407 controls). Second, we used as an instrument a genetic score consisting of 14 common SNPs that exclusively associate with HDL cholesterol and tested this score in up to 12,482 cases of myocardial infarction and 41,331 controls. As a positive control, we also tested a genetic score of 13 common SNPs exclusively associated with LDL cholesterol.

FINDINGS

Carriers of the LIPG 396Ser allele (2·6% frequency) had higher HDL cholesterol (0·14 mmol/L higher, p=8×10(-13)) but similar levels of other lipid and non-lipid risk factors for myocardial infarction compared with non-carriers. This difference in HDL cholesterol is expected to decrease risk of myocardial infarction by 13% (odds ratio [OR] 0·87, 95% CI 0·84-0·91). However, we noted that the 396Ser allele was not associated with risk of myocardial infarction (OR 0·99, 95% CI 0·88-1·11, p=0·85). From observational epidemiology, an increase of 1 SD in HDL cholesterol was associated with reduced risk of myocardial infarction (OR 0·62, 95% CI 0·58-0·66). However, a 1 SD increase in HDL cholesterol due to genetic score was not associated with risk of myocardial infarction (OR 0·93, 95% CI 0·68-1·26, p=0·63). For LDL cholesterol, the estimate from observational epidemiology (a 1 SD increase in LDL cholesterol associated with OR 1·54, 95% CI 1·45-1·63) was concordant with that from genetic score (OR 2·13, 95% CI 1·69-2·69, p=2×10(-10)).

INTERPRETATION

Some genetic mechanisms that raise plasma HDL cholesterol do not seem to lower risk of myocardial infarction. These data challenge the concept that raising of plasma HDL cholesterol will uniformly translate into reductions in risk of myocardial infarction.

FUNDING

US National Institutes of Health, The Wellcome Trust, European Union, British Heart Foundation, and the German Federal Ministry of Education and Research.

Exploring the relationship between maternal iron status and offspring's blood pressure and adiposity: a Mendelian randomization study

Background

Iron deficiency is the most common micronutrient deficiency worldwide. Experimental animal studies suggest that mothers deficient in iron during pregnancy are more likely to have offspring who become obese with high blood pressure. C282Y mutation carriers are more likely to have higher iron stores.

Methods

We undertook an instrumental variable (IV) analysis, using maternal C282Y as an indicator for the mother’s iron status, to examine its association with offspring blood pressure (BP), waist circumference (WC), and body mass index (BMI), and compared the results to that of ordinary least squares (OLS) regression. Offspring of a sub-cohort of mothers from the UK Women’s Cohort Study (UKWCS) were recruited in 2009–2010 (n = 348, mean age = 41 years). Their blood pressure, height, and weight were measured at their local general medical practice, and they were asked to self-measure their waist circumference. About half were offspring of C282Y carriers. Maternal ferritin was used as a biomarker of maternal iron status.

Results

Maternal C282Y was strongly associated with maternal ferritin (mean difference per allele = 84 g/L, 95% confidence interval: 31–137, P = 0.002). Using IV analyses, maternal ferritin was not linked to offspring’s BP, BMI, or WC. The first stage F-statistic for the strength of the instrument was 10 (Kleibergen–Paap rk LM P = 0.009). Maternal ferritin was linked to offspring diastolic BP, WC, and BMI in univariable, but not in multivariable OLS analysis. There was no difference between the OLS and the IV models coefficients for any of the outcomes considered.

Conclusion

We found no association between maternal iron status and adult offspring’s BP and adiposity using both multivariable OLS and IV modeling. To our knowledge, this is the first study examining this relationship. Further exploration in larger studies that have genetic variation assessed in both mother and offspring should be considered.

Causation and causal inference for genetic effects

Over the past three decades, substantial developments have been made on how to infer the causal effect of an exposure on an outcome, using data from observational studies, with the randomized experiment as the golden standard. These developments have reshaped the paradigm of how to build statistical models, how to adjust for confounding, how to assess direct effects, mediated effects and interactions, and even how to analyze data from randomized experiments. The congruence of random transmission of alleles during meiosis and the randomization in controlled experiments/trials, suggests that genetic studies may lend themselves naturally to a causal analysis. In this contribution, we will reflect on this and motivate, through illustrative examples, where insights from the causal inference literature may help to understand and correct for typical biases in genetic effect estimates.

Circulating 25-hydroxyvitamin D concentration and the risk of type 2 diabetes: results from the European Prospective Investigation into Cancer (EPIC)-Norfolk cohort and updated meta-analysis of prospective studies

AIMS/HYPOTHESIS

Epidemiological evidence is suggestive, but limited, for an association between circulating 25-hydroxyvitamin D (25[OH]D) and risk of type 2 diabetes. We conducted a systematic review and meta-analysis that included new data from previously unpublished studies.

METHODS

Using a nested case-cohort design in the European Prospective Investigation into Cancer (EPIC)-Norfolk study, we identified a random subcohort and incident type 2 diabetes cases occurring between baseline (1993-1997) and 2006. In the Ely prospective study we identified incident type 2 diabetes cases between 1990 and 2003. We conducted a systematic review of prospective studies on 25(OH)D and type 2 diabetes published in MEDLINE or EMBASE until 31 January 2012, and performed a random-effects meta-analysis combining available evidence with results from the EPIC-Norfolk and Ely studies.

RESULTS

In EPIC-Norfolk, baseline 25(OH)D was lower among incident type 2 diabetes cases (mean [SD] 61.6 [22.4] nmol/l; n=621) vs non-case subcohort participants (mean 65.3 [23.9] nmol/l; n=826). There was an inverse association between baseline 25(OH)D and incident type 2 diabetes in multivariable-adjusted analyses: HR (95% CI) 0.66 (0.45, 0.97), 0.53 (0.34, 0.82), 0.50 (0.32, 0.76), p trend <0.001, comparing consecutive increasing 25(OH)D quartiles with the lowest. In Ely, 37 incident type 2 diabetes cases were identified among 777 participants. In meta-analysis, the combined RR of type 2 diabetes comparing the highest with lowest quartile of 25(OH)D was 0.59 (0.52, 0.67), with little heterogeneity (I (2) =2.7%, p=0.42) between the 11 studies included (3,612 cases and 55,713 non-cases).

CONCLUSIONS/INTERPRETATION

These findings demonstrate an inverse association between circulating 25(OH)D and incident type 2 diabetes. However, causal inference should be addressed through adequately dosed randomised trials of vitamin D supplementation or genetic Mendelian randomisation experiments.

Lipoprotein(a) and cardiovascular disease in diabetic patients

Lipoprotein(a) (Lp[a]) is a LDL-like particle consisting of an ApoA moiety linked to one molecule of ApoB(100). Recent data from large-scale prospective studies and genetic association studies provide highly suggestive evidence for a potentially causal role of Lp(a) in affecting risk of cardiovascular disease (CVD) in general populations. Patients with Type 2 diabetes display clustered metabolic abnormalities and elevated risk of CVD. Lower plasma Lp(a) levels were observed in diabetic patients in several recent studies. Epidemiology studies of Lp(a) and CVD risk in diabetic patients generated inconsistent results. We recently found that Lp(a)-related genetic markers did not predict CVD in two diabetic cohorts. The current data suggest that Lp(a) may differentially affect cardiovascular risk in diabetic patients and in the general population. More prospective studies, Mendelian randomization analysis and functional studies are needed to clarify the causal relationship of Lp(a) and CVD in diabetic patients.

Serum 25-hydroxyvitamin D3 and D2 and non-clinical psychotic experiences in childhood

Objective

Non-clinical psychotic experiences are common and distressing. It has been hypothesized that early life vitamin D deficiency may be a risk factor for psychosis-related outcomes, but it is not known if circulating concentrations of 25-hydroxyvitamin D (25(OH)D) during childhood are associated with psychosis-related outcomes or whether the two different forms of 25(OH)D, (25(OH)D₃ and 25(OH)D₂, have similar associations with psychosis-related outcomes.

Methods

We investigated the association between serum 25(OH)D₃ and 25(OH)D₂ concentrations and psychotic experiences in a prospective birth cohort study. Serum 25(OH)D₃ and 25(OH)D₂ concentrations were measured at mean age 9.8 years and psychotic experiences assessed at mean age 12.8 years by a psychologist (N = 3182).

Results

Higher 25(OH)D₃ concentrations were associated with lower risk of definite psychotic experiences (adjusted odds ratio: OR (95% confidence interval: CI) 0.85 (0.75–0.95)). Higher concentrations of 25(OH)D₂ were associated with higher risk of suspected and definite psychotic experiences (adjusted odds ratio: OR (95% confidence interval: CI) 1.26 (1.11, 1.43)). Higher 25(OD)D₂ concentrations were also weakly associated with definite psychotic experiences (adjusted OR (95% CI) 1.17 (0.96, 1.43), though with wide confidence intervals including the null value.

Conclusions

Our findings of an inverse association of 25(OH)D₃ with definite psychotic experiences is consistent with the hypothesis that vitamin D may protect against psychosis-related outcomes.

Cerebrospinal fluid APOE levels: an endophenotype for genetic studies for Alzheimer's disease

The apolipoprotein E (APOE) genotype is the major genetic risk factor for Alzheimer's disease (AD). We have access to cerebrospinal fluid (CSF) and plasma APOE protein levels from 641 individuals and genome-wide genotyped data from 570 of these samples. The aim of this study was to test whether CSF or plasma APOE levels could be a useful endophenotype for AD and to identify genetic variants associated with APOE levels. We found that CSF (P = 8.15 × 10(-4)) but not plasma (P = 0.071) APOE protein levels are significantly associated with CSF Aβ(42) levels. We used Mendelian randomization and genetic variants as instrumental variables to confirm that the association of CSF APOE with CSF Aβ(42) levels and clinical dementia rating (CDR) is not because of a reverse causation or confounding effect. In addition the association of CSF APOE with Aβ(42) levels was independent of the APOE ε4 genotype, suggesting that APOE levels in CSF may be a useful endophenotype for AD. We performed a genome-wide association study to identify genetic variants associated with CSF APOE levels: the APOE ε4 genotype was the strongest single-genetic factor associated with CSF APOE protein levels (P = 6.9 × 10(-13)). In aggregate, the Illumina chip single nucleotide polymorphisms explain 72% of the variability in CSF APOE protein levels, whereas the APOE ε4 genotype alone explains 8% of the variability. No other genetic variant reached the genome-wide significance threshold, but nine additional variants exhibited a P-value <10(-6). Pathway mining analysis indicated that these nine additional loci are involved in lipid metabolism (P = 4.49 × 10(-9)).

Two-step epigenetic Mendelian randomization: a strategy for establishing the causal role of epigenetic processes in pathways to disease

The burgeoning interest in the field of epigenetics has precipitated the need to develop approaches to strengthen causal inference when considering the role of epigenetic mediators of environmental exposures on disease risk. Epigenetic markers, like any other molecular biomarker, are vulnerable to confounding and reverse causation. Here, we present a strategy, based on the well-established framework of Mendelian randomization, to interrogate the causal relationships between exposure, DNA methylation and outcome. The two-step approach first uses a genetic proxy for the exposure of interest to assess the causal relationship between exposure and methylation. A second step then utilizes a genetic proxy for DNA methylation to interrogate the causal relationship between DNA methylation and outcome. The rationale, origins, methodology, advantages and limitations of this novel strategy are presented.

The Promises and Pitfalls of Genoeconomics*

This article reviews existing research at the intersection of genetics and economics, presents some new findings that illustrate the state of genoeconomics research, and surveys the prospects of this emerging field. Twin studies suggest that economic outcomes and preferences, once corrected for measurement error, appear to be about as heritable as many medical conditions and personality traits. Consistent with this pattern, we present new evidence on the heritability of permanent income and wealth. Turning to genetic association studies, we survey the main ways that the direct measurement of genetic variation across individuals is likely to contribute to economics, and we outline the challenges that have slowed progress in making these contributions. The most urgent problem facing researchers in this field is that most existing efforts to find associations between genetic variation and economic behavior are based on samples that are too small to ensure adequate statistical power. This has led to many false positives in the literature. We suggest a number of possible strategies to improve and remedy this problem: (a) pooling data sets, (b) using statistical techniques that exploit the greater information content of many genes considered jointly, and (c) focusing on economically relevant traits that are most proximate to known biological mechanisms.

Methods for meta-analysis of individual participant data from Mendelian randomisation studies with binary outcomes

Mendelian randomisation is an epidemiological method for estimating causal associations from observational data by using genetic variants as instrumental variables. Typically the genetic variants explain only a small proportion of the variation in the risk factor of interest, and so large sample sizes are required, necessitating data from multiple sources. Meta-analysis based on individual patient data requires synthesis of studies which differ in many aspects. A proposed Bayesian framework is able to estimate a causal effect from each study, and combine these using a hierarchical model. The method is illustrated for data on C-reactive protein and coronary heart disease (CHD) from the C-reactive protein CHD Genetics Collaboration (CCGC). Studies from the CCGC differ in terms of the genetic variants measured, the study design (prospective or retrospective, population-based or case-control), whether C-reactive protein was measured, the time of C-reactive protein measurement (pre- or post-disease), and whether full or tabular data were shared. We show how these data can be combined in an efficient way to give a single estimate of causal association based on the totality of the data available. Compared to a two-stage analysis, the Bayesian method is able to incorporate data on 23% additional participants and 51% more events, leading to a 23-26% gain in efficiency.

Integrating genomic data and social science: challenges and opportunities

Why should social scientists be interested in using molecular genetic data? Here are five reasons:

Associations between serum uric acid and markers of subclinical atherosclerosis in young adults. The cardiovascular risk in Young Finns study

BACKGROUND AND METHODS

Serum uric acid (SUA) is a suggested biomarker for established coronary artery disease, but the role of SUA in early phases of atherosclerosis is controversial. The relations of SUA with vascular markers of subclinical atherosclerosis, including carotid artery intima-media thickness (cIMT), carotid plaque, carotid distensibility (Cdist) and brachial flow-mediated dilatation (FMD) were examined in 1985 young adults aged 30-45 years. In addition to ordinary regression, we used Mendelian randomization techniques to infer causal associations.

RESULTS

In women, the independent multivariate correlates of SUA included BMI, creatinine, alcohol use, triglycerides, glucose and adiponectin (inverse association) (Model R(2) = 0.30). In men, the correlates were BMI, creatinine, triglycerides, C-reactive protein, alcohol use, total cholesterol and adiponectin (inverse) (Model R(2) = 0.33). BMI alone explained most of the variation of SUA levels both in women and men (Partial R(2) ∼ 0.2). When SUA was modeled as an explanatory variable for vascular markers, it directly associated with cIMT and inversely with Cdist in age- and sex-adjusted analysis. After further adjustments for BMI or glomerular filtration rate, these relations were reduced to non-significance. No associations were found between SUA and FMD or the presence of a carotid plaque. Mendelian randomization analyses using known genetic variants for BMI and SUA confirmed that BMI is causally linked to SUA and that BMI is a significant confounder in the association between SUA and cIMT.

CONCLUSION

SUA is associated with cardiovascular risk markers in young adults, especially BMI, but we found no evidence that SUA would have an independent role in the pathophysiology of early atherosclerosis.

Instrumental variable estimation of the causal effect of plasma 25-hydroxy-vitamin D on colorectal cancer risk: a mendelian randomization analysis

Vitamin D deficiency has been associated with several common diseases, including cancer and is being investigated as a possible risk factor for these conditions. We reported the striking prevalence of vitamin D deficiency in Scotland. Previous epidemiological studies have reported an association between low dietary vitamin D and colorectal cancer (CRC). Using a case-control study design, we tested the association between plasma 25-hydroxy-vitamin D (25-OHD) and CRC (2,001 cases, 2,237 controls). To determine whether plasma 25-OHD levels are causally linked to CRC risk, we applied the control function instrumental variable (IV) method of the mendelian randomization (MR) approach using four single nucleotide polymorphisms (rs2282679, rs12785878, rs10741657, rs6013897) previously shown to be associated with plasma 25-OHD. Low plasma 25-OHD levels were associated with CRC risk in the crude model (odds ratio (OR): 0.76, 95% Confidence Interval (CI): 0.71, 0.81, p: 1.4×10(-14)) and after adjusting for age, sex and other confounding factors. Using an allele score that combined all four SNPs as the IV, the estimated causal effect was OR 1.16 (95% CI 0.60, 2.23), whilst it was 0.94 (95% CI 0.46, 1.91) and 0.93 (0.53, 1.63) when using an upstream (rs12785878, rs10741657) and a downstream allele score (rs2282679, rs6013897), respectively. 25-OHD levels were inversely associated with CRC risk, in agreement with recent meta-analyses. The fact that this finding was not replicated when the MR approach was employed might be due to weak instruments, giving low power to demonstrate an effect (<0.35). The prevalence and degree of vitamin D deficiency amongst individuals living in northerly latitudes is of considerable importance because of its relationship to disease. To elucidate the effect of vitamin D on CRC cancer risk, additional large studies of vitamin D and CRC risk are required and/or the application of alternative methods that are less sensitive to weak instrument restrictions.

The genetics of coronary heart disease

BACKGROUND

Coronary heart disease (CHD) is a leading cause of death worldwide, yet many areas of its pathogenesis remain unknown or poorly understood, leaving potential for novel preventive and therapeutic interventions. Recent major advances in genomic science and technology have opened new avenues of investigation in the pathogenesis of CHD, some of which are leading to clinical translation.

SOURCES OF DATA

The published literature in CHD genetics has burgeoned in the last 5 years with the reporting of genome-wide association studies (GWASs) and many other findings.

AREAS OF AGREEMENT

Identification of many genetic variants with small effects on CHD risk has been a common finding. These have included several predicted loci, such as those involved in conventional CHD risk factors (e.g. plasma lipids) and many novel loci, where their mechanism of action is unclear. The need for large, collaborative approaches to research has also become clear and is now an accepted modus operandi.

AREAS OF CONTROVERSY

The clinical utility of novel GWAS findings remains uncertain. In particular, the relative contribution of common variants of modest effect and rare variants of larger effects to risk of CHD or response to drugs is unclear.

GROWING POINTS

As a greater number of larger GWASs are conducted in CHD and its related phenotypes, much effort is being made to find translational applications for their findings. Therapeutics, prediction and pathology are major areas of research endeavour.

Using multiple genetic variants as instrumental variables for modifiable risk factors

Mendelian randomisation analyses use genetic variants as instrumental variables (IVs) to estimate causal effects of modifiable risk factors on disease outcomes. Genetic variants typically explain a small proportion of the variability in risk factors; hence Mendelian randomisation analyses can require large sample sizes. However, an increasing number of genetic variants have been found to be robustly associated with disease-related outcomes in genome-wide association studies. Use of multiple instruments can improve the precision of IV estimates, and also permit examination of underlying IV assumptions. We discuss the use of multiple genetic variants in Mendelian randomisation analyses with continuous outcome variables where all relationships are assumed to be linear. We describe possible violations of IV assumptions, and how multiple instrument analyses can be used to identify them. We present an example using four adiposity-associated genetic variants as IVs for the causal effect of fat mass on bone density, using data on 5509 children enrolled in the ALSPAC birth cohort study. We also use simulation studies to examine the effect of different sets of IVs on precision and bias. When each instrument independently explains variability in the risk factor, use of multiple instruments increases the precision of IV estimates. However, inclusion of weak instruments could increase finite sample bias. Missing data on multiple genetic variants can diminish the available sample size, compared with single instrument analyses. In simulations with additive genotype-risk factor effects, IV estimates using a weighted allele score had similar properties to estimates using multiple instruments. Under the correct conditions, multiple instrument analyses are a promising approach for Mendelian randomisation studies. Further research is required into multiple imputation methods to address missing data issues in IV estimation.