Assessing causal relationships using genetic proxies for exposures: an introduction to Mendelian randomization

Abstract

Background and aimsStudying the consequences of addictive behaviours is challenging, with understanding causal relationships from observational data being particularly difficult. For example, people who smoke or drink excessively are often systematically different from those who do not, are less likely to participate in research and may misreport their behaviours when they do. Furthermore, the direction of causation between an addictive behaviour and outcome may be unclear. Mendelian randomization (MR) offers potential solutions to these problems.MethodsWe describe MR's principles and the criteria under which it is valid. We identify challenges and potential solutions in its application (illustrated using two applied examples) and describe methodological extensions in its application.ResultsMR is subject to certain assumptions, and requires the availability of appropriate genetic data, large sample sizes and careful design and conduct. However, it has already been applied successfully to the addiction literature. The relationship between alcohol consumption (proxied by a variant in the ADH1B gene) and cardiovascular risk has been investigated, finding that alcohol consumption increases risk, with no evidence of a cardioprotective effect at moderate consumption levels. In addition, heaviness of smoking (proxied by a variant in the CHRNA5‐A3‐B4 gene cluster) and risk of depression and schizophrenia have been investigated, with no evidence of a causal effect of smoking on depression but some evidence of a causal effect on schizophrenia.ConclusionsMendelian randomization analyses are already producing robust evidence for addiction‐related practice and policy. As genetic variants associated with addictive behaviours are identified, the potential for Mendelian randomization analyses will grow. Methodological developments are also increasing its applicability.

Cannabis use and suicide attempts among 86,254 adolescents aged 12–15 years from 21 low- and middle-income countries

“…The strong and independent associations between cannabis use and suicide attempts in a large sample of adolescents from 21 LMICs herein reported open relevant research and clinical implications. First, the causality of this association warrant confirmation in subsequent prospective and Mendelian Randomization studies [29]. Our findings suggest that cannabis use should be taken into consideration when assessing suicide risk in adolescents particularly those living in LMICs.…”

Calcium and Vitamin D Supplementation Are Not Associated With Risk of Incident Ischemic Cardiac Events or Death: Findings From the UK Biobank Cohort

“…(22,23) However the genetic instruments only explained around 0.8% of the variance in calcium concentrations, and these analyses are strongly dependent on the underlying assumptions. (35) They also represent lifelong exposure to calcium concentration and this design cannot be used to imply that dietary supplementation in older age, or transient increases in calcium concentration lead to these outcomes. Furthermore, it is undocumented, to our knowledge, whether the transient rise in serum calcium concentration following ingestion of a calcium load (which is modest and remains below the saturation point of the calcium x phosphate product) is specifically associated with adverse cardiovascular events.…”

The Role of Genetic Variation of BMI, Body Composition, and Fat Distribution for Mental Traits and Disorders: A Look-Up and Mendelian Randomization Study

“…We performed a power analysis to estimate whether our analysis, given sample size, proportion of cases in the study and the proportion of variance explained, is sufficient to detect a true casual effect (Brion et al, 2013). In order to investigate the relationship between study accuracy and effect size, we created a funnel plot (Haycock et al, 2016;Katikireddi et al, 2018). To examine whether an individual data point (SNP) has a large influence on the regression coefficients, we calculated the IVW regression by leaving each genetic variant out in turn .…”