Mendelian randomisation in type 2 diabetes and coronary artery disease
Introduction
Metabolic diseases, such as type 2 diabetes (T2D), coronary artery disease (CAD) and hypertension affect large proportions of the world’s population. For example, diabetes was estimated to effect 4.25 million adults globally (population approximately 4.84 billion) in 2017, whilst hypertension and CAD affected 1.13 billion and 110 million people worldwide in 2015 respectively [1, 2, 3]. Many factors are associated with these metabolic diseases including altered lipid, glucose and metabolite levels, differential adiposity measures including waist-to-hip-ratio (WHR), fatty liver markers, glucose intolerance, insulin resistance, differential gene expression and epigenetic patterns [4, 5, 6]. Identifying which factors are causal and not a result of disease or its treatment is difficult. Many prospective studies imply causality by showing clear associations between these factors and disease by indicating that these changes occurred before disease onset. However, prospective associations do not prove causality because changes can occur long before disease and may be associated with unaccounted for confounding factors [7].
Genetic variation influences metabolism and can help us understand the true causal effects between traits and disease to deepen our understanding of the pathophysiology of metabolic diseases [8]. Genetic studies have resulted in an improved understanding of disease for several reasons. These reasons include the use of Mendelian randomisation (MR). The discovery and availability of many tens, sometimes hundreds, of genetic variants associated with a risk factor have facilitated more powerful MR studies. These variants, discovered by genome-wide association studies (GWAS), have provided important `instruments’ for MR studies that help define the role of these risk factors in disease and related traits.
Section snippets
Mendelian randomisation
MR can serve as a useful alternative or complement to randomised control trials (RCTs). Well-designed RCTs provide the most reliable evidence for causality. RCTs involve random allocation of interventions or placebos to participants eligible for study to determine efficacy of treatments. Randomisation reduces bias and confounding but is often costly and time-consuming and may be high-risk and ethically unjust [9•]. MR can be used to test for causal associations between a risk factor and outcome
Methods
We performed an extensive literature search using Web of Science (https://apps.webofknowledge.com/) and PubMed databases [20]. The search terms sequentially used were “Mendelian Randomisation” OR “Mendelian Randomization” and the risk factor and disease name with all variations included (Table 1). We next filtered on year of publication; January 2015–present day inclusive. We then eliminated studies involving other traits. We scan-read the remaining papers and chose those using MR to find
Genetic variants associated with BMI to assess causality with cardiometabolic diseases
GWAS have identified 97 genetic variants associated with BMI. These 97 variants account for 2.7% of BMI variation [37••]. These variants, and subsets of them identified in earlier studies, have been used in MR tests to assess the causal relationship between BMI and disease.
Summary
Over the last two years, MR has provided evidence of causal genetic associations between anthropometric traits and cardiometabolic diseases. In most cases, MR has helped to support results obtained through observational studies and RCTs by providing genetic evidence for the causal relationships. The limitations of these MR studies include a lack of knowledge about the mechanistic actions of the genetic variants used as instruments. Careful consideration of the three assumptions (Figure 1) is
Conflict of interest statement
Charli Stoneman is an MRC CASE student partly funded by GSK. Tim Frayling has consulted for several pharmaceutical companies including Sanofi and Boehringer-Ingelheim.
References and recommended reading
Papers of particular interest, published within the period of review, have been highlighted as
• of special interest
•• of outstanding interest
References (84)
- et al.
Epigenetic reprogramming in metabolic disorders: nutritional factors and beyond
J Nutr Biochem
(2018) - et al.
Variation in PCSK9 and HMGCR and risk of cardiovascular disease and diabetes
New Engl J Med
(2016) - et al.
Comparison of waist-to-hip ratio and other obesity indices as predictors of cardiovascular disease risk in people with type 2 diabetes: a prospective cohort study from ADVANCE
Eur J Cardiovasc Prev Rehabil
(2011) - et al.
Measures of general and central obesity and risk of type 2 diabetes in a Ghanaian population
Trop Med Int Health
(2013) - et al.
Anthropometric predictors of coronary heart disease in Chinese women
Int J Obes Relat Metab Disord
(2004) IDF Diabetes Atlas
(2017)- et al.
Global, regional, and national incidence, prevalence, and years lived with disability for 310 diseases and injuries, 1990–2013: a systematic analysis for the Global Burden of Disease Study 2015
Lancet
(2015) Global Health Observatory (GHO) Data
(2015)- et al.
Genetic and epigenetic studies of adiposity and cardiometabolic disease
Genome Med
(2017) - et al.
The association between obesity, cardiometabolic disease biomarkers, and innate immunity-related inflammation in Canadian adults
Diabetes Metab Syndr Obes
(2012)