Deciphering correlation and causation in risk factors for heart disease with Mendelian randomization

Abstract

In this analysis, we studied the use of Mendelian randomization to identify the risk factors of coronary artery disease (CAD), a major cause of cardiovascular disease. Identifying risk factors of CAD are critical to understanding and managing the disease. Our analysis combined results from 28 genetic analyses from 12 unique studies. For each genetic variation, we obtained the variant ID, chromosome, base-pair position, reference and alternative alleles of the genetic variation, and estimated effect and p-value of the genetic variation on the outcome. We hypothesized that traits which are correlated with CAD outcomes will be causally associated with CAD risk in a genetic Mendelian randomization analysis. Our analysis showed that several traits such as blood pressure readings (systolic, OR 0.51 (95% CI: 0.34-0.69), p-value = 5.4x10-9) and (diastolic, OR 0.56 (95% CI: 0.41-0.71), p-value = 7.6x10-14), LDL cholesterol levels (OR 0.54 (95% CI: 0.47-0.60), p-value = 4.4x10-56), and BMI (OR 0.41 (95% I: 0.35-0.48), p-value = 6.30x10-33) were significant risk factors for CAD. C-reactive protein (OR -0.09 (95% CI: -0.18–0.00), p-value = 0.05) was a protective risk factor of CAD due to its negative odds ratio. In contrast, eosinophil count (OR -0.007 (95% CI: -0.06-0.04), p-value = 0.79) had no statistically significant association. Blood cells had weak associations with CAD, and uric acid’s role as a causal or reversible risk factor of CAD was inconclusive, requiring further study.