Elevated morning plasma cortisol is associated with multiple cardiovascular risk factors in metabolic syndrome. Epidemiological studies have reported positive associations between plasma cortisol and coronary heart disease (CHD) however traditional observational studies do not allow causality to be determined and are unable rule out the possibility of confounding. A two-sample Mendelian randomisation approach was used to estimate the causal effect of plasma cortisol on risk of CHD. A genetic instrument for plasma cortisol comprised three SNPs which were associated with plasma cortisol in the recent Cortisol Network (CORNET) genome wide association meta-analysis (GWAMA) (n = 12,597). We investigated the association between this genetic instrument and risk of CHD in 22,223 cases/64,762 controls from the CARDIOGRAM consortium. Each standard deviation rise in genetically predicted plasma cortisol was associated with an odds ratio of 1.27 (95% CI: 1.01–1.60) for CHD. These results are compatible with a causal effect for the observational association between plasma cortisol and CHD. The inconsistent results from observational studies may be explained by: the inverse association between cortisol and obesity, which confounded the positive association of cortisol with other cardiovascular risk factors; and the use of single ‘snapshot’ plasma cortisol measurement rather than cumulative measure of cortisol exposure provided by genetic prediction. A bidirectional Mendelian randomisation analysis between plasma cortisol and BMI may yield greater clarity. To improve the strength of our genetic instrument an expanded CORNET GWAMA is currently underway. Measurements of cortisol may add value to predictions of CHD risk.