Background
Major depressive disorder (MDD) is a highly heterogeneous mental illness and a major public health problem worldwide. A large number of observational studies have demonstrated a clear association between MDD and coronary heart disease (CHD), and some studies have even suggested that the relationship is bidirectional. However, it was unknown whether any causal relationship existed between them and whether causality was bidirectional in such an instance. Thus, we aimed to determine whether there is a bidirectional causal relationship between major depressive disorders and coronary heart disease.
Methods
Our two-sample Bidirectional Mendelian Randomization Study consisted of two parts: forward MR analysis regarded MDD as exposure and CHD as the outcome, and reverse MR analysis considered CHD as exposure and MDD as the outcome. Summary data on MDD and CHD were obtained from the IEU Open GWAS database. After screening criteria(P < $$5\times {10}^{-8}$$
5
×
10
-
8
), 47 MDD-associated SNPs and 39 CHD-associated SNPs were identified. The inverse-variance weighted (IVW) method, ME-Egger regression, and weighted median method were used to estimate causality. In addition, sensitivity methods, including the heterogeneity test, horizontal pleiotropy test, and leave-one-out method, were applied to ensure the robustness of causal estimation.
Results
Based on the MR-Egger regression intercept test results, there did not appear to be any horizontal pleiotropy in this study (MDD: intercept = -0.0000376, P = 0.9996; CHD: intercept = -0.0002698, P = 0.920). Accordingly, IVW results suggested consistent estimates of causal effect values. The results showed that people with MDD increased the risk of CHD by 14.7% compared with those without MDD (OR = 1.147, 95%CI: 1.045–1.249, P = 0.009). But there was no direct evidence that CHD would increase the risk of MDD(OR = 1.008, 95%CI: 0.985–1.031, P = 0.490). The heterogeneity test and funnel plot showed no heterogeneity in 47 SNPs of MDD (Q = 42.28, $${I}^{2}$$
I
2
=0, P = 0.629), but there was heterogeneity in 39 SNPs of CHD (Q = 62.48, $${I}^{2}$$
I
2
=39.18%, P = 0.007). The leave-one-out method failed to identify instances where a single SNP was either biased toward or dependent on the causation.
Conclusion
Our study supports a one-way causal relationship between MDD and CHD, but there is no bidirectional causal relationship. MDD increases the risk of CHD, but there is no evidence that CHD increases the risk of MDD. Therefore, the influence of psychological factors should also be considered in the prevention and treatment of CHD. For MDD patients, it is necessary to prevent cardiovascular diseases.
