Background: More than 100 million cases of COVID-19 have been reported worldwide. A number of risk factors for infection or severe infection have been identified, however observational studies were subject to confounding bias. In addition, there is still limited knowledge about the complications or medical consequences of the disease. Methods: Here we performed bi-directional Mendelian randomization (MR) analysis to evaluate causal relationships between liability to COVID-19 (and severe/critical infection) and a wide range of around 30 cardiometabolic disorders (CMD) or traits. Genetic correlation (rg) was assessed by LD score regression(LDSC). The latest GWAS summary statistics from the COVID-19 Host Genetics Initiative was used, which comprised comparisons of general population controls with critically ill, hospitalized and any infected cases. Results: Overall we observed evidence that liability to COVID-19 or severe infection may be causally associated with higher risks of type 2 diabetes mellitus(T2DM), chronic kidney disease(CKD), ischemic stroke (especially large artery stroke[LAS]) and heart failure(HF) when compared to the general population. On the other hand, our findings suggested that liability to atrial fibrillation (AF), stroke (especially LAS), obesity, diabetes (T1DM and T2DM), low insulin sensitivity and impaired renal function (low eGFR and diabetic kidney disease) may be causal risk factors for COVID-19 or severe disease. In genetic correlation analysis, T2DM, CAD, obesity, fasting insulin, CKD, gout, stroke and urate showed positive rg with critical or hospitalized infection. All above findings passed multiple testing correction at a false discovery rate (FDR)<0.05. Conclusions: In summary, this study provides evidence for tentative bi-directional causal associations between liability to COVID-19 and severe disease and a number of CM disorders. Further replications and prospective studies are required to verify the findings.

Statin is one of the most commonly prescribed medications worldwide. Besides reduction of cardiovascular risks, statins have been proposed for the prevention or treatment of other disorders, but results from clinical studies are mixed. There are also controversies concerning the adverse effects caused by statins.In this study we employed a Mendelian randomization (MR) approach across a wide range of complex traits to explore repositioning opportunities and side-effects of statins. MR is analogous to a "naturalistic" randomized controlled trial (RCT), which is much less susceptible to confounding and reverse causation as compared to observational studies.We employed two genetic instruments (rs12916 and rs17238484) in the HMGCR gene which have been shown to provide reliable estimates of the risk of statins on type 2 diabetes and weight gain. We observed in the single-and joint-SNP analysis that low density lipoprotein cholesterol (LDL-C) reduction from HMG-CoA reductase inhibition results in increased depressive symptoms. This finding appeared to be supported by nominally significant results of raised major depression risk in single-SNP MR analysis of rs17238484, and analyses using LDL-C as the exposure. Several other outcomes also reached nominal significance (p < 0.05) in single-or joint-SNP analyses; for example, we observed causal associations of LDL-C lowering from HMG-CoA reductase inhibition with reduced risks of schizophrenia, anorexia nervosa, Alzheimer disease, Parkinson disease, as well as increased forearm bone mineral density, sleep duration and extreme longevity (highest q-value = 0.289). We also found evidence of casual relationships of LDL-C levels with schizophrenia, anorexia, sleep duration and longevity, following the same association directions as in analyses of HMGCR variants. These findings were at least partially supported by previous clinical studies. We did not observe associations with cognitive test profiles, renal outcomes, autoimmune diseases or cancers. While MR has its limitations and our findings remain to be confirmed in further studies, this work demonstrates the potential of a phenome-wide approach to reveal novel therapeutic indications and unknown drug side-effects.