Evaluation of bidirectional causal association between cardiovascular diseases and pneumonia: A Mendelian randomization study.

Abstract

Observational evidence suggests a bidirectional relationship between cardiovascular diseases (CVDs) and pneumonia. However, the causality between CVDs and pneumonia remains undetermined. Thus, we aimed to investigate the bidirectional causality between CVDs and pneumonia using Mendelian randomization (MR) analysis. Global genetic correlation analysis and bidirectional two-sample MR analysis were performed to infer the genetic correlation and causality between CVDs and pneumonia by using genome-wide association study (GWAS) summary data from GWAS meta-analysis study, FinnGen or UK Biobank consortium. Post-hoc power calculation was conducted to assess the power for detecting the causality. The linkage disequilibrium score regression analysis suggested a positive significant genetic correlation between CVDs and pneumonia. In the MR analysis, only genetically predicted ischemic stroke was causally associated with any pneumonia (odds ratio [OR]: 1.119, 95% confidence interval [CI]: 1.031-1.393), bacterial pneumonia (OR: 1.251, 95% CI: 1.032-1.516), and pneumococcal pneumonia (OR: 1.308, 95% CI: 1.093-1.565), but the causality was attenuated to non-significance after adjusting for deep venous thrombosis. However, the causal effects of pneumonia on CVDs were not detected. Post-hoc power calculations supported strong power (more than 80%) to detect the causality. Ischemic stroke is causally associated with an increased risk of pneumonia, but there is no evidence for the causal effect of pneumonia on CVDs. Our findings have important implications as they provide further support for the thrombosis risk screening as a strategy to reduce the incidence of pneumonia in patients with ischemic stroke.