The causal relationship between bacterial pneumonia and diabetes: a two-sample mendelian randomization study

Abstract

The escalating worldwide occurrence of diabetes mellitus, recognized as a chronic metabolic ailment contributing to an amplified global disease burden, has stimulated researchers to explore its etiology. Consequently, the study employed a two-sample Mendelian randomization (MR) methodology to examine the causal connection between bacterial pneumonia and diabetes, drawing upon the existing observational study that identified a potential association between bacterial pneumonia and diabetes. Furthermore, MR investigations suggest a reciprocal causal relationship between bacterial pneumonia and gestational diabetes mellitus(GDM), and a plausible causal link between bacterial pneumonia and T1DM. Background: Previous observational studies have established the high prevalence of bacterial pneumonia in diabetic patients, which in turn leads to increased mortality. However, the presence of a causal connection between bacterial pneumonia and diabetes remains unobserved. Methods: We chose genome-wide significant (Ρ < 1 × 10−5 and Ρ < 1 × 10−6) and independent (r2 < 0.001) single-nucleotide polymorphisms (SNPs) as instrumental variables (IVs) to proceed a bidirectional two-sample MR study. The extracted SNPs explored the relationship between bacterial pneumonia and diabetes by Inverse variance weighted (IVW), MR-Egger, and weighted median methods. In addition, we conducted the Heterogeneity test, the Pleiotropy test, MR-presso and the Leave-one-out (LOO) sensitivity test to validate the reliability of results. Results: In an MR study with bacterial pneumonia as an exposure factor, four different types of diabetes as outcome. It was observed that bacterial pneumonia increases the incidence of GDM (OR = 1.150 (1.027–1.274, P = 0.011) and T1DM (OR = 1.277 (1.024–1.531), P = 0.016). In the reverse MR analysis, it was observed that GDM (OR = 1.112 (1.023–1.201, P = 0.009) is associated with an elevated risk of bacterial pneumonia. However, no significant association was observed bacterial pneumonia with T1DM and other types of diabetes (P > 0.05). Conclusion: This study utilizing MR methodology yields robust evidence supporting a bidirectional causal association between bacterial pneumonia and GDM. Furthermore, our findings suggest a plausible causal link between bacterial pneumonia and T1DM.