A Causal Inference Study Of Chronic Obstructive Pulmonary Disease And Several Common Diseases

Chronic obstructive pulmonary disease(COPD)is a chronic inflammatory respiratory system disease with a large number of patients and a high mortality rate,imposing a significant health care burden on patients,families and the nation.However,current researches are still not fully clear about the underlying molecular mechanisms and risk factors characterizing COPD.Observational studies have shown correlations between some common chronic diseases and COPD,but the existence of correlations does not necessarily prove a causal association.Therefore,causal relationships between diseases need to be explored by means of causal inference methods.This study guarantees the robustness of the results through a two-sample MR methods under several different hypotheses,and tests the heterogeneity and horizontal pleiotropy of the selected instrumental variables,excludes weak instrumental bias by F-statistics,and determines their direction of causality by reverse MR exploration.The heritability and genetic correlation of each trait are measured by Cross-trait linkage disequilibrium score regression(LDSC).And the sample overlap is examined by the intercept term.For multiple exposures with potential correlation,if they all have significant causal effects on the outcome,a multivariable mendelian randomization(MVMR)analysis method is selected for correction and testing.Finally,a colocalization analysis approach,focusing on the chromosome level and testing for genetic regions,is adopted to provide additional evidence for a causal association between exposure and outcome.Cardiovascular diseases(CVD)and hypertension(HTN)are found to have a positive causal effect on COPD,increasing the risk of COPD by an average of 1.5%(p=0.0014)and 0.8%(p=0.0061),respectively.After correction for MVMR,CVD still exerts a positive causal effect on COPD(p=0.0264),with consistent results across MVMR methods,whereas the causal effect of HTN on COPD is more influenced by MVMR methods under different modeling,and therefore more studies are needed to investigate the causal association.It is also found that COPD has a positive causal effect on diabetes mellitus(DM)after correction and attenuation for heterogeneity and horizontal pleiotropy in the instrumental variables,with an average 5.52-fold increased risk of developing DM in COPD patients(p=0.0028)compared to those without COPD.For psychiatric disorders,a positive causal association for depressive disorder(DD)and COPD is observed,with an average 1.2%increased risk of COPD in people with DD(p=0.0327)compared to those without DD.The findings suggest that there is no causal association between generalized anxiety disorder(GAD)and COPD(p=0.0854).LDSC regression excludes the possibility of sample overlap between the data groups and concludes that all traits are genetically correlated,while further colocalization analysis concludes that no genetic variation is shared between pairs of traits.Therefore,studying the causal relationship between COPD and its comorbidities through statistical methods and exploring the pathogenesis of COPD can also help to better prevent and treat COPD and reduce the disease burden.