Mendelian randomization study of Alzheimer’s disease and lung cancer

Abstract

Background: Observational studies show that Alzheimer's disease (AD) and lung cancer is negatively associated in both directions. However, whether there is a causal relationship between AD and lung cancer remains uncertain. Mendelian Randomization (MR) is a novel approach for the assessment of causality, using genome‐wide significant single‐nucleotide polymorphisms (SNP) as instrument variables (IV) to eliminate the bias caused by confounders. In this study, we utilized a bi‐directional two‐sample MRanalysis to estimate the potential causality between AD and lung cancer. Methods: The genome‐wide association study (GWAS) summary data of AD and lung cancer were derived from the International Genomics of Alzheimer's Project (IGAP,17,008 AD cases and 37,154 controls) and the International Lung Cancer Consortium (ILCCO, 11,348 lung cancer cases and 15,861 controls), respectively. We performed a bi‐directional two‐sample MRanalysis to determine the causal relationship between AD and lung cancer. To be specific, Inverse‐Variance Weighted (IVW), MR‐Egger and Weighted Median methods were utilized for deriving causal estimates. Single SNP analysis and leave‐one‐out analysis were also conducted to identify if a single SNP is driving the association. We also performed the MR‐Egger regression and MR‐PRESSO analysis for the evaluation of horizontal pleiotropy. Results: There was no causal relationship between AD and lung cancer in bi‐directional MR. The results of Single SNP analysis and leave‐one‐out analysis indicated that no single SNP was driving the association. No directional horizontal pleiotropy was detected in the MR‐Egger regression analysis. The MR‐PRESSO analysis did not detect horizontal pleiotropy. There was no evidence for violations of MR assumptions in the sensitivity analysis. Conclusions: We found no causality between AD and lung cancer through the twosample MRanalysis. However, further study is warranted. (Table Presented).