No cause-effect relationships between psychiatric disorders and amyotrophic lateral sclerosis: a bidirectional Mendelian randomization study

Abstract

Background Previous reports have been inconsistent about the associations between amyotrophic lateral sclerosis (ALS) and psychiatric disorders. More importantly, it needs to be elucidated whether these associations are causal or confounded. Objective We aimed to determine the presence of cause-effect relationships between ALS and four psychiatric disorders (schizophrenia; bipolar disorder; anxiety, nerves, tension or depression disorders (ADD); and neuroticism) using a bidirectional Mendelian randomization (BDMR) analysis based on gene associations. Methods First, we extracted genome-wide association studies (GWAS) summaries for the four psychiatric disorders and ALS from the Integrative Epidemiology Unit (IEU) and the Psychiatric Genomics Consortium (PGC) GWAS database. We then identified single-nucleotide polymorphisms (SNPs) that are strongly associated with exposure, are independent of confounders, and are related to exposure-outcome, as instrumental variables (IVs) for Mendelian randomization (MR) analyses. In the MR analysis, the cause-effect relationships were analyzed using psychiatric disorders as the exposure variable and ALS as the outcome, with six methods (primary inverse-variance weighted (IVW), MR-Egger, weighted median estimator, simple mode, weighted mode, and robust adjusted profile score (RAPS)analyses). Moreover, the results were subjected to sensitivity analyses, namely, IVW and MR-Egger analyses (for potential heterogeneity and horizontal pleiotropy) and MR leave-one-out analysis (to determine whether one SNP drove the causal signal). Finally, reverse MR analyses were conducted using ALS as the exposure variable and four psychiatric disorders as outcome variables; sensitivity was assessed. Results After extracting the GWAS summaries, we established IVs, including 147, 13, 37, and 105 SNPs for schizophrenia, bipolar disorder, ADD, and neuroticism, respectively. Subsequently, the MR analyses conducted using the six methods revealed no cause-effect relationships of the four psychiatric disorders with ALS. Moreover, we set up six SNPs as IVs for ALS in reverse MR analyses. Finally, no statistically significant cause-effect relationship of ALS with the four psychiatric disorders was found. These associations were robust as shown by the sensitivity analyses. Conclusions BDMR analyses revealed no significant genetic evidence for a cause-effect relationship between ALS and four psychiatric disorders.