Gastroesophageal reflux disease (GERD) is characterized by symptoms or complications resulting from the reflux of acidic stomach contents into the esophagus. Approximately 20% of adults in the Western world are estimated to experience GERD (Maret-Ouda et al., 2020). The comorbidities between GERD and various neurodegenerative and psychiatric disorders have been widely reported (Gau et al., 2022; Jansson et al., 2007; Kessing et al., 2015; Zhang et al., 2012). It is essential to clarify the causality between GERD and the neurodegenerative and psychiatric disorders, which might contribute to better diseases prevention and management. However, most of these studies failed to address the issue of medical surveillance bias that might affect the results, and observational studies cannot fully explain the causality because of possible residual confounders and reverse causation. Mendelian randomization (MR) analyses use genetic variants correlated to exposure as instrumental variables (IVs) to identify the potential causality of risk factors related to outcome (Sanderson et al., 2022). Since genotypes cannot be altered by disease, this method could minimize the bias caused by confounding or reverse causality. Previous MR studies have only evaluated the causal relationships between insomnia, depressive disorder, anxiety disorder, and GERD (Chen, D. et al., 2023; Chen, G. et al., 2023; Zeng et al., 2023; Zhao et al., 2023). Nevertheless, several limitations may undermine the validity of their findings, such as the sample overlap problem, imprecise estimates caused by the low power in sensitivity analyses, and misleading findings due to the very few single nucleotide polymorphisms (SNPs) used in the MR study. More importantly, previous MR studies were unable to stratify the analyses by different subtypes of the diseases, which vary greatly in physiopathology involving both genetic and environmental factors. Stratification by drug usage is also necessary but has yet to be addressed by previous MR studies. For instance, proton pump inhibitors (PPIs), which are commonly prescribed medications for managing GERD, have been implicated in various preclinical studies for potentially elevating Aβ levels, interacting with tau protein, and impacting the neuronal microenvironment through multiple mechanisms (Ortiz-Guerrero et al., 2018). Thus, PPI usage might act as an important confounder affecting the causality between GERD and Alzheimer's disease (AD). Similarly, benzodiazepine hypnotics and sedatives used for treating psychiatric disorders also have the effect of muscle relaxation (e.g., diazepam), which might contribute to GERD risk. Therefore, benzodiazepine hypnotics and sedatives might be confounders between psychiatric disorders and GERD.

Linkage disequilibrium score regression (LDSC) is a powerful method used for genetic correlation analysis (Bulik-Sullivan et al., 2015), providing insights into the shared genetic foundations of distinct traits. By examining the relationship between test statistics and LDSC, the contribution of each genetic element can be quantified. This approach also enables estimation of heritability and genetic correlation from genome-wide association study (GWAS) summary statistics, helping to differentiate between real polygenicities and mixed biases (e.g., implicit association and population stratification) (Bulik-Sullivan et al., 2015). In the present study, LDSC analysis based on GWAS summary statistical data was used to assess the genetic correlation of 6 neurodegenerative and psychiatric disorders with GERD as affected by heredity. We also conducted bidirectional MR analysis to determine the causality between GERD and 6 neurodegenerative and psychiatric disorders, including AD, amyotrophic lateral sclerosis (ALS), depressive disorder, anxiety disorders, bipolar disorder (BD), and insomnia. Additional subgroup analyses according to subtypes were also performed in AD (subtypes: early-onset AD, late-onset AD, and dementia in AD) and anxiety disorders (subtypes: generalized anxiety disorder and phobic anxiety disorders). To further discover the gene regulation mechanisms underlying these diseases, we applied a range of bioinformatics analyses to investigate the hub shared genes, the common pathways related to the pathogenesis of GERD with 6 neurodegenerative and psychiatric disorders. A flow chart of this study design is presented in Fig. 1.