Depression Promotes Gastroesophageal Reflux Disease: New Evidence Based on Mendelian Randomization

Background: Although observational studies have reported that depression is a risk factor for gastroesophageal reflux disease, it is difficult to determine the potential causal correlation. Thus, this study investigated the causal relevance of depression for gastroesophageal reflux disease using Mendelian randomization and provided new evidence for their association. Methods: Based on data from the UK Biobank, we assessed the causality of the 2 diseases by analyzing 135 458 severe depressive disorder cases and 41 024 gastroesophageal reflux disease cases. The causal inference was assessed using inverse-variance weighting, weighted median, Mendelian randomization–Egger, and weighted median methods. Simultaneously, pleiotropy and sensitivity analyses were used for quality control. Finally, we also explored whether depression affects gastroesophageal reflux disease through other risk factors. Results: A positive causal relationship between depression and gastroesophageal reflux disease was found in the inverse-variance weighted and weighted median methods, both of which were statistically significant [odds ratio = 1.011, 95% CI: 1.004-1.017, P  = .001; odds ratio = 1.011, 95% CI: 1.004-1.020, P  = .002)]. Sensitivity analyses were consistent with a causal interpretation, and the main deviation of genetic pleiotropy was not found (Intercept β = 0.0005; SE = 0.005, P  = .908). The genetic susceptibility to depression was also associated with smoking, insomnia, and sleep apnea (odds ratio = 1.166, 95% CI: 1.033-1.316, P  = .013; odds ratio = 1.089, 95% CI: 1.045-1.134; and odds ratio = 1.004, 95% CI: 1.001-1.006, P  = .001, respectively). Conclusion: Our results verified a causal correlation that depression could slightly increase the risk of gastroesophageal reflux disease.


INTRODUCTION
Gastroesophageal reflux disease (GORD) is a common disease caused by the reflux of gastric contents into the esophagus, causing discomfort symptoms and/or complications inside and outside the esophagus. 1,2 Several observational studies have reported that genetic predisposition, smoking, drinking, insomnia, sleep apnea, hypertension, type 2 diabetes, obesity, and psychological abnormalities are common risk factors for developing GORD. 1,[3][4][5][6] Major depressive disorder (MDD) is a common psychological disease characterized by depression, decreased interest, and impaired cognitive function. 7 Both the incidence and prevalence of depression and GORD have been increasing, eventually becoming a global health concern in recent years. 8,9 Previous studies have found that depression and GORD are closely related and that there is an interaction between the 2 diseases. 10-13 Systematic reviews have shown that the severity and frequency of GORD symptoms are closely related to the severity of depression. 14 However, there are also unanswered key questions in depression and GORD, particularly regarding potential mechanisms underlying this comorbidity. Currently, it is unclear whether the association between GORD and depression comes from common genetic or environmental factors and whether these associations are causal.
Mendelian randomization (MR) with genetic association is a new epidemiological tool for assessing causality that cannot be confused with other risk factors. 15 This method can provide strong causal evidence for the research and can effectively avoid the deviation caused by observational research. 16 An MR study has been used to evaluate the effect of higher central adiposity on GORD, indicating that a higher waist-hip ratio increases the risk of GORD. 17 In this study, we evaluated the potential causal role of depression in GORD and explored the potential mediators of the causal association between the two.

MATERIALS AND METHODS Genetic Variants Associated with Major Depressive Disorder
A meta-analysis of MDD in the genome-wide association studies (GWAS) was applied as an exposure set in this study, with 135 458 MDD cases and 344 901 controls among participants. 18 The MDD cases were diagnosed using the Diagnostic and Statistical Manual of Mental Disorders-IV or the International Classification of Diseases-10 (ICD-10). Based on strict inclusion criteria (P < 5 × 10 −8 ), Wray et al 18 reported that 44 single nucleotide polymorphisms (SNPs) played a significant role in MDD (Table S1). The 0.23% change in MDD was explained by these associated SNPs, and the instruments with large F-statistics (value of 156) can strongly predict GORD. 19

Genome-Wide Association Study Summary Data on Gastroesophageal Reflux Disease
Gastroesophageal reflux disease's genome-wide association study summary data were from the UK Biobank (https ://ww w.ukb ioban k.ac. uk/). 20 The GORD cases were defined by medical record ICD10. Totally, there were 41 024 GORD cases and 410 073 controls. There were 15 SNPs removed from the 44 SNPs in MDD, including palindromic with intermediate allele frequency and the result set lacked.

Summary Data of Gastroesophageal Reflux Disease Risk Factors
The association of MDD with previously reported risk factors for GORD was assessed by inverse variance weighting (IVW) to explore whether depression affects GORD through other risk factors. Smoking, drinking, body mass index, daytime sleeping, sleep lessn ess/i nsomn ia, sleep apnea, hypertension, and type 2 diabetes were considered as possible mediating factors of GORD. The data of these factors were obtained from several widely used databases (Table 1).

Statistical Analyses
According to MR guidelines, the IVW method has maximum statistical power when the instrumental variables (IVs) are all effective. Therefore, it was mainly used to evaluate the causal relationship between MDD and GORD risk. 21 If at least 50% of the weights in the analysis is from IVs, the weighted median method provides a causal estimation. 22

RESULTS
A total of 29 SNPs associated with MDD were used. Patients with MDD were 1.011-fold more likely to have GORD (95% CI: 1.004-1.017, P = .001) than those without MDD. In addition, a similar positive result was also found in the weighted median method (odds ratio [OR] = 1.011, 95% CI: 1.004-1.020, P = .002) (Figure 1). At the same time, the MR-Egger method and weighted mode obtained similar risk estimates, but neither was found to be statistically significant.
The MR-Egger method is usually used to test whether multiple IVs have horizontal pleiotropy. No significant intercept was found in this study (intercept β = 0.0005; SE = 0.005, P = .908), indicating the lack of horizontal pleiotropy. (Figure 2). The funnel plot can intuitively show the heterogeneity of SNPs; the funnel plot in this study was symmetrical as a whole, showing that there was no pleiotropy ( Figure 3). The omission sensitivity analysis showed that the MR results were actually robust and not affected by the SNP (Figure 4).
Furthermore, the IVW also assessed the relationship between depression and previously reported risk factors for GORD to explore the potential factors that serve as mediators (

DISCUSSION
In this study, the causal relationship between depression and GORD was verified using publicly available summary statistics from several genetic consortia. The MR analysis showed that depression could slightly promote GORD. This result was conducive to a clear understanding of the relationship between depression and GORD, which may contribute to better prevention and treatment of GORD.   Observational studies have shown that depression can lead to an increased risk of reflux symptoms. 25,26 Studies have shown that mental disorders are associated with the onset of reflux esophagitis, but no correlation was found between the severity of depression and the severity of reflux esophagitis. 27 In addition, a study showed that patients with anxiety and depression had a 2.8-fold higher risk of GORD than those without anxiety and depression. 25 In one study, patients with GORD with underlying mental disorders had postoperative satisfaction of 11.1%, compared with the 95.3% in patients without these diseases. 28 Therefore, depression is usually considered to be a potential factor of GORD in observational studies. Unfortunately, due to the inevitable limitations of observational studies, the causal relationship between these diseases is difficult to be confirmed. Based on this background, MR analysis may be an important causal reasoning research tool. Through MR analysis, our results supported the hypothesis that depression is a risk factor for GORD.
It should be noted that the pathophysiological mechanism of the relationship between depression and GORD remains unclear. Existing studies point to the following possible explanations. Due to the influence of psychological factors, patients with depression can reduce the sensory threshold of the body and increase the sensation of esophageal stimulation. 29,30 Studies have also shown that there is a close relationship between the brain and the gastrointestinal tract. 31 Patients with depression may have increased lower esophageal sphincter relaxation and then aggravated GORD. 32 The mechanism of psychological factors affecting reflux symptoms has also been confirmed in animal studies. The damage to esophageal epithelial tight junction in stressed rats leads to the decreased functionality of the esophageal mucosal barrier, which then increases the vulnerability of the esophagus to reflux. 33  Our study had several limitations. First, the related GWAS data were from populations of European ancestry, and GORD-related individuals were middle-aged and elderly people in the UK. Second, more result datasets related to GORD in MR analysis had not been further verified. Third, there were multiple subtypes in patients with depression and GORD, and further subgroup analysis may be necessary.

CONCLUSIONS
This study demonstrated that there is a causal relationship between depression and GORD and that depression could slightly increase the risk of GORD.