Causal relationships between depression, emotional changes, and hiatal hernia: A Mendelian randomization analysis Authors: Chaofan Pan, Mingzhi Lin, Wenbin Luo, Ruoyun Li, Changjiang Luo Published: 2024 ABSTRACT Hiatal hernia (HH) is a common gastrointestinal disorder characterized by the displacement of abdominal contents, particularly the stomach, into the thoracic cavity. This condition is frequently associated with gastroesophageal reflux disease (GERD) and can lead to various symptoms, including chronic cough and respiratory issues. Despite its prevalence, the mechanisms linking psychological factors to HH are not well understood. Observational studies have suggested correlations between mental health issues – such as stress, anxiety, and depression – and gastrointestinal disorders, indicating that emotional states may influence the development of HH. This study aims to clarify the causal relationships between mood swings, depression, and the risk of developing HH using Mendelian randomization (MR), a robust method that utilizes genetic variants as instrumental variables (IVs) to infer causality. Data for this MR analysis were obtained from publicly available genome-wide association studies (GWAS). We employed a bidirectional, 2-sample MR approach, using IVs associated with mood swings, depression, feelings of tension, and feelings of misery as exposures, with HH as the outcome. A reverse MR analysis was also conducted, treating HH as the exposure and the aforementioned emotional states as outcomes. The primary analytical method used was inverse variance weighting (IVW), supplemented by sensitivity analyses, including MR-Egger and weighted median methods. Our analysis revealed significant associations: mood swings (OR = 1.014; 95% CI = 1.001–1.027; P = .032), depression (OR = 1.019; 95% CI = 1.006–1.033; P = .003), feelings of tension (OR = 1.012; 95% CI = 1.004–1.020; P = .001), and feelings of misery (OR = 1.007; 95% CI = 1.003–1.010; P = .0001) significantly increased the risk of HH. Importantly, reverse MR analysis indicated no causal influence of HH on these emotional states. This study provides evidence that mood swings, depression, feelings of tension, and feelings of misery are significant risk factors for developing HH. These findings highlight the need to address psychological factors in the clinical management and prevention strategies for HH, potentially improving patient outcomes. 1. Introduction A hiatus hernia (HH) is a structural condition in which abdominal contents, especially the stomach, push through the esophageal hiatus into the mediastinum. It has a high prevalence of up to 20% in endoscopic examinations, and HH is also the most prevalent form of diaphragmatic hernia.[1] HH, which stands for HH, refers to the upward displacement of the stomach or other organs into the chest cavity through the esophageal hiatus. It is divided into 4 subtypes according to the placement of the gastroesophageal junction (GOJ) and the involvement of additional organs.[2] Typically, HH can lead to gastroesophageal reflux and a range of extra-esophageal symptoms. According to the Montreal classification, these extra-esophageal symptoms include both definitively associated symptoms and those that are presumed to be associated. Symptoms definitively linked to the condition include reflux cough, reflux laryngitis, reflux asthma, and reflux-related dental erosion. Symptoms that are presumed to be associated include sinusitis, pulmonary fibrosis, pharyngitis, otitis media, and others.[3,4] The number of patients seeking medical attention for this condition is increasing in Western countries. Previously, it was thought to be less common in Asian regions. However, with economic development and changes in lifestyle habits, the incidence of this disease in Asia has significantly risen, attracting attention. According to current research, the prevalence of HH is increasing worldwide.[5,6] Mendelian randomization (MR) is a growing epidemiological technique for determining causality, utilizing genetic variation as instrumental variables (IVs). It has played a crucial role in identifying or ruling out particular risk factors for diseases.[7] With the rapid progress in genome-wide association studies (GWAS), MR studies have gained increased flexibility in dual-sample settings.[8] In the research on HH, this method has been comparatively underutilized, so we adopted it to identify the etiology that has not yet been determined to be causally related to HH. In current studies, the relationship between HH and mood changes, depression remains undetermined; however, most agree that patients experiencing anxiety or stress are more likely to develop the condition.[9,10] However, there is no definitive conclusion about the relationship between depression, mood changes, and the occurrence of HH, which may be due to the influence of confounding factors. Traditional observational studies are influenced by various confounding factors and may fail to accurately determine causal relationships, whereas MR (Mendelian randomization) studies reduce the risk of reverse causation bias because genetic variants are unaffected by disease status and are randomly allocated at conception. Our MR analysis uses genetic variants to estimate causality between mood changes, depression, and HH, employing a causal inference framework. This approach is appropriate for making causal inferences when randomized controlled trials (RCTs) are not feasible or practical.[11,12] In our paper, we use MR to examine whether genetically determined mood changes causally affect the risk of various forms of HH. We seek to elucidate their causal relationship. 2. Method 2.1. Mendelian randomization design Mood swings, feelings tense, feelings miserable, and depression are considered exposures, with HH as the outcome. Additionally, reverse MR was conducted where HH was the exposure and mood swings, feelings tense, feelings miserable, and depression were the outcomes, aiming to clarify reverse causal relationships. Information about the MR experimental process can be found in Figure 1. The Mendelian randomization study process used in the article. (A) Outlines the fundamental assumptions underlying the MR design. (B) Is the reverse MR design. MR = Mendelian randomization. When a single nucleotide polymorphism (SNP) achieves genome-wide significance (P < 5e−8), we select those SNPs for further analysis while considering linkage disequilibrium (r2 < 0.001). SNPs with overlapping exposures are excluded from subsequent analysis. The primary method for estimating the effect of exposures on outcomes is the Inverse Variance Weighted (IVW) method. Additionally, MR-Egger regression and the weighted median method are used as supplementary approaches. To address potential biases from heterogeneity and horizontal pleiotropy, multiple methods are employed. Cochran’s Q statistic is computed to evaluate heterogeneity.[13] To assess horizontal pleiotropy, 2 methods are employed, including MR-Egger intercept and MR-PRESSO.[14,15] We also use an F-value test, F = ((N − 2) * (R2/(1 − R2))), where N represents the sample size and R2 is the coefficient of determination between the instrumental variable and the exposure variable. Through the F-value test, we can select robust IVs, which not only enhances the accuracy of causal inference but also increases the credibility of study results.[16] 2.2. Data sources We define mood changes and adverse mood states as swings in emotions, feelings tense, feelings miserable, and depression experienced by patients. All exposure-level GWAS summary data are sourced from large-scale studies on European populations, obtained from open GWAS datasets (https://gwas.mrcieu.ac.uk/). The sources ieu, ebi, and ukb are maintained jointly by EMBL-EBI (European Bioinformatics Institute) and NHGRI (National Human Genome Research Institute). Their IDs are “ukb-b-14180,” “ebi-a-GCST006952,” “ebi-a-GCST006943,” and “ieu-b-102.”[17,18] The GWAS summary statistics for HH are also sourced from an open GWAS dataset with the ID “ebi-a-GCST90038691,” involving 484,598 Europeans, including 10,810 cases and 473,788 controls. All participants have a uniform genetic background typical of Europeans, which helps ensure that causal estimates are not skewed by ancestry-related heterogeneity.[19] Please note that the data in this study are not sourced from specific countries; all genome-wide association study (GWAS) data used are summary-level data. The phenotypic data and associated GWAS information are summarized in Table 1. Table 1All GWAS data sources and information used in this study.Phenotypes/IDData sourceStudy information/PMIDCases/controlsAuthor/yearsFeeling tenseebi-a-GCST006952European/29500382NANagel M/2018Feeling miserableebi-a-GCST006943European/29500382NANagel M/2018Mood swingsukb-b-14180European/NA204,412/47,207BenElsworth/2018Major depressionieu-b-102European/30718901170,756/329,443Howard DM/2018Hiatus herniaebi-a-GCST90038691European/3395972310,810/473,788NA/2021 All GWAS data sources and information used in this study. 2.3. Statistical analysis and data visualization To ensure the stability and reliability of the MR analysis, a series of sensitivity analyses were performed. The primary analytical method used was the IVW approach, which offers the most precise estimates when all instruments are valid. To address potential issues with invalid instruments, supplementary methods such as MR-Egger and weighted median were also applied, providing more robust estimates. This is important because invalid instruments can introduce heterogeneity and horizontal pleiotropy, which might bias MR estimates if only IVW is used. Data visualization techniques were employed to clearly and transparently present the results of these analyses.[15,20] Additionally, we used Cochran’s Q statistic to assess heterogeneity and applied MR-Egger intercept and MR-PRESSO to evaluate horizontal pleiotropy.[14,15] We conducted all tests using software packages within R version 4.3.1, including “TwoSampleMR” and “MR-PRESSO” from the “devtools” package. The results of the univariable MR analysis are shown in Figure 2 The forest plot denotes univariable MR analysis. OR is the odds ratio; 95% CI is the lower limit of 95% confidence interval; P is the P-value of OR; P for H is the P-value of heterogeneity test; P for P is the P-value of pleiotropy test. 3. Results 3.1. Hiatal hernia as the outcome Based on MR analysis using the IVW method, which provides the most effective numerical estimates, the study results indicate that mood swings, feelings tense, feelings miserable, and depression are risk factors for HH (IVW, OR = 1.014; 95% CI = 1.001–1.027; P = .032; IVW, OR = 1.012; 95% CI = 1.004–1.020; P = .001; IVW, OR = 1.019; 95% CI = 1.006–1.033; P = .003; IVW, OR = 1.007; 95% CI = 1.003–1.010; P = .0001). Two-sample MR analysis supports the causal role of these factors in the development of HH. Cochran’s Q test and MR-Egger intercept tests did not show significant heterogeneity or pleiotropy. Furthermore, to assess the impact of each SNP on the overall results, we conducted sensitivity analyses using leave-one-out methodology on positive results. This method involves systematically removing each SNP and observing its impact on the overall outcome. The analysis demonstrated that no single SNP had an undue influence on the results, ensuring the robustness of the study. For a more intuitive presentation of the MR analysis results, scatter plots were generated. For more detailed information and visual representations of the data analysis, please refer to Figure 3, which includes scatter plots for heterogeneity analysis, forest plots and funnel plots using the leave-one-out method. The results obtained are presented in Figure 3. The scatter plots exhibit the IVW regression direction tested by 5 methods. IVW = inverse variance weighted. 3.2. Hiatal hernia as the exposure Reverse MR analysis indicates that genetic susceptibility to HH is not associated with adverse mental states. Additionally, the results from MR-Egger were not statistically significant (P > .05) during the discovery phase, suggesting no reverse causal relationship exists. 4. Discussion This study used large-scale GWAS data to investigate the risk of HH associated with mood changes and adverse mood states. Generally, our MR analysis demonstrated a link between mental states and HH. Additionally, reverse MR analysis indicated that HH does not influence mental states. In current research on HH, the impact of swings in mental factors and adverse psychological states remains inconclusive. However, in this study, we observed that prolonged exposure to mood swings, tension, and similar adverse mental states could increase the risk of developing HH. Studies on gastroesophageal reflux disease (GERD) associated with HH have shown that socioeconomic status, occupation, and psychological factors are linked to GERD.[21,22] We hypothesize that these factors contribute to the increased risk of HH by potentially causing chronic gastrointestinal hyperactivity associated with adverse psychological states. Additionally, these factors may affect the clearance ability of the LES, thereby exerting a significant influence on the development of HH.[23,24] Existing studies indicate that patients who are in a prolonged state of anxiety and similar conditions experience an increase in gastric acid levels.[25,26] Furthermore, elevated gastric pH increases the risk of developing HH, and this risk is synergistically heightened by prolonged periods of inadequate rest.[21,26,27] Depression is also closely associated with an increased risk of HH. Depression can have detrimental effects on both psychological and physiological factors, potentially leading to HH. However, according to our current literature review and MR results, the complications associated with depression do not appear to affect the incidence of HH.[28] For most people, HH can largely be avoided or managed through lifestyle changes, particularly by adopting healthy habits and maintaining good mental well-being. Both physiological and psychological factors significantly influence the incidence and prognosis of HH.[29,30] And individuals who maintain mood stability year-round, without experiencing depression, are generally more inclined and better able to take proactive measures to maintain their physical health. This includes regular checkups and adjusting lifestyle habits promptly upon clinical symptoms to reduce their risk of developing HH. To our knowledge, no research has yet investigated the impact of HH on mental states. Our study suggests that HH does not significantly alter an individual’s mental state. This finding may be explained by the fact that mental states are primarily influenced by lifestyle, work environment, and genetics, with HH contributing minimally to these factors. Additionally, our analysis of data from European databases indicates no significant association between the severity of HH and incidence rates. This lack of association could be due to Europe’s advanced medical infrastructure and differing lifestyle habits compared to developing countries. This study has several strengths: Each GWAS had a relatively large sample size, providing sufficient statistical power; and MR is a robust method for causal inference that minimizes the risk of reverse causation. However, there are important limitations to consider horizontal pleiotropy remains an inherent challenge in MR studies. Despite employing various methods to address this issue, it may still subtly influence our results; our analysis revealed no significant differences in outcomes among control subjects, which helps mitigate concerns about selection bias; all participants in the GWAS were of European descent, which may limit the generalizability of our findings to other populations. Future research should aim to include diverse cohorts to validate the universality of this causal relationship; previous studies suggest that patients with psychiatric disorders often exhibit abnormal BMI and body fat percentages, factors that could also influence the occurrence of HH and may relate to dietary habits; and finally, the absence of individual-level data restricts our ability to assess specific causal relationships between mental states and HH, and we lack information regarding participants’ statuses concerning gastroesophageal reflux disease. This lack of detailed data precludes a thorough mediation analysis, and we cannot rule out potential indirect causal relationships between gastroesophageal reflux and the risk of HH, warranting further exploration. In conclusion, while our MR analysis effectively evaluates the direct causal link between mental states and the risk of developing HH, the limitations outlined highlight the need for cautious interpretation of our findings and suggest avenues for future research. 5. Conclusion In summary, this study utilized aggregated genetic data to investigate the complex causal relationship between mood swings, adverse psychological states, and the risk of HH. Our research indicates that these adverse factors influence the occurrence of HH. Current studies suggest that prolonged exposure to conditions like depression may decrease esophageal acid clearance, gradually affecting lower esophageal sphincter (LES) pressure and contributing to the formation of HH. Therefore, improving mental well-being could be beneficial in preventing HH. These findings provide new insights for developing strategies to prevent HH. Acknowledgments All individuals mentioned in this section have provided their consent to be named. Author contributions Conceptualization: Mingzhi Lin. Data curation: Mingzhi Lin. Formal analysis: Wenbin Luo. Funding acquisition: Changjiang Luo. Investigation: Wenbin Luo. Resources: Ruoyun Li. Software: Ruoyun Li. Supervision: Changjiang Luo. Validation: Ruoyun Li, Changjiang Luo. Visualization: Ruoyun Li. Writing – original draft: Chaofan Pan. Writing – review & editing: Chaofan Pan.