Introduction

Diverticular disease is a prevalent condition in Western nations, with its prevalence steadily rising alongside the aging demographic. The United States, Western Europe, and Australia exhibit the highest rates of incidence, particularly among individuals aged 60 and older, reaching levels as high as 50%. Nonetheless, contemporary research suggests a growing occurrence of diverticular disease among younger patients1.The manifestations of diverticular disease are characterized by nonspecific signs and symptoms, such as abdominal pain, bloating, and alterations in bowel habits. The spectrum of severity ranges from a benign, self-limiting condition to potentially fatal complications2,3.The etiology of diverticular disease remains uncertain, with diet and environmental factors being identified as primary risk factors. However, epidemiological research has demonstrated that genetic predisposition also significantly contributes to the pathogenesis of the disease4,5.

Liver cirrhosis is a condition characterized by liver damage, inflammation, and fibrosis, ultimately leading to liver failure. It is widely recognized as a terminal illness in medical practice and is a prevalent contributor to rising rates of incidence and mortality worldwide6,7.

Research has shown that the presence of liver cirrhosis significantly impacts the prognosis of patients with diverticular disease, leading to a higher likelihood of postoperative complications in this population8.Andrew J. Kruger and colleagues conducted a nationwide sample survey of hospitalized patients diagnosed with diverticulitis between 2007 and 2013, examining the impact of comorbid liver cirrhosis on mortality rates. The findings of the study revealed thatpatients with liver cirrhosis faced a higher risk of mortality during treatment for diverticulitis compared to those without liver cirrhosis. This discovery highlights the significance of accurately diagnosing and treating liver cirrhosis in individuals with diverticulitis, as the presence of liver cirrhosis may elevate the likelihood of complications, consequently influencing the overall health status and survival rate of patients[9].Prior research has suggested a correlation between the two variables, however, the precise mechanisms and causality are not fully understood. Consequently, further investigation is warranted. Mendelian Randomization (MR) is an epidemiological technique that employs genetic variation as instrumental variables to evaluate the causal link between exposure factors (e.g. lifestyle, environmental factors, biomarkers) and health outcomes (e.g. diseases)9.In accordance with Mendelian genetics principles, the stochastic distribution of alleles effectively mitigates the potential confounding and reverse causation biases that pose challenges to quantification within conventional epidemiological studies10.At present, there is a lack of Mendelian Randomization (MR) studies investigating the potential causal link between diverticular disease and liver cirrhosis. Recent Genome-Wide Association Studies (GWAS) have uncovered various genetic loci linked to both conditions, paving the way for more rigorous MR analyses.

In this study, we employed the most recent Genome-Wide Association Study (GWAS) summary-level data to perform a two-sample Mendelian Randomization (MR) analysis with the objective of elucidating the causal correlation between diverticular disease and liver cirrhosis.

Materials and methods

Mendelian randomization

Mendelian Randomization involves the utilization of instrumental variables to evaluate potential causal correlations between an exposure and an outcome11, it is assumed that the genetic variants chosen for their correlation with diverticular disease are not influenced by confounding variables or alternative pathways, with the exception of the exposure factor and liver cirrhosis12, as shown in Fig. 1.This study employed two-sample single nucleotide polymorphisms (SNPs) from the Genome-Wide Association Study (GWAS) database as the foundation for analysis in order to evaluate the causal correlation between diverticular disease and liver cirrhosis. Sensitivity analysis was performed to assess the robustness of the data.

Fig. 1
figure 1

The three corresponding main hypotheses in Mendelian randomization studies.

Full size image

Data source

This research employed data acquired from extensive Genome-Wide Association Studies (GWAS) to investigate the relationship between diverticular disease and liver cirrhosis. Specifically, the data for diverticular disease were extracted from FinnGen R10 (finngen_R10_K11_DIVERTIC), encompassing a sample size of 363,000 individuals and 19,344,832 single nucleotide polymorphisms (SNPs), and this population comprised individuals of Finnish ancestry. The diagnosis of diverticular disease was confirmed based on standardized clinical diagnostic criteria outlined by the American Gastroenterological Association. This included the identification of characteristic symptoms, such as abdominal pain, alongside imaging findings from CT scans that confirmed the presence of diverticula. The data for liver cirrhosis were obtained from the IEU Open GWAS project database, comprising a sample size of 347,406 individuals and 19,079,888 SNPs13,this dataset included participants primarily of European ancestry.The diagnosis of liver cirrhosis was established by adhering to standardized clinical criteria, which involved comprehensive clinical assessments, relevant laboratory tests (including liver function tests), and imaging studies such as abdominal ultrasound, consistent with the guidelines provided by the World Health Organization.Only SNPs with complete genotype data and confirmed diagnoses of diverticular disease and cirrhosis were included. Individuals with incomplete data, non-European ancestry were excluded.All data used in this study were obtained from publicly available databases, thereby eliminating any ethical approval considerations, As shown in Table 1.

Table 1 Data sources.
Full size table

Selection of instrumental variables

We excluded genetic variants (single nucleotide polymorphisms, SNPs) with missing data from the genome-wide association study (GWAS) database. Subsequently, we selected SNPs that were significantly associated with the trait of interest as instrumental variables, employing the following criteria: p-value threshold (p1) of 5 × 10^-8, linkage disequilibrium threshold of 0.001, and a region width of 10,000 kilobases. To meet the requirements of the third hypothesis, we set the p-value threshold to less than 5 × 10^-5 to filter out SNPs that were significantly associated with the outcome.We then utilized the Phenoscanner database14 to identify and exclude potential confounding factors.

Furthermore, we assessed the genetic instrument strength of all single nucleotide polymorphisms (SNPs) by calculating the F-statistic, and subsequently focused on instrumental variables with an F-statistic exceeding 10. The F-statistic is determined by the formula F = R^2(N-2) / (1 - R^2), where R^2 denotes the proportion of variance elucidated by the SNPs associated with the exposure factor, and N represents the sample size of the genome-wide association study (GWAS) for the exposure factor. The formula for calculating R^2 in this study is expressed as [2 × MAF × (1 - MAF) × beta^2] / [2 × MAF × (1 - MAF) × beta^2 + 2 × MAF × (1 - MAF) × N × sx(beta)^2], with MAF denoting the minor allele frequency and beta representing the effect size of the single nucleotide polymorphism on the exposure factor15.

Mendelian randomization analysis

A two-sample Mendelian randomization study used the inverse variance weighted (IVW) method16 to explore the causal link between diverticular disease and cirrhosis, chosen for its strong statistical power when pleiotropy was absent. To validate the findings, seven additional methods were employed, including MR-Egger regression (MR-Egger)17 to adjust for systematic bias in the instrumental variables. The weighted median (WM)18 method provides robust, bias-resistant estimates using instrumental variables. The Weighted mode technique19 ensures consistency in estimates. The constrained maximum likelihood (cML-MA) approach20 addresses heterogeneity in instrumental variable effects via constrained maximum likelihood estimation. The contamination mixture (ConMix) method21 handles instrumental variabe heterogeneity using a mixture model.The the Robust adjusted profile score (MR-RAPS)22 method reliably estimates causal effects, especially with effect heterogeneity, while the Debiased inverse-variance weighted method(DIVW) method23 effectively reduces bias in instrumental variables, making it ideal for heterogeneous instrument effects. Cochran’s Q value was calculated to assess heterogeneity in the causal correlation. Horizontal pleiotropy was assessed using the intercept term of the MR-Egger method. with a P-value exceeding 0.05 demonstrating the lack of horizontal pleiotropy within the chosen SNPs. Additionally, a global test using the MR-PRESSO method was performed to further evaluate the presence of horizontal pleiotropy and identify any potential outliers.Sensitivity analysis was performed utilizing the leave-one-out method, in which each single nucleotide polymorphism (SNP) was systematically excluded and the inverse variance-weighted (IVW) analysis was subsequently re-executed. All analyses were performed using the TwoSampleMR (version 0.6.0), MendelianRandomization (version 0.8.0), and MRPRESSO (version 1.0) packages in R software version 4.3.3 (https://www.R-project.org).

Power calculations

To determine the sample size required for this Mendelian Randomization study, we conducted power calculations. Based on prior literature and preliminary findings, we estimated that a sample of 847,200 participants would be necessary to achieve 80% statistical power (β = 0.20) at a significance level of 0.05 (α). However, our study was limited to 122 cases and 34,284 controls, resulting in an actual power of only 8%. This low power significantly increases the risk of false negatives and may lead us to overlook a genuine effect between the exposure and the outcome.

Despite this limitation, we implemented optimized analytical methods to mitigate the influence of confounding factors. We made considerable efforts to enhance the reliability of our findings through rigorous adjustments for potential confounders, aiming to provide meaningful insights even in the context of insufficient power.

Results

Instrumental variables

In the process of conducting a genome-wide association study (GWAS) database search, a total of 58 single nucleotide polymorphisms (SNPs) were initially identified in the outcome data. Following the removal of SNPs in the outcome data that showed significant correlations with exposure (p-value < 5 × 10^-5), the number of outcome SNPs remained at 58. Subsequent harmonization of the outcome data, which involved removing data that did not conform to a palindrome pattern, resulted in a total of 56 SNPs, with the exclusion of SNP rs2337106. Notably, no outliers were detected in the dataset. A thorough examination of the Phenoscanner database revealed that no SNPs were excluded. All SNPs satisfied the criteria of independence and exclusivity, as evidenced by F-statistics exceeding 10, thereby confirming the reliability of the findings.

The analysis results of mendelian randomization

The primary analysis was conducted using the Inverse Variance Weighted (IVW) method. To complement and enhance the reliability of the results, we conducted seven other MR methods.IVW (OR = 0.849; 95%CI: 0.743 ~ 0.971; P = 0.016), MR-Egger (OR = 0.678; 95%CI: 0.466 ~ 0.988; P = 0.048), WM (OR = 0.791; 95%CI: 0.650 ~ 0.962; P = 0.019), Weighted mode (OR = 0.766; 95%CI: 0.603 ~ 0.973; P = 0.033), ConMix (OR = 0.856; 95%CI: 0.740 ~ 0.990; P = 0.036), MR-RAPS (OR = 0.851;95%CI: 0.742 ~ 0.976P = 0.021), DIVW (OR = 0.846; 95%CI: 0.739 ~ 0.969; P = 0.016), cML-MA(OR = 0.852; 95%CI: 0.736 ~ 0.987; P = 0.033).The results of the Mendelian Randomization analysis can be found in Table 2, Table 3; Fig. 2.

Fig. 2
figure 2

Scatter plots assessing the causal relationship between Diverticular disease of intestine and Cirrhosis using eight methods.

Full size image
Table 2 The results of the four MR menthods.
Full size table
Table 3 Specific information on the 56 SNPs used for MR analysis.
Full size table

And the MR-PRESSO test did not identify any significant outliers, reinforcing the robustness of our Mendelian randomization results.

Sensitivity analysis

Cochran’s Q values for the Inverse Variance Weighted (IVW) and MR-Egger methods were calculated as 57.23 (P-value = 0.39) and 55.62 (P-value = 0.41), respectively, suggesting no significant heterogeneity. Consequently, based on the results of the Cochran’s Q test, we conducted the primary analysis using a fixed-effect IVW model in this study.Please refer to Fig. 3 for visualization. The intercept term of the MR-Egger regression was utilized to assess horizontal pleiotropy, yielding a P-value of 0.215 (> 0.05), suggesting the robustness of the chosen SNPs.The funnel plot exhibited a symmetrical distribution of the included single nucleotide polymorphisms (SNPs), indicating the absence of bias. Sensitivity analysis utilizing the leave-one-out method, wherein MR analysis is iteratively conducted by excluding one SNP at a time, further confirmed the robustness of the Mendelian Randomization findings. Please refer to Fig. 4 for details.

Fig. 3
figure 3

Funnel plot based on the results of the analysis. The horizontal coordinate mainly refers to the degree of variability and the vertical coordinate mainly refers to the total effect size. The positions of the values of the main effect sizes are marked with a line, and the points on the left and right sides can be distributed symmetrically, indicating the absence of bias.

Full size image
Fig. 4
figure 4

Results of “leave-one-out” sensitivity analysis in MR. MR results were calculated for the remaining SNPs after removing them one by one.

Full size image

Conclusions

Diverticular disease is characterized by the protrusion of weakened areas of the intestinal wall. The majority of individuals with this condition are asymptomatic, though symptomatic patients commonly exhibit symptoms such as diverticulitis or gastrointestinal bleeding24.Colonic diverticula are more prevalent in the sigmoid colon in Western countries, while in Asia, they are predominantly located in the right colon, a phenomenon believed to be influenced by dietary and genetic factors24.This disease has a global impact on the elderly population and has been confirmed to have a genetic susceptibility. Nevertheless, the genetic mechanisms underlying the disease have not been extensively investigated25.

As individuals age, there is a notable exacerbation of tissue inflammation and oxidative stress in liver cirrhosis. There is a growing body of evidence suggesting a potential link between diverticular disease and liver diseases. For instance, Abdurrahman Sahin et al.observed a significant decrease in liver steatosis among patients with diverticular disease undergoing colonoscopy examinations compared to those without the condition. The presence of diverticular disease may potentially serve as a protective factor against liver steatosis; however, the precise mechanisms responsible for this phenomenon are not yet fully understood26.

We examine the verification of the causal link between diverticular disease and liver cirrhosis from various viewpoints.

Chronic liver injury is characterized by the progression of liver fibrosis, characterized by the deposition of type I collagen within the liver parenchyma, leading to advanced fibrosis and ultimately resulting in the manifestation of liver cirrhosis27.The gut-liver axis represents a reciprocal interaction between the intestinal microbiota and the liver, impacting metabolic processes and signaling pathways via various factors including dietary intake, genetic predisposition, and environmental influences28.The disruption of the gut-liver axis and impairment of intestinal barrier function can result in dysbiosis of the gut microbiota, heightened bacterial translocation, and the activation of inflammatory factors, ultimately leading to the development of severe complications29,30,31.The gut microbiota and barrier function play a direct role in the development of compensated cirrhosis and are correlated with the severity and complications of decompensated cirrhosis28.

Study identified a deficiency of Bifidobacterium as a risk factor for cirrhosis, whereas an elevated abundance of Bifidobacterium was observed in patients with acute diverticulitis. The increased presence of Bifidobacterium in the gut can benefit cirrhosis patients by acidifying the gastrointestinal environment, enhancing the mucosal barrier, and improving immune regulation32,33,34,35,36. These findings provide evidence from the perspective of gut microbiota that the increase in Bifidobacterium observed in patients with diverticular disease may exert a protective effect against cirrhosis.

Furthermore, in patients with diverticular disease, encompassing both simple diverticular disease and acute diverticulitis, there is a notable upregulation in the expression levels of tumor necrosis factor-alpha (TNF-α). In the context of cirrhosis, TNF-α has been shown to downregulate the expression of type I collagen genes and directly counteract certain fibrogenic effects of transforming growth factor beta 1 (TGF-β1), thereby demonstrating an anti-fibrotic effect37,38,39,40,41.This cytokine-mediated mechanism may elucidate how the elevated levels of TNF-α observed in diverticular disease could contribute to an anti-fibrotic effect in cirrhosis.

This research utilized Mendelian Randomization techniques with eight distinct models to genetically assess the causal correlation between diverticular disease and liver cirrhosis. The findings indicated an inverse relationship between diverticular disease and liver cirrhosis, implying that diverticular disease may act as a protective factor against the development of liver cirrhosis. Diverticular disease, as a benign indication for colorectal surgery, may have a positive impact on patients with cirrhosis in certain situations. In clinical practice, the treatment of cirrhotic patients with concomitant diverticular disease should be based on an individual assessment, weighing the necessity and potential risks of surgery.For patients with cirrhosis and mild symptoms of diverticular disease, clinical intervention may be avoided to delay the progression of cirrhosis.

The link between diverticular disease and cirrhosis involves gut microbiota and cytokine levels, potentially leading to new therapeutic targets. These could pave the way for new treatments for cirrhosis, but further validation in medical practice is needed.Future research should prioritize the investigation of the precise mechanisms underlying their correlation and the translation of this knowledge into clinical practice to enhance patient outcomes.

This study demonstrates several notable strengths. Primarily, it breaks new ground in investigating the potential causal link between diverticular disease and liver cirrhosis, employing Mendelian Randomization analysis to evaluate the relationship between exposure variables and outcomes, effectively controlling for confounding variables and bolstering the evidence base. Furthermore, through a comprehensive analysis of data from both the Finnish database and the IEU database, this study effectively mitigates the impact of genetic and environmental factors, thereby improving the internal validity of the research, facilitating comparability of results, and minimizing bias resulting from sample overlap.Finally, this study presents initial evidence of the correlation between diverticular disease and liver cirrhosis, contributing to a more comprehensive comprehension of the interplay between these two medical conditions.This offers novel directions and insights for future scientific research and clinical interventions, facilitating a deeper understanding of the biological mechanisms underlying this negative correlation and enabling the exploration of innovative approaches for the treatment of cirrhosis.

The study is limited by several factors. Firstly, the data on exposure and outcomes are sourced from European populations, potentially introducing racial bias and limiting the generalizability of the findings. Secondly, Mendelian randomization depends on the validity of the instrumental variable assumptions; however, ensuring these assumptions are met can be challenging, particularly in the presence of pleiotropy within the instrumental variables. Furthermore, the sample size was suboptimal, leading to reduced statistical power and an elevated risk of Type II errors. While sensitivity analyses can provide insight into the robustness of our findings under certain conditions, they do not compensate for the limitations posed by insufficient sample size and the resulting implications for statistical power.In future research, it is imperative to incorporate samples from a broader range of ethnic backgrounds and to increase the sample size in order to enhance the generalizability and applicability of the findings.

In summary, this research substantiates the genetic basis for the causal correlation between diverticular disease and liver cirrhosis using Mendelian Randomization analysis, indicating that diverticular disease may confer a protective effect against liver cirrhosis. These findings offer novel perspectives on the clinical correlation between these diseases; however, additional investigation is warranted due to the limited support from clinical research evidence.