1 INTRODUCTION

Non-alcoholic fatty liver disease (NAFLD), recently renamed metabolic dysfunction-associated steatotic liver disease (MASLD), has become the leading cause of chronic liver disease globally. The recently proposed change in nomenclature from NAFLD to MASLD responds to our better understanding of the pathophysiological processes that lead to the development and progression of MASLD (where metabolic dysfunction plays a key role in the pathophysiology of this common liver disease).¹ The global prevalence of MASLD is estimated to be around 38% among adult individuals and is projected to continue to rise with increasing global rates of obesity and type 2 diabetes.^1-3

MASLD poses a significant health and economic burden worldwide.^{4, 5} Substantial epidemiological evidence indicates that MASLD is a multisystem disease^{6, 7} that is associated not only with liver-related complications (cirrhosis, end-stage liver disease and hepatocellular carcinoma) but also with an increased risk of developing extrahepatic cancers (mainly non-liver gastrointestinal cancers)⁸ and cardiometabolic diseases, such as new-onset type 2 diabetes,⁹ fatal and non-fatal cardiovascular events¹⁰ and adverse renal outcomes.¹¹

Among the least investigated MASLD-related extrahepatic complications, an ever-increasing number of observational cross-sectional and prospective studies have assessed the association between MASLD and the risk of severe bacterial infections, such as pneumonia, meningitis, sepsis, gastrointestinal/abdominal infections or other bacterial infections requiring hospital admission (as extensively discussed below). However, the magnitude of this risk and whether the risk varies with the severity of MASLD remain uncertain.

Therefore, we have performed a comprehensive systematic review and meta-analysis of relevant observational studies to quantify the magnitude of the association between MASLD and the risk of serious bacterial infections requiring hospital admission. Moreover, we have also examined whether there was an association between the severity of MASLD and the risk of serious bacterial infections. Clarifying the magnitude of the risk of serious bacterial infections requiring hospital admission in patients with MASLD might favour the development of prevention strategies for serious bacterial infections in this patient population.

2 MATERIALS AND METHODS

3 RESULTS

3.2 Cross-sectional studies on the association between MASLD and serious bacterial infections

The distribution of cross-sectional studies (involving 26 434 377 adult individuals from different countries) by estimate of the association between MASLD and the risk of serious bacterial infections requiring hospital admission is plotted in Figure 1.^15-20 We found that MASLD was significantly associated with higher odds of serious bacterial infections (n = 6 studies; pooled random-effects odds ratio 1.93, 95% CI 1.44–2.58; I² = 93%). The serious bacterial infections requiring hospital admission considered in each eligible study are reported in Table 1. Since we have always used the fully adjusted OR estimates for each eligible study, this pooled random-effects OR was independent of age, sex, ethnicity, obesity, diabetes and other potential confounders (as specified in Table 1).

3.4 Prospective studies on the association between the severity of MASLD and the risk of developing serious bacterial infections

The distribution of prospective cohort studies by estimate of the association between the severity of MASLD and the risk of developing serious bacterial infections is plotted in Figure 2. Incidence rates of serious bacterial infections requiring hospital admission were further increased with more advanced liver disease, especially with higher fibrosis stage, that is, non-cirrhotic fibrosis or cirrhosis (n = 2 studies; pooled random-effects hazard ratio 2.42, 95% CI 1.89–2.29; I² = 92%). The ICD-9/ICD-10 codes used for diagnosing all subtypes of serious bacterial infections in each cohort study are specified in Supplementary Documents 1 and 2 in the online-only Supplementary Material.

3.6 Publication bias

As shown in Figure 3, the visual inspection and the Egger's regression test (although less than 10 studies were included) did not show any statistically significant asymmetry of the funnel plot for the included studies (p = .336), thus suggesting that the publication bias was low.

4 DISCUSSION

In this comprehensive meta-analysis that incorporated eight observational studies (six cross-sectional hospital-based studies and two population-based cohort studies) with aggregate data on ~26.6 million adult individuals from different countries (Israel, Croatia, Sweden and the USA), we found that MASLD was significantly associated with a ~ twofold higher prevalence of serious bacterial infections that required in-hospital or emergency department care (n = 6 studies; pooled random-effects odds ratio 1.93, 95% CI 1.44–2.58). The association between MASLD and the risk of having serious bacterial infections requiring hospital admission remained significant in those studies where statistical analysis was adjusted for age, sex, ethnicity, obesity, type 2 diabetes and other clinical comorbidities (as specified in Table 1). Meta-regression analyses did not show effect modification by age, sex, obesity or type 2 diabetes on the association between MASLD and the risk of serious bacterial infections. Furthermore, this association was more robust in studies where the diagnosis of MASLD was based on imaging techniques than in those using ICD codes. Meta-analysis of data from the two population-based cohort studies showed that MASLD assessed by liver biopsy or ICD codes was significantly associated with a ~ twofold higher incidence of serious bacterial infections requiring hospital admission (pooled random-effects hazard ratio 1.80, 95% CI 1.62–2.0), independent of age, sex, obesity, diabetes and other baseline clinical comorbidities. Notably, the incidence rates of bacterial infections were further increased with more advanced liver disease, especially with higher fibrosis stage, that is, non-cirrhotic fibrosis and cirrhosis (pooled random-effects hazard ratio 2.42, 95% CI 1.89–2.29).

This is the first and most comprehensive meta-analysis of observational studies examining the magnitude of the association between MASLD and the risk of serious bacterial infections requiring hospital admission. Concordant with the overall results on serious bacterial infections, all subtypes of bacterial infection showed a higher risk in patients with MASLD than in those without MASLD. As reported in Table 1, MASLD was associated with a substantially higher prevalence and incidence of serious bacterial infections, especially urinary tract infections, upper or lower respiratory tract infections, and gastrointestinal/abdominal infections, followed by sepsis, meningitis, peritonitis, and musculoskeletal, skin or connective tissue infections. As reported in the two Swedish cohort studies included in the meta-analysis, there was a graded positive relationship between the severity of MASLD and the risk of developing serious bacterial infections (with the highest risk observed for patients with advanced fibrosis or cirrhosis).^{21, 22} Specifically, in the Swedish nationwide cohort study by Ebrahimi et al.,²¹ the risk of serious bacterial infections increased progressively and independently with worsening MASLD histological severity over a median of 14.1 years (isolated steatosis (adjusted HR 1.64, 95% CI 1.55–1.73), non-fibrotic steatohepatitis [adjusted HR 1.84, 95% CI 1.60–2.12], non-cirrhotic fibrosis [aHR 1.77, 95% CI 1.56–2.0] and cirrhosis [adjusted HR 2.32, 95% CI 1.92–2.82]). Similarly, in the cohort study by Shang et al.,²² both non-cirrhotic MASLD (adjusted HR 1.90, 95% CI 1.7–2.0) and MASLD without diabetes (adjusted HR 1.90, 95% CI 1.8–2.1) were independently associated with higher risk of developing serious bacterial infections, although this risk was stronger for MASLD-related cirrhosis (adjusted HR 3.60, 95% CI 2.7–4.6). In this latter study,²² MASLD was also associated with a higher risk of infection-related mortality (adjusted HR 1.80, 95% CI 1.6–2.2). Based on subgroup analyses, these investigators also noted that attention should be paid to young and female patients with MASLD, whose liver disease may have a comparatively stronger adverse impact on the incidence of severe bacterial infections.²² The findings of our meta-analysis complement also the results of previously published meta-analyses showing that MASLD was associated with an increased risk of H. pylori infection and more severe COVID-19 infection.^{23, 24}

The findings of our meta-analysis may have some important clinical implications. These results apply to serious bacterial infections requiring in-hospital or emergency department care and, therefore, cannot be extrapolated to milder bacterial infections. That said, clinicians need to be aware of the increased risk in most, if not all, types of bacterial infections among patients with MASLD (especially among those with advanced fibrosis or cirrhosis) and should also have increased clinical vigilance for serious bacterial infections and consider preventive measures, such as regular checks of vaccination status.^{21, 22} Patients should be advised to seek medical attention in case of acute infections. With an approximately twofold increase in the risk of developing serious bacterial infections and infection-related death, patients with MASLD should receive broad-spectrum antibiotics if hospitalized for bacterial infection, with the choice of antibiotic therapy refined when culture test results become available.²⁵ Furthermore, given the greater risk of colonization by multidrug-resistant bacteria (MDR bacteria) in patients with liver failure, it could also be helpful to perform rectal and/or nasal swabs for MDR bacteria in the presence of serious infection. Future studies should examine whether drug treatment of MASLD will reduce the incidence rates of serious bacterial infections in this patient population. Although not specifically tested in the present meta-analysis, it is also reasonable to assume that patients with MASLD, especially those with advanced MASLD or with coexisting diabetes/obesity, are at increased risk of developing viral, fungal and parasitic infections. This implies that clinicians need to be aware that the risk of infectious diseases is high in this patient population and should be monitored especially in patients with MASLD-related cirrhosis and those who have coexisting type 2 diabetes or obesity.

A detailed discussion of the putative mechanisms underpinning the association between MASLD and the risk of serious bacterial infections (e.g. pneumonia, meningitis, urinary tract infections, gastrointestinal/abdominal infections, peritonitis and sepsis) is beyond the scope of the meta-analysis. Briefly, the liver is essential for host defence against infectious diseases. Bacterial infections are common among patients with cirrhosis (including spontaneous bacterial peritonitis) and may result in hepatic decompensation and liver-related mortality.²⁶ MASLD is closely associated with obesity and type 2 diabetes. While these are two metabolic diseases known to increase the susceptibility to bacterial infections (and other types of infections), neither obesity nor pre-existing type 2 diabetes modified the strength of the association in univariable meta-regression analyses we undertook. Other MASLD-related factors, such as systemic low-grade chronic inflammation and impaired immune system function, may also contribute to the increased susceptibility to bacterial infections in patients with MASLD.^{25, 27} Accumulating research also supports a role of intestinal barrier disruption and bacterial translocation from leaky gut in MASLD as an important contributing factor to development of bacterial infections and other infection types (viral, fungal or parasitic), both at the intestinal and extra-intestinal level.²⁸ However, the precise mechanisms and the interplay of the mentioned factors remain not fully understood. Further mechanistic studies are needed to better elucidate these issues.

Our meta-analysis has some important limitations, which are inherent to the design of the included studies. First, since most of the included studies had a retrospective cross-sectional design, we cannot establish a causal association between MASLD and serious bacterial infections requiring hospital admission. Second, although most eligible studies adjusted the results for age, sex, ethnicity, obesity, diabetes and other clinical comorbidities, the possibility of residual confounding by unmeasured factors cannot be entirely excluded. For instance, although the diagnosis of MASLD requires the exclusion of significant alcohol consumption, alcohol consumption is often under-reported in individuals classified as having MASLD (thus inducing a possible misclassification bias) and alcohol and metabolic factors may interact to exacerbate the progression of liver disease and the risk of developing bacterial infections.²⁹ Third, although we used a random-effects model, the interpretation of some results requires some caution because of the observed medium-high heterogeneity for the pooled primary analysis of cross-sectional studies and the moderate overall quality of the studies, suggesting a medium-high risk of bias according to the NOS. Detailed subgroup analyses by types of serious bacterial infection (that could partly explain the observed high between-study heterogeneity) cannot be performed. Fourth, although all eligible studies have used the NAFLD nomenclature, for this meta-analysis, we have assumed that NAFLD and MASLD are synonymous, as the two fatty liver disease nomenclatures share a very high level of overlap (>96% of individuals with NAFLD meet the diagnostic criteria for MASLD).^30-32 Fifth, an assessment of the severity of liver fibrosis (using non-invasive fibrosis biomarkers/scores) was not available for the eligible cross-sectional studies; it was only available for the two prospective cohort studies. Sixth, although a selective reporting bias of eligible studies could not be excluded, our comprehensive search has made it unlikely that any published studies were missed. Sixth, although in one large cross-sectional study by Patel et al. the diagnosis of NAFLD was based on ICD codes (so the liver disease could be underdiagnosed), it should be noted that the one-study remove (leave-one-out) approach to test the influence of each study on the overall effect size showed that eliminating of the study by Patel et al. (Figure S5) from the pooled analysis increased the magnitude of the association between MASLD and risk of serious bacterial infections (pooled random-effects odds ratio 2.11, 95% CI 1.55–2.88). Finally, these results derived from studies from high-income countries may not be directly transferable to low-income countries (where the risk of severe bacterial infections is higher) because poor communities were largely underrepresented in the meta-analysis. Further prospective cohort studies will be needed in low- and lower-middle-income countries.³³

Despite these limitations, the present meta-analysis also has important strengths. Our meta-analysis is the first and largest assessment of the association between MASLD and the risk of serious bacterial infections requiring hospital admission. The large number of individuals included in the meta-analysis (although the limited number of observational studies available so far) provided sufficient statistical power to quantify the magnitude of the association between MASLD and the risk of serious bacterial infections. Moreover, we have used standardized risk estimates from all included studies to allow a consistent combination of estimates across eligible studies. Finally, the visual inspection of the funnel plot did not reveal any significant asymmetry, thus suggesting that the risk of publication bias was low.

In conclusion, this comprehensive meta-analysis provides evidence of a significant association between MASLD and an increased risk of serious bacterial infections requiring hospital admission (such as pneumonia, meningitis, urinary tract infections, gastrointestinal/abdominal infections or sepsis). This association remained statistically significant in those studies whose results were adjusted for age, sex, ethnicity, obesity, type 2 diabetes and other important clinical comorbidities. Incidence rates of serious bacterial infections increased further with greater severity of MASLD, principally higher fibrosis stage. Future well-designed large prospective cohort studies from different countries (including Asian populations and low-income countries) are needed to further corroborate these results, and mechanistic studies are also required to better decipher the complex molecular pathways linking MASLD with increased susceptibility to developing serious bacterial infections requiring hospital admission.

AUTHOR CONTRIBUTIONS

All listed authors contributed greatly to this study and fulfilled the stated guidelines for authorship, as described by the published criteria for authorship by the ICMJE. Each author participated sufficiently in the work to take public responsibility for appropriate portions of the content. Below is a detailed account of each author's contribution: AM and GT were involved in the conception of the study and the analysis and interpretation of the results. GT wrote the first draft of the manuscript. RM, VF and MGL were involved in the conduct of the study and searched the published articles. AG, LM, FG, HT and CDB were involved in the interpretation of the results and contributed to the discussion. All authors edited, reviewed and approved the final version of the manuscript.

FUNDING INFORMATION

There was no funding source for this specific study. GT is supported in part by grants from the University School of Medicine of Verona, Verona, Italy. CDB is supported in part by the Southampton NIHR Biomedical Research Centre (NIHR203319), UK. This study was partly supported by the Italian Ministry of Health under Fondi Ricerca Corrente — Linea 1 (IRCCS Sacro Cuore - Don Calabria Hospital).

CONFLICT OF INTEREST STATEMENT

The authors have no potential conflicts of interest to disclose.

ETHICS STATMENT

This study involves human participants but was not approved by an Ethics Committee. Approval from an Ethics Committee is not necessary as this is a meta-analysis of published observational studies that had already obtained both informed consent from participants and ethical approval by their local Ethics committees.