Helicobacterpylori Infection and Metabolic Dysfunction-Associated Steatotic Liver Disease

To the Editor,

We read with considerable interest the article by Zheng et al. [1] on the association between H. pylori (Hp) infection and nonalcoholic fatty liver disease (NAFLD). Specifically, the authors performed a retrospective cohort study based on the records of the Department of Geriatrics, Union Hospital, Huazhong University of Science and Technology, Wuhan City, China: they retrieved data from individuals undergoing annual health checkups and managed to collect a large population (n = 2,063), aged 48.6 ± 14.9 years, without NAFLD at baseline (2011–2012). The diagnosis of Hp (exposure) was based on ¹³C-urea breath test (UBT) records, thus indicating an active Hp infection, and the diagnosis of new-onset NAFLD (main outcome) was based on abdominal ultrasonography records [1].

At baseline, 796 individuals (38.6%) were infected with Hp [Hp(+)], based on UBT. After a mean follow-up of 1.7 years, 332 individuals (16.1%) developed NAFLD; Hp(+) individuals had significantly higher rates (18.6%) of new-onset NAFLD compared with those without Hp infection [Hp(−)] at baseline (14.5%; p = 0.014); however, after adjustment for multiple potential confounders, Hp infection was reportedly not independently associated with NAFLD [1]. The authors also observed that age, body mass index (BMI), systolic and diastolic blood pressure, as well as circulating glucose, triglycerides, total cholesterol, low-density lipoprotein-cholesterol (LDL-C), alanine aminotransferase (ALT), and aspartate aminotransferase (AST) were positively associated, whereas high-density lipoprotein-cholesterol (HDL-C) was inversely associated with the development of NAFLD. Therefore, they inserted all these variables into the same Cox regression model to adjust the association of Hp infection with the development of NAFLD for all these variables. Nonetheless, by inserting into the same model ALT with AST, or systolic and diastolic blood pressure, or triglycerides and total cholesterol, LDL-C and HDL-C, there was high possibility of overadjustment; in essence, the association between Hp infection and the development of NAFLD is adjusted more than once for blood pressure, lipid profile, and liver function tests. Moreover, both ALT and AST are highly associated with NAFLD, which is the main outcome of the study. Thus, their insertion into the Cox regression model with NAFLD as outcome may also lead to overadjustment. Furthermore, it seems that the authors used either baseline or endpoint values of BMI, systolic and diastolic blood pressure, circulating glucose, triglycerides, total cholesterol, LDL-C, HDL-C, ALT, AST, which may also have distorted the results, because all these variables were reportedly associated with the development of NAFLD in the univariate analysis. Therefore, their deltas (endpoint value – baseline value), or an alternative way integrating the change of these variables over time, might have been inserted into the Cox regression model, instead of the respective baseline or endpoint variables. Considering all the above, it would be highly valuable, if the authors repeated the analysis by adjusting the association between Hp infection and the development of NAFLD for age, delta (BMI), delta (systolic blood pressure) (or delta [diastolic blood pressure]), delta (LDL-C) (or delta [total cholesterol] or delta [HDL-C] or delta [triglycerides]), and delta (glucose), while excluding ALT and AST from the model.

Another important issue that may have affected the results of this study is eradication therapy that may have been administered to some individuals who were Hp(+) at the baseline. Thus, the development of NAFLD in the long term might have been mitigated in those subjected to successful eradication therapy, i.e., those shifted from Hp(+) to Hp(−) during the follow-up. However, the authors reported that, according to the “Fourth Consensus on Hp Management in China,” the only indications strongly recommended for Hp eradication therapy were peptic ulcer and gastric muco-associated lymphoid tissue lymphoma until 2017, a period that includes their study, and that patients with gastric malignancies were excluded from their study [1]. Nonetheless, the patients with peptic ulcer were not excluded, so some of the initially Hp(+) individuals may have been subjected to eradication therapy. Vice-versa, some of the individuals being Hp(−) at the baseline might have been infected by Hp, i.e., being shifted from Hp(−) to Hp(+) during the follow-up; this swift might also have affected the results in the long term. Indeed, the high rates of Hp recurrence and re-infection in Asia, including China, should be considered. Therefore, in case that relevant data exist, it would be valuable, if the authors performed a sensitivity analysis with four groups: (1) Hp(−) at the baseline that remained Hp(−) during the follow-up; (2) Hp(−) at the baseline that were shifted to Hp(+) during the follow-up; (3) Hp(+) at the baseline that remained Hp(+) during the follow-up (because they did not receive eradication therapy or eradication therapy was unsuccessful); and (4) Hp(+) at the baseline that were shifted to Hp(−) after eradication therapy during the follow-up.

Regarding the UBT used for the diagnosis of Hp infection in the study by Zheng et al. [1], the authors claimed that it was performed at least 2 weeks after the discontinuation of proton pump inhibitors to avoid false-negative results, according to the “Fourth Consensus on Hp Management in China” [1]. However, false-negative results can occur due to a decrease in Hp gastric diversity, reported for antibiotics, bismuth compounds, and proton pump inhibitors. Reduced sensitivity occurs in the setting of active gastrointestinal bleeding and recent usage of bismuth-containing compounds, antibiotics, or antisecretory drugs. Therefore, it is also recommended to stop antibiotics and bismuth-containing compounds at least 4 weeks before UBT. Additionally, false-positive results can also occur in the existence of microbiome capable of producing urease, such as Helicobacter heilmannii, that is frequent in Chinese people, in contamination with oral flora and/or in achlorhydria.

The association of Hp infection with NAFLD, a highly prevalent disease, which was recently proposed to be renamed to metabolic dysfunction-associated steatotic liver disease (MASLD) [2], continues to be an intriguing topic. In 2012, we showed for the first time that NAFLD patients had higher rates of Hp IgG seropositivity compared with BMI-, sex- and age-matched controls [3]. The percentage of Hp eradication treatment history was also higher in patients with simple hepatic steatosis than in those with nonalcoholic steatohepatitis [3], currently renamed to metabolic dysfunction-associated steatohepatitis [2], which may imply a negative association of Hp eradication with the progression of the disease. Since then, many relevant studies have been published on this topic with variable and sometimes conflicting results, as summarized elsewhere [4]. However, most of the existing studies are cross-sectional or case-control [5], thus being low at the pyramid of evidence-based medicine. Therefore, there is high need for relevant cohort studies and, in this regard, the cohort study of Zheng et al. [1] may add to the literature. Furthermore, it is important to note that most relevant studies, including that of Zheng et al. [1], were conducted in Asia, whereas only a few of them in America or Europe, which makes apparent an issue of generalizability of their results. Additionally, the different methods of diagnosis of both Hp infection and MASLD in different studies make hard the comparison of each one with the others. It is of note that, by using the histological diagnostic gold standard methods for the diagnosis of both active Hp infection and MASLD, we previously showed that active Hp infection was independently associated with MASLD and its severity in morbidly obese patients subjected to bariatric surgery [6].

The importance of the association between Hp infection and MASLD is also obvious from the multiple relevant meta-analyses of observational studies that have been published on the topic, the newest being those of Mantovani et al. and Liu et al. [7, 8], which all indicate a positive association between Hp infection and MASLD [5].

MASLD and Hp infection are highly prevalent diseases; MAFLD prevalence is about 30%, and Hp infection prevalence is 50% in the global general population [9, 10]. Apart from hepatic and gastric morbidity, MASLD and Hp infection are also associated with extrahepatic and extragastric, respectively, morbidity [9, 10]. MASLD is a multifactorial disease, meaning that multiple factors contribute to its development and progression; in this regard, Hp infection may be one of the pathogenic contributors to the pathophysiology of MASLD [11]. Apart from Hp, the dysbiosis of gut microbiota may also be associated with the pathogenesis of MASLD [12]. However, Hp infection could not be the single or the strongest factor contributing to the pathogenesis of MASLD, but it may interplay with other factors, e.g., diet, BMI, diabetes, gut microbiota, thus increasing the risk of MASLD development and/or progression. If the association between Hp infection and MASLD comes true, this may have certain therapeutic implication, i.e., by eradicating Hp, the development of MASLD may be attenuated. However, this largely remains to be shown since existing relevant interventional studies are limited and inconclusive [13]. A major limitation of existing interventional studies is the lack of paired liver biopsies (at the baseline and at the endpoint after successful eradication of Hp), which would have provided the most robust evidence.

Many hypotheses on the molecular connections between Hp infection and MASLD have been provided. Hp infection increases the permeability of the gastrointestinal mucosal barrier, allowing the passage of unfavorable, gut-derived bacterial products to the liver, thus contributing to the development of MASLD [4, 5]; human β-defensin-1, which is also induced by Hp infection, is increased, and it has been proposed to serve as a biomarker of bacterial translocation in chronic liver disease, including MASLD. Hp infection may also contribute to the low-grade, chronic systemic inflammation and oxidative stress, which are hallmarks of the pathogenesis of MASLD, by triggering the release of proinflammatory cytokines and by decreasing adiponectin [4, 5]. Moreover, Hp infection is positively linked with insulin resistance (IR), a key pathogenic contributor to NAFLD [14]. Additionally, Hp infection may also be involved in the pathophysiology of MASLD by promoting: coagulation via stimulation of mononuclear cells toward the production of a tissue factor-like procoagulant activity that converts fibrinogen into fibrin; platelet aggregation via the binding of von Willebrand factor and/or formation of L- and P-selectin-dependent platelet-leukocyte aggregates; augmentation of plasma triglycerides and various atherosclerotic risk factors, such as homocysteine; elevation of other thrombotic and inflammatory mediators, including plasma plasminogen activator inhibitor concentration and the proinflammatory cytokines tumor necrosis factor-α and interleukin-6. All the aforementioned prothrombotic/inflammatory mediators are involved in the pathogenesis of MASLD and its related complications [14]. Likewise, circulating fetuin A, a multifunctional protein associated with Hp infection and the metabolic syndrome (MetS), was also shown to be correlated with IR and the severity of MASLD. Moreover, Hp-related galectin-3 seems to be connected with MetS and may play a role in the pathophysiology of MetS-related MASLD, since the number of galectin-3-positive liver cells was shown to be associated with the severity of MASLD. Other authors also supported that MASLD and Hp infection share genetic susceptibility (95 genes), and, notably, 21 of these 95 genes were shown to be mutually regulated in a genetic network analysis [15]. All these speculations, however, warrant mechanistic studies to explore whether Hp infection may causally affect the development of MASLD.

In conclusion, the study of Zheng et al. [1] may add on the intriguing association between Hp infection and MASLD, but a post hoc analysis of the same database, as recommended above, may provide even more valuable results in the field. If cohort studies validate the association between Hp and MASLD, which was supported by all meta-analyses of cross-sectional and cohort studies, then well-designed clinical trials are warranted to explore whether the eradication of Hp may beneficially affect MASLD. However, MASLD is a multifactorial disease, so the eradication of Hp, even if shown beneficial to MASLD, may alleviate but cannot “eradicate” MASLD because all other contributing factors to MASLD (e.g., diet, obesity, diabetes) should also be appropriately managed.