Association between cagA negative Helicobacter pylori status and nonalcoholic fatty liver disease among adults in the United States

We investigated the relationship of H. pylori stratified by cytotoxin-associated gene A (cagA) status with nonalcoholic fatty liver disease (NAFLD) in the general population of the United States (US). We utilized the Third National Health and Nutrition Examination Survey from 1988 to 1994 in this study. NAFLD was defined by ultrasonographic detection of hepatic steatosis in the absence of other known causes of liver diseases and significant alcohol consumption. Hepatic steatosis was assessed by parenchymal brightness, liver to kidney contrast, deep beam attenuation, bright vessel walls and gallbladder wall definition. Antibodies to H. pylori and cagA of participants were measured using H. pylori IgG and anti-cagA IgG enzyme-linked immunosorbent assays. Among 5,404 participants, the prevalence of NAFLD was higher in H. pylori positive subjects (33.5±1.8%) compared to H. pylori negative subjects (26.1±1.7%, p <0.001). In terms of cagA protein status stratification, while cagA positive H. pylori group did not demonstrate an association with NAFLD (OR: 1.05; 95% CI: 0.81–1.37), cagA negative H. pylori group was noted to have a significant association with NAFLD in a multivariable analysis (OR: 1.30; 95% CI: 1.01–1.67). In conclusion, our study demonstrated that cagA negative H. pylori infection was an independent predictor of NAFLD in the US general population.


Introduction
Nonalcoholic fatty liver disease (NAFLD) is the most prevalent liver disease in the United States (US) and in many parts of the world [1,2]. NAFLD is a clinical and pathological entity that encompasses nonalcoholic fatty liver and nonalcoholic steatohepatitis complicated by progressive stages of fibrosis, cirrhosis, and hepatocellular carcinoma [3]. Because the mechanism underlying the development of NAFLD has been linked to insulin resistance and metabolic syndrome, NAFLD is associated with many risk factors of cardiovascular disease, such as obesity, diabetes, and dyslipidemia [4][5][6][7].
Helicobacter pylori (H. pylori) infection has been associated with a variety of extra-digestive conditions, including cardiovascular, lung, hematologic, neurologic and hepatobiliary diseases [8,9]. In the Finnish Alpha-Tocopherol, Beta-Carotene Cancer Prevention (ATBC) study, H. pylori seropositivity was associated with an increased risk of biliary tract cancer [10]. Bacterial infection of intestinal tract may compromise mucosal integrity, allowing passive or facilitated entry of harmful bacteria or their products into the circulation [11]. Notably, cytotoxin-associated gene A (cagA) positive strains can introduce H. pylori proteins into the epithelial cells and induce an antibody response to the cagA protein. These strains was found to be associated with an increased risk of gastric cancer and peptic ulcer disease and a decreased risk of esophageal reflux disease [12]. Until now, several small-scale epidemiological studies have reported a relationship between NAFLD and H. pylori infection [8,13]. However, association between NAFLD and H. pylori infection with regard to cagA status has not been studied in a large, general population.
Therefore, in this study, we used the National Health and Nutrition Examination Survey (NHANES) data to investigate the impact of H. pylori infection with regard to cagA status on NAFLD.

Study population and study design
This study represents an analysis of the third NHANES data (1988)(1989)(1990)(1991)(1992)(1993)(1994), the National Center for Health Statistics, the Centers for Disease Control and Prevention [CDC]). NHANES employed a stratified, multistage, clustered probability sampling design to obtain a representative sample of the non-institutionalized civilian population in the US. In total, 7,275 adults from 20 to 74 years of age, who participated in the NHANES III survey Phase I, underwent laboratory tests at a mobile examination center (Fig 1). Subjects with significant alcohol consumption (> 21 drinks/week in men and > 14 drinks/week in women) [14], viral hepatitis (positive serum hepatitis B surface antigen and/or positive serum hepatitis C antibody), iron overload (transferrin saturation ! 50%), and pregnant women were excluded (n = 816). Among the remaining 6,459 participants, hepatic steatosis could be evaluated in 5,910 (91.5%). We further excluded participants whose tests did not examine H. pylori status. Finally, 5,404 subjects were included and analyzed in this study. The original NHANES survey was approved by the CDC's Institutional Review Board, and all participants provided written informed consent to participate. This retrospective study was exempted by the Institutional Review Board of the Seoul National University Hospital because the dataset used in this analysis was completely deidentified. This study was conducted according to the Guideline for Good Clinical Practice (GCP) and the provisions of the Helsinki declaration.

H. pylori status
Antibodies to H. pylori were measured in participants 20 years and older using a H. pylori IgG enzyme-linked immunosorbent assay (ELISA) (Wampole Laboratories, Cranbury, New Jersey) in serum surplus samples [15]. Additionally, anti-cagA IgG was also measured in surplus sera among participants positive for H. pylori, using a method developed and standardized by Vanderbilt University, as described previously [16]. On the basis of the H. pylori and cagA status, we classified participants into three groups: H. pylori positive and cagA positive, H. pylori positive and cagA negative, and H. pylori negative [17].

Assessment of NAFLD
We used a previously described method to assess NAFLD [2]. Briefly, the original NHANES III survey included gallbladder ultrasonography as part of the assessment for digestive disease in adults aged 20 and above. Three board-certified radiologists reviewed the archived gallbladder ultrasonography images to assess fatty liver [18]. Assessment of fatty liver was performed using the following criteria: (1) parenchymal brightness; (2) liver to kidney contrast; (3) deep beam attenuation; (4) bright vessel walls; and (5) gallbladder wall definition. Overall assessment, made using an algorithm based on these five criteria, defined normal versus mild, moderate, or severe hepatic steatosis [18]. For our study, NAFLD was diagnosed as the presence of any degree of fatty liver (mild to severe).

Advanced fibrosis by noninvasive panels
The noninvasive panels used for the evaluation of advanced fibrosis have been previously described [2,19]. Briefly, advanced fibrosis was assessed by NAFLD fibrosis score, FIB-4 score, and aspartate aminotransferase to platelet ratio index. Due to the small sample size of subjects with high probability of advanced fibrosis, 'suspected advanced fibrosis' was evaluated and defined as presence of high probability for advanced fibrosis on at least one of the three noninvasive fibrosis panels.

Clinical and laboratory evaluation
A vast array of demographic, lifestyle, dietary, anthropometric and comprehensive laboratory data were available in this NHANES dataset. Hypertension was defined as systolic blood pressure ! 140 mmHg or diastolic blood pressure ! 90 mmHg and/or previous use of antihypertensive medication. Diabetes mellitus was defined as having a history of diabetes and/or treatment with a hypoglycemic agent or insulin. Current smokers were participants who reported ongoing smoking or those who had smoked at least 100 cigarettes in the preceding 5 years.

Statistical analyses
The outcome variable in this study was the presence of NAFLD. Analyses were conducted using the SAS-callable SUDAAN 10.0.1 (Research Triangle Institute, Research Triangle Park, NC), which allows for the stratified sampling scheme by NHANES to project these data to the US population [20]. We analyzed proportions of categorical variables and means ± standard error of continuous variables (PROC CROSSTAB, PROC DESCRIPT). We used chi-squared tests for the categorical variables and Student's t-test and one-way analysis of variance (ANOVA) for the continuous variables (PROC CROSSTAB, PROC REGRESS). A multivariable logistic regression model was used to investigate the independent association between NAFLD and H. pylori status (PROC LOGISTIC). All subjects in NHNAES III aged 17 years and over were followed for mortality through December 2006. We utilized Cox proportional hazards regression testing for survival analysis of overall mortality.

Baseline characteristics according to NAFLD
The prevalence of NAFLD among the eligible subjects was 31.9%. Demographic and clinical characteristics of subjects with NAFLD are shown in Table 1. Subjects with NAFLD were more likely to be older, male, hypertensive, and diabetic than those without NAFLD. Similarly, body mass index (BMI), waist circumference, lipid panels, liver enzymes, and fasting glucose were higher in subjects with NAFLD than those without NAFLD. Table 2 shows the distribution of participants based on H. pylori status by demographic and clinical risk factors in the overall study population, including those with data on H. pylori cagA status. Subjects with positive H. pylori serology results were more likely to be older, male and with a higher BMI and waist circumference than that of H. pylori-negative participants. Subjects of a race/ethnicity other than non-Hispanic white and those who had a history of diabetes or hypertension were more likely to be H. pylori-positive. Similarly, lipid panels, liver enzymes, and fasting glucose were higher in subjects with positive H. pylori serology than without. There was no apparent association of H. pylori positivity with current smoking status. Among subjects with positive H. pylori serology, those without cagA were likely to be older with higher prevalence of hypertensive versus those with cagA. Subjects of a race/ethnicity other than non-Hispanic white were more likely to be cagA positive.
When we evaluated the association of H. pylori seropositivity and/or cagA presence with advanced fibrosis among subjects with NAFLD (S2 Table), no statistically significant  H. pylori infection was associated with an increased risk of NAFLD in all ethnicities. However, H. pylori infection was significantly associated with NAFLD in non-Hispanic black participants. Regarding cagA status, cagA negative H. pylori was associated with an increased risk for NAFLD in the non-Hispanic white and non-Hispanic black populations. However, cagA negative H. pylori status was not associated with NAFLD in Mexican-Americans (Fig 3).

Discussion
The main finding in this nationally representative, population-based study was that the prevalence of NAFLD was higher in subjects with H. pylori than in those without. In subjects with H. pylori, the prevalence of NAFLD was greater in cagA negative subjects compared to cagA positive subjects. Additionally, cagA negative H. pylori positivity was significantly associated with NAFLD after adjustment for multiple classical risk factors, confirming that a relevant clinical relationship exists between these two conditions.
Several studies have investigated the association between H. pylori and NAFLD. Polyzos et al. reported that higher rates of anti-H. pylori IgG were observed in the NAFLD group (n = 28) compared to that of the healthy control group (n = 25) [8]. They also showed that biopsy-proven NAFLD patients had higher insulin and tumor necrosis factor (TNF)-α levels compared to those of controls. However, there were no significant differences in steatosis grade, fibrosis stage, lobular or portal inflammation, or balloon degeneration when NAFLD patients were divided based on H. pylori IgG seropositivity or 13 C urea breath test positivity [8]. This study was also limited by a small sample size. Dogan et al. showed that fatty liver determined by ultrasonographic score was observed more frequently in H. pylori positive patients, but they did not consider other metabolic confounders [13]. NAFLD is considered as the hepatic manifestation of metabolic syndrome and is strongly associated with cardiovascular disease. [21]. Several studies reported that an association between H. pylori and metabolic syndrome is significant as shown in case of NAFLD. In a large cross-sectional study from Japan, H. pylori seropositivity was significantly associated with metabolic syndrome [22]. On the contrary, studies from highly endemic areas of H. pylori revealed no association between positive H. pylori serology and NAFLD [23,24]. In a recent meta-analysis, H. pylori infection and NAFLD demonstrated a modest, significant association (pooled OR 1.21 (95% CI 1.07-1.37) [25]. Recently, Fan et al. found that prevalence of NAFLD was significantly higher in subjects with H. pylori infection (36.0% vs. 33.3%, p <0.05) [26]. However, after adjusting for confounding factors, H. pylori infection was not significantly associated with NAFLD. In our study using representative US population, we found H. pylori and cagA negativity was associated with NAFLD, which may partly explain the inconsistent results in the literature. In the present study, we found only infection with cagA negative H. pylori was significantly associated with NAFLD. CagA plays an important role not only in gastric inflammation but also in the development of gastric cancer. Regarding the effect of cagA on cardiovascular and metabolic risk factors, a German population-based cohort study showed decreased incidence of cardiovascular mortality in the cagA positive population compared with that of H. pylori- negative subjects. The difference of incidence between the two groups was 33% -38%, but the cardiovascular mortality of the cagA negative population was similar to that of the H. pylorinegative group [27]. Analysis of the association between H. pylori and mortality in the NHANES III study showed an inverse relationship between mortality from stroke and cagA positive H. pylori infection [28]. The inverse relation between cagA and cardiovascular mortality can be explained by regulatory T cells (T-reg). In mice model, cagA plays an important role in the migration of CD4+ T cells in the gastric mucosa and cagA-dependent T-cell priming induces T-reg differentiation [29]. The presence of T-reg in the gastric mucosa of H. pyloriinfected subjects suggested their involvement in suppressing mucosal immune responses, contributing to the infection persistence and modulation of the H. pylori-induced gastritis [30]. Reduced risk of asthma and allergies in H. pylori-infected populations appears to be related to the induction of T-reg responses [31]. T-reg is also known to be important in regulating inflammatory processes in NAFLD [31]. The highly stimulated regulatory immune system in the setting of cagA may be the reason why only cagA negative H. pylori infection is related with NAFLD as shown in this study. Insulin resistance can be another possible explanation for the association between H. pylori and NAFLD. Insulin resistance is a key factor in both metabolic syndrome and NAFLD. One study showed that insulin resistance, which is measured as HOMA-IR, was significantly higher in the H. pylori positive group than in the H. pylori negative counterparts [32]. A systematic review of nine studies investigating the association between H. pylori infection and insulin resistance indicated that H. pylori infection was positively associated with insulin resistance [33]. It is well known that even chronic subclinical inflammation is associated with cardiovascular disease [34]. Long-term H. pylori infection induces a 'state of inflammation' by stimulating excessive release of pro-inflammatory cytokine and vasoactive substances, such as interleukin (IL)-6, IL-8, IL-1β, and TNF-α [35,36]. TNF-α is a key mediator of both direct and indirect effects of H. pylori infection on NAFLD. TNF-α interferes with insulin signaling, thereby favoring steatosis, and may play a pro-inflammatory role in the pathogenesis of NAFLD [37,38]. In addition, TNF-α can accelerate lipolysis, leading to an increase in free fatty acids. This can result in hepatocyte dysfunction, including increased oxidative stress, induction of endoplasmic reticulum dysregulation and subsequent expression of pro-inflammatory cytokines [39,40]. We noted that H. pylori positivity was associated with increased risk for allcause mortality among subjects with NAFLD, though after adjusting for traditional risk factors, the association remained at marginal significance and thus needs further confirmation. This result perhaps suggests that the effect of H. pylori positivity in NAFLD on all-cause mortality may be mediated in part by diabetes, hypertension, and smoking.
In case of a disruption of the intestinal barrier, lipopolysaccharide from gram-negative bacteria increases in the portal circulation and is accompanied by increased levels of endotoxinmediated cytokines in the liver. So bacterial constituents and cytokines enhance hepatic inflammation and fibrosis [41]. Miele et al. also found that in NAFLD patients, increased gut permeability and the prevalence of small intestine bacterial overgrowth correlated with the severity of steatosis but not with the presence of steatohepatitis [42]. H. pylori is believed to colonize not only the stomach and duodenal epithelium but also the biliary epithelium [43]. So H. pylori colonization in stomach and duodenum can increase in lipopolysaccharides and endotoxin in portal circulation and can promote hepatic inflammation and fibrosis.
The prevalence of H. pylori infection showed racial disparities. Both non-Hispanic blacks and Mexican-Americans had a significantly higher prevalence of H. pylori than non-Hispanic whites [44]. The prevalence of cagA positive H. pylori was higher in non-Hispanic blacks and Mexican-Americans versus non-Hispanic whites [44]. However, in this study, a significant association between cagA negative H. pylori strains and NAFLD was found in non-Hispanic whites and non-Hispanic blacks but not in Mexican-Americans. Further research is needed to confirm these ethnic differences.
Currently, only two studies have reported the effect of H. pylori eradication on metabolic profile or NAFLD. One prospective study revealed that in the group with successful eradication, levels of total adiponectin and each multimer form were significantly increased 12 weeks after completion of treatment [45]. This study suggested that eradication could have a beneficial effect on metabolic risk by increasing adiponectin levels. A recent small-scale randomized open label study showed that lifestyle modification plus H. pylori eradication did not result in significant differences in liver fat content by ultrasonography and in insulin resistance compared with that of lifestyle modification alone [46]. However, the primary outcome of this study is relatively subjective, and the open label design may be a pitfall. Therefore, further large prospective studies are needed to confirm the effect of H. pylori eradication on NAFLD.
There are strengths and limitations in our study. We studied a large representative sample from the US general population; thus, the results can be generalizable. Second, this study performed analysis according to cagA seropositivity and studied the differential impact of cagA presence (versus cagA absence) on the association between NAFLD and H. pylori. However, there may be other predictive factors that may play an influential role in impacting the association of H. pylori and NAFLD. Third, we included sufficient confounders from the NHANES III data, which increased the accuracy of this study. Finally, our study included various ethnic/ racial groups, and thus, our conclusions may be applicable to the general population. This study also has some limitations. First, because of its cross-sectional design, a causal relationship between H. pylori infection and NAFLD could not be identified. Second, our diagnosis of NAFLD was based on ultrasonographic examination, which is not able to differentiate nonalcoholic steatohepatitis from nonalcoholic fatty liver. In addition, hepatic ultrasonography cannot identify fatty infiltration below 30% and has intra-and inter-observer variability in making a diagnosis. Third, our study is unable to offer a mechanistic explanation to associate cagA negative H. pylori seropositivity with the risk of NAFLD. Further studies are warranted to determine mechanism(s) explaining this association.
In conclusion, the present study suggests that H. pylori infection is related to an increased prevalence of NAFLD in the general population. CagA negative H. pylori infection was associated with an increased risk for NAFLD, confirming that a relevant clinical relationship exists between these two conditions. However, since pathogenic mechanisms have not been fully elucidated, future studies are needed to clearly understand the clinical implications of our results.
Supporting information S1