Gallstones increase the prevalence of Barrett’s esophagus
- First Online:
- Cite this article as:
- Matsuzaki, J., Suzuki, H., Asakura, K. et al. J Gastroenterol (2010) 45: 171. doi:10.1007/s00535-009-0153-4
- 79 Views
Bile and acid exposures are thought to be major risk factors for Barrett’s esophagus in Western countries. The association of gallstones with Barrett’s esophagus has not been fully evaluated. The present study was designed as a case-control study for determining the possible factors associated with endoscopically suspected esophageal metaplasia (ESEM), defined as an endoscopic finding suggestive of Barrett’s esophagus, in Japanese patients.
A total of 528 patients with ESEM were allocated to the case group, while 528 age- and gender-matched patients without ESEM were allocated to the control group. Findings on esophagogastroduodenoscopy and clinical background factors were compared using a multivariate logistic regression model.
The presence of gallstones and hiatus hernia and the severity of gastric mucosal atrophy were independently associated with the presence of ESEM [odds ratio (OR) 1.67, 95% confidence interval (CI) 1.03–2.69; OR 2.75, 95% CI 1.75–4.33; OR 1.25, 95% CI 1.01–5.6, respectively]. Compared with subjects with neither gastric corpus atrophy nor gallstones, although subjects with gallstones alone were not associated with the presence of ESEM (OR 1.59, 95% CI 0.87–2.92), having both gastric corpus atrophy and gallstones was strongly associated with the presence of ESEM (OR 2.94, 95% CI 1.40–6.17).
The presence of gallstones was independently associated with the presence of ESEM in the Japanese outpatient population, suggesting a causal association of distal esophageal bile exposure with the development of ESEM. Further studies are needed to confirm our findings in cases with histologically confirmed Barrett’s esophagus.
KeywordsBarrett’s esophagusESEMGallstoneGastric mucosal atrophy
The incidence of esophageal adenocarcinoma is rising more rapidly than that of any other malignancy in Western countries [1–3]. Esophageal adenocarcinoma remains highly lethal, with a 5-year survival rate of less than 15% [4, 5]. In most patients, esophageal adenocarcinoma arises from its premalignant precursor lesion, Barrett’s esophagus (BE) . BE is a transformation of the esophageal squamous epithelium into gastric metaplasia or specialized intestinal metaplasia . Investigating the risk factors of BE was important to clarify the cause of BE and esophageal adenocarcinoma.
BEs are classically attributed to reflux of the acid component of gastric content into the esophagus. Meanwhile, a number of observation support the concept that bile reflux, as occurs in duodeno-gastro-esophageal reflux (DGER), also contributes to the development of BE [8–11]. Izbeki et al.  reported that prevalence of gallstone disease was increased and gallbladder motility was impaired in patients with BE. Since microscopic cholesterol crystals are regularly washed out of the gallbladder if its contractions are effective enough, impaired motility of the gallbladder is one of the factors contributing to the development of gallstones . Gallbladder dysfunction or nonfunction is closely associated with the efficiency of continuous bile flow into the duodenum and the occurrence of duodeno-gastric reflux [14, 15]. Therefore, the formation of gallstones has been considered to be associated with duodeno-gastric reflux .
According to the Montreal definitions, the term endoscopically suspected esophageal metaplasia (ESEM) is used for an endoscopic finding consistent with BE that awaits histological evaluation . The prevalence of long-segment ESEM, where the length of the circumferential ESEM is more than 3 cm, has been estimated to be 0.2–0.4%, while that of short segment ESEM, where the length of the circumferential ESEM is less than 3 cm, has been estimated to be 6.0–20.6% in Japan [17, 18]. To diagnose for BE, an intensive biopsy protocol of four quadrant biopsies, taken at 1–2 cm intervals, for all patients with ESEM has been recommended [19, 20]. Although histological confirmations are needed to clarify the risk factors for BE, lower esophageal adenocarcinoma is still a rare disease in Japan, and biopsies from sites of ESEM are not routinely obtained at the screening EGD due to cost, risk and complexity. Since the presence of ESEM is a prerequisite condition for the development of BE , risk factors for ESEM also extend to BE and esophageal adenocarcinoma.
Several risk factors for BE are reported, such as gastro-esophageal reflux symptoms, hiatus hernia, aging, male gender, obesity, alcohol and tobacco use [22–27]. However, it is unknown whether the presence of gallstones is an independent risk factor for BE. Therefore, the authors performed a multivariate analysis on the relationship between ESEM and gallstones.
Characteristics of subjects with or without ESEM
ESEM (+) cases (n = 528)
ESEM (−) controls (n = 528)
Mean ± SD (years)
65.3 ± 12.3
65.3 ± 12.3
Gastric mucosal atrophy
Clinical background factorsa [no./total no. (%)]
BMI ± SD (kg/m2)
22.6 ± 4.0
22.4 ± 3.5
Assessment by endoscopy
The presence/absence of ESEM and hiatus hernia was determined, and gastric mucosal atrophy and reflux esophagitis were graded retrospectively by three gastrointestinal endoscopists according to the criteria listed below.
The presence/absence of ESEM was examined in the lower portion of the esophagus, including the EGJ, during inflation of the esophagus. The EGJ was defined as the oral side end of the fold continuous with the gastric lumen , as well as the anal side end of the palisade vessel, because the veins in the lower part of the esophagus were distributed uniformly, running parallel and longitudinally in the lamina propria [21, 29]. The squamo-columnar junction (SCJ) was defined by a clear change in the color of the mucosa. ESEM was defined as the area between the SCJ and the EGJ . Most patients were evaluated for the presence/absence of ESEM using the anal side end of the lower esophageal palisade vessel, since such a definition of the EGJ was more suitable for the retrospective evaluation.
Gastric mucosal atrophy
The severity of gastric mucosal atrophy was assessed endoscopically by the Kimura–Takemoto classification of the atrophic pattern [21, 30, 31]. This classification divides the severity of gastric mucosal atrophy into seven types (C-0, C-1, C-2, C-3, O-1, O-2 and O-3) according to the location of the atrophic border as detected by endoscopy as follows: C-0, absence of atrophy; C-1, pyloric mucosal atrophy; C-2, atrophy extending over the lesser curvature of the lower third of the stomach; C-3, the atrophy extending over the lesser curvature of the middle third of the stomach; O-1, border of the atrophy between the lesser curvature and anterior wall of the stomach; O-2, atrophy within the limits of the anterior wall of the stomach; O-3, atrophic area extending from the anterior wall to the major curvature of the stomach. Using this classification, the severity of the gastric mucosal atrophy was divided into four grades: none (C-0), mild (C-1 and C-2), moderate (C-3 and O-1) and severe (O-2 and O-3).
The presence/absence of hiatus hernia was examined by the valvular appearance of the cardia visualized from below using the retroflexed endoscope during gastric inflation .
Reflux esophagitis was defined as the presence of gross mucosal injury, ranging from red longitudinal streaks with associated friability to erosion or ulceration in the distal esophagus or breakage in the lower portion of the esophagus. The severity of reflux esophagitis was graded according to the Los Angeles classification .
Clinical background factors
The alcohol consumption status, smoking status and the presence/absence of H. pylori infection, obesity, hypertension, diabetes mellitus, dyslipidemia and gallstones were determined from the medical records. The alcohol consumption status was defined as a positive/negative history of daily alcohol consumption. The smoking status was defined as a positive/negative history of smoking cigarettes. The presence of H. pylori infection was defined as a history of H. pylori infection, including both pre- and post-eradication. The presence of H. pylori infection was detected by serological test, 13C-urea breath test, culture or histology of the gastric mucosal biopsy specimen. Obesity was defined as a body mass index of more than 25 kg/m2. Hypertension was defined as systolic blood pressure of over 140 mmHg and/or diastolic blood pressure of over 90 mmHg, or a history of use of antihypertensive drugs for the treatment of hypertension. Diabetes mellitus was defined as a serum hemoglobin A1c (HbA1c) value of over 6.5% or a history of use of antidiabetic agents. Dyslipidemia was defined as a serum level of low-density lipoprotein cholesterol (LDL-C) of over 140 mg/dl, high-density lipoprotein cholesterol (HDL-C) of under 40 mg/dl, a fasting triglyceride level of over 150 mg/dl or a history of use of lipid-lowering agents. The presence/absence of gallstones was determined by abdominal CT or ultrasonography. The presence/absence of gallstones of patients who had received cholecystectomy was determined by the previous record of abdominal CT or ultrasonography before cholecystectomy.
The associations of endoscopic findings or clinical background factors with the presence of ESEM were evaluated by a logistic regression model with adjustment for age and gender. Trends of association of the severity of reflux esophagitis and gastric mucosal atrophy with the presence of ESEM were evaluated by a logistic regression model that assigned scores to the level of the independent variable. The multivariate logistic regression model was conducted with adjustment for age, gender, hiatus hernia, reflux esophagitis, gastric mucosal atrophy, obesity and gallstones. All of the statistical analysis was performed using STATA 9.1 (Stata Corporation, College Station, TX). A two-sided p value of <0.05 was considered statistically significant.
Association between endoscopic findings and ESEM
Results for logistic regression analysis
ESEM (+) case no./total no. (%)
ESEM (−) control no./total no. (%)
Age- and gender-adjusted analysis
Odds ratio (95% CI)
Odds ratio (95% CI)
Gastric mucosal atrophy
Association between clinical background factors and ESEM
All of the clinical background factors of the subjects could not be determined from the medical records. Therefore, the associations between the clinical background factors and the presence of ESEM were analyzed using partial subjects whose background could be determined. All of the determined rates of the background factors were not different between cases and controls. Among these background factors, only the presence of gallstones was significantly associated with that of ESEM (OR 1.56, 95% CI 1.09–2.25). The presence of obesity was possibly associated with that of ESEM (OR 1.43, 95% CI 0.98–2.10, p = 0.06) (Table 2).
Multivariate analysis for endoscopic findings, obesity, gallstones and ESEM
The age- and gender-adjusted analysis showed that the presence of hiatus hernia, gallstones and obesity, and the severity of reflux esophagitis and gastric mucosal atrophy were associated with ESEM. Therefore, the authors analyzed the association of these factors with the presence of ESEM by a multivariate logistic regression model (Table 2). The presence of hiatus hernia, gallstones and the severity of gastric mucosal atrophy were independently associated with the presence of ESEM.
Association of gastric mucosal atrophy and gallstones and ESEM
Association of gastric mucosal atrophy and gallstones with ESEM
ESEM (+) cases
ESEM (−) controls
Odds ratio (95% CI)a
No./total no. (%)
No./total no. (%)
Factor 1: gastric corpus atrophy (+)b, gallstone (−)
Factor 2: gastric corpus atrophy (−)c, gallstone (+)
Factor 3: gastric corpus atrophy (+)b, gallstone (+)
Although the group with gallstones alone was not associated with the presence of ESEM (OR 1.59, 95% CI 0.87–2.92), the group with both gastric corpus atrophy and gallstones was strongly associated with the presence of ESEM (OR 2.94, 95% CI 1.40–6.17). The group with gastric corpus atrophy alone was also associated with the presence of ESEM (OR 1.63, 95% CI 1.03–2.57).
The result of the present case control study showed that the presence of gallstones, hiatus hernia and the severity of gastric mucosal atrophy were independently associated with the presence of ESEM. It has been suggested that both gallstones and previous cholecystectomy contribute to the occurrence of duodeno-gastric reflux [16, 34–39]. The underlying mechanism is thought to be a dysfunction of the antroduodenal motor unit that favors the reflux of duodenal contents into the stomach . According to the large population-based study, a moderately increased risk for esophageal adenocarcinoma following cholecystectomy was observed (OR 1.3, p < 0.05) . In the present study, independent association of the presence of gallstones with that of ESEM was shown, suggesting that the duodeno-gastric bile reflux plays an important role in the development of ESEM, although the association of previous cholecystectomy with the presence of ESEM was not directly examined.
The present study also showed that the presence of hiatus hernia was associated with that of ESEM. Previous studies have shown a strong association of hiatus hernia with BE [41–43], suggesting the occurrence of gastro-esophageal reflux in hiatus hernia. On the other hand, the multivariate analysis revealed the severity of reflux esophagitis was not independently associated with the presence of ESEM. This result was consistent with that of the recent Korean study , which showed that 77.7% of BE patients did not have reflux esophagitis.
Although obesity is known to be associated with BE [55, 56], the multivariate analysis to investigate the association of gallstones and obesity with BE has not yet been reported. The result of the present study revealed that the presence of obesity was not an independent risk factor for ESEM, which means that the presence of obesity was confounding variable. Recent studies suggested that low plasma adiponectin is associated with the development of BE [26, 57]. On the other hand, low plasma adiponectin is thought to be also associated with the development of gallstones [58, 59]. In addition, obesity is known to increase the prevalence of hiatus hernia [60, 61]. Therefore, obesity may confound hiatus hernia or gallstones.
In conclusion, not only the presence of hiatus hernia, but also the presence of gallstones and the severity of gastric mucosal atrophy appear to be risk factors in the presence of ESEM in the Japanese outpatient population. The injurious potential of gallstones complications might be exacerbated by the presence of gastric corpus atrophy with low acid. These results suggest a causal association of distal esophageal bile exposure with the development of ESEM, which could be enhanced by severe gastric atrophy.
This study was supported by Graduate School Doctoral Student Aid Program, Keio University (to J.M.), Grant-in-Aid for Exploratory Research from the Japan Society for the Promotion of Science (JSPS) (19659057, to H.S.) and Keio Gijuku Academic Development Fund (to H.S.). The authors thank Prof. Nimish Vakil, University of Wisconsin Medical School, for his valuable instructions and suggestions for this report.