Barrett Esophagus

Abstract

Barrett esophageal cancer has the fastest growing incidence of any cancer in Western countries. In Asian countries, most cases of esophageal cancer consist of squamous cell carcinomas, not adenocarcinomas. Recently, however, the increase in the number of Barrett esophagus cases with subsequent Barrett cancer has become worrisome in Asian countries, as the number of patients with gastro-esophageal reflux disease has been increasing in these countries. In this review, recent reports regarding Barrett esophagus in Asian countries have been collected and this problem is discussed from various perspectives. In Asia, long-segment Barrett esophagus is much less prevalent than in Western countries, whereas short-segment Barrett esophagus is frequently found. In epidemiologic studies, evaluation of the prevalence of Barrett esophagus is limited by poor interob-server diagnostic agreement. Standard criteria for the endoscopic diagnosis of Barrett esophagus in Asian patients, especially of the short-segment type, should be established as soon as possible. A high prevalence of hiatal hernia and a decreasing prevalence of Helico-bacter pylori infection may increase the number of Barrett esophagus cases and subsequent Barrett cancer in Asian countries in the near future. Therefore, a strategy for the clinical management of Barrett esophagus in Asian countries should be devised.

Barrett esophagus is defined as intestinal metaplasia of the esophageal mucosa, leading to replacement of the esophageal squamous epithelium with columnar epithelium.^1,2 Esopha-geal adenocarcinoma, namely Barrett cancer, derived from Barrett esophagus has been rapidly increasing during the last two decades in North America, Australia, and Europe. Approximately 60% of Caucasian patients with esophageal cancer living in the United States have adenocarcinoma in Barrett esophagus.³ Columnar-lined esophagus with intestinal metaplasia expressing MUC2 core protein can develop into esophageal adenocarcinoma.^4–7 Therefore, in Western countries, it is quite reasonable to define columnar-lined esophagus with goblet cell metaplasia as Barrett esophagus with malignant potential, as goblet cells express MUC2 core protein. In Asian countries, the majority of esophageal cancer is squamous cell carcinoma. The number of patients with Barrett cancer is small, and only approximately 1% of cases with esophageal cancer are histo-logically adenocarcinomas.⁸ One reason for the low prevalence of Barrett cancer in Asia may be that Barrett esophagus cases are mainly short-segment (<3 cm). However, as with long-segment Barrett esophagus (LSBE; >3 cm), some short-segment Barrett esophagus (SSBE) cases also show malignant potential.^9–11 Reports of a gradually increasing incidence of Barrett cancer in Asia suggest that particular attention should be paid to these patients.^12,13 Accordingly, the strategy for clinical management of Barrett esophagus in Asia, as well as in Western countries, should be established. This review article evaluates the present status of Barrett esophagus in Asia and discusses perspectives on its diagnosis and treatment.

Prevalence of Barrett Esophagus in Asia

According to recent reports from Western countries, the prevalence of Barrett esophagus (histologically confirmed by the presence of goblet cell metaplasia/specialized columnar epithelium), LSBE, and SSBE is 1.6–25.0%, 0.5–7.2%, and 1.1–17.2%, respectively.^14–19 In these reports, most of the patients were Caucasian. In patients with gastroesophageal reflux disease (GERD), the prevalence of Barrett esophagus is over 10%.^20,21 Cameron and associates reported that LSBE showed a fairly rapid evolution to its full length with little subsequent change. They observed no age-related increase in the prevalence of LSBE.²² On the other hand, in our study, we found that change in the prevalence of Barrett esophagus (which was 99% SSBE) was age-related, as shown in Figure 1. In our study, Barrett esophagus was significantly more prevalent in patients over 70 years of age than in younger patients. This finding suggests that the pathophysiology of Barrett esophagus (mainly SSBE) may be different in Asia than in the West (which has a high LSBE prevalence).

Figure 1.

The prevalence of patients with Barrett esophagus (indicated by the bars) and hiatal hernia (indicated by the black dots) correlates with increasing age. During the study period of 2004– 2006, 333 patients with histologically confirmed Barrett esophagus were found in 1,997 patients who underwent eso-phagogastroduodenoscopy. In patients over 70 years of age, the prevalence of Barrett esophagus and hiatal hernia was higher than that in the younger population.

*P<.05, compared to the patients younger than 40 years of age.

Reports examining the prevalence of Barrett esophagus in Asian countries are shown in Table 1.^{7,8,12,23–39} LSBE prevalence in Asia is extremely low (<1.0% of all Barrett esophagus patients in most reports). In contrast, SSBE prevalence in Asia is greater than 96% of all Barrett esophagus patients. The prevalence data vary, as demonstrated in Table 1, possibly due to racial differences in Asian countries²³ or differences in how the Barrett esophagus was diagnosed (endoscopically vs histologically). The prevalence of endoscopic versus histologic Barrett esophagus also varied widely. Armstrong was the first researcher to recommend that “endoscopic Barrett esophagus” be considered synonymous with esophageal columnar-appearing mucosa or columnar-lined esophagus.⁴⁰ The most important reason for the differentiation is the difficulty in diagnosing Barrett esophagus endoscopically. In particular, the endoscopic diagnosis of SSBE is less reliable than that of LSBE.^41,42

Table 1.

Prevalence of Barrett Esophagus in Asian Countries

Study	Study Period	Country or Ethnicity	Sample Size (N)	Study Method	Prevalence of Barrett Esophagus
Rajendra S, et al.23	1997–2000	Chinese, Indian, Malay	1,985	Single center	LSBE: 1.6%, SSBE: 4.6%
Hongo M, Shoji T8	1999–2000	Japan, Korea, Taiwan	7,134	Multicenter	LSBE: 0.4%,* SSBE: 0.9%*
Wong WM, et al.24	1997–2001	China	16,606	Single center	LSBE: 0.02%, SSBE: 0.04%
Zhang J, et al.25	2001–2002	China	391	Single center	LSBE: 6.6%,* SSBE: 24.0%*
Amarapurkar AD, et al.26	1997	India	150 with FD	Single center	2.6%
Dhawan PS, et al.27	2000	India	271	Single center	5.9% (by 1 biopsy specimen)
Punia RS, et al.28	1999–2002	India	55 with GERD	Single center	23.6%
Bafandeh Y, et al.29	2001–2003	Iran	1,248 with GERD	Single center	LSBE: 0.8%, SSBE: 1.6%, BE: 8.3%*
Rezailashkajani M, et al.30	2005–2006	Iran	501		0.2%
Azuma N, et al.31	1996–1998	Japan	650	Single center	LSBE: 0.6%,* SSBE: 15.1%*†
Hongo M, Skoji T8	1999–2000	Japan	18,400	Multicenter	LSBE: 0.2%,* SSBE: 1.0%*
Fujiwara Y, et al.32	2001–2003	Japan	548	Multicenter	LSBE: 0.2%, SSBE: 12.0%
Kawano T, et al.33	2003	Japan	2,577	Multicenter	LSBE: 0.2%,* SSBE: 20.8%*†
Amano Y, et al.7	2003–2004	Japan	1,699	Single center	LSBE: 0.2%, SSBE: 19.7% LSBE: 0.4%,* SSBE: 37.7%*
Lee JI, et al.34	2000	Korea	1,553	Multicenter	LSBE: 0.32%
Kim JH, et al.35	1997–2004	Korea	70,103	Multicenter	LSBE: 0.01%, SSBE: 0.14% LSBE: 0.01%,* SSBE: 0.53%*
Choi DW, et al.36	2002 (publication)	Korea	847	Single center	LSBE: 0.5%,* SSBE: 16.5%*
Kim JY, et al.37	2005 (publication)	Korea	992	Multicenter	LSBE : 0.1%, SSBE : 3.5% LSBE: 0.3%,* SSBE: 10.9%*
Rosaida MS, Goh KL38	2004 (publication)	Malaysia	1,000	Single center	2.0%
Gadour MO, Ayoola EA39	1999 (publication)	Saudi Arabia	159	Single center	0.3%
Yeh C, et al.12	1991–1992	Taiwan	464	Single center	1.98%

BE

Barrett esophagus

FD

functional dyspepsia

GERD

gastroesophageal reflux disease

LSBE

long-segment Barrett esophagus

SSBE

short-segment Barrett esophagus.

^*Prevalence of endoscopic Barrett esophagus.

^†Endoscopic Barrett esophagus diagnosed by palisade vessel criteria.

A multicenter, prospective, nationwide, epidemio-logic study conducted in Japan by the Japanese GERD Society found an overall prevalence of endoscopically suspected Barrett esophagus of 24.1%, of which 99.1% was SSBE.³³ Some patients were symptomatic during their annual health check-up, whereas others were asymptomatic. The prevalence of Barrett esophagus was higher in this study than in other studies. Moreover, the diagnostic criteria for endoscopic Barrett esophagus were those widely used in Japan, but not in Western countries. The distal end of the lower esophageal palisade vessels was used to define the gastroesophageal junction.

Endoscopic Diagnosis of Barrett Esophagus

Marked variation in the prevalence of Barrett esophagus in Asian countries may stem from differences in endoscopic diagnostic criteria among studies. In the study by the Japan GERD Society cited above, the endoscopic criteria proposed by the Japan Esophageal Society were utilized.⁴³ In Japan, the distal end of the lower esophageal palisade vessels is believed to coincide with, and is used to define, the esophagogastric junction (EGJ).^44–48 In Japan, Barrett esophagus is defined by the presence of columnar epithelium between the squamocolumnar junction (SCJ) and the EGJ, irrespective of the type of columnar epithelium and its length (ie, LSBE or SSBE), depending upon the distance of the SCJ from the EGJ.

In Western countries, the landmark for the EGJ is the proximal end of the gastric longitudinal folds. According to the Prague C & M Criteria for Barrett Esophagus, which were proposed by a subgroup of the International Working Group for the Classification of Oesophagitis (IWGCO) in order to standardize the objective diagnosis of endoscopic Barrett esophagus using conventional endoscopy,⁴⁰ the landmark for the EGJ is the proximal end of the gastric folds. As interobserver disagreement regarding endoscopic diagnosis of Barrett esophagus, in particular SSBE, results from the difficulty of determining the location of the EGJ,^41,42 the question arises as to which landmark is appropriate: the proximal end of the gastric folds or the distal end of the esophageal palisade vessels?

Usually, palisade vessels can be found easily when the lower esophagus is adequately distended by air insufflation. In cases with superficial dysplastic lesions and/or intense inflammation, palisade vessels sometimes cannot be endoscopically detected, suggesting that the palisade vessels criterion may not be useful for endoscopic diagnosis of Barrett esophagus.

On the other hand, an investigation of the Prague C & M Criteria using video clip images of the EGJ found interobserver disagreement in Barrett esophagus cases less than 1 cm in length, but it also found interobserver agreement in Barrett esophagus cases greater than 1 cm.⁴¹ Similarly, in our own study, kappa values were higher for endoscopic diagnosis of Barrett esophagus longer than 1 cm in length,⁴² suggesting the unsuitability of the C & M criteria for diagnosing SSBE.

Utilizing only palisade vessels to determine the endoscopic definition of the EGJ resulted in a low kappa coefficient of reliability. When the upper end of the longitudinal gastric folds was used, the diagnostic concordance, though initially low, improved to an acceptable level after the complete application of Prague C & M Criteria for Barrett Esophagus.⁴² In April 2007, the Japan Esophageal Society added gastric folds as an additional landmark of the EGJ to the guidelines for clinical and pathologic studies on carcinoma of the esophagus.⁴³ Therefore, it is acceptable for Japanese endoscopists to determine the EGJ using both landmarks: the distal end of the palisade vessels and the proximal end of the gastric folds when the palisade vessels are invisible.

Histologic Diagnosis of Barrett Esophagus

Another reason for variation in the prevalence of Barrett esophagus may stem from the lack of a worldwide consensus on the histologic definition of Barrett esophagus. According to the guidelines of the American College of Gastroenterology,¹ Barrett esophagus is defined by the change of the esophageal epithelium of any length recognized at endoscopy and histologically confirmed to contain intestinal metaplasia. Furthermore, a definite histologic diagnosis of Barrett esophagus requires proof of intestinal metaplasia with goblet cells. Kerkhof and colleagues showed that intestinal metaplasia can be found easily in biopsy specimens if the length of endoscopic Barrett esophagus is longer than 2 cm.⁴⁹ However, intestinal metaplasia cannot be found easily in many SSBE cases with uneven distribution of intestinal metaplasia.

According to the proposed guidelines of the British Society of Gastroenterology,² the presence of specialized columnar epithelium is less important for the diagnosis of Barrett esophagus than the presence of a proper esopha-geal gland, squamous island, and/or double muscularis mucosae.^50–53 The efficacy and usefulness of their criteria have also been sufficiently discussed at the Paris Workshop on Columnar Metaplasia in the Esophagus and the Esophagogastric Junction.⁵⁴ Like the British Society of Gastroenterology, the Japan Esophageal Society defines Barrett esophagus as the presence of at least 1 of the following factors: a proper esophageal gland, squamous island, and double muscularis mucosae.⁴³ Therefore, differences among studies regarding epidemiologic prevalence data for Barrett esophagus may be due to differences among these studies regarding their histologic criteria.

Barrett Esophagus and Helicobacter pylori Infection in Asia

Predictors for the presence of Barrett esophagus include old age, male gender, hiatal hernia, and long-lasting reflux symptoms.^16,55–59 Alcohol and smoking have also been reported to be predictors for Barrett esophagus and Barrett cancer in the West,^15,18,49,60 but not in Asia.^7,35,37

However, hiatal hernia and Helicobacter pylori infection are important predictors for Barrett esophagus in Asia. The presence of hiatal hernia plays an important role for the pathogenesis of reflux esophagitis and Barrett esophagus, as hiatal hernia induces long-lasting reflux of gastric acid contents.^23,38,61 The prevalence of hiatal hernia increases with age in Japan; hiatal hernia was found in approximately 70% of patients over 70 years of age, as shown in Figure 1.

H. pylori infection is well known to be more prevalent in Asian than in Western countries and to protect against the development of GERD and Barrett esophagus.^62–69 In Japan, the prevalence of H. pylori infection was approximately 70% among patients born before 1950.⁷⁰ The prevalence of H. pylori infection, however, has decreased during the last two decades.⁷¹ The gradual decrease of H. pylori seroprevalence was also found in many Asian countries.^72–74 H. pylori infection protects against the development of Barrett esophagus via the decrease of gastric-acid secretion. Therefore, the high rate of H. pylori infection is believed to account for the lower prevalence of Barrett esophagus in Asian countries. Vieth and coworkers determined the prevalence of H. pylori infection from a meta-analysis of 23 studies, including their own, and concluded that 35.6% of 2,084 patients with Barrett esophagus in Western countries were infected with H. pylori.⁷⁵ As shown in Table 2,^69,75–83 the prevalence of H. pylori infection in patients with Barrett esophagus is higher in Asian than in Western countries.

Table 2.

Prevalence of Helicobacter pylori Infection in Patients with Barrett Esophagus in Asian Versus Western Countries

Study	Study Period	Country	Sample Size (N)	Study Method	Prevalence
Asian Countries
Abe Y, et al.76	1996-2001	Japan	112 with RE	Single center	SSBE: 18.7% LSBE: 0%
Zhang J, et al.77	2001-2002	China	375 with GERD	Single center	SSBE: 48.7%, LSBE: 41.8%
Amano Y, et al.78	2002-2003	Japan	400 with BE	Single center	SSBE: 51.2%
Chinuki D, et al.79	2003-2004	Japan	266 with BE	Single center	SSBE: 45.5%
Rajendra S, et al.69	2003-2005	Malaysia	188 with GERD	Single center	SSBE: 72.0%, LSBE: 26.7%
Western Countries
Vieth M, et al.75	1990-1997	Germany	1,054 with BE	Single center	43.9%
Loffeld RJ, van der Putten AB80	1994-2004	Netherlands	11,691	Single center	30.7%
Wani S, et al.81	2000-2002	US	46 with BE	Single center	6.5%
Weston AP, et al.82	2000 (publication)	US	289 with BE	Single center	35.1%
Veldhuyzen van Zanten SJ, et al.83	2006 (publication)	Canada	1,040	Multicenter	28.0%

BE

Barrett esophagus

GERD

gastroesophageal reflux disease

LSBE

long-segment Barrett esophagus

RE

reflux esophagitis

SSBE

short-segment Barrett esophagus.

Schenk and associates found that the prevalence of H. pylori infection was 20.4% and 44.3% in patients with and without Barrett esophagus, respectively.⁸⁴ Weston and colleagues also found that the incidence of H. pylori infection is related to the incidence of Barrett carcino-genesis (ie, 35.1%, 14.3%, and 15.0% in patients without dysplastic lesion, with high-grade dysplasia, and with Barrett cancer, respectively).⁸² Moreover, SSBE developed in 10.3% of patients who underwent H. pylori eradication therapy,⁸⁵ and Barrett esophagus developed after spontaneous healing of atrophic gastritis in patients with H. pylori antibody titers that had fallen to normal levels.⁸⁶ These findings clearly suggest that H. pylori infection is protective for the development of Barrett esophagus.

Vicari and colleagues reported that cytotoxin-associ-ated gene A (CagA)-positive strains of H. pylori, which are known to induce intense inflammation in the gastric mucosa and are closely correlated with the incidence of gastric carcinogenesis, protect against the development of Barrett esophagus.⁸⁷ Although the prevalence of H. pylori infection was not significantly lower in patients with Barrett esophagus (approximately 34%) than in the control group, the prevalence of CagA-positive H. pylori infection was significantly decreased in patients with Barrett esophagus (13.3%). This protective effect was confirmed by other investigations in Asian and Western countries.^88–91 Thus, it is suggested that H. pylori infection affects the prevalence of Barrett esophagus and may account for the lower prevalence of this disease in Asia than in the West.

On the other hand, several investigators found no relationship between H. pylori infection and the development of Barrett esophagus or Barrett cancer.^{75,77,92–94} Henihan and coworkers found that H. pylori infection induces severe chronic inflammation in Barrett esophagus and, thereby, is not protective against development of Barrett esophagus.⁹⁵ Thus, the relationship between H. pylori infection and Barrett esophagus is still ambiguous. Koike and associates demonstrated that the prevalence of H. pylori infection was significantly lower in patients with adenocarcinoma at the EGJ than in patients with gastric cancer, though not as low as in patients with erosive esophagitis and Barrett esophagus.⁹⁶ This finding suggests that preservation of gastric acid secretion may be important for the development of adenocarcinoma at the EGJ, regardless of the presence of H. pylori infection. Thus, a high infection rate of H. pylori does not completely explain the lower prevalence of Barrett esophagus and shorter Barrett segments in Asian than in Western countries.

Issues Related to Incidence of Barrett Esophagus in Asia

According to the comprehensive registry of the Japan Esophageal Society during 1995–1997, esophageal aden-ocarcinoma accounted for only 1.4% of all esophageal cancers.⁸ However, as GERD is gradually increasing in Asia,⁹⁷ Barrett esophagus may increase. Most Barrett esophagus in Asia is SSBE, which also has a malignant potential. Therefore, the establishment of strategies for diagnosing and treating Barrett esophagus is an important goal in Asia, as well as in the West.

First, new criteria need to be established for the endoscopic diagnosis of Barrett esophagus, as the Prague C & M criteria proposed by IWGCO are not helpful for the diagnosis and classification of SSBE.^41,42 Second, a systematic method for the endoscopic surveillance of SSBE should also be established. Most Barrett cancer in Asian countries develops from SSBE. Although a stepwise 4-quadrant biopsy procedure is the gold standard for the surveillance of dysplastic Barrett esophagus in many Western countries, it is not an ideal method from the standpoint of safety, labor, or cost-effectiveness, and it should be avoided if guidance methods such as crystal violet chromoendoscopy,^98,99 magnifying endoscopy with acetate enhancement^100,101 or narrow band imaging,^102,103 or autofluorescence endoscopy¹⁰⁴ can be used to obtain the target biopsy. Crystal violet chromoendoscopy can detect dysplastic Barrett lesions with high sensitivity and specificity in both LSBE and SSBE cases. Magnifying endoscopy combined with acetate enhancement or narrow band imaging is also useful, though it is time-consuming in LSBE cases. These new endoscopic procedures may have advantages over the random biopsy diagnostic method in Asian patients with SSBE.

Strategies for the treatment and management of Barrett esophagus and the prevention of Barrett cancer should also be established in Asia. Proton pump inhibitors (PPIs) are used to shorten the length of Barrett esophagus. Nonsteroidal anti-inflammatory drugs (NSAIDs), such as aspirin, or selective cyclooxygenase-2 (COX-2) inhibitors are used to prevent Barrett carcinogenesis. PPIs suppress cellular proliferation^105,106 and facilitate the regression of Barrett esophagus.^107–109 NSAIDs such as aspirin are known to decrease the risk of many types of cancer, and in patients with Barrett esophagus, they decrease the risk of carcinogenesis from Barrett esophagus.^6,110–112 Recent studies have discussed the prophylactic effects of selective COX-2 inhibitors for Barrett carcinogenesis.^113–115 The efficacy of their drugs must be evaluated in Asian patients with Barrett esophagus, as little data from Asian countries are available.

Conclusion

Barrett esophagus is not rare, even in Asia, where the majority of it is in the form of SSBE. Endoscopic diagnosis of SSBE has several problems, and an endoscopic surveillance policy should be established. As patients with SSBE have a somewhat higher risk for developing cancer, an appropriate treatment strategy for SSBE should also be established in Asia.