Specialized intestinal metaplasia, dysplasia, and cancer of the esophagus and esophagogastric junction: prevalence and clinical data.

BACKGROUND & AIMS: Adenocarcinoma of the esophagus and esophagogastric junction (EGJ) is increasing, the earliest lesion being specialized intestinal metaplasia (SIM). This study determined the prevalence and demographic features of patients with SIM, dysplasia, and cancer in the esophagus and EGJ.

METHODS

Two antegrade biopsy specimens were taken distal to the squamocolumnar junction (SCJ) and any tongues of pink mucosa proximal to the SCJ. Patients were categorized endoscopically and histologically as having long-segment (LSBE) or short-segment Barrett's esophagus (SSBE), EGJ-SIM, or a normal EGJ.

RESULTS

Of 889 patients studied, 56 were undergoing esophagoduodenoscopy screening or surveillance and were not included in the prevalence calculation. The overall prevalence of SIM was 13.2%, with 1.6% LSBE, 6.0% SSBE, and 5.6% EGJ-SIM. Dysplasia or cancer was noted in 31% of LSBE, 10% of SSBE, and 6.4% of EGJ-SIM patients (P </= 0.043). Two cancers were associated with LSBE, 1 with SSBE, and 1 with EGJ-SIM. Patients with LSBE and SSBE were predominantly white (P </= 0.001), male (P </= 0. 009), and smokers (P </= 0.004), with LSBE patients having a longer history of heartburn (P </= 0.009). In contrast, patients with EGJ-SIM were similar in gender and ethnicity to the reference group, tended to be older (P </= 0.05), drank less alcohol (P </= 0.02), and had a higher prevalence of Helicobacter pylori infection (P </= 0.05).

CONCLUSIONS

The prevalence of SSBE and EGJ-SIM is similar, but each entity is 3.5 times more prevalent than LSBE. However, the prevalence of dysplasia in LSBE is 2 times greater than in SSBE and 4 times greater than in EGJ-SIM. Demographically, EGJ-SIM patients are different from patients with Barrett's esophagus and have a higher prevalence of H. pylori infection. These data help to explain the increasing incidence of adenocarcinoma of the distal esophagus and EGJ.