Barrett's esophagus and the risk of adenocarcinoma

Abstract: Barrett’s esophagus is the precursor lesion of esophageal adenocarcinoma, a tumor with increasing incidence and poor prognosis. The overall aim of the thesis was to assess risk and prognosis in patients with Barrett’s esophagus and esophageal adenocarcinoma. Four studies were conducted based on data from Swedish nationwide registers and medical records from 71 Swedish hospitals. Study I was a population-based cohort study which assessed the risk of esophageal adenocarcinoma among patients with Barrett’s esophagus. Among 7,932 study participants with Barrett’s esophagus (median age 66 years, 68% men), 89 developed esophageal adenocarcinoma. After excluding prevalent adenocarcinomas (70%), 27 adenocarcinomas developed over a period of 18,415 person-years, which corresponded to an incidence rate of 1.5 (95% CI 0.9-2.0) cases per 1,000 person-years at risk and a standardized incidence ratio of 9.4 (95% CI 6.2-13.6). Study II was a population-based, nested case-control study designed to identify a prediction model for progression from Barrett’s esophagus to adenocarcinoma or high-grade dysplasia. All adenocarcinoma and high-grade dysplasia in patients with Barrett’s esophagus in Sweden were included as cases (n=279). Four randomly selected non-progressors per case were included as controls (n=1,089). For the included patients, endoscopy and histopathology records were collected and reviewed. Older age, male sex and longer Barrett’s esophagus segment length were associated with increased risk of adenocarcinoma/high-grade dysplasia. In contrast, hiatal hernia and esophagitis were not associated with tumor progression. A model based on age, sex and segment length predicted 71% of adenocarcinoma/high-grade dysplasia. Study III was a population-based cohort study which evaluated the adherence to surveillance and treatment guidelines for Barrett’s esophagus. All patients with dysplastic Barrett’s esophagus in Study II were included and followed for median 3.9 years using nationwide registers. Among 211 participants (71% low-grade dysplasia, 29% high-grade dysplasia), 84% had a follow-up endoscopy, 17% received endoscopic therapy and 8% underwent esophagectomy. However, 60% were not managed in accordance with clinical guidelines, mainly due to under-surveillance. Risk factors for deviation from surveillance and treatment recommended in guidelines were low-grade dysplasia compared to high-grade dysplasia and longer segment length compared to shorter segment length, while treatment in surgical compared to gastroenterological departments was associated with recommended surveillance and treatment. Study IV was a population-based cohort study which assessed whether endoscopy screening improves the prognosis of esophageal adenocarcinoma. Among 6,600 study participants with adenocarcinoma (mean age 70 years, 79% male) followed for 9,138 person-years, 7% had a history of gastroesophageal reflux disease and 9% underwent endoscopy before cancer diagnosis. The 5-year mortality was decreased in patients with history of gastroesophageal reflux disease (HR 0.71, 95% CI 0.64-0.80), and this decrease was only slightly attenuated by adjustment for prior endoscopy (HR 0.79, 95% CI 0.70-0.90). The 5-year mortality was unchanged in patients with 1-2 screening endoscopies (compared to patients without screening endoscopy), while those with ≥3 endoscopies for gastroesophageal reflux disease had improved survival in esophageal adenocarcinoma (HR 0.55, 95% CI 0.36-0.85). To conclude, the overall risk of adenocarcinoma in Barrett’s esophagus is low, but it is possible to predict a clearly higher risk of tumor progression based on a few clinically available risk factors, enabling tailored endoscopy surveillance in these patients. Currently, adherence to recommended surveillance and treatment guidelines is poor, and efforts to implement these guidelines in clinical practice are needed. Use of endoscopy screening has a limited impact on survival in adenocarcinoma unless performed frequently.