Endoscopic Surveillance for Barrett’s Esophagus: Two-Year Results

After more than 10 years of follow-up, scheduled endoscopy surveillance every two years failed to demonstrate an improvement in overall or cancer-specific survival, according to findings from the randomized BOSS trial published in Gastroenterology. As a result, at-need endoscopies were recommended by the study authors for low-risk patients with Barrett's esophagus.

"The burden of Barrett’s surveillance on endoscopic resources can be significant, and it may be appropriate to lengthen the surveillance interval or consider offering at-need endoscopy to well-informed patients, with an awareness of the risks and benefits of each option," said co-senior study author Paul Moayyedi, MD, Professor of Medicine, Division of Gastroenterology, McMaster University Medical Centre, Hamilton, Ontario, Canada.

Barrett's esophagus is a precursor lesion for esophageal adenocarcinoma. Population-based studies have estimated the annual risk of developing esophageal adenocarcinoma to be about 0.12% to 0.22%. Gastroenterology societies have advocated for regular surveillance endoscopies to detect esophageal adenocarcinomas early, but its use has not previously been examined in a randomized trial.

Observational studies have shown conflicting results regarding the benefits of surveillance for patients with Barrett's esophagus. A systematic review of all-cause mortality showed a reduction in mortality risk among patients in surveillance programs (hazard ratio [HR] = 0.75; 95% confidence interval [CI] = 0.59-0.94). Additionally, meta-analysis of esophageal adenocarcinoma‒related death showed that risk of mortality was reduced with surveillance (relative risk = 0.6; 95% CI = 0.5-0.71). The patients in surveillance programs were also more likely to be diagnosed with cancer at an earlier stage than individuals not under surveillance. 

The multicenter, randomized, controlled BOSS trial was conducted at 109 centers in the United Kingdom. A total of 3,453 patients were recruited and randomized to receive either two-yearly surveillance endoscopy or "at-need" endoscopy for symptom management only. All patients were diagnosed with Barrett's esophagus within 2 years of study enrollment. Patients with low-grade dysplasia were also eligible for study participation. 

In the endoscopic surveillance arm, patients underwent endoscopy every two years with quadrantic biopsies taken every 2 cm within the area of Barrett's esophagus. In the at-need arm, no endoscopies were scheduled until symptoms presented. If a patient developed dysphagia, unexplained weight loss, iron-deficiency anemia, recurrent vomiting, or worsening upper gastrointestinal symptoms, an endoscopy was offered regardless of the study arm. 

The primary endpoint was overall survival, and secondary outcomes included cancer-specific survival, time to diagnosis of esophageal adenocarcinoma, stage of gastroesophageal adenocarcinoma at diagnosis, serious adverse events related to endoscopy. For the overall survival endpoint, the study assumed that with 3,400 participants, a hazard ratio of 1.3 at 93% power would be allowed. 

Patients were followed for a median of 12.8 years, at which point 19.2% of patients in the surveillance arm had died as well as 20.7% of patients in the at-need arm. The mortality rates were not significantly different with surveillance endoscopy versus at-need endoscopy (HR = 0.95; 95% CI = 0.82-1.10; P = .503).  

In terms of cancer, 32.4% of patients in the surveillance arm died from any cancer and 29.8% from the at-need arm (HR = 1.01; 95% CI = 0.77-1.33; log rank P = .761). Only 2.3% of patients in the surveillance arm and 1.8% from the at-need arm were diagnosed with esophageal adenocarcinoma (HR = 1.32; 95% CI = 0.82-2.11; log rank P = .210). This amounted to a 0.23% risk per patient per year. The stage of esophageal adenocarcinoma at diagnosis was also similar between the two arms, with the exception of a trend toward more T1a cancer diagnoses in the surveillance arm. About 50% of patients with esophageal adenocarcinoma in each arm died during the trial period. 

A total of 6,124 endoscopies were performed in the surveillance arm compared with 2,424 in the at-need arm, for an adjusted endoscopy rate ratio of 1.62 (95% CI = 1.55-1.70; P < .01). Fifty-nine percent of patients in the at-need arm received at least one endoscopy during the study. The median time between endoscopies was 24.8 months in the surveillance arm and 25.7 months in the at-need arm, though the study authors noted that this could have been due to COVID-19. 

Serious adverse events were reported in eight patients in the surveillance arm and in seven patients in the at-need arm, with a total of four events likely related to endoscopy (three in the surveillance arm and one in the at-need arm). 

The Data and Safety Monitoring Committee recommended that all patients in the at-need arm be offered an end-of-trial endoscopy. A total of 448 end-of-trial endoscopies were performed, with an additional eight cases of esophageal adenocarcinoma detected plus nine cases of high-grade dysplasia. 

The overall risk for patients with Barrett's esophagus developing esophageal adenocarcinoma in the study was in line with the incidence rates from large population-based studies that have explored progression in Barrett's esophagus, noted Dr. Moayyedi and co-authors. The findings of the BOSS study challenge current clinical practice that encourages the use of surveillance for all patients with Barrett's esophagus by showing that there was no long-term advantage for patients randomized to endoscopic surveillance every two years over at-need endoscopy. Surveillance did not reduce the number of cases of esophageal adenocarcinoma or the risk of cancer death.

Although implemented after the initiation of this trial, current guidelines recommend an endoscopy every three to five years for patients with Barrett's esophagus, so the interventional arm does not reflect current clinical practice. The study authors noted that "our data indicate [three to five years] may be too aggressive, as any benefit is likely to be modest for low-risk patients with nondysplastic Barrett's esophagus." They added that risk stratification needs to be assessed by experts in guideline groups. 

The authors disclosed having no conflicts of interest.

Source: Gastroenterology