Patients With Barrett's Esophagus and Persistent Low-grade Dysplasia Have an Increased Risk for High-grade Dysplasia and Cancer

Background & aims: In some patients with Barrett's esophagus (BE) and a confirmed diagnosis of low-grade dysplasia (LGD), the LGD is not detected during follow-up examinations. We would like to avoid the unnecessary risks and costs of ablative treatment for these patients. Therefore, we investigated whether persistent LGD increases risk for high-grade dysplasia (HGD) or esophageal adenocarcinoma (EAC) and what proportion of patients are no longer found to have dysplasia after an initial diagnosis of LGD.

Methods: In a retrospective study, we collected information on 1579 patients with BE and LGD from 2005 through 2010 by using a nationwide registry of histopathology diagnoses in the Netherlands (PALGA). Confirmed LGD was defined as a diagnosis of LGD that was confirmed by any other pathologist. Persistent LGD was defined as LGD detected at the first and follow-up endoscopy. Data were collected on patients until treatment for HGD, detection of EAC, or the last endoscopy at which a biopsy was collected (through July 2014). We evaluated whether persistent LGD was a risk factor for malignant progression by using univariable and multivariable Cox regression analyses.

Results: Of individuals with BE and LGD in the database, the diagnosis of LGD was confirmed for 161 patients (10% of total). In these patients, the incidence of HGD and/or EAC was 5.18/100 person-years (95% confidence interval [CI], 4.32-8.10/100 person-years) compared with 1.85/100 person-years (95% CI, 1.52-2.22/100 person-years) in patients for whom LGD was not confirmed at the first endoscopy. The incidence of EAC alone in patients with confirmed LGD was 2.51/100 person-years (95% CI, 1.46-3.99/100 person-years), compared with 1.01/per 100 person-years (95% CI, 0.41-2.10/100 person-years) in patients for whom LGD was not confirmed at the first endoscopy. Of patients in whom LGD was confirmed at the first endoscopic examination, 51% were not found to have dysplasia at the first follow-up endoscopy, and 30% had persistent LGD. In patients with persistent LGD, the incidence of HGD and/or EAC was 7.65/100 person-years (95% CI, 4.45-12.34) and of only EAC was 2.04/100 person-years (95% CI, 0.65-4.92); in patients without persistent LGD, the incidence of HGD and/or EAC was 2.32/100 person-years (95% CI, 1.08-4.40/100 person-years) and of only EAC was 1.45 (95% CI, 0.53-3.21/100 person-years). Persistent LGD was found to be an independent risk factor for the development of HGD and/or EAC, with hazard ratio of 3.5 (95% CI, 1.48-8.28).

Conclusions: In a large population-based cohort study of patients with BE and LGD, the risk of progression to HGD and/or EAC was higher in patients with confirmed LGD and highest in those with confirmed and persistent LGD.