Risk of progression of Barrett's esophagus in patients with cirrhosis

AIM To study Barrett's esophagus (BE) in cirrhosis and assess progression to esophageal adenocarcinoma (EAC) compared to non-cirrhotic BE controls. METHODS Cirrhotic patients who were found to have endoscopic evidence of BE confirmed by the presence of intestinal metaplasia on histology from 1/1/2000 to 12/1/2015 at Cleveland Clinic were included. Cirrhotic patients were matched 1: 4 to BE controls without cirrhosis. Age, gender, race, BE length, hiatal hernia size, Child-Pugh (CP) class and histological findings were recorded. Cases and controls without high-grade dysplasia (HGD)/EAC and who had follow-up endoscopies were studied for incidence of dysplasia/EAC and to assess progression rates. Univariable conditional logistic regression was done to assess differences in baseline characteristics between the two groups. RESULTS A total of 57 patients with cirrhosis and BE were matched with 228 controls (BE without cirrhosis). The prevalence of dysplasia in cirrhosis and controls were similar with 8.8% vs 12% with low grade dysplasia (LGD) and 12.3 % vs 19.7% with HGD or EAC (P = 0.1). In the incidence cohort of 44 patients with median follow-up time of 2.7 years [interquartile range 1.0, 4.8], there were 7 cases of LGD, 2 cases of HGD, and 2 cases of EAC. There were no differences in incidence rates of HGD/ EAC in nondysplastic BE between cirrhotic cases and noncirrhotic controls (1.4 vs 1.1 per 100 person-years, P = 0.8). In LGD, cirrhotic patients were found to have higher rates of progression to HGD/EAC compared to control group though this did not reach statistical significance (13.7 vs 8.1 per 100 person-years, P = 0.51). A significant association was found between a higher CP class and neoplastic progression of BE in cirrhotic patients (HR = 7.9, 95% CI: 2.0-30.9, P = 0.003). CONCLUSION Cirrhotics with worsening liver function are at increased risk of progression of BE. More frequent endoscopic surveillance might be warranted in such patients.