Sa1068 Predictors of Progression in Barrett's Esophagus: A Systematic Review and Meta-Analysis

Background: The overall risk of progression to esophageal adenocarcinoma (EAC) in BE subjects without dysplasia is lower than previously estimated. There is hence lack of consensus on the effectiveness of current surveillance programs. Stratification of risk of progression is appealing, as it may help identify BE patients at higher risk who can be monitored more intensively or offered endoscopic therapy. Though several studies have identified factors predicting progression, evidence on several clinically relevant risk factors influencing progression is not consistent. We aimed to identify the predictors of progression in BE subjects without high grade dysplasia by performing a systematic review and meta-analysis of relevant studies. Methods: A comprehensive search of several databases including PubMed, EMBASE, and Web of Science databases (earliest inception to July 28th, 2014) was performed and studies reporting outcomes of BE cohorts under surveillance were identified. We included studies that reported predictors of progression to HGD/ EAC as an outcome in patients with non-dysplastic BE (NDBE) and BE-low grade dysplasia (LGD). A pooled estimate (odds ratio) of the potential of age, sex, smoking, alcohol use, BMI, baseline LGD (vs NDBE), BE length, and medication use in predicting progression to HGD/EAC in NDBE/ LGD patients was calculated. Results: 23 studies reporting predictors of progression as an outcome were identified from the systematic review. 1231 events in 74943 patients in 204926.2 patient years of follow-up were analyzed. Table 1 shows the pooled odds ratio (OR) of predictors analyzed. Increasing age, male sex, ever smoking (current or past), and increasing BE segment length were predictive of increased risk of progression to HGD/EAC with low heterogeneity (9-13 studies). Patients with LGD had an almost four fold increased risk of progression compared to those with no dysplasia. NSAID and statin use were protective against progression to HGD/EAC without heterogeneity. Alcohol use and obesity as measured by BMI did not influence progression rates. Conclusion: Age, male sex, ever smoking (current or past), and BE length are predictive of increased risk of progression to HGD/ EAC. Subjects with baseline LGD are at 4 times increased risk of progression compared to those without dysplasia. These factors may be used to risk stratify BE subjects as part of a prediction score, to identify those who could potentially benefit from intensive surveillance or endoscopic therapy. Smoking is a modifiable risk factor for cancer prevention in BE subjects. Table 1: Pooled odds ratio (OR) of predictors of progression to HGD / EAC.