Predictors of esophageal stricture formation post endoscopic mucosal resection.

BACKGROUND/AIMS: Stricture formation is a common complication after endoscopic mucosal resection. Predictors of stricture formation have not been well studied. METHODS: We conducted a retrospective, observational, descriptive study by using a prospective endoscopic mucosal resection database in a tertiary referral center. For each patient, we extracted the age, sex, lesion size, use of ablative therapy, and detection of esophageal strictures. The primary outcome was the presence of esophageal stricture at follow-up. Multivariate logistic regression was used to analyze the association between the primary outcome and predictors. RESULTS: Of 136 patients, 27% (n=37) had esophageal strictures. Thirty-two percent (n=44) needed endoscopic dilation to relieve dysphagia (median, 2; range, 1 to 8). Multivariate logistic regression analysis showed that the size of the lesion excised is associated with increased odds of having a stricture (odds ratio, 1.6; 95% confidence interval, 1.1 to 2.3; p=0.01), when controlling for age, sex, and ablative modalities. Similarly, the number of lesions removed in the index procedure was associated with increased odds of developing a stricture (odds ratio, 2.3; 95% confidence interval, 1.3 to 4.2; p=0.007). CONCLUSIONS: Stricture formation after esophageal endoscopic mucosal resection is common. Risk factors for stricture formation include large mucosal resections and the resection of multiple lesions on the initial procedure.

Survival in esophageal high-grade dysplasia/adenocarcinoma post endoscopic resection.

BACKGROUND: Endoscopic resection followed by ablative therapy is frequently used to treat esophageal high-grade dysplasia or early esophageal adenocarcinoma. AIMS: To study outcomes in patients with high-grade dysplasia compared to those with esophageal adenocarcinoma after endoscopic resection. METHODS: Retrospective, observational, descriptive, single-centre study from a prospective database. We extracted data from 116 endoscopic resections. Survival was plotted using Kaplan-Meier curves multivariable Cox-proportional hazard assess for possible predictors of survival post-endoscopic resection was performed. RESULTS: 116 patients (64 esophageal adenocarcinoma, 52 high-grade dysplasia) underwent endoscopic resection from May 2003 to June 2010. Mean age was 71 ± 11 years for high-grade dysplasia and 72 ± 10 years for esophageal adenocarcinoma. Median follow-up was 17 months. Eighty-five patients had negative margins on endoscopic resection. Five-year survivals for high-grade dysplasia and esophageal adenocarcinoma were 86% (range 68-100%) and 78% (59-96%), respectively. Survival was not significantly different between groups (p=0.20). Overall mortality rate was 10.6% (9/85). At multivariable Cox regression increased Barrett's oesophagus length was associated with worse survival (HR 1.18 [1.06-1.33], p=0.0039). Survival was not affected by the pathology before resection: HR 2.4 [95%CI, 0.70-8.4], p=0.16. CONCLUSIONS: Survival in patients with high-grade dysplasia of the oesophagus is similar to those with esophageal adenocarcinoma. Longer Barrett's oesophagus segments are associated with decreased survival.

Factors associated with increased bleeding post-endoscopic mucosal resection.

OBJECTIVE: Our aim was to identify patient and procedure characteristics that correlate with increased likelihood of bleeding during and after endoscopic mucosal resection (EMR), and thus anticipate the need for preventative therapy. METHODS: This was a retrospective, observational, descriptive study using a prospective EMR database, performed in a tertiary-care center. A total of 935 EMR of various locations within the gastrointestinal tract were collected. The main outcome measurement was early bleeding (occurring during the procedure) and delayed bleeding (occurring after the completion of the procedure and up to 30 days after). RESULTS: Early bleeding occurred in 5.3% (50/935) of the procedures. In multivariate logistic regression analysis, esophageal EMR (OR 2.5, 95% CI 1.2-5, P = 0.0009) and increase in lesion size (OR 1.24, 95% CI 1.1-1.5, P = 0.003) were both associated with higher odds of early bleeding in EMR when controlling for age, gender and non-steroidal anti-inflammatory drugs (NSAIDs)/clopidogrel use. Delayed bleeding occurred after 3.1% (n = 29) of the procedures. Of these, 86.2% (25/29) required hospital admission and endoscopic intervention to confirm and/or treat bleeding site. In multivariate logistic regression analysis, increased lesion size (OR 1.3, 95% CI 1.1-1.5, P = 0.004) was associated with higher incidence of delayed bleeding post-EMR. CONCLUSIONS: In experienced hands bleeding during and after EMR appears to be uncommon. Larger lesions are at increased risk of early and delayed bleeding.