Introduction EMR is an established treatment for large colorectal polyps, yet the data regarding efficacy and outcome are principally limited to single centre experience. We present a multicentre study to determine prognostic factors of short-term (3 month) and long-term (1 year) outcomes following an index (intention to treat (ITT)) EMR for large sessile colorectal polyps.

Methods Endoscopy databases & hospital coding records identified patients that had an ITT EMR for colonic polyps 2 cm or greater between 2005 and 2010 in five North-West teaching hospitals. Patients were audited longitudinally for 1 year. Multivariate analysis (Logistic Regression Model) determined significant prognostic factors (OR [95% CI]; p<0.05 as significant).

Results Demographics: 313 patients (mean [SD] age 69.7[10.4] years, 65.5% male, 86% ASA grade 1–2) underwent EMR for mean polyp size of 33.5 [11.5] mm. Morphologically; 63.8% were flat lesions and 21.7% were located in the right colon. Procedure: 26% were performed by inexperienced endoscopists (Outcomes: Cancer diagnosis was in 9.5% (6.3% invasive & 3.2% intra-mucosal). Complications; Perforation rate 0.5%, bleeding rate 0.5% & all cause 30-day mortality 0.6 %, with no procedure related mortality. Recurrence rate was 26% & treated endoscopically in 64.6%, & surgically in 8.5%. Overall, surgery was required in 7.3%, of which 8.7% were emergencies (to treat perforation) and 56% was for cancer. ITT short-term success rate was 68% and 1-year success rate was 82.4%, with an adjusted rate of 87.2%. Prognostic Factors: Predictors of recurrence were cancer histology (OR 9 [95% CI 4 to 22]=<0.05), piecemeal resection (OR 4 [95% CI 1.5 to 11] p<0.005) and EMR session >1 (OR 22 [95% CI 10 to 50] p<0.005). Poor prognostic indicators for long-term success were cancer (OR 11 [95% CI 4.5 to 28] p<0.005) and EMR session >1 (OR 3.6 [95% CI 1.5 to 8.4] p=0.003). While endoscopist inexperience, increasing polyp size, no adjuvant APC were poor prognostic factors univariately, on multivariate analysis they were insignificant. Gender, age, ASA, Training, site, morphology and complications were not significant factors.

Conclusion While recurrence rates in EMR for large colonic lesions were high (>1/4), long-term outcomes were good (cure rate 87.2%) with complications similar to previous series. Most important poor prognostic factors were cancer histology and requirement of more than 1 session.

Competing interests J Geraghty grant/research support from: Cook Medical, K Bodger: None declared, S Alam: None declared, W Jafar: None declared, M Gillon: None declared, C Babbs: None declared, J Ramesh: None declared, R Willert: None declared, S Lal: None declared, S Sarkar: None declared.