Original article Patterns and predictive factors of complications after endoscopic retrograde cholangiopancreatography T. Glomsaker 1,4 , G. Hoff 5,6,7 , J. T. Kvaløy 3 , K. Søreide 1,4 , L. Aabakken 2 and J. A. Søreide 1,4 , on behalf of the Norwegian Gastronet ERCP Group Departments of 1 Gastroenterological Surgery and 2 Medicine, Stavanger University Hospital, and 3 Department of Mathematics and Natural Sciences, University of Stavanger, Stavanger, 4 Department of Surgical Sciences, University of Bergen, Bergen, 5 Department of Medicine, Telemark Hospital, Skien, and 6 Faculty of Medicine, University of Oslo, and 7 Cancer Registry of Norway, Oslo, Norway Correspondence to: Professor J. A. Søreide, Department of Gastroenterological Surgery, Stavanger University Hospital, PO Box 8100, N-4068 Stavanger, Norway (e-mail: jon.soreide@kir.uib.no) Background: With an increased use of magnetic resonance imaging, the indications for endoscopic retrograde cholangiopancreatography (ERCP) have changed. Consequently, the patterns and factors predictive of complications after ERCP performed during current routine clinical practice are not well known. Methods: A prospective multicentre cohort study was undertaken in 11 Norwegian hospitals. Complications and mortality within 30 days after ERCP were analysed by univariable and multivariable regression analysis. Results: There were 2808 ERCP procedures, of which 2573 (91·6 per cent) were therapeutic. More than half of the patients were aged 70 years or more. Common bile duct cannulation was achieved in 2557 procedures (91·1 per cent). Complications occurred in 327 (11·6 per cent) of the procedures, including cholangitis in 100 (3·6 per cent), pancreatitis in 88 (3·1 per cent), bleeding in 66 (2·4 per cent), perforation in 25 (0·9 per cent) and cardiovascular–respiratory events in 32 (1·1 per cent). In the multivariable regression analysis, older age, increasing American Society of Anesthesiologists fitness score, centre ERCP volumes of more than 150 procedures annually and precut sphincterotomy were predictive factors for severe complications. The overall 30-day mortality rate was 2·2 per cent (63 patients), with a procedure-related mortality rate of 1·4 per cent (39 patients). Malignancy was diagnosed in 46 (73 per cent) of the patients who died. Conclusion: ERCP is a procedure with considerable risk for complications. Morbidity and mortality are related to patient age and co-morbidity, as well as hospital volume of ERCP procedures and the type of intervention. Paper accepted 27 September 2012 Published online 6 December 2012 in Wiley Online Library (www.bjs.co.uk). DOI: 10.1002/bjs.8992 Introduction In recent years various clinical features of endoscopic retrograde cholangiopancreatography (ERCP), including technical performance and complications 1–4 , quality assessment 5–8 and aspects of education 9 , have been addressed. However, outcomes reported from highly dedicated centres do not necessarily mirror general ERCP practice 2 . To evaluate complications and possible risk factors for these complications in unselected populations, several multicentre studies have been performed 3,4,10–17 . Nevertheless, a number of concerns, including study design and differences in study populations, definitions and their interpretations, as well as the quality and completeness of the registrations, make it difficult to evaluate and compare reported figures in the various studies 18 . A meta-analysis of patient-, procedure- or operator-dependent factors associated with ERCP complications has supported these concerns 19 . The evaluation of general ERCP practice remains important because of possible changes over time, such as the implementation of other imaging modalities, and different clinical traditions across continents, as well as  2012 British Journal of Surgery Society Ltd British Journal of Surgery 2013; 100: 373–380 Published by John Wiley & Sons Ltd