WITHDRAWN. Endoscopic retrograde cholangiopancreatography in gallstone-associated acute pancreatitis.

Background: Early endoscopic retrograde cholangio-pancreatography with or without endoscopic sphincterotomy (ERCP+/-ES) has been advocated to reduce complications in patients presenting with a severe attack of gallstone-associated acute pancreatitis (GAP). However, a recent trial has reported contradictory results. Importantly, patients with acute cholangitis were excluded suggesting it may be a major confounding factor affecting previous studies.

Objectives: To assess the effectiveness of early ERCP+/-ES compared to conservative management stratified according to severity of disease, concealment of randomisation, acute cholangitis and bilirubin level in the reduction of mortality, morbidity, length of hospitalisation and cost in adults suspected of having GAP.

Search Strategy: We searched - Cochrane Library (Issue 4 2003), Medline (1966-2004), EMBASE (1980-2004) and LILACS. 'Grey literature' was sought by looking at cited references and hand searched to identify further relevant trials. Conference proceedings of United European Gastroenterology Week (published in Gut) and Digestive Disease Week (published in Gastroenterology) were also hand searched.

Selection Criteria: Randomized controlled trials (RCT) of adult patients, from 15 years old or greater, presenting with gallstone-associated acute pancreatitis (GAP) comparing ERCP +/- ES versus Conservative management within 72 hours of admission.

Data Collection And Analysis: Data were assessed for quality independently by two reviewers. Wherever appropriate, results were pooled together and sub-grouped by predicted severity of disease. Fixed and random effects models were applied. Sensitivity analysis was performed to test the fragility of results.

Main Results: Three trials, involving 511 patients, met inclusion criteria. The test for heterogeneity yielded statistically non-significant results (p-value 0.1 to 0.63) suggesting all comparisons were above the established threshold for combinability (p<0.1). Fixed effect and random effect meta-analyses gave identical results. Early ERCP +/- ES was associated with non-significant effect on reduction of mortality in predicted mild (OR = 0.62, 95% CI = 0.27 to 1.41) and severe GAP (OR = 0.62, 95% CI = 0.27 to 1.41). Reduction in complications was non-significant in predicted mild (OR = 0.89, 95% CI = 0.53 to 1.49), but significant in severe GAP (OR = 0.27, 95% CI = 0.14 to 0.53). There was insufficient evidence to draw any conclusions about hospital stay and cost.

Authors' Conclusions: Odds of having complications are reduced in predicted severe disease by early ERCP +/- ES. This effect was however, non-significant in predicted mild disease and for reduction of mortality in either predicted mild or severe disease. These results are controlled for confounding due to associated acute cholangitis and are robust for clinical and statistical heterogeneity.