BACKGROUND: Acute pancreatitis after ERCP is a severe side effect. AIM: To evaluate the preventive effect of nitroglycerin on post-ERCP pancreatitis by a meta-analysis of randomized clinical studies. METHODS: We searched on Pubmed, Embase, Cochrane Library and all abstracts presented at Digestive Disease Week and United European Gastrointestinal Week from 2004 to 2008. We used the MeSH terms 'pancreatitis' together (AND) with the terms: 'glyceryl trinitrate', 'glyceryl dinitrate', 'isosorbide dinitrate' or 'nitroglycerin'. RESULTS: Five clinical studies evaluating the incidence of post-ERCP pancreatitis after administration of nitroglycerin were identified. Meta-analysis including all five studies showed a relative risk (RR) of 0.61 (95% CI; 0.44, 0.86) with the number needed to treat (NNT) of 26 (95% CI: 16, 82). Three studies evaluated nitroglycerin administered by a dermal patch reaching together an RR of 0.66 (95% CI; 0.43, 1.01). The use of nitroglycerin is associated with a significantly increased risk of hypotension (RR 2.25) and headache (RR 3.64). No difference in mortality was observed. CONCLUSIONS: Overall, our meta-analysis supports the use of nitroglycerin in the prevention of post-ERCP pancreatitis, but administration of nitroglycerin by the dermal route, which is the preferred route of administration, did not reach statistical significance.