Abstract
Background and study aims Post-ERCP pancreatitis (PEP) is the most common complication attributed to the procedure, its incidence being approximately 9.7 %. Numerous studies have evaluated the predictive efficacy of post-procedure serum amylase and lipase levels but with varied procedure-to-test time intervals and cut-off values. The aim of this meta-analysis was to present pooled data from available studies to compare the predictive accuracies of serum amylase and lipase for PEP. Patients and methods A total of 18 studies were identified after a comprehensive search of various databases until June 2021 that reported the use of pancreatic enzymes for PEP. Results The sample size consisted of 11,790 ERCPs, of which PEP occurred in 764 (6.48 %). Subgroups for serum lipase and amylase were created based on the cut-off used for diagnosing PEP, and meta-analysis was done for each subgroup. Results showed that serum lipase more than three to four times the upper limit of normal (ULN) performed within 2 to 4 hours of ERCP had the highest pooled sensitivity (92 %) for PEP. Amylase level more than five to six times the ULN was the most specific serum marker with a pooled specificity of 93 %. Conclusions Our analysis indicates that a lipase level less than three times the ULN within 2 to 4 hours of ERCP can be used as a good predictor to rule out PEP when used as an adjunct to patient clinical presentation. Multicenter randomized controlled trials using lipase and amylase are warranted to further evaluate their PEP predictive accuracy, especially in high-risk patients.