Purpose: Review our experience with AP in children and identify predictors of Complications (C). Methods: Retrospective review of charts of patients admitted to LL Children's Hosp with AP between 02/02 and 03/09. AP was diagnosed if patient had abdominal pain &/or vomiting and elevated serum amylase (SA) or lipase (SL) (at least 2x). Records were reviewed in detail including history, physical examination, lab tests, imaging studies and C. Univariate (UV) and multivariate (MV) logistic regression analyses were done to determine if any clinical parameter could predict occurrence of C. Results: There were total 108 patients (50M: 58F) with AP during study period with age ranging from 9mo to 20yr (average 10.6 yr, 15% were <5yr and 58% were >10yr). Presenting symptoms were abdominal pain (82%), vomiting (71%), nausea (51%), diarrhea (19%) and abdominal distension (2%). Peak SA was 496 IU/L and SL was 1173 IU/L. Of 108 cases, etiology could be identified in 63 (gall stones-9, hereditary pancreatitis with abnormal PRESS or SPINK -9, trauma-9, autoimmune dis-9, cystic fibrosis-6, misc-23) while 45 (42%) were idiopathic. 75 (69%) patients developed at least one of the following C: abnormal liver enzymes-53, low serum albumin-40, ascites-20, pleural effusion-24, infection-14, hyperglycemia-9, pseudocyst-8, necrotizing pancreatitis-7 and others-17. 1 patient died of cause unrelated to pancreatitis. Average length of stay was 9.3(1-66) days. All except 2 patients had pancreatic imaging by either ultrasound (US), CT, MRCP or ERCP. US was abnormal in 47/90(52%), CT in 42/56(75%), MRCP in 12/20(60%) and ERCP in 7/7 (100%) cases. UV analysis of the effect of initial clinical measure on the occurrence of C revealed high AST on admission (OR1.08, p 0.005), low serum albumin (OR 0.24, p 0.003), number of imaging abnormalities (OR 2.83, p 0.001) and first episode of recurrent pancreatitis (OR 0.27, p 0.005 for subsequent episode) were associated with occurrence of C. When analyzed using MV models, recurrence had no effect on C. Age and sex also had no effect on C. Conclusion: 1. AP in children is associated with significant occurrence of C. 2. Following variables were associated with occurrence of C: high AST, low serum albumin and imaging abnormality (higher number of abnormalities were associated with increased risk of C). 3. Prospective studies are needed to confirm usefulness of above in predicting complications in pediatric patients with AP.