a tube to decompress in SAP is uncertain. Therefore, a randomized trial was conducted to compare the effects of oral feeding without decompression with NG feeding in patients with predicted SAP. METHODS A total of 83 patients with predicted SAP were randomized to oral feeding (OF) group (n=41) or NG feeding (NGF) group (n=42). Predicted SAP was defined as AP patients admitted to hospital within 48 hours of onset, with one of the followings: 1) APACHE II score ≥8; 2) massive peripancreatic fluid collection on CT scan. All patients were under fasting once admitted. Patients in NGF group underwent insertion of NG tube for gastrointestinal decompression and tube feeding, while patients in OF group accepted oral feeding without the decompression as above. Feeding in both groups initiated when abdominal pain was decreasing and bowel movement was recovering. The primary outcomes were the occurrence of infectious complications and mortality, secondary outcomes were throat discomfort measured by numerical rating scale (0~10 score), length and cost of hospitalization. RESULTS No significant differences were found between the groups in demographics, APACHE II score on admission and timing of feeding. One patient in OF group underwent insertion of NG tube because of gastric retention and recurrent vomiting on day 2. The occurrence of pulmonary infection was significantly lower in OF group (7.32% vs. 23.81%, P<0.05), whereas the incidence of pancreatic infection and peripancreatic necrosis was similar in the two groups (2.44% vs. 4.76%, P>0.05). One patient in each group died during the study period (2.38% vs. 2.44%, P>0.05). The throat discomfort score in OF group was significantly lower than that of NGF group (1.87±0.76 vs. 7.13±2.05, P<0.05). The patients in OF group had shorter hospital stay (13.26±3.29 days vs. 16.48±4.13 days, P<0.05) and less cost of hospitalization (RMB 40269.15±7238.29 vs. 54286.27±10253.48, P<0.05) when compared with NGF group. CONCLUSIONS This study suggests that the efficacy of gastrointestinal decompression in the treatment of predicted SAP is limited. Compared with NG feeding, oral feeding in patients with predicted SAP is more effective in prevention of pulmonary infection, more tolerable since it decreases throat discomfort and presents better cost-effect. KEY WORDS oral feeding, nasogastric feeding, predicted severe acute pancreatitis, randomized clinical trial Mo1346 Peripancreatic Necrosis: Prevalence and Clinical Outcomes in a Large Cohort of Acute Pancreatitis Patients Efstratios Koutroumpakis, Alessandro Furlan, Anil Dasyam, Adam Slivka, David C. Whitcomb, Dhiraj Yadav, Georgios I. Papachristou Background: Acute pancreatitis (AP) is a common inflammatory process with variable course and severity. A revised AP severity classification was recently published, which modified the definitions of local complications. Based on the Revised Atlanta Classification (RAC), necrotizing AP refers to pancreatic parenchymal necrosis, but also pure necrosis of the peripancreatic tissues with normal pancreatic gland, as determined on contrast-enhanced CT (CECT). The natural history of peripancreatic necrosis (PeriPNec) is poorly described. The aim of our study is to report the prevalence of pancreatic necrosis (PNec) and PeriPNec in a large cohort of AP patients from a referral center in the US and compare their clinical outcomes. Methods: Patients with AP admitted at the University of Pittsburgh Medical Center were prospectively enrolled to the Pancreatitis-Associated Risk of Organ Failure (PROOF) study between the years 2003-14. Demographics and clinical data were collected prospectively. Initial and follow-up CECT scans were reviewed by abdominal radiologists with subspecialty training, who were blinded to patients' outcomes, following RAC defini- tions. Detailed data on the management of AP was abstracted retrospectively. Results: Four hundred patients with AP were prospectively recruited [mean age 52 (SD: 17), male 52%, median BMI 28 (IQR: 25-33), biliary etiology 39%, transfers 54%]. Two hundred fifty two patients (63%) underwent an initial and 141 (35%) a follow-up CECT. Initial and follow- up CECTs were performed at a median of 3 (1-9) and 37 (21-82) days from admission respectively. Among patients with CECT scans, normal pancreas was found in 23 (9%), interstitial pancreatitis in 90 (36%), and necrotizing pancreatitis in 139 (55%) patients. Among patients with necrotizing pancreatitis, PeriPNec was seen in 30 (12%) and PNec in 109 (43%; Table). Thirty-three % (10/30) of patients with PeriPNec developed persistent organ failure vs. 46% in the PNec group (p=0.31). Fewer patients with PeriPNec progressed to walled-off necrosis (WON) (10/30, 33%) and required drainage/debridement (7%) when compared to PNec (64%, 48%, respectively; p<0.001). PeriPNec patients stayed at the hospital for a median of 16 days (7-23 ) and 3 of them (10%) died during hospitalization, whereas the median length of stay for PNec patients was 20 days (10-35, p=0.08) and 5 died (5%). Conclusions: In contrast to previous reports, the prevalence of PeriPNec in our cohort was lower in comparison to PNec with a ratio of 1:4; this likely reflects interobserver variability in radiologists' interpretation. PeriPNec appears to have a favorable outcome with lower rates of progression to WON and interventions when compared to PNec. Table: Demographics and clinical outcomes in patients with necrotizing acute pancreatitis *WON: Walled-off necrosis; **LOS: Length of hospital stay S-679 AGA Abstracts Mo1347 Soluble Urokinase-Type Plasminogen Activator Receptor (suPAR) in Patients With Acute Pancreatitis; Progress or ‘suPAR' Revolution in Predicting Acute Pancreatitis Outcome Michal Lipinski, Alicja Rydzewska-Rosolowska, Andrzej Rydzewski, Grazyna Rydzewska Background: Serum soluble urokinase plasminogen activator receptor (suPAR) is a glycopro- tein secreted during inflammation and infections. Moreover, increased levels of suPAR are observed after hypoxia and ischemia. It has been suggested that suPAR could be used as a marker of severity and mortality in sepsis. The aim of the study was to assess whether suPAR could represent a useful marker of AP severity. Patients and Methods: We have observed a cohort of 47 prospectively enrolled patients. Based on the presence of persistent organ failure (more than 48 h) and local complications (diagnosis of moderate AP (MoAP)), patients were classified into three groups, mild AP (MAP), moderate (MoAP) and severe AP (SAP). BISAP score was determined in all patients within the first 24 h from admission. The blood samples were taken on admission for detecting suPAR concentrations from the 1st. day. suPAR testing was performed using suPARnostic © assay. Results: AP was considered severe in 10 patients (21.3%) , MoAP was found in 16 patients (34%) and MAP in 21 patients (44.7%). There were statistically significant trends for increasing severity (P<0.001) and mortality (P<0.02) with increasing suPAR concentration. The AUC for SAP predicted by suPAR was 0.966 (95% CI 0.919- 1). The calculated cut-off point for prognosis SAP is 4.3 ng/mL with 90% sensitivity and 94.6% specificity. BISAP score ≥ 3 for detection of SAP had sensitivity and specificity of 70%, 100%, respectively. The AUC for severity predicted by BISAP amounted to 0.975 (95% CI, 0.945-1). For a cut-off value of >3 ng/mL, suPAR had 73 % sensitivity and 81 % specificity for predicting MAP [AUC=0.81 (95% CI, 0.685- 0.935)]. Additionally suPAR turned out to be a good predictor of fatal AP - for the cut-off point 4.3 ng/mL the AUC was 0.894 (95%CI 0.698-1) with sensitivity 100% and specificity 80%. The AUC for death prediction in AP patients based on BISAP score ≥ 3 was 0.844 (95% CI 0.689-0.999), with sensitivity 50% and specificity 86.6%. Conclusions: suPAR concentration is a promising new diagnostic and prognostic indicator in SAP obtainable in the early stage of disease. suPAR could prove to be reliable tool for early prediction of AP, which may open new avenues for direct management and improve outcome. Larger studies are recommended to evaluate this role further. Mo1348 Obesity As a Predictor of Severity in Acute Triglyceride Pancreatitis in Hispanic and African-American Population Jasbir S. Makker, Bharat Bajantri, Kanthi Rekha Badipatla, Hafiz Rizwan Talib Hashmi, Mohammad Alshelleh, Bhavna Balar, Sridhar Chilimuri BACKGROUND: Hypertriglyceridemia is the third most common cause of acute pancreatitis after gallstone and alcohol. Little is known about the effect of bodyweight on the course of acute triglyceride pancreatitis. METHODS: We reviewed the electronic medical records of 304 patients admitted with acute pancreatitis to an inner city hospital in New York from 2002 to 2013. Patients with acute pancreatitis secondary to hypertriglyceridemia were identified. These patients were divided into two groups based on their BMI. One group included patients with BMI <25 while the other group included all patients with BMI 25 and above (overweight and obese). The demographic and laboratory data were collected for two groups. Severity of pancreatitis was also determined and categorized into three types - mild, moderately severe and severe pancreatitis. Statistical analysis was done using fisher's exact t-test for categorical data and t-test for continuous data. RESULTS: A total of 74 patients were identified with 57 patients (mean age 47, 53% male, 51% Hispanics, 35% Blacks) in the overweight and obese group and 17 patients (mean age 48, 65% male, 53% Hispanics, 35% Blacks) in the normal BMI group. In the overweight and obese group, 42 patients had mild pancreatitis and 15 patients (11+4) had moderately severe or severe pancreatitis as compared to 17 patients with mild pancreatitis and none with moderately severe or severe pancreatitis in the normal BMI group (p 0.016). Length of stay was signifi- cantly longer for patients with BMI more than 25 group as compared to patients with normal BMI (8.9 in BMI <25 vs 6.2 in BMI ≥ 25, p 0.03). There was no significant difference in length of stay between men and women (8.4 vs 8.0, p 0.82), between different ethnic groups (7.6 in Hispanics vs 7.8 in Blacks, p 0.86) or between different serum triglyceride levels (6.4 in TG 1000-2000 vs 8.9 in TG >2000, p 0.07). CONCLUSION: Overweight and obese patients admitted with acute triglyceride pancreatitis tend to have a more complications as compared to patients with normal BMI. BMI can serve as a marker of severity in patients with acute triglyceride pancreatitis and further larger studies are needed to validate the effect of BMI on course of acute triglyceride pancreatitis. Mo1349 Validation of Japanese Severity Score for Acute Pancreatitis Using Japan National Administrative Database 2010-2012 Kazuki Someya, Keiji Muramatsu, Naoki Takahashi, Tatsuo Takama, Hiroki Otsubo, Takashi Kido, Shinya Matsuda, Toshihiko Mayumi Background The criteria for severity assessment of acute pancreatitis in Japan were deter- mined in 1990 (of which a partial revision was made in 1999). In 2008, an overall revision was made and the diagnosis of severe acute pancreatitis can be made according to 9 prognostic factors and/or the contrast-enhanced computed tomography (CE-CT) grades based on con- trast enhanced CT. But they have not been assessed in large studies. Here we verified accuracy of them using Japan National Administrative Database. Aim To evaluate the accuracy of Japanese Severity Score (JSS) for acute pancreatitis. Methods Using Japan national adminis- trative database 2010-2012, we compare JSS (prognostic factors, CE-CT grades, Table 1) and hospital mortality of acute pancreatitis. Prognostic factors consist of the following 9 items: (1) base excess (BE) % -3 mEq/L or shock: (systolic blood pressure % 80 mmHg), (2) PaO 2 % 60 mmHg (room air) or requiring respirator management, (3) blood urea nitrogen (BUN) ^ 40 mg/dl (or creatinine [Cr] ^2.0 mg/dl) or oliguria after fluid replacement, (4) lactic dehydrogenase (LDH) ^ 2 times of upper limit of normal, (5) platelet count % AGA Abstracts