of opioid delivery were recorded. Primary endpoints included length of stay (LOS), readmis- sion rates, and subsequent outpatient opioid use. Secondary endpoints included time to oral intake and adverse drug events. Data was analyzed using General Linear Mixed Models with links corresponding to the distribution of outcomes. Results: A total of 763 patients with AP were identified (116 receiving PCA and 647 patients receiving IVP analgesia). There was no significant difference in age, sex, race, or etiology between the two groups but patients who received PCA had greater disease severity (p<0.001, Table 1). Patients receiving PCA had increased LOS compared with patients receiving IVP (least-squares mean of 9.9 days [95% CI: 8.7-11.3] vs. 6.4 days [95% CI: 5.7-7.1], respectively; p<0.001, Table 2). In addition, patients within the PCA group took longer to transition to oral opioids (least- squares mean of 6.8 days [95% CI: 5.9-7.9] vs. 4.1 days [95% CI: 3.6-4.6], respectively; p<0.001), had a greater time to oral feeding (6.3 days [95% CI: 5.5-7.3] vs. 3.5 days [95% CI: 3.1-3.9]; p<0.001), and required opioids upon discharge more often (OR: 2.7; 95% CI: 1.6-4.3) compared with patients receiving IVP despite disease severity. While altered mental status (OR: 5.9; 95% CI: 2.1-16.7) and opioid-induced constipation/ileus (OR: 3.2; 95% CI: 1.8-5.8) occurred more often with PCA, there were no differences in readmission rates (OR: 1.4; 95% CI: 0.88-2.2), respiratory depression (OR: 1.4; 95% CI: 0.44-4.5), or naloxone usage (OR: 1.3; 95% CI: 0.76-2.1). Conclusion: In the setting of AP, the use of PCA resulted in increased LOS, longer time to oral intake, greater outpatient opioid use, and higher rates of adverse drug events independent of disease severity suggesting that traditional IVP opioid administration may be optimal. Future randomized controlled trials comparing IVP and PCA are needed to identify the most effective method of opioid administration in AP. Sa1362 INCREASED FLUID RATE HASTENS RECOVERY OF ACUTE KIDNEY INJURY IN ACUTE PANCREATITIS Liam Hilson, Brent Hiramoto, Selena Zhou, Alexander Tonthat, Carlos Buitrago, Nicole Evans, Lee Helen, James L. Buxbaum INTRODUCTION: While clinical guidelines suggest that aggressive hydration should be administered for acute pancreatitis, it is based on limited evidence. Theoretically, ample fluids may improve pancreatic and renal perfusion. Our aim was to assess whether there is an association between the rate of fluid resuscitation and clinical outcome in those with concomitant pancreatitis and acute kidney injury. METHODS: A granular, prospectively ascertained data set of patients with acute pancreatitis presenting between 2015 and 2018 was used for the analysis. Acute kidney injury on presentation was defined as a creatinine increase of ≥0.3mg/dL from baseline. The primary endpoint, early recovery of renal function, was defined as a return to baseline creatinine within 48 hours. The primary predictor was S-333 AGA Abstracts the weight-based intravenous fluid administration rate (mL/kg/hr) within the first 24 hours from the time of pancreatitis diagnosis. Logistic regression models were used for hypothesis testing. RESULTS: Among 742 unique patients presenting with acute pancreatitis, 64 pre- sented with concomitant acute kidney injury. The mean age of the patients was 52.2 (+/- 15.5) years and 44 (69%) were male. The mean admission creatinine in the cohort was 2.06mg/dL. There was a significant correlation between early recovery of renal function and weight-based rate of intravenous fluids. The odds of recovery increased 1.6 fold for each 1mL/kg/hr increase in fluid rate (OR 1.6, 95% CI[1.01-2.42]). Receiver operating characteris- tic (ROC) analysis indicated that the minimal rate of resuscitation to favor recovery was 2mL/kg/hr. CONCLUSIONS: Early aggressive intravenous fluid administration within the first 24 hours of diagnosis hastens recovery from acute kidney injury in the setting of acute pancreatitis. The rate of hydration should be at least 2mL/kg/hr. Sa1363 USE OF ACOUSTIC RADIATION FORCE IMPULSE ELASTOGRAPHY TO PREDICT THE SEVERITY OF ACUTE PANCREATITIS (BASED ON BEDSIDE INDEX FOR SEVERITY IN ACUTE PANCREATITIS (BISAP) SCORE): A CASE-CONTROL STUDY Kuan-Chih Chen, Hsiu-Po Wang, Tien-Yu Huang, Cheng-Kuan Lin, Tzong-Hsi Lee, Chien-Chu Lin, Cheng-Lu Lin, Chen-Shuan Chung Background and Aim: Diagnosis of uncomplicated acute pancreatitis (AP) by trans-abdomi- nal ultrasonography is not sensitive. We try to demonstrate the potential value of acoustic radiation force impulse (ARFI) elastography on prediction of AP. Methods: We performed ARFI elastography (Siemens Acuson S2000™ ultrasound system: median value of 3 valid measurements and median value, meters/second (m/s)) on 20 AP patients and 43 control patients who received ultrasonography for indications other than AP (non-AP group) between March 2016 and March 2017. Measurement of ARFI at pancreatic body was carried out via transhepatic route. The ARFI mean values, clinical, laboratory data and bedside index for severity in acute pancreatitis (BISAP) score of two groups were analyzed. Results: More male patients were noted in AP group (male 80.0% vs. 37.2%, p=0.002). The mean age of AP and non-AP groups were 42.80 ± 10.67 years and 49.70 ± 15.79 years (p=0.08), respectively. Etiologies for AP were alcohol (13/20, 65%), followed by biliary tract stones (3/20, 15%), hypertriglyceridemia (2/20, 10%), and unknown cause (2/20, 10%). Most (18/ 20, 90%) of ARFI elastography in AP group were performed within 24 hours after the hospitalization. The mean values of ARFI in AP and non-AP groups were 1.88 ± 0.72 m/s and 1.13 ± 0.29 m/s (p<0.001), respectively. Higher ARFI values were found in alcohol (2.08 ± 0.80, p<0.001) and hypertriglyceridemia (1.66 ± 0.10, p=0.003) related AP than control group. Using cut-off value of ARFI 1.32 m/s for diagnosis of AP had sensitivity and specificity of 85.0% and 79.1%, respectively (AUC ROC 0.86). ARFI elastography had no value in prediction of serum lipase level (P=0.412), BISAP score (P=0.936) and the length of hospitalization days (P=0.831). Conclusions: Higher ARFI was found significantly in AP patients, especially for alcohol and hypertriglyceridemia related causes. However, the value of ARFI was not associated with clinical severity. Further study with larger number of study subjects is warranted to demonstrate the diagnostic and prognostic role of ARFI elastography in AP patients. Sa1364 PANCREATIC CANCER IN ACUTE PANCREATITIS: A CASE CONTROL STUDY Maria B. Baldursdottir, Berglind Magnusdottir, Evangelos Kalaitzakis, Einar Bjornsson Background: Pancreatic cancer (PaC) is one of the most common malignancies in the western world. Five year survival is poor as the disease is consistently diagnosed at an advanced stage. While chronic pancreatitis (CP) is a well established risk factor for PaC, the association between acute pancreatitis (AP) and PaC is less clear. AP is a rare manifestation of PaC but it is unclear whether AP is an indicator of pancreatic malignancy in patients diagnosed with PaC following AP. This study aimed to investigate the frequency of PaC following first-time AP in a population-based setting and to compare patients diagnosed with PC with or without a prior episode of AP in terms of outcome. Methods: This was a comparative case-control study. All patients with first-time AP in Iceland during 2006-2015 were retrospectively enrolled (n=1102). Patients with PaC after a diagnosis of first time AP were compared to patients diagnosed with PaC without prior AP. Medical files were scrutinized and relevant data including etiology of AP and diagnosis of PaC extracted. Each patient was matched randomly with two PaC patients of the same gender and age (±5 years) from a cohort of 239 patients diagnosed with PaC between 1998-2009 in Iceland, also identified retrospec- tively. Data on location of pancreatic tumors were extracted for both cohorts. Both cohorts were complemented with data from the national cancer registry in Iceland which is essentially complete, in order to confirm the diagnosis of cancer (searched until 2016). Results: A total of 23 of 1102 (2%) patients with first-time AP were diagnosed with PaC (52% female, mean age 67 yrs for cases and 63 yrs for PC controls (p=0.4). PaC presented concomitantly with AP in 39% of patients whereas in the majority (61%) PaC was diagnosed after a median of 28 (range 2-80) months after the initial AP had resolved. Location of pancreatic tumors within the pancreas did not differ significantly between groups (data not shown). In all, 74% of patients with PaC after AP had metastatic disease at diagnosis compared to 56% of controls (p=0.09). Survival after diagnosis of PaC was 4.4 months (95% CI 3.29-5.42) in the case group and 3.5 months (95% CI 1.97-4.96) in the control group and did not differ between the two groups (log rank test, p=0.49). Conclusion: PaC was diagnosed in 2% of patients with a first-time AP. AP as a presentation of PaC did not appear to affect survival when compared to patients diagnosed with PaC without prior AP. Location of pancreatic tumors did not differ between the two groups. The majority of patients in both groups had metastatic, non-resectable disease at diagnosis. Early diagnosis of pancreatic cancer remains a challenge and the clinical relevance of AP in patients who are later diagnosed with PaC needs further study. AGA Abstracts