1062 CFTR MUTATIONS RELEVANCE ON PANCREATITIS OUTCOME: GENOTYPE-PHENOTYPE CORRELATION AND IMPACT OF ASSOCIATED FACTORS Solène Dermine, Emmanuelle Girodon-Boulandet, Paul Calame, Marie-Pierre Vullierme, Thierry Bienvenu, Philippe Ruszniewski, Philippe Lévy, Vinciane Rebours Introduction: CFTR gene, coding for a chloride channel, may be responsible for pancreatitis. More than 2,000 mutations were described with variable functional impact and their clinical relevance on pancreatitis outcome is still debated. The objectives were: to describe the natural history of CFTR-associated pancreatitis; to assess a genotype-phenotype correlation; to specify the role of environmental risk factors; and to assess the risk of pancreatic cancer. Patients and Methods: Between 2002 and 2016, 1,092 patients were referred to an expert center to research for the cause of acute recurrent or chronic non-alcoholic non-biliary pancreatitis. After a standardized work-up, 240 patients with one or more CFTR mutations (without PRSS1 and SPINK1 mutation) were compared to 103 controls (idiopathic pancreati- tis). Clinical, biological and morphological data were collected retrospectively. Imaging procedures (CT scan and MRI) were all reviewed by an expert radiologist: calcifications, pancreatic ducts dilatation and pancreatic atrophy. CFTR genotypes were classified into 3 categories (causal, contributory or neutral) by an expert geneticist according to the assumed clinical impact on epidemiological and/or functional arguments. (1) Results: The median follow-up after the first symptom was 4.4 years. No clinical and morphological difference between CFTR and controls was found, except for a lower proportion of intraparenchymal calcifications in the CFTR group. Sixteen patients with a "causative" genotype developed chronic pancreatitis earlier than those with a "contributory" (n=139) or "neutral" genotype (n= 78) (29, 45 and 46 years respectively, p=0.003), with a similar pancreatic symptomatology. Pancreatitis symptoms and morphological abnormalities (calcifications, duct dilations) were more frequent in case of tobacco intake (52.5%), tobacco-alcohol association (84.6%) versus no toxic consumption (43.4%, p= 0.019). Pancreas divisum was more common in "causal" or "contributory" genotype (50 and 27.3% respectively, versus 16.5% in controls) but didn't change the age at onset, frequency of pancreatic symptoms or morphological abnormalities. No over-risk of pancreatic cancer in the CFTR group compared to the control group was found, (n=7 versus 2, p=NS, respectively). Conclusion: The genotype-phenotype relation in CFTR-associated pancreatitis is complex due to molecular heterogeneity and environmental interactions. CFTR mutation types did not influence natural history, except for the "causal" genotype. Pancreas divisum is most common in "causal" or "contributory" genotype, but cannot be considered as a cofactor. Associated environmental factors are probably the more deleterious factors influencing natural history. Ref:(1) Masson et al . PlosOne 2013. 1063 EXTRACORPOREAL SHOCK WAVE LITHOTRIPSY FOR CHRONIC PANCREATITIS PATIENTS WITH STONES AFTER PANCREATIC SURGERY Lu Hao, Yu Liu, Dan Wang, Jun-Tao Ji, Lei Xin, Zhuan Liao, Zhao-Shen Li, Liang-Hao Hu Background: Both ESWL/ERCP and surgery are first-line treatments for chronic pancreatitis (CP). However, parts of CP patients still suffer from relapsed pancreatic stones and pain after surgery. A redo operation is much challenging. For these patients, endotherapy may be a potential choice. This study aimed to evaluate the safety and efficacy of pancreatic extracorporeal shock wave lithotripsy (P-ESWL) for painful chronic pancreatitis (CP) patients with pancreatic stones after previous pancreatic surgery. Methods: This is a single center cohort study prospectively conducted in pained CP patients undergoing P-ESWL. Patients with pancreatic surgery history (PSH) were included in the PSH group, which was then further divided into decompression, resection and debridement subgroups. Patients without PSH in the corresponding period were assigned to the control group. Patients in PSH and control group were matched by age, sex, type of pain, and location of stone (s) in 1:1 proportion. The primary outcomes were complications of P-ESWL and pain relief. Secondary outcomes included stone clearance, physical and mental health, quality-of-life score and changes in exocrine and endocrine pancreatic function. Results: From March 2011 to October 2014, P-ESWLs were performed on 1,017 patients (50 in the PSH group, 967 in the control group). In the PSH group, the decompression, resection and debridement subgroups enrolled 36, 6 and 8 patients respectively. No significant difference was observed in the occurrence of P-ESWL complications between the PSH group and control group (14.0% vs. 13.2%, P=0.877, Table 1). Among the 48 patients (96.0%) of the PSH group who underwent follow-up for 2.6 years (range: 1.0-4.5), pain relief was achieved in 36 patients (75.0%) and complete clearance of the stones was seen in 37 patients (77.1%), with no significant differences compared to the matched controls. The visual analogue score decreased, scores for the quality of life and physical health from the SF-36 improved significantly (all P<0.001). Conclusions: P-ESWL is observed to be safe and effective for pained CP patients with stones after pancreatic surgery. For these patients, P-ESWL can safely achieve significant pain relief and stone clearance and as a result, prevent the need for a redo surgery (Figure 1). S-203 AGA Abstracts Table 1. P-ESWL complications of PSH group and control group Figure 1. Diagrams of stone clearance before and after P-ESWL Stones are fragmented into small pieces #3 mm in diameter by P-ESWL. (a) Pre-ESWL: relapsed stones in the MPD after LPJ procedure; (b) Post-ESWL: fragmented stones spontaneously flow out of the pancreatic duct through PJA of LPJ procedure or are extracted by a retrieval basket during ERCP; (c) Pre-ESWL: relapsed stones in the MPD after Whipple procedure; (d) Post-ESWL: fragmented stones spontaneously flow out of the pancreatic duct through PJA of Whipple procedure (illustrated by Ji JT). Abbreviations: MPD, main pancreatic duct; LPJ, longitudinal pancreatico-jejunostomy; PJA, pancreaticojejunal anastomosis 1064 SELENIUM DEFICIENCY IS ASSOCIATED WITH EXOCRINE PANCREATIC INSUFFICIENCY IN PATIENTS WITH CHRONIC PANCREATITIS Wei Zhan, Annette Wilson, Celeste Shelton, Julia B. Greer, Phil J. Greer, Dhiraj Yadav, David C. Whitcomb Background: Chronic pancreatitis (CP) is a pathologic fibro-inflammatory syndrome of the pancreas in individuals with genetic, environmental and/or other risk factors who develop persistent pathologic responses to parenchymal injury or stress. Features of progression to end-stage CP, such as exocrine pancreatic insufficiency (EPI), are highly variable and may be modified by the response to oxidative stress. The pancreas expresses multiple antioxidant enzymes that utilize selenium at the catalytic site. Selenium deficiency in animals results in reduced expression of selenium-containing enzymes and significant pancreatic damage and/ or dysfunction. Our aim is to determine if low selenium levels in patients with CP correlates with EPI. Methods: We utilized clinical data and biological samples from North American Pancreatitis Study II (NAPS2). Serum samples were available on 278 CP cases and 263 controls. Serum trypsinogen levels were measured in a commercial laboratory by radioimmu- noassay as a marker of pancreatic mass. Levels < 10 ng/ml were used to classify CP patients with EPI. Serum selenium concentrations were determined using inductively coupled plasma mass spectrometry (ICP-MS). C-reactive protein (CRP) was measured by the Luminex plat- form, and vitamin levels were measured by high-performance liquid chromatography. Groups were compared using the Wilcoxon rank sum test. Result: Serum selenium levels were significantly lower in CP patients than in controls (127.49 ± 32.52 μg/L vs. 148.31 ± 27.15 μg/L, p<0.001). Subset analysis of CP patients indicated that CP patients had serum trypsinogen <10 ng/ml (>2 SD below the mean of normal) as a marker of EPI. CP patients with serum trypsinogen <10 ng/ml had significantly lower selenium levels than other CP patients (121.85 ± 32.25 μg/L vs. 129.90 ± 32.42 μg/L, p=0.014). CRP levels were measured to determine the effects of active inflammation on selenium levels. High CRP (>0.7mg/dl) levels correlated with low serum selenium level in CP patients (119.05 ± 33.4 μg/L vs. 131.28 ± 31.48 μg/L, p=0.004) but not controls (p>.05) suggesting that low selenium was AGA Abstracts