Comparison of endoscopic ultrasound chronic pancreatitis criteria to the endoscopic secretin-stimulated pancreatic function test.

Endosonography (EUS) has emerged as a major diagnostic tool in pancreatic imaging. Direct tests of pancreatic function are considered the most sensitive and accurate method to establish a diagnosis of chronic pancreatitis (CP), particularly when imaging studies are inconclusive. The aim of this study was to compare current EUS CP criteria with our newly described, purely endoscopic, secretin-stimulated pancreatic function test (ePFT). Fifty-six patients (25 male, mean age = 44 years) who were referred for evaluation/treatment of chronic abdominal pain with or without CP underwent both EUS and ePFT. The EUS protocol included the following: (1) EUS images were obtained in a standardized fashion from both gastric and duodenal stations, and (2) EUS images were scored independently by one of three therapeutic endoscopists for 0--9 parenchymal/ductal criteria as follows: 0-3 = normal, 4-5 = equivocal, >/=6 = definite CP. Endoscopic pancreatic function test (ePFT) protocol included the following: (1) upper endoscopy, (2) intravenous synthetic porcine secretin (0.2 mcg/kg, ChiRhoClin, Inc.) after test dose, (3) duodenal fluid aspirated every 15 min for 1 h, and (4) autoanalyzed for [HCO3] cutpoint of 80 mEq/L. According to EUS, 33 were normal, 13 equivocal, and 10 definite for CP. The mean peak [HCO3 -] range (in mEq/L) for each group was normal CP (83.7, range = 58-118), equivocal CP (68, range = 30-88), and definite CP (56, range=19-84). Using a peak [HCO3 -] of </=80 mEq/L as diagnostic for CP, the referent values (sensitivity%/specificity%) for EUS in the diagnosis of CP were normal (60/72), equivocal (36/94), and definite (26/100), respectively. An EUS score or greater than 5 had the best specificity (100%) and negative predictive value (100%). We conclude that endoscopic pancreatic function testing with secretin confirms that as EUS score increases, the peak pancreatic fluid bicarbonate decreases. We also conclude that EUS has excellent statistical inferences for diagnosing CP when at least 6 or more criteria are present. EUS as a screening test in patients with chronic abdominal pain and equivocal imaging studies may be of limited value.