Rapid endoscopic secretin stimulation test and discrimination of chronic pancreatitis and pancreatic cancer from disease controls

Abstract

Background & aims: The cholecystokinin (CCK)/secretin pancreatic function tests to diagnose pancreatic exocrine insufficiency are time consuming and invasive. Our aim was to develop a rapid pancreatic function test performed during upper endoscopy that could discriminate between patients with normal from impaired exocrine pancreatic secretion.

Methods: We prospectively evaluated 412 patients for possible pancreatic diseases. During upper endoscopy, 1 CU/kg of secretin was given intravenously and duodenal juice (collected for 10 min) was assayed for concentrations of bicarbonate and lipolytic and trypsin activity. Final diagnosis was by histology, imaging, and a previously validated scoring system (for chronic pancreatitis). Of 412 patients, 117 patients had normal pancreas, 72 patients had chronic pancreatitis, and 116 patients had pancreatic adenocarcinoma. The remaining 107 patients had miscellaneous disease of the peripancreatic region. In 28 patients we also validated the secretin test with the standard CCK pancreatic function test.

Results: There was no difference between bicarbonate or trypsin concentrations among the groups. Lipolytic concentration was significantly lower in chronic pancreatitis (115 +/- 18) and in pancreatic adenocarcinoma (87 +/- 10) compared with patients with normal pancreas (229 +/- 23; P < 0.03 and P < 0.0001, respectively). The overall accuracy of the endoscopic secretin test was 79%, with positive and negative predictive values of 73% and 85%, respectively. The concentration of lipolytic activity obtained by the endoscopic secretin test in 28 patients correlated moderately well (r = 0.41, P < 0.03) with lipolytic output obtained by the CCK pancreatic function test.

Conclusions: Lipolytic concentration in duodenal juice after intravenous secretin collected for 10 minutes during upper endoscopy was significantly lower in chronic pancreatitis and pancreatic adenocarcinoma compared with patients with normal pancreas, but was not accurate enough for routine clinical use.