Duodenal secretion and fecal excretion of pancreatic elastase-1 in healthy humans and patients with chronic pancreatitis

Fecal elastase-1 is a candidate for a sensitive noninvasive test detecting chronic pancreatitis. This prospective study enrolled 10 healthy male controls and 23 patients referred for tube testing of pancreatic function. It was designed (a) to correlate duodenal outputs and fecal concentrations of elastase-1 with duodenal outputs of amylase, lipase, trypsin, and chymotrypsin in the fed state (duodenal perfusion of a mixed liquid meal at 2.5 kcal/min for 150 min), (b) to compare the diagnostic accuracy of fecal elastase-1 and fecal chymotrypsin, and (c) to characterize the cyclical pattern of postprandial pancreatic secretion in healthy subjects and patients with chronic pancreatitis. Based on their enzyme responses to duodenal meal perfusion and imaging procedures, 12 patients were classified as having normal pancreatic function and 11 patients as having chronic pancreatitis. Duodenal enzyme outputs of elastase-1 were markedly lowered in chronic pancreatitis (p < 0.0001) and correlated well with the outputs of the other four enzymes (r > 0.71, p < 0.00001). Fecal concentrations of elastase-1 were also clearly reduced in chronic pancreatitis (p < 0.0001). Fecal chymotrypsin was less strongly associated with duodenal enzyme outputs (r = 0.33 to r = 0.587), whereas fecal elastase-1 correlated more precisely with the duodenal outputs of all five enzymes (r = 0.637 to r = 0.830, p < 0.00001). Sensitivity and specificity in the detection of chronic pancreatitis amounted to 0.64 and 0.95 for fecal elastase-1 and 0.27 and 0.95 for fecal chymotrypsin, respectively. In the postprandial state, peaks of enzyme secretion occurred at a frequency of about 1 peak/150 min. The amplitude but not the frequency of secretory peaks was markedly reduced in chronic pancreatitis (p < 0.01). We conclude that fecal elastase-1 clearly exceeds the sensitivity of fecal chymotrypsin in the diagnosis of chronic pancreatitis but does not reliably detect all cases with mild to moderate disease. The pattern of postprandial pancreatic secretion is cyclical, even with minimal secretory outputs in chronic pancreatitis.