OBJECTIVE There is controversy about the cholecystokinin (CCK)-induced decrease of plasma amino acids for the diagnosis of pancreatic insufficiency in that it is unclear whether the conflicting results are due to different degrees of stimulation by CCK. We have therefore evaluated the amino acid consumption test by examining the dose-response relation among increasing doses of cerulein and plasma amino acids compared with the release of pancreatic polypeptide in six healthy volunteers and six patients with severe pancreatic insufficiency proven by a pathological para-aminobenzoic acid test. METHODS Stepwise increasing doses of cerulein (10-80 pmol/kg/h) were given i.v., each for 60 min with a secretin background (1 CU/kg/h). In the volunteer group, an additional experiment with infusion of placebo (saline 0.9%) was performed. RESULTS CCK-induced amino acid changes were small in volunteers (maximum decrease: 8.4 +/- 0.9%; mean +/- SEM), tended to be more pronounced in patients (maximum decrease: 13.8 +/- 2.8%; NS), and did not permit a distinction among volunteers and patients. There was no dose-response relationship between CCK and plasma amino acids in either group. In contrast, pancreatic polypeptide levels increased markedly and dose-dependently in volunteers and patients and tended to be lower in patients with pancreatic insufficiency. CONCLUSIONS The decrease of plasma amino acid levels in response to cerulein does not reflect pancreatic function and does not permit the diagnosis of pancreatic insufficiency.