To assess exocrine pancreatic function in patients before and after pancreatoduodenectomy (PD), we used the breath test, with nonradioactive 13C-labeled trioctanoin, in 14 patients before pancreatic resection because of localized pancreatic mass (preop-PD group), and in 13 patients who had undergone pancreatoduodenectomy more than 5 yr before (post-PD group). The results were compared with those of the secretin test, N-benzolyl-L-tyrosyl-p-amino benzoic acid (BT-PABA) test, and fecal chymotrypsin. Means +/- SD and frequencies of low values of the recovery of the breath test were 42.0 +/- 3.4%, 0/5 in the control; 24.2 +/- 10.5%, 14/14 in the preop-PD group; and 18.6 +/- 8.0, 13/13 in the post-PD group. The overall sensitivities in the preop- and post-PD groups were 100% for the recovery and 93% for the maximal mass ratio of the breath test, 89% for the secretin test, 67% for the BT-PABA test, and 64% for fecal chymotrypsin. The recovery of the breath test correlated significantly with the duodenal outputs of lipase, amylase, and chymotrypsin, and was not affected in patients with obstructive jaundice or with low D-xylose absorption. The breath test is as sensitive as the secretin test, more reliable than the conventional tubeless tests, and is available to follow up the exocrine pancreatic function before and after pancreatoduodenectomy.