Effective intraductal delivery of the enzyme collagenase into the pancreas is crucial to the subsequent ability to isolate viable islets. Most clinical islet transplant centers load the enzyme into the pancreas by retrograde injection using a syringe following cannulation of the pancreatic duct. An alternative approach is to perfuse the pancreas via the pancreatic duct with collagenase solution using a recirculating perfusion device system. This provides control over perfusion pressures and collagenase temperature. This study reports on our evaluation of the delivery of Liberase-HI into the pancreas of 14 consecutive adult multiorgan cadaveric donors. Alternate glands were procured and processed using an identical protocol with the exception of collagenase delivery. The first group of pancreases was loaded using the perfusion technique where cold (4 degrees C) Liberase-HI was perfused at 80 mmHg for 5 min after which the pressure was increased to 180 mmHg. The collagenase solution was then slowly warmed to 35 degrees C, transferred to the dissociation chamber and mechanically dissociated, and then purified using discontinuous gradients of Ficoll. Pancreases in the second group were loaded with collagenase (28-32 degrees C) using the syringe technique before mechanical dissociation and purification. There were no significant differences in pancreas cold ischemia, donor age, body mass index, maximum blood glucose, or serum amylase of the donors between the two groups. Mean collagenase digestion time in the digestion chamber was not different between the two groups; however, the amount of undigested tissue remaining after dissociation was significantly higher in the syringe-loaded group (15.3 +/- 2.6 g vs. 4.6 +/- 2.1 g, mean +/- SEM, p < 0.05). Postdigestion recovery of islets was 471 +/- 83 x 10(3) IE in the perfusion group compared with 391 +/- 57 x 10(3) IE for the syringe-loaded group. Postpurification recovery was higher in the perfused group (379 +/- 45 vs. 251 +/- 28 x 10(3) IE, p < 0.05, two-tailed paired t-test). No difference in in vitro islet viability was observed between the two groups following glucose perifusion with the calculated stimulation index of 4.6 +/- 0.6 for the perfusion group and 4.2 +/- 0.7 for the syringe-loaded group. Controlled perfusion via the pancreatic duct allows the effective delivery of the enzyme achieving maximal distension to all regions of the pancreas leading to an increased recovery of the islets with no detrimental effect on subsequent in vitro islet function.