The metabolic consequences of hemipancreatectomy in living pancreas donors were tested in a dog model in which a 50% lobe-specific pancreatectomy was performed. Removal of the dorsal lobe (analogous to a donor, n = 4) resulted in a progressive increase in fasting glucose during 12 mo from 5.32 +/- 0.16 to 8.17 +/- 0.28 mM and a decrease in fasting insulin from 54 +/- 3 to 6.0 +/- 2.4 pM and glucose clearance (Kg) from 3.00 +/- 0.22 to 1.00 +/- 0.06 mM. Removal of the ventral lobe (analogous to a recipient, n = 5) did not result in a change in fasting glucose or Kg during 12 mo, although fasting insulin was reduced from 36.0 +/- 1.8 to 18.00 +/- 1.93 pM. In vitro perfusion of the remnants after 1 yr showed a deterioration in glucose-stimulated insulin secretion (5-11 mM) by the dorsal remnant (18 +/- 11 vs., 232 +/- 37%) and the ventral remnant (2.6 +/- 19 vs. 87 +/- 13%). The dorsal remnant had a higher response than the ventral remnant (46 +/- 23 vs. -16 +/- 10%, respectively) to severe hyperglucosuria (11-27.7 mM). Insulin content was unchanged in the dorsal remnant (224 +/- 16 vs. 180 +/- 14 micrograms/g), but was reduced in the ventral remnant (65 +/- 14 vs. 154 +/- 15 micrograms/g). In vitro insulin pulse intervals were reduced in both remnants (5.3 +/- 0.2 min vs. control 7.00 +/- 0.18 min). Because of the above effects on the donor when the dorsal lobe is removed, the continued use of hemipancreatectomy as a source of transplantable tissue must be questioned.