Extreme maldistribution and immediate establishment of severe cellular injury are typical features of traditional bile salt models of acute pancreatitis; both factors complicate assessment and interpretation of therapeutic benefits in trials of experimental therapy. Even more important, both are indications not of the desired induction of pancreatitis but rather of local injury by barotrauma and noxious chemicals. This study contrasts the severity and regional variability of cellular injury in traditional high-dose bile salt models with that seen in a new preparation employing the combination of intravenous (iv) caerulein (CAE) and intraductal (id) low-dose glycodeoxycholic acid (GDOC). Thirty-six male Sprague-Dawley rats (350-450 g) were induced with (group A) high-dose GDOC id (34 mmol/L), low-dose GDOC id (10 mmol/L) (group B), or low-dose GDOC id combined with caerulein iv for 6 h (group C). The regional distribution of histopathologic injury within the pancreas was assessed in 20 fields/organ by two pathologists unaware of the induction technique used. High-dose GDOC id (group A) resulted in extremely heterogenous distribution of injury for all variables (edema, p = 0.001; acinar necrosis, p = 0.0001; inflammation, p = 0.0001; and hemorrhage p = 0.001). The lesions were confined to the head of the pancreas, which showed large areas of necrosis involving entire lobules, whereas adjacent areas were unaffected. Low-dose GDOC id (group B) was more homogenously distributed, but the injury was mild and regional variability (edema, p = 0.0001; acinar necrosis, p less than 0.04; inflammation, p = 0.0001; and hemorrhage p less than 0.05) was still demonstrable. In contrast, low-dose GDOC id combined with CAE iv (group C) produced moderately severe pancreatitis, which equally affected all areas of the gland. There were no geographical differences in acinar necrosis or inflammation. This feature of the new model provides a desirable prerequisite for accurate and reproducible assessment of histopathology in studies aimed at detecting effects of therapy. We suggest that it replace traditional id bile salt infusion models.