Clinical islet allotransplantation is dependent on the ability to achieve a high yield and purity of islets isolated from human cadaver pancreas donors. The aim of this study was to determine the factors influencing the pancreas prior to islet isolation that may alter yield and purity. The results of 50 consecutive islet isolations from cadaver donor human pancreati at the University of Chicago Medical Center from December 1991 to April 1993 were analyzed. All pancreati were first offered for whole pancreas transplantation before being considered for islet isolation. Human pancreatic islet isolation was accomplished by a modified automated method. Some islet isolations resulted in a high islet yield but low islet purity. Other resulted in well-purified islets, but a low yield. Arbitrarily, successful islet isolation is defined as that yielding over 250,000 islet equivalents (EQN) with a purity of at least 80%. The success rate of human pancreatic islet isolation was 70%. The mean final islet yield obtained from these 50 pancreati was 300,000 +/- 131,000 islet EQN. The mean purity of the final preparation was 73% +/- 25%. By univariate analysis, five factors were found to affect significantly the yield, purity, or overall success rate of islet isolation: organ cold ischemic time, donor age, donor plasma glucose levels, donor body weight, and cause of donor death. Even when islet isolation was successful, the function of islets from hyperglycemic and older donors appear to be impaired both in vitro and in vivo. These results suggest that islet yield and purity are affected by multiple donor-related factors. Even when adequate yield and purity are obtained, islet function is also dependent on donor variables.