Complications of insulin infusion pump therapy include ketoacidosis related to interruption of insulin delivery and infected infusion sites. To determine if the type of insulin used in insulin pumps is a factor influencing these complications, we compared the occurrence of inflammation and infection at infusion sites and obstruction of infusion tubing in a 6-mo crossover study of 28 patients using two different insulin preparations. The use of buffered purified pork insulin was associated with less infusion site inflammation (P = 0.0039) and infusion set obstruction (P = 0.0215) than use of unbuffered beef-pork insulin. Infected infusion sites occurred less often with buffered pure pork insulin than with unbuffered beef-pork insulin, but the difference was not statistically significant. We conclude that the type of insulin selected for use in insulin pumps influences the occurrence of adverse reactions to insulin pump therapy: buffered purified pork insulin is associated with less infusion site inflammation and infusion set obstruction than unbuffered beef-pork insulin.