This review considers subcutaneous insulin pharmacokinetics, with emphasis on those aspects relevant to subcutaneous insulin infusion devices. These devices are in increasing use for diabetes therapy, although reliable data on subcutaneous insulin absorption is required to optimise their programming. Techniques for obtaining and interpreting pharmacokinetic data are considered. Recent studies would suggest that it is possible to simulate the major physiological fluctuations in blood insulin levels via continuous subcutaneous insulin infusion, combined with appropriate bolus insulin delivery. Most insulin infused subcutaneously will reach the systemic circulation, and in the majority of diabetics, subcutaneous insulin degradation is low. However, the absorption rate is slow and it may take 6 to 8 hours to reach a steady-state following a change in the subcutaneous infusion rate. Thus there is little to be gained from hour by hour adjustments to the basal insulin infusion rate.Basal rate supplementation, and meal insulin requirements, are best met by bolus delivery. It is particularly important to provide increased systemic delivery of insulin with the start of a meal. The most appropriate adjustment for additional exercise may be a small reduction in the preceding meal bolus, rather than a reduction in the basal rate. Studies of the pharmacokinetics of insulin are an important contributing factor to the optimisation of subcutaneous insulin infusion therapy.