Background: Insulin pharmacokinetics models describe the distribution and elimination of insulin in the body. These models can be useful in designing infusion schemes to produce a desired plasma insulin profile.
Methods: We evaluated the pharmacokinetics of intravenous insulin delivery in five human subjects with type 1 diabetes; one subject was studied on three separate occasions for a total of seven experiments. Each subject consumed an identical 829 Kcal meal of solid food for breakfast and lunch, followed by exercise on a stationary bicycle. Regular human insulin was infused into a peripheral vein at a basal rate (0.5 U/h) for the entire study. Coinciding with the consumption of each meal, insulin was infused at a higher rate for 120 min (average dose 8.5 +/- 2.1 U). The bolus dose was based upon preprandial and 60-min postprandial blood glucose measurements and a sliding scale regimen. Blood glucose levels were allowed to fluctuate throughout the protocol (mean 174 mg/dL, range 55-340 mg/dL). Human insulin levels were measured from plasma samples obtained every 10 min throughout the 8.5-h protocol. A total of 346 plasma samples were assayed to measure the concentration of human insulin.
Results and conclusions: Data were used to compare (previously developed) insulin kinetics models consisting of one, two, and three first-order linear differential equations. The ability of each model to simulate the input/output data (intravenous insulin infusion rate/plasma insulin concentration) was assessed. Our main finding is that the first-order linear insulin kinetics model was sufficient to describe the clinical data. The model had a fractional insulin loss rate of 0.112 +/- 0.063 min(1) and a distribution volume of 15.6 +/- 4.0 L.