Development and testing of a simple algorithm for a glucose clamp

The glucose clamp technique is widely used in clinical research studies to maintain a constant blood glucose (BG) level during metabolic perturbation. An algorithm is used to set the rate of an intravenous glucose infusion according to measured BG concentrations. Currently used clamp algorithms are unnecessarily complex. This paper describes the development and use of a new algorithm which is simpler to implement. The algorithm employs proportional and differential feedback control. Discrete BG sampling with a sampling period of 5 min is used. Computer simulation with a two-compartment model of glucose dynamics was used to refine the algorithm. A Monte Carlo approach was adopted to examine the effects of random variations in measured BG concentrations and the elapsed time between blood sampling and adjustment of the glucose infusion rate. A coefficient of variation in BG concentration of less than 4 per cent was predicted for a euglycemic clamp. This was confirmed when the algorithm was applied in clinical research studies. This degree of BG control compares well with other published algorithms.