M. Horibe, M.D., J.D. Lang, M.D., B. G. Nair, Ph.D.
Hyperglycemia is associated with increased mortality and morbidity in critically ill patients even though intensive insulin therapy with existing protocols has shown inconsistent effects on mortality and increased rates of hypoglycemia. We aim to evaluate a novel glucose control technique that utilizes a computerized algorithm in post-surgical ICU. A total of 42 consented patients will be randomly assigned to receive insulin dose based on either the existing "insulin drip" or a new algorithm when the blood glucose levels exceed 140 mg/dL during their ICU stay after elective surgery. Blood glucose levels will be measured every hour and a computer program will prescribe an insulin dose based on the selected algorithm. The new scientific algorithm, called the Fading Memory algorithm, will produce an insulin dose response that is quite similar to that of normal pancreas. In addition to evaluating the feasibility of the new algorithm, its performance will be compared against the existing "insulin drip" protocol based on their abilities to maintain target glycemic levels. It is anticipated that the Fading Memory algorithm will provide better glucose control more efficiently than the existing protocol thus leading to better clinical outcome without severe hypoglycemia. It is hoped that the proposed study will pave the way for an integrated, closed-loop, automated glucose control system.