Automatisation du contrôle strict de la glycémie en réanimation

Since the publication of the first Leuven study highlighting the association between tight glycemic control and reduced mortality in surgical intensive care patients, many protocols have been developed and used in clinical practice. Recent multicentrer studies that did not find the positive results of the landmark study lead us to question the methods used to achieve tight glycemic control. New computerized protocols must take in account the nurse workload, the risk for severe hypoglycaemia and the unnecessary variation in clinical care. There are two opposite approaches: the “closed-loop” approach coming from research in diabetes but necessitating reliable near-continuous glucose sensors and involving legal and regulatory concerns and the “human in the loop” approach easier to implement in clinical practice. Whatever the approach, a standardized multifactorial evaluation in order to select the most efficient computer protocols is required.