Our Scientific Journey through the Ups and Downs of Blood Glucose Control in the ICU

This article tells the story of our long search for the answer to one question: is stress hyperglycemia in critically illness adaptive or mal-adaptive? Our earlier work had suggested lack of hepatic insulin effect and hyperglycemia jointly predicting poor outcome. We therefore hypothesized that insulin infusion to reach normoglycemia, tight-glucose-control, improves outcome. In 3 RCTs, we found morbidity and mortality benefit with tight-glucose-control. Moving from the bed to the bench, we attributed benefits to prevention of glucose toxicity in cells taking up glucose in an insulin-independent, glucose concentration gradient-dependent manner, counteracted rather than synergized by insulin. Several subsequent RCTs did not confirm benefit and the large 'NICE-SUGAR' trial found increased mortality with tight-glucose-control associated with severe hypoglycemia. Our subsequent clinical and mechanistic research revealed that early use of parenteral nutrition, the context of our initial RCTs, had been a confounder. Early parenteral nutrition aggravated hyperglycemia, suppressed vital cell damage removal and hampered recovery. Therefore, in our next and largest 'TGC-fast' RCT, we re-tested our hypothesis, without use of early parenteral nutrition and with a computer algorithm for tight-glucose-control that avoided severe hypoglycemia. In this trial, tight-glucose-control prevented kidney and liver damage though with much smaller effect size than in our initial RCTs without affecting mortality. Our quest ends with the strong recommendation to omit early parenteral nutrition for ICU patients, as this reduces need of blood glucose control and allows cellular housekeeping systems to play evolutionary selected roles in the recovery process. Once again, less is more in critical care.