Intensive Insulin Therapy in the Medical ICU

Abstract & Commentary

By David J. Pierson, MD, Editor, Professor, Pulmonary and Critical Care Medicine, Harborview Medical Center, University of Washington, is Editor for Critical Care Alert.

Synopsis: The use of an insulin infusion protocol to keep blood glucose between 80 and 110 mg/dL in adult medical ICU patients had no overall effect on mortality but appeared to decrease selected complications. Survival was improved by tight glucose control among patients remaining in the ICU beyond 3 days.

Source: Van den Berghe G, et al. Intensive insulin therapy in the medical ICU. N Engl J Med. 2006;354:449-461.

In a previous study involving patients in a surgical ICU1, Van den Berghe and associates at the Catholic University of Leuven, Belgium, showed that tight control of serum glucose levels by means of a strict insulin infusion protocol decreased both morbidity and mortality. In the present study, these same investigators sought to determine whether the same results would be found in medical ICU patients, who tended to have greater severity of illness, more co-morbidities, and higher mortality rates. In a prospective, controlled clinical trial, the authors randomized medical ICU patients who were predicted to require at least 3 days of ICU care to either conventional therapy with respect to serum glucose or a tight-control protocol to keep values between 80 and 110 mg/dL. Patients randomized to receive conventional therapy were administered insulin when blood glucose exceeded 215 mg/dL, but the infusion was tapered when levels fell below 180 mg/dL.

Of 2110 adult medical ICU patients screened during a 3-year period, 1200 were randomized to conventional vs tight-control glucose management. Also, 767 of these patients remained in the ICU for at least 3 days. Among all 1200 patients in the study (intention-to-treat analysis), in-hospital mortality was 40% in the conventional-treatment group vs 37% in the intensive-treatment group, a non-significant difference (P = 0.33). However, among the patients who remained in the ICU for at least 3 days, there was a significantly lower in-hospital mortality among those who received the tight-glucose-control therapy (52% vs 43%; P = 0.009). Morbidity, in the form of newly acquired renal dysfunction, delayed weaning from ventilatory support, and prolongation of ICU stay, was less with the intensive-therapy regimen, both in the overall group and in those patients staying in the ICU for at least 3 days.


In this large clinical trial conducted in the medical ICU of a single institution, the use of an intensive insulin therapy protocol to keep blood glucose levels between 80 and 110 mg/dL had no effect on overall mortality. The fact that mortality was significantly reduced among the 64% of patients who remained at least 3 days in the ICU (the investigators’ target population) means that mortality had to be correspondingly increased among patients who spent less time in the ICU. This was apparently the case, a finding for which the authors offered no ready explanation. That the investigators were unable to predict in advance which patients would remain in the ICU at least 3 days—and thus, which patients would presumably be helped rather than harmed by intensive insulin therapy—makes potential application of the study’s findings problematic for the clinician.

The introduction, investigation, and adoption of new ICU therapies often seem to follow a distinct sequence. An initial study—usually relatively small and from a single center—is published showing markedly positive results from the new treatment. In a number of such initial studies, mortality rates in the control population—patients receiving standard care rather than the new treatment—have been unusually high. Often, aspects of patient selection, potentially confounding conditions and therapies, and other factors have raised questions about the study and about the generalizability of its findings. These cautions notwithstanding, the new therapy tends to be widely embraced by intensivists, sometimes even to the point of becoming a standard of care. Subsequently, larger, multi-center studies are published. Sometimes these more extensive investigations confirm the effects of the new therapy and validate the findings of the initial paper. This was the case with lung-protective ventilation for acute lung injury and ARDS, and also for noninvasive ventilation for acute hypercapnic respiratory failure in patients with COPD. However, in numerous other instances the larger-scale trials have failed to demonstrate the benefit found in the initial study. Therapies such as late corticosteroids and inhaled nitric oxide for ARDS, and supra-physiologic oxygen-delivery goals for severe shock come to mind here, among others.

What will the "bottom line" prove to be with respect to intensive insulin therapy when the more extensive studies called for by Van den Berghe et al are done? Such studies are reportedly underway at the present time.2 How should the clinician manage blood glucose levels in critically ill patients, based on current information? In an editorial accompanying the paper by Van den Berghe and colleagues, Malhotra3 lists 3 possible courses of action. One would be to go along with current trends and simply adopt the "tight glucose control" approach, on the rationale that it seems to benefit some patients and its adverse effects do not seem to be too severe. This approach has already become a standard of care in many ICUs around the world, and is included in the most recent guidelines for the management of sepsis.4 Another approach would be to withhold intensive insulin therapy until its evidence base is more complete and more confident conclusions about when and how to use it can be drawn. A reasonable case could be made for either of these approaches.

Malhotra suggests a third, more middle-of-the-road approach.3 During the first 3 days in the ICU, try to keep blood glucose levels below 150 mg/dL. If critical illness continues and the patient remains in the ICU beyond 3 days, consider tighter glucose control to blood levels between 80 and 110 mg/dL, as employed by Van den Berghe et al. While the effects of tight glucose control starting at ICU day 3 have not specifically been investigated, this approach seems reasonable, and would be supported by the current study as well as other currently available data.


  1. Van den Berghe G, et al. Intensive insulin therapy in critically ill patients. N Engl J Med. 2001;345:1359-1367.
  2. Angus DC, Abraham E. Intensive insulin therapy in critical illness. Am J Respir Crit Care Med. 2005;172:1358-1359.
  3. Malhotra A. Intensive insulin in intensive care. N Engl J Med. 2006;354:516-518.
  4. Dellinger RP, et al. Surviving Sepsis Campaign Management Guidelines Committee. Surviving Sepsis Campaign guidelines for management of severe sepsis and septic shock. Crit Care Med. 2004;32:858-873 Erratum in: Crit Care Med. 2004;32:1448. Correction of dosage error in text. Crit Care Med. 2004;32:2169-2170.