Intensive insulin therapy in the neurocritical care setting is associated with poor clinical outcomes.

Published

Journal Article

BACKGROUND: Studies devoted to intensive glucose control suggested that the intensive insulin therapy (IIT) approach could effectively reduce complications associated with critical illness. A program of IIT with the goal of achieving a blood glucose of 80-120 mg/dL was, therefore, adopted in this study. To explore the impact of this approach in patients admitted to a neurocritical care unit, we compared the short-term outcomes of patients treated before and after our policy change. METHODS: Retrospectively extracted data from the electronic medical records of 1913 patients admitted between Feb 1, 2005 and Aug 30, 2006 were included in the standard insulin therapy group (SIT targeted to maintain blood glucose levels less than 150 mg/dL) and 1796 patients admitted between Sept 1, 2006 and March 30, 2008 were included in the IIT group (IIT-targeted to maintain blood glucose levels between 80 and 120 mg/dL). RESULTS: Mean glucose levels were lower in the IIT compared to SIT (136.9 mg/dL, SD 47.6 vs. 143.8 mg/dL, SD 46.4); however, this strategy was also associated with an increased incidence of hypoglycemia (OR: 1.8, and 95% CI: 1.5-2.3). The likelihood of mortality increased proportionally as the severity of hypoglycemia worsened (any blood glucose value <70 mg/dL, OR: 3.26, and 95% CI: 2.52-4.22, any blood glucose <40, OR: 3.65, and 95% CI: 2.21-6.02, any blood glucose <20, OR: 6.25, and 95% CI: 2.41-16.23). CONCLUSION: IIT was not only able to reduce overall mean glucose levels, but also resulted in significantly more episodes of hypoglycemia, increased mortality, and increased length of stay. The relationship between hypoglycemia and mortality indicates that efforts to control glucose levels should also aggressively avoid induction of hypoglycemia.

Full Text

Duke Authors

Cited Authors

  • Graffagnino, C; Gurram, AR; Kolls, B; Olson, DM

Published Date

  • December 2010

Published In

Volume / Issue

  • 13 / 3

Start / End Page

  • 307 - 312

PubMed ID

  • 21086066

Pubmed Central ID

  • 21086066

Electronic International Standard Serial Number (EISSN)

  • 1556-0961

Digital Object Identifier (DOI)

  • 10.1007/s12028-010-9469-4

Language

  • eng

Conference Location

  • United States