Introduction of intensive glycaemic control into a neurosurgical intensive care unit: a retrospective cohort study.

Background: There is evidence that intensive glycaemic control decreases morbidity and mortality in surgical intensive care unit patients. In traumatic brain injury, hyperglycaemia is a prognostic indicator.

Objective And Study Design: This retrospective cohort study describes the blood glucose level outcomes before and after the introduction of an intensive insulin protocol to a neurosurgical ICU.

Methods: We analysed data on all patients admitted to a neurosurgical ICU during the year before, and the year after, the introduction in July 2003 of an intensive insulin protocol targeting patient blood glucose level (BGL) to 4.4- 6.1mmol/L. Patients who underwent fewer than 20 BGL measurements during the ICU admission were excluded. Data were obtained from an electronic clinical information system.

Results: 121 patients were enrolled: 64 before and 57 after introduction of the protocol. The groups were similar in sex, age, and severity of illness on ICU admission. After introduction of the protocol, BGL was significantly lower (weighted mean BGL, 6.73mmol/L v 7.77mmol/L before; P < 0.001), and proportion of BGL readings in the target range was greater (37% v 21% before; P < 0.001). In addition, BGL variability was greater (weighted mean standard deviation, 2.03 v 1.88 before; P < 0.001), hypoglycaemia was more common (0.58% of BGL readings v 0.20% before; P = 0.06), and ICU stay was significantly longer, but there was no difference in median Glasgow Coma Scale score or ICU discharge status (alive or dead).

Conclusions: The intensive insulin protocol was effective in lowering blood glucose in neurosurgical ICU patients.