Objective: The introduction of strict glycemic control in the intensive care unit has increased the risk for hypoglycemia. In this study we examined the association between predefined circumstances and the occurrence of hypoglycemia in the intensive care unit.
Design: : Retrospective cohort study.
Setting: Academic medical center.
Patients: All episodes of hypoglycemia (glucose value <45 mg/dL) in our intensive care unit between September 2002 and September 2004 were identified. Presence of predefined circumstances previously associated with hypoglycemia was scored around the moment of hypoglycemia using a patient data management system and in-hospital charts. Patients with a first hypoglycemic event were contrasted to controls from the same cohort, who were matched for time since admission, to correct for the effect of length of stay. Data were analyzed using conditional logistic regression analysis.
Interventions: None.
Measurements and main results: Of 2,272 patients, 156 (6.9%) experienced at least one episode of hypoglycemia. Continuous venovenous hemofiltration with bicarbonate-based substitution fluid (odds ratio [OR], 14; 95% confidence interval [CI], 1.8-106), a decrease of nutrition without adjustment for insulin infusion (OR, 6.6; 95% CI, 1.9-23), diabetes mellitus (OR, 2.6; 95% CI, 1.5-4.7), insulin use (OR, 5.3; 95% CI, 2.8-11), sepsis (OR, 2.2; 95% CI, 1.2-4.1), and inotropic support (OR, 1.8; 95% CI, 1.1-2.9) were associated with hypoglycemia. Simultaneous octreotide and insulin use (OR, 6.0; 95% CI, 0.72-50) may also be associated with hypoglycemia. Gastric residual during enteral nutrition without adjusting insulin infusion, liver failure, continuous venovenous hemofiltration with lactate-based substitution fluid, diminished glomerular filtration rate, dose diminishment of glucocorticoids or catecholamines, and use of beta-blocking agents were not associated with hypoglycemia. Adjusting for age, gender, and Acute Physiology and Chronic Health Evaluation II score at admission did not materially change ORs.
Conclusion: Use of bicarbonate-based substitution fluid during continuous venovenous hemofiltration, a decrease of nutrition without adjustment for insulin infusion, a prior diagnosis of diabetes mellitus, sepsis, and need for inotropic support were found to be associated with hypoglycemia. Simultaneous use of insulin and octreotide may be associated with hypoglycemia.