A Comparison Study of Continuous Insulin Infusion Protocols in the Medical Intensive Care Unit: Computer-Guided Vs. Standard Column-Based Algorithms

• Authors: Christopher A. Newton, Emory University; Dawn Smiley, Emory University; Bruce W. Bode, Piedmont Hospital; Abbas E. Kitabchi, University of Tennessee; Paul C. Davidson, Piedmont Hospital; Sol Jacobs, Emory University; R. Dennis Steed, Southeastern Endocrine & Diabetes; Frankie Stentz, University of Tennessee; Limin Peng, Emory University; Patrick Mulligan, Emory University; Amado X. Freire, University of Tennessee; Angel Temponi, Emory University; Guillermo Umpierrez, Emory University
• Journal Title: Journal of Hospital Medicine
• Volume: Volume 5, Number 8
• Publisher: Wiley: 12 months | 2010-10, Pages 432-437
• Type of Work: Article | Post-print: After Peer Review
• Publisher DOI: http://doi.org/10.1002/jhm.816
• Subjects: Health Sciences, Medicine and Surgery
• Keywords: diabetes; hospital; hyperglycemia; hypoglycemia; insulin infusion; intensive care unit

Abstract

PURPOSE To compare the safety and efficacy of continuous insulin infusion (CII) via a computer-guided and a standard paper form protocol in a medical intensive care unit (ICU). METHODS Multicenter randomized trial of 153 ICU patients randomized to CII using the Glucommander (n = 77) or a standard paper protocol (n = 76). Both protocols used glulisine insulin and targeted blood glucose (BG) between 80 mg/dL and 120 mg/dL. RESULTS The Glucommander resulted in a lower mean BG value (103 ± 8.8 mg/dL vs. 117 ± 16.5 mg/dL, P < 0.001) and in a shorter time to reach BG target (4.8 ± 2.8 vs.7.8 hours ± 9.1 hours, P < 0.01), and once at target resulted in a higher percentage of BG readings within target (71.0 ± 17.0% vs. 51.3 ± 19.7%, P < 0.001) than the standard protocol. Mean insulin infusion rate in the Glucommander was similar to the standard protocol (P = 0.12). The percentages of patients with ≥1 episode of BG <40 mg/dL and <60 mg/dL were 3.9% and 42.9% in the Glucommander and 5.6% and 31.9% in the standard, respectively [P = not significant (NS)]. Repeated measures analyses show that the probabilities of BG reading <40 mg/dL or <60 mg/dL were not significantly different between groups (P = 0.969, P = 0.084) after accounting for within-patient correlations with or without adjusting for time effect. There were no differences between groups in the length of hospital stay (P = 0.704), ICU stay (P = 0.145), or inhospital mortality (P = 0.561). CONCLUSION Both treatment algorithms resulted in significant improvement in glycemic control in critically ill patients in the medical ICU. The computer-based algorithm resulted in tighter glycemic control without an increased risk of hypoglycemic events compared to the standard paper protocol.