This presentation is part of: Poster Session - Clinical Strategies; Judgment and Decison Making
VALIDATION AND FEASIBILITY OF THE "VAN DEN BERGHE" INSULIN PROTOCOL TO MAINTAIN A TIGHT BLOOD GLUCOSE CONTROL IN THE CRITICALLY ILL
Sandra GR Oeyen, MD, Jan Poelaert, PhD, Kirsten Colpaert, MD, and Johan Decruyenaere, PhD. University Hospital Ghent, SICU, Ghent, Belgium
Purpose: The Van den Berghe group developed an insulin protocol (IP) to target glycemia (G) between 81-110 mg/dL which resulted in a mortality reduction of 50% (N Engl J Med 2001 345:1359-67). We evaluated the compliance with the IP in the ICU of an university hospital and thereby the efficacy and feasibility of this protocol. Methods: Evaluation of the IP was started after lessons and a 1 month initiation period. All patients with an arterial line and an expected ICU-stay (ILOS) > 48 h were included. Nurses measured G on undiluted arterial blood samples using a bedside glucometer (GlucoTouch, LifeScan, Benelux). The according rate of the continuous insulin infusion (Humulin Regular, Lilly) and the frequency of G measurements were exclusively managed by ICU-nurses. A correct compliance with the IP was seen as a right insulin dose adjustment according to the protocol and the condition of the patient. Efficacy was established by comparing the number of G values within the correct range to the total number of G measurements. We defined 2 groups adjusted to the median percentage of correct G range: (1) G values ≤ 39% and (2) > 39% in the normal range. Data are presented as mean ± SD or as percentage. Statistical analysis was performed using an independent samples T-test. Significance was accepted when p<.05. Results: Out of 6016 G measurements, (30 patients - APACHE II 26±8) a good compliance with the IP in 70.9% (N=4267) was found. Only 41.6% (N=2504) of all G values were regulated in the desired range despite hourly or two-hourly G measurements (17 ± 7 per day). Group 1 showed higher G values G>250 mg/dL 1.5±1.5 vs 0.7±0.9; p<.006. In group 2 more G measurements were <40 mg/dL 0.1±0.4 vs 0.5±0.9; p<.01. The 2 groups were comparable concerning APACHE II score, number of diabetic patients, ILOS, compliance with the IP and number of daily G measurements. Conclusion: The compliance with the paper-based "Van den Berghe" IP was acceptable but can be improved. A good compliance does not imply a high percentage of G values within the target range. Further study is needed to see if computer-based and more complex guidelines, implemented bedside in an Intensive Care Information System, could lead to a better blood glucose control.
