LAUSR

Introduction/Hypothesis: The approach and glycemic target range for adequate glycemic control in the critically ill patient is a common challenge and the optimal insulin infusion protocol is still uncertain. Methods: We conducted a pilot study after implementing a multidisciplinary insulin infusion protocol for a glycemic target of 140-180mg/dL. We did a cohort retrospective study of ICU patients that required insulin infusion before (Jun/2016Jun/2017) and after (Jul/2017-Jul/2018) the implementation of the institutional protocol. Demographics, APACHE II & SOFA, total insulin dose (TID), hours of infusion (TINF), time to achieve glycemic range (Trange), hypoglycemia (HypoG), hyperglycemia values (HyperG), episodes of hyperglycemia (NHyperG), percentage of infusion-time inside range (%range), ICU stay (ICUS), vasopressor days (TVP), mechanical ventilation days (TMV) and mortality were obtained from medical charts. A sub-group analysis with SPSS v25 was done dividing the group they belonged to (protocol (P) vs no-protocol (NP)). Results: Baseline characteristics, APACHE II & SOFA scores of both groups were similar. TID (245,4 vs 136,6), TINF (65 vs 50,3), HypoG (6,7% vs 4,6%), NhyperG (6,7 vs 3,3), ICUs (12,2 vs 9,22), TVP (6,16 vs 4,79), TMV (14,1 vs 7,2) and mortality (40% vs 27,7%) were greater in the NP group compared to the P group. While Trange (5,8 vs 4,9) and HyperG (276,7 vs 259,9) were greater in P compared to NP. The means of said groups were compared by a two-tailed t-test after Levene’s test for equality of variances was done. T-test of student or Welch t-Test was chosen accordingly. No statistically significant differences were found. A bivariate correlation test found the following associations with a p=<0,01: TID with ICUs (0,31), NHyperG with ICUs (0,27) and TID with TVP (0,49); while %range with ICUs (0,22) and TID with TVM (0,28) had a p=<0,05. Conclusions: While not having statistical significance, possibly due to small sample size (n=95), some variables show a trend towards difference between the P and NP groups. Also, there is a statistically significant relationship between said variables and relevant ICU outcomes (ICUs, TVP and TMV). We encourage implementing of institutional insulin infusion protocols for glycemic control between 140-180mg/dL as part of a quality improvement program in the ICU.