PURPOSE: We have reported that highest, lowest, and more variable blood glucose values predict unfavorable cognitive sequelae in ARDS survivors managed with a paper protocol (target 120-180 mg/dL). We now retrospectively compare blood glucose values and variability in medical/surgical ICU patients managed with three blood glucose protocols.
METHODS: We included critically ill patients treated using glucose protocols in the LDS Hospital STRICU during three time periods. Group1: Paper protocol from January–December 2002 (target 120-180 mg/dL). Group2: two electronic protocols from January-March 5, 2004 (targets 120-180 mg/dL and 80-110 mg/dL). Group3: electronic protocol (eProtocol-insulin) from March 6, 2004–December 2005, target 80-110 mg/dL. We report means ± SDs.
RESULTS: Group1: N=306, age 53±18 years, initial APACHE II score 22.1±8.1, and ICU LOS 7.4±7.6 days. Group2: N=512, age 52±19 years, initial APACHE II score 21.6±7.9, and ICU LOS 6.4±6.9 days. Group3: N=992, age 52±18 years, initial APACHE II score 19.7±8.2, and ICU LOS 6.0±7.4 days. Group1 blood glucose was 148±53 mg/dL (median 141 mg/dL), Group2 136±47 mg/dL (median 128 mg/dL), and Group3 115±37 mg/dL (median 109 mg/dL). Blood glucose values <40 mg/dL occurred in 0.23% of measurements in Group1 and 0.12% in Groups 2 and 3. Blood glucose values <60 mg/dL were not different between the groups and occurred in 0.89% (81/9091) of measurements in Group1, 0.92% (164/17,862) in Group2, and 1.1% (510/46,067) in Group3. However, the percent of patients with blood glucose values <60% mg/dL significantly increased from 12.8% in Group1 to 15.4% in Group2 and 26.1% in Group3 (p<0.001). General estimating equations comparing the mean blood glucose levels found significant differences for the 3 groups (p<0.0001).
CONCLUSION: A management strategy that targets glucose levels of 80-110 mg/dL resulted in a significant decrease in blood glucose variability but increased the percentage of patients with blood glucose values <60 mg/dL.
CLINICAL IMPLICATIONS: Glucose management using a target of 80-110 mg/dL reduces blood glucose variability in critically ill patients.
DISCLOSURE: Ramona Hopkins, None.