Randomized Controlled Trial to Evaluate Blood Glucose Control by the Model Predictive Control Algorithm With Variable Sampling Rate (eMPC) vs. Routine Glucose Management Protocol in Peri- and Postoperative Period in Cardiac Surgery Patients

Context. Increased blood glucose levels frequently occur in critically ill patients and its normalization by intensive insulin treatment markedly improves clinical outcome.

Objective and Design: Randomized controlled trial to compare blood glucose control by the model predictive control algorithm with variable sampling rate (eMPC) with routine glucose management protocol (RMP) in peri- and postoperative period in cardiac surgery patients.

Setting: Department of Cardiac Surgery, University Hospital. Patients. 60 cardiac surgery patients. Interventions. Elective cardiac surgery and treatment with continuous insulin infusion to maintain euglycemia (target range 4.4 - 6.1 mmol/l). 30 patients were randomized for eMPC and 30 for RMP treatment. Blood glucose was measured in 1-4 hour intervals depending on request of each algorithm during surgery and post-operation period for 24 hours.

Main Outcome Measures. Mean blood glucose, percentage of time in target range. Results. Mean blood glucose was 6.15 ± 1.11 mmol/l in eMPC vs. 7.21 ± 1.08 mmol/l in RMP group (p<0.05); percentage of time in target range was 60.4 ± 22.8% for eMPC vs. 27.5 ± 16.2% for RMP group (p<0.05). No severe hypoglycemia (blood glucose bellow 2.9 mmol/l) was observed during the study. Average insulin infusion rate was 4.67 ± 3.34 in eMPC vs. 2.57 ± 1.66 IU/h in RMP (p<0.05), average sampling interval was 1.46 ± 0.31 vs. 2.10 ± 0.22 hours (p<0.05).

Conclusions. eMPC algorithm was more effective and comparably safe as compared to RMP in maintaining euglycemia in cardiac surgery patients.