Blood glucose control by a model predictive control algorithm with variable sampling rate versus a routine glucose...

Area of Science:

• Critical care medicine
• Endocrinology
• Medical device technology

Background:

• Critically ill patients frequently experience elevated blood glucose levels.
• Intensive insulin therapy for tight glucose control significantly improves outcomes in critically ill patients.

Purpose of the Study:

• To compare the efficacy and safety of a laptop-based enhanced model predictive control (eMPC) algorithm with a variable sampling rate against a routine glucose management protocol (RMP) for peri- and postoperative blood glucose control.
• To evaluate glycemic control in elective cardiac surgery patients using eMPC versus RMP.

Main Methods:

• A randomized controlled trial involving 60 elective cardiac surgery patients.
• Patients were randomized to either eMPC (n=30) or RMP (n=30) treatment groups.
• Continuous insulin infusion was used, with RMP also incorporating IV insulin boluses, to maintain a target blood glucose range of 4.4-6.1 mmol/liter over 24 hours postoperatively.

Main Results:

• The eMPC group achieved a significantly lower mean blood glucose (6.2 ± 1.1 mmol/liter) compared to the RMP group (7.2 ± 1.1 mmol/liter; P < 0.05).
• Patients in the eMPC group spent a significantly greater percentage of time within the target glucose range (60.4 ± 22.8%) versus the RMP group (27.5 ± 16.2%; P < 0.05).
• No severe hypoglycemia (< 2.9 mmol/liter) was observed in either group; however, the eMPC group required a higher mean insulin infusion rate (4.7 ± 3.3 IU/h vs. 2.6 ± 1.7 IU/h; P < 0.05).

Conclusions:

• The eMPC algorithm demonstrated superior effectiveness in maintaining euglycemia for cardiac surgery patients compared to the RMP.
• The eMPC algorithm provided comparable safety to RMP in managing blood glucose levels during the peri- and postoperative periods.