Stress-induced hyperglycemia is prevalent in critical care, even in patients with no history of diabetes. Control of blood glucose level with tight insulin therapy has been shown to reduce incidences of hyperglycemia leading to reduced mortality and improved clinical outcomes. STAR is a tablet-based glucose control protocol with a specialized user interface into which insulin and nutrition information can be entered and predicted. This research describes the first clinical pilot trial of STAR approach in International Islamic University Hospital, Kuantan, Pahang. The clinically specified target for blood glucose level is between 4.4 and 8.0 mmol/L. Seven episodes (359 hours) were recruited based on the need for GC. Overall, 43.93% of measurement are in the range of 4.4-8.0 mmol/L band. The BG median is 8.30 [6.32-10.00] mmol/L with only 1 patient having below than 2.22 mmol/L which is the guaranteed minimum risk level. This pilot study shows that STAR protocol is a patient specific approach that provides a good glycemic control in critically ill patients. Nevertheless, its implementation in Malaysian intensive care environments require modifications and improvements in certain areas.

No abstract

Web of Science [v.5.32]-Web of Science Core Collection Full Record apps.webofknowledge.com.ezproxy.um.edu.my/full_record.do?product=WOS&search_mode=GeneralSearch&qid=1&SID=D2NrCuzuHodxl9wclyT&pa… 1/2

Tight glycaemic management has been shown to be beneficial to the outcomes of patients receiving intensive care. However, tight glycaemic control (TGC) protocol within intensive care (ICU) comes with a high clinical demand, namely high nursing effort. Thus, there is a need for a protocol that is safe, effective, robust, yet does not require a high nursing effort. A less intensive protocol is designed to use a combination of subcutaneous long-acting insulin (glargine) with IV insulin bolus and only requires blood glucose (BG) measurements every 4 hours while maintaining measurement within 4.0-6.1 mmol/L.

A clinically verified patient-specific glucose-insulin metabolic model known as ICING is used to account for time-varying insulin sensitivity. ICING was developed and validated from critically-ill patients with various medical conditions in the intensive care unit in Christchurch Hospital, New Zealand. Hence, it is interesting and vital to analyse the compatibility of the model once fitted to Malaysian critically-ill data. Results were assessed in terms of percentage of model-fit error, both by cohort and per-patient analysis. The ICING model accomplished median fitting error of<1% over data from 63 patients. Most importantly, the median per-patients is at a low fitting error of 0.34% and per cohort is 0.35%. These results provide a promising avenue for near future simulations of developing tight glycaemic control protocol in the Malaysian intensive care unit.