Risk-based Dosing of Insulin and Nutrition Improves Glycaemic Control Outcomes

Objective:
Hyperglycaemia and insulin resistance are common in critically ill patients and associated with worsened outcomes. STAR (Stochastic TARgeted) glycaemic control (GC) has proven effective over different units and clinical practices. Unlike many protocols, STAR also modulates nutrition with insulin, using a patient-specific risk-based dosing approach to provide greater flexibility in control. This study compares and assesses safety and efficacy of the ongoing STAR clinical trial results at the University Hospital of Liège, Belgium.
Method:
Two arms are compared: the first uses an insulin only version of STAR (STAR-IO), and the second the full insulin+nutrition version of STAR. The target band is 80-145mg/dL. Insulin is administered IV and nutrition is administered enterally. GC was stopped after 72h or if BG was stable at insulin rate ≤2U/h. Safety is assessed by %BG <80mg/dL below target and hyperglycaemia (%BG>180mg/dL). Performance is evaluated by %BG within target band and median BG. Clinical data from 11 patients on STAR-IO and 10 patients on STAR totalling 1100 hours of control is used. Ethics approval was granted by the University Hospital of Liège Ethics Committee.
Results:
STAR performance is statistically significantly better compared to STAR-IO (89% vs. 78% for %BG in target band, p<0.01 using Fisher Exact test). Median [IQR] BG is similar but tighter in STAR (118[109 129] vs. 120[107 138]mg/dL, p=0.19 using Wilcoxon rank sum test). STAR is also safer compared to STAR-IO with 0.7% vs. 1.4% for %BG<80 mg/dL and only 2.0% vs. 9.8% for %BG>180mg/dL. This outcome was achieved using less insulin and nutrition rates for STAR vs STAR-IO (3.0[2.0 4.0] U/h vs. 3.5[1.5 6]U/h and 7.0[4.7 8.2] vs. 8.1[4.9 9.2]g/h).
Conclusions:
Modulating nutrition in addition to insulin can significantly improve GC outcomes, especially by reducing nutrition rates for highly resistive patients.