Intensive Insulin Therapy in the Intensive Care Unit: Assessment by Continuous Glucose Monitoring

De Block C, Manuel-Y-Keenoy B, Van Gaal L, Rogiers P. Diabetes Care. 2006;29:1750-1756.

Hyperglycemia is a common occurrence in critically ill patients in the medical intensive care unit (MICU), leading to high rates of morbidity and mortality. Clinical interventions (eg, enteral or parenteral nutrition, dialysis, corticosteroids) can promote hyperglycemia. Additionally, changes in carbohydrate metabolism and frequent changes in insulin requirements may induce hyperglycemia in both diabetic and nondiabetic patients. Intensive insulin therapy to sustain normoglycemic levels of 80 to 110 mg/dL reduces morbidity and mortality in patients in the MICU. Accordingly, the aim of this study is to assess the reliability of continuous glucose monitoring (CGM) in the MICU and to use CGM to evaluate whether insulin regimens based on discontinuous glucose monitoring are effective for achieving normoglycemia.

A prospective, observational study comprised of 30 diabetic and 20 nondiabetic critically ill patients was conducted at a 13-bed MICU in which the nurse:patient ratio was 1:2.5. Exclusion criteria included pregnancy, a “do not reanimate” code, or had an anticipated MICU stay <3 days. Insulin regimens varied according to glycemic state at admission. Subcutaneous insulin therapy was administered if the admission glucose level was 150-199 mg/dL (5 units), 200-249 mg/dL (10 units), 250-299 mg/dL (15 units), or ≥300 mg/dL (20units); blood sugar values were repeatedly tested every 3 hours. An intravenous insulin regimen was prescribed if the patient had an admission glucose level >200 mg/dL. Illness severity, assessed by the Acute Physiology and Chronic Health Evaluation II (APACHE-II) and the Sequential Organ Failure Assessment (SOFA) scores, also determined the requirement for intravenous therapy. All patients initially received intravenous fluids upon admission to the MICU, after which parenteral or enteral nutrition began in 24 hours. Within 3 days post-admission, 48-hour CGM was initiated using the glucose sensor GlucoDay, which measures glucose concentration from subcutaneous interstitial fluid. GlucoDay devices were calibrated using 2-point and 6-point calibration. Error grid analysis (EGA) was used for point-to-point comparison of both sets of values to arterial blood glucose measurements. The accuracy of CGM and its potential suitability for directing clinical action was investigated using CG-EGA. This method was applied in 5 patients and consisted of using blood glucose values collected every 15 minutes in 3 separate 4-hour blocks.

The study included 22 patients on intravenous insulin therapy and 28 on subcutaneous insulin therapy. Median APACHE-II and SOFA scores were 22 and 8, respectively. Patients on the subcutaneous regimen spent more time at glycemia >110 mg/dL or <60 mg/dL, but had a mean blood glucose level similar to the intravenous group. However, patients in the intravenous regimen group received higher insulin doses than those in the subcutaneous regimen group (eg, 141 ± 100 units/day vs 15 ± 26 units/day, respectively. On the whole, only 22 ± 18% of patients attained the target normoglycemic range of 80-110 mg/dL, and glycemic values >140 mg/dL and <60 mg/dL were observed, respectively, 39 ± 27% and 5 ± 10% of the time. Moreover, although age, sex, and body mass index (BMI) were the same in both of the insulin therapy groups, septic shock and diabetes were more frequently seen in the intravenous regimen; further, these patients were more often receiving intravenous corticosteroids, parenteral nutrition, or renal replacement therapy. No difference was noted between diabetics and nondiabetics regarding age, sex, or severity of illness. However, patients with diabetes had a higher BMI, were more in need of a higher insulin dose, and had a higher mean A1C value. Patients who died in the MICU were older, and had higher APACHE-II and SOFA scores and higher lactate levels but similar glycemic values compared to their counterparts; mortality was not influence by diabetes, insulin regimen, or insulin dose. CGM correlated linearly with discontinuous arterial blood glucose measurements. The correlation was stronger for 6-point (r =.85) vs 2-point calibration (r =.76). Importantly, rapid changes in blood glucose levels were noted immediately with CGM, whereas changes were noted after approximately 1 to 2 hours using discontinuous glucose monitoring.

Many key issues could be concluded from this study. Hyperglycemia is prevalent in the MICU. In this study, 74% of MICU patients experienced hyperglycemia, and patients were normoglycemic only 22% of the time. Individuals on the subcutaneous insulin regime had poorer glycemic control than the intravenous group; this shows that the current methodology of insulin therapy is not achieving targeted glycemic levels and warrants the possibility of abdicating subcutaneous protocol. In addition, discontinuous glucose monitoring does not result in adequate glycemic control. CGM may promote recognition of glycemic excursions and facilitate optimization of ICU insulin therapy.