[Our experience in continuous administration of insulin using infusion dispenser at a metabolic intensive care unit (MICU)].

Introduction: The article summarises the experience of the authors in the administration of insulin using an infusion dispenser at a metabolic intensive care unit (MICU) both to patients with decompensated diabetes mellitus and to patients admitted with a sepsis or other life-threatening condition of internal nature.

Patients And Results: Insulin was administered with the use of a dispenser to a total of 50 patients during 6 months of 2005. 13 of those patients showed signs of ketoacidotic or hyperosmolar coma in the course of diabetes mellitus. All of these patients were compensated within 24 hours and transferred to a standard ward for further treatment within 48 hours. 37 patients were admitted in a critical condition, the cause ofwhich was sepsis and a serious internal disease in 29 and 8 patients, respectively. 12 of the patients died during their hospitalisation at MICU, of which 8 in the first 3 days after admission. No significant correlation between the age, diabetes mellitus diagnosis or an associated cardiovascular morbidity and the death at MICU was discovered, but there was a very close ling between the mortality at the intensive care unit and the baseline blood level of C-reactive protein (160 mg/l; 32-352 in the patients who died, and 111 mg/l 15-168 in the patients who survived), p < 0.01. Glycaemia at admission did not differ significantly for the patients who dies and those who survived, but average glycaemia for all three measurements at MICU was significantly higher in the patients who died (10.4 mmol/l; 6.2-22.4) as compared with those who survived (7.8 mmol/l; 5.8-16.6), p < 0.01. The time of insulin administration was significantly shorter in patients who died (3.3 days; 1-6) as compared with those who survived (5.2 days; 3-10), p < 0.01. There was no significant difference between hourly insulin dose in the patients who died (2.8 j/hour; 0.6-8.6) and in those who survived (2.6 j/hour; 0.8-7.6). A trend towards lower mortality was recorded for the group of patients with average glycaemia below 8 mmol/l and/or those in whom glycaemia mostly ranged between 4.4 and 8.0 mmol/l, but the difference was not statistically significant. A significantly lower consumption of insulin was recorded for the patients with average glycaemia below 8 mmol/l and/or those whose glycaemia measurements mostly ranged between 4.4 and 8.0 mmol/l. Hypoglycaemia defined as glycaemia below 4.4 mmol/l was present in 2% of all measurements, in 11 patients on the total, and their results were not significantly associated with mortality at MICU.

Conclusion: Mortality of patients admitted with sepsis or other life-threatening condition of internal nature was significantly higher in the group of patients with higher average glycaemia among all the measurements performed at MICU. In patients who died, the total time of insulin administration was significantly shorter, but there was no difference between the average hourly insulin dose in the group of the patients who died and those who survived.