Hyperglycemia is very common among critically ill patients, even in the absence of diagnosed diabetes or pre-diabetes.

We aimed to determine the prevalence of occult glucose metabolism abnormalities in a general intensive care unit (ICU) and hypothesized that hyperglycemia severity, as reflected by insulin requirements for maintenance of normoglycemia, could be used as a tool to identify high risk patients.

100 consecutive adults patients with an expected stay of 24 hours or more admitted to the Sacré-Coeur Hospital ICU between November 2005 and March 2006 were eligible for enrollment. Our hospital is a 28 beds ICU tertiary center affiliated with the University of Montreal. We have a mixed medical and surgical ICU units treating patients with a variety of conditions including sepsis, respiratory failure, circulatory insufficiency, trauma, as well as cardiac, thoracic and vascular surgery. We excluded patients with a previous diagnosis of diabetes or pre-diabetes. Patients admitted for diabetic ketoacidosis or hyperglycemic hyperosmolar states, treated with chronic corticosteroid therapy at supraphysiological doses or enrolled in another exclusive study were also excluded.

Data collection

At time of admission, demographic data were collected and classical risk factors for diabetes, including high blood pressure, dyslipidemia, coronary artery disease, vascular disease and family history of type 2 diabetes were assessed. Corticosteroid exposure in the 72 hours preceding admission was noted. Physiology and laboratory data were collected for calculation of the Acute Physiologic and Chronic Health Evaluation (APACHE-II) score of illness severity. For each patient, body mass index (BMI) was calculated and waist circumference was measured according to standard technique .

All patients had strict glycemic control according to local regular insulin protocol. Bedside glucose measurements were made every two to four hours using a calibrated plasma glucometer (EliteTM by Bayer). We calculated the daily and cumulative insulin dose administered during the first 72 hours in the ICU in order to maintain a blood glucose level between 72 and 126 mg per deciliter. Co-treatment with other forms of insulin and oral hypoglycaemic agents was prohibited during the study. The daily caloric intake, including feeding, dextrose solution, medication diluent and propofol, as well as the corticosteroids and vasopressors received during that period of time were also calculated.

Random plasma glucose level and glycated hemoglobin (A1C, Homogenous Immunoturbimetric Assay) were measured at baseline. Three months after ICU admission, a second value of A1C was obtained and a standard 75g Oral Glucose Tolerance Test (OGTT) was performed. Based on the OGTT results, patients were classified as having normal or abnormal glucose metabolism according to the criteria published by the American Diabetes Association.

Inclusion Criteria:

• All consecutive ICU adult patients with an expected LOS of at least 24 hours.

Exclusion Criteria:

• We excluded patients with a previous diagnosis of diabetes or pre-diabetes. Patients admitted for diabetic ketoacidosis or hyperglycemic hyperosmolar states, treated with chronic corticosteroid therapy at supraphysiological doses or enrolled in another exclusive study were also excluded.

All adults patients with an expected stay of 24 hours or more who are admitted to our 28 beds ICU tertiary center Sacré-Coeur Hospital ICU.