Increasing evidence from observational studies in hospitalized patients with and without diabetes indicates that hyperglycemia is a predictor of poor outcome. Over the short-term, hyperglycemia can adversely affects fluid balance (through glycosuria and dehydration), impairs immunologic response to infection, and promotes inflammation and endothelial dysfunction. Blood glucose control with intensive insulin therapy in patients with acute critical illness reduces the risk of multiorgan failure and systemic infections, and decreases short- and long-term mortality .

- Hypotheses: we hypothesize that management of inpatient hyperglycemia with a computer-guided intravenous infusion protocol (Glucommander) will facilitate a smoother glycemic control with a lower rate of hypoglycemic events than treatment following a standard insulin infusion algorithm or a simple calculated infusion protocol in critically ill patients in the medical and surgical ICU.

I. RESEARCH OBJECTIVES AND SPECIFIC AIMS

A. Introduction:

Increasing evidence from observational studies in hospitalized patients with and without diabetes indicates that hyperglycemia is a predictor of poor outcome. Over the short-term, hyperglycemia can adversely affects fluid balance (through glycosuria and dehydration), impairs immunologic response to infection, and promotes inflammation and endothelial dysfunction. Blood glucose control with intensive insulin therapy in patients with acute critical illness reduces the risk of multiorgan failure and systemic infections, and decreases short- and long-term mortality .

The use of intravenous insulin infusion is the preferred route of insulin administration for the management of diabetic subjects with diabetic ketoacidosis and nonketotic hyperosmolar state, intraoperative and postoperative care, the postoperative period following heart surgery and organ transplantation, acute myocardial infarction, stroke, and critical care illness. Some of these settings may be characterized by, or associated with, severe or rapidly changing insulin requirements, generalized patient edema, impaired perfusion of subcutaneous sites, requirement for pressor support, and/or use of total parenteral nutrition. In these settings, the intravenous route for insulin administration has been considered superior to the subcutaneous injection of split-mixed regimen of intermediate and regular insulin with respect to rapidity of effect in controlling hyperglycemia, overall ability to achieve glycemic control, and most importantly, preventing hypoglycemic episodes. Recently, several insulin infusion protocols have been reported in the literature . In general, orders to "titrate drip" are given to achieve a target blood glucose range using an established algorithm or by the application of mathematical rules by nursing staff. These algorithms and formulas, however, may be confusing and difficult to follow and may increase the risk of dosing errors. To facilitate patient care, insulin algorithms could be placed on a computer and used at the patient bedside to direct the nursing staff administering the intravenous insulin. The Glucommander is one such computer-derived insulin infusion protocol which has been used successfully in over 5,802 patients with diabetes between 1984 and 1998.

B. Hypotheses: we hypothesize that management of inpatient hyperglycemia with a computer-guided intravenous infusion protocol (Glucommander) will facilitate a smoother glycemic control with a lower rate of hypoglycemic events than treatment following a standard insulin infusion algorithm or a simple calculated infusion protocol in critically ill patients in the medical and surgical ICU.

C. Specific Aim: to determine differences in glycemic control between treatment with computer-guided intravenous infusion protocol (Glucommander), a standard insulin infusion algorithm and a simple calculated infusion protocol in critically ill patients in the medical and surgical ICU.

II. BACKGROUND AND CURRENT STATUS OF WORK IN THE FIELD.

The result of several observational and interventional studies indicate that hyperglycemia is associated with poor hospital outcomes including prolonged hospital stay, infections, disability after hospital discharge, and death , and that improvement in outcomes can be achieved with improved glycemic control in patients with critical and surgical illness. Although there are still no proven mechanisms to explain the detrimental effects of hyperglycemia, there are increasing efforts worldwide to improve and maintain strict glycemic control in subjects with critical illness.

In 2001, a large prospective, randomized controlled trial from Leuven, showed that near normalization of blood glucose levels using an intensive insulin protocol improved clinical outcomes in patients admitted to a surgical intensive care area . In that study, insulin administration to maintain blood glucose levels between 80-110 mg/dl, reduced ICU mortality by 34%, and reduced the risk of multiorgan failure, systemic infection, incidence of acute renal failure, and the need for blood transfusions and prolonged mechanical ventilatory support. Interventional studies in the setting of acute coronary events have shown that intensive insulin therapy resulted in decreased short- and long-term mortality. Similarly, attainment of targeted glucose control in the setting of cardiac surgery is associated with reduced mortality and with a significant reduction in deep sternal wound infections. Similarly, in the setting of acute neurological illness, stroke, and head injury extensive observational and interventional studies indicates that hyperglycemia is associated with increased mortality and with diminished neurological recovery . Based in these observational and interventional studies, aggressive control of blood glucose is recommended in patients with critical illness. A recent position statement of the American Association of Clinical Endocrinologists recommended glycemic targets for hospitalized patients in the intensive care unit between 80 - 110 mg/dL.

The American College of Endocrinology position statement supports the following indications for intravenous insulin therapy in hospitalized patients with diabetes:

• Prolonged fasting (> 12 hours) in type 1 diabetes
• Critical illness
• Before major surgical procedures
• After organ transplantation
• Diabetic ketoacidosis
• Total parenteral nutrition therapy
• Labor and delivery
• Myocardial infarction
• Other illnesses requiring prompt glucose control.

Institutions around the world use a variety of insulin infusion algorithms that can be implemented by nursing staff. These algorithms facilitate communication between physicians and nurses, achieve correction of hyperglycemia in a timely manner, provide a method to determine the insulin infusion rate required to maintain blood sugars within a defined target range, include a rule for making temporary corrective increments or decrements of insulin infusion rate without under- or overcompensation, and allow for adjustment of the maintenance rate as patient insulin sensitivity or carbohydrate intake changes. In most insulin infusion protocols, orders to "titrate drip" are given to achieve a target blood glucose range using an established algorithm or by the application of mathematical rules by nursing staff. These algorithms may be confusing and difficult to follow which may increase the risk of errors.

To facilitate patient care, Drs. Davidson and Steed developed the Glucommander in 1984, a computer based system for glycemic control in hospitalized patients. This computer-guided insulin infusion system directs the administration of intravenous insulin in response to blood glucose measurement at the patient's bedside. Prior to starting the insulin infusion, the physician must specify the following parameters: the low end and high end of the target range for blood glucose, an initial factor or multiplier, and the maximum time interval between blood glucose measurements. During the infusion, the nurse enters blood glucose levels and the computer recommends the insulin infusion rate and a time to check the next blood glucose testing. The starting insulin infusion follows the formula: insulin / hour = multiplier x (BG - 60). The "multiplier" is a parameter that is automatically adjusted based on the glucose pattern and response to insulin. For most adults, the initial multiplier is 0.02. The Glucommander is programmed to adjust the multiplier upwards or downwards if blood glucose levels are above or below target levels.

This study will compare three methods of comparing how glucose is regulated with IV insulin. The first method uses a sliding scale of insulin based on glucoses drawn at set intervals. The second method uses an algorithm that uses the glucose drawn at a set interval and also how much that glucose changed from the last time the glucose was drawn to adjust the amount of insulin to be given. The algorithm takes into count how fast the glucose is changing and also how far the glucose is from the expected values to improve the accuracy of the amount of insulin to be given. The glucocommander is a handheld based algorithm that again takes in how fast the glucose has changed and how far the glucose is from where you want to be and adds another layer in that it tells the nurse when to draw the next glucose so that if the glucose is very far away from the expected value drawing glucose in a shorter time period will help get to the expected glucose faster. The comparison is does the method reach the expected glucose faster and how often is the glucose going to be to low. Methods that improve the way to get the glucose to the expected value are not better for the patient if they increase the chance of hypoglycemia. This study will compare three methods of regulating glucose with insulin.

Inclusion Criteria:

• Patients who are admitted to Medical or Surgical ICU.
• History of diabetes mellitus
• Newly diagnosed hyperglycemia (defined as a blood glucose greater than 140 mg/dl on ≥ 2 occasions)
• Subjects must have an admission blood glucose < 500 mg/dL, without laboratory evidence of diabetic ketoacidosis (serum bicarbonate < 18 mEq/L or positive serum or urinary ketones).

Exclusion Criteria:

• Non-Diabetic patients
• Subjects with acute hyperglycemic crises such as diabetic ketoacidosis (DKA) and hyperosmolar hyperglycemic state.
• Patients with known HIV
• Patients with severely impaired renal function (serum creatinine ≥3.0 mg/dl).
• Patients with mental condition rendering the subject unable to understand the nature, scope, and possible consequences of the study.
• Female subjects who are pregnant or breast feeding at time of enrollment into the study

Limitations of trial. Variable length of stay the first 48 hours was analyzed as the sample size decreased after that.