Atlanta Diabetes Associates Original Title A SEMI-CLOSED LOOP INTRAVENOUS INSULIN ALGORITHM, SHOWN TO BE SAFE, SIMPLE, AND EFFECTIVE IN 82,078 HOURS OF OPERATION New Title GLUCOMMANDER: AN ADAPTIVE, COMPUTER- DIRECTED SYSTEM FOR IV INSULIN, SHOWN TO BE SAFE, SIMPLE, AND EFFECTIVE IN 120,618 HOURS OF OPERATION Atlanta Diabetes Associates Paul C. Davidson, R. Dennis Steed, and Bruce W. Bode

Atlanta Diabetes Associates Glucommander Practical Alternative to Complex IV Insulin Protocols l Computer Based Algorithm for IV Insulin l Invented by Davidson and Steed in 1984 l 19 Years Experience l Developed for Marketing by MiniMed and Boehringer Manheim Corp. l Glucose Management System (GMS) l Shelved Pending FDA Approval of IV Insulin l Useful and Safe for Any Application of IV Insulin

Atlanta Diabetes Associates Intravenous Insulin with Severe Illness Three major recent studies DIGAMI: Prospective Randomized Study of Intensive Insulin Treatment on Long Term Survival After Acute Myocardial Infarction in Patients with Diabetes Mellitus Malmberg, et al. BMJ. 1997;314: Portland: Continuous Insulin Infusion Reduces Mortality in Patients with Diabetes Undergoing Coronary Artery Bypass Grafting Fumary et al J Thorac Cardiovasc Surg 2003;123: Leuven: Intensive Insulin Therapy in Critically Ill Patients Van den Berghe et al N Engl J Med 2001; 345:

Atlanta Diabetes Associates Cardiovascular Risk Mortality After MI Reduced by Insulin Therapy in the DIGAMI Study Malmberg, et al. BMJ. 1997;314: All Subjects (N = 620) Risk reduction (28%) P =.011 Standard treatment Years of Follow-up 2345 Low-risk and Not Previously on Insulin (N = 272) Risk reduction (51%) P =.0004 IV Insulin 48 hours, then 4 injections daily Years of Follow-up

Atlanta Diabetes Associates Mortality of DM Patients Undergoing CABG Fumary et al J Thorac Cardiovasc Surg 2003;123:

Atlanta Diabetes Associates P= P=0.026 BG< <BG<150 BG>150 ICU Mortality Effect of Average BG Van den Berghe et al (Crit Care Med 2003; 31: )

Atlanta Diabetes Associates All three have IV insulin protocols Complex Require ICU housing Specially trained nurses Dedicated supervision Consequently not widely accepted IV Insulin Based Studies DIGAMI, Portland, Leuven

Atlanta Diabetes Associates 1. Start Portland protocol during surgery and continue through 7 AM of the third POD. Patients who are not receiving enteral nutrition on the third POD should remain on this protocol until receiving at least 50% of a full liquid or soft American Diabetes Association diet. 2. For patients with previously undiagnosed DM who have hyperglycemia, start Portland protocol if blood glucose is greater than 200 mg/dL. Consult endocrinologist on POD 2 for DM workup and follow-up orders. 3. Start infusion by pump piggyback to maintenance intravenous line as shown in Appendix Table Test blood glucose level by finger stick method or arterial line drop sample. Frequency of blood glucose testing is as follows: a. When blood glucose level greater than 200 mg/dL, check every 30 minutes. b. When blood glucose level is less than 200 mg/dL, check every hour. c. When titrating vasopressors, (eg, epinephrine) check every 30 minutes. d. When blood glucose level is 100 to 150 mg/dL with less than 15 mg/dL change and insulin rate remains unchanged for 4 hours (“stable infusion rate”), then you may test every 2 hours. e. You may stop testing every 2 hours on POD 3 (see items 1 and 8). f. At night on telemetry unit, test every 2 hours if blood glucose level is 150 to 200 mg/dL; test every 4 hours if blood glucose level is less than 150 mg/dL and “stable infusion rate” exists. 5. Insulin titration according to blood glucose level is performed as follows a. When blood glucose level is less than 50 mg/dL, stop insulin and give 25 mL 50% dextrose in water. Recheck blood glucose level in 30 minutes. When blood glucose level is greater than 75 mg/dL, restart with rate 50% of previous rate. b. When blood glucose level is 50 to 75 mg/dL, stop insulin. Recheck blood glucose level in 30 minutes; if previous blood glucose level was greater than 100 then give 25 mL 50% dextrose in water. When blood glucose level is greater than 75 mg/dL, restart with rate 50% of previous rate. c. When blood glucose level is 75 to 100 mg/dL and less than 10 mg/dL lower than last test, decrease rate by 0.5 U/h. If blood glucose level is more than 10 mg/Dl lower than last test, decrease rate by 50%. If blood glucose level is the same or greater than last test, maintain same rate. d. When blood glucose level is 101 to 150 mg/dL, maintain rate. e. When blood glucose level is 151 to 200 mg/dL and 20 mg/dL lower than last test, maintain rate. Otherwise increase rate by 0.5 U/h. f. When blood glucose level is greater than 200 mg/dL and at least 30 mg/dL lower than last test, maintain rate. If blood glucose level is less than 30 mg/dL lower than last test (or is higher than last test), increase rate by 1 U/h and, if greater than 240 mg/dL, administer intravenous bolus of regular insulin per initial intravenous insulin bolus dosage scale (see item 3). Recheck blood glucose level in 30 minutes. g. If blood glucose level is greater than 200 mg/dL and has not decreased after three consecutive increases in insulin, then double insulin rate. h. If blood glucose level is greater than 300 mg/dL for four consecutive readings, call physician for additional intravenous bolus orders. 6. American Diabetes Association 1800-kcal diabetic diet starts with any intake by mouth. 7. Postmeal subcutaneous Humalog insulin supplement is given in addition to insulin infusion when oral intake has advanced beyond clear liquids. a. If patient eats 50% or less of servings on breakfast, lunch, or dinner tray, then give 3 units of Humalog insulin subcutaneously immediately after that meal. b. If patient eats more than 50% of servings on breakfast, lunch, or supper tray, then give 6 units of Humalog insulin subcutaneously immediately after that meal. 8. On third POD, restart preadmission glycemic control medication unless patient is not tolerating enteral nutrition and is still receiving an insulin drip. Portland Protocol Furnary et al J Thorac Cardiovasc Surg 2003;123:

Atlanta Diabetes Associates Complexity versus Simplicity. Arterial BG q 1-2 hours, then q 4 hours if stable. If BG >220 give 4 units/hr. If BG >110 mg/dl give 2 units/hr.. If F/U BG in 1-2 hours >140 mg/dl Increase insulin 1-2 units/hr.. If F/U BG in 1-2 hours mg/dl increase insulin unit/hr.. If F/U BG mg/dl increase insulin units/hr.. If BG mg/dl then do not change.. If BG decreases >50% decrease insulin 50%.. If BG mg/dl decrease insulin “reduced as dictated by previous BG level.. Repeat BG in one hour.. If B mg/dl discontinue insulin.. If BG >40 mg/dl give 10 Gm glucose IV. Repeat q 1 hr until BG mg/dl.. If BGT decreases >20% in mg/dl range decrease insulin 20%.. If patient transferred from ICU and insulin <2 units/hr, DC insulin.. If patient transferred from ICU and insulin >2 units/hr get endocrine consult. Van den Berghe Orders Glucommander Orders Requires ICU nurses trained in protocol and study physicianAdministered by floor nurse and any physician

Atlanta Diabetes Associates Glucommander. Summary of Performance Glucose Averages for 3404 Patients Glucose mgm/dl 50 Percentiles90 10 Hours Percentiles

Atlanta Diabetes Associates A System for the Maintenance of Overnight Euglycemia and the Calculation of Basal Insulin Requirements in Insulin-Dependent Diabetics NEIL H. WHITE, M.D., DONALD SKOR, M.D., JULIO V. SANTIAGO, M.D.; Saint Louis, Missouri Ann Int Med 1982 ;97: Practical Closed Loop Insulin Delivery 1/slope = Multiplier = Glucose (mg/dl) Insulin Rate (U/hr) INSPIRATION FOR GLUCOMMANDER

Atlanta Diabetes Associates Historical Perspective l IV Insulin Algorithm –Insulin = (BG-60) x Multiplier l “White’s” Multiplier Not Applicable for Majority –Based on Type 1 Pediatric Pump Patients –IV Insulin Used Frequently in Stressed Type 2 Only 14% Stabilized at 0.02 Glucommander Multipliers N=2364 Runs White = 0.02

Atlanta Diabetes Associates Glucommander 5802 Runs and 120,618 BG’s

Atlanta Diabetes Associates Glucommander Principles Insulin Units / Hour Glucose mgm / dl

Atlanta Diabetes Associates Hours Glucose Multiplier Insulin Glucose Typical Glucommander Run Hi Low

Atlanta Diabetes Associates Glucommander Average and Standard Deviation of of All Runs 1985 to 1998; 5808 runs, 120,618 BG’s

Atlanta Diabetes Associates Glucommander ………………….……………………………….. Complete Data Set 1985 to 1998 Beyond Data Analyzed by Boehringer Manheim/MiniMed in years of data from Glucommander Runs over 120,618 hours. Correction of hyperglycemia: Mean starting BG=259 mg/dL (SD 127). Mean stable <150 after three hours. Subsequent stability in target range for 60 hrs. 90% of patients achieved BG<180 within 8 hrs. Experience with Hypoglycemia:  BG’s <50 were 0.6% of total BG’s.  2.6% all runs had one BG <40. All were immediately corrected to 100 with IV glucose, insulin held 30 min, then modified.  No severe hypoglycemia.

Atlanta Diabetes Associates Lauren <40 mg/dl 5.2% Hypoglycemia on Glucommander 5772 Runs

Atlanta Diabetes Associates Glucommander. Correction of Hypoglycemia IV 50% Glucose: ( 100-BG) X 0.15 Grams Time (min) Glucose (mg/dl) N = 886

Atlanta Diabetes Associates Conformity of Blood Glucose to Glucommander Target

Atlanta Diabetes Associates IV Insulin Protocols l Correct with minimal insulin –Least reactive hypoglycemia –Cut insulin quickly l Correct hyperglycemia quickly –Limit intracellular dehydration –Start insulin aggressively l Avoid prolonged hyperglycemia –Less intracellular edema with correction l Many protocols in use –Few with outcomes ADA Diabetes Care 26:S109-S117,2003 Watts Diabetes Care 10:722-28,1987 Umpierrze Personal Commication Markovitz Endocr Pract 8:10-18,2002 Metchick Am J Med 133: , 2002 Van den Berghe N Engl J Med 346:1586-8, 2002 Fumary J.Thor CV Surg 125: , 2003

Atlanta Diabetes Associates Glucommander Comparsion to Other Systems Insulin Units / Hour Glucose mgm / dl Glucommander 33 u ADA 38 u MARKS 52 u FUMARY 19 u METCHICK 37 u VAN DEN BERGHE 41 u IV DRIP 38 u UMPIERRZE 34 u MARKOVITZ 33 u WATTS 46 U LEVETAN 32 u

Atlanta Diabetes Associates Insulin Units / Hour Glucose mgm / dl Glucommander 33 u ADA 38 u IV DRIP 38 u MARKOVITZ 33 u Glucommander Similar Systems Features in Common Early high dose Decrease in parallel with BG End up at common dose Similar total dose

Atlanta Diabetes Associates Glucommander. Surgical Series Compared to Watts Algorithm Watts Glucommander Watts et al Diab Care :

Atlanta Diabetes Associates Glucommander. Surgical Series Compared to Watts Algorithm Glucommander Watts

Atlanta Diabetes Associates How has the Glucommander been used? l Treatment of ketoacidosis l Hyperosmolar non-ketotic state l Perioperative glucose management l Labor and delivery l Myocardial infarction l Critically ill patients in ICU l Hyperalimentation l Gastroparesis with intractable nausea and vomiting l Estimating a patient’s insulin sensitivity –A guide for dosing insulin Estimating total insulin dose, correction factor, CHO/Ins

Atlanta Diabetes Associates Clinical Experience with Glucommander l Simple, safe, and effective method for maintaining glycemic control l Extensively studied l Standardized treatment method applicable in a wide variety of conditions l Available for review, l Opportunity to improve clinical outcome now not when and if

Atlanta Diabetes Associates Glucommander Available for review on internet Slides available at