Bruce W' Bode, MD, FACE

1
Continuous Insulin Infusion
Bruce W. Bode, MD, FACE Atlanta Diabetes
Associates Member of the Diabetes Special
Interest Group of the Georgia Hospital
Association Andrew Ahmann, MD Harold Schnitzer
Diabetes Health Center Oregon Health Science
University
2
Potential Inpatient Protocols

• Insulin infusions
• ICU
• Non-ICU
• DKA
• Labor and delivery
• Subcutaneous insulin order forms
• Incorporate basal/prandial concepts
• Transition to subcutaneous insulin
• Supplementary insulin scale
• Hypoglycemia protocol

3
Potential Inpatient Protocols

• Insulin infusions
• ICU
• Non-ICU
• DKA
• Labor and delivery
• Subcutaneous insulin order forms
• Incorporate basal/prandial concepts
• Transition to subcutaneous insulin
• Supplementary insulin scale
• Hypoglycemia protocol

Vary by goals Prefer consistency of format
4
Protocol Driven Insulin Infusions in the Hospital

• Evidence of improvement in glucose control for
  hospital systems
• Evidence of improved outcomes
• Proven relatively safe
• Accepted by nursing staff (when aware)
• Are most successful with attention to potential
  improvements over time.
• Multiple protocols now published

5
Intravenous Insulin TherapyCurrent Uses

• Diabetic ketoacidosis
• Non-ketotic hyperosmolar state
  
• Critical care illness
• Myocardial infarction or cardiogenic shock
• Post-operative period following heart surgery
  
• NPO status in type 1 diabetes
• Peri-operative care of most surgeries

(continued)
6
Intravenous Insulin TherapyCurrent Uses (cont)

• Exacerbated hyperglycemia during high-dose
  glucocorticoid therapy
• Labor and delivery
• Organ Transplantation
• Dose-finding strategy
• Any illness for which prompt glycemic
  control is important to recovery
  

7
Successful IV Insulin Infusion
110
80
Desired Pattern
Desired Pattern
Undesirable Pattern
8
Insulin Infusion ProtocolsNecessary Components

• The following are evident and well-defined
• When to start the infusion
• A target glucose range
• Precautions for hypoglycemia and its treatment
• Considers initial expected insulin sensitivity
• Reaches goal glucose rapidly and maintains
• Is managed by nurses after a single order
• Adjusts for changes in insulin sensitivity
• Adjusts timing of glucose testing and infusion
  rate change according to glucose stability
• Ultimately includes transition to SC

9
Selecting An IV Algorithm

• No comparative studies of IV insulin therapy
  approaches1-4 There is no one best protocol.
• Review available protocols considering
  comparative elements
• Time to control
• Apparent risk of hypoglycemia
• Complexity relative to staffing situation
• Relevancy to patient population
• Any previous experience of key stakeholders

1. Bode et al. Endocr Pract. 200410(suppl
2)71-80 2. Furnary et al. Endocr Pract. 200410
(suppl 2)21-33 3. Trence et al. J Clin
Endocrinol Metab. 2003882430-2437 4. Goldberg
et al. Diabetes Care. 200427461-467.
10
Initiating An Intravenous Insulin Protocol

• Gather stakeholders to develop a plan
• One or more individuals to choose several
  protocols for review
• Select a protocol and suggest minor modifications
• Inservice the nursing and physician staff
• Do pilot of the protocol on a selected unit
• Collect pre-identified results on each patient
• Time to target
• Hypoglycemia and reason for hypoglycemia
• Protocol errors
• Nursing Satisfaction

11
Initiating An Intravenous Insulin Protocol (cont)

• Group should review results carefully and modify
  procedure if indicated
• May need to change process
• Implement protocol in one unit at a time
• With appropriate inservicing
• Have a clear outline of the review process
• Who does it, how often
• Carefully designed collection tool
• Predetermined dates for data presentation
• Modification, re-education or other corrections
  if needed
• Ongoing review of all serious hypoglycemia

12
Types of Insulin Infusion Protocols

• Those using a column method
• Based on various columns of insulin sensitivity
• Move from one column to another as sensitivity
  changes
• Nurses do no calculations
• Easily computerized
• Those based on present and past glucose
• Tend to be quite flexible/ responsive
• Generally require more nursing calculations
• Somewhat more complicated
• Can also be computerized
• Basic titration
• Less consistent unless nurse adjustments applied

13
Column Method Updated
Braithwaite SS, et al. Diabetes Technology
Therapeutics 2006 8476-488.
14
Column Method Modified by University of Washington

• Target BG level usually 80180 mg/dL
• Protocol 4 Algorithms based on insulin
  sensitivity
• Algorithm 1 for most insulin-sensitive patients
• Algorithm 4 for most insulin-resistant patients
• Most patients start with Algorithm 1
• Insulin-resistant patients start with Algorithm 2
• Insulin infusion rate determined by BG level
• Varies hourly until patient stable in target
  range
• Patient is moved up to next algorithm if BG
  targets not achieved or BG not decreased by at
  least 60 mg/dL in preceding hour

Trence et al. J Clin Endocrinol Metab.
2003882430-2437.
15
Glycemic Management ProtocolBased on Insulin
Sensitivity Example
Trence et al. J Clin Endocrinol Metab.
2003882430-2437.
16
Protocol Based on Insulin Sensitivity Example
(cont)
Trence et al. J Clin Endocrinol Metab.
2003882430-2437.
17
Continuous IV Insulin Infusion Based on BG Levels

• Portland Protocol (2004)start insulin infusion
  via pump piggybacked to normal saline IV as
  follows

Furnary et al. Endocr Pract. 200410(suppl
2)21-33.
18
Continuous IV Insulin Infusion Based on BG Levels

• Insulin titration (Portland 2004)

Furnary et al. Endocr Pract. 200410 (suppl
2)21-33.
19
Continuous IV Insulin Infusion Based on BG Levels
(cont)
Insulin titration (Portland 2004)
Furnary et al. Endocr Pract. 200410 (suppl
2)21-33.
20
Implementation of a Safe and Effective Insulin
Infusion Protocol in a MICU

• Yale University 14 bed MICU
• Had attempted Leuven protocol but were
  unsuccessful
• Developed their own protocol with goal
  100-139mg/dl
• Protocol was initiated when BG exceeded 200 mg/dl
• Did inservicing and follow-up evaluation
• Protocol was predicated on three points
• Current glucose
• Previous glucose value
• Current infusion rate

Goldberg PA et al. Diabetes Care, 27461, 2004 1.
21
Implementation of a Safe and Effective Insulin
Infusion Protocol in a MICU

• RESULTS
• Initial BG 299 mg/dl - - - - mean 10.1 hours to
  reach goal
• After reaching goal
• 52 of values fell within goal range
• 66 in range of 80-139 mg/dl
• 93 in range of 80-199 mg/dl
• 0.3 of values dropped below 60 mg/dl
• Had better glucose levels than historical
  controls
• 73 of nurses characterized the protocol as very
  easy or somewhat easy to use

Goldberg PA et al. Diabetes Care, 27461, 2004 2.
22
Yale Protocol Based On Glucose, Rate of Change,
and Current Insulin Rate

• Bolus and Initial Infusion Rate
• Initial BG level
• 70
• Round to nearest 0.5 U
• If the BG is lt150 mg/dL, the bolus is withheld

Initial BG 335 mg/dL, 335 70 4.78, rounded
up to 5 IV bolus 5 U and infusion
started at 5 U/hr
Goldberg. Endocr Pract. 200612(suppl 3)79-85.
23
Protocol for Changing Rate of Insulin Infusion
(New Yale)
Changes (?) in infusion rate are determined by
the current infusion rate. (Shown on the next
slide.)
D/C INSULIN DRIP ? BG q 30 min when BG 90,
restart drip _at_75 of most recent rate.
Goldberg. Endocr Pract. 200612(suppl 3)79-85.
24
Yale Insulin Infusion Protocol ?s
Goldberg. Endocr Pract. 200612(suppl 3)79-85
25
Van den Berghe ICU Insulin ProtocolAbbreviated

• If gt 100 mg/dl, 2 U/h If gt 200 mg/dl, 4 U/h
• If gt 140 mg/dl, increase by 1 2 U/h
• If 121 to 140 mg/dl, increase by 0.5 1 U/h
• If 111 to 120 mg/dl, increase by 0.1 0.5 U/h
• If 81 to 110 mg/dl, no change
• If 61 to 80 mg/dl, change back to prior rate

• Van den Berghe et al, NEJM 2001345(19)1359

26
Bruce Bode, MDObjectives

• Understand the need for nurse mandated protocols
  for identifying and managing hyperglycemia and
  diabetes in the hospital
• Realize the benefit of a computerized system
• Present what Piedmont Hospital has done to date
  and what tools we are using to accomplish this
  task

27
Patient Presents With Hyperglycemia
Previously diagnosed DM
Diabetic ketoacidosis or hyperglycemic crisis
follow DKA protocol
No previous diagnosis of DM and BG gt140 mg/dL
Modification of therapy to keep BG at goal
Begin BG testing
BG is gt110 mg/dL for a critically ill patient
notify physician for initiation of IV insulin
therapy
BG is gt140 mg/dL for noncritically ill patient,
notify physician for initiation of subcutaneous
therapy
All patients with hyperglycemia should have an
A1C drawn to aid in transition and discharge
therapy
28
Target blood glucose in mg/dL

• 80 110 in ICU patients
• 80 140 in other Surgical and Medical Patients
• 70 100 in Pregnancy

Bode et al Endocrine Practice July 2004
29
Threshold blood glucose in mg/dL
for starting IV insulin
infusion

• Peri-operative care gt 110 - 140
• ICU care gt 110 - 140
• Non-surgical illness gt 140 - 180
• Pregnancy gt 100

Van den Berghes study supports 110
Finneys study supports 145 If drip
indication is failure of SQ therapy, use 180
if indication is specific condition ( DM 1/
NPO, MI, etc ), use 140
30
The Ideal IV Insulin Protocol

• Easily ordered (signature only)
• Effective (Gets to goal quickly)
• Safe (Minimal risk of hypoglycemia)
• Easily implemented
• Able to be used hospital wide

31
Are There Risks to Tight Glucose Control With IV
Insulin?

• Hypoglycemia
• Stimulation of sympathetic nervous system
• ? stroke volume (myocardial O2 demand)
• Potassium / phosphate shift

32
Hypoglycemia Rates (lt40 mg/dl) per pt? Or event
in Randomized ICU Glucose Control Trials
Independent predictor of mortality
33
(No Transcript)
34
Glucose in ACS

• 713 DM patients with ACS
• 2-year all-cause mortality (34)
• Hyperglycemia on admission
• HR 2.66 (95 CI, 1.83 3.86)
• Hypoglycemia during hospitalization
• HR 1.77 (95 CI, 1.09 2.86)

Svensson AM, et al. Eur Heart J.
2005261255-1261.
35

Blood Glucose Post-AMI OutcomesA U-Shaped
Relationship?
P lt 0.001 for each endpoint
Pinto DS, et al. J Am Coll Cardiol.
200546178-180.
36
Hypoglycemia Rates With Published IV Insulin
Protocols
Van den Berghe G, et al. N Engl J Med.
20013451359-1367. Van den Berghe G, et
al. N Engl J Med. 2006354449-461.
Goldberg PA, et al. Diabetes Care.
200427461-467. Goldberg PA, et al. J
Cardiothorac Vasc Anesth. 200418690-697.
Malmberg K, et al. Eur Heart J.
200526650-661. Mehta SR, et al. JAMA.
2005293437-446. Davidson PC, et al.
Diabetes Care. 2005282418-2423.
37
Practical Closed Loop Insulin Delivery
Multiplier Method
A System for the Maintenance of Overnight
Euglycemia and the Calculation of Basal Insulin
Requirements in Insulin-Dependent Diabetics
1/slope Multiplier 0.02
6
5
4
Insulin Rate (U/hr)
3
2
1
0
0
100
200
300
400
Glucose (mg/dl)
NEIL H. WHITE, M.D., DONALD SKOR, M.D., JULIO V.
SANTIAGO, M.D. Ann Int Med 1982 97210-214
38
Continuous Variable Rate IV Insulin DripAtlanta
Multiplier Method

• Starting Rate Units / hour (BG 60) x 0.02
  where BG is current Blood Glucose and 0.02 is
  the multiplier
• Check glucose every hour and adjust drip
• Adjust Multiplier to keep in desired glucose
  target range (80 to 110 in ICU 100 to 140 on
  floor)

39
MultiplierPrinciples
Insulin Units / Hour
Glucose mg/ dl
Davidson et al, Diabetes Care 28(10) 2418-2423,
2005
40
(No Transcript)
41
(No Transcript)
42
(No Transcript)
43
HOUR BG 0 308





1 315
2 288 3 240 4 200 5 173
6 120 7 104 8 71 9 98
44
Ideal Solution

• Computer directed insulin infusion
• Complexity is moved to the computer
• Standardization is achieved
• Hypoglycemia is minimized

Davidson et al, Diabetes Care 28(10) 2418-2423,
2005
45
Computerized Insulin Delivery

• In 1984, R Dennis Steed computerized our insulin
  drip orders into Glucommander
• Product never commercialized because IV insulin
  had not been FDA approved
• On October 2005, FDA approved Novolin R and
  Novolog for IV use
• Glucommander followed by EndoTool have received
  FDA approval

46
Typical Glucommander Run
Glucose
Glucose
Hi
Low
Multiplier
Multiplier
Insulin
Insulin
Hours
Davidson et al, Diabetes Care 28(10) 2418-2423,
2005
47
Computerized IV Insulin Delivery
(Glucommander) Average and Standard Deviation of
of All Runs 1985 to 1998 5808 runs, 120,618 BGs
Davidson et al, Diabetes Care 28(10) 2418-2423,
2005
48
Treating to Target Range
Davidson et al, Diabetes Care 28(10) 2418-2423,
2005
49
Low Range
80 to 120
100 to 120
50
Physician View Writing orders

• High Target Glucose (120 mg/dl)
• Low Target Glucose (90 mg/dl)
• Multiplier (0.02 unless post CABG, then 0.06)
• Maximum interval (120 minutes)
• Insulin concentration (1 unit per ml)

51
Nurse View of Computerized Delivery

• Computer periodically alarms
• Check blood glucose
• Enter glucose into computer
• Set insulin drip to rate from computer
• Eliminates calls to the physician

52
Benefits of a Nurse Mandated Protocol

• All patients are screened and treated the same
• Standardization can be achieved
• Modification of the protocols can easily be done
  based on outcomes and analysis of the data

53

• Nurse Mandated Protocol in CV Patients
• Piedmont Hospital
• Reason Only 50 of our CV patients met the IHI
  criteria on no BG gt 200 mg/dl in the first 48
  hours post surgery
• The protocol was developed by a committee of
  surgeons, nurse and endocrinologists.
• The goal of the protocol is a mean BG of 105
  mg/dl with minimal hypoglycemia and all post-op
  BGs lt200 mg/dl.
• The criteria for treating patients with IV
  insulin was a history of diabetes, BG gt140 mg/dl,
  or two BGs post-op gt110 mg/dl.

54
Presented at the European Association for the
Study of Diabetes September 14.2006
Use of Glucommander Following Cardiovascular
Surgery Resulted in Average Post-Op BG of 107
mg/dl 94 of Patients Had No BG gt200 mg/dl and
Only 2 Had Transient BG lt50 mg/dl
P.C.Davidson, L.Prevosti, H.R.Hebblewhite, V.
Cheekati, R.D.Steed, and B.W. Bode
Atlanta Diabetes Associates
55

• Results of Nurse Mandated Protocol in CV Patients
• At this point 470 CV surgery patient have been
  studied.
• On IV insulin BG stabilized at lt120 mg/dl in
  mean of 3 hrs.
• IV insulin was continued for an average of 37
  hrs.
• 98 were controlled so that no BG was gt200
  mg/dl in the 48 hours following surgery.
• 2 had transient BG lt50 mg/dl.
• The mean for all BG in all patients was 107
  mg/dl.

56
Patients with All BG lt200 mg/dl for 48 hrs
Following CV Surgery Piedmont Hospital
February-July 2006
Start of Glycemic Protocol Using Glucommander
57
Average BGs of All Glycemic Protocol Runs With
Standard Deviation (N 470)
58
Changes in Patient's Insulin Regimen continued
Patients at Admission
59
Changes in Patient's Insulin Regimen (All SubQ
was Basal/Bolus with Glargine Aspart)
60
SOME EXAMPLES OF IV INSULIN SYSTEM
61
A Typical Glucommander Run
62
Transition to SubQ
Managed by Anesthesiology in Operating Room
SubQ Basal-Bolus
Glucommander
0 12 24 36
48 60
hours
63
Intravenous Insulin Infusion Under Basal
Conditions Correlates Well With Subsequent
Subcutaneous Insulin Requirement
Total Intravenous vs. Subcutaneous 24-hour
Insulin Requirements, units
275 250 225 200 175 150 125 100 75 50 25 0
Units SQ
275
250
225
200
175
150
125
100
75
50
25
0
Units IV
Hawkins et al. Endocr Pract. 19951385389.
64
Nurse Mandated Transition from IV insulin to SC
Basal Bolus Insulin

• Criteria for Transition
• History of diabetes
• HbA1c gt6
• Methodology
• Glargine SC given at HS POD 1 if able to eat
• IV insulin discontinued at noon POD2 post am
  meal insulin

65
Transition from Multiplier Method to
Subcutaneous Insulin

• 24-hour insulin requirement
• Multiplier X 1000 TDD
• Give one-half TDD as basal (Glargine)
• Multiplier X 500 BI
• Give rapid acting insulin based on CHO consumed
• 0.5 / multiplier CIR (Gms CHO / unit) or
• 30 X multiplier units / CHO exchange
• Monitor BG a.c. t.i.d., h.s., and 3 am
• Correct all BG gt 140 mg/dL
• (BG - 100) / (1.7 / multiplier)

66
Transition from Glucommander to Basal-Bolus
Insulin Glargine and Aspart Basal Multiplier
500 CIR 0.5 / Multiplier Correction Factor
1.7 / Multiplier
n209
Blood Glucose (mg/dl)
Hours after IV insulin
Breakfast
Breakfast
Breakfast
Bedtime
Bedtime
Bedtime
Last GM
300AM
Lunch
300AM
300AM
Dinner
Dinner
Dinner
Lunch
Lunch
67
CVS Hyperglycemia ProtocolPiedmont Hospital 2006

• Reduced Post-op Length of Stay 0.8 Days
• ICU Cost Per Day 1244.99
• CVS at Piedmont Per Year 998
• Annual Savings 0.8 1245 998 994,000
• Sites in Database of Society of Thoracic Surgeons
  860
• Projected Total National Savings 855 Millions