Stuart Weinzimer, MD

1
(No Transcript)
2
Stuart Weinzimer, MD

• Associate Professor, Pediatrics
• Yale University School of Medicine

3
The Dawn of the Artificial Pancreas
2006 Annual JDRF ConferenceGlobal Diabetes
Research Forum

• Stuart A Weinzimer, MD
• Associate Professor of Pediatrics
• 12 May 2006

4
Why do we need an artificial pancreas?

• Present methods of diabetes treatment improve,
  but dont normalize, blood glucose levels
• Burden of care extremely high
• Islet cell replacement therapies limited to small
  segments of population and not well-suited for
  children with T1D

5
Design elements for artificial b-cell

• Insulin pump to accurately and precisely deliver
  variable amounts of insulin
• Continuous glucose sensor to accurately determine
  ambient glucose levels
• Effective algorithms to vary insulin delivery
  rates based on real-time glucose sensor outputs

6
Components of an Algorithm for Insulin
Delivery(the Non-Engineers Approach)

• When glucose is high, give more insulin
• When glucose is low, give less insulin
• When glucose is rising, increase insulin delivery
• When glucose is falling, reduce insulin delivery

7
External Closed-loop DevelopmentStep 1 Sensor
Augmented Pump Rx

• external pump
• real-time glucose sensor used 24/7
• clinicians, patients parents devise and adjust
  algorithm
• sensors downloaded for retrospective insulin
  dosage adjustments

8
Example of a real-time algorithm

• calculate of the amount of insulin needed to
  cover the carbohydrates and correct for high BG
• check sensor arrow make following adjustments
• ? (90 Up) increase meal dose by 20
• ? (45 Up) increase meal dose by 10
• ? (No change) no change in meal dose
• ? (45 Down) decrease meal dose by 10
• ? (90 Down) decrease meal dose by 20

9
(No Transcript)
10
Why do we need an artificial pancreas?

• We are built to make errors
• Lewis Thomas

11
In Vivo Physiological Insulin Secretion
Blood Glucose
Insulin
12
Algorithm Components
Blood Glucose
Insulin
Derivative
Integrative
Proportional
13
External Closed-loop DevelopmentStep 2
Algorithm for Insulin Delivery

• Proportional to the glucose level
• Derivative rate-of-change of the glucose
• Integrative slowly adaptive basal rate

14
Prototypic closed-loop system

• Sensor signals transmitted to a laptop computer
  that displays the sensor glucose and calculates
  rate of insulin delivery

• Rate of insulin delivery is transmitted to the
  insulin pump

15
(No Transcript)
16
First PEDIATRIC Closed-Loop Patient
17
Mean glucose levels in first 5 subjects
18
Closed-loop control compared to traditional
open-loop pump therapy
400
300
BG (mg/dL)
200
100
0
Pre-B
Post-B
Pre-L
Post-L
Pre-D
Post-D
19
Effect of peripheral insulin delivery
20
Closed-loop vs. hybrid control
21
Where do we go from here?

• Improving closed-loop / hybrid algorithms
• Optimize prandial control
• Demonstrate persistent reduction in nocturnal
  hypoglycemia
• Characterize performance under dynamic conditions
• Feasibility studies of ambulatory CL control

22
Financial Support

• Juvenile Diabetes Research Foundation
• Medtronic MiniMed
• NIH NICHD / NIDDK
• Novo Nordisk

23
Acknowledgements

• Yale Research Team
• Karena Swan
• Elizabeth Doyle
• Kristin Sikes
• Heather Mokotoff
• Amy Steffen
• Melody Martin
• GCRC Staff

24
(No Transcript)