Evaluation of Factors Affecting CGMS Calibration

1
Evaluation of Factors Affecting CGMS Calibration
Bruce Buckingham,1 Craig Kollman,2 Roy W Beck,2
Andrea Kalajian,2 Rosanna Fiallo-Scharer,3
Michael J Tansey,4 Larry A Fox,5 Darrell M
Wilson,1 Stuart A Weinzimer,6 Katrina J Ruedy,2
and William V Tamborlane6 for the DirecNet Study
Group 1. Stanford, CA 2. Tampa, FL 3. Denver,
CO. 4. Iowa City, IA 5. Jacksonville, FL 6.
New Haven, CT
2
Abstract_title Evaluation of Factors Affecting
CGMS CalibrationAbstract Objective To explore
the optimal number and timing of calibration
values entered into the Continuous Glucose
Monitoring System ("CGMS" Medtronic MiniMed,
Northridge, CA).Research Design and Methods 
Fifty subjects with T1DM (aged 10-18y) were
hospitalized in a clinical research center for
24h on two different days.  CGMS and One-Touch
Ultra Meter ("Ultra" Lifescan, Milpitas, CA)
data were obtained.  The CGMS was retrospectively
recalibrated using the Ultra data varying the
number and timing of calibration values. 
Resulting CGMS values were compared against
laboratory reference values.Results  There was
a modest improvement in accuracy with increasing
number of calibrations.  The median relative
absolute deviation (RAD) was 14, 15, 13 and
13 when using 3, 4, 5 and 7 calibration values,
respectively (plt0.001).  Corresponding
percentages of CGMS-reference pairs meeting the
ISO criteria were 66, 67, 71 and 72,
respectively (plt0.001).  Nighttime accuracy
improved when daytime calibrations (pre-lunch and
pre-dinner) were removed leaving only two
calibrations at 9p.m. and 6a.m. (median
difference -2 vs. -9mg/dL, plt0.001 median RAD
12 vs. 15, plt0.001 ISO 73 vs. 67,
p0.003).  Accuracy was significantly better on
visits where the average absolute rate of glucose
change at the times of calibration was lower.  On
visits with average absolute rates lt0.5,
0.5-lt1.0, 1.0-lt1.5 and 1.5mg/dL/min, median RAD
values were 11 vs. 15 vs. 18 vs. 19 (p0.01)
and ISO percentages were 76 vs. 68 vs. 58 vs.
59 (p0.02), respectively.Conclusions 
Although accuracy is slightly improved with more
calibrations, the timing of the calibrations
appears more important.  Modifying the algorithm
to put less weight on daytime calibrations for
nighttime values and calibrating during times of
relative glucose stability may have greater
impacts on accuracy.
3

• The Medtronic Minimed continuous glucose
  monitoring system (CGMS) uses a retrospective
  calibration based upon 3-4 glucose meter test
  results/day. Our studies were designed to answer
  the following questions
• Will more than 3 or 4 calibration values
  improve sensor accuracy?
• Will the addition of postprandial to
  preprandial calibration values improve accuracy
  by providing a broader glucose range?
• Does the rate of change of glucose values at
  the time of calibration affect sensor accuracy?
• How does the use of daytime calibrations affect
  nighttime readings?

4


Fifty subjects with type 1 diabetes (average age
was 14.8 1.7 years) were hospitalized in a
clinical research center on two different days.
During one visit, they exercised on a treadmill.
A CGMS Gold sensor was calibrated using a
One-Touch Ultra Meter. Intravenous blood
samples for reference serum glucose values were
obtained every 20 minutes during the exercise
session, and hourly overnight from 10 p.m. to 6
a.m., and measured in a central laboratory
(University of Minnesota). Glucose
measurements with the Ultra meter were also
obtained before lunch and dinner and on the hour
starting at 2 p.m. A computer program provided
by Medtronic MiniMed was used to recalibrate the
CGMS using 3, 4, 5 or 7 Ultra values
5
Table 1. Schedule of Recalibrations
6

• Standard measurements
• Difference CGMS minus laboratory reference
• Relative absolute difference (RAD absolute
  value of the difference divided by the reference
  expressed as a percentage)
• ISO criteria (CGMS within 15 mg/dL for
  reference 75 mg/dL and within 20 for reference
  gt75 mg/dL)

Analyses were limited to sensors functioning for
at least 15 hours and visits where both Ultra and
CGMS measurements were available at all 7
calibration time points listed in Table 1. This
resulted in 12 of the 100 visits being excluded
from analysis. The bootstrap resampling
technique was used to account for correlated data
from the same subject.
7
Analyses of false positives for hypoglycemia
based on events rather than discrete
points. Episode At least 2 CGMS values 70
mg/dL with no intermediate values gt80 mg/dL.
Distinct episodes had to be separated by at least
30 min. True positive at least one reference
value 70 mg/dL during episode. False positive
reference value gt10 mg/dL higher than the
concurrent CGMS value. Non-evaluable If
neither of these conditions were met.
Rate of change during calibration was calculated
using CGMS values 10 minutes apart (5 minutes
prior to and following the time of each
calibration). Absolute rate of change was
averaged over 4 calibration times (pre-lunch,
pre-dinner, 9 p.m., 6 a.m.) for each visit. For
analyses of rate of change during calibration,
accuracy measures were adjusted for the rate of
change during the reference measurement by giving
equal weight to each category (lt0.5, 0.5-lt1.0,
1.0-lt1.5 and 1.5 mg/dL/min).
8

• Original Calibrations Median number of
  calibrations 5.
• Median RAD 15 overall, and 18 when reference
  values 70 mg/dL,
• ISO criteria met 64.

Table 2 Effect of of Calibrations on CGMS
accuracy
Additional calibration values made a modest
improvement in accuracy (plt0.001) This trend
appeared more pronounced during hypoglycemia
Increasing the number of calibrations did not
correct the tendency for the CGMS to read low
overnight
9
Table 3 Effect of rate of change during
calibration of blood glucose on sensor
calibration and accuracy
The CGMS was more accurate when the rate of
change was lower (p0.001)
Table 4 False positive alarm rates
With additional calibrations, there was a trend
towards lower overnight false positive rates
(p0.08), and there was trend for lower false
alarm rates when the rate of change was slower
(p0.19).
10

• Timing of Calibrations
• Additional re-calibration schemes were run to
  explore how the timing of the calibrations might
  affect accuracy. The 4-calibration scheme
  (pre-lunch, pre-dinner, 9 p.m., 6 a.m.) was used
  as a baseline for comparison.
• Removing the two daytime (pre-lunch and
  pre-dinner) calibrations actually improved
  nighttime accuracy (median difference 2 vs. 9
  mg/dL, plt0.001 median RAD 12 vs. 15, p0.001
  ISO 73 vs. 67, p0.004.
• Changing the 3 daytime calibration values from
  pre-prandial to postprandial (i.e., calibrating
  at 2 p.m., 7 p.m., 10 p.m. and 6 a.m.) did not
  affect accuracy. The median RAD 14 vs. 15,
  p0.50 ISO percentage 68 vs. 67, p0.62

11

• Our results showed a slight improvement in
  sensor accuracy when additional calibration
  values are added with the median RAD improving
  from 14-15 with 3 or 4 calibrations each day, to
  13 with 5 or 7 calibrations each day.
• Additional calibration values also tended to
  improve sensor performance during exercise

• Our results confirm that sensor accuracy is
  decreased when blood glucose levels are changing
  rapidly during calibrations
• Accuracy was not affected by using calibration
  values obtained either pre-prandially or about 2
  hours post-prandially
• A separate calibration algorithm for overnight
  readings might improve accuracy at night
• - The median bias during the day was 4 mg/dL,
  and overnight the bias was 9 mg/dL
• - When we limited calibration values to those
  only obtained overnight, this bias was resolved
  (2 mg/dL)

12
(No Transcript)