Alarms based on real-time sensor glucose values alert patients to

Alarms based on real-time sensor glucose values
alert patients to hypo- and hyperglycemia: the
Guardian Continuous Monitoring System
Background information
•
•
•
•
Continuous glucose monitoring (CGM) allows glucose levels to be measured and
recorded 24 hours a day. This technology enables patients to have an unprecedented
level of control over their glucose levels, creating the potential for long-term
improvements in clinical outcomes.
The Guardian Continuous Monitoring SystemTM (‘the Guardian’) is a first-generation
device that can alert the wearer in real-time when glucose levels fall below, or exceed,
programmed threshold values.
This device is distinct from the more recent Guardian® RT Continuous Glucose
Monitoring System, which displays real-time glucose values in addition to having alarms.
It is anticipated that real-time alarms will facilitate the management of
hypo- and hyperglycemic events, and prevent the development of serious or
life-threatening episodes.
Aims
•
•
To determine whether use of real-time alarms is associated with reductions in
hypo- or hyperglycemic events.
To assess the accuracy of the Guardian, and to assess its sensitivity and specificity in
detecting hypo- and hyperglycemia.
Methods
•
•
•
Duration of hypoglycemic episodes
(minutes per event)
•
Patients aged > 18 years with type 1 diabetes were randomized to one of two groups.
In the first group (the alert group), patients wore the Guardian for two periods of up
to 6 days each. During the first period, the alerts were turned off; during the second
period, the alerts were turned on.
In the second group (the control group), patients wore the Guardian but had the
alerts switched off for both study periods.
The lower and upper glucose thresholds were set at 70 mg/dL (3.8 mmol/L) and
250 mg/dL (13.8 mmol/L), respectively.
Blood glucose (BG) levels were measured 4 times per day (before meals and at
bedtime) using the Accu-Chek® BG meter (Roche Diagnostics). Additional
measurements were made in response to the monitor’s alarm, and whenever
symptoms of hypo- or hyperglycemia were experienced.
80
p = 0.004
p = 0.03
Figure 1. Duration of
hypoglycemic episodes
(< 65 mg/dL or 3.6 mmol/L)
70
60
69.6
64.4
63.8
50
40
Period 1
30
33.6
Period 2
20
10
0
Alerts
OFF
Alerts
ON
Alert group (n = 32)
Alerts
OFF
Alerts
OFF
Control group (n = 36)
For more information, speak to your Medtronic Diabetes representative.
©2006 Medtronic Diabetes. All rights reserved. UC200700039 EE
Medtronic International Trading Sàrl, Route du Molliau 31, Case Postale 84, CH-1131 Tolochenaz, Switzerland.
Tel. +41 21 802 7000 Fax +41 21 802 7900 www.medtronic-diabetes.com
Key point
■
Use of real-time
alarms produced a
48% reduction in
the mean duration
of hypoglycemic
events (p = 0.004).
•
•
•
Absolute relative error (the overall agreement between the paired Guardian and BG
meter readings) was calculated and used as a measure of accuracy.
Sensitivity (defined as the ability of the device to correctly identify true hypo- or
hyperglycemia) and specificity (defined as the ability of the device to correctly
identify the absence of these states) were measured at different sensor glucose levels.
The impact of real-time alarms on the frequency and duration of hypoglycemic
(defined as glucose levels < 65 mg/dL [3.6 mmol/L]) and hyperglycemic
(> 275 mg/dL [15.1 mmol/L]) episodes was assessed by comparing data from the
two periods in both groups.
Results
•
•
•
•
Seventy-one patients (23 male; mean age 44.0 ± 11.4 years) were enrolled; 35 were
randomized to the alert group and 36 to the control group.
4453 paired Guardian and BG meter values were analyzed. The median absolute
relative error was 17.3%.
There was good agreement between the values: the correlation co-efficient was 0.87.
The main results of the sensitivity and specificity analyses are shown in Table 1.
Sensitivity and specificity were found to be optimal at respective lower and upper
glucose thresholds of 80 mg/dL (4.4 mmol/L; sensitivity 83%, specificity 86%) and
190 mg/dL (10.4 mmol/L; sensitivity 88%, specificity 87%).
Alert threshold
Sensitivity (%)
Specificity (%)
False alert rate (%)
< 70 mg/dL
67
90
47
> 250 mg/dL
63
97
19
Table 1. Sensitivity and specificity of the Guardian
•
•
In the alert group, the duration of hypoglycemic episodes was significantly reduced,
by 27 minutes, during the ‘alarm on’ period compared with the ‘alarm off’ period
(Figure 1). No significant change was seen in the control group.
Use of the alarm did not significantly affect the duration of hyperglycemic episodes,
although there was a trend towards a higher frequency of hyperglycemic episodes
during the ‘alarm on’ period. This trend may have been caused by over-treatment
of hypoglycemia.
Conclusions
• Real-time alarms enable patients to respond quickly to
hypoglycemia, reducing the duration of hypoglycemic episodes
and preventing the development of more serious events.
• The Guardian provides reasonably accurate CGM data.
From the authors
‘‘
‘‘
The Guardian will have a significant
role in managing excessive
daily variation in glycemic control
Authors
Bode B, Gross K,
Rikalo N, et al.
Reference
Diabetes Technol Ther
2004;6:105–13
Type of study
Randomized,
controlled trial
Parameters assessed
Accuracy, sensitivity,
specificity, duration of
hypo- and hyperglycemic
episodes