Comparison of self-monitor blood glucose with meter with the laboratory standard measurement of blood glucose

StudyPopulationInterventionOutcomesResultsCommentsDesignEL
Trajanoski et al, 199633415 patients during hypoglycaemic– hyperinsulinaemic clamps (6 patients with type 1 diabetes and 9 healthy volunteers)
Austria
663 blood glucose monitor readings and 119 reference values ranging from 2.28 to 3.89 mmol/l
Analysed by six blood glucose monitors, compared with laboratory reference method (Beckman Glucose Analyzer 2):
One Touch II®
Companion II®
Reflolux®
Accutrend®
Elite®
HemoCue®
  1. Comparison of correlation between laboratory standard and self-monitor blood glucose meter
  2. % of readings 10% and 20% within reference range
  3. % of readings 40% outside reference range
  1. Correlation coefficients:
    One Touch II: 0.91
    Companion II: 0.81
    Reflolux: 0.78
    Accutrend: 0.88
    Elite: 0.78
    HemoCue: 0.93
  2. % of readings 10% and 20% within reference range (10% then 20%):
    One Touch II: 78.0%, 99.2%
    Companion II: 58.0%, 88.2%
    Reflolux: 34.5%, 85.0%
    Accutrend: 14.5%, 46.0%
    Elite: 39.5%, 75.6%
    HemoCue: 80.5, 96.6%
  3. % of readings 40% outside reference range:
    One Touch II: 0%
    Companion II: 2.5%
    Reflolux: 0.9%
    Accutrend: 6.6%
    Elite: 4.2%
    HemoCue: 0%
Test evaluation studyIIb
Rayman et al, 1984335180 blood samples
UK
Analysis of blood glucose monitor Reflocheck® compared with laboratory reference method, by ‘skilled operator’Comparison of correlation between laboratory standard and self-monitor blood glucose meterCorrelation coefficient: r = 0.996Test evaluation studyIIb
Lehmann et al, 200133655 patients with type 1 diabetes and 40 patients with type 2 diabetes
760 blood glucose monitor readings ranging from 2.6 to 20.0 mmol/l
Aged 19–79 years
Switzerland
Analysis of blood glucose monitor (MediSense Precision QID®) compared with laboratory reference method
Blood glucose monitor measurement was performed by patient and by nurse
Comparison of correlation between laboratory standard and self-monitor blood glucose meter
Patient preference
Correlation coefficient: r = 0.97, p < 0.0001 (n = 95)
Of blood glucose meter readings < 3.9 mmol/l (but > 2.6 mmol/l), 64% were within ± 10% and 100% were within ± 20% of the reference reading
The glucose measuring system and test strips received an overall rating of good to very good (although this was not compared with any other glucose monitoring systems)
Test evaluation studyIIb
Laus et al, 198433712 patients with type 1 diabetes
28 blood glucose monitor readings
USA
Analysis of blood glucose monitors (Glucoscan I®) compared with laboratory reference method
Blood glucose monitor measurement was performed by nurse
Comparison of correlation between laboratory standard and self-monitor blood glucose meter
Number of acceptable values < 15% deviation from reference
Glucoscan I: correlation coefficient: r = 0.847 (n = 28)
10/28 values were considered unacceptable (> 15% deviation from reference, maximum deviation seen 55%)
Test evaluation studyIIb
Laus et al, 19843373 blood samples of different glucose concentrations, tested 9 different timesAnalysis of blood glucose monitors (Glucoscan I® and Glucoscan II®) looking at variations in test procedure:
From: 1) completely coat the reaction pad with blood, 2) wait exactly 60 seconds, 3) wash for 2 seconds, 4) blot reaction pad firmly, 5) inset test strip into meter and read
For Glucoscan I the procedure was varied in several ways: 1) sample volume varied from 5 μl to 50 μl (but so pad is completely covered at all times), 2) reaction time varied from 40 to 70 seconds, using 5 μl sample, 3) wash time varied from 1 to 4 seconds, 4) blotting pressure and technique were interpreted three different ways
For Glucoscan II the procedure was varied in several ways: 1) sample volume varied from 5 μl to 50 μl (but so pad is completely covered at all times), 2) timing of blotting varied from 35 to 45 seconds, using 5 μl sample, 3) blotting pressure and technique were interpreted in two different ways, 4) total reaction time or reading time varied from 60 to 80 seconds
Variation of the different methods and significant differencesGlucoscan I:
  1. Increasing sample volume from 5 to 50 μl led blood glucose measurements to vary from 38.0 ± 6.5 to 86.0 ± 1.5 mg/dl, p < 0.05. However, there was no difference in blood glucose measurements at sample volumes of 10, 20, 30, 40 or 50 μl, p > 0.05
  2. Varying reaction time from 40 to 70 seconds led blood glucose measurements to vary from 60.3 ± 2.4 to 90.7 ± 4.4 mg/dl, p < 0.05
  3. Varying washing time from 1 to 4 seconds led blood glucose measurements to vary from 88.3 ± 0.9 to 62.7 ± 1.2 mg/dl, p < 0.05
  4. Varying blotting pressure and technique in three different ways led blood glucose measurements to vary from 85.3 ± 3.5 (gentle) to 64.7 ± 1.2 mg/dl (extra firm), p < 0.05
Glucoscan II:
  1. Increasing sample volume from 5 to 50 μl led blood glucose measurements to vary from 91 to 97 mg/dl, p > 0.05
  2. Varying reaction times from 35 to 45 seconds led blood glucose measurements to vary from 89.0 ± 1.0 to 96.0 ± 3.5 mg/dl, p > 0.05
  3. Varying blotting pressure and technique in three different ways led blood glucose measurements to vary from 67.5 ± 1.7 (gentle) to 69.5 ± 2.8 mg/dl (firm), p > 0.05
  4. Varying delay in reading led blood glucose measurements to vary from 94.0 ± 4.6 to 97.7 ± 4.2 mg/dl, p > 0.05
Test evaluation studyIIb
Nelson et al, 198333841 adults with diabetes
101 blood glucose monitor readings
Canada
Analysis of blood glucose monitors (Glucometer®, Glucoscan® and Hypocount B®) compared with laboratory reference method venous serum glucose assayed by the hexokinase method and the capillary whole blood glucose determined by the glucose oxidase method (Yellow Springs Instrument)
Blood glucose monitor measurement was performed by experienced nurse
Comparison of correlation between laboratory standard (hexokinase method) and self-monitor blood glucose meter
Sensitivity and specificity
Correlation coefficients:
Glucometer: r2= 0.92
Glucoscan: r2= 0.86
Hypocount B: r2= 0.92
Glucose oxidase method: r2= 0.95
No significant difference between them
Sensitivity and specificity:
Glucometer: 95.9%, 82.4%
Glucoscan: 87.8%, 90.2%
Hypocount B: 89.8%, 96.1%
Accuracy of meters within 10% of the laboratory hexokinase method < 100 mg/dl:
Glucometer: 45.5%
Glucoscan: 54.6%
Hypocount B: 18.2%
Accuracy of meters within 10% of the laboratory hexokinase method > 180 mg/dl:
Glucometer: 86.1%
Glucoscan: 36.1%
Hypocount B: 69.4%
Test evaluation studyIIb
Kolopp et al, 198333930 blood glucose monitor readings for each monitor
France
Analysis of blood glucose monitors (Hypocount® Glucocheck®, Glucometer® Dextrometer® and Eyetone®) compared with laboratory reference method (Beckman BGA II®)
Two monitors were tested for each blood glucose monitor type, except for Eyetone where only one monitor was tested
Coefficients of variation for 0–3.82, 3.83–8.38 and ≥ 8.39 mmol/l, and for all blood glucose concentrationsIn all meters there was a statistically significant correlation with the reference measure (p < 0.001)
Coefficients of variation for 0–3.82 mmol/l (for each meter):
Hypocount: 10.8, 7.1
Glucocheck: 30.3, 11.2
Glucometer: 8.6, 5.1
Dextrometer: 8.8, 12.6
Eyetone: 17.5
Beckman BGA II: 2.4
Coefficients of variation for 3.83–8.38 mmol/l:
Hypocount: 7.9, 8.4
Glucocheck: 23.9, 15.2
Glucometer: 5.6, 7.5
Dextrometer: 11.5, 11.3
Eyetone: 11.1
Beckman BGA II: 1.7
Coefficients of variation for ≥ 8.39 mmol/l:
Hypocount: 6.2, 6.8
Glucocheck: 14.1, 12.9
Glucometer: 7.2, 5.8
Dextrometer: 8.5, 8.0
Eyetone: 7.1
Beckman BGA II: 1.3
Coefficients of variation for all blood glucose concentrations:
Hypocount: 8.3, 7.4
Glucocheck: 22.8, 13.1
Glucometer: 7.1, 6.1
Dextrometer: 9.6, 10.6
Eyetone: 11.9
Beckman BGA II: 1.8
Test evaluation studyIIb
Gifford-Jorgensen et al, 198634037 adult and paediatric patients with diabetes
USA
Analysis of blood glucose monitors (Accu-Chek®, Glucochek II®, Glucochek Dextro®, Glucometer® and Glucoscan II®) compared with laboratory reference method (Beckman Astra 8®)
Blood samples tested 8 times on each meter within 30 min of blood being drawn
Coefficients of variation blood glucose 60–180, 181–300 and 301–400 mg/dlCoefficients of variation for 60–180 mg/dl:
Accu-Chek: 3.31
Glucochek II: 1.38
Glucochek Dextro: 3.85
Glucometer: 4.09
Glucoscan II: 5.47
Coefficients of variation for 181–300 mg/dl:
Accu-Chek: 2.84
Glucochek II: 2.84
Glucochek Dextro: 1.65
Glucometer: 3.13
Glucoscan II: 6.41
Coefficients of variation for 301–400 mg/dl:
Accu-Chek: 2.61
Glucochek II: 4.71
Glucochek Dextro: 1.81
Glucometer: 4.09
Glucoscan II: 7.84
Test evaluation studyIIb
Kyvik et al, 199034150 patients with insulin-treated diabetes
Aged 18–79 years
Denmark
  1. 3 blood samples of approximately 4, 10 and 20 mmol/l measured with 6 ExacTech® meters compared with laboratory standard
  2. blood sample from each patient analysed with ExacTech® meters compared with laboratory standard, tested by study staff
  3. blood sample from each patient analysed with ExacTech® meters compared with laboratory standard, tested by patient
Coefficient of variation
  1. Coefficient of variation: 10.28%, p > 0.05 (n = 63)
  2. Coefficient of variation: 4.45%, p > 0.05 (n = 45)
  3. Coefficient of variation: 4.46%, p > 0.05 (n = 60)
Test evaluation studyIIb
Merino-Torres et al, 200334225 adults with type 1 diabetes
Aged 17 to 34 years
Spain
62 measurements of HbA1c
Patients monitored for 3–9 months
Linear regression calculated between HbA1c and mean blood glucose for the 60 days before HbA1c samplingr = 0.71, r2= 0.497, p = 0.000Test evaluation studyIIb

From: Evidence tables

Cover of Type 1 Diabetes
Type 1 Diabetes: Diagnosis and Management of Type 1 Diabetes in Children and Young People.
NICE Clinical Guidelines, No. 15.2.
National Collaborating Centre for Women's and Children's Health (UK).
London: RCOG Press; 2004 Sep.
Copyright © 2004, National Collaborating Centre for Women’s and Children’s Health.

Apart from any fair dealing for the purposes of research or private study, criticism or review, as permitted under the Copyright, Designs and Patents Act, 1988, no part of this publication may be reproduced, stored or transmitted in any form or by any means, without the prior written permission of the publisher or, in the case of reprographic reproduction, in accordance with the terms of licences issued by the Copyright Licensing Agency in the UK. Enquiries concerning reproduction outside the terms stated here should be sent to the publisher at the UK address printed on this page.

The use of registered names, trademarks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant laws and regulations and therefore for general use.

The rights of National Collaborating Centre for Women’s and Children’s Health to be identified as Author of this work have been asserted by them in accordance with the Copyright, Designs and Patents Act, 1988.

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.