Download Relationship between mean blood glucose and glycated

dme_2379.fm Page 174 Tuesday, February 12, 2008 11:12 AM
DIABETICMedicine
DOI: 10.1111/j.1464-5491.2007.02379.x
Original Article: Clinical Care and Delivery
Blackwell Publishing Ltd
Relationship between mean blood glucose and glycated
haemoglobin in Type 2 diabetic patients
Original article
K. Makris, L. Spanou*, A. Rambaouni-Antoneli†, K. Koniari*, I. Drakopoulos, D. Rizos‡ and
A. Haliassos§
Clinical Biochemistry Department, KAT General Hospital, Kifissia, Greece, *Internal Medicine Department, KAT General Hospital, Kifissia, Greece, †Diabetes Care
Centre, IKA-Halandri, Athens, Greece, ‡‘Aretaieion’ University Hospital, Medical School, University of Athens, Greece and §Greek External Quality Assessment
Scheme (GEQAS), Athens, Greece
Accepted 23 August 2007
Abstract
Aims To correlate the values of MBG to HbA1c in Greek patients with Type 2 diabetes and/or metabolic syndrome.
Methods We followed up 140 Greek adult patients: 92 patients with Type 2 diabetes treated with insulin or oral
glucose-lowering medication, and 48 patients with newly diagnosed Type 2 diabetes or metabolic syndrome not receiving
any treatment. MBG was calculated for each patient from self-measurements of blood glucose using a portable glucometer,
made six times a day (before eating and 2 h after a meal), three times a week for 1 month. HbA1c was determined by HPLC
at 0 and 12 weeks.
HbA1c at 0 (x) and 12 weeks (y) correlated strongly (y = 0.790x + 1.115, r = 0.92), confirming that the patient’s
glycaemic status remained stable during the whole period of follow-up. Linear regression was performed on MBG values;
HbA1c at 12 weeks, sex, age, body mass index (BMI) and patient status (Type 2 diabetes treated or not) were used as
independent variables. None of the independent variables reached statistical significance in the model, with the exception
of HbA1c at 12 weeks. The final model was: MBG (mg/dl) = (34.74 × HbA1c) – 79.21, r = 0.93; or MBG (mmol/l)
= 1.91 × HbA1c – 4.36, r = 0.93.
Results
Conclusions Our results establish for the first time a strong correlation between MBG and HbA1c in Type 2 diabetic
patients and support the idea of expressing HbA1c results as MBG. This will help patients to gain a clearer interpretation
of the result, with less confusion. This simplification will allow every person with diabetes using home glucose-monitoring
to understand his or her own target level.
Diabet. Med. 25, 174–178 (2008)
Keywords
HbA1c, mean blood glucose, Type 2 diabetes
Abbreviations DCCT, Diabetes Control and Complications Trial; EDTA, ethylenediaminetetraacetic acid; HPLC,
high-performance liquid chromatography; IFCC, International Federation of Clinical Chemistry; MBG, mean blood
glucose; NGSP, National Glycohemoglobin Standardization Program; SMBG, self-monitoring of blood-glucose levels;
UKPDS, UK Prospective Diabetes Study
Introduction
Monitoring of glycaemic status is considered a cornerstone of
diabetes care. It is used to assess the efficiency of therapy and
to guide adjustments in lifestyle and/or therapy in order to
achieve the best possible blood-glucose control. Blood-glucose
testing by patients and measurement of glycated haemoglobin
by healthcare providers are of major importance.
Correspondence to: K. Makris, Clinical Biochemistry Department, KAT General
Hospital, 2 Nikis Street, 14651, Kifissia, Greece. E-mail: [email protected]
174
SMBG provides real-time feedback of blood-glucose levels
and is useful for the day-to-day management of diabetes [1,2].
It is generally accepted as an integral part of self-management
in insulin-treated diabetic patients, although its use in
non-insulin-treated patients remains controversial [3 –5].
Measurement of glycated haemoglobin, and especially
HbA1c (which is its major constituent), is widely used in
patients with diabetes as a monitor of long-term glycaemic
control [1,2]. Two clinical trials (the DCCT and the UKPDS)
have demonstrated a relationship between HbA1c levels and
diabetic complications in patients with Type 1 and Type 2
© 2008 The Authors.
Journal compilation © 2008 Diabetes UK. Diabetic Medicine, 25, 174–178
dme_2379.fm Page 175 Tuesday, February 12, 2008 11:12 AM
Original article
diabetes, respectively [6,7]. The lack of standardization of
HbA1c measuring methods and the variation in reference
ranges and results has led many national organizations to
develop standardization programmes in order to reduce
inter-laboratory variability and to harmonize GHb results.
The IFCC working group on HbA1c standardization prepared
a primary reference material of pure HbA1c and HbA0 and
developed a reference method for the measurement of HbA1c
[8–10]. The link between the IFCC reference method values
and the US NGSP values is provided by the regression equation
NGSP – HbA1c = 0.915 × (IFCC – HbA1c) + 2.15%. The IFCC
method produces significantly lower results than the NGSP
method (1.3–1.9%). These findings created considerable
debate as to how HbA1c should be reported. A working group
established in 2004 on behalf of ADA/EASD/IDF to harmonize
HbA1c reporting globally [11] suggested that HbA1c results be
expressed in MBG units in clinical use, provided that prospective
studies in various populations prove that the relationship
between HbA1c and MBG is strong enough to warrant this
[12–14]. The largest set of data relating plasma glucose to
HbA1c comes from the DCCT trial. Results showed a linear
relationship between HbA 1c and MBG expressed by the
formula MBG (in mmol/l) = (1.98 × HbA1c) – 4.29, r = 0.82
and when HbA1c is traceable to the DCCT reference. These
data were derived from the retrospective examination of
seven-point capillary blood-glucose profiles obtained from
Type 1 diabetic patients [14]. In this study, we examine the
relationship between MBG and HbA1c in Greek patients with
Type 2 diabetes or metabolic syndrome.
Subjects and methods
The study was approved by the hospital ethics committee
and all patients gave written, informed consent.
One hundred and forty Greek adults (84 men, 56 women)
were followed up for a period of 30 days. Serial blood-glucose
measurements were made. The mean age of our population
was 61.9 years (range 41–81), 63 years (range 41–81) for men
and 61 years (range 44 –80) for women. Two groups of patients
were enrolled in our study: 92 patients (44 men, 48 women)
with Type 2 diabetes treated with insulin or oral glucose-lowering
medication (group A), and 48 patients (40 men, eight women)
with newly diagnosed Type 2 diabetes and/or metabolic
syndrome, who did not receive any glucose-lowering medication
during the follow-up period (group B).
Body mass index (BMI) was calculated as the ratio of body
weight (in kg) to the square of the height (in m2) and expressed
in kg/m2. Mean (± SD) BMI was 28.9 ± 3.9 kg/m2 for men and
29.1 ± 5.0 kg/m2 for women.
MBG for each patient was derived from six daily measurements
(before eating and 2 h after one of the three main daily meals),
three times a week for 1 month (30 days). MBG for each patient
was calculated from at least 72 measurements. Self-performed
measurements of capillary whole blood glucose were performed
using a portable glucometer (Ascensia-Cotnour; Bayer HealthCare, UK). The measurement of glucose is based on the
measurement of electrical current caused by the reaction of
© 2008 The Authors.
Journal compilation © 2008 Diabetes UK. Diabetic Medicine, 25, 174–178
DIABETICMedicine
blood glucose with the reagents on the electrode on the strip.
The Ascencia contour instrument has been calibrated by the
manufacturer to produce results expressed in mg/dl or in mmol/l
that are equivalent to plasma/serum glucose measurements.
Each participant was given a portable glucometer and a
notebook, in which he or she was asked to record all measurements according to our instructions. At the end of the follow-up
period, the data from the glucometer were downloaded to the
program WinGlucofacts Professional (Bayer HealthCare,
Tarrytown, NY, USA), where the MBG was calculated. MBG
values were also verified from the recorded measurements in
the notebook.
HbA1c was measured by an automated analyser (Menarini–
Akray HA8160; Menarini, Florence, Italy) using the HPLC
technique, calibrated using standards traceable to NGSP.
HbA1c was measured twice in each patient, at the beginning
of the follow-up period (0 weeks) and 2 months after the
completion of the self-monitoring period (12 weeks from the
beginning of the follow-up period). All samples were collected
using Vaccutainer® (Becton-Dickinson, New Jersey, USA)
tubes with EDTA anticoagulant.
Statistical analysis was performed using the SPSS software
(SPSS Inc., Chicago, IL, USA). A probability level ≤ 0.05 was
considered significant.
Results
MBG ( ± SD) in group A patients (9.1 ± 2.5 mmol/l) was
significantly higher (P < 0.05, t-test) than in group B patients
(7.0 ± 0.8 mmol/l). MBG values are presented in Table 1
according to sex, age and BMI in the two groups of patients.
Mean MBG was similar in men and women in both patient
groups and did not vary with age. However, MBG increased
with increasing BMI in Group A while MBG decreased with
increasing BMI in group B (Table 1).
HbA1c values at 0 and 12 weeks were strongly correlated
(y = 0.790x + 1.115, r = 0.92) proving that the glycaemic
status of our patient population remained relatively stable
during the follow-up period.
In order to determine predictors of MBG, linear regression
was performed on MBG values using HbA1c at 12 weeks, sex,
age, BMI and patient group as independent variables. Only
HbA1c at 12 weeks was included in the final model:
MBG (mg/dl) = (34.74 × HbA1c) – 79.21, r = 0.93
(95% confidence interval [CI] for slope: 32.47–37.02)
or
MBG (mmol/l) = (1.91 × HbA1c) – 4.36, r = 0.93
(95% CI for slope: 1.79 –2.04)
As estimated from the equation, the mean increase of MBG
per 1% increase in HbA1c is 1.9 mmol/l (35.2 mg/dl), which is
almost similar to that observed in the DCCT study of Type 1
diabetic patients (2.0 mmol/l or 35.6 mg/dl).
A scatter-plot of MBG values vs. HbA1c at 12 weeks,
together with the regression line, its 95% CI and the 95%
prediction interval, are presented in Fig. 1. The 95% prediction
175
dme_2379.fm Page 176 Tuesday, February 12, 2008 11:12 AM
DIABETICMedicine
HbA1c and mean blood glucose • K. Makris et al.
Table 1 MBG in groups A and B
Group A (n = 92)
Group B (n = 48)
MBG, mmol/l (mg/dl)
Sex
Men (n = 84)
Women (n = 56)
t-test
Age (years)
≤ 50
51–60
61–70
≥ 71
ANOVA
BMI (kg/m2)
Normal weight (BMI < 25)
Overweight (25 ≤ BMI ≤ 30)
Obese (BMI > 30)
ANOVA
Total (n = 140)
ANOVA
8.4 ± 2.2 (152.4 ± 40.6)
9.7 ± 2.5 (177.0 ± 45.1)
NS
7.0 ± 0.9 (126.8 ± 16.2)
6.9 ± 2.5 (124.8 ± 5.8)
NS
9.8 ± 2.5 (177.7 ± 51.3)
9.7 ± 3.1 (176.0 ± 56.1)
8.3 ± 2.0 (150.6 ± 36.2)
9.3 ± 1.9 (168.4 ± 33.7)
NS
7.1 ± 0.9 (129.0 ± 15.4)
7.0 ± 0.9 (126.4 ± 16.7)
7.0 ± 0.8 (127.6 ± 15.2)
6.6 ± 0.4 (120.3 ± 7.2)
NS
7.4 ± 1.2 (133.8 ± 22.6)
8.7 ± 2.0 (158.6 ± 35.8)
10.9 ± 2.6 (198.0 ± 47.4)
P < 0.001
9.1 ± 2.5 (165.2 ± 44.5)
7.5 ± 0.7 (135.5 ± 13.3)
7.1 ± 0.7 (129.2 ± 13.3)
6.7 ± 0.9 (122.3 ± 15.9)
NS
7.0 ± 0.8 (126.4 ± 15.0)*
= analysis of variance, NS = non-significant. *P < 0.005 (t-test).
FIGURE 2 Regression lines and equations calculated in our study
(total population, group A, group B) and the DCCT study.
FIGURE 1 Scatter-plot of MBG values vs. HbA1c at 12 weeks
(open circles), the regression line (solid line) the 95% confidence interval
of the regression line (inner dotted lines) and the 95% prediction interval
(outer lines).
interval for a patient’s MBG in our study is ±1.37 mmol/l, at
HbA1c range of 5.1–10.9%. In Fig. 2, the regression lines and
the equations calculated for the total population as well as for
groups A and B of our study and for the DCCT study are
presented. The slope of the regression line in group A (1.94,
95% CI: 1.78 –2.11) approximates the slope of the regression
line in the combined group (1.91, 95% CI: 1.79–2.04), but is
significantly greater (P < 0.001) than that in group B (1.25,
95% CI: 0.99–1.51).
176
Discussion
For more than 25 years, the HbA1c test has been the most
widely accepted outcome measure for assessing glycaemic
control in individuals with diabetes mellitus [2,15]. The
test provides an index of a patient’s average blood-glucose
level during the past 60–90 days and is therefore an independent
parameter of carbohydrate metabolism. It is considered
as the most objective and reliable marker of long-term
glycaemic control. Reproducibility (i.e. long-term comparability
of the values within one individual patient) is therefore an
absolute necessity. At present, more than 20 different HbA1c
methods are in use, based on three different assay principles
(cation exchange chromatography, affinity chromatography
© 2008 The Authors.
Journal compilation © 2008 Diabetes UK. Diabetic Medicine, 25, 174–178
dme_2379.fm Page 177 Tuesday, February 12, 2008 11:12 AM
Original article
and immuno-turbidimetry). Harmonization, comparability
and standardization of HbA1c results have therefore become
an issue. Local initiatives have generally addressed this [16],
although the NGSP programme has been highly successful in
harmonizing HbA1c methods worldwide to the results of the
DCCT and UKPDS, and has demonstrated that results from
methods that utilize different assay principles can in fact be
harmonized to produce results that are equivalent to those of
the DCCT [17].
The recent development of a reference method by IFCC
solved the standardization problem and allowed routine
methods to be traceable to a true accuracy base but created a
great controversy over the way the results should be reported
from the clinical laboratories. The high specificity of the
reference method results in lower HbA 1c values in blood
samples, because the non-specific components falsely identified
as HbA1c in routine methods are not measured by the reference
method. Such change implies that the reported HbA1c would
be 1–2% less than that currently reported. One problem with
such a change in the reporting of HbA1c results is confusion on
the part of patients, which could result in the deterioration of
glycaemic control. This was confirmed in a recent study [18]
that showed that there is a considerable risk of deterioration in
glycaemic control when reporting results to patients on a lower
scale (as would happen if the new IFCC number scale were used
to report HbA1c results to patients). However, in the same study
a positive effect on glycaemic control was observed when the
HbA1c reference level was adjusted to higher levels (DCCT).
Another problem is that nearly every guideline for diabetes has
based its intervention and target levels on the standard of the
DCCT and the UKPDS. The question is how to report standardized IFCC results. Some advocate converting IFCC values
to DCCT values and others propose expressing HbA1c as an
average blood-glucose equivalent. One option of the IFCC
working group is to express HbA1c in mmol HbA1c per mol
total Hb; this would result in ten times higher values; thus the
reference range would be 29–43 mmol HbA1c per mol Hb [19].
To resolve these controversies, a working group was formed
by ADA/EASD/IDF to review the opportunities arising from
the development of a new IFCC reference method for the
measurement of HbA1c, and to make recommendations on its
implementation [11]. The conclusion was that there is now a
great opportunity to redefine the entire assay into something
that reflects the mean blood glucose. The relationship between
plasma glucose and HbA1c is complex. Several previous studies
have tried to analyse this relationship, and several investigators
have correlated HbA1c with blood-glucose measurements at
various times within a day. They have concluded that HbA1c
is an index of mean plasma glucose (calculated from all these
measurements) over the previous weeks to months [13,14,20].
Single plasma glucose measurements as an indication of
long-term glycaemia should be used with caution: previous
studies have shown that either they tend to underestimate
HbA1c (at increasing plasma glucose levels) or their contribution
is variable (post-meal glucose levels) [14,21,22].
© 2008 The Authors.
Journal compilation © 2008 Diabetes UK. Diabetic Medicine, 25, 174–178
DIABETICMedicine
This relationship between MBG and HbA1c was documented
by Rohlfing et al., based on retrospective examination of sevenpoint glucose assays from Type 1 diabetic patients during the
DCCT study. If this relationship can also be verified in a prospective study and in other types of diabetic patients, then it
might be possible to report the new IFCC figures as MBG. Our
results confirm the following points. Firstly, there is a close
relationship between HbA1c and mean blood glucose of the
previous 2–3 months in Type 2 diabetic patients. Secondly, this
relationship can be described by a linear regression equation that
is very close to the one described by the retrospective analysis of
the Type 1 diabetic patients. Thirdly, this relationship is
not affected by type of diabetes, type of treatment or the sex
or age of the patient.
However, there are drawbacks when multiple observations
per patient are used in order to calculate a pathological quantity,
such as improper meter use, laboratory errors, pathological
conditions that alter the normal red cell lifespan or variant
haemoglobins that can interfere with HbA1c. Also, discrepancies
between calculated MBG and measured HbA 1c can arise.
Although many studies have shown that intra-individual
biological variation of HbA1c is minimal, there is evidence of
wide fluctuations in HbA1c between individuals that are
unrelated to glycaemic status, suggesting the existence of high
and low glycators [14,23 –25]. High glycators have consistently
higher HbA1c than expected for their MBG, whereas low
glycators have lower HbA1c than their MBG would suggest
[14,23 –25]. The proposed reasons for this between-individual
variability in haemoglobin glycation rate include differences in
erythrocyte survival and other genetic elements [26,27].
In our study, the observed strong correlation between the
calculated MBG and the measured HbA1c suggests that bloodglucose levels measured over a period of time by patients
themselves can provide a reasonably accurate estimation of
HbA1c. This finding is in agreement with previous results
derived from Type 1 diabetic patients [14].
However, there are some limitations that must be mentioned.
Firstly, although a correlation coefficient of 0.93 is statistically
impressive, the 95% CIs are wide. For example, an HbA1c of
6.5% corresponds to an MBG of 8.0 mmol/l, with a prediction
interval of 6.7–9.4 mmol/l. This rather wide dispersion must
be taken into account when discussing the expression of
HbA 1c in MBG units.
A second point of concern is the accuracy of portable
glucose meters because we have to depend on such instruments
for the calculation of MBG. Although these instruments claim
a wide measurable range, linearity at the extremes of this range
can be questioned. In the case of the two patients with MBG
of 18.4 mmol/l and 17.9 mmol/l (the two highest points in
Fig. 1), several individual measurements were at the upper
end of the instrument’s measuring range and possibly were
overestimated by the instrument. This is also supported by the
fact that these patients had measured HbA1c of 10.9% and
10.3%, respectively, which correspond to much lower MBG
(16.5 mmol/l and 15.3 mmol/l, respectively).
177
dme_2379.fm Page 178 Tuesday, February 12, 2008 11:12 AM
DIABETICMedicine
A final point is the observed difference in the regression
line and correlation coefficient between group A (patients
receiving treatment) and group B (a heterogeneous group of
newly diagnosed patients and/or patients with metabolic
syndrome) (Fig. 2). This difference could be attributed to
the limited range of HbA 1c values in group B (5.1–6.9%)
compared to group A (5.9–10.9%).
In summary, the relationship between blood glucose and
HbA1c can be described by a single linear regression equation
in both Type 1 and Type 2 diabetic patients. This relationship
only applies when HbA1c is measured by methods that are
certified by the NGSP as traceable to the DCCT reference
method. Because the relationship between IFCC reference
method and the DCCT reference is already defined [10], this
equation can be converted to apply to IFCC reference traceable methods. Using this approach to calculate MBG can allow
laboratories to report, along with the measured HbA1c, the
corresponding calculated MBG, which is independent from
the calibration and the measuring system and reference system
traceability. This may give healthcare professionals and
patients the ability to set day-to-day blood-glucose targets and
understand glycaemic status in a more conceivable way.
Competing interests
None to declare.
References
1 IDF Clinical Guidelines Task Force. Global guideline for Type 2
diabetes. Brussels: International Diabetes Federation, 2005.
2 ADA Position Statement. Standards of Medical Care in Diabetes.
Diabetes Care 2007; 30 (Suppl 1): 4–41.
3 Welschen LMC, Bloemendal E, Nijpels G, Dekker JN, Heine RJ,
Stalman WAB et al. Self-monitoring of blood glucose in patients who
are not using insulin: a systematic review. Diabetes Care 2005; 28:
1510–1517.
4 Ipp E, Aquino RL, Christenson P. Point: self-monitoring of blood
glucose in Type 2 diabetic patients not receiving insulin: the sanguine
approach. Diabetes Care 2005; 28: 1528–1530.
5 Davidson MB. Counterpoint: self-monitoring of blood glucose in
Type 2 diabetic patients not receiving insulin: a waste of money.
Diabetes Care 2005; 28: 1531–1533.
6 The Diabetes Control and Complications Trial Research Group. The
effect of intensive treatment of diabetes on the development and
progression of long term complications in insulin-dependent diabetes
mellitus. N Engl J Med 1993; 329: 977–986.
7 UK Prospective Diabetes Study (UKPDS) Group. Intensive bloodglucose control with sulphonylureas or insulin compared with
conventional treatment and risk of complications in patients with
Type 2 diabetes (UKPDS 33). Lancet 1998; 352: 837–853.
8 Finke A, Kobold U, Hoelzel W, Weykamp C, Miedema K, Jeppsson
JO. Preparation of a candidate primary reference material for the
international standardization of HbA1c determinations. Clin Chem
Lab Med 1998; 36: 299–308.
178
HbA1c and mean blood glucose • K. Makris et al.
9 Jeppsson JO, Kobold U, Barr J, Finke A, Hoelzel W, Hoshino T et al.
Approved IFCC reference method for measurement of HbA1c in
human blood. Clin Chem Lab Med 2002; 40: 78–89.
10 Hoelzel W, Weykamp C, Jeppsson JO, Miedma K, Barr JR, Goodall
I et al. IFCC reference system for measurement of hemoglobin A1c in
human blood and the national standardization schemes in the United
States, Japan and Sweden: a method comparison study. Clin Chem
2004; 50: 166–174.
11 ADA/EASD/IDF. Report of the ADA/EASD/IDF working group of
the HbA1c assay. Diabetologia 2004; 47: R53–54.
12 Svendson PA, Lauritzen T, Soegaard U, Nrup J. Glycosylated haemoglobin and steady-state mean blood glucose concentration in Type 1
(insulin-dependent) diabetes. Diabetologia 1982; 23: 403–405.
13 Nathan DM, Singer DE, Hurxthal K, Goodson JG. The clinical
information value of the glycosylated hemoglobin assay. N Eng J
Med 1984; 310: 341–346.
14 Rohlfing CL, Wiedmeyer H-M, Little RR, England JD, Tennill A,
Goldstein DE. Defining the relationship between plasma glucose
and HbA1c: analysis of glucose profiles and HbA1c in the Diabetes
Control and Complications Trial. Diabetes Care 2002; 25: 275–278.
15 Sacks DB, Bruns DE, Goldstein DE, Maclaren NK, McDonald JM,
Parrott M. Guidelines and recommendations for laboratory analysis
in the diagnosis and management of diabetes mellitus. Clin Chem
2002; 48: 436–472.
16 Midema K. Towards worldwide standardization of HbA1c determination. Diabetologia 2004; 47: 1143–1148.
17 Little R. Glycated Hemoglobin Standardization—National Glycohemoglobin Standardization Program (NGSP) Perspective. Clin
Chem Lab Med 2003; 41: 1191–1198.
18 Hanas R. Psychological impact of changing the scale of reported HbA1c
results affects metabolic control. Diabetes Care 2002; 25: 2110–2111.
19 Nordin G, Dydkær-International Federation of Clinical Chemistry
and Laboratory Medicine—Scientific Division. Recommendation
for term and measurement unit for ‘HbA1c’. Clin Chem Lab Med
2007; 45: 1081–1082.
20 Svendson PA, Lauritzen T, Soegaard U, Nerup J. Glycosylated haemoglobin and steady-state mean blood glucose concentration in Type 1
(insulin-dependent) diabetes. Diabetologia 1982; 23: 403–405.
21 Avignon A, Radauceanu A, Monnier L. Nonfasting plasma glucose
is a better marker of diabetic control than fasting plasma glucose in
Type 2 diabetes. Diabetes Care 1997; 20: 1822–1826.
22 Bonora E, Calcaterra F, Lombardi S, Bonfante N, Formentini G,
Bonadonna R et al. Plasma glucose levels throughout the day and
HbA1c interrelationships in Type 2 diabetes: implications for treatment and monitoring of metabolic control. Diabetes Care 2001; 24:
2023–2029.
23 Yudkin JS, Forrest RD, Jackson CA, Ryle AJ, Davie S, Gould BJ.
Unexplained variability of glycated hemoglobin in non-diabetic subjects not related to glycemia. Diabetologia 1990; 33: 208–215.
24 Kilpatrick ES, Maylor PW, Keevil BG. Biological variation of glycated
hemoglobin: implications for diabetes screening and monitoring.
Diabetes Care 1998; 21: 261–264.
25 Kilpatrick ES, Rigby AS, Atkin SL. Variability in the relationship
between mean plasma glycose and HbA1c: implications for the assessment of glycaemic control. Clin Chem 2007; 53: 897–901.
26 Virtue MA, Furne JK, Nuttall FQ, Levitt MD. Relationship between
GHb concentration and erythrocyte survival determined from breath
carbon monoxide concentration. Diabetes Care 2004; 27: 931–935.
27 Snieder H, Sawtell PA, Ross L, Walker J, Spector TD, Leslie RD.
HbA1c levels are genetically determined even in Type 1 diabetes: evidence
from healthy and diabetic twins. Diabetes 2001; 50: 2858–2863.
© 2008 The Authors.
Journal compilation © 2008 Diabetes UK. Diabetic Medicine, 25, 174–178