Glycated haemoglobin, diabetes, and mortality in men in Norfolk cohort of European Prospective Investigation of Cancer and Nutrition (EPIC-Norfolk)
BMJ 2001; 322 doi: https://doi.org/10.1136/bmj.322.7277.15 (Published 06 January 2001) Cite this as: BMJ 2001;322:15
All rapid responses
Rapid responses are electronic comments to the editor. They enable our users to debate issues raised in articles published on bmj.com. A rapid response is first posted online. If you need the URL (web address) of an individual response, simply click on the response headline and copy the URL from the browser window. A proportion of responses will, after editing, be published online and in the print journal as letters, which are indexed in PubMed. Rapid responses are not indexed in PubMed and they are not journal articles. The BMJ reserves the right to remove responses which are being wilfully misrepresented as published articles or when it is brought to our attention that a response spreads misinformation.
From March 2022, the word limit for rapid responses will be 600 words not including references and author details. We will no longer post responses that exceed this limit.
The word limit for letters selected from posted responses remains 300 words.
EDITOR: The study by Khaw et al1 have elegantly demonstrated that glycated
haemoglobin (HbA1C) is strongly associated with increased mortality from
cardiovascular disease in men with diabetes mellitus. Even more
interestingly, they have demonstrated that HbA1C is also strongly
associated with excess cardiovascular disease related mortality in men
without diabetes mellitus.
It is now widely accepted that the risk factors that underlie the
development of cardiovascular disease are identical to those that
predispose to vasculogenic erectile dysfunction2. As such, a major risk
factor for both atherogenesis and erectile dysfunction is diabetes
mellitus3. To this effect, it is estimated that the prevalence of erectile
dysfunction in diabetic men is as high as 50%4.
Tissues are continually exposed to damaging reactive oxygen species5. In
diabetes mellitus, this phenomenon is greatly increased6. Hence, there is
evidence that diabetes mellitus can lead to vascular endothelial
dysfunction7. This may play a role in the excess cardiovascular mortality
noted in diabetics as we have recently demonstrated, in an animal model of
diabetes mellitus, that oxidative stress via reactive oxygen species is
associated with impaired cavernosal smooth muscle relaxation8. As,
cardiovascular and cavernosal smooth muscle share very similar properties,
it is reasonable to hypothesize that factors affecting the cavernosa may
also be affecting the cardiovascular system.
Recently, there has been evidence emerging to support the view that HbA1C
is an independent predictor of diabetic erectile dysfunction. As such,
erectile dysfunction, in diabetic men, has been shown to correlate
strongly with the level of HbA1C9 . However, to date, there has been no
investigation to determine whether HbA1C is associated with erectile
dysfunction in men without diabetes mellitus. As Khaw et al1 have now
demonstrated that HbA1C is associated with cardiovascular disease related
mortality in men without diabetes, it is worthwhile investigating whether
it is also associated with erectile dysfunction in men without diabetes
mellitus.
Masood Khan
Specialist Registrar in Urology
Dept. of Urology, St. Bart's Hospital, London E1, UK.
Dimitri Mikhailidis
Reader and Honorary Consultant
Dept. of Chemical Pathology, Royal Free Hospital, London NW3, UK.
Robert Morgan
Consultant Urological Surgeon
Dept. of Urology, Royal Free Hospital, London NW3, UK.
1. Khaw K-T, Wareman N, Luben R, Bingham S, Oakes S, Welch A, et al.
Glycated haemoglobin, diabetes, and mortality in men in Norfolk cohort of
European Prospective Investigation of Cancer and Nutrition (EPIC-Norfolk).
BMJ 2001; 322:15-18.
2. Azadzoi KM, De Tejada IS. Hypercholesterolemia impairs endothelium -
dependent relaxation of rabbit corpus cavernosum smooth muscle. J Urol
1991; 146:238-40.
3. De Tejada IS, Goldstein I, Azadzoi K, Krane RJ, Cohen RA. Impaired
neurogenic and endothelium - mediated relaxation of penile smooth muscle
from diabetic men with impotence. N Engl J Med 1989; 320:1025-30.
4. McCulloch DK, Campbell IW, Fu FC, Prescott RJ, Clarke BF. The
prevalence of diabetic impotence. Diabetologia 1980; 18:279-83.
5. Halliwell BH, Gutteridge JMC. Free Radicals in Biology and Medicine.
Oxford: Oxford University Press, 1989.
6. Honing MLH, Morrison PJ, Banga JD, Stroes ESK, Rabelink TJ. Nitric
oxide availability in diabetes mellitus. Diabetes Matab Rev 1998; 14:241-
49.
7. Cohen RA. Dysfunction of vascular endothelium in diabetes mellitus.
Circulation 1993; 87: V67.
8. Khan MA, Thompson CS, Jeremy JY, Mumtaz FH, Mikhailidis DP, Morgan RJ.
The effect of superoxide dismutase on nitric oxide-mediated and electrical
field stimulated diabetic rabbit cavernosal smooth muscle relaxation. Br J
Urol Int 2001; 87:98-103.
9. Romeo JH, Seftel AD, Madhun ZT, Aron DC. Sexual function in men with
diabetes type 2: association with glycemic control. J Urol 2000; 163:788-
91.
Competing interests: No competing interests
Analytical information is required to enable generalisation of data
The data presented by Khaw et al,(1) suggesting that the relationship
between cardiovascular disease and glycaemia is a continuum extending
throughout the non-diabetic population is potentially fascinating.
Unfortunately, as the paper stands, the results are not generalisable
since neither HbA1c methodology nor calibration data have been included.
Professional organisations throughout the UK(2) are in agreement that
percentage HbA1c should be harmonized in relation to a common standard and
that, in the absence of a primary calibrant, this standard should be that
employed in the Diabetes Control and Complications Trial (DCCT).(3) Without
this information, the results of the present study cannot be compared
against others and it is inappropriate to discuss "a threshold commonly
accepted for diagnosis of diabetes."
For example, with an ion exchange (HA-8140, Menarini Diagnostics)
method widely used in Europe and the UK, the upper limit of the reference
range observed in non-diabetic subjects has been reported as 5.l%,(4)
equivalent to 5.9% following DCCT standardisation. If either of these cut-
off values were applicable to the present study, a significantly large
(although vastly differing) number of their subjects could not be
considered to have 'normal' glycaemic control. This could be explicable in
terms of the age range of their cohort (45-79 years) and the known
association of increased age and decreased glucose tolerance.(5) To suggest
that a large percentage of their cohort may have 'abnormal' glycaemic
control would not be surprising: Harris et al, (5) for example, have shown
27.1% and 42.9% of the male US population aged 55-64 and 65-74 years
respectively to have either impaired glucose tolerance or diabetes. The
present data could therefore be explained on the basis that a large number
of patients with impaired glucose tolerance were included in the upper two
-thirds of their 'non-diabetic' population and that these have an
increased rate of cardiovascular disease: this of course would not be a
new observation.
The authors should be encouraged to provide at least a non-diabetic
reference range for their assay and preferably some data relating their
results to a DCCT aligned method. This study can then be compared to other
work in this area.
Finally, it is stated that "HbA1c may provide a practical screening
tool for diabetes or impaired glucose tolerance." Although this could be
true, no evidence is provided to support this statement. The World Health
Organisation has categorically stated that HbA1c should not be used to
establish the diagnosis of diabetes and this has been re-iterated by
Diabetes UK.
1. Khaw KT, Wareham N, Luben R, Bingham S, Oakes S, Welch A, Day N.
Glycated haemoglobin, diabetes, and mortality in men in Norfolk cohort of
European Prospective Investigation of Cancer and Nutrition (EPIC-Norfolk).
BMJ 2001; 322: 15-18.
2. Marshall SM, Barth JH. Standardization of HbA1c measurements - a
consensus statement. Diabetic Medicine 2000; 17: 5-6.
3. 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: 977-86.
4. John WG, Braconnier F, Miedema K, Aulesa C, Piras G. Evaluation of
the Menarini-Arkray HA 8140 hemoglobin A1c analyzer. Clin Chem 1997; 43:
968-975.
5. Harris MI, Hadden WC, Knowler WC, Bennett PH. Prevalence of
diabetes and impaired glucose tolerance and plasma glucose levels in US
population aged 20-74 yr. Diabetes 1987; 36: 523-534.
Dr Edmund Lamb,
Consultant Clinical Biochemist,
East Kent Hospitals NHS Trust,
Kent and Canterbury Hospitals,
Ethelbert Road,
Canterbury,
Kent
e-mail: edmund.lamb@kch-tr.sthames.nhs.uk
Competing interests: No competing interests
Editor - The paper of Prof Khaw's and colleagues' very important
study on the relationship between percentage of glycated haemoglobin and
mortality in diabetic and non-diabetic men [1] mentions at best by way of
intimation a very likely cause of elevated HbA1c: psychosocial stress.
Activation of both the corticoid system from the hypothalamic-pituitary-
adrenal axis and the epinephrine system from the sympathetic-adrenal
medullary axis increases blood glucose.
Environmental circumstances e.g. such as job strain or belonging to
indigenous populations, stressed by westernised living conditions, were
associated with higher proportions of glycated haemoglobin [2, 3]. Longer
examination periods increased significantly HbA1c percentage in healthy
medical students; several months after the stressful examinations the
values were significantly lower [4].
Elevated extra- and intra-cellular glucose concentrations result in what
biochemists call "oxidative stress", defined as imbalance between
prooxidants and antioxidants [5]. Glucose autoxidation, formation of
usually irreversible, so-called "advanced glycation endproducts" (ACEs),
and the polyol pathway, seem to be involved in the genesis of this
oxidative stress, which is reported both in experimental diabetes in
animals as well as in type 1 and type 2 diabetic patients [5].
Many secondary diseases of diabetes are nowadays explained by these
mechanisms. HbA1c is one of these ACEs; and the observed positive
correlation between glycated haemoglobin concentration and mortality below
the diabetic threshold [1] might perhaps be explained by the same
mechanisms as in diabetics.
Peter Schuck, MD, PhD
Egerstr. 19,
D-08258 Markneukirchen, Germany.
References:
1 Khaw KT, Wareham N, Luben R, Bingham S, Oakes S, Welch A, et al.
Glycated haemoglobin, diabetes, and mortality in men in Norfolk cohort of
European Prospective Investigation of Cancer and Nutrition (EPIC-Norfolk).
BMJ 2001;322:1-6.
2 Netterstrøm B, Kristensen TS, Damsgaard MT, Olsen O, Sjol A. Job
strain and cardiovascular risk factors: a cross sectional study of
employed Danish men and women. Br J Ind Med 1991;48:684-9.
3 Daniel M, O'Dea K, Rowley KG, McDermott R, Kelly S. Glycated
hemoglobin as an indicator of social environmental stress among indigenous
versus westernized populations. Prev Med 1999;29:405-13.
4 Schuck P. Glycated hemoglobin as a physiological measure of stress
and its relation to some psychological stress indicators. Behav Med
1998;24:89-94.
5 Bonnefont-Russelot D, Bastard JP, Jaudon MC, Delattre J.
Consequences of the diabetic status on the oxidant/antioxidant balance.
Diabetes Metab 2000;26:163-76.
Competing interests: No competing interests
Correction to article BMJ 2001;322:1-6.
I am grateful to Dr. Henry Kahn, CDC Atlanta for pointing out an
error in Table 3 of the above paper. For the Cox multivariate regression
for ischaemic heart disease, model 3 including diabetes and HbA1C in the
same model, the relative risk (95% CI) for Diabetes history should
be 2.00 (95% CI 0.67 - 5.96) P=0.21. The figure of 2.69 was incorrect.
Kay-Tee Khaw
Competing interests: No competing interests
After years of observing diabetics and their test results, I found
that on numerous occasions, patients with poor control (as measured by
their home glucose meters)produced remarkably low A1c results. For
example, one patient with an average meter reading of 170 had an A1c of
5.6. This leads me to question the interpretation of the A1c.
One possibility is that the A1c is related to blood sugar levels, but
is partially independent of them. It also is reasonable to postulate other
mechanisms accounting for A1c variations, either alone or in concert with
blood sugar. We can accept that glycated hemoglobin is a destructive
biological entity (not only for diabetics) and look for other ways of
controlling it besides tight sugar control.
It would be just another error in the history of diabetes treatment
if we were satisfied with A1c as the benchmark of care, only to discover
its shortcomings after the current generation of patients took false
refuge in its comfort. Lkewise, some with poor A1c's may have relatively
good glucose control. This latter group will suffer from the lack of
research into alternative methods of lowering their glycated hemoglobin
levels.
Competing interests: No competing interests
Has there been any striation with different centrifuge speeds?
Regardzs
D
Competing interests: No competing interests
Dear Sir,
Khaw and colleagues (1) convincingly demonstrate an association
between high HbA1C levels and mortality in 4,702 men aged 45 to 79.
However, their claim that 25% of the "excess deaths" could potentially be
avoided by a reduction in the population mean HbA1C level of 0.2% is
premature, since the authors have not demonstrated causation. That is to
say, they have not demonstrated that, in patients without diabetes,
whatever it is that causes increased levels of HbA1C also causes increased
risk of premature death. Even if these factors do share a common genesis,
it is therapeutic optimism in the extreme to believe that the detrimental
effects of a lifetime's glycaemia can be reduced by a quarter - within 4
years, based on the duration of their follow up - by changing the HBA1C.
The failure to discriminate cross sectional (distribution of HbA1C
levels and mortality across a population) from longitudinal (what happens
to an individual when their HbA1C level changes) data is reminiscent of
the early enthusiasm for aggressive cholesterol reduction in the primary
prevention of heart disease (2), and will I am sure engender a similar
enthusiastic echo from the pharmaceutical industry.
Both impaired glucose tolerance and NIDDM (3,4) are associated with
low birth weight; this association is apparent by the age of 7 (5) and is
seen in monozygotic twins (6). Low birth weight is also associated with
increased cardiovascular mortality (7) and with maternal socioeconomic
deprivation (as determined by the Carstairs Deprivation Category of
maternal postcode (8)). Furthermore, interventions to ease maternal
deprivation have been shown in a randomized, controlled study (9) to have
a beneficial effect on birth weight in those most at risk.
Perhaps a small proportion of the resource allocated to the
pharmacological treatment of cardiovascular risk could instead be used to
tackle the social circumstances that appear to lie at the heart of that
increased risk. In the UK, starting Child Benefit payments from the time
of ante-natal booking rather than the time of delivery would reduce
poverty during pregnancy, would encourage early contact with ante-natal
services - and would cost about the same as a year's supply of lipid
lowering therapy.
References
1. Khaw K-T, Wareham N, Luben et al (2001) BMJ 322, 1-6.
2. Sudlow C and Macleod M (1994) BMJ 308, 1039-40.
3. Levitt NS, Lambert EV, Woods D et al (2000) J Clin Endocrinol Metab 85,
4611-8.
4. Carlsson S, Persson PG, Alvarsson M et al (1999) Diabetes Care 22, 1043
-7.
5. Law CM, Gordon GS, Shiell AW et al (1995) Diabet Med 12, 24-9.
6. Poulsen P, Vaag AA, Kyvik KO et al (1997) Diabetologia 40, 439-46.
7. Barker DJP, Osmond C, Winter PD et al (1989) Lancet 2, 577-80.
8. Information and Statistics Department, Scottish Health Service
(personal communication)
9. Kehrer BH and Wolin CM (1979) J Human Resources 14, 434-462.
Competing interests: No competing interests
Have the authors or others studied the serum Insulin levels in the
"normal" pts with elevated HgA1C? Have the normals been otherwise tested
for Insulin resistance (e.g. with an Insulin challenge). Could these be
underlying factors for the elevated HgA1C and its implied risk?
Competing interests: No competing interests
In reporting lipid values, most laboratories now reference an optimal
range, rather than a statistical reference range for normal in the
population. Clearly with lipids, normal is not equal to optimal. With
electrolytes normal is pretty much equal to optimal. Similarly blood
pressure has optimal ranges that are not equal to statistical normal
ranges. Is it time now to stop reporting a reference range for HgA1C
measurements, and instead report an optimal range?
John S. Muchmore MD, PhD
Oklahoma City, OK, USA
muchjs@integris-health.com.us
http://www.drmuchmore.com.us
Competing interests: No competing interests
Glycated haemoglobin, diabetes, and explanations
Dear Editor,
The study by Khaw et al. demonstrated that glycated haemoglobin
(HbA1C) is associated with increased mortality in men without diabetes
mellitus.1 The discussion was centered around the perspective of
populational intervention, and explanations of the correlation between
HbA1C and mortality. These accents are possibly misleading.
The result of this study is another illustration of the general rule
for any continuous variable (e.g. blood cholesterol concentration): 1) for
subjects with characteristics lying outside the 'normal range' the
incidence of new acute and chronic conditions and mortality is elevated;
2) for subjects with characteristics inside the 'normal range' the
incidence of diseases and mortality is not equal; in most cases the
minimum of diseases' incidence and mortality is not close to median or
mean values.2 It is shown for many variables, from height and weight,
blood red blood cells count, to HbA1C now. It is true not only for
population samples, but for very selective healthy subjects - military
pilots - for whom it was shown in relation to blood albumin, RBC etc.3 It
may be shown for any variable, if to study it carefully enough. For
selected pathology(cause of death) the shape of the curve, describing
incidence VS variable, is special, e.g. curve of cancer mortality VS blood
cholesterol is different from curve of cardiovascular mortality VS
cholesterol.4 Every curve is explainable, if to try hard enough.
Last 50 years medicine is drifting from the use of reference range to
use of diagnostic criteria. For continuous variables, such as HbA1C or
blood glucose concentration(BGC), it means that procedure of setting the
diagnostic threshold is based not on the distribution of the values in
'healthy subjects' but on the trade between underdiagnosed sick
people(false negatives) and overdiagnosed healthy (false positives).5
Current diagnostic criteria trade over probability of complication of
diabetes. Also there is a tendency to expand intervention into the non-
diabetes range of BGC (so-called impaired glucose tolerance) and now we
see the proposal to involve all population in the interventions. The
similar 'threshold' is used for diagnosis of obesity, and interventions
are expanded to the non-obesity range of body weight (so-called
overweight), and all population is now under pressure to reduce body
weight. Authors of 1 are wrong stating that 'it is uncertain whether the
relation between blood glucose concentration and ... diseases has a
threshold or is a continuum'. Threshold exists only for simplicity of
diagnostic decision making. Blurred border between healthy subjects and
subjects who are 'sick' or 'under risk' makes explanation hard, if
possible, but open prospects for speculations because loss of specificity.
Medical practice is historically limited to more or less definite
groups of suffering people, whom we call 'sick' or 'diseased'. The
opposing tendency is a shadow of medicine - it is a medicalization of
traits, habits and risks. Frequently expansion of the intervention to the
majority of the population is based on the understanding that only small
proportion of 'subjects at risk' are in the 'diagnostic' group. It is a
simple consequence of the low effectiveness of the 'diagnostic' criteria.
Indiscriminate interventions (reduce weight of all people, reduce BGC of
all people etc.) may lead to the desired effect, but usually they are of
low efficiency. We can find better application of our limited funds, e.g.
pour them into the educational programs or support of the poor. Anyway,
the efficacy of any action must be proved first. Even if some one likes
the idea to isolate children from cakes and chocolates, or want to sell
the BGC-reducing drug-for-all.
References
1. Khaw KT, Wareham N, Luben R, Bingham S, Oakes S, Welch A et al.
Glycated haemoglobin, diabetes, and mortality in men in Norfolk cohort of
European Prospective Investigation of Cancer and Nutrition (EPIC-Norfolk).
Brit.Med.J. 2001;322:15.
2. Vlassov VV. Reaction of the organism to external stimuli: General
patterns of reaction's
development and related methodological problems of research [Russian].
Irkutsk: Irkutsk University, 1994: 343 p.
3. Vlassov VV. Age changes of the haemoglobin and erythrocyte
concentrations in pilots [Russian]. Aviakosmich Ecologich Med 1992; #4:19-
23.
4. Schuit AJ, Van Dijk CEMJ, Dekker JM, Schouten EG, Kok FJ. Inverse
association between serum total cholesterol and cancer mortality in Dutch
civil servants. Am.J.Epidemiol. 1993;137:966-76.
5. Remein QR,.Wilkerson HL. The efficiency of screening tests for
diabetes. J.Chron.Dis. 1961;13:6-21.
Competing interests: No competing interests