Reference values of skin autofluorescence as an estimation of tissue accumulation of advanced glycation end products in a general Slovak population

Prévia do material em texto

Short Report: Complications
Reference values of skin autofluorescence as an
estimation of tissue accumulation of advanced glycation
end products in a general Slovak population
K. Simon Klenovics1, R. Kolla´rova´2, J. Hodosy2, P. Celec2 and K. Sˇebekova´2
1Institute of Physiology, Comenius University Medical Faculty, Bratislava, Slovakia and 2Institute of Molecular Biomedicine, Comenius University Medical Faculty,
Bratislava, Slovakia
Accepted 19 September 2013
Abstract
Aims For decades, Slovakia has maintained a prominent place in mortality rates from cardiovascular diseases among
European Union (EU-27) countries. Determination of skin autofluorescence serves as an estimate of tissue accumulation
of advanced glycation end products—substances accumulating in tissues and body fluids that play a pathophysiological
role in age-related diseases and their complications, such as diabetes.
Methods In 1385 apparently healthy Slovakian subjects aged from a few days old to 77 years, skin autofluorescence
was determined using an advanced glycation end product reader and compared with reference data from Dutch
Caucasians. The impact of the weekly frequency of recreational physical exercise on skin autofluorescence was
investigated in the adults, and the impact of feeding regimen in the infants.
Results With the exception of 10- to 19-year-olds, Slovaks had lower skin autofluorescence values in comparison with
the Dutch Caucasians. In healthy non-smokers, physical exercise for > 30 min/day performed ≥ 3 times/week was
associated with lower skin autofluorescence levels. In infants, breastfeeding (advanced glycation end product-poor diet)
was associated with lower skin autofluorescence levels in comparison with consumption of infant formulas (advanced
glycation end product-rich diet).
Conclusions Reference ranges of skin autofluorescence in Slovak Caucasians, detailed for paediatric age groups, are
provided. Our data show that, in healthy adults, regular physical exercise associates with lower skin autofluorescence.
Infants fed or weaned from infant formulas (advanced glycation end product-rich diet) have higher skin autofluorescence
than their breast milk-consuming counterparts. It is unclear why Slovaks have lower skin autofluorescence compared
with a Dutch population with lower cardiovascular mortality rates. Reference data on skin autofluorescence from
diverse populations are needed for the precise clinical interpretation of skin autofluorescence measurements.
Diabet. Med. 31, 581–585 (2014)
Introduction
Advanced glycation end products are formed via the reaction
of free amino groups of proteins with reactive carbonyls.
Hyperglycaemia and oxidative and carbonyl stress accelerate
their formation [1] and decreased renal function [2],
smoking [3] and an advanced glycation end product-rich
diet [4,5] enhance accumulation of these products. Physical
exercise reduces levels of advanced glycation end products
[6–8].
Advanced glycation end products play a pathophysiolog-
ical role in the development of age-related diseases and
their complications [9,10], via modification of protein
structure and function, and an interaction with a specific
cell surface receptor for advanced glycation end products
inducing detrimental responses via nuclear factor kappa B
(NF-jB) and other pathways [11].
Intrinsic fluorescence of some advanced glycation end
products enables non-invasive estimation of advanced gly-
cation end product accumulation in tissues as skin autoflu-
orescence. Increased skin autofluorescence is indicative of
widespread atherosclerosis [12]. It associates with cardio-
vascular and renal risk factors and is a predictor of overall
and/or cardiovascular mortality [13–16]. Skin autofluores-
cence reference values have been published for Dutch
Caucasians and Chinese people [17,18].Correspondence to: Katarı´na Sˇebekova´. E-mail: kata.sebekova@gmail.com
ª 2013 The Authors.
Diabetic Medicine ª 2013 Diabetes UK 581
DIABETICMedicine
DOI: 10.1111/dme.12326
According to Eurostat 2010 data, Slovakia holds the
fifth-highest position in the standardized death rate from
cardiovascular disease and the second-highest from ischaemic
heart disease. In comparison with the Netherlands, which has
the third-lowest mortality among the 27 countries that were
members of the European Union (EU-27) at the time the
data were compiled, in Slovakia, the standardized death rate
from the above-mentioned diseases was 3- and 6.5-fold
higher, respectively (http://epp.eurostat.ec.europa.eu/statistics_
explained/index.php/Causes_of_death_statistics).
Weaskedwhether: (1) the general Slovakpopulationdisplays
higher skin autofluorescence, or its steeper age-dependent rise,
in comparison with the healthier Dutch Caucasians; (2) the
frequency of physical exercise associates with skin autofluores-
cence; (3) in infants, skin autofluorescence reflects the die-
tary advanced glycation end product-load, as breastfeeding in
comparisonwith infant formula-feeding represents a long-term
humanmodel of advanced glycation end product-poor vs. -rich
diet [4,19].
Subjects and methods
Skin autofluorescence
Skin autofluorescence was measured on intact healthy skin on
the volar side of the lower arm or, in infants, on the buttocks,
using the advanced glycated end product (AGE) Reader
(DiagnOptics Technologies BV, Groningen, the Netherlands),
and was calculated as the ratio of mean fluorescence
intensities detected at 420–600 and 300–420 nm [17].
Subjects with skin phototypes V and VI, and those who had
recently applied skin oils/ointments on the site of measure-
ment were excluded from the evaluation [20].
Subjects
In the summer of 2008, skin autofluorescence was measured
in 1573 Caucasians aged from a few days old to 77 years
(62.0% female), recruited from among students, employees
and participants of epidemiological studies at the Comenius
University and the Slovak Medical University, visitors of
scientific exposition, and in cooperation with paediatricians.
Retrospective analysis of anonymous data was approved by
the institutional review board in 2012.
Age, gender, current smoking status and presence of
age-related diseases (hypertension, diabetes, hyperlipopro-
teinemia, cardiovascular and chronic renal disease, cancer)
were recorded.
In 226 healthy non-smoker adults, who declared that they
did not exercise or were they recreational sportspersons, BMI
and self-reported weekly frequency of exercise longer than
30 min were recorded. Subjects were classified as not
exercising, exercising 1–2 times/week and exercising ≥ 3
times/week.
In the ≤ 12-month-old infants, gestational age and current
feeding regimen were recorded. Infants ≤ 6 months old were
classified as fully breastfed, fully infant formula-fed or
mixed-fed; those of age between 6 and 12 months old were
classified as weaned from breast milk or from infant formula.
Statistical analyses
Data are given as mean � SD. Groups were compared using
unpaired two-sided Student’s t-test or Mann–Whitney U-test
or analysis of variance (ANOVA) with post-hoc Scheffe’s
test, as appropriate. The general linear model was employed
to study the impact of independent variables on skin
autofluorescence levels. Pearson correlation coefficients were
calculated. A P-value of < 0.05 was considered significant.
SPSS version 16 for Windows (SPSS Inc., Chicago, IL, USA)
was used.
Results
After the exclusion of 188 subjects, skin autofluorescence
was evaluated in 1385 healthy subjects (62% female).
Children and adolescents
The mean skin autofluorescence in paediatric subjects is given
in Table 1. In multivariate analysis of ≤ 12-month-olds, both
age and feeding regimen independentlydetermined skin
autofluorescence (P < 0.001 for both); gender, gestational
age and their interaction had no significant impact (corrected
model P < 0.001; R2 = 37%). Infants ≤ 6 months old and
What’s new?
• Slovak adults, presenting one of the highest rates of
cardiovascular mortality within European Union
(EU-27) countries, display lower skin autofluorescence
(a widely used estimate of cardiovascular risk in
high-risk patient groups) than age-matched Dutch
adults, who have the third-lowest cardiovascular mor-
tality among EU-27 countries.
• In adults, regular physical exercise is associated with
lower skin autofluorescence.
• Formula-fed (advanced glycation end product-rich diet)
infants have higher skin autofluorescence than their
breast milk-fed counterparts.
• Data on reference ranges of skin autofluorescence in
diverse populations and knowledge of different (life-
style) factors affecting it could facilitate the interpre-
tation of data on skin autofluorescence in clinical
practice
582
ª 2013 The Authors.
Diabetic Medicine ª 2013 Diabetes UK
DIABETICMedicine Skin autofluorescence in Slovaks: effects of exercise and diet � K. Simon Klenovics et al.
those between 6 and 12 months old who consumed breast
milk had lower skin autofluorescence than their infant
formula-fed counterparts.
Whole cohort
With the exception of 10- to 19-year-old Slovaks, mean skin
autofluorescence was lower in all age groups when compared
with the Dutch subjects [17] (Fig. 1). Accordingly, the slope
and constant in an equation describing skin autofluorescence
in Slovaks (0.022 9 age + 0.76, R2 = 65%, P < 0.001)
were lower than in the Dutch subjects (0.024 9 age + 0.83,
R2 = 60%, P < 0.001). In our study, 20- to 29-year-old and
30- to 39-year-old smokers had significantly lower skin
autofluorescence when compared with the Dutch subjects
(P < 0.009 and P < 0.05, respectively) [17].
In other respects, our data showed similar associations to
those reported for the Dutch and Chinese subjects [17,18]:
female gender independently determined skin autofluores-
cence in multivariate analysis (see also Supporting Informa-
tion, Fig. S1); smokers displayed higher skin autofluorescence
and a steeper age-dependent rise than non-smokers (see also
Supporting Information, Fig. S2) (although no interaction
between gender and smoking being present in the Dutch study
was confirmed); and healthy subjects displayed lower skin
autofluorescence than those presenting age-related diseases
(1.53 � 0.37 vs. 1.94 � 0.48 arbitrary units, respectively;
P < 0.001). If corrected for age, gender and smoking; age
(P < 0.001), presence of disease (P = 0.042) and interaction
of smoking with disease (P = 0.010) were independent signif-
icant contributors (overall model P < 0.001, R2 = 52%).
Effect of exercise
Skin autofluorescence was lower in subjects exercising ≥ 3
times/week (1.24 � 0.25 arbitrary units) than in those exer-
cising 1–2 times/week (1.36 � 0.22 arbitrary units,P < 0.05)
and in subjects who did not exercise (1.38 � 0.29 arbitrary
units, P < 0.01). Age (P < 0.001) and exercise frequency
(P = 0.005), but not BMI, gender or their interaction, were
significantly contributing independent variables (overall
model P = 0.001,R2 = 45%). Group characteristics are given
in the Supporting Information (Table S1).
Table 1 Skin autofluorescence values in paediatric age categories
n
Skin
autofluorescence P*
Infants (< 12 months) 86 0.63 � 0.19
≤ 6 months 63 0.58 � 0.18 < 0.001†
Mother’s milk†† 29 0.52 � 0.16 < 0.001‡
Mother’s milk +
infant formula††
12 0.57 � 0.12 NS§
Infant formula†† 22 0.67 � 0.20 0.004¶;
0.023**
> 6–12 months 23 0.76 � 0.14
Mother’s milk†† 12 0.69 � 0.11
Infant formula†† 11 0.82 � 0.15 0.028
Toddlers (1–3 years) 40 0.80 � 0.14 < 0.001
Pre-school children
(4–5 years)
17 0.88 � 0.14 NS
Younger school age
(6–10 years)
20 0.93 � 0.19 NS
Older school age
v(11–15 years)
40 0.97 � 0.22 NS
Adolescents
(16–19 years)
111 1.16 � 0.22 < 0.001
*P denotes significance vs. the lower age category, if not
indicated differently, i.e. † vs. all infants aged > 6–12 months;
‡ vs. > 6- to 12-month-old mother’s milk-fed infants; § vs.
mother’s milk- and infant formula-fed groups of the same age
category; ¶ vs. mother’s milk-fed infants of the same age
category; ** vs. infant formula-fed > 6–12 months old. ††Feed-
ing regimen categories within the age group.
NS, not significant.
4
3
2
1
0
0–9
AR
F 
(a
rb
itr
ar
y 
un
its
)
Age group (years) 
10–19 20–29 30–39 40–49 50–59 60–69 70–79
n = 27/8n = 45/39n = 64/132n = 72/199n = 86/214n = 62/479n = 50/154n = 45/160
P < 0.001
P < 0.001
P < 0.001
P < 0.001
P < 0.001
P < 0.001 P < 0.009
P < 0.001
P < 0.001 P < 0.006
P < 0.002
P < 0.003
P < 0.013
FIGURE 1 Mean skin autofluorescence (ARF) per age group in the Dutch and the Slovak population. Mean � sd of skin autofluorescence per age
group as published for the Caucasian population in the Dutch study [17] (——); mean � SD of skin autofluorescence per age group in the present
(Slovak) study (- - - -). n, number of subjects in the Dutch/Slovak studies, respectively, in the corresponding age groups. Significance indicated above
the solid line: Dutch vs. Slovak study, corresponding age groups; significance indicated below the dashed line: p pertinent to 2 indicated age groups
(both 2-sided Student’s t-test).
ª 2013 The Authors.
Diabetic Medicine ª 2013 Diabetes UK 583
Research article DIABETICMedicine
Discussion
We present here the first reference data on skin autofluores-
cence for the paediatric and general Slovak population.
Infants consuming infant formulas (rich in advanced glyca-
tion end products) had higher skin autofluorescence than
their breastfed counterparts. Surprisingly, the general Slovak
population characterized by a high cardiovascular mortality
[21] displayed lower skin autofluorescence when compared
with the healthier Dutch Caucasians [17]. In healthy
subjects, regular recreational physical activity was associated
with lower skin autofluorescence. Being cross-sectional in
nature, our study allows us to comment only on associations
between skin autofluorescence and other factors.
Adults
With the exceptionof10- to19-year-olds, theSlovaksdisplayed
significantly lower skin autofluorescence when compared with
the Dutch subjects [17]. Lower levels in ≤ 9-year-olds may be
because our cohort had a higher number of infants aged
≤ 5 years in comparisonwith theDutch study [17].The reasons
behind lower skin autofluorescence in the adults remain
unclear. It cannot be attributed to gender disparity—two-thirds
of our participants were women, who also display higher
plasma advanced glycation end product levels thanmen [22]—
or to a lower proportion of smokers in comparison with the
Dutch study (11 vs. 22%, respectively), as our smokers
displayed skin autofluorescence similar to or lower than the
age-matched Dutch subjects (smokers and non-smokers) [17].
In rats, long-termconsumptionof anadvanced glycation end
product-richdiet leads to an accumulationof theproduct in the
skin [23]. In adult humans, no significant relationship between
dietary intake of advanced glycation end products and skin
autofluorescence has been found, suggesting that, in healthy
adults, the natural fluctuation in consumedadvanced glycation
end products is insufficient to impact on skin autofluorescence
significantly [24]. Skin autofluorescence inversely correlated
with alcohol consumption [24], which is high in Slovakia [25],
but we did not record alcohol intake in our study.
The lower skin autofluorescence in subjects exercising
≥ 3 times/week is similarto observations on the amelioration
of circulating and/or tissue advanced glycation end products
following an intervention with physical training [6–8] and
confirms the assumption that physical activity must be
performed regularly to show association with skin autoflu-
orescence [26]. Several mechanisms may account for this. An
efficient utilization of reactive intermediates of the glycolytic
pathway for energy production may decrease the pool
available for glycation [6]. Exercise-induced stimulation of
endogenous antioxidant enzyme activities may in turn
decrease formation of advanced glycation end products via
oxidative pathways [8]. Physical activity increases the
expression and production of heat shock proteins involved
in the degradation of misfolded and possibly glycated
proteins [27,28]. The reported frequency of exercising was
lower and inactivity higher in the Slovaks than in the Dutch
subjects [25], therefore the lower autofluorescence levels in
Slovaks could not be accounted for by exercise frequency.
Lower skin autofluorescence in this Slovak cohort might
simply reflect a selection bias, such that our cohort represents
an ‘over-healthy’, not a general, population. In addition, we
cannot exclude the possibility that our results paradoxically
reflect the higher cardiovascular mortality of Slovak adults,
as high-risk persons with higher skin autofluorescence die
earlier; neither can we exclude the existence of a geograph-
ical latitude-associated variation in skin autofluorescence.
Infants
Infant formulas may contain up to 100s-fold more advanced
glycation end products than human breast milk [4,29]. Their
intake is associated with higher plasma and urinary levels of
advanced glycation end products [4,19]. Higher skin auto-
fluorescence in formula-fed infants than in breastfed infants
suggests that high long-term consumption of dietary (infant
formula-derived) advanced glycation end products might
enhance skin autofluorescence.
Conclusions
In conclusion, our data suggest that, except for current
well-known factors (ageing, smoking, gender, age-related
diseases), regular physical exercise, and in infants the
consumption of infant formulas, affect skin autofluorescence
levels. More data on reference ranges of skin autofluores-
cence in different general population cohorts, and studies
targeting lifestyle or geographical factors are needed to
elucidate which factors play a role in the determination of
skin autofluorescence.
Funding sources
None.
Competing interests
None declared.
Acknowledgements
The authors wish to express their gratitude to all participat-
ing subjects, cooperating paediatricians and Drs Katarı´na
Volkovova´, Bozˇena Smolkova´ and Marta Staruchova´ from
the Slovak Medical University in Bratislava for their help in
recruiting the subjects and DiagnOptics Technologies BV,
Groningen, the Netherlands for providing the AGE Reader to
perform the study.
584
ª 2013 The Authors.
Diabetic Medicine ª 2013 Diabetes UK
DIABETICMedicine Skin autofluorescence in Slovaks: effects of exercise and diet � K. Simon Klenovics et al.
References
1 Miyata T, Kurokawa K. van Ypersele de Strihou C. Relevance of
oxidative and carbonyl stress to long-term uremic complications.
Kidney Int Suppl 2000; 76: S120–125.
2 Gugliucci A, Bendayan M. Renal fate of circulating advanced
glycated end products (AGE): evidence for reabsorption and
catabolism of AGE-peptides by renal proximal tubular cells.
Diabetologia 1996; 39: 149–160.
3 Cerami C, Founds H, Nicholl I, Mitsuhashi T, Giordano D,
Vanpatten S, et al. Tobacco smoke is a source of toxic reactive
glycation products. Proc Natl Acad Sci U S A 1997; 94:
13915–13920.
4 Sebekova K, Saavedra G, Zumpe C, Somoza V, Klenovicsova K,
Birlouez-Aragon I. Plasma concentration and urinary excretion of
N epsilon-(carboxymethyl)lysine in breast milk- and formula-fed
infants. Ann N Y Acad Sci 2008; 1126: 177–180.
5 Somoza V, Wenzel E, Weiss C, Clawin-Radecker I, Grubel N,
Erbersdobler HF. Dose-dependent utilisation of casein-linked
lysinoalanine, N(epsilon)-fructoselysine and N(epsilon)-carbo-
xymethyllysine in rats. Mol Nutr Food Res 2006; 50: 833–841.
6 Boor P, Celec P, Behuliak M, Grancic P, Kebis A, Kukan M, et al.
Regular moderate exercise reduces advanced glycation and ame-
liorates early diabetic nephropathy in obese Zucker rats. Metabol
Clin Exp 2009; 58: 1669–1677.
7 Panteleeva IG, Rogozkin VA. Effect of physical load on serum
protein glycation in rats with induced diabetes. Ross Fiziol Zh Im I
M Sechenova 2001; 87: 1202–1207.
8 Goon JA, Aini AHN, Musalmah M, Anum MYY, Nazaimoon
WMW, Ngah WZW. Effect of tai chi exercise on DNA damage,
antioxidant enzymes and oxidative stress in middle-age adults.
J Phys Act Health 2009; 6: 43–54.
9 Goh SY, Cooper ME. The role of advanced glycation end products
in progression and complications of diabetes. J Clin Endocrinol
Metab 2008; 93: 1143–1152.
10 Busch M, Franke S, Ruster C, Wolf G. Advanced glycation
end-products and the kidney. Eur J Clin Invest 2010; 40: 742–755.
11 Bierhaus A, Humpert PM, Morcos M, Wendt T, Chavakis T,
Arnold B, et al. Understanding RAGE, the receptor for advanced
glycation end products. J Mol Med (Berl) 2005; 83: 876–886.
12 Noordzij MJ, Lefrandt JD, Loeffen EAH, Saleem BR, Meerwaldt R,
Lutgers HL, et al. Skin autofluorescence is increased in patients
with carotid artery stenosis and peripheral artery disease. Int
J Cardiovasc Imag 2012; 28: 431–438.
13 McIntyre NJ, Fluck RJ, McIntyre CW, Taal MW. Skin autofluo-
rescence and the association with renal and cardiovascular risk
factors in chronic kidney disease stage 3. Clin J Am Soc Nephrol
2011; 6: 2356–2363.
14 Smit AJ, Gerrits EG. Skin autofluorescence as a measure of
advanced glycation end product deposition: a novel risk marker
in chronic kidney disease. Curr Opin Nephrol Hypertens 2010; 19:
527–533.
15 Meerwaldt R, Lutgers HL, Links TP, Graaff R, Baynes JW, Gans
ROB, et al. Skin autofluorescence is a strong predictor of cardiac
mortality in diabetes. Diabetes Care 2007; 30: 107–112.
16 Gerrits EG, Lutgers HL, Smeets GHW, Groenier KH, Smit AJ,
Gans ROB, et al. Skin autofluorescence: a pronounced marker of
mortality in hemodialysis patients. Nephron Extra 2012; 2:
184–191.
17 Koetsier M, Lutgers HL, de Jonge C, Links TP, Smit AJ, Graaff R.
Reference values of skin autofluorescence. Diabetes Technol Ther
2010; 12: 399–403.
18 Yue X, Hu H, Koetsier M, Graaff R, Han C. Reference values for
the Chinese population of skin autofluorescence as a marker of
advanced glycation end products accumulated in tissue. Diabet
Med 2011; 28: 818–823.
19 Klenovics KS, Boor P, Somoza V, Celec P, Fogliano V, �Sebekov�a K.
Advanced glycation end products in infant formulas do not
contribute to insulin resistance associated with their consumption.
PLoS One 2013; 8: e53056.
20 Mulder DW, van de Water T, Lutgers HL, Graaff R, Gans RO,
Zijlstra F et al. Skin autofluorescence, a novel marker for glycemic
and oxidative stress-derived advanced glycation endproducts: an
overview of current clinical studies, evidence, and limitations.
Diabetes Technol Ther 2006; 8: 523–535.
21 Ginter E, Havelkova B, Rovny I, Hlava P, Barakova A, Kudlackova
M. Health status of the Slovakia population at its entry to the
European Union. Brat Lek List 2005; 106: 45–54.
22 Sˇebekova´ K, Krivosˇ�ıkova´ Z, Gajdosˇ M. Total plasma Ne-(carb-
oxymethyl)lysine and sRAGE levels are inversely associated with a
number of metabolic syndrome risk factors in non-diabetic
young-to-middle-aged medication-free subjects. Clin Chem Lab
Med 2014; 52: 139–149.
23 Peppa M, Brem H, Ehrlich P, Zhang JG,Cai WJ, Li Z, et al.
Adverse effects of dietary glycotoxins on wound healing in
genetically diabetic mice. Diabetes 2003; 52: 2805–2813.
24 Jochemsen BM, Mulder DJ, Van Doormaal JJ, Mulder G, Volmer
M, Graaff R, et al. Relation between food and drinking habits, and
skin autofluorescence and intima media thickness in subjects at high
cardiovascular risk. J Food Nutr Res 2009; 48: 51–58.
25 Nichols M, Townsend N, Luengo-Fernandez R, Leal J, Gray A,
Scarborough P, et al. European Cardiovascular Disease Statistics
2012. Brussels/Sophia Antipolis: European Heart Network/
European Society of Cardiology, 2012.
26 Hansen AL, Carstensen B, Helge JW, Johansen NB, Gram B,
Christiansen JS et al. Combined heart rate- and accelerome-
ter-assessed physical activity energy expenditure and associations
with glucose homeostasis markers in a population at high risk of
developing diabetes: The ADDITION-PRO Study. Diabetes Care
2013; 36: 3062–3069.
27 Morton JP, Kayani AC, McArdle A, Drust B. The exercise-induced
stress response of skeletal muscle, with specific emphasis on
humans. Sports Med 2009; 39: 643–662.
28 Schmid K, Haslbeck M, Buchner J, Somoza V. Induction of heat
shock proteins and the proteasome system by casein-N epsi-
lon-(carboxymethyl)lysine and N epsilon-(carboxymethyl)lysine in
Caco-2 cells. Ann N Y Acad Sci 2008; 1126: 257–261.
29 Pischetsrieder M, Henle T. Glycation products in infant formulas:
chemical, analytical and physiological aspects. Amino Acids 2012;
42: 1111–1118.
Supporting Information
Additional Supporting Information may be found in the
online version of this article:
Figure S1. Mean skin autofluorescence per age group in male
and female subjects.
Figure S2. Mean skin autofluorescence per age group in
non-smokers and smokers.
Table S1. BMI and age of probands evaluated in regard to
physical exercise.
ª 2013 The Authors.
Diabetic Medicine ª 2013 Diabetes UK 585
Research article DIABETICMedicine