You are on page 1of 7

Diabetes Care 1

Oana Maftei,1 Alexia S. Pena,1,2


Early Atherosclerosis Relates to Thomas Sullivan,3 Timothy W. Jones,4,5,6
Kim C. Donaghue,7 Fergus J. Cameron,8,9,10
Urinary Albumin Excretion and Elizabeth Davis,4,5,6 Andrew Cotterill,11
Maria E. Craig,7 Roger Gent,1
Cardiovascular Risk Factors in Neil Dalton,12 Denis Daneman,13
David Dunger,14 John Deaneld,15 and
Adolescents With Type 1 Diabetes: Jenny J. Couper,1,2 on behalf of the AdDIT

Adolescent Type 1 Diabetes cardio- Study Group

renal Intervention Trial (AdDIT)


DOI: 10.2337/dc14-0700

1
Departments of Endocrinology and Diabetes
and Medical Imaging, Womens and Childrens
Hospital, Adelaide, Australia
2
Robinson Institute and Discipline of Paediatrics,
University of Adelaide, Adelaide, Australia
3
School of Population Health, University of Ade-
laide, Adelaide, Australia
4
Department of Endocrinology and Diabetes,
Princess Margaret Hospital for Children, Subiaco,
OBJECTIVE Australia
5
Telethon Institute for Child Health Research,
The origins of cardiovascular and renal disease in type 1 diabetes begin during University of Western Australia, Subiaco,
childhood. We aimed to evaluate carotid (cIMT) and aortic intima-media thickness Australia
6
(aIMT) and their relationship with cardiovascular risk factors and urinary albumin School of Paediatrics and Child Health, Univer-
excretion in adolescents with type 1 diabetes in the Adolescent Type 1 Diabetes sity of Western Australia, Subiaco, Australia
7
Institute of Endocrinology and Diabetes, The
cardio-renal Intervention Trial (AdDIT). Childrens Hospital at Westmead, Sydney,
Australia
RESEARCH DESIGN AND METHODS 8
Department of Endocrinology and Diabetes,
A total of 406 adolescents with type 1 diabetes, who were 14.1 6 1.9 years old Royal Childrens Hospital, Melbourne, Australia
9
with type 1 diabetes duration of 6.7 6 3.7 years, and 57 age-matched control Department of Paediatrics, University of Mel-

CARDIOVASCULAR AND METABOLIC RISK


bourne, Australia
subjects provided clinical and biochemical data and ultrasound measurements of 10
Murdoch Childrens Research Institute, Mel-
vascular structure (cIMT and aIMT). Vascular endothelial and smooth muscle bourne, Australia
11
function was also measured in 123 of 406 with type 1 diabetes and all control Department of Paediatric Endocrinology, Ma-
subjects. ter Childrens Hospital, Brisbane, Australia
12
WellChild Laboratory, St. Thomas Hospital,
RESULTS London, U.K.
13
Department of Paediatrics, The Hospital for
In type 1 diabetic subjects, mean/maximal aIMT (P < 0.006; <0.008), but not Sick Children, University of Toronto, Toronto,
mean/maximal cIMT, were greater than in control subjects. Mean/maximal aIMT Canada
14
related to urinary albumin-to-creatinine ratio (multiple regression coefcient [SE], University Department of Paediatrics,
Addenbrookes Hospital, Cambridge, U.K.
0.013 [0.006], P = 0.03; 0.023 [0.007], P = 0.002), LDL cholesterol (0.019 [0.008], P = 15
National Centre for Cardiovascular Disease
0.02; 0.025 [0.011], P = 0.02), and age (0.010 [0.004], P = 0.004; 0.012 [0.005], P = Prevention and Outcomes University College
0.01), independent of other variables. Mean/maximal cIMT was greater in males London, London, U.K.
(0.023 [0.006], P = 0.02; 0.029 [0.007], P = <0.0001), and mean cIMT related in- Corresponding author: Jenny J. Couper, jennifer.
dependently to systolic blood pressure (0.001 [0.001], P = 0.04). Vascular smooth couper@adelaide.edu.au.
muscle function related to aIMT and cIMT but not to urinary albumin excretion. Received 18 March 2014 and accepted 30 June
2014.
CONCLUSIONS J.D. and J.J.C. contributed equally as cosenior
authors.
aIMT may be a more sensitive marker of atherosclerosis than cIMT in type 1
2014 by the American Diabetes Association.
diabetes during mid-adolescence. Higher urinary albumin excretion, even within
Readers may use this article as long as the work
the normal range, is associated with early atherosclerosis and should direct clin- is properly cited, the use is educational and not
ical attention to modiable cardiovascular risk factors. for prot, and the work is not altered.
Diabetes Care Publish Ahead of Print, published online July 28, 2014
2 Early Atherosclerosis in Type 1 Diabetes Diabetes Care

Children with type 1 diabetes are at risk factors in adolescents with type 1 showed a close correlation, and the me-
greatly increased risk for the develop- diabetes with structural arterial wall dian of the six urinary ACR measure-
ment of both renal and cardiovascular changes. We hypothesized that even mi- ments was used in analysis (17).
disease in later life (1,2). Evidence is ac- nor increases in albumin excretion would Exclusion criteria were other types of
cumulating that these two complications be associated with early atherosclerosis diabetes without detected islet autoanti-
may have a common pathophysiology, but that this would be detectable only bodies at diagnosis, severe hyperlipidemia
with endothelial dysfunction a key early in the abdominal aorta. Secondary aims and/or familial hypercholesterolemia, hy-
event. were to compare cIMT and aIMT in type pertension, exposure to ACE inhibitors or
Microalbuminuria is a recognized 1 diabetes and control subjects and in a statins, other comorbidities considered
marker of endothelial damage (3) and subgroup examine the relationship of unsuitable by the investigator (excluding
predicts progression to proteinuria and vascular endothelial and smooth muscle treated hypothyroidism and celiac dis-
diabetic nephropathy, as well as to ath- function with IMT and urinary albumin ease), and proliferative retinopathy.
erosclerosis (4) and increased cardio- excretion. Healthy age- and sex-matched control
vascular risk (5). It is, however, rare in subjects (n = 57) were recruited at the
adolescents with type 1 diabetes who RESEARCH DESIGN AND METHODS central site, Adelaide, South Australia,
more often have higher urinary albumin Subjects and Study Design
from relatives and school friends of the
excretion rates within the normal range, A total of 406 adolescents, aged 1016 participants with type 1 diabetes. They
which are associated with later progres- years, with type 1 diabetes for more had the same clinical, biochemical, and
sion to microalbuminuria and protein- than 1 year, recruited in ve centers ultrasound assessments. They were com-
uria (6). Renal decline may, however, across Australia, were enrolled in this pared with 167 adolescents with type 1
precede microalbuminuria, at least in cross-sectional study (Table 1). Some diabetes for IMT and 123 adolescents
adults with type 1 diabetes (7,8), and participants (n = 337) had been with type 1 diabetes for ow-mediated
the Diabetes Control and Complications screened for urinary albumin excretion dilatation (FMD) and glyceryl trinitrate
Trial/Epidemiology of Diabetes Inter- as part of AdDIT, a multicenter multina- mediated dilatation (GTN), and all had
ventions and Complications (DCCT/ tional (U.K., Canada, and Australia) ran- their images measured on the same ul-
EDIC) study irrefutably conrms the im- domized controlled trial (clinical trial trasound machine in Adelaide.
portance of intensive therapy to reduce reg. no. ISRCTN91419926) (10). An addi- The study was approved by the Hu-
risk at each stage of the progression to tional 69 participants were not screened man Research Ethics Committee of
proteinuria (9). for the AdDIT study but were recruited each of the ve participating centers in
The Adolescent Type 1 Diabetes cardio- using the same inclusion criteria and in- Australia. Written informed consent was
renal Intervention Trial (AdDIT) (10) is de- vestigations, except for having one early obtained from all parents/guardians and
signed to examine the impact of minor morning urinary albumin-to-creatinine the study participants.
differences in albumin excretion in ado- ratio (ACR) only.
lescents on the initiation and pro- The AdDIT screening measured ACR in Clinical Assessments
gression of cardiovascular and renal two sets of three consecutive early-morning All participants were required to be well
disease. The primary cardiovascular end urine samples. The two screening visits without fever, intercurrent infection, or
point in AdDIT is carotid intima-media
thickness (cIMT). Subclinical athero-
sclerosis can be detected noninvasively Table 1Characteristics of adolescents with type 1 diabetes and control subjects
using high-resolution ultrasound to mea- Type 1 diabetic subjects Control subjects P
sure the intima-media thickness (IMT) of
n 406 57
the carotid arteries, which predicts car-
diovascular morbidity and mortality Age (years) 14.2 6 1.9 14.0 6 2.9 0.6
(11,12).Whereas cIMT is increased by Diabetes duration (years) 6.7 6 3.7 d d
late adolescence and during early adult- BMI (kg/m2) 22.0 6 3.6 20.9 6 4.1 0.03
hood in type 1 diabetes (13), this nding BMI z score 0.61 6 0.8 0.34 6 0.96 0.02
is inconsistent in earlier childhood, when Waist circumference (cm) 74.3 6 9.1 71.3 6 9.2 0.03
the rst changes of atherosclerosis be- Mean SBP (mmHg) 117.1 6 12.0 109.6 6 7.4 ,0.0001
gin. Autopsy studies in children show Mean DBP (mmHg) 65.7 6 7.6 60.8 6 6.1 ,0.0001
that the earliest atherosclerotic changes HbA1c (%, mmol/mol) 8.5 6 1.4, 5.2 6 0.3, ,0.0001
(fatty streaks) occur in the distal abdom- 69 6 11.4 33 6 1.9
inal aortic wall (14); aorta IMT (aIMT) is Total cholesterol (mmol/L) 4.5 6 0.9 4.0 6 0.7 ,0.0001
measurable and increased in children HDL cholesterol (mmol/L) 1.56 6 0.35 1.46 6 0.34 0.05
and adolescents with type 1 diabetes Triglycerides (mmol/L) 1.05 6 0.71 0.78 6 0.34 0.01
(15,16). It is also possible that drivers of LDL cholesterol (mmol/L) 2.42 6 0.75 2.17 6 0.57 0.02
atherosclerosis differ between different eGFR (mL/min/1.73 m2) 128.79 6 24.75 d d
vascular beds. Urinary ACR (mg/mmol) 1.03 6 0.95 d d
The primary aim of this study was to Race (% European/Asian/other) 91/6/3 93/7/0 0.7
examine the relationship of increased uri-
Data are means 6 SD unless otherwise indicated.
nary albumin excretion and cardiovascular
care.diabetesjournals.org Maftei and Associates 3

ketosis on the days of investigation. the central site who were blinded to the Nitrolingual Pump spray, Sano Aventis).
Data on age, duration of diabetes, an- subjects clinical characteristics; and 5) For each scan, measurements were
thropometry (weight, height, and waist intraobserver coefcient of variation made over four consecutive cardiac
circumference), pubertal (Tanner) (CV) in 33 subjects, within this cohort, cycles, incident with the R wave on the
stage, and blood pressure were col- was 0.6% for cIMT and 1.4% for aIMT, ECG, by observers blinded to the subject
lected. Height was measured on wall- and interobserver CV in 20 subjects, group using ultrasonic calipers. Measure-
mounted stadiometers and weight on within this cohort, was 1.8% for cIMT ments were averaged and expressed as
electronic scales. Waist circumference and 2.3% for aIMT. percentages of the resting vessel diame-
was measured in triplicate with a exi- Ultrasound images of left and right ter. Interobserver CV between 20 sub-
ble tape at the midpoint between the common carotid arteries and abdominal jects studied on two occasions was
lower margin of the last palpable rib aorta were acquired using our standard- 3.9% for FMD and 4.0% for GTN (18).
and the top of the iliac crest following ized protocol (16). For cIMT, images of
World Health Organization guidelines. the posterior wall of the distal 10-mm- Biochemical Assessments
Brachial blood pressure was mea- long arterial segment of the common In AdDIT participants, urinary biochem-
sured three times (Omran M6 Blood carotid artery, just 1 cm proximal to ical assessments were performed in a
Pressure Monitor; Kyoto), separated the carotid bulb, were recorded. For central laboratory (WellChild Labora-
by 5-min intervals using the appropriate aIMT, images of the straight, most distal tory, London, U.K.). Urine albumin was
cuff size. The average of the three 10 mm of the abdominal aorta just be- measured using nephelometric immu-
measurements was used in the analysis. fore the bifurcation were recorded. A noassay (Siemens BN Prospec). Urine al-
The cuff was chosen to be of the appro- minimum of three images at end dias- bumin concentrations below the limit of
priate size for the adolescents upper tole, triggered on the R wave of the elec- quantitation of nephelometry, ,2.1
arm, with a bladder width that was at trocardiogram (ECG), for each of the mg/L, were measured using ELISA.
least 40% of the arm circumference at a common carotid arteries and abdominal Between-batch imprecision for the BN
point midway between the olecranon aorta was taken and digitally stored for Prospec was 3.7% at 4.16 mg/L (n =
and the acromion and a bladder length later analysis. The greatest distance be- 51), 2.9% at 19.0 mg/L (n = 55), and
to cover 80100% of the circumference tween the lumen-intima interface and 2.9% at 144 mg/L (n = 54). Between-
of the arm. media-adventitia interface was mea- batch imprecision on the ELISA at ,2.1
A blood sample was collected for sured at a minimum of 100 points mg/L was ,15%. Urine creatinine was
HbA1c and lipid prole (total cholesterol, using a semiautomated edge detection measured using a chromatographic
triglycerides, HDL cholesterol, and LDL and measurement computer software stable isotope dilution electrospray mass
cholesterol) measurement at the local package (B Bailey; Royal Prince Alfred spectrometrymass spectrometry meth-
laboratories. Hospital, NSW, Australia). Three best od on an AB SCIEX API5000. Between-
quality images for cIMT (right and left batch imprecision (n = 48) was 2.6%
Ultrasound Assessment of Vascular carotids) and aIMT were selected and at 6.89 mmol/L and 3.3% at 17.4
Structure and Function analyzed for mean and maximal IMT. mmol/L. Plasma creatinine was mea-
Ultrasound images of vascular struc- The mean of the readings was recorded sured using a reference stable isotope
ture (aIMT and cIMT) were collected to give the nal result for each subject. dilution electrospray mass spectrometry
at ve Australian vascular centers by Ultrasound assessments of vascular mass spectrometry.
accredited vascular sonographers. All endothelial function measured by FMD In those 69 of 406 subjects who were
images were read at the central site in and vascular smooth muscle function not screened for AdDIT, urinary albumin
Adelaide. Reproducibility and quality measured by GTN were performed at was measured by an immunoturbidimetric
control among the measurements of the central site only. FMD and GTN assay and urinary creatinine by an en-
IMT in each center and between centers were assessed as previously described zyme colorimetric method (Roche Cobas
were ensured as follows: 1) sonographer (18). Brachial artery diameter was mea- C501; Hitachi) at the Adelaide site. There
training conducted through the Vascular sured in a longitudinal section 215 cm was a close correlation between mea-
Imaging Unit at the Women and Child- above the elbow using B mode ultra- surements performed in London and
rens Hospital, Adelaide, which has ex- sound with a 17-MHz linear array trans- Adelaide in 106 samples (r = 0.99 [P ,
tensive experience over 15 years in ducer (Philips iU22; Philips, Bothell, 0.0001], 0.81 [P , 0.0001], and 0.99 [P ,
running vascular health studies in chil- WA). An ECG was recorded simulta- 0.0001] for albumin, creatinine, and ACR,
dren; 2) intrasonographer reproducibil- neously with the ultrasound images. respectively).
ity on scans produced on two separate Each study included four scans: 1) a rest- Lipids were measured using commer-
occasions in the same ve subjects for ing scan, after which reactive hyperemia cial enzymatic assays on Roche Hitachi
aIMT was 2.8% and on scans produced was induced by occluding arterial blood cobas C systems. HbA1c was assessed in
on two separate occasions in the same ow using a sphygmomanometer in- the ve local laboratories using a Van-
13 subjects for cIMT was 3.6%; 3) three ated to 250 mmHg for 4 min; 2) an tage analyzer (Siemens Diagnostics,
trial images were sent and assessed for FMD scan recorded between 45 and 75 s Camberley, U.K.), or a Variant analyzer
quality before the sonographer was ac- after cuff deation; 3) a recontrol scan (Bio-Rad Laboratories, Hercules, CA),
credited to be part of the study; 4) all 1015 min later; and 4) the last scan, both of which showed high correlations
images were read by one of two indepen- taken 4 min after the sublingual admin- with DCCT-standardized control sub-
dent observers (O.M. and M. La Forgia) at istration of the GTN spray (400 mg; jects (r = 0.98) .
4 Early Atherosclerosis in Type 1 Diabetes Diabetes Care

Calculations Some (21.7%) of the adolescents with aIMT images collected only) was 0.49 6
BMI was calculated as weight in kilo- type 1 diabetes met the target meta- 0.08/0.59 6 0.09 mm. Univariable anal-
grams divided by the square of height bolic control (HbA1c ,7.5%), and 135 ysis showed that males and those with
in meters, and z scores were calculated of 406 (33.3%) adolescents with type 1 longer diabetes duration, greater waist
using the Centers for Disease Control diabetes and 14 of 57 (24.5%) control circumference, higher SBP and DBP,
and Prevention, National Center for subjects (P , 0.05) were overweight or higher HbA1c, higher eGFR, and lower
Health Statistics 2000 growth charts, obese (BMI .85th centile for sex and GTN had greater mean and maximal
U.S. (19). Estimated glomerular ltra- age). Pubertal staging by self-report in cIMT. In the multivariable analysis,
tion rate (eGFR) (milliliters per minute 347 of 406 type 1 diabetic subjects mean and maximal cIMT were indepen-
per 1.73 meters squared) was calculated showed that 11% were Tanner stage 1, dently associated with sex, and mean
as 42 3 height (centimeters)/plasma 55% Tanner 24, and 34% Tanner 5. cIMT independently associated with
creatinine (micromoles per liter). SBP (Table 3). As for aIMT, center effect
Associations With IMT was included in the multivariable anal-
Statistical Analysis aIMT and cIMT was measurable in 92% ysis (Tables 2,3). In type 1 diabetes,
Data are summarized as mean 6 SD. and 100% of the subjects studied, mean cIMT related to mean aIMT (r =
Clinical characteristics (Table 1) were com- respectively. 0.37, P , 0.0001) and maximal cIMT re-
pared between groups using independent- lated to maximal aIMT (r = 0.41,
sample t tests and x2 tests as appropriate. aIMT P , 0.0001).
Univariable linear regression models Mean/maximal aIMT in 373 subjects with
were used to identify predictors of type 1 diabetes was 0.61 6 0.11/0.76 6 IMT in Type 1 Diabetic Versus Control
each IMT outcome in adolescents with 0.16 mm. Univariable analysis showed Subjects
type 1 diabetes. Variables that were sta- that older subjects and those with longer Adolescents with type 1 diabetes in
tistically signicant (P , 0.05) on univari- diabetes duration, greater waist cir- whom IMT images were taken on the
able analysis and that had data available cumference, higher systolic (SBP) and one ultrasound machine at the Ade-
on .75% of observations (all except for diastolic blood pressure (DBP), LDL cho- laide central site (n = 167) had greater
GTN and eGFR) were entered into multi- lesterol and ACR, and lower GTN had aIMT, but not cIMT, than healthy age-
variable models for each IMT outcome in greater mean and maximal aIMT. In the and sex-matched control subjects (n =
order to identify independent predictors multivariable analysis, mean and maxi- 57) (Table 4). Respectively, 60.7% and
of the respective outcome. To avoid po- mal aIMT were independently associated 61.3% of adolescents with type 1 diabe-
tential issues with multicollinearity, BMI with age, LDL cholesterol, and ACR (Ta- tes had mean/maximal aIMT above the
(correlated with waist circumference) ble 2). An increase in ACR by 1 unit was control mean; 52.1% and 55.1% had
and total cholesterol (correlated with associated with a signicant increase in mean/maximal cIMT above the control
LDL cholesterol) were not considered mean aIMT by 0.013 mm and an in- mean.
for inclusion into multivariable models. crease in maximal aIMT by 0.023 mm,
Comparisons of the four IMT outcomes independent of age, diabetes duration, Vascular Endothelial and Smooth
(mean/maximal cIMT and mean/maximal waist circumference, blood pressure, Muscle Function
aIMT) between the adolescents with type LDL cholesterol, and center effect. The FMD and GTN were lower in type 1 di-
1 diabetes and control subjects were per- signicant relationship between ACR abetic than control subjects (Table 4).
formed using simple linear regression and mean/maximal aIMT was not al- Lower GTN related to greater mean/
models. tered by excluding the 69 participants maximal aIMT (regression coefcient =
Two-tailed P values ,0.05 were con- who had only one ACR measurement. 20.004, P = 0.004; 20.005, P = 0.005)
sidered statistically signicant. Statisti- and greater mean/maximal cIMT (regres-
cal analyses were performed using SAS, cIMT sion coefcient = 20.002, P = 0.006;
version 9.3 (SAS Institute, Cary, NC). Mean/maximal cIMT in 397 subjects 20.002, P = 0.01). Neither GTN nor FMD
with type 1 diabetes (nine subjects had related to urinary albumin excretion.
RESULTS
Subject Characteristics Table 2Multivariable models for aIMT in type 1 diabetes
A total of 406 adolescents (211 boys) Mean aIMT Maximal aIMT
with type 1 diabetes (47 from New
Predictor variable Coefcient (SE) P Coefcient (SE) P
South Wales, 35 from Queensland, 167
from South Australia, 75 from Victoria, Age 0.010 (0.004) 0.004 0.012 (0.005) 0.01
and 82 from Western Australia) and 57 Diabetes duration 0.001 (0.002) 0.70 0.002 (0.002) 0.38
healthy adolescents (26 boys) from Waist circumference 0.001 (0.001) 0.41 0.001 (0.001) 0.28
South Australia were included in the Mean SBP 0.001 (0.001) 0.12 0.002 (0.001) 0.07
analysis (Table 1). There were no signif- Mean DBP 20.001 (0.001) 0.38 20.001 (0.001) 0.71
icant clinical or biochemical differences LDL cholesterol 0.019 (0.008) 0.02 0.025 (0.011) 0.02
between the type 1 diabetic subjects Urinary ACR 0.013 (0.006) 0.03 0.023 (0.007) 0.002
screened for AdDIT (n = 337) and the Center* ,0.0001 ,0.0001
rest of the type 1 diabetic group (n = *Comparing all ve centers simultaneously.
69) (data not shown).
care.diabetesjournals.org Maftei and Associates 5

Table 3Multivariable models for cIMT in type 1 diabetes measurable in 100%. The association be-
Mean cIMT Maximal cIMT
tween the two was relatively weak,
which is consistent with the possibility
Predictor variable Coefcient (SE) P Coefcient (SE) P
that they are increasing at different ages
Sex (boys vs. girls) 0.023 (0.006) 0.0001 0.029 (0.007) ,0.0001 and driven by potentially different risk
Diabetes duration 0.001 (0.001) 0.29 0.001 (0.001) 0.16 factors. The main technical difculties
Waist circumference 0.001 (0.001) 0.29 0.001 (0.001) 0.37 that interfere with a satisfactory aIMT
Mean SBP 0.001 (0.001) 0.04 0.001 (0.001) 0.15 image are obesity and the extended
Mean DBP 20.001 (0.001) 0.08 20.001 (0.001) 0.23 study time by 15 min; these two factors
HbA1c (%) 20.003 (0.002) 0.12 20.003 (0.002) 0.29 are limitations to its use outside re-
Center* ,0.0001 ,0.0001 search studies with current ultrasound
*Comparing all ve centers simultaneously.
technology.
We have shown recently that adoles-
cents enrolled in AdDIT with urinary al-
bumin excretion rates in the upper
CONCLUSIONS recognized for .20 years (3,20,21). En-
tertile of the normal range, which pre-
We report the largest study of changes dothelial dysfunction is the rst critical
dicts microalbuminuria in 85% of cases
in two vascular beds in adolescents with step in the development of atheroscle-
(6), have a small but signicant in-
type 1 diabetes. Structural changes in rosis (22). Early rises in urinary albumin
crease in arterial stiffness (17). This
the aorta and carotid arteries could be excretion precede the development of
Australian AdDIT cohort provides cor-
detected in .50% of adolescents with microalbuminuria and proteinuria (23).
roborative evidence for the early link
type 1 diabetes, but there was a differ- It follows that the rst structural changes
between urinary albumin excretion
ent related risk prole, and changes in of atherosclerosis could relate to the
and subclinical cardiovascular disease.
the aorta were more common at this rst biochemical changes of diabetic
The cross-sectional analysis prevents
age. Aortic IMT was therefore able to nephropathy. To our knowledge, this is
the unraveling of cause/effect relation-
better differentiate adolescents with the rst study to provide evidence of
ships, and only longitudinal follow-up
type 1 diabetes from control subjects this.
will clarify the temporal relationship
than was carotid wall changes. Aortic We also provide more evidence that
between early nephropathy and early
IMT enabled detection of the very early atherosclerosis begins in the abdominal
atherosclerosis.
wall changes that are present with even aorta (14). More subjects had raised
The relationship between IMT and
small differences in urinary albumin ex- aIMT above the control mean, and
overweight may also have an impact
cretion. This not only supports the con- aIMT, but not cIMT, was signicantly
on these longitudinal outcomes. The
cept of early intervention but provides a greater in type 1 diabetic than in control
overweight epidemic has resulted in up
link between renal and cardiovascular subjects. This conrms our and others
to one-third of adolescents with type 1
disease. smaller studies (15,16) suggesting that
diabetes in Australia being overweight
The independent relationship be- aIMT is a more sensitive measure in chil-
or obese, consistent with international
tween aIMT and urinary albumin excre- dren and adolescents and therefore
trends (24), and this could add to the
tion extends our knowledge of the potentially a more sensitive outcome
cardiovascular burden of diabetes. BMI
pathogenesis of cardiovascular and renal measure in intervention trials in this
and waist circumference had direct re-
disease in type 1 diabetes by showing age-group. The difference in aIMT be-
lationships with IMT, but these did not
that the rst signs of the development tween type 1 diabetic patients and
remain on multivariate analysis of the
of cardiovascular disease and diabetic age- and sex-matched control subjects
cohort, 33% of whom were overweight
nephropathy are related. The concept was equivalent to that seen with a 5-
or obese.
that microalbuminuria is a marker of a to 6-year age increase in the type 1 di-
There was a signicant center effect
generalized endothelial damage, as well abetic patients. aIMT was measurable in
on the measurement of aIMT and cIMT.
as a marker of renal disease, has been 92% of subjects, whereas cIMT was
We chose to adjust for center differ-
ences in the multivariate models as op-
posed to standardizing the outcomes by
Table 4IMT and vascular function in type 1 diabetic and control subjects
center. However, a phantom image ex-
Type 1 diabetic subjects Control subjects P
change had also been performed be-
n 167 57 tween the ve centers, and this
Mean cIMT (mm) 0.43 6 0.06 0.43 6 0.06 0.81 method of standardization did not sig-
Maximal cIMT (mm) 0.52 6 0.07 0.51 6 0.07 0.41 nicantly alter the multivariate analysis
Mean aIMT (mm) 0.56 6 0.11 0.51 6 0.10 0.008 results. Importantly, there was no evi-
Maximal aIMT (mm) 0.67 6 0.13 0.61 6 0.12 0.005 dence for an effect modication by any
FMD (%) 5.77 6 4.3 7.16 6 4.48 0.001 center for the predictors (albumin ex-
GTN (%) 23.27 6 7.51 25.43 6 8.61 0.001 cretion, LDL cholesterol, age, sex, and
Vessel diameter (cm) 0.28 6 0.04 0.27 6 0.04 0.01 systolic blood pressure) of IMT.
The SEARCH for Diabetes in Youth
Data are means 6 SD unless otherwise indicated. n = 123.
Cardiovascular Disease (SEARCH CVD)
6 Early Atherosclerosis in Type 1 Diabetes Diabetes Care

study cohort, which recently investi- addition to efforts to maximize meta- Cardiovascular Nursing; American Heart Associ-
gated cIMT in type 1 diabetes, was of bolic control. ation Council on the Kidney in Heart Disease;
Interdisciplinary Working Group on Quality of
similar size, but subjects were older Care and Outcomes Research. Cardiovascular
with longer duration of disease, as in- risk reduction in high-risk pediatric patients:
vestigation was during later adoles- Acknowledgments. The authors thank a scientic statement from the American Heart
cence and early adulthood, and of Dr. Jennifer Harrington and Dr. Jemma Anderson, Association Expert Panel on Population and Pre-
Adelaide, South Australia, for contributions to vention Science; the Councils on Cardiovascular
mixed ethnic origins (13). aIMT was recruitment of subjects. The authors thank David Disease in the Young, Epidemiology and Preven-
not measured and urinary albumin Celermajer, Jason Harmer, and Dr. Jim Ramsay for tion, Nutrition, Physical Activity and Metabolism,
excretion was not included in their supervision of training and collection of IMT High Blood Pressure Research, Cardiovascular
analysis. Our study demonstrates that images. The authors are thankful for the support Nursing, and the Kidney in Heart Disease; and
atherosclerosis can be detected by of the AdDIT research nurses (Meredith Krieg, the Interdisciplinary Working Group on Quality
Alison Pryke, Julie Kendall, Claire Bingley, Julianne of Care and Outcomes Research: endorsed by
early adolescence after a shorter dura- Wilson, Alison Roberts, and Julie Dart) and the American Academy of Pediatrics. Circulation
tion of disease. In SEARCH CVD, only sonographers (Melissa La Forgia [lead], Yukari 2006;114:27102738
HbA1c independently related to cIMT Newman, Katie Maslin, Jane Koleff, Amanda 2. Jrgensen ME, Almdal TP, Carstensen B.
(13). As in SEARCH CVD, the majority Crowe, Sinh Le, and Rachel Tarte) in this study. Time trends in mortality rates in type 1 diabetes
of our type 1 diabetic subjects (78%) The authors thank all participants for their in- from 2002 to 2011. Diabetologia 2013;56:2401
volvement and commitment. 2404
did not achieve target metabolic con- Funding. The study was funded by the National 3. Deckert T, Feldt-Rasmussen B, Borch-
trol of HbA1c ,7.5% (,58 mmol/mol), Health and Medical Research Council, Australia, Johnsen K, Jensen T, Kofoed-Enevoldsen A.
consistent with a recent audit of Aus- no. 632521; Diabetes UK; the JDRF; and the Albuminuria reects widespread vascular
tralian pediatric diabetes clinics (25), British Heart Foundation. damage. The Steno hypothesis. Diabetologia
representing .80% of children with The funders had no role in the design or conduct 1989;32:219226
of the study; collection, management, analysis, 4. Krantz JS, Mack WJ, Hodis HN, Liu CR, Liu CH,
type 1 diabetes in Australia, including or interpretation of the data; or preparation, Kaufman FR. Early onset of subclinical athero-
the ve centers participating in this review, or approval of the manuscript. sclerosis in young persons with type 1 diabetes.
study. Metabolic control is an estab- Duality of Interest. No potential conicts of J Pediatr 2004;145:452457
lished risk factor of cIMT in type 1 diabe- interest relevant to this article were reported. 5. Marcovecchio ML, Tossavainen PH, Dunger
Author Contributions. O.M. acquired data, DB. Status and rationale of renoprotection stud-
tes (26), but at this early stage of
analyzed and interpreted data, and critically ies in adolescents with type 1 diabetes. Pediatr
atherosclerosis, we emphasize the addi- revised the manuscript. A.S.P. developed the Diabetes 2009;10:347355
tional importance of other cardiovascu- study concept and design, acquired data, criti- 6. Dunger DB, Schwarze CP, Cooper JD, et al.
lar risk factors. cally revised the manuscript, and supervised Can we identify adolescents at high risk for
The study is not without its limita- the study. T.S. analyzed and interpreted data, nephropathy before the development of
critically revised the manuscript, and performed microalbuminuria? Diabet Med 2007;24:131
tions. The relatively small number of
statistical analysis. T.W.J., K.C.D., and E.D. de- 136
control subjects provided comparative veloped the study concept and design, acquired 7. Molitch ME, Steffes M, Sun W, et al.; Epide-
data for IMT and vascular function data, critically revised the manuscript, and miology of Diabetes Interventions and Complica-
only, as measured on the same machine obtained funding. F.J.C. acquired data and tions Study Group. Development and progression
by the same sonographer. However, our obtained funding. A.C. and M.E.C. acquired of renal insufciency with and without albumin-
data, critically revised the manuscript, and uria in adults with type 1 diabetes in the diabe-
primary purpose was to examine the obtained funding. R.G. acquired data and su-
relationship between IMT and risk fac- tes control and complications trial and the
pervised the study. N.D. developed the study epidemiology of diabetes interventions and
tors in type 1 diabetes. We would re- concept and design and analyzed and inter- complications study. Diabetes Care 2010;33:
quire very large numbers of control preted data. D.Da. developed the study concept 15361543
subjects to examine a relationship be- and design. D.Du. developed the study concept 8. Krolewski AS, Niewczas MA, Skupien J,
and design, critically revised the manuscript, and et al. Early progressive renal decline precedes
tween ACR and IMT in the normal child- obtained funding. J.D. developed the study
hood population. the onset of microalbuminuria and its progression
concept and design and critically revised the
to macroalbuminuria. Diabetes Care 2014;37:
In conclusion, atherosclerosis is de- manuscript. J.J.C. contributed to the study con-
226234
tectable from early adolescence in type cept and design, acquired data, critically revised
9. de Boer IH; DCCT/EDIC Research Group. Kid-
the manuscript, analyzed and interpreted data,
1 diabetes. Its early independent associ- ney disease and related ndings in the diabetes
obtained funding, and supervised the study.
ations are male sex, age, systolic blood control and complications trial/epidemiology of
O.M., T.S., and J.J.C. are the guarantors of this
pressure, LDL cholesterol, and, impor- diabetes interventions and complications study.
work and, as such, had full access to all the data
Diabetes Care 2014;37:2430
tantly, urinary albumin excretion. in the study and take responsibility for the
10. Adolescent type 1 Diabetes cardio-renal In-
Changes appear to occur rst in the integrity of the data and the accuracy of the
tervention Trial Research Group. Adolescent
data analysis.
aorta in type 1 diabetes, as has been type 1 Diabetes Cardio-renal Intervention Trial
detected in normal childhood. Early (AdDIT). BMC Pediatr 2009;9:79
rises in urinary albumin excretion during References 11. Amato M, Montorsi P, Ravani A, et al. Ca-
1. Kavey RE, Allada V, Daniels SR, et al.; Amer- rotid intima-media thickness by B-mode ultra-
adolescence not only are important for ican Heart Association Expert Panel on Popula- sound as surrogate of coronary atherosclerosis:
determining risk of progression to mi- tion and Prevention Science; American Heart correlation with quantitative coronary angiog-
croalbuminuria and diabetic nephropa- Association Council on Cardiovascular Disease raphy and coronary intravascular ultrasound
thy but also may alert the clinician to in the Young; American Heart Association Coun- ndings. Eur Heart J 2007;28:20942101
increased risk of cardiovascular disease. cil on Epidemiology and Prevention; American 12. Lorenz MW, Markus HS, Bots ML, Rosvall M,
Heart Association Council on Nutrition, Physical Sitzer M. Prediction of clinical cardiovascular
Their detection during adolescence Activity and Metabolism; American Heart As- events with carotid intima-media thickness:
should prompt extra attention to modi- sociation Council on High Blood Pressure Re- a systematic review and meta-analysis. Circula-
able cardiovascular risk factors in search; American Heart Association Council on tion 2007;115:459467
care.diabetesjournals.org Maftei and Associates 7

13. Urbina EM, Dabelea D, DAgostino RB Jr, 18. Wiltshire EJ, Gent R, Hirte C, Pena A, 23. Schultz CJ, Neil HA, Dalton RN, Dunger DB;
et al. Effect of type 1 diabetes on carotid Thomas DW, Couper JJ. Endothelial dysfunction Oxforn Regional Prospective Study Group. Risk
structure and function in adolescents and relates to folate status in children and adoles- of nephropathy can be detected before the on-
young adults: the SEARCH CVD study. Diabetes cents with type 1 diabetes. Diabetes 2002;51: set of microalbuminuria during the early years
Care 2013;36:25972599 22822286 after diagnosis of type 1 diabetes. Diabetes Care
14. McGill HC Jr, McMahan CA, Herderick EE, 19. Ogden CL, Kuczmarski RJ, Flegal KM, et al. 2000;23:18111815
Malcom GT, Tracy RE, Strong JP. Origin of ath- Centers for Disease Control and Prevention 24. MacKenzie KE, Wiltshire EJ, Gent R, Hirte C,
erosclerosis in childhood and adolescence. Am J 2000 growth charts for the United States: im- Piotto L, Couper JJ. Folate and vitamin B6 rapidly
Clin Nutr 2000;72(Suppl):1307S1315S provements to the 1977 National Center for normalize endothelial dysfunction in children
15. Jarvisalo MJ, Jartti L, Nanto-Salonen K, et al. Health Statistics version. Pediatrics 2002;109: with type 1 diabetes mellitus. Pediatrics 2006;
Increased aortic intima-media thickness: 4560 118:242253
a marker of preclinical atherosclerosis in high- 20. Viberti GC, Hill RD, Jarrett RJ, Argyropoulos A, 25. Cameron F, Cotterill A, Couper J, et al. Short
risk children. Circulation 2001;104:29432947 Mahmud U, Keen H. Microalbuminuria as a pre- report: Care for children and adolescents with
16. Harrington J, Pe~ na AS, Gent R, Hirte C, dictor of clinical nephropathy in insulin-dependent diabetes in Australia and New Zealand: have we
Couper J. Aortic intima media thickness is an early diabetes mellitus. Lancet 1982;1:14301432 achieved the dened goals? J Paediatr Child
marker of atherosclerosis in children with type 1 21. Rossing P, Hougaard P, Borch-Johnsen K, Health 2013;49:E258E262
diabetes mellitus. J Pediatr 2010;156:237241 Parving HH. Predictors of mortality in insulin 26. Nathan DM, Lachin J, Cleary P, et al.; Diabetes
17. Marcovecchio ML, Woodside J, Jones T, et al.; dependent diabetes: 10 year observational fol- Control and Complications Trial; Epidemiology of
AdDIT Investigators. Adolescent Type 1 Diabetes low up study. BMJ 1996;313:779784 Diabetes Interventions and Complications Research
Cardio-Renal Intervention Trial (AdDIT): urinary 22. Ross R. The pathogenesis of atherosclero- Group. Intensive diabetes therapy and carotid
screening and baseline biochemical and cardiovas- sis: a perspective for the 1990s. Nature 1993; intima-media thickness in type 1 diabetes mellitus.
cular assessments. Diabetes Care 2014;37:805813 362:801809 N Engl J Med 2003;348:22942303

You might also like