Pathophysiology/Complications

O R I G I N A L A R T I C L E

Impaired Endothelial Function

in Preadolescent Children With
Type 1 Diabetes
GHUFRAN S. BABAR, MD, MSC1 RAYMOND G. HOFFMANN, PHD5,6 In addition, adverse carotid remodel-
HANAA ZIDAN, MD2 MARWAN DAOUD, PHD7 ing, known to portend future cardiovas-
MICHAEL E. WIDLANSKY, MD, MPH3,4 RAMIN ALEMZADEH, MD5,6 cular risk, also has been consistently
EMON DAS, MD4 reported in this population (10–13).
However, these studies have not rigor-
ously assessed pubertal status, and it re-
OBJECTIVE—We evaluated the prevalence of endothelial dysfunction as measured by flow- mains unknown whether the adverse
mediated dilatation (FMD) of the brachial artery and carotid intima-media thickness (c-IMT) in
relationship to vascular inflammatory biomarkers in preadolescent children with type 1 diabetes.
effects of type 1 diabetes on vascular ho-
meostasis are apparent even during the
RESEARCH DESIGN AND METHODS—We studied 21 type 1 diabetic children (aged preadolescent stage. We hypothesized
8.3 6 0.3 years with diabetes duration of 4.3 6 0.4 years) and 15 group-matched healthy siblings that prepubertal children with type 1 di-
(aged 7.6 6 0.3 years). Fasting plasma glucose (FPG), lipid profile, HbA1c, high-sensitivity abetes would also manifest early signs of
C-reactive protein (hs-CRP), fibrinogen, homocysteine, and erythrocyte (red blood cell [RBC]) abnormal vascular homeostasis, includ-
folate were evaluated in all subjects. Each subject underwent c-IMT and brachial artery FMD ing impaired endothelial function, in-
percentage (FMD%) measurements using high-resolution vascular ultrasound. creased carotid intima-media thickness
RESULTS—Type 1 diabetic children had higher FPG (173.4 6 7.9 mg/dL vs. 81.40 6 1.7 mg/ (c-IMT), and elevated circulating markers
dL; P , 0.0001), HbA1c (8.0 6 0.2% vs. 5.0 6 0.1%; P , 0.0001), and hs-CRP (1.8 6 0.3 vs. of inflammation. We evaluated our hy-
0.70 6 0.2; P = 0.017) than control children without significant differences in BMI, homocys- pothesis in a cross-sectional study of
teine, and fibrinogen levels; RBC folate content; and c-IMT between the groups. Children with type 1 diabetes and healthy matched sib-
type 1 diabetes had lower FMD% than control children (7.1 6 0.8% vs. 9.8 6 1.1%; P = 0.04), ling control subjects.
whereas c-IMT did not differ between groups.
CONCLUSIONS—Preadolescent children with type 1 diabetes and mean diabetes duration RESEARCH DESIGN AND
of 4 years displayed evidence of low-intensity vascular inflammation and attenuated FMD mea- METHODS—Twenty-one prepubertal
surements. These data suggest that endothelial dysfunction and systemic inflammation, known children with type 1 diabetes, aged 8.5 6
harbingers of future cardiovascular risk, are present even in preadolescent children.
0.3 years (diabetes duration of 4.3 6 0.3
Diabetes Care 34:681–685, 2011 years), were recruited from the Children’s
Hospital of Wisconsin Diabetes Clinic,

P
which is affiliated with the Medical Col-
atients with type 1 diabetes have young-adult autopsy findings. Their lege of Wisconsin. Children with type 1
two to four times the risk of de- work prompted multiple small studies diabetes were either on multiple daily in-
veloping cardiovascular disease rel- (5–9) that have consistently demon- sulin, consisting of bedtime insulin glar-
ative to the nondiabetic population (1). strated abnormal vascular homeostasis gine and premeal aspart insulin, or
Type 1 diabetes causes endothelial dys- and inflammation in children with type continuous subcutaneous insulin infusion
function and early atherosclerosis (2). En- 1 diabetes. These studies have consis- (CSII) with insulin aspart. We reviewed
dothelial dysfunction and alterations in tently demonstrated that children and 2-week, seven-point, self-monitored blood
vascular structure are early indicators of adolescents with type 1 diabetes have en- glucose logs to determine mean blood glu-
future cardiovascular events (3). Berenson dothelial dysfunction relative to nondia- cose and SDs as well as rates of moderate
et al. (4) observed that atherosclerotic betic age-matched control children, as (blood glucose ,60 mg z dL21 z week21) or
changes begin much earlier than the ap- measured by flow-mediated dilation severe hypoglycemia (blood glucose ,50
pearance of clinical disease, as shown by (FMD) in the brachial artery (5,7,8). mg z dL 21 z week21 with altered mental
status). In addition, 15 group-matched
c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c healthy siblings of the diabetic cohort
From the 1Department of Pediatric Endocrinology, Children’s Mercy Hospital, Kansas City, Missouri; the were recruited as control subjects. Inclu-
2
Section of Pediatric Endocrinology, Department of Pediatrics, Children’s Hospital of Michigan, Detroit, sion criteria consisted of prepubertal chil-
Michigan; the 3Departments of Medicine and Pharmacology, Medical College of Wisconsin, Milwaukee, dren aged 6–9 years. Exclusion criteria
Wisconsin; the 4Cardiovascular Research Center, Medical College of Wisconsin, Milwaukee, Wisconsin; the
5
Department of Pediatrics, Medical College of Wisconsin, Milwaukee, Wisconsin; the 6Children’s Research
included known dyslipidemia, hyperten-
Institute, Children’s Hospital of Wisconsin, Milwaukee, Wisconsin; and the 7Department of Pediatrics, sion, microvascular complications, anemia
Wayne State School of Medicine, Detroit, Michigan. (hemoglobin ,11.0 g/dL), congenital heart
Corresponding author: Ramin Alemzadeh, ralemzad@mcw.edu. disease, allergy to ultrasound gel, or family
Received 12 November 2010 and accepted 23 December 2010. history of hypercholesterolemia or prema-
DOI: 10.2337/dc10-2134
© 2011 by the American Diabetes Association. Readers may use this article as long as the work is properly ture cardiovascular disease. The study pro-
cited, the use is educational and not for profit, and the work is not altered. See http://creativecommons.org/ tocol was approved by the Children’s
licenses/by-nc-nd/3.0/ for details. Hospital of Wisconsin Institutional Review

care.diabetesjournals.org DIABETES CARE, VOLUME 34, MARCH 2011 681

Endothelial function in children with diabetes

Board. Informed consent and assent was Homocysteine kit (Diagnostic Products, carotid bulb was imaged, and the image
obtained from the study parents or guard- Flanders, NJ). The intra-assay and inter- was magnified. The c-IMT for each sub-
ians and the subjects. assay CVs for homocysteine were 2.7 and ject was the average of 10 measurements
3.4%, respectively. HbA 1c was deter- (five from the right and five from the left
Laboratory studies mined by the Bayer DCA (Bayer Diagnostic, common carotid artery) (17). All images
Peripheral venous blood samples were Tarrytown, NY) 2000 instrument (nondia- were recorded and analyzed with carotid
obtained to determine the complete betic range of 4.5–5.7%). Erythrocyte imaging software (Medical Imaging, Iowa
blood-count plasma glucose, HbA1c, lip- folate concentrations were measured by City, IA). A single vascular sonographer,
ids, high-sensitive C-reactive protein the use of a radioimmunoassay technique blinded to the participant’s diagnosis, an-
(hs-CRP), fibrinogen, chemistry panel, (Quest Diagnostics, Nichols Institute) with alyzed all the recorded ultrasound scans.
homocysteine, and erythrocyte folate (red intra-assay and interassay CVs of 3.9 and
blood cell [RBC] folate) between 0800 h 5.6%, respectively. Chemistry profile Statistical analysis
and 1000 h after an overnight 12-h fast. testing was done on the VITROS 5,1 FS Our primary outcome was an observed
The study procedures were rescheduled if (Ortho Clinical Diagnostics). difference in FMD%. With 15 type 1
the patients had a self-monitored blood diabetic and 21 control subjects, we had
glucose $200 mg/dL or ,80 mg/dL on over 90% power to detect a 25% differ-
FMD
the morning of the study day. Each subject ence in FMD% between groups at a =
High-resolution ultrasound (GE Pro
had breakfast after the completion of all 0.05. Data are expressed in means 6 SE,
Logiq 500) was used to assess reactive
studies. Children with type 1 diabetes unless stated otherwise. Differences in the
hyperemia (FMD) in each subject after
received their bolus insulin aspart doses means were evaluated with the Student
they had remained in the supine position
according to their home regimen. t test. Univariate associations between
for 10 min in a stable room temperature
Complete blood-count testing was the study variables were estimated by cal-
(14,15). The methods for determining,
done on the Abbott automated Cell-Dyn culating the Spearman rank correlation
analyzing, and reporting the FMD per-
instrument (probably the 4000 model). because some of the data were skewed
Fasting plasma glucose (FPG) concentra- centage (FMD%) in the brachial artery or nonnormal. All statistical analyses
as a surrogate of endothelial function
tions were measured with a Glucose were performed using the SAS 9 statistical
were performed in our vascular labora-
Analyzer II (Beckman Instruments, Brea, analysis system (SAS Institute, Cary, NC).
CA), using a glucose oxidase procedure. tory, as previously described (16). A P value ,0.05 was defined as statisti-
Replicate readings were repeated to cally significant.
within 3 mg/dL in triplicates. Fasting Carotid artery studies
plasma triglyceride, total cholesterol, All studies were performed according to a RESULTS—Table 1 summarizes the de-
HDL cholesterol, and LDL cholesterol predetermined, standardized scanning mographic, metabolic, and vascular char-
levels were determined by spectrophotom- protocol for the far wall of the common acteristics of the participants. Children
etry using kits from Stanbio Laboratory carotid artery, as described previously with type 1 diabetes had higher FPG,
(San Antonio, TX), Roche-Boehringer (17). A longitudinal section of the com- HbA1c, and hs-CRP (P , 0.05). Interest-
(Indianapolis, IN), Roche-Boehringer mon carotid artery 1 cm proximal to the ingly, the overall plasma HDL cholesterol
(after phosphotungstic acid/MgCl 2 pre-
cipitation), and Trinity Biotech (Berkeley
Heights, NJ), respectively. All determina- Table 1—Characteristics of preadolescents with type 1 diabetes and healthy
tions were performed in triplicates. Quality control subjects
controls were performed to assure stability
and reliability of the assays. The intra-assay Type 1 diabetic Control
and interassay coefficients of variation Parameters subjects subjects P
(CVs) for the lipid analyses were 4.7 and
5.3% for triglycerides, 5.5 and 6.7% for n 21 15
cholesterol, 5.7 and 6.1% for HDL choles- Age (years) 8.3 6 0.3 7.6 6 0.3 0.118
terol, and 6.9 and 7.5% for LDL choles- Sex (% female) 57.1 60.0 0.857
terol, respectively. BMI z score 0.52 6 0.19 0.31 6 0.37 0.588
hs-CRP was determined using a solid- Mean SBP (mmHg) 97.4 6 1.8 90.4 6 3.2 0.053
phase enzyme-linked immunosorbent as- Mean DBP (mmHg) 58.1 6 1.3 56.7 6 1.7 0.510
say from MP Biomedicals (West Chester, FPG (mg/dL) 173.4 6 7.9 81.4 6 1.7 ,0.0001*
PA), with a sensitivity of 0.1 mg/L and an HbA1c (%) 8.0 6 0.2 5.0 6 0.1 ,0.0001*
intraassay CV ranging from 4.1 to 2.3% Triglycerides (mg/dL) 61.8 6 7.3 64.0 6 8.3 0.844
with increasing concentrations. Plasma Cholesterol (mg/dL) 168.6 6 5.8 160.9 6 7.2 0.406
fibrinogen was determined by the Clauss LDL cholesterol (mg/dL) 104.9 6 6.5 103.3 6 5.6 0.861
method (Quest Diagnostics, Nichols In- HDL cholesterol (mg/dL) 51.4 6 3.1 44.6 6 2.2 0.108
stitute, San Clemente, CA), with intra- Cholesterol–to–HDL cholesterol 3.5 6 0.2 3.7 6 0.2 0.0056†
assay and interassay CVs of 2.6 and 4.2%, hs-CRP (mg/L) 1.8 6 0.4 0.70 6 0.20 0.036‡
respectively. Homocysteine (mmol/L) 4.3 6 0.30 3.9 6 0.40 0.419
A homocysteine assay was performed RBC folate (ng/mL) 353.6 6 13.5 331.1 6 19.0 0.327
by using the Siemens Immulite platform Fibrinogen (mg/dL) 313.0 6 11.7 301.8 6 11.9 0.517
(chemiluminescence) and the Immulite Data are means 6 SE. *P , 0.0001; †P , 0.01; ‡P , 0.05.

682 DIABETES CARE, VOLUME 34, MARCH 2011 care.diabetesjournals.org

 Babar and Associates

concentration tended to be higher among 191 6 157%; P = 0.84). In addition, the Relationships between measures of
type 1 diabetic subjects than control sub- type 1 diabetic and control groups had vascular homeostasis and systemic
jects, whereas the cholesterol-to-HDL similar c-IMT (0.48 6 0.01 mm vs. inflammation
ratio was significantly lower in type 1 di- 0.48 6 0.02 mm; P = 0.976) (Fig. 1B). Table 3 summarizes the Spearman corre-
abetic subjects than in control subjects. Table 2 summarizes the clinical and lation between FMD% and other clinical
There were no significant differences in vascular characteristics of preadolescents and biochemical parameters among type
BMI, BMI z score, mean systolic (SBP) with type 1 diabetes stratified according 1 diabetic children. We noted positive
and diastolic (DBP) blood pressure, tri- to glycemic control. Children with sub- correlations between FMD% and HbA1c
glycerides, total cholesterol, LDL choles- optimal glycemic control had higher (r = 0.47, P = 0.033) and FMD% and 2-
terol, homocysteine, RBC folate, and 2-week mean blood glucose and SDs, week blood glucose SD (r = 0.50, P =
fibrinogen between type 1 diabetic and HbA1c, and FMD% compared with those 0.021), adjusted for diabetes duration.
control subjects. with optimal glycemic control (P , 0.05). However, there was no correlation be-
There were no significant differences in tween FMD% and HbA 1c and 2-week
duration of diabetes, insulin requirement, blood glucose SD among control subjects
Measures of vascular homeostasis in insulin regimen, FPG, rate of moderate (data not shown). There were no signifi-
prepubescent children with type 1
hypoglycemia, BMI z score, mean SBP cant correlations between FMD% and
diabetes
and DBP, lipid profile, hs-CRP, homocys- BMI z score, diabetes duration, FPG,
As shown in Fig. 1A, children with type 1
teine, RBC folate, fibrinogen, and c-IMT plasma lipids, hs-CRP, homocysteine,
diabetes displayed lower brachial artery
between subgroups. Also, there were no RBC folate, fibrinogen, mean c-IMT,
FMD% change than control subjects
sex (male vs. female) differences with re- mean SBP and DBP, and baseline brachial
(7.1 6 0.8% vs. 9.8 6 1.1%; P = 0.04).
gard to glycemic control (8.1 6 0.3% vs. artery diameter.
Although the overall brachial diameter in
7.9 6 0.3%; P = 0.649) control and FMD% Table 4 summarizes the Spearman
the type 1 diabetic group was larger than
(8.0 6 0.9% vs. 6.4 6 1.21%; P = 0.359) rank correlations between c-IMT and
that of the control group (2.4 6 0.1 mm
and glucose variability and hypoglycemia clinical and biochemical parameters
vs. 2.2 6 0.1 mm; P = 0.04), there was no rates (data not shown). among type 1 diabetic children. There
significant correlation between brachial
diameter and FMD% in the study group
(r = 0.054, P = 0.841), which suggests that
group differences in baseline diameter did Table 2—Characteristics of preadolescents with type 1 diabetes according to glycemic
not significantly influence this relation- control
ship. There were no statistically signifi-
cant differences in peak hyperemic flow
between the two groups (249 6 270% vs. Optimal Suboptimal
Parameters (HbA1c ,8.0%) (HbA1c $8.0%) P
n 11 10
Age (years) 8.5 6 0.3 8.1 6 0.4 0.438
Sex (% female) 72.7 40.0 0.287
Diabetes duration 4.8 6 0.6 3.8 6 0.4 0.190
Insulin regimen (% CSII) 63.6 30 0.72
Insulin dose (units z kg21 z day21) 0.77 6 0.17 0.88 6 0.19 0.670
BMI (kg/m2) 17.5 6 0.5 17.6 6 1.2 0.937
BMI z score 0.56 6 0.22 0.47 6 0.32 0.816
Mean SBP (mmHg) 100.3 6 1.7 95.6 6 3.2 0.198
Mean DBP (mmHg) 57.4 6 1.9 59.9 6 1.5 0.321
FPG (mg/dL) 169.9 6 11.3 176.3 6 12.7 0.709
2-week MBG (mg/dL) 174.8 6 8.1 231.7 6 12.8 0.001*
2-week blood glucose SD (mg/dL) 43.1 6 3.6 66.6 6 8.1 0.013†
Hypoglycemia rate/week
(blood glucose ,60 mg/dL) 1.7 6 0.3 1.8 6 0.3 0.817
HbA1c (%) 7.4 6 0.1 8.7 6 0.3 0.0004‡
Triglycerides (mg/dL) 58.0 6 5.6 66.2 6 14.7 0.595
Cholesterol (mg/dL) 163.6 6 6.8 176.1 6 9.4 0.288
LDL cholesterol (mg/dL) 103.0 6 7.9 112.4 6 9.6 0.455
HDL cholesterol (mg/dL) 49.2 6 2.9 50.4 6 4.9 0.831
Cholesterol–to–HDL cholesterol 3.5 6 0.3 3.8 6 0.4 0.551
hs-CRP (mg/L) 2.5 6 0.7 1.3 6 0.40 0.163
Homocysteine (mmol/L) 4.4 6 0.30 4.2 6 0.40 0.690
RBC folate (ng/mL) 349.5 6 11.8 353.3 6 22.3 0.878
Figure 1—Evaluation of endothelial function Fibrinogen (mg/dL) 333.1 6 18.4 294.9 6 10.9 0.097
in type 1 diabetic (T1DM) and control sub- FMD% 5.5 6 0.9 9.1 6 1.1 0.019†
jects. A: Change in FMD%, *P = 0.04. B: c-IMT c-IMT (mm) 0.48 6 0.02 0.46 6 0.02 0.497
in type 1 diabetic and control subjects, P = 0.98. Data are means 6 SE. MBG, mean blood glucose. *P , 0.01; †P , 0.02; ‡P , 0.001.

care.diabetesjournals.org DIABETES CARE, VOLUME 34, MARCH 2011 683

Endothelial function in children with diabetes

Table 3—Correlation between FMD% and Table 4—Correlation between c-IMT and differences in the time of exposure to type 1
subject characteristics subject characteristics diabetes and chronic glycemic control (19).
Interestingly, our data show a trend
Spearman correlation Spearman correlation toward increased HDL cholesterol levels
and significantly reduced cholesterol–to–
FMD% Correlation P Mean c-IMT Correlation P HDL cholesterol ratios in preadolescent
Age at study 0.165 0.473 Age at study 20.138 0.549 children with type 1 diabetes. These
BMI z score 20.131 0.571 BMI z score 20.050 0.827 data are in line with two prior reports
Diabetes duration 0.022 0.924 Diabetes duration 20.160 0.487 (8,10) in older children with type 1 dia-
FPG 0.271 0.233 FPG 0.162 0.473 betes. Although higher plasma HDL cho-
2-week BGSD 0.50 0.021* 2-week BGSD 0.111 0.631 lesterol levels are widely thought to be
HbA1c 0.470 0.033† HbA1c 20.138 0.549 atheroprotective, in the setting of type 1
Triglycerides 20.287 0.205 Triglycerides 0.009 0.966 diabetes, HDL cholesterol may be dys-
Cholesterol 20.07 0.754 Cholesterol 20.166 0.473 functional in combating the adverse,
LDL cholesterol 20.177 0.441 LDL cholesterol 20.259 0.255 proinflammatory, and proatherogenic ef-
HDL cholesterol 0.351 0.119 HDL cholesterol 20.183 0.425 fects of oxidized LDL cholesterol (20).
Cholesterol–to–HDL Cholesterol–to–HDL The presence of dysfunctional HDL cho-
cholesterol 20.349 0.121 cholesterol 0.017 0.993 lesterol would make type 1 diabetic sub-
hs-CRP 20.421 0.092 hs-CRP 20.272 0.289 jects more vulnerable to oxidative
Homocysteine 20.497 0.070 Homocysteine 20.112 0.702 vascular damage despite higher absolute
RBC folate 20.044 0.871 RBC folate 20.117 0.664 levels. Additional work is necessary to elu-
Fibrinogen 20.129 0.597 Fibrinogen 20.343 0.151 cidate whether this mechanism is at work
Mean c-IMT 0.159 0.491 FMD% 0.159 0.490 in the development of endothelial dysfunc-
Mean SBP 20.084 0.738 Mean SBP 20.133 0.596 tion in children with type 1 diabetes.
Mean DBP 20.148 0.558 Mean DBP 20.271 0.277 In addition, we observed positive
Baseline brachial BGSD, blood glucose SD.
correlations between FMD% and HbA1c
artery diameter 0.054 0.814 and FMD% and glucose variability, as
BGSD, blood glucose SD. *P = 0.021; †P = 0.033.
measured by blood glucose SD. One pre-
vious (9) study has reported a positive
that circulating levels of hs-CRP were correlation between HbA1c and reactive
higher in preadolescents with type 1 di- hyperemia. Although mechanistic links
were no significant correlations between abetes relative to control subjects. These between glucose variability and glycemic
c-IMT and BMI z score, diabetes duration, findings are consistent with previous control and endothelial function related
FPG, HbA1c, plasma lipids, hs-CRP, homo- work (11,18) that demonstrated in- to changes in oxidative stress levels have
cysteine, RBC folate, fibrinogen, FMD%, creased systemic inflammation in ado- been suggested in diabetic subjects
and mean SBP and DBP. lescents with type 1 diabetes. Taken (21,22), we were not able to identify a
together, our data extend previous find- correlation between either HbA1c or glu-
CONCLUSIONS—We observed sig- ings in type 1 diabetes adolescents to pre- cose variability and FMD% when looking
nificant increased levels of systemic vas- adolescents and suggest that conditions at diabetic and nondiabetic children sep-
cular inflammation and impaired for the early clinical manifestation of ath- arately. Therefore, the observed correla-
endothelial function, as measured by erosclerosis are evident among very tions between FMD% and glycemic
FMD, in type 1 diabetic preadolescent young children with type 1 diabetes. control and variability may not represent a
children compared with age-matched Several studies (5,10–13) in children true physiological relationship. Mecha-
control siblings. However, there were no and adolescents with type 1 diabetes have nistic studies evaluating the relationship
differences in c-IMT between the two consistently reported increased c-IMT between short- and long-term glycemic
groups. These data suggest that adverse compared with healthy control subjects. control and endothelial function in chil-
changes in vascular homeostasis observed However, we did not observe any signifi- dren with type 1 diabetes are needed to
in preadolescent children with type 1 cant difference in c-IMT in preadolescent better elucidate the biological plausibility
diabetes are evident during the earliest children with type 1 diabetes compared of these findings.
stages of their life, heralding future car- with control subjects. The average ages, Our study has several limitations.
diovascular risk in this population. duration of diabetes, and HbA1c values First, the small sample size of the cohort
Previous published data (5,7,8) in these studies were 11.8–15.0 years, may have underpowered the study to
clearly demonstrate that vascular homeo- 3.8–6.8 years, and 7.5–8.6%, respec- detect alterations in some of the vascular
stasis is disrupted in type 1 diabetes. tively. These studies enrolled older pop- biomarkers. Additional studies to corrob-
Singh et al. (5) observed that FMD% ulations with either a longer duration of orate our findings and extend them to
was significantly impaired in a group of type 1 diabetes and/or poorer glycemic other vascular biomarkers are necessary. In
adolescents with a mean age of 15 years. control compared with our study cohort addition, we did not administer nitroglyc-
Also, other studies (7,8) have demon- (5,12). The absence of a c-IMT difference erin to our preadolescents for assessment of
strated impaired endothelial function in between type 1 diabetic and control sub- smooth-muscle reactivity. Although differ-
children and adolescents with type 1 di- jects in our study may be secondary to the ences in smooth-muscle reactivity between
abetes compared with age-matched significantly younger age and prepubertal groups cannot be completely excluded,
healthy control subjects. We also found status of our study population as well as FMD% is a well-established measure of

684 DIABETES CARE, VOLUME 34, MARCH 2011 care.diabetesjournals.org

 Babar and Associates

endothelium-dependent vasodilatation (23). type 1 diabetics. Eur J Vasc Endovasc Surg population-based study. Diabetes Care
Balanced against these limitations, we re- 2009;37:611–615 2010;33:2043–2048
port our novel findings of adverse vascular 3. Widlansky ME, Gokce N, Keaney JF Jr, 14. Celermajer DS, Sorensen KE, Gooch VM,
effects of type 1 diabetes in preadolescents. Vita JA. The clinical implications of en- et al. Non-invasive detection of endothe-
dothelial dysfunction. J Am Coll Cardiol lial dysfunction in children and adults at
In conclusion, our data suggest that
2003;42:1149–1160 risk of atherosclerosis. Lancet 1992;340:
vascular endothelial dysfunction and sys- 4. Berenson GS, Wattigney WA, Tracy RE, 1111–1115
temic inflammation are present in preado- et al. Atherosclerosis of the aorta and 15. Järvisalo MJ, Rönnemaa T, Volanen I, et al.
lescents with type 1 diabetes. Long-term coronary arteries and cardiovascular risk Brachial artery dilatation responses in
prospective studies are needed to evaluate factors in persons aged 6 to 30 years and healthy children and adolescents. Am J
the progression of vascular changes in re- studied at necropsy: the Bogalusa Heart Physiol Heart Circ Physiol 2002;282:
lation to duration of diabetes, glycemic Study. Am J Cardiol 1992;70:851–858 H87–H92
control, and progression through the stages 5. Singh TP, Groehn H, Kazmers A. Vascular 16. Kizhakekuttu TJ, Gutterman DD, Phillips
of puberty. function and carotid intimal-medial thick- SA, et al. Measuring FMD in the brachial
ness in children with insulin-dependent artery: how important is QRS-gating? J
diabetes mellitus. J Am Coll Cardiol 2003; Appl Physiol 2010;109:959–965
41:661–665 17. Aggoun Y, Szezepanski I, Bonnet D.
Acknowledgments—Funding support for
6. Järvisalo MJ, Jartti L, Näntö-Salonen K, et al. Noninvasive assessment of arterial stiff-
this study was received from the National In-
Increased aortic intima-media thickness: ness and risk of atherosclerotic events in
stitutes of Health General Clinical Research
a marker of preclinical atherosclerosis in children. Pediatr Res 2005;58:173–178
Center Grant (M01-RR-00058) to the Medical
high-risk children. Circulation 2001;104: 18. Snell-Bergeon JK, West NA, Mayer-Davis
College of Wisconsin. M.E.W. is supported by
2943–2947 EJ, et al. Inflammatory markers are in-
1K23HL089326, the Elsa Shoeneich Medical
7. Järvisalo MJ, Raitakari M, Toikka JO, et al. creased in youth with type 1 diabetes: the
Research Fund (Greater Milwaukee Founda-
Endothelial dysfunction and increased SEARCH Case-Control Study. J Clin En-
tion), and a T. Franklin Williams Scholars
arterial intima-media thickness in chil- docrinol Metab 2010;95:2868–2876
Award provided by Atlantic Philanthropies,
dren with type 1 diabetes. Circulation 19. Elhadd TA, Khan F, Kirk G, et al. In-
the American Heart Association (Grant-in-Aid
2004;109:1750–1755 fluence of puberty on endothelial dys-
10GRNT3880044), the John A. Hartford
8. Jin SM, Noh CI, Yang SW, et al. Endo- function and oxidative stress in young
Foundation, and the Association of Specialty
thelial dysfunction and microvascular patients with type 1 diabetes. Diabetes
Physicians.
complications in type 1 diabetes mellitus. Care 1998;21:1990–1996
G.S.B. researched data and wrote the man-
J Korean Med Sci 2008;23:77–82 20. Perségol L, Foissac M, Lagrost L, et al.
uscript. H.Z. wrote and reviewed the manu-
9. Ladeia AM, Ladeia-Frota C, Pinho L, HDL particles from type 1 diabetic pa-
script. M.E.W. researched data, oversaw the
Stefanelli E, Adan L. Endothelial dysfunc- tients are unable to reverse the inhibitory
sonographic data acquisition, and edited the
tion is correlated with microalbuminuria effect of oxidised LDL on endothelium-
manuscript. E.D. researched data and per-
in children with short-duration type 1 di- dependent vasorelaxation. Diabetologia
formed sonographic vascular studies. R.G.H.
abetes. Diabetes Care 2005;28:2048–2050 2007;50:2384–2387
helped design and perform the biostatistical
10. Stakos DA, Schuster DP, Sparks EA, 21. Ceriello A, Esposito K, Piconi L, et al.
analyses and contributed to the manuscript.
Wooley CF, Osei K, Boudoulas H. Car- Oscillating glucose is more deleterious to
M.D. researched data and performed bio-
diovascular effects of type 1 diabetes endothelial function and oxidative stress
statistical analyses. R.A. researched data and
mellitus in children. Angiology 2005;56: than mean glucose in normal and type 2
wrote and edited the manuscript.
311–317 diabetic patients. Diabetes 2008;57:1349–
Parts of this work were presented in abstract
11. Schwab KO, Doerfer J, Krebs A, et al. Early 1354
form (no. P3-438, 2009) at the 91st Annual
atherosclerosis in childhood type 1 di- 22. Wentholt IM, Kulik W, Michels RP,
Meeting of the Endocrine Society, Washington,
abetes: role of raised systolic blood pres- Hoekstra JB, DeVries JH. Glucose fluctu-
D.C., 13–10 June 2010.
sure in the absence of dyslipidaemia. Eur J ations and activation of oxidative stress in
Pediatr 2007;166:541–548 patients with type 1 diabetes. Diabetologia
12. Rabago Rodriguez R, Gómez-Díaz RA, 2008;51:183–190
References Tanus Haj J, et al. Carotid intima-media 23. Corretti MC, Anderson TJ, Benjamin EJ,
1. Laing SP, Swerdlow AJ, Slater SD, et al. thickness in pediatric type 1 diabetic pa- et al. Guidelines for the ultrasound as-
Mortality from heart disease in a cohort of tients. Diabetes Care 2007;30:2599–2602 sessment of endothelial-dependent flow-
23,000 patients with insulin-treated di- 13. Margeirsdottir HD, Stensaeth KH, Larsen JR, mediated vasodilation of the brachial
abetes. Diabetologia 2003;46:760–765 Brunborg C, Dahl-Jørgensen K. Early artery: a report of the International Brachial
2. Hurks R, Eisinger MJ, Goovaerts I, et al. signs of atherosclerosis in diabetic chil- Artery Reactivity Task Force. J Am Coll
Early endothelial dysfunction in young dren on intensive insulin treatment: a Cardiol 2002;39:257–265

care.diabetesjournals.org DIABETES CARE, VOLUME 34, MARCH 2011 685