Atherosclerosis 237 (2014) 129e134
Contents lists available at ScienceDirect
Atherosclerosis journal homepage: www.elsevier.com/locate/atherosclerosis
Blood pressure disturbances and endothelial dysfunction markers in children and adolescents with type 1 diabetes L. Machnica a, *, G. Deja b, J. Polanska c, P. Jarosz-Chobot b a
Upper-Silesia Child Health Center, ul. Medykow 16, 40-752 Katowice, Poland Dept. of Pediatrics, Endocrinology and Diabetes, Medical University of Silesia, ul. Medykow 16, 40-752 Katowice, Poland c Faculty of Automatic Control, Electronics and Computer Science, Silesian University of Technology, ul. Akademicka 16, 44-101 Gliwice, Poland b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 23 February 2014 Received in revised form 16 July 2014 Accepted 7 September 2014 Available online 8 September 2014
Objective: Being the earliest step on the way to atherosclerosis, endothelial dysfunction is particularly escalated in diabetes. This study aimed at assessing endothelial dysfunction and blood pressure disturbances in young patients with type 1 diabetes mellitus (T1DM) and defining their interrelations. Methods: The study group comprised 52 children and adolescents aged 14.07 ± 3.03 years, with T1DM duration 5.13 ± 2.18 years. 20 healthy controls with similar age and sex distribution were included. Chosen serum biochemical markers of endothelial damage: intercellular adhesion molecule-1 (sICAM-1), vascular cell adhesion molecule-1 (sVCAM-1), sE-selectin, tumor necrosis factor-alpha (TNF-a), interleukin-6 (IL-6) as well as ambulatory blood pressure monitoring (ABPM) were performed in all subjects. Results: Patients with T1DM displayed significantly higher concentrations of chosen markers of endothelial dysfunction compared to controls (sVCAM-1 (ng/ml): 951.56 ± 330.68 vs. 710.35 ± 162.12, TNF-a (pg/ml): 16.63 ± 8.32 vs. 9.41 ± 4.23, IL-6 (pg/ml): 3.38 ± 1.31 vs. 2.45 ± 0.81; p < 0.05). Within the study group subjects with an abnormal ABPM reading had significantly higher concentrations of sEselectin compared with subjects with normal ABPM (in ng/ml: 45.71 ± 15.63 vs. 32.42 ± 11.95; p < 0.01). The study revealed a significant positive correlation between sE-selectin and systolic as well as diastolic pressure loads during the day period (respectively: r ¼ 0.46, r ¼ 0.60; p < 0.01). Conclusions: Endothelium dysfunction may be present early in the course of T1DM in children and adolescents. It seems to be related with blood pressure disturbances which highlights the need to intensify treatment in this group of patients. © 2014 Elsevier Ireland Ltd. All rights reserved.
Keywords: Type 1 diabetes Endothelial dysfunction Blood pressure ABPM
1. Introduction Endothelium forms an internal lining of all vessels in human body. Being strategically situated between the blood stream and surrounding tissues it has got numerous functions crucial for maintaining intravascular homeostasis [1]. Endothelial cells play an important role in vascular tone regulation, hemostasis and firinolysis [2e4]. With the negative charge of its intact surface and produced substances this tissue actively resists thrombosis [3,4]. Another essential feature of endothelium is the interaction with circulating white blood cells facilitating their transmigration through the vessel wall [2,5,6]. In diabetes hyperglycemia and related pathological biochemical processes trigger damage to the endothelial cells causing their dysfunction. The dysfunctional
* Corresponding author. Present address: Lukasz Machnica, os. Bohaterow Wrzesnia 81/59, 31-620 Krakow, Poland. E-mail address: [email protected] (L. Machnica). http://dx.doi.org/10.1016/j.atherosclerosis.2014.09.006 0021-9150/© 2014 Elsevier Ireland Ltd. All rights reserved.
endothelium adopts prothrombotic, proinflammatory and vasoconstrictive phenotype promoting the development of atherosclerosis [2,3,7]. Many different measures of endothelium status assessment are available. One of them is a constantly increasing group of biochemical markers like intercellular adhesion molecule1 (sICAM-1), vascular cell adhesion molecule-1 (sVCAM-1), von Willebrandt factor (vWF), plasminogen activator inhibitor-1 (PAi1), tissue plasminogen activator (TPA), thrombomodulin (sTM), tumor necrosis factor-alpha (TNF-a), interleukin-6 (IL-6), interleukin-1 (IL-1), high-sensitive C-reactive protein (hsCRP), endothelin-1 (ET-1). A growing number of studies showed elevated levels of those substances even in children with diabetes [8e11]. Hypertension (HT) is a classical risk factor of the peripheral vessels disease. Its prevalence is known to be 1.5e3 times higher in patients with diabetes than in general population [12,13]. It was previously thought to occur mostly in adults but similar data come from developmental population as well [14e16]. HT produces shear
130
L. Machnica et al. / Atherosclerosis 237 (2014) 129e134
stress to the endothelial cells damaging them physically and causing endothelial dysfunction [1,2,17]. The accumulation of different endothelium-hostile pathological factors in diabetes makes it possible for the endothelial dysfunction to appear even in children. Most of the existing data come from an adult population of patients with type 2 diabetes. Knowledge about endothelium dysfunction in children with type 1 diabetes (T1DM) is still limited. The available studies usually focus on single markers of endothelial dysfunction and vary in results. Moreover data on the influence of high blood pressure on the endothelium status in children with type 1 diabetes are scarce. Only few studies focus on this problem, mostly not analyzing the blood pressure profiles of the studied groups in a complex way [8,20]. The aim of this study was to assess endothelial function in children and adolescents with type 1 diabetes in relation to blood pressure. We hypothesized that the endothelium dysfunction could be connected with blood pressure disturbances. 2. Subjects, materials and methods Study group consisted of 52 children and adolescents with type 1 diabetes (33 girls) with the mean age 14.07 ± 3.03 years, mean diabetes duration 5.13 ± 2.18 years and mean HbA1c e 7.18 ± 1.04% (55 ± 10.8 mmol/mol). The subjects were hospitalized in the Department of Pediatrics, Endocrinology and Diabetes of the Medical University of Silesia in Katowice (Poland) during 2007e2010. The medical records of those individuals from the Diabetes Outpatient Clinic of the Upper Silesia Child Health Center in Katowice were also used in the analysis. The inclusion criteria were: age between 8 and 18 years, duration of diabetes at least 3 years for prepubertal children and at least 2 years for pubertal individuals, no known chronic disease (especially Hashimoto or celiac disease), no acute inflammation for 3 weeks before and at the day of the examination, no drugs taken except insulin, no smoking, informed consent of the guardians and participants. A control group comprised 20 healthy peers (12 girls) with similar age and sex distribution as in the study group. Volunteers were invited to participate in the study after completing the same inclusion criteria as in the study group (apart from diabetes). The subjects were recruited from friends of patients with diabetes. The number of controls was limited because of the substantial difficulties with the recruitment and limitations set by Bioethical Committee. Medical history including data about previous and current course of diabetes treatment (i.e. mean HbA1c from the diabetes onset done once every three months, daily insulin dose in units/kg body weight/24 h), physical examination and laboratory tests were performed in the study group. Fasting concentrations of chosen adhesion molecules (sICAM-1, sVCAM-1, sE-selectin) and proinflammatory cytokines (TNF-a, IL-6) were tested using commercial kits (R&D Systems). Microalbuminuria was determined by means of 10-h night collection samples and immunologic chemiluminescence method. Thyroid function tests (TSH, fT4), thyroid antibodies: thyroid peroxidase antibodies (ATA) and thyroglobulin antibodies (ATG) as well as tissue transglutaminase antibodies (IgAtTG) in order to fulfill inclusion criteria were also evaluated. Moreover a 24-h blood pressure monitoring (ABPM) using Spacelabs 90217 device with age-fitted cuffs was performed. It was set to measure blood pressure (BP) every 20 min during the day period (6:00 a.m.e10:00 p.m.) and every 30 min during the night period (10:00 p.m.e6:00 a.m.). The cut-off values of blood pressure were set on the basis of the centile charts for age, sex and height and according to the guidelines of “The fourth report on the diagnosis, evaluation, and treatment of high blood pressure in children and adolescents” [21]. Patients whose systolic and/or diastolic blood
Table 1 Clinical characteristics of the study and control group. Study group N ¼ 52 14.07 ± 3.03 [13.23 ÷ 14.92] Height (cm) 164.49 ± 16.37 [159.93 ÷ 169.05] Body weight (kg) 58.15 ± 15.26 [53.90 ÷ 62.40] 21.02 ± 2.67 BMI (kg/m2) [20.28 ÷ 21.77] Age at onset of diabetes 9.02 ± 0.60 (years) [8.82 ÷ 9.22] Diabetes duration 5.09 ± 1.97 (years) [4.39 ÷ 5.79] Mean HbA1c (%/mmol/ 7.17 ± 0.99/ mol) 55 ± 10.8 [6.82 ÷ 7.52] 0.84 ± 0.18 Daily insulin requirement (units/kg [0.78 ÷ 0.90] body weight/24 h) Age (years)
Control group N ¼ 20
Statistical significance of differences
13.09 ± 3.05 [11.66 ÷ 14.51] 157.38 ± 15.10 [150.31 ÷ 164.44] 50.64 ± 19.44 [41.54 ÷ 59.73] 19.81 ± 4.09 [17.89 ÷ 21.72]
p ¼ NS p ¼ NS p < 0.01 p < 0.05
Data are presented as mean ± standard deviation (SD) and [95% CI].
pressure load during any period was beyond 40% were considered BP-abnormal. Those with blood pressure fall during the night lower than 10% were classified as non-dippers. HbA1c in our Department as well as in aforementioned Diabetes Outpatient Clinic is measured using the HPLC method according to DCCT standards. After physical examination the same laboratory tests including: TSH, fT4, ATA, ATG, IgAtTG, sICAM-1, sVCAM-1, sE-selectin, TNF-a, IL-6 and ABPM were performed in the control group (except microalbuminuria). The following descriptive statistics and their 95% confidence intervals (noted as 95%CI) were estimated per every continuous variable under investigation: mean, standard deviation, median, lower and upper quartile, minimum and maximum value. Lilliefors test was applied to verify the hypothesis on distribution normality, while homogeneity of variances was checked with F test. Hypotheses on equality of variances and/or median values were verified by t test (for Gaussian distributions) or ManneWhitney U test (in case of non-Gaussian distributions). The association between variables was measured by Pearson's or Spearman's correlation coefficients (for Gaussian and non-Gaussian distributions respectively) and the tests on their significance. The logistic regression algorithm was applied to find the discriminant function distinguishing between patients and healthy controls. Due to the small sample sizes, the forward selection algorithm was applied to the prefiltered set of features accompanied by Akaike information criterion (AIC) and likelihood ratio test for model selection. The initial set of prefiltered features was constructed of those with p-value less than 0.2 in SISO comparative analysis. The linear regression technique was used to construct an adjusted model of interactions between a vascular cell adhesion molecule-1 marker and some clinical parameters. In case of the skewed distributions of the analyzed variables, the BoxeCox power transformation was applied to correct for the departures from normality. The age dependent standardization was used to the anthropometric measurements (patient height and weight) in the form SDS ¼ (X Xsr)/SD per every age group independently. The F statistics was used to assess the quality of model fitting. The results were assumed as statistically significant if p-values were less than 0.05. The study gained a positive opinion of the Bioethical Committee of the Medical University of Silesia in Katowice no. NN-6501-82/07.
L. Machnica et al. / Atherosclerosis 237 (2014) 129e134
3. Results
Table 3 Comparative analysis of ABPM results (study versus control group).
The anthropometric and clinical characteristics of the study and control groups are summarized in Table 1. The analysis of chosen biochemical markers of endothelial dysfunction revealed significantly higher concentrations of sVCAM-1, TNF-a and IL-6 in the study group and significantly higher sE-selectin concentrations in the control group (Table 2). No significant differences in the concentrations of investigated endothelial markers in terms of sex were found in the study group. On the contrary boys from the control group had significantly higher concentrations of sVCAM-1 than girls (respectively mean ± SD in ng/dl: 816.57 ± 114.02 [711.12 ÷ 922.02] versus 636.00 ± 152.12 [527.18 ÷ 744.82]; p < 0.05). The comparison of 24-h blood pressure monitoring results between patients with diabetes and controls displayed significantly higher values of diastolic blood pressure during the whole record period (day þ night) as well as higher diastolic blood pressure load during the day period in the study group (Table 3). Based on the ABPM results (blood pressure load) the participants were divided into two subgroups (BP e normal and BP e abnormal). The comparison between subgroups revealed significantly higher concentrations of sE-selectin in BP e abnormal subgroup of the study group. Similar significant differences were observed in the control group (data not shown). Furthermore, we found significantly higher microalbuminuria value in children and adolescents with diabetes and abnormal blood pressure (Table 4). We did not reveal any significant differences in measured parameters of endothelial damage between dippers and non-dippers. The logistic regression based classifier was constructed to find the most discriminative features between study and control groups. The final regression equation obtained was as follows:
logit ðGroupÞ ¼ a0 þ a1 $ðTNF alfaÞ þ a2 $ðE_selectinÞ þ a3 $ðDBPÞ þ a4 $Sex TNF-a, E-selectin and DBP (24 h) were the variables adjusted for sex that differentiated between groups best. No significant correlations in the study group between biochemical markers of endothelial dysfunction and BMI, mean HbA1c as well as daily insulin requirement were shown. Moreover, Table 2 Comparative analysis of chosen markers of endothelial dysfunction (study versus control group).
sICAM (ng/ ml)
sVCAM (ng/ ml) TNF-a (pg/ ml)
IL-6 (pg/ml)
sE-selectin (ng/ml)
131
Study group N ¼ 52
Control group N ¼ 20
Statistical significance of differences
382.81 ± 142.15 [342.83 ÷ 422.79] 338.40 [287.50; 474.25] 951.56 ± 330.68 [856.58 ÷ 1046.55] 16.63 ± 8.32 [14.24 ÷ 19.02] 14.40 [9.99; 23.22] 3.38 ± 1.31 [2.99 ÷ 3.78] 3.17 [2.12; 4.55] 42.01 ± 21.12 [36.13 ± 47.89]
417.55 ± 190.36 [328.46 ÷ 506.64] 370.00 [250.00; 631.00] 710.35 ± 162.12 [627.00 ÷ 793.71] 9.41 ± 4.23 [7.42 ÷ 11.39] 8.05 [6.85; 12.90] 2.45 ± 0.81 [2.05 ÷ 2.85] 2.41 [1.90; 2.98] 59.26 ± 31.23 [44.64 ÷ 73.87]
p ¼ 0.650a
117.71 ± 9.47 [115.04 ÷ 120.37] 66.51 ± 3.57 [65.44 ÷ 67.58] 67.00 [65.00; 69.00] SBP (day) 120.61 ± 9.58 (mmHg) [117.91 ÷ 123.30] DBP (day) 68.93 ± 4.01 (mmHg) [67.74 ÷ 70.13] 32.90 ± 27.75 SBP load [25.18 ± 40.63] (day) 23.30 (%) [10.85; 48.50] 14.23 ± 12.97 DBP load [10.58 ± 17.88] (day) 11.30 (%) [4.25; 21.95] SBP (night) 108.98 ± 9.71 (mmHg) [106.28 ± 111.68] DBP (night) 57.80 ± 4.48 (mmHg) [56.53 ± 59.07] 32.33 ± 29.37 SBP load [24.16 ± 40.51] (night) 22.70 (%) [6.70; 53.65] 9.32 ± 8.77 DBP load [6.82 ± 11.81] (night) 6.70 (%) [0.00; 13.30]
SBP (24-h) (mmHg) DBP (24-h) (mmHg)
Control group N ¼ 20
Statistical significance of differences
114.22 ± 8.29 [110.10 ÷ 118.35] 64.37 ± 4.47 [62.21 ÷ 66.53] 63.00 [61.00; 67.00] 115.94 ± 4.92 [113.32 ÷ 118.56] 66.85 ± 5.04 [64.99 ÷ 69.21] 32.20 ± 29.66 [18.32 ÷ 46.09] 20.50 [9.70; 61.15] 4.62 ± 4.13 [2.50 ÷ 6.74] 2.80 [1.70; 8.05] 107.55 ± 9.85 [102.94 ÷ 112.16] 57.10 ± 5.99 [54.30 ÷ 59.90] 18.95 ± 19.06 [9.47 ÷ 28.43] 12.00 [6.30; 26.70] 8.38 ± 7.27 [4.98 ÷ 11.78] 7.20 [0.00; 13.80]
p ¼ 0.172 p ¼ 0.038a
p ¼ 0.067 p ¼ 0.078 p ¼ 0.841a
p ¼ 0.003a
p ¼ 0.579 p ¼ 0.594 p ¼ 0.128a
p ¼ 0.879a
Data are presented as mean ± standard deviation (SD) and [95% CI] ], and median with [interquartile range] depending on the distribution. a p value for the nonparametric test.
mean HbA1c did not correlate significantly with ABPM results. The study displayed significant positive associations between chosen adhesion molecules and proinflammatory cytokines and vice versa. Additionally, concentration of sE-selectin correlated positively with systolic and diastolic blood pressure loads during the day (Table 5). We attempted to identify the variables that could influence the endothelial damage. On the basis of the literature review and univariate analyses sVCAM-1 was chosen as the dependant variable. A stepwise linear regression combined with BoxeCox power transformation applied to the analyzed parameters gave the best final model in the form:
1 ln ðVCAMÞ ¼ b0 þ b1 $ðheight_SDSÞ2 þ b2 $ IL6 þ b3 $ln ðmicroalbuminuriaÞ
4. Discussion
p < 0.001 p < 0.001
Study group N ¼ 52
a
p ¼ 0.016a
p ¼ 0.009
Data are presented as mean ± standard deviation (SD) and [95% CI], and median with [interquartile range] depending on the distribution. a p value for the nonparametric test.
The analysis of chosen biochemical markers of endothelial dysfunction showed significantly higher concentrations of sVCAM1, IL-6 and TNF-a in the study group. However, the evidence from literature is somewhat inconsistent. As highlighted before most of the available data about biochemical markers of endothelium dysfunction comes from the adult patients with type 2 diabetes. Data from patients with type 1 diabetes especially children are much more sparse. EURODIAB Prospective Complications Study in 543 young adults with type 1 diabetes revealed significantly higher concentration of sVCAM-1, sE-selectin, CRP, IL-6 and TNF-a in those with chronic complications compared to those without them [22].
132
L. Machnica et al. / Atherosclerosis 237 (2014) 129e134
Table 4 Comparative analysis of chosen markers of endothelial dysfunction and microalbuminuria in the study group on the basis of ABPM results. BP e normal N ¼ 23 sE-selectin (ng/ml) 32.42 ± 11.95 [27.59 ÷ 37.24] Microalbuminuria 4.93 ± 3.58 (ug/ml) [3.38 ÷ 6.48]
BP e abnormal N ¼ 24
Statistical significance of differences
45.71 ± 15.63 [38.95 ÷ 52.47] 7.63 ± 3.59 [6.12 ÷ 9.15]
p < 0.01 p < 0.01
Data are presented as mean ± standard deviation (SD) and [95% CI].
Table 5 Correlation coefficients (Pearson or Spearman depending on the distribution) and p values in the study group.
sICAM (ng/ml) sVCAM (ng/ml) TNF-a (pg/ml) IL-6 (pg/ml) sE-selectin (ng/ml)
Correlation coefficients
P value
sE-selectin: r ¼ 0.43 TNF-a: r ¼ 0.39 IL-6: r ¼ 0.41 sVCAM: r ¼ 0.39 IL-6: r ¼ 0.45 sVCAM: r ¼ 0.39 TNF-a: r ¼ 0.45 sICAM: r ¼ 0.43 SBP load (day): r ¼ 0.46 DBP load (day): r ¼ 0.60
p p p p p p p p p p
< < < < < < < < <
Contents lists available at ScienceDirect
Atherosclerosis journal homepage: www.elsevier.com/locate/atherosclerosis
Blood pressure disturbances and endothelial dysfunction markers in children and adolescents with type 1 diabetes L. Machnica a, *, G. Deja b, J. Polanska c, P. Jarosz-Chobot b a
Upper-Silesia Child Health Center, ul. Medykow 16, 40-752 Katowice, Poland Dept. of Pediatrics, Endocrinology and Diabetes, Medical University of Silesia, ul. Medykow 16, 40-752 Katowice, Poland c Faculty of Automatic Control, Electronics and Computer Science, Silesian University of Technology, ul. Akademicka 16, 44-101 Gliwice, Poland b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 23 February 2014 Received in revised form 16 July 2014 Accepted 7 September 2014 Available online 8 September 2014
Objective: Being the earliest step on the way to atherosclerosis, endothelial dysfunction is particularly escalated in diabetes. This study aimed at assessing endothelial dysfunction and blood pressure disturbances in young patients with type 1 diabetes mellitus (T1DM) and defining their interrelations. Methods: The study group comprised 52 children and adolescents aged 14.07 ± 3.03 years, with T1DM duration 5.13 ± 2.18 years. 20 healthy controls with similar age and sex distribution were included. Chosen serum biochemical markers of endothelial damage: intercellular adhesion molecule-1 (sICAM-1), vascular cell adhesion molecule-1 (sVCAM-1), sE-selectin, tumor necrosis factor-alpha (TNF-a), interleukin-6 (IL-6) as well as ambulatory blood pressure monitoring (ABPM) were performed in all subjects. Results: Patients with T1DM displayed significantly higher concentrations of chosen markers of endothelial dysfunction compared to controls (sVCAM-1 (ng/ml): 951.56 ± 330.68 vs. 710.35 ± 162.12, TNF-a (pg/ml): 16.63 ± 8.32 vs. 9.41 ± 4.23, IL-6 (pg/ml): 3.38 ± 1.31 vs. 2.45 ± 0.81; p < 0.05). Within the study group subjects with an abnormal ABPM reading had significantly higher concentrations of sEselectin compared with subjects with normal ABPM (in ng/ml: 45.71 ± 15.63 vs. 32.42 ± 11.95; p < 0.01). The study revealed a significant positive correlation between sE-selectin and systolic as well as diastolic pressure loads during the day period (respectively: r ¼ 0.46, r ¼ 0.60; p < 0.01). Conclusions: Endothelium dysfunction may be present early in the course of T1DM in children and adolescents. It seems to be related with blood pressure disturbances which highlights the need to intensify treatment in this group of patients. © 2014 Elsevier Ireland Ltd. All rights reserved.
Keywords: Type 1 diabetes Endothelial dysfunction Blood pressure ABPM
1. Introduction Endothelium forms an internal lining of all vessels in human body. Being strategically situated between the blood stream and surrounding tissues it has got numerous functions crucial for maintaining intravascular homeostasis [1]. Endothelial cells play an important role in vascular tone regulation, hemostasis and firinolysis [2e4]. With the negative charge of its intact surface and produced substances this tissue actively resists thrombosis [3,4]. Another essential feature of endothelium is the interaction with circulating white blood cells facilitating their transmigration through the vessel wall [2,5,6]. In diabetes hyperglycemia and related pathological biochemical processes trigger damage to the endothelial cells causing their dysfunction. The dysfunctional
* Corresponding author. Present address: Lukasz Machnica, os. Bohaterow Wrzesnia 81/59, 31-620 Krakow, Poland. E-mail address: [email protected] (L. Machnica). http://dx.doi.org/10.1016/j.atherosclerosis.2014.09.006 0021-9150/© 2014 Elsevier Ireland Ltd. All rights reserved.
endothelium adopts prothrombotic, proinflammatory and vasoconstrictive phenotype promoting the development of atherosclerosis [2,3,7]. Many different measures of endothelium status assessment are available. One of them is a constantly increasing group of biochemical markers like intercellular adhesion molecule1 (sICAM-1), vascular cell adhesion molecule-1 (sVCAM-1), von Willebrandt factor (vWF), plasminogen activator inhibitor-1 (PAi1), tissue plasminogen activator (TPA), thrombomodulin (sTM), tumor necrosis factor-alpha (TNF-a), interleukin-6 (IL-6), interleukin-1 (IL-1), high-sensitive C-reactive protein (hsCRP), endothelin-1 (ET-1). A growing number of studies showed elevated levels of those substances even in children with diabetes [8e11]. Hypertension (HT) is a classical risk factor of the peripheral vessels disease. Its prevalence is known to be 1.5e3 times higher in patients with diabetes than in general population [12,13]. It was previously thought to occur mostly in adults but similar data come from developmental population as well [14e16]. HT produces shear
130
L. Machnica et al. / Atherosclerosis 237 (2014) 129e134
stress to the endothelial cells damaging them physically and causing endothelial dysfunction [1,2,17]. The accumulation of different endothelium-hostile pathological factors in diabetes makes it possible for the endothelial dysfunction to appear even in children. Most of the existing data come from an adult population of patients with type 2 diabetes. Knowledge about endothelium dysfunction in children with type 1 diabetes (T1DM) is still limited. The available studies usually focus on single markers of endothelial dysfunction and vary in results. Moreover data on the influence of high blood pressure on the endothelium status in children with type 1 diabetes are scarce. Only few studies focus on this problem, mostly not analyzing the blood pressure profiles of the studied groups in a complex way [8,20]. The aim of this study was to assess endothelial function in children and adolescents with type 1 diabetes in relation to blood pressure. We hypothesized that the endothelium dysfunction could be connected with blood pressure disturbances. 2. Subjects, materials and methods Study group consisted of 52 children and adolescents with type 1 diabetes (33 girls) with the mean age 14.07 ± 3.03 years, mean diabetes duration 5.13 ± 2.18 years and mean HbA1c e 7.18 ± 1.04% (55 ± 10.8 mmol/mol). The subjects were hospitalized in the Department of Pediatrics, Endocrinology and Diabetes of the Medical University of Silesia in Katowice (Poland) during 2007e2010. The medical records of those individuals from the Diabetes Outpatient Clinic of the Upper Silesia Child Health Center in Katowice were also used in the analysis. The inclusion criteria were: age between 8 and 18 years, duration of diabetes at least 3 years for prepubertal children and at least 2 years for pubertal individuals, no known chronic disease (especially Hashimoto or celiac disease), no acute inflammation for 3 weeks before and at the day of the examination, no drugs taken except insulin, no smoking, informed consent of the guardians and participants. A control group comprised 20 healthy peers (12 girls) with similar age and sex distribution as in the study group. Volunteers were invited to participate in the study after completing the same inclusion criteria as in the study group (apart from diabetes). The subjects were recruited from friends of patients with diabetes. The number of controls was limited because of the substantial difficulties with the recruitment and limitations set by Bioethical Committee. Medical history including data about previous and current course of diabetes treatment (i.e. mean HbA1c from the diabetes onset done once every three months, daily insulin dose in units/kg body weight/24 h), physical examination and laboratory tests were performed in the study group. Fasting concentrations of chosen adhesion molecules (sICAM-1, sVCAM-1, sE-selectin) and proinflammatory cytokines (TNF-a, IL-6) were tested using commercial kits (R&D Systems). Microalbuminuria was determined by means of 10-h night collection samples and immunologic chemiluminescence method. Thyroid function tests (TSH, fT4), thyroid antibodies: thyroid peroxidase antibodies (ATA) and thyroglobulin antibodies (ATG) as well as tissue transglutaminase antibodies (IgAtTG) in order to fulfill inclusion criteria were also evaluated. Moreover a 24-h blood pressure monitoring (ABPM) using Spacelabs 90217 device with age-fitted cuffs was performed. It was set to measure blood pressure (BP) every 20 min during the day period (6:00 a.m.e10:00 p.m.) and every 30 min during the night period (10:00 p.m.e6:00 a.m.). The cut-off values of blood pressure were set on the basis of the centile charts for age, sex and height and according to the guidelines of “The fourth report on the diagnosis, evaluation, and treatment of high blood pressure in children and adolescents” [21]. Patients whose systolic and/or diastolic blood
Table 1 Clinical characteristics of the study and control group. Study group N ¼ 52 14.07 ± 3.03 [13.23 ÷ 14.92] Height (cm) 164.49 ± 16.37 [159.93 ÷ 169.05] Body weight (kg) 58.15 ± 15.26 [53.90 ÷ 62.40] 21.02 ± 2.67 BMI (kg/m2) [20.28 ÷ 21.77] Age at onset of diabetes 9.02 ± 0.60 (years) [8.82 ÷ 9.22] Diabetes duration 5.09 ± 1.97 (years) [4.39 ÷ 5.79] Mean HbA1c (%/mmol/ 7.17 ± 0.99/ mol) 55 ± 10.8 [6.82 ÷ 7.52] 0.84 ± 0.18 Daily insulin requirement (units/kg [0.78 ÷ 0.90] body weight/24 h) Age (years)
Control group N ¼ 20
Statistical significance of differences
13.09 ± 3.05 [11.66 ÷ 14.51] 157.38 ± 15.10 [150.31 ÷ 164.44] 50.64 ± 19.44 [41.54 ÷ 59.73] 19.81 ± 4.09 [17.89 ÷ 21.72]
p ¼ NS p ¼ NS p < 0.01 p < 0.05
Data are presented as mean ± standard deviation (SD) and [95% CI].
pressure load during any period was beyond 40% were considered BP-abnormal. Those with blood pressure fall during the night lower than 10% were classified as non-dippers. HbA1c in our Department as well as in aforementioned Diabetes Outpatient Clinic is measured using the HPLC method according to DCCT standards. After physical examination the same laboratory tests including: TSH, fT4, ATA, ATG, IgAtTG, sICAM-1, sVCAM-1, sE-selectin, TNF-a, IL-6 and ABPM were performed in the control group (except microalbuminuria). The following descriptive statistics and their 95% confidence intervals (noted as 95%CI) were estimated per every continuous variable under investigation: mean, standard deviation, median, lower and upper quartile, minimum and maximum value. Lilliefors test was applied to verify the hypothesis on distribution normality, while homogeneity of variances was checked with F test. Hypotheses on equality of variances and/or median values were verified by t test (for Gaussian distributions) or ManneWhitney U test (in case of non-Gaussian distributions). The association between variables was measured by Pearson's or Spearman's correlation coefficients (for Gaussian and non-Gaussian distributions respectively) and the tests on their significance. The logistic regression algorithm was applied to find the discriminant function distinguishing between patients and healthy controls. Due to the small sample sizes, the forward selection algorithm was applied to the prefiltered set of features accompanied by Akaike information criterion (AIC) and likelihood ratio test for model selection. The initial set of prefiltered features was constructed of those with p-value less than 0.2 in SISO comparative analysis. The linear regression technique was used to construct an adjusted model of interactions between a vascular cell adhesion molecule-1 marker and some clinical parameters. In case of the skewed distributions of the analyzed variables, the BoxeCox power transformation was applied to correct for the departures from normality. The age dependent standardization was used to the anthropometric measurements (patient height and weight) in the form SDS ¼ (X Xsr)/SD per every age group independently. The F statistics was used to assess the quality of model fitting. The results were assumed as statistically significant if p-values were less than 0.05. The study gained a positive opinion of the Bioethical Committee of the Medical University of Silesia in Katowice no. NN-6501-82/07.
L. Machnica et al. / Atherosclerosis 237 (2014) 129e134
3. Results
Table 3 Comparative analysis of ABPM results (study versus control group).
The anthropometric and clinical characteristics of the study and control groups are summarized in Table 1. The analysis of chosen biochemical markers of endothelial dysfunction revealed significantly higher concentrations of sVCAM-1, TNF-a and IL-6 in the study group and significantly higher sE-selectin concentrations in the control group (Table 2). No significant differences in the concentrations of investigated endothelial markers in terms of sex were found in the study group. On the contrary boys from the control group had significantly higher concentrations of sVCAM-1 than girls (respectively mean ± SD in ng/dl: 816.57 ± 114.02 [711.12 ÷ 922.02] versus 636.00 ± 152.12 [527.18 ÷ 744.82]; p < 0.05). The comparison of 24-h blood pressure monitoring results between patients with diabetes and controls displayed significantly higher values of diastolic blood pressure during the whole record period (day þ night) as well as higher diastolic blood pressure load during the day period in the study group (Table 3). Based on the ABPM results (blood pressure load) the participants were divided into two subgroups (BP e normal and BP e abnormal). The comparison between subgroups revealed significantly higher concentrations of sE-selectin in BP e abnormal subgroup of the study group. Similar significant differences were observed in the control group (data not shown). Furthermore, we found significantly higher microalbuminuria value in children and adolescents with diabetes and abnormal blood pressure (Table 4). We did not reveal any significant differences in measured parameters of endothelial damage between dippers and non-dippers. The logistic regression based classifier was constructed to find the most discriminative features between study and control groups. The final regression equation obtained was as follows:
logit ðGroupÞ ¼ a0 þ a1 $ðTNF alfaÞ þ a2 $ðE_selectinÞ þ a3 $ðDBPÞ þ a4 $Sex TNF-a, E-selectin and DBP (24 h) were the variables adjusted for sex that differentiated between groups best. No significant correlations in the study group between biochemical markers of endothelial dysfunction and BMI, mean HbA1c as well as daily insulin requirement were shown. Moreover, Table 2 Comparative analysis of chosen markers of endothelial dysfunction (study versus control group).
sICAM (ng/ ml)
sVCAM (ng/ ml) TNF-a (pg/ ml)
IL-6 (pg/ml)
sE-selectin (ng/ml)
131
Study group N ¼ 52
Control group N ¼ 20
Statistical significance of differences
382.81 ± 142.15 [342.83 ÷ 422.79] 338.40 [287.50; 474.25] 951.56 ± 330.68 [856.58 ÷ 1046.55] 16.63 ± 8.32 [14.24 ÷ 19.02] 14.40 [9.99; 23.22] 3.38 ± 1.31 [2.99 ÷ 3.78] 3.17 [2.12; 4.55] 42.01 ± 21.12 [36.13 ± 47.89]
417.55 ± 190.36 [328.46 ÷ 506.64] 370.00 [250.00; 631.00] 710.35 ± 162.12 [627.00 ÷ 793.71] 9.41 ± 4.23 [7.42 ÷ 11.39] 8.05 [6.85; 12.90] 2.45 ± 0.81 [2.05 ÷ 2.85] 2.41 [1.90; 2.98] 59.26 ± 31.23 [44.64 ÷ 73.87]
p ¼ 0.650a
117.71 ± 9.47 [115.04 ÷ 120.37] 66.51 ± 3.57 [65.44 ÷ 67.58] 67.00 [65.00; 69.00] SBP (day) 120.61 ± 9.58 (mmHg) [117.91 ÷ 123.30] DBP (day) 68.93 ± 4.01 (mmHg) [67.74 ÷ 70.13] 32.90 ± 27.75 SBP load [25.18 ± 40.63] (day) 23.30 (%) [10.85; 48.50] 14.23 ± 12.97 DBP load [10.58 ± 17.88] (day) 11.30 (%) [4.25; 21.95] SBP (night) 108.98 ± 9.71 (mmHg) [106.28 ± 111.68] DBP (night) 57.80 ± 4.48 (mmHg) [56.53 ± 59.07] 32.33 ± 29.37 SBP load [24.16 ± 40.51] (night) 22.70 (%) [6.70; 53.65] 9.32 ± 8.77 DBP load [6.82 ± 11.81] (night) 6.70 (%) [0.00; 13.30]
SBP (24-h) (mmHg) DBP (24-h) (mmHg)
Control group N ¼ 20
Statistical significance of differences
114.22 ± 8.29 [110.10 ÷ 118.35] 64.37 ± 4.47 [62.21 ÷ 66.53] 63.00 [61.00; 67.00] 115.94 ± 4.92 [113.32 ÷ 118.56] 66.85 ± 5.04 [64.99 ÷ 69.21] 32.20 ± 29.66 [18.32 ÷ 46.09] 20.50 [9.70; 61.15] 4.62 ± 4.13 [2.50 ÷ 6.74] 2.80 [1.70; 8.05] 107.55 ± 9.85 [102.94 ÷ 112.16] 57.10 ± 5.99 [54.30 ÷ 59.90] 18.95 ± 19.06 [9.47 ÷ 28.43] 12.00 [6.30; 26.70] 8.38 ± 7.27 [4.98 ÷ 11.78] 7.20 [0.00; 13.80]
p ¼ 0.172 p ¼ 0.038a
p ¼ 0.067 p ¼ 0.078 p ¼ 0.841a
p ¼ 0.003a
p ¼ 0.579 p ¼ 0.594 p ¼ 0.128a
p ¼ 0.879a
Data are presented as mean ± standard deviation (SD) and [95% CI] ], and median with [interquartile range] depending on the distribution. a p value for the nonparametric test.
mean HbA1c did not correlate significantly with ABPM results. The study displayed significant positive associations between chosen adhesion molecules and proinflammatory cytokines and vice versa. Additionally, concentration of sE-selectin correlated positively with systolic and diastolic blood pressure loads during the day (Table 5). We attempted to identify the variables that could influence the endothelial damage. On the basis of the literature review and univariate analyses sVCAM-1 was chosen as the dependant variable. A stepwise linear regression combined with BoxeCox power transformation applied to the analyzed parameters gave the best final model in the form:
1 ln ðVCAMÞ ¼ b0 þ b1 $ðheight_SDSÞ2 þ b2 $ IL6 þ b3 $ln ðmicroalbuminuriaÞ
4. Discussion
p < 0.001 p < 0.001
Study group N ¼ 52
a
p ¼ 0.016a
p ¼ 0.009
Data are presented as mean ± standard deviation (SD) and [95% CI], and median with [interquartile range] depending on the distribution. a p value for the nonparametric test.
The analysis of chosen biochemical markers of endothelial dysfunction showed significantly higher concentrations of sVCAM1, IL-6 and TNF-a in the study group. However, the evidence from literature is somewhat inconsistent. As highlighted before most of the available data about biochemical markers of endothelium dysfunction comes from the adult patients with type 2 diabetes. Data from patients with type 1 diabetes especially children are much more sparse. EURODIAB Prospective Complications Study in 543 young adults with type 1 diabetes revealed significantly higher concentration of sVCAM-1, sE-selectin, CRP, IL-6 and TNF-a in those with chronic complications compared to those without them [22].
132
L. Machnica et al. / Atherosclerosis 237 (2014) 129e134
Table 4 Comparative analysis of chosen markers of endothelial dysfunction and microalbuminuria in the study group on the basis of ABPM results. BP e normal N ¼ 23 sE-selectin (ng/ml) 32.42 ± 11.95 [27.59 ÷ 37.24] Microalbuminuria 4.93 ± 3.58 (ug/ml) [3.38 ÷ 6.48]
BP e abnormal N ¼ 24
Statistical significance of differences
45.71 ± 15.63 [38.95 ÷ 52.47] 7.63 ± 3.59 [6.12 ÷ 9.15]
p < 0.01 p < 0.01
Data are presented as mean ± standard deviation (SD) and [95% CI].
Table 5 Correlation coefficients (Pearson or Spearman depending on the distribution) and p values in the study group.
sICAM (ng/ml) sVCAM (ng/ml) TNF-a (pg/ml) IL-6 (pg/ml) sE-selectin (ng/ml)
Correlation coefficients
P value
sE-selectin: r ¼ 0.43 TNF-a: r ¼ 0.39 IL-6: r ¼ 0.41 sVCAM: r ¼ 0.39 IL-6: r ¼ 0.45 sVCAM: r ¼ 0.39 TNF-a: r ¼ 0.45 sICAM: r ¼ 0.43 SBP load (day): r ¼ 0.46 DBP load (day): r ¼ 0.60
p p p p p p p p p p
< < < < < < < < <
