Impact of Type 1 Diabetes and Body Weight Status on Cardiovascular Risk Factors in Adolescent Children
Sowmya Krishnan, MD, University of Oklahoma Health Science Center, Pediatrics, 1200, North Phillips Avenue, Suite 4500, Oklahoma City, OK 73104
Type 1 diabetes (T1D) is a risk factor for cardiovascular disease. However, it is unclear whether increased body weight amplifies that risk in T1D patients. This is a cross-sectional study examining the presence of cardiovascular risk factors in normal and overweight children, both with and without T1D. Sixty-six children (aged 16±2.2 years) were included in one of the following groups: (T1D and normal weight, T1D and overweight, healthy and normal weight, and healthy and overweight). A fasting blood sample was analyzed for lipid profile (triglyceride, cholesterol, high-density lipoprotein cholesterol, and low-density lipoprotein cholesterol), apolipoprotein B (apoB), and apolipoprotein C-III (apoC-III) levels. Body composition was determined by dual energy x-ray absorptiometry and vascular elasticity by HDI/Pulsewave CR-2000 (Hypertension Diagnostics, Eagan, MN). Statistical analyses examined the effect of T1D and body weight status and their interactions on cardiovascular risk parameters. In this study, the authors were unable to demonstrate an additive effect of body weight status and T1D on cardiovascular risk profile. However, subgroup analysis of patients with T1D revealed higher apoC-III levels in overweight patients with T1D (P=.0453) compared with normal-weight diabetic children. Most notably, there was a direct relationship of small artery elasticity to body weight status. This seemingly paradoxical observation supports recent data and warrants further investigation.
J Clin Hypertens (Greenwich). 2011;13:351–356. ©2010 Wiley Periodicals, Inc.
Type 1 diabetes (T1D) is a common disease of childhood and is increasing in prevalence worldwide.1 Cardiovascular (CV) disease occurs at a higher frequency and at a younger age in patients with T1D compared with the general population.2,3 Men and women with T1D have a cumulative mortality rate of 35% from coronary artery disease by the age of 55 years, compared with only 4% to 8% in nondiabetic persons.4 Since T1D is a disease mostly occurring in childhood, patients face a lifelong CV burden. Identifying risk factors for CV disease in childhood allows for earlier intervention and possible amelioration of risk.
The Diabetes Control and Complications Trial (DCCT) examined the effect of an intensive insulin regimen compared with a conventional insulin regimen on subsequent microvascular and macrovascular complications associated with T1D. The trial demonstrated that tight control of diabetes with an intensive insulin regimen resulted in a decreased incidence of microvascular complications.5 The use of intensive insulin resulted in better control of diabetes but had some deleterious side effects, notably increased risk of hypoglycemia and increased weight gain. The prevalence of overweight, defined by body mass index (BMI) of >27.8 kg/m2 in men and >27.3 kg/m2 in women, is almost 2-fold higher in intensively treated patients compared with patients on a conventional regimen.6 However, it is still unclear whether macrovascular complications that occur with weight gain7 offset the benefits of intensive insulin therapy. The current study was performed to examine the presence of CV risk factors (body composition, lipid profile, and apolipoproteins B and C-III) in children with and without T1D, both normal weight and overweight, between the ages of 13 and 20 years. We hypothesized that children with T1D will have a worse CV risk profile than children without T1D, and that there will be a synergistic or additive effect of overweight status and T1D on CV risk profile as defined by less favorable body composition and lipid profile and decreased arterial compliance.
This was a cross-sectional study of children with and without T1D between the ages of 13 and 20 years who were either of normal weight or were overweight. Patients with T1D were mainly recruited from our diabetes clinics, and nondiabetic patients were recruited through recruitment fliers and campus-wide emails. A total of 77 children were enrolled, and 68 completed the study. Data from 2 participants were not included in the analysis: 1 patient was taking oral contraceptive pills (inadvertently consented and screened, despite having an exclusion criterion per protocol, and the other because of markedly abnormal lipid data, suggesting a possible genetic mutation in lipoprotein metabolism, which was also an exclusion criterion). Study patients were defined as being overweight if their BMI was above the 85th percentile for age and sex and normal weight if their BMI was between the 3rd and 85th percentile for age and sex. The research protocol was approved by the institutional review board at the University of Oklahoma Health Sciences Center, and all patients signed an assent form prior to testing.
Inclusion criteria for children with T1D included a diagnosis of T1D for more than 3 years and an average hemoglobin HbA1c (HbA1c) between 6.5% and 10.7% for the past 6 months. To control for potentially independent effects of extreme hyperglycemia, a cut-off HbA1c level of 10.7% was selected, which represents one standard deviation above the mean HbA1c achieved in the adolescent population during the DCCT. To control for the well-described changes in the rates of diabetic complications after puberty, children who were prepubertal or early pubertal (Tanner stages 1 and 2) were excluded. Children were excluded from the study if they had any other coexisting endocrine, genetic, or metabolic disease or if they were taking any medications, including those that could affect substrate metabolism (excluding insulin), psychotropic medications, weight loss medications, and oral contraceptives for female patients. Inclusion criteria for Tanner stage and ages were identical for non-T1D children. Children were excluded from the study if they had impaired fasting glucose or had diabetes based on fasting glucose values. Additional exclusion criteria were otherwise identical to those listed above for the group with T1D. Two children with coexisting hypothyroidism and 1 child with Addison disease, all well-controlled on physiologic replacement hormonal therapy, were included in the study. Three diabetic participants treated for urinary microalbuminuria with angiotensin-converting enzyme inhibitors (for an average of 3 years prior to entry into the study) were included.
After obtaining appropriate consent and assent, each child underwent a history and physical examination by a board certified pediatrician. Height and weight were used to calculate BMI, waist and hip circumference were obtained for each patient, and the presence and degree of acanthosis nigricans were noted if present. They then underwent testing for body composition, vascular elasticity indices, and a blood draw for lipid profile and apolipoprotein values. All testing was performed in the morning after an overnight fast and was by an experienced nurse assigned to the study at the General Clinical Research Center at the University of Oklahoma.
Body composition was measured using dual energy x-ray absorptiometry scan (DXA; Hologic QDR 4500, Waltham, MA). Pulse wave analysis determination was made by HDI/Pulsewave CR-2000 (Hypertension Diagnostics, Eagan, MN).8 This technique uses a modified Windkessel model to derive information on proximal and distal arteries by analyzing the diastolic part of the arterial wave form.9 Testing was performed in fasting patients when rested in the supine position. An average of 3 readings was calculated to derive the mean small artery elasticity index and large artery elasticity index. Age-related norms for small artery and large artery compliance have recently been published.8
Blood was analyzed for fasting lipids, apolipoprotein B (apo B), and apolipoprotein C-III (apoC-III) levels. Total cholesterol, triglycerides, and high-density lipoprotein cholesterol (HDL-C) were measured by standardized enzymatic procedure. Very low-density lipoprotein cholesterol (VLDL-C) and low-density lipoprotein cholesterol (LDL-C) were estimated by Freidewald formula and non–HDL-C was calculated as total cholesterol minus HDL-C.
Blood was frozen for later analysis of apolipoprotein levels. Apolipoprotein levels were measured by immunoturbidimetry as previously described.10 Since heparin manganese has a high affinity for apoB, apoC-III in the heparin manganese precipitate (HP) is apoC-III bound to apoB-containing lipoproteins, and apoC-III in the heparin-manganese supernate (HS) is apoC-III bound to apoA-containing lipoproteins. ApoC-III is measured in the total plasma sample and in the precipitate following reconstitution to the original volume to obtain apoC-III HP. The value for apoC-III HS was derived by subtracting apoC-III HP from total plasma apoC-III.
Demographic variables including age, weight, BMI, BMI centile, height, Tanner stage, and HbA1c were compared across groups using the Fisher exact test and the Kruskal–Wallis test. Children without T1D who were overweight had a much higher BMI on average compared with children with T1D who were overweight, preventing us from doing a simple 4-group comparison. Thus, the effect of body weight status, diabetes status, and the interaction between the two on each CV variable was examined by regression analysis after controlling for age, sex, and ethnicity. Least square means along with standard error for both are presented. If the interaction was not statistically significant then the final model included only the main effects (diabetes or body weight status) and control variables (age, sex, and ethnicity). Additionally, a subgroup analysis was conducted on only those children with T1D to examine the effect of body weight status after controlling for age, sex, and ethnicity on CV risk profile.
Secondary analyses were performed on other indices of body fat. Additional models were created as described above, with the substitution of trunk fat mass, fat-free mass, total fat mass, and percent fat separately in place of the main effect of overweight status. The effect of these indices of body fat was summarized in each regression model by the model coefficient pertaining to the body fat index, along with the standard error and 95% confidence interval around this coefficient. All analyses were performed using SAS (version 9.1) software (SAS Institute, Cary, NC) with a significance level of .05.
Demographic information for all study patients and groups are listed in Table I. There were no differences in age or Tanner stages across the groups. As expected, there was a significant difference in weight and BMI between overweight and normal-weight groups (P<.0001). There was no significant difference in HbA1c between the diabetic patients (overweight and normal weight), and, as expected, HbA1c values were higher in diabetic patients compared with nondiabetic patients (P<.0001). There was a significant difference in ethnic distribution across all 4 groups (P=.0207). Since there were a higher proportion of Caucasian children in the T1D group compared with the other groups, the data were controlled for ethnicity. The effects of T1D, body weight status, and their interaction were examined for all CV risk parameters.
Table I. Demographics
|BMI centile, %ª||93.1±4.2||50.7±21.6||96.6±3.2||47.3±17.1|
Effects of diabetes on the CV risk factors are presented in Table II. These associations are adjusted by age, sex, and ethnicity due to imbalances across groups and known clinical associations of these variables with the CV risk factors. HDL-C levels were consistently higher in T1D patients compared with nondiabetic patients (P=.0023). None of the other lipid variables or any of the vascular measures showed a statistically significant association with diabetes status. A significant interaction was found (P<.05) between the effects of diabetes and body weight status for waist circumference, total fat mass, and trunk fat mass. The overweight nondiabetic patients had higher values than the overweight diabetic patients, whereas little differences were seen between the two normal-weight groups. A significant interaction was also found (P<.05) for percent body fat, as the normal-weight diabetic patients had higher values than the normal-weight nondiabetic patients, whereas little difference was noted between the two overweight groups.
Table II. Diabetes Effect on Cardiovascular Risk Factors from Multiple Regression Models, Adjusted for Age, Sex, and Ethnicity
| Waist circumference, cma|
| Normal weight||71.7±1.8||70.0±1.7||.2347|
| Waist to hip ratio, cm||0.79±0.02||0.79±0.01||.9747|
| Body fat, %a|
| Normal weight||24.3±1.7||21.4±1.6||.1814|
| Total fat mass, kga|
| Normal weight||13.4±1.1||11.3±1.1||.1619|
| Fat-free mass, kg||48.5±2.0||50.5±1.5||.3860|
| Trunk fat mass, kga|
| Normal weight||4.7±0.6||4.4±0.6||.6673|
| Triglycerides, mg/dL||79.9±9.9||89.6±7.2||.3826|
| Cholesterol, mg/dL||165.9±8.4||151.5±6.1||.1310|
| HDL-C, mg/dL||46.6±2.5||37.5±1.8||.0023b|
| LDL-C, mg/dL||103.3±7.4||96.0±5.4||.3863|
| LDL-C:HDL-C ratio||2.3±0.2||2.7±0.2||.1311|
| ApoC-III HS||4.2±0.3||3.8±0.2||.3292|
| ApoC-III HP||2.4±0.3||2.5±0.2||.7185|
| Mean arterial pressure, mm Hg||82.3±2.0||79.2±1.4||.1755|
| Systolic BP, mm Hg||116.6±2.4||115.4±1.7||.6690|
| Diastolic BP, mm Hg||62.9±1.7||59.3±1.2||.0619|
| Large artery elasticity index||16.9±1.3||17.1±0.9||.8788|
| Small artery elasticity index||8.4±0.6||9.4±0.4||.1331|
Effect of body weight status on the CV risk variables are presented in Table III. As expected, waist to hip ratio was significantly higher in the overweight patients than their normal-weight counterparts (P=.0001). Small arterial elastic index was higher in overweight patients (P=.0007). When substituting trunk fat mass, fat-free mass, percent fat, or total fat mass as indices of body fat for body weight status (Table IV), similar significance was found for each body fat measure with small artery elasticity index. Additionally, fat-free mass was found to have a significant negative association with HDL-C levels and significant positive association with systolic blood pressure. Total fat mass and percent fat were the only body composition variables that had a significant and positive association with large artery elasticity index (P=.0456 and P=.0457, respectively).
Table III. Weight Effect on Cardiovascular Risk Factors from Multiple Regression Models, Adjusted for Age, Sex, and Ethnicity
| Waist circumference, cma||N/A||N/A||N/A|
| Waist to hip ratio, cm||0.75±0.01||0.82±0.01||.0001b|
| Body fat, %a||N/A||N/A||N/A|
| Total fat mass, kga||N/A||N/A||N/A|
| Fat-free mass, kg||43.1±1.8||55.9±1.6||<.0001b|
| Trunk fat mass, kga||N/A||N/A||N/A|
| Triglycerides, mg/dL||77.1±8.8||92.4±8.1||.1518|
| Cholesterol, mg/dL||156.7±7.4||160.6±6.9||.6625|
| HDL-C, mg/dL||42.9±2.2||41.2±2.1||.5320|
| LDL-C, mg/dL||98.4±6.6||100.9±6.1||.7519|
| LDL-C:HDL-C ratio||2.4±0.2||2.6±0.2||.5192|
| ApoC-III HS||3.9±0.3||4.1±0.3||.6618|
| ApoC-III HP||2.1±0.3||2.7±0.3||.0826|
| Mean arterial pressure, mm Hg||79.5±1.8||82.0±1.6||.2476|
| Systolic BP, mm Hg||114.7±2.2||117.4±2.0||.2914|
| Diastolic BP, mm Hg||60.7±1.5||61.4±1.4||.6852|
| Large artery elasticity index||16.3±1.1||17.7±1.0||.3056|
| Small artery elasticity index||7.8±0.5||10.0±0.5||.0007b|
Table IV. Effect of Body Composition Variables on Lipid Profile and Vascular Measures from Multiple Regression Models, Adjusted for Age, Sex, Ethnicity, and Diabetes Status
|Mean arterial pressure, mm Hg||0.1±0.2||0.1±0.1||0.1±0.1||0.1±0.1|
|Systolic BP, mm Hg||0.1±0.2||0.3±0.1a||0.0±0.1||−0.0±0.1|
|Diastolic BP, mm Hg||−0.1±0.1||−0.1±0.1||−0.0±0.1||−0.0±0.1|
|Large artery elasticity index||0.2±0.1||0.1±0.1||0.1±0.1a||0.2±0.1a|
|Small artery elasticity index||0.2±0.0a||0.1±0.0a||0.1±0.0a||0.1±0.0a|
A subgroup analysis was performed to delineate the relationship between weight status and CV risk profile in children with T1D (Table V). Children with T1D who were overweight had higher apoC-III levels (P=.0453) compared with their normal-weight counterparts. Also, the triglyceride and apoC-III HP levels were higher in the overweight group (P=.0541 and P=.0761, respectively), although this did not reach statistical significance. These findings could not be explained by diabetes duration (P=.5877) or insulin dosage per kilogram of body weight (P=.7400).
Table V. Weight Effect on Cardiovascular Risk Factors from Multiple Regression Models in T1D Patients, Adjusted for Age, Sex, and Ethnicity
| Waist circumference, cm||74.1±3.9||91.5±4.6||.0002a|
| Waist to hip ratio, cm||0.8±0.03||0.8±0.03||.0809|
| Body fat, %||25.5±2.2||36.2±2.6||<.0001a|
| Total fat mass, kg||14.6±2.7||28.3±3.2||<.0001a|
| Fat-free mass, kg||40.9±3.7||50.6±4.3||.0138a|
| Trunk fat mass, kg||5.6±1.4||12.3±1.7||<.0001a|
| Triglycerides, mg/dL||78.0±16.2||111.2±19.4||.0541|
| Cholesterol, mg/dL||173.4±11.4||176.9±13.6||.7610|
| HDL-C, mg/dL||47.4±4.5||47.2±5.4||.9565|
| LDL-C, mg/dL||110.4±9.5||107.4±11.4||.7590|
| LDL-C:HDL-C ratio||2.4±0.3||2.3±0.4||.8743|
| ApoC-III HS||4.6±0.6||5.3±0.8||.2744|
| ApoC-III HP||2.4±0.5||3.3±0.6||.0761|
| Mean arterial pressure, mm Hg||83.7±3.2||83.2±3.8||.8834|
| Systolic BP, mm Hg||117.7±3.5||119.2±4.2||.6851|
| Diastolic BP, mm Hg||62.5±3.0||61.1±3.7||.6626|
| Large artery elasticity index||16.7±1.5||19.9±1.8||.0507|
| Small artery elasticity index||7.8±1.0||9.7±1.2||.0645|
The results of this study did not support our hypothesis that CV risk factors (unfavorable body composition, lipid values, and decreased arterial compliance) would be greater in T1D children and still greater in overweight children with T1D.
In our study, diabetes status was not associated with higher CV risk profile. Children with T1D had consistently and paradoxically higher HDL-C levels than non-T1D patients irrespective of their overweight status, although this is not a novel finding. Maahs and colleagues11 reported earlier the lower prevalence of decreased HDL-C levels in T1D children compared with the National Health and Nutrition Examination Survey (NHANES) cohort, irrespective of their glycemic status. The SEARCH for Diabetes in Youth Case-Control Study has shown12 higher LDL-C and total cholesterol levels in patients with T1D who had poorly controlled disease. It is possible that we did not see higher LDL-C levels in the children with T1D in our study because only children with relatively well-controlled diabetes were included, and thus the patients with poorly controlled diabetes and worse dyslipidemia were not represented.
As expected, overweight status conferred a worse body composition profile, with children who were overweight having higher waist to hip ratio and more fat-free mass compared with nonobese children. However, in this study, overweight status was not associated with a worse lipid profile or higher blood pressure. The interaction of overweight status and T1D was significant for several body composition variables. Overweight children without T1D had higher waist circumference, total fat mass, and trunk fat mass compared with overweight children with T1D, and this likely can be best explained from our sample of convenience: the overweight children without T1D in our sample were somewhat more overweight than our sample of overweight children with T1D. Interestingly, normal-weight T1D patients had a higher percentage of body fat compared with normal-weight nondiabetic patients, whereas little difference was noticed between the two overweight groups.
A subgroup analysis of the children with T1D showed that overweight children with T1D had significantly higher apoC-III levels compared with their normal-weight counterparts. Also these children had higher triglyceride and apoC-III HP levels that trended toward significance. Various apolipoprotein abnormalities have been described in patients with T1D, including elevated apoB and apolipoprotein C-III levels,13,14 and these have been postulated to be related to the increased CV risk seen even in diabetic patients who have a relatively normal lipid profile. ApoC-III has been associated with the metabolic syndrome in adults15,16 and adolescents16 and with insulin resistance in young children and adolescents.17 Furthermore, it has been shown to be a predictor of coronary events, progression of CV disease, and CV mortality.18,19 ApoB levels also were strongly linked to CV risk in patients with T1D in a 15-year follow-up study in Switzerland.20 It is unclear why we did not observe a relationship between T1D and apoB levels in our study, although it is possible that the negative consequences of diabetes on apoB may be confined to older patients.
We did not observe a statistically significant adverse effect of diabetes on arterial compliance nor an interaction between diabetes and overweight status. Other studies by Haller and colleagues21 and Jarvisalo and colleagues22 have demonstrated increased arterial stiffness and endothelial dysfunction, respectively, in children with T1D, although the apparently discordant findings between our findings and those reports may be related to differences in methods used. Haller and colleagues demonstrated arterial stiffness using radial tonometry in 98 diabetes patients compared with controls. In our study, we used applanation tonometry using the HDI pulsewave analyzer to characterize arterial stiffness. Jarvisalo used flow-mediated dilatation to detect endothelial dysfunction and ultrasound to measure carotid intimal thickness in children with T1D. He found endothelial dysfunction (75 diabetic children) and increased carotid artery intimal thickness compared with healthy controls. In our report, the trend for T1D patients to have lower small and large arterial compliance compared with the nondiabetic patients (although not statistically significant) is concerning and supports existing literature about the negative influence of T1D on vascular health.
Small artery elasticity index was higher in overweight patients. Although surprising, these data are consistent with a reported slower pulse wave velocity measured in obese compared with control children.23 This seemingly paradoxical observation was further supported by the subgroup analysis of patients with T1D, in that the overweight diabetic patients also tended to have higher small arterial elasticity index compared with their normal-weight counterparts (P=.0645). One might speculate whether this increased compliance might reflect a state of tonic vascular dilatation, either secondary to relative hyperinsulinemia or simply an initial compensating mechanism for the CV burden placed by excessive weight gain. This could possibly even reflect a higher insulin sensitivity in these overweight children. We observed no effect of overweight status on large arterial elasticity index.
Despite the fact that this is a relatively small study with all the well-described limitations of a cross-sectional study, several observations are key, including the lack of obvious potentiation of overweight status on risk factors in children with T1D and the seemingly improved arterial compliance in both overweight diabetic and nondiabetic children. Our study was limited by the fact that only well-controlled children with T1D were included in the study; thus, the observations may not apply to all children with T1D. However, the effect of poorly controlled T1D on CV risk factors is already well documented. Additional longitudinal studies are needed to further delineate the interaction between overweight status and T1D, including the effect of glycemic control on these measures.
These data do not support the notion that well-controlled children with T1D have worse CV risk factors compared with nondiabetic children. We could not demonstrate an additive effect of T1D and overweight status on CV risk profile. However, when children with T1D were considered separately, overweight diabetic children had higher apoC-III levels and tended to have higher triglyceride and apoC-III HP levels compared with their normal-weight counterparts. The direct relationship of small artery elasticity to BMI in both diabetic and nondiabetic children is paradoxical but consistent with recently published and newly emerging vascular data and warrants further investigation.
Acknowledgments and disclosures: We would like to gratefully acknowledge the helpful contributions of Amy Wisniewski, PhD, Steven D. Chernausek, MD, and Peter Alaupovic, PhD, to this project. This work was supported in part by grant M01 RR14467 from the National Center for Research Resources, National Institutes of Health and by Novo-Nordisk (C7042301). AWG is supported by grants from the National Institute on Aging (R01-AG-24296), National Center on Minority Health and Health Disparities (P20-MD-000528), National Center for Research Resources (M01-RR-14467 and P20-RR-024215), and Oklahoma Center for the Advancement of Science and Technology (HR09-035). KCC is supported by research funding from the National Institutes of Health (U01-DK061230-09). The authors have no competing interests to declare. Sowmya Krishnan, MD, and David A. Fields, MD, have research support through an investigator-initiated grant from Novo-Nordisk (C7042301).