Acute Respiratory Viral Infections Aggravate Arterial Endothelial Dysfunction in Children With Type 1 Diabetes

A total of 26 children (aged 6–18 years, mean [±SD] age 14 ± 3 years) with type 1 diabetes (duration 6 ± 3 years) were recruited. Of these, 11 children had a clinically manifested upper respiratory tract infection (body temperature >38°C and flu-like symptoms) 6–8 weeks before the study day. None of the children who had the infection required antibiotic treatment or hospitalization. Ten age- and weight-matched healthy children served as control subjects. None of the control subjects and remaining diabetic children had a history of infection or other acute inflammatory disease within the preceding 3 months. Exclusion criteria were clinically manifested angiopathy, hypertension, obesity (BMI >25 kg/m²), smoking, pregnancy, and early coronary heart disease among first-degree relatives. The study was approved by the Lund University ethics committee.

Weight and blood pressure were measured on the ultrasound day. Venous samples were collected from the diabetic children following the ultrasound study. Data on age, diabetes duration, and mean HbA_1c (3 months before the study) were obtained from the registry of the outpatient diabetes clinic.

LDL, HDL, and total cholesterol were analyzed by enzymatic methods using a Hitachi Modular-P system (Roche Diagnostics). Plasma von Willebrand antigen (vWF) was measured by a latex immunoassay. Plasma high-sensitivity C-reactive protein (CRP) was measured by an enzyme immunoassay using polyclonal antibodies (Dako Diagnostics, Glostrup, Denmark). An electroimmunoassay was used to assess plasma orosomucoid. A modified enzyme-linked immunosorbent assay was used to determine the antibodies against oxidized LDL cholesterol. The oxidized LDL cholesterol data are expressed as the ratio of binding to oxidized to binding to native LDL.

The methodology is described in detail elsewhere (10). Briefly, electrocardiogram-gated end-diastolic longitudinal scans of the brachial artery were imaged through a 15-mHz linear probe of an Acuson Sequoia C256. All ultrasound frames were transferred via frame grabber to a computer for offline edge-detected measurement of the arterial diameter (Data Acquisition and Analysis; Information Integrity 1.51). Flow-mediated dilation (FMD) and glyceryl-trinitrate (GTN)-induced dilatation were expressed as maximum percent dilatation following cuff deflation and GTN administration, respectively. All scans were taken by the same sonographer. The analyses of the subjects’ clinical characteristics were done blindly.

StatView (SAS Institute, Cary, NC) was used for statistical calculations. Differences in the studied variables were assessed by ANOVA. Eventual correlations between the dependent variable (i.e., FMD) and the hypothesized predictor variables (LDL, HDL, and total cholesterol and vWF, CRP, orosomucoid, mean and latest HbA_1c values, and diabetes duration) were assessed by simple regression analysis. Data are means ± SD. Given the skewed distribution of CRP, values derived from log-transformed means, coincident with geometric means, were compared. Statistical significance was inferred at P < 0.05.

With the exception of vWF, which was higher in the diabetic children who had infection (P < 0.05), no differences in the studied variables were noted between the infection and noninfection diabetic children (Table 1). Overall, in diabetic children, a direct correlation was noted between oxidized LDL and CRP (r = 0.5, P = 0.03).

Neither the baseline diameter nor the GTN-induced dilatory responses of the brachial artery (Table 1) differed between the groups (P > 0.5). In contrast, diabetic children who had infection had significantly lower FMD than diabetic children without infection (P < 0.01) and control subjects (P < 0.001) (Table 1). In the infection group, FMD was inversely related to orosomucoid (r = −0.8, P < 0.05). The FMD of the entire diabetic group (infected and noninfected children) was 5.6 ± 2% and inversely correlated with vWF (r = −0.4, P < 0.05), CRP (r = −0.6, P = 0.01), and LDL cholesterol (r = −0.6, P < 0.01).

The present study provides suggestive evidence that acute respiratory viral infections could aggravate the arterial endothelial dysfunction in children with type 1 diabetes, as indicated by the significantly lower FMD and higher vWF in the postinfection diabetic children. In keeping with other studies (3,4), we found that FMD of the brachial artery is profoundly altered in children with diabetes, and this appears to be mediated in part by a low-grade systemic inflammation (defined by high-sensitivity CRP and elevated LDL cholesterol levels), as suggested by their inverse correlation with FMD.

Recurrent acute respiratory infections are more common in diabetic patients, seemingly as a consequence of impaired immunity (8). Acute infections could cause short-lasting inflammatory responses with subsequent endothelial damage, which, depending on the number of infections in time and presence of additional cardiovascular risk factors (e.g., diabetes), might be critical to atherosclerosis development (11). Indeed, earlier pediatric studies have suggested that various types of acute infections could elicit proatherogenic changes in arterial structure (12).

The main limitation of this study resides in the small sample size, which could have influenced the statistical power. The cross-sectional nature of this study precludes conclusions with regard to the possible implications of the infection-related endothelial changes in the pathogenesis of atherosclerosis in diabetic children. However, in view of the central role of endothelial dysfunction in the development of atherosclerosis and its multifactorial etiology, it is conceivable that augmentation of endothelial dysfunction by infections could further increase the susceptibility to accelerated atherosclerosis in the diabetes milieu. This issue needs to be addressed by future large-scale prospective studies.

This study was supported by grants from the Swedish Royal Physiographic Society (to P.L.), Anna-Lisa and Sven-Eric Lundgren’s Foundation for Medical Research (to P.L.), Barndiabetesfonden (to S.S.), and Lund University Hospital (to P.L.).

We thank Annica Maxedius, RN, for excellent technical assistance. We also acknowledge Irén Tiberg, RN, and Annette Leger, RN, for their support in patient recruitment.