The classical cardiovascular risk factors explain only a part of the increased macrovascular disease in subjects with type 2 diabetes (1). Reduction of the elastic properties of the aorta is an early marker of atherosclerosis, and it is diagnosed early in the course of type 2 diabetes (2). The detrimental effects of the classical risk factors on aortic distensibility are well known. However, the potential effect of some newer markers that have been proposed as risk factors for atherosclerosis, including increased plasma homocysteine (Hcy), high-sensitivity C-reactive protein (hs-CRP), fibrinogen, and uric acid levels, increased white blood cell (WBC) count, and low plasma adiponectin levels on aortic distensibility, has not been studied so far. The present cross-sectional study deals with the effect of the classical and newer cardiovascular risk factors on aortic distensibility in subjects with and without type 2 diabetes.
RESEARCH DESIGN AND METHODS
We studied 100 subjects (49 male and 51 female, mean age 59.4 ± 1.0 years, duration of diabetes 9.8 ± 7.1 years) with type 2 diabetes consecutively attending the outpatient diabetes clinic of our hospital. In addition, 100 subjects (49 male and 51 female, aged 59.2 ± 1.0 years) without diabetes who had attended the hospital for minor problems were also recruited. Type 2 diabetes was diagnosed according to the American Diabetes Association criteria (3). Subjects with clinically apparent macrovascular disease and renal impairment were excluded.
Blood was collected after an overnight fast of at least 12 h. Plasma glucose, lipids, HbA1c (A1C), creatinine, adiponectin, Hcy, hs-CRP, fibrinogen, uric acid, WBC count, and neutrophil count were measured. LDL cholesterol was calculated. Albumin-to-creatinine ratio (ACR) was measured in a first morning urine sample.
Aortic distensibility was determined noninvasively using a Hewlett Packard Sonos 1000 ultrasound system, based on the relationship between changes in aortic diameter and pressure with each cardiac pulse, as described previously (2).
Univariate and multivariate linear regression analyses were performed to look for associations between aortic distensibility and the variables of interest (Table 1). Several models of multivariate analysis were created with the addition of each one of the newer risk factors in a basic model (model 1) that included the classical risk factors for atherosclerosis and ACR, which was associated with aortic distensibility in univariate analysis.
RESULTS
Subjects without diabetes
Univariate linear regression analysis showed significant relationships between aortic distensibility and age (P < 0.001), BMI (P = 0.006), smoking status (P = 0.03), presence of dyslipidemia (P = 0.004), mean arterial blood pressure (P < 0.001), history of hypertension (P < 0.001), uric acid (P = 0.007), ACR (P = 0.001), and WBC count (P = 0.02). No significant relationships were found with Hcy, hs-CRP, adiponectin, fibrinogen, neutrophil, or WBC count. Models 1–7 of multivariate analysis demonstrated that age, mean arterial blood pressure, smoking status, and dyslipidemia were constantly and independently associated with aortic distensibility. These variables explained 50% of the variance of aortic distensibility values. Addition of the log10WBC count eliminated the role of smoking as a confounding factor. The total explained variance of the model increased by only 2% with the addition of the log10WBC count.
Subjects with type 2 diabetes
Univariate linear regression analysis showed significant relationships between aortic distensibility and sex (P = 0.005), age (P < 0.001), duration of diabetes (P < 0.001), history of hypertension (P < 0.001), fibrinogen (P = 0.03), uric acid (P = 0.01), and ACR (P = 0.04). No significant association was found between aortic distensibility and the other classical as well as the new risk factors for atherosclerosis or A1C. In all models, age (P = 0.001), duration of diabetes (P < 0.001), and history of hypertension (P < 0.02) were constantly and independently associated with aortic distensibility. These variables explained 40% of the variance of the dependent variable. The total explained variance of the models increased by 6% with the addition of uric acid.
When both participants with and without diabetes were examined together, diabetes status per se was the strongest predictor of aortic distensibility (β = −0.81, P < 0.001), explaining 64% of the variance of aortic distensibility. Moreover, in both study groups, substitution of the continuous independent variables with categorical variables (values above the mean value versus values below the mean value of each parameter in each study group) did not change the results of multivariate analysis. Furthermore, the kind of antihypertensive treatment was not associated significantly with aortic distensibility in univariate analysis in either group.
CONCLUSIONS
In agreement with previous reports (4,5), this study has shown that in the nondiabetic subjects, age, blood pressure, and dyslipidemia are associated with reduced aortic distensibility. A novel finding was the association between aortic distensibility and WBC count. Data on the relationship between WBC and coronary artery disease are controversial (6–8). The independent association between WBC and aortic distensibility (only in subjects without diabetes) is probably explained by the fact that WBC count reflects underlying inflammation. This association might be more apparent in nondiabetic individuals, because diabetes status per se masks the effect of other factors.
Also in agreement with previous reports, we found that in the diabetes group, age, presence of hypertension, and duration of diabetes were constantly associated with aortic distensibility (4,5,9). Long duration of diabetes is associated with a decline in aortic elastin and glycation-induced molecular cross-links, which compromise elasticity of the aorta (9,10).
This study demonstrated an independent relationship between aortic distensibility and uric acid in type 2 diabetes. There are limited data on the association between uric acid and macrovascular disease in type 2 diabetes; two such studies showed that high uric acid predict macrovascular disease (11,12). Moreover, recent data suggest that uric acid is an independent risk factor for cardiovascular disease but only in high-risk individuals, such as subjects with type 2 diabetes (13).
Our study does not have enough power to confirm or reject all the relationships between some of the newer markers studied and aortic distensibility. This is especially true for hs-CRP, Hcy, and adiponectin. It is noted that in order to achieve a strong power, at least 400 subjects per group would be needed; therefore, the results are indicative and not conclusive. Furthermore, it is worth mentioning that recent studies have questioned the role of these parameters as cardiovascular risk markers (14–16).
In conclusion, this study has shown that the classical cardiovascular risk factors are associated with aortic distensibility in both subjects with and without type 2 diabetes. As it concerns the newer markers, it indicates that only WBC count and uric acid are associated with aortic distensibility, although they add little burden to the reduction of aortic distensibility.
Multivariate linear regression analysis: the effect of some newer risk factors for atherosclerosis on the aortic distensibility in subjects with and without type 2 diabetes
. | Subjects without diabetes . | . | Subjects with type 2 diabetes . | . | ||
---|---|---|---|---|---|---|
. | β . | P . | β . | P . | ||
Model 1 | ||||||
Age (years) | −0.38 | <0.001 | −0.34 | <0.001 | ||
Male vs. female | 0.12 | 0.13 | 0.02 | 0.75 | ||
BMI (kg/m2) | −0.10 | 0.19 | −0.12 | 0.13 | ||
Current smokers vs. nonsmokers | −0.15 | 0.04 | −0.08 | 0.30 | ||
Mean arterial blood pressure (mmHg) | −0.35 | <0.001 | −0.10 | 0.18 | ||
History of hypertension (yes) | −0.21 | 0.01 | −0.20 | 0.01 | ||
Presence of dyslipidemia (yes) | −0.17 | 0.02 | −0.12 | 0.10 | ||
Duration of diabetes (years) | −0.39 | <0.001 | ||||
ACR (mg/mmol) | −0.18 | 0.01 | −0.11 | 0.16 | ||
Model 2 | ||||||
Uric acid (mmol/l) | −0.10 | 0.45 | −0.28 | 0.001 | ||
Model 3 | ||||||
Homocysteine (μmol/l) | −0.01 | 0.89 | −0.11 | 0.19 | ||
Model 4 | ||||||
hs-CRP (mg/l) | −0.004 | 0.96 | −0.05 | 0.51 | ||
Model 5 | ||||||
log10Adiponectin (μg/ml) | −0.08 | 0.50 | 0.14 | 0.16 | ||
Model 6 | ||||||
Fibrinogen (g/l) | 0.08 | 0.25 | 0.02 | 0.75 | ||
Model 7 | ||||||
log10Neutrophils count (109/l) | −0.09 | 0.24 | −0.03 | 0.69 | ||
Model 8 | ||||||
log10WBC count (109/l) | −0.18 | 0.02 | −0.05 | 0.49 |
. | Subjects without diabetes . | . | Subjects with type 2 diabetes . | . | ||
---|---|---|---|---|---|---|
. | β . | P . | β . | P . | ||
Model 1 | ||||||
Age (years) | −0.38 | <0.001 | −0.34 | <0.001 | ||
Male vs. female | 0.12 | 0.13 | 0.02 | 0.75 | ||
BMI (kg/m2) | −0.10 | 0.19 | −0.12 | 0.13 | ||
Current smokers vs. nonsmokers | −0.15 | 0.04 | −0.08 | 0.30 | ||
Mean arterial blood pressure (mmHg) | −0.35 | <0.001 | −0.10 | 0.18 | ||
History of hypertension (yes) | −0.21 | 0.01 | −0.20 | 0.01 | ||
Presence of dyslipidemia (yes) | −0.17 | 0.02 | −0.12 | 0.10 | ||
Duration of diabetes (years) | −0.39 | <0.001 | ||||
ACR (mg/mmol) | −0.18 | 0.01 | −0.11 | 0.16 | ||
Model 2 | ||||||
Uric acid (mmol/l) | −0.10 | 0.45 | −0.28 | 0.001 | ||
Model 3 | ||||||
Homocysteine (μmol/l) | −0.01 | 0.89 | −0.11 | 0.19 | ||
Model 4 | ||||||
hs-CRP (mg/l) | −0.004 | 0.96 | −0.05 | 0.51 | ||
Model 5 | ||||||
log10Adiponectin (μg/ml) | −0.08 | 0.50 | 0.14 | 0.16 | ||
Model 6 | ||||||
Fibrinogen (g/l) | 0.08 | 0.25 | 0.02 | 0.75 | ||
Model 7 | ||||||
log10Neutrophils count (109/l) | −0.09 | 0.24 | −0.03 | 0.69 | ||
Model 8 | ||||||
log10WBC count (109/l) | −0.18 | 0.02 | −0.05 | 0.49 |
β, standardized regression coefficient. In subjects without diabetes, additional variables tested in models 2–8 include sex, age, BMI, smoking status, mean arterial blood pressure or presence of hypertension, presence of dyslipidemia, and ACR. In subjects with type 2 diabetes, additional variables tested in models 2–8 include sex, age, BMI, duration of diabetes, smoking status, history of hypertension, presence of dyslipidemia, and ACR.
Article Information
This work was supported by a research grant (code 03EΔ/95) from the European Union (The Operational Programme “Competitiveness” 2000–2006, European Social Fund), the General Secretariat for Research and Technology, Greece, and the National and Kapodistrian University of Athens.
References
A table elsewhere in this issue shows conventional and Système International (SI) units and conversion factors for many substances.
DOI: 10.2337/dc06-0154
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.