The Aging of Elastic and Muscular Arteries: A comparison of diabetic and nondiabetic subjects

The study was approved by the institutional ethics committee and performed in accordance with the Declaration of Helsinki. All subjects gave written informed consent.

In 169 subjects (100 men [39 type 2 diabetic] and 69 women [18 type 2 diabetic]), age range 34–90 years, we assessed the mechanical properties of three arterial segments of differing wall composition. PWV was measured using noninvasively accessible superficial pulses and the Complior device (Artech Medicine, Pantin, France) in the predominantly elastic thoraco-abdominal aorta (carotid femoral PWF [PWV_cf]) and in an upper-limb muscular artery, over the carotid radial segment (carotid radial PWV [PWV_cr]). A different method was used to determine the PWV from the aorta to the finger (PWV_fin) by the timing between the R-wave of the electrocardiogram and arrival of the pressure pulse in the index finger (Finapress; Ohmeda) using a purpose-devised software. In addition, as an indicator of left-ventricular load and total systemic arterial compliance, central arterial compliance (CAC) was also assessed.

All methods have been described previously. Briefly, central PWV is determined from the time interval between the identified foot of the carotid and femoral arterial pulse waves and peripheral PWV from the carotid to radial arterial pulse interval (11). Distance was measured using a tape measure between transducer points, as specified by the manufacturer, and the PWV was automatically calculated by the Complior unit. In the case of PWF from the aorta to the finger (PWV_fin), distance was taken as that between the suprasternal notch and the tip of the index finger. CAC was assessed from simultaneous measurements of ascending aortic blood velocity using Doppler velocimetry and surrogate estimates of aortic root pressure obtained by applanation tonometry transducer (Millar Instruments, TX) of the proximal right carotid artery. CAC was then derived from the area method described by Liu et al. (12) and previously reported from our laboratory (13).

Analysis was performed using SPSS for Windows (release 10.0.5; SPSS, Chicago, IL) with significance taken as P < 0.05. The association of parameters with age was investigated by correlation using Pearson’s correlation coefficient and univariate linear regression. Differences in association of vascular parameters with age were investigated by comparison of the slope (β) of the appropriate regression lines (y = α + β × age).

Subject demographics and mean results are shown in Table 1. Compared with the nondiabetic subjects, the diabetic group was younger and heavier with increased SBP and DBP. Mean CAC was significantly higher in the nondiabetic group, and PWV_fin was significantly lower. Mean PWV_cf and PWV_cr were not significantly higher in the diabetic group (P = 0.580 and P = 0.233, respectively). Correlations of the measured parameters with age are shown in Table 2, with univariate associations illustrated graphically in Fig. 1. Of note is the disparity of findings between the nondiabetic and diabetic groups. None of the unadjusted measures of arterial stiffness in any segment was associated with age in the diabetic group (β did not differ significantly from zero), whereas PWV_cf (β = 0.12 ms⁻¹/year ± 0.021, P < 0.001) varied significantly with age in the nondiabetic subjects. Association of specific parameters with blood pressure (DBP, mean blood pressure [MBP], and SBP) varied between groups, but PWV_cf and PWV_fin were both positively associated with SBP in diabetic and nondiabetic subjects (Table 3). After adjusting for SBP, PWV_cf remained significantly correlated with age in the nondiabetic subjects [PWV_cf = 0.186 + 0.109 (age) + 0.035 (SBP), r² = 0.31, coefficient for age significant at the 0.001% level and for SBP at the 0.01% level]. Multiple regression analysis including potential determinants of PWV (age, MBP, DBP, SBP, height, weight, and sex) was performed. MBP was a significant determinant of all PWV except PWV_cr in the diabetic group.

Expression of results as a ratio involving predominantly nonelastic to elastic arterial segments improved the relationship with age (Fig. 2). In the nondiabetic group the absolute correlation with age tended to improve from 0.483 for PWV_cf to 0.594 when the ratio of PWV_cr to PWV_cf was considered (P = 0.064 for the difference). When PWV_fin was used in the numerator of the ratio, correlation with age also tended to improve in relation to the original parameter (nondiabetic subjects 0.483 to 0.571 [P = 0.145 for the difference]; 0.116 to 0.235 [P = 0.156 for the difference] for the diabetic group).

Significant differences existed in the rate of aging (β, slope of line) of the aortic segment (PWV_cf) compared with the carotid radial (PWV_cr) segment in nondiabetic subjects (slope = 0.12 vs. −0.02 ms⁻¹/year, respectively; P < 0.01). The changes for year of age for the nondiabetic group were −0.011 (PWV_cr-to-PWV_cf, r² = 0.35) and −0.0053 (PWV_fin-to-PWV_cf, r² = 0.33), both significant at the 0.001% level, whereas in the diabetic group the PWV_fin-to-PWV_cf slope was −0.0024 (r² = 0.05, P = 0.13) and PWV_cr-to-PWV_cf slope was −0.003 (r² = 0.002, P = 0.30).

Correction of the ratios for measured SBP, DBP, and MBP did not alter nondiabetes associations, but the association between PWV_fin/PWV_cf and age in diabetes was attenuated (r = 0.24, P = 0.10). Multivariate analysis on age and SBP with the calculated ratios as the dependant variable was performed. The ratios were associated with age and not SBP. In nondiabetic subjects PWV_cr/PWV_cf = 1.704 − 0.011(age) −0.000(SBP), r² = 0.35, and the coefficient for age was significant at the 0.001% level; and PWV_fin/PWV_cf = 0.844 − 0.005(age) − 0.001(SBP), r² = 0.36, and the coefficient for age was significant at the 0.001% level.

This study specifically investigates the relationship between arterial mechanical properties and aging in predominantly elastic compared with predominantly muscular arteries. In this context a significant finding is the apparent loss of the normal age-associated deterioration in any of the indices in the diabetic group. This is likely related to the fact that even though the diabetic subjects were on average younger, they had a higher BP and their mean values of PWV were higher in all segments than in nondiabetic subjects, suggesting that in the diabetic group age changes had occurred at an accelerated rate at an earlier age and had perhaps reached a functional plateau. Biologically, the diabetic group (mean chronological age 61 years) seems to have reached an “arterial” age approximating 75 years in those without diabetes. This is consistent with a relative lack of further deterioration from the elderly to very elderly and also consistent with reports suggesting that diabetic arteries stiffen earlier (14,15) and the fact that earlier onset of clinical arterial disease in diabetes is well known.

Our findings also confirm the known difference in age-related change between elastic and muscular arteries (10,16). Our nondiabetic subject group showed a nonsignificant and small decrease in PWV with age in the carotid to radial segment (Table 2), possibly associated with a decrease in arterial stiffness. Conversely, the predominantly elastic aortic artery demonstrated a considerable age-related increase in PWV. This is consistent with previously reported changes and the location of clinically related arterial disease. CAC predominantly represents the buffering capacity of the proximal aorta. The arterial segment represented by CAC is largely precluded from involvement in PWV_cf assessment, as the time taken for the pressure pulse to reach the femoral artery is long, and the time taken to propagate to the transduction point at the carotid artery corresponds to passage of the aortic pressure pulse around the aortic arch. There was no statistically significant age-associated changes in CAC in our study group. However, despite their relative youth, we did demonstrate the expected lower CAC in those with diabetes compared with those without diabetes (17).

Change in specific arterial mechanical properties has been shown to bear a closer relation to chronological age than many other parameters, such as graying of hair, skin elasticity, etc., that are currently used (18). In this context it has been proposed that the assessment of “biological” as opposed to “chronological” age of an individual’s artery may be of use in the management or prevention of arterial disease. Our results suggest that use of a ratio incorporating the differences between age effects on elastic and muscular arteries may be a useful means of improving the age association of arterial change in nondiabetic subjects and thus a means of improving the accuracy of assessment of arterial health. Our results further suggest that if a ratio is to be used, then either PWV_cr or PWV_fin may be used as the numerator in nondiabetic subjects. For diabetic subjects, however, the numerator should be PWV_fin. This is most likely due to the segment represented by PWV_fin, including small arteries such as those characteristically affected by diabetes.

There were apparent differences in the pressure dependence of PWV in the arterial segments studied between groups, with the effect of pressure on PWV_cr seemingly less in the diabetic group. One potential explanation for this could be that in this group intrinsic stiffening changes (structural or functional) had occurred to such an extent that further pressure dependence was irrelevant. However, this did not apply to PWV_fin and PWV_cf. In the PWV ratios calculated here pressure was not a significant determinant, and this further suggests that a ratio of muscular to elastic artery PWV may be a useful parameter across individuals.

There were differences in the baseline characteristics of the diabetic and nondiabetic groups. This is not a significant drawback in the current study because the aim was not to compare absolute parameters between groups but to look at relationships within groups. Nevertheless, the results demonstrate why there is much controversy regarding the effects of diabetes on arterial compliance, positive deleterious results arising from studies in younger persons, and negative studies in older subjects in which the normal control subjects also have impaired compliance (11,18,19). All arterial mechanical properties vary with blood pressure level. As expected, all measures of BP in this study increased with age. This could have a confounding effect on our measurements of age changes in arterial properties. We do not feel that this is a significant influence, however, as conclusions were not altered by adjustment for DBP, SBP, or MBP.

In addition, the method used for measuring PWV also has limitations. The distance traveled by the pulse wave is measured on the surface of the body. The vessels in the human body do not travel in a straight line, therefore possibly confounding the measurements. The PWV_fin, which is a time interval measurement, has a latent period due to intraventricular activation time. This may vary in different people and may also be affected by diabetes. Diabetes affects small vessels, thus this may be a confounder in measuring PWV_fin.

In conclusion, we have shown significant differences in the rate of age-related deterioration in elastic arteries of nondiabetic compared with diabetic arteries. Further, we have shown that an improvement in the assessment of age-related change in elastic arteries may be obtained by the use of a ratio measure incorporating values obtained from assessment of elastic and muscular arteries separately.

J.D.C. was supported by a Wellcome Trust Travel Grant.

We thank Mr. Winston Banya, MSc, for statistics support.