Assessment of Endothelial Function by Means of Flow‐Mediated Changes Using Pulse Wave Velocity

Abstract

J Clin Hypertens (Greenwich). 2010;12:495–501. © 2010 Wiley Periodicals, Inc.

Endothelial dysfunction is a leading cause of early development of cardiovascular diseases. Endothelial function can be assessed using ultrasound methods to watch the arterial flow‐mediated dilation. It is also possible to find changes in pulse wave velocity (PWV) after induced ischemia related to the vessel diameter changes. Pre‐ and post‐induced ischemia carotid‐radial PWV was recorded in 226 hypertensive patients (150 women [63.5±12.4 years old] and 76 men [63.2±11.8 years old] and 55 healthy patients (38 women [63.1±12.6 years old] and 17 men [54.8±12.8 years old]). The authors considered normal endothelial function a PWV reduction of 5% from baseline. To assess nondependent endothelial dilation the authors performed carotid‐radial PWV after sublingual administration of 5 mg of isosorbide dinitrate in a group of patients with abnormal flow‐mediated dilation. A significant PWV reduction of 9.8% in normal patients and only 1.2% among hypertensive patients (P<.0005) was found. After sublingual isosorbide dinitrate intake the authors observed a greater fall in PWV (14%) than that observed in healthy people after induced ischemia. Carotid‐radial PWV after induced ischemia decreased significantly in normal participants. No significant changes were observed in hypertensive patients. These results may offer a reliable tool to assess endothelial function in medium‐size arteries.

The endothelium is a large paracrine organ; it is extremely active in the regulation of blood vessel tone and cellular growth by secreting a number of dilator and constrictor substances. In addition to the influence on vascular tone, it acts producing modification on platelet aggregation, thrombogenicity of the blood, vascular inflammation, and oxidative stress, and more importantly it influences cells’ migration and growth leading to the development and progression of atherosclerosis. According to this fact, endothelial dysfunction is an early sign of atherosclerosis and represents the pathway to cardiovascular (CV) disease development. Early detection of endothelial dysfunction is a milestone to prevent atherosclerosis and related CV events. Therefore, improvement of endothelial dysfunction is capable of improving CV risk. ¹

Recently, noninvasive ultrasonographic (US) procedures have been developed to easily assess flow‐mediated dilation (FMD) in forearm arteries. The arteries’ ability to respond to physical and chemical stimuli allows them to regulate their smooth muscle tone. Many of the vessels respond to flow enhancement—shear stress—with dilation. Ischemia produced by brachial artery cuff compression during at least 5 minutes leads to local nitric oxide (NO) production induced by shear stress after flow liberation. Under normal conditions, this response produces post‐ischemia dilation easily measured by the above‐mentioned method, which is considered an accurate index of endothelial function. ¹

Dilation produced by NO lowers pulse wave velocity (PWV); it can be noninvasively assessed by means of US or by computerized methods such as the Complior (Colson AS, Paris, France) with a high reproducibility intra‐ and interobserver index. ² This method also allows the follow‐up of aortic arterial distensibility in hypertensive patients or other patients with high CV risk factors. ³

Arterial distensibility assessed by aortic PWV is a powerful independent predictor of morbidity and mortality, more accurate than systolic blood pressure (SBP). ⁴ Nevertheless, aortic PWV is not an indicator of endothelial function while FMD is.

According to the Moens–Korteweg equation (Figure 1), PWV is directly proportional to the arterial wall width and Young modulus—a measurement of elastin and collagen concentration or arterial stiffness—and inversely proportional to vessel diameter and blood viscosity. Hence, higher arterial wall width and stiffness will lead to higher PWV. In contrast, a larger vessel lumen diameter and higher blood density decreases PWV. Due to this mechanism, the arterial FMD after induced ischemia produces a slower PWV. ⁵

Figure 1.

 Moens–Korteweg equation. PWV indicates pulse wave velocity; E, young modulus; h, vessel wall thickness; R, vessel radius; δ, blood density.

Based on these concepts, the aim of our study was to assess endothelial function in hypertensive patients by means of carotid‐radial PWV measurement before and after brachial artery induced ischemia through sphygmomanometer cuff inflation.

Material and Methods

Carotid‐radial PWV was measured before and after induced ischemia in 281 patients seen at our hypertension clinic, to assess endothelial function similar to endothelial dependent dilation assessed with US. Subjects were divided into 2 groups according to their blood pressure (BP). Hypertension was defined by an SBP >140 mm Hg and/or a diastolic BP (DBP) >90 mm Hg in nondiabetic patients, and an SBP >130 mm Hg or a DBP >85 mm Hg in diabetics. Group 1 included 226 hypertensive patients aged 63.4±12.1 years old (range between 31 and 86) of which 150 were women aged 63.5±12.4 years old (31–86) and 76 were men aged 63.2±11.7 years old (32–86). Patients in this group had several CV risk factors usually considered as a cause of endothelial dysfunction. All of them had hypertension. Among them there were patients with type 2 diabetes, and other CV risk factors such as dyslipidemia, obesity, metabolic syndrome, hypothyroidism, chronic kidney disease, and smoking (Table I). Group 2 included 55 normotensive patients aged 60.5±13.1 years old (32–84), of which 38 were women aged 63.1±12.6 (42–84) and 17 were men aged 54.8±12.8 (32–71) without any detected CV risk factors related to endothelial dysfunction. Most patients were receiving antihypertensive drugs, usually angiotensin‐converting enzyme inhibitors (ACEI) or angiotensin receptor blockers (ARB) or another treatment for their CV risk factors. All patients were studied fasting and without taking medication in the morning of the procedure.

Table I.

 Prevalent Risk Factors and General Characteristics of Both Groups at Baseline

	Group 1			Group 2			P Value
	All	Men	Women	All	Men	Women
Hypertension	100%			0%
Type 2 diabetes	31.6%			0%
Dyslipidemia	24.0%			0%
Metabolic syndrome	20.4%			0%
Hypothyroidism	8.9%			0%
Chronic kidney disease	3.6%			0%
Smokers	6.2%			0%
Number of subjects	226	76	150	55	17	38
Age (years)	63.4±12.1	63.2±11.8	63.5±12.4	60.5±13.1	54.8±12.8	63.1±12.6	NS
Height (cm)	163.0±8.9	171.9±6.2	158.5±6.2	162.6±8.5	171.1±6.1	158.8±6.5	NS
Weight (kg)	74.3±15.9	84.2±16.8	69.3±12.8	68.3±10.2	77.3±6.0	64.3±9.0	<.005
BMI (kg/m2)	27.9±5.0	28.4±5.3	27.6±4.8	25.8±3.2	26.4±2.0	25.5±3.6	<.005
Waist circumference (cm)	96.1±12.5	100.2±12.7	94.0±11.9	90.2±8.6	93.2±6.4	88.8±9.2	<.001
SBP (mm Hg)	138.8±18.3	138.5±17.5	138.9±18.8	120.3±9.5	119.9±11.5	120.6±8.6	<.001
DBP (mm Hg)	84.3±9.8	85.1±11.1	83.9±9.1	76.6±6.7	76.5±8.0	77.7±6.1	<.001
Heart rate (beats/min)	69.1±10.0	70.0±10.0	68.7±10.1	67.0±8.1	64.6±7.1	68.0±8.4	NS

Values are expressed in absolute numbers.

Abbreviations: BMI, body mass index; DBP, diastolic blood pressure; NS, not significant; SBP, systolic blood pressure.

The Ethics Committee and Teaching and Research Committee of our institution approved the protocol. Each patient signed a consent form in which the nature of the research and its purpose was clearly explained.

BP Measurement

To establish a baseline value, SBP and DBP were measured at least 3 times in the supine position after 10 minutes of rest using a certified aneroid sphygmomanometer (Tycos lifetime certified, Welch Allyn Tycos, Arden, NC) and also before each PWV assessment. According to clinical use, SBP was considered as the first Korotkoff sound (K1) and DBP when all sounds disappear (K5). The cuff was deflated at a rate of 2 mm/s.

Determination of PWV

Pre‐ and post‐induced ischemia carotid‐radial PWV was assessed by means of a noninvasive automatic device, Complior (Colson AS, Paris, France) which allowed an online pulse wave recording and automatic calculation of PWV. The validation of this automatic method and its reproducibility has been previously published. ² , ⁶ Briefly, common carotid artery and radial artery pressure waveforms were recorded noninvasively using a pressure‐sensitive transducer (TY‐306, Fukuda, Tokyo, Japan). The digitized waveforms were sampled at 500 Hz. Both waveforms were stored in a memory buffer. When the operator observed on the computer screen a pulse wave of sufficient quality, digitization was suspended and calculation of the time delay between the 2 pressure upstrokes was initiated. Measurement was repeated over more than 20 different cardiac cycles, and the calculated mean value was used for the final analysis. The distance between the 2 measurement positions was measured over the body surface as the distance between the 2 recording sites (D), whereas pulse transit time (t) was automatically determined by the Complior as PWV = D/t. The concept of PWV as an indicator of arterial compliance was previously shown. ⁷ The same observer, who was blind to the previously assessed BP of the subject, performed all measurements. We measured carotid‐radial PWV and then we compressed the brachial artery for 5 minutes using the sphygmomanometer cuff 50 mm Hg above SBP. Thirty seconds after the artery flow release we performed the assessment of PWV again to register the changes after FMD.

Carotid‐radial post‐ischemia changes were considered normal after a PWV reduction of 5% or more from baseline and abnormal when the reduction was <5% or with a paradoxical response (increasing PWV). Furthermore, to confirm the relation between FMD and endothelial function, in patients with abnormal response we performed the measurement after sublingual isosorbide dinitrate administration to evaluate nondependent endothelial dilation. ⁸

Statistical Analysis

Values were expressed as mean ± standard deviation. We compared pre‐ and post‐ischemia PWV in both groups and we performed the Student t test to determine the statistical signification and multivariate analysis of variance (MANOVA) to confirm the results. We also performed a multivariate analysis including age and SBP at baseline. We considered a P value of <.05 and <.025 in the multivariate analysis as statistically significant.

Results

We compared anthropometrical characteristics in both groups (Table I). Waist circumference and body mass index (BMI) were significantly less in normotensive subjects. SBP and DBP were also significantly lower in normotensives. Heart rate (HR) had no significant differences between groups (Table I). There were not significant differences in pre‐ and post‐ischemia BP and HR in each group (Table II).

Table II.

 Pre‐ and Post‐Ischemia Systolic and Diastolic Blood Pressure and Heart Rate According to Study Group

	All		Men		Women		P Value
	Pre	Post	Pre	Post	Pre	Post
Group 1
SBP (mm Hg)	138.8±18.3a	137.1±18.5a	138.5±17.5a	136.1±16.7a	138.9±18.8a	137.7±19.3a	NS
DBP mm Hg)	84.3±9.8a	83.4±9.9a	85.1±11.1a	83.6±10.6a	83.9±9.1a	83.2±9.6a	NS
HR (beats/min)	69.1±10.0	68.1±10.5	70.0±10.0	68.9±12.1	68.7±10.1	67.7±9.6	NS
Group 2
SBP (mm Hg)	120.3±9.5a	120.9±10.2a	119.9±11.5a	119.3±11.9a	120.6±8.6a	121.6±9.5a	NS
DBP mm Hg)	76.6±6.7a	76.5±6.6a	76.5±8.0a	76.4±8.2a	76.7±6.1a	76.5±5.9a	NS
HR (beats/min)	67.0±8.1	66.8±8.1	64.6±7.1	64.6±7.1	68.0±8.4	67.5±8.4	NS

Values are expressed in average ± standard deviation.

Abbreviations: DBP, diastolic blood pressure; HR, heart rate; NS, not significant; SBP, systolic blood pressure. ^a P<.0001.

Post‐ischemia PWV was significantly lower in normotensive patients in comparison with hypertensive patients (6.9±1.2 vs 7.7±1.2 m/s (Figure 2, Table III). The percent of variation in both groups was −1.2% and −9.8%, respectively (Table III, Figure 2). When we analyzed by gender, the percent of variation was significantly lower among men (−3.4% and −6.5%) than among women (−0.04% and −11.2%) both in hypertensive patients and normotensive patients, respectively (Table III, Figure 3B1–C1). After considering the analysis by age, we found a more significant difference between genders indeed. Men above 65 years old had a more significant difference between hypertensive patients and normotensive patients than men below this age. The women cohort showed a more significant fall in PWV among the younger group than the oldest (Table IV; Figure 3B2/3–C2/3). Nevertheless, the multivariate analysis performed over the whole population, including age and/or SBP at baseline has not shown a significant influence in the results observed over the single groups of hypertensive or normotensive patients.

Figure 2.

 Pre‐ and post‐ischemia pulse wave velocity (PWV).

Table III.

 Carotid‐Radial Pre‐ and Post‐Ischemia Pulse Wave Velocity and Percent of Variation According to Groups

	Total			Men			Women			P Value
	Pre	Post	%	Pre	Post	%	Pre	Post	%
Group 1	7.8±0.4	7.7±0.4	−1.2	8.3±0.4	8.0±0.4	−3.4	7.6±0.4	7.5±0.4	−0.04	NS
Group 2	7.7±0.2	6.9±0.4	−9.1	8.1±0.4	7.5±0.4	−6.5	7.5±0.3	6.6±0.4	−11.2	<.001
P value	NS	<.00005		NS	NS		NS	<.0005

Values are expressed in average ± standard deviation.

Abbreviation: NS, not significant.

Figure 3.

 Percentage of pulse wave velocity variation according to groups. IDN indicates isosorbide dinitrate.

Table IV.

 Age‐Related Pre‐ and Post‐Ischemia Percent Differences According to Groups

	All		Men		Women
	<65	≥65	<65	≥65	<65	≥65
Group 1	−1.8	−1.0	−3.00	−3.8	+0.3	−0.4
Group 2	−11.9	−6.6	−4.50	−12.9	−14.4	−8.0
P value	<.005	.07	NS	NS	<.001	<.001

Values are expressed in absolute percent of variation.

Abbreviation: NS, not significant.

The isosorbide dinitrate test was performed in 20 hypertensive patients with abnormal response. Post‐ischemia PWV showed a decrease of 14% from baseline (8.2±0.4 vs 7.1±0.4 m/s)(P<.001) (Figure 3A).

Discussion

Traditional methods assess FMD using US, which allows measuring the vessel dilation. Our study demonstrates a novel method to evaluate endothelial function by measurements of the carotid‐radial PWV as the vasodilatation response to the locally induced ischemia. It showed significant differences between hypertensive and normotensive people independent of age or SBP at baseline demonstrating the importance of the illness on the development of vascular structural and functional changes. Moreover, there was an important difference between genders in regard to FMD. Our results showed a significant difference between normotensive and hypertensive women in regard to FMD while no significant changes were observed in the male population. Normotensive men showed a less pronounced dilation after induced ischemia than women. We also found a gender difference in hypertensive patients: women had a severe impairment in FMD while men did not evidence significant differences. Interestingly, this fact was observed much more pronounced among younger women than in the older female population. Hence, we consider that normotensive women have a greater FMD in comparison with normotensive men. In contrast, we observed that the vascular damage in females was more evident than in males among the hypertensive subjects. This fact could be an explanation of the increased frequency of CV diseases in men and the greater severity of CV illness in women. Further investigations are needed about our finding.

The indirect assessment of FMD by measurements of PWV post‐ischemia changes could be a simplified and reliable tool due to the high inter‐ and intraobserver reproducibility of the method. ⁴ , ⁸

Different methods have been used to establish the structural and functional integrity of the endothelium. Measurements of the carotid intima‐media thickness (IMT) using US in subjects with severe risk factors is the most noninvasive method to assess changes in arterial wall leading to endothelial impairment. Moreover, IMT alteration closely correlates with increased CV risk. ¹⁰ , ¹¹

Endothelial dysfunction is a common factor in different pathologic processes leading to oxidative stress, inflammation, and atherosclerosis. Early finding of abnormal endothelial function is very important to consider in risk factor treatment, since any of them may alter the endothelial‐dependent dilation by avoiding the NO availability. ⁸

Vessels with normal anatomy and function respond with a significant dilation after 5 minutes of induced ischemia and a decreased PWV. Abnormal structural and functional arteries do not dilate their vascular smooth muscle significantly or have a paradoxical response (vasoconstriction instead of dilation). Hence, hypertensive and also diabetic patients have endothelial dysfunction with impaired ability to dilate after induced ischemia and PWV does not change or is faster than normal. These changes also involve other major abnormalities related to endothelial function. ¹²

There is a growing knowledge about the pathophysiological role of endothelium in the vessel tone regulation and a number of diseases may impair this response. Endothelial NO is the most important vasodilation mediator and also has antithrombotic and antiinflammatory properties (ie, inhibition of leukocyte and platelet adhesion and plasminogen activator inhibitor 1 expression). ⁹ , ¹³

The burden of data suggests that impaired endothelial dilation is an important predictor of the CV outcome independently of other risk factors. This impairment links with reactive oxygen species (ROS) production and inflammation. These findings indicate that endothelial function assessment could be used in clinical practice to identify high risk patients. ¹⁴ , ¹⁵

Endothelial function assessed by means of PWV measured before and after induced ischemia appears as a complementary tool to estimate CV risk in addition to the aortic stiffness assessed by carotid‐femoral PWV. Nevertheless, abnormal aortic PWV is a closer indicator of arterial impairment and represents a more accurate index related to CV risk. ⁴ , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ , ²¹ Naka et al. ²² in a previous study with a small number of participants showed a reduction of PWV, for the first time, after post‐ischemic hyperemia. Our study was performed in a larger number of hypertensive patients and normal patients using the computerized Complior method. The present data are in concordance with those of Naka et al and additionally showed a lack of correlation between aortic PWV with FMD assessed by pre‐ and post‐ischemic induced carotid‐radial PWV. The difference is due to the different vascular wall composition between elastic arteries (such as aorta) and muscular arteries (such as the brachial artery).

Subjects evaluated with other CV scoring methods such as European Risk Chart or Framingham score may have normal risk and abnormal PWV and FMD (also IMT) or vice versa. We consider both methods to assess endothelial function—FMD by US or PWV assessment—as complementary but independent measurements of CV risk and would have to be integrated with the already existing scoring methods to establish more accurate treatments to prevent the development of CV pathologic events.

Conclusions

Carotid‐radial pre‐ and post‐induced ischemia showed similar results to those obtained by US in previously published studies. After induced ischemia, hypertensive patients had less FMD represented by a higher PWV than healthy people. Normotensive women had a more significant FMD after induced ischemia than normotensive men. This finding could be related to a better endothelial function among healthy women than men. In contrast hypertensive women exhibited a very important endothelial dysfunction than hypertensive men. Anyway, we consider that this measurement is a reliable method to evaluate endothelial function.

Limitations of the Study

The population in our study is based in an over sample of female patients and on a scarce sample of normotensive men above 65 years old. This fact leads to a more difficult analysis between genders in older people.

Future Perspectives

The use of this method as a predictor of vascular diseases may lead to early diagnosis and treatment of patients with endothelial dysfunction. This approach, if adopted by physicians, is likely to improve the management of patients at risk for vascular diseases.

Disclosures:  The authors have no conflict of interest to report.