The reactive hyperemia test (RHtest) evokes a transient increase in shear stress as a stimulus for endothelial-dependent flow-mediated vasodilation (EDFMD). We developed a noninvasive method to create controlled elevations in brachial artery (BA) shear rate (SR, estimate of shear stress), controlled hyperemia test (CHtest), and assessed the impact of this vs. the RHtest approach on EDFMD. Eight healthy subjects participated in two trials of each test on 3 separate days. For the CHtest, SR was step increased from 8 to 50 s−1, created by controlled release of BA compression during forearm heating. For the RHtest, transient increases in SR were achieved after 5 min of forearm occlusion. BA diameter and blood flow velocity (ultrasound) were measured upstream of compression and occlusion sites. Both tests elicited significant dilation (RHtest: 6.33 ± 3.12%; CHtest: 3.00 ± 1.05%). The CHtest resulted in 1) reduced between-subject SR and EDFMD variability vs. the RHtest [SR coefficient of variation (CV): 4.9% vs. 36.6%; EDFMD CV: 36.16% vs. 51.80%] and 2) virtual elimination of the impact of BA diameter on the peak EDFMD response (peak EDFMD vs. baseline diameter for RHtest, r2 = 0.64, P < 0.01, vs. CHtest, r2 = 0.14, P < 0.01). Normalization of the RHtest EDFMD response to the magnitude of the SR stimulus eliminated test differences in between-subject response variability. Reductions in trial-to-trial and day-to-day SR variability with the CHtest did not reduce EDFMD variability. Between-subject SR variability contributes to EDFMD variability with the RHtest. SR controls with the CHtest or RHtest response normalization are essential for examining EDFMD between groups differing in baseline arterial diameter.
- reactive hyperemia
one of the most important mechanical stimuli that evokes vasodilator release from the vascular endothelium is shear stress, resulting from the movement of the blood along the endothelial cells (11). An increase in blood flow causes an increase in shear stress; the vasodilation that results from elevated shear stress is thus termed endothelial-dependent flow-mediated dilation (EDFMD) (17).
Endothelial dysfunction results in an impaired ability to dilate in response to a shear stress stimulus. Limited dilation in the face of increased flow is clinically important in coronary arteries because this may contribute to ischemic events (6, 16, 18). In addition, accumulating evidence shows that endothelial dysfunction precedes and plays a role in the pathogenesis of atherosclerosis (8, 12, 21, 33). The response in peripheral conduit arteries is well correlated with the coronary response in humans (1, 33).
The noninvasive assessment of EDFMD in humans presently involves the creation of a shear stress stimulus in the brachial or radial artery via occlusion of the forearm or hand vessels, respectively, for 5 min to create ischemia. Release of forearm occlusion results in instantaneous peak brachial artery blood velocity that quickly decays over a period of 1–2 min (reactive hyperemia). The peak brachial artery vasodilation in response to this stimulus is used as an index of endothelial function (3, 7, 10, 32).
Although this technique has proved to be a very useful tool in assessing endothelial function in humans, it has substantial limitations due to the uncontrolled nature of the evoked shear stress stimulus. First, reactive hyperemia presents as an immediate “peak” stimulus that rapidly decays over time. The duration of that decay can be manipulated by increasing the duration of occlusion (4, 28) or the amount of limb vasculature occluded (3, 4); however, it is not possible to maintain a given level of stimulus over time. In addition, the rate of onset of the stimulus cannot be controlled. Given recent evidence that mechanisms of EDFMD may vary depending on rate of onset and duration of the shear stress stimulus (5, 28), controlling these aspects of the EDFMD stimulus seems advantageous. Furthermore, with the reactive hyperemia test (RHtest), only a peak vasodilation can be clearly characterized, providing only basic information on the nature of the vasoregulatory mechanisms mediating the EDFMD response. Although changing occlusion durations to create a range of stimuli can be used to form a crude dose-response relationship, the threshold increase in the shear stress stimulus required to evoke EDFMD and the slope of the EDFMD response across a range of shear stress stimuli are difficult to assess with the RHtest protocol. Furthermore, it is impossible to quantify the time course of the EDFMD response to a given change in shear stress with reactive hyperemia. It is presently unknown 1) whether changes in these characteristics of the EDFMD response are manifest earlier than changes in peak dilation and would therefore provide a better diagnostic assessment of the presence of endothelial dysfunction and 2) to what extent the nature of underlying mechanisms of EDFMD is reflected by these response parameters.
Second, the magnitude of the shear stress imparted on the conduit arteries cannot be controlled. Conduit vessel blood velocity after temporary limb ischemia is a function of the magnitude of downstream vascular resistance, affected by the total resistance vasculature available, and the conduit vessel diameter. Thus the magnitude of the shear stress stimulus will inevitably vary between individuals and across groups with distinct differences in these two parameters. This makes comparison of the magnitude of response between such populations (e.g., gender, spinal cord injured vs. control, children vs. adults) problematic. Some researchers have begun to “normalize” the EDFMD response to some quantification of the reactive hyperemia stimulus; however, the effectiveness of this practice is unclear (23, 27).
Third, an uncontrolled stimulus may have poor trial-to-trial and/or day-to-day repeatability, and this may in turn contribute to increased response (EDFMD) variability. A large degree of EDFMD variability makes it difficult to assess changes in an individual's endothelial function after intervention, increasing the number of subjects required to detect differences between groups.
Given the key role of endothelial vasoactive and vasoprotective factors in the function and structure of blood vessels (9, 13, 14), understanding the response characteristics of EDFMD and the mechanisms that mediate them is essential to developing a comprehensive understanding of cardiovascular function and health. The purpose of our study was therefore to 1) develop a new, noninvasive technique that allows precise control of the brachial artery shear rate (approximation of shear stress without accounting for blood viscosity) stimulus, possibly leading to a more comprehensive characterization of the EDFMD response, and 2) assess the impact of a controlled shear rate (new technique) compared with the uncontrolled shear rate (reactive hyperemia) on the EDFMD response.
Eight healthy, disease-free subjects (7 men and 1 woman) between 22 and 37 yr of age from the Queen's University Physical and Health Education graduate program volunteered to participate. Health status of the subjects was confirmed with a medical screening questionnaire for risk factors associated with endothelial dysfunction. Each subject served as his/her own control. All subjects completed a consent form that was approved by the Health Sciences Human Research Ethics board at Queen's University.
Subjects repeated the experimental protocol on 3 separate days, each separated by at least 1 mo and not more than 6 wk. Two trials of each test were performed on each test day, with brachial artery diameter being allowed to return to baseline between trials. Subjects were instructed to abstain from alcohol and caffeine ingestion for 24 h before the study and to abstain from food for 2 h before the study. All experiments were performed in the afternoon between 1:00 and 3:30 PM in a temperature-controlled room (24°C).
All subjects lay supine with their arm supported at a 90° angle from their body at heart level. Continuous heart rate measurements were made throughout the experiment by three-lead electrocardiogram. Subject's forearm skin temperature was monitored with a skin temperature probe (Barnant thermistor thermometer 600-1070) secured to the heated arm. Heated air flow within an enclosed Plexiglas box sealed at the level of the olecranon process was controlled to maintain a forearm skin temperature of 42–45°C. Forearm blood flow velocity was measured with Doppler ultrasound (GE Vingmed system 5). This technique provides accurate, reliable, noninvasive, beat-by-beat measures of mean brachial artery inflow velocity to the forearm and compares favorably with both strain gauge plethysmography and thermodilution techniques (29, 30, 34). The Doppler probe was calibrated for velocity such that the voltage output was plotted for a range of known velocities (measured volumetrically as the rate of change of fluid volume in a collecting container) passing through a known tube diameter. The linear relationship between voltage output and velocity was highly repeatable.
The ultrasound probe was placed over the brachial artery in a position that produced the strongest arterial blood velocity signal, with no interference from adjacent veins. Once in position, the probe was secured with a clamp stand and a guide adhered to the skin. The brachial artery was imaged by two-dimensional gray-scale ultrasound imaging (GE Vingmed system 5). The probe operator was able to make minor corrections to probe placement to maintain an optimal velocity signal and vessel image throughout the experiment. Images were stored on videotape for future analysis of diameters.
EDFMD stimulus quantification.
Shear stress is a function of vessel diameter, blood flow velocity at the vessel lumen, and blood viscosity, such that shear stress = ηV/D (17), where V is mean blood flow velocity, D is vessel diameter, and η is blood viscosity.
Shear rate is an estimate of shear stress without accounting for blood viscosity (shear rate = mean blood velocity/vessel diameter). Our laboratory is presently not equipped to measure blood viscosity; therefore, we examined the effect of controlling the shear rate estimate of shear stress on EDFMD responses. Future references to the stimulus for EDFMD will refer to shear rate estimates of shear stress, with the understanding that shear stress is the actual EDFMD stimulus.
Reactive hyperemia test.
RHtest is a standard endothelial function test in which baseline brachial artery blood flow and diameter are recorded for 1 min, followed by 5-min cuff inflation at 300 mmHg to occlude brachial artery blood flow. In our study, the cuff was positioned just above the elbow and ∼10 cm distal to the site of brachial artery ultrasound measurements. After 5 min of occlusion, the cuff was released, after which blood velocity and diameter recordings were continued for another 5 min. This method is described by Celermajer et al. (8). This test was performed twice in succession on each experiment day, with sufficient time between trials to allow brachial artery diameter to return to baseline (Fig. 1).
Controlled hyperemia test.
After the two trials of the RHtest were completed, the pressure cuff was removed and replaced by a pneumatic piston placed over the same region of the brachial artery. Care was taken to ensure that piston position allowed compression of the artery, reducing blood flow velocity adequately to achieve a baseline target shear rate of 8 s−1 without resulting in more than minimal discomfort (Fig. 2). Brachial artery blood flow velocity and diameter were monitored throughout this controlled hyperemia text (CHtest) protocol. The blood flow velocity values were displayed on the (online) data acquisition system as a 5-s moving average, acting as feedback for the piston operator to guide regulation of brachial artery compression. Maintenance of the desired shear rate in the CHtest was achieved by measuring the brachial artery diameter and calculating a target mean blood velocity (target velocity = desired shear rate × measured brachial artery diameter). The piston was pressurized, compressing the brachial artery to control blood flow velocity at a target that created a shear rate of 8 s−1. The target velocity was maintained by carefully adjusting the compression of the brachial artery based on the continuous 5-s average mean blood velocity feedback. This compression was maintained for 5 min at which time a “compression baseline” was recorded. The elevated temperature air flow was then applied for 30 min, maintaining the skin temperature between 42 and 45°C. After 29 min of heating, 1 min of “heated baseline” brachial artery diameter was recorded. The shear rate was then increased to 50 s−1 and maintained for 5 min. The shear rate was then reduced back to 8 s−1 and maintained until diameter returned to baseline. When the diameter returned to the baseline value, the shear was increased to 50 s−1 again and maintained for 5 min (Fig. 1).
Full piston release.
In a subset of six subjects, all arterial compression after the second trial of the CHtest was removed. Diameter and velocity data were recorded for 4 min after release. Data were analyzed as for all other trials.
Blood pressure was analyzed in a subset of four subjects after the completion of the main experiment because subject position prohibited blood pressure measurement during the study. Subjects underwent the same protocol as described above but without recording of diameter data. Blood pressure measurements were made via arterial tonometry (Colin 7000, Trudell Medical, London, ON).
Blood velocity was analyzed offline, averaged over 3-s time bins. Vessel diameter was measured manually from the videotaped two-dimensional B-mode ultrasound images. Briefly, on-screen calipers were placed on a clear section of the vessel wall at three points on a frozen screen image of the brachial artery during diastole and then averaged to generate the vessel diameter for that time interval. Diameter measurements were taken every 10 s for the first minute and every 30 s for the remaining 4-min post-cuff/piston release. One operator performed all of the diameter analyses. Baseline diameters were the average of caliper measurements from three frozen screens during a 1-min baseline diameter recording. The repeatability of our baseline diameter measurements (see results) illustrates this laboratory's ability to obtain accurate diameters.
To assess the impact of “normalizing” RHtest responses to the stimulus elicited by reactive hyperemia, the RHtest peak dilation response was divided by either the shear rate 1-min area under the curve (AUC) or the peak shear rate stimulus.
We plotted raw diameter data over time and fit these with an exponential function using custom curve-fitting software. This software allowed the user to select the type of exponential function (2 components for the RHtest, 1 component for the CHtest) and manipulate the function parameters (time delay and time constant) to obtain a curve fit with minimized mean squared error. The residuals of this fit were visually examined to ensure equal positive and negative distribution along the entire length of the fit. With the use of the exponential function parameters, a diameter estimate every 3 s was then obtained. This was combined with the 3-s average of the velocity data in the equation V/D to produce a shear rate 3-s average.
The most accurate characterization of the true stimulus for EDFMD when using reactive hyperemia is unknown. However, both the peak shear rate attained and the duration of the stimulus appear to be important (22). For comparison between tests, we elected to quantify the stimulus as the AUC of the shear rate over time profile in the first minute after cuff/piston release. The 1-min interval was selected because peak diameter in the RHtest was achieved by ∼60 s. The response was defined as the peak percent change in diameter from baseline.
To assess both stimulus and response variability, individual coefficients of variation (CVs) were calculated as the standard deviation (SD) of a subject's data across trials or days divided by the mean of the same data × 100. The average of these individual CVs was then reported (±SD). Between-subject CVs were calculated as the SD of all subject data divided by the mean of all subject data × 100.
Two- and three-way repeated-measure ANOVAs were used to compare the magnitude and variability of the stimulus and response as well as to compare the diameters before and after arterial compression and heat. Linear regression analysis was used to assess the relationship between the stimulus and the response and the relationship of baseline diameter to peak response for both tests.
Compression and upstream vasodilation.
To rule out a mechanically induced brachial artery vasodilation as a result of piston pressure, the brachial artery diameter was measured before compression and after 5 min of compression before the introduction of the heat stimulus. It was found that this postcompression diameter (mm) was slightly but significantly reduced vs. the precompression diameter (4.46 ± 0.51 vs. 4.511 ± 0.50 mm, P = 0.008). This indicates that there was no upstream dilation of the brachial artery with arterial compression.
Heat and upstream vasodilation.
To rule out a heat-induced, conducted signal from the forearm as a cause of brachial artery vasodilation, the brachial artery diameter was measured after 5 min of compression without the heat stimulus (compressed baseline) and then after the 30 min of combined compression and heat (heated baseline). Brachial artery diameter was not significantly different in the compressed baseline vs. heated baseline conditions (4.48 ± 0.51 vs. 4.46 ± 0.51 mm, P = 0.299), indicating that forearm heating did not induce upstream brachial artery dilation.
Shear rate stimulus profiles.
To compare the sustainability of the shear rate stimulus over time between the RHtest and the new CHtest, the baseline and postrelease shear rates were expressed as 20-s average time bins. In the RHtest, the average shear rate was only significantly different from baseline during the first 80-s postcuff release (first four 20-s average time bins, see Fig. 3A) (P < 0.001). In addition, each of the first four 20-s time bins were significantly different from each other (P < 0.001). These results speak to the short-lived and constantly changing nature of the RHtest shear stress stimulus (Fig. 3A). In addition, the peak EDFMD occurred well after the peak shear rate had occurred and was approaching resting levels (Fig. 4A).
In contrast, the average shear rate for all postcompression release 20-s periods for the CHtest was significantly different from baseline (baseline of 8.58 ± 2.74 s−1 vs. postrelease average of 48.36 ± 1.90 s−1, P = 0.001) (Fig. 3B). In addition, there was no significant difference in the average shear rate during any 20-s period postrelease (P > 0.05). This indicates that the shear rate stimulus was controlled and sustained throughout the postcompression release period. There was no effect of day (1, 2, or 3) or trial (1 or 2) on the average 20-s time bin shear rate for either test.
Stimulus and response magnitudes.
The AUC of the shear rate stimulus was quantified for the first 60-s postcuff/piston release because this represented the stimulus preceding the peak EDFMD response in the RHtest. There was a significantly greater shear rate stimulus provided by the RHtest vs. the CHtest in this time period (5,812.94 ± 2,027.75 vs. 2,844.25 ± 147.28, P = 0.003) (Fig. 3).
There was a significant difference between the baseline vs. peak brachial artery diameter in both tests; thus each evoked a significant degree of dilation (base vs. peak for the RHtest = 4.45 ± 0.52 vs. 4.72 ± 0.46 mm, P < 0.001; base vs. peak for the CHtest = 4.47 ± 0.50 vs. 4.62 ± 0.50 mm, P < 0.001). The day of the test (1, 2, or 3) and the trial (1 or 2) had no effect on the diameter. The degree of dilation (percent change in diameter) evoked by the RHtest vs. the CHtest was not significantly different (6.33 ± 3.12% for the RHtest vs. 3.00 ± 1.05% for the CHtest; P = 0.011).
During 4 min of full piston release after the second 5-min 50 s−1 shear rate step in a subset of six subjects, the average shear rate achieved was significantly greater than the average shear rate during the CHtest (90.87 ± 17.42 s−1 vs. 48.36 ± 1.90 s−1; P = 0.002). Correspondingly, the peak percent change in diameter during full piston release was significantly greater than that during the CHtest (8.74 ± 2.91% vs. 2.91 ± 0.80%; P < 0.001) and not different from that achieved during the RHtest (7.23 ± 3.08%; P = 0.374) in the same subset of six. Thus the CHtest technique may be capable of creating a shear rate stimulus that evokes as much dilation as the traditional RHtest method.
Subject's baseline diameter is considered to be constant within one data collection session (20); therefore, the repeatability of this measurement may be used to represent the accuracy of diameter measurements. The trial-to-trial CV for baseline diameter was 0.44 ± 0.62%, which represents an average absolute difference between trials of 0.027 mm. This value represents only 0.61% of the 4.46-mm average baseline diameter.
Stimulus response relationship.
Linear regression analysis of the peak EDFMD response vs. the shear rate 1-min AUC in the RHtest across all trials and days revealed a significant, positive relationship (slope = 1.05, r2 = 0.47, P < 0.001) (Fig. 5A). Thus stimulus magnitude contributed substantially to response magnitude in the RHtest. In contrast, regression analysis of the peak EDFMD response vs. mean shear rate over 5 min in the CHtest across all trials and days demonstrated no relationship (slope = −1.02, r2 = 0.018, P < 0.36) (Fig. 5B). This is to be expected because we controlled the shear rate magnitude in the CHtest.
Effect of shear rate control on between-subject EDFMD response variability.
As expected, there was considerable between-subject variability in the RHtest shear rate stimulus (expressed as a CV = 36.61% calculated from the 1-min AUC data across all trials and days) (Fig. 6E). In contrast, efforts to control shear rate in the CHtest resulted in much lower between-subject variability (CV of 4.87%) (Fig. 6E). This substantial reduction in between-subject shear rate variability with the CHtest resulted in a reduction in the between-subject variability for the EDFMD response (expressed as CV, 36.16% for CHtest vs. 51.80% for RHtest) (Fig. 6F).
When the EDFMD response was normalized to our measure of the magnitude of the stimulus (60-s AUC) using the equation of peak percent change in diameter/60-s AUC, the between-subject CV was shown to be similar in the RHtest and CHtest (36.87 vs. 35.67%). Normalization of the response to the peak shear rate stimulus provided similar results (RHtest between-subject CV of 37.11%).
A further aspect of between-subject variability explored in this study was that related to baseline brachial artery diameter. Subjects demonstrated a range of brachial artery diameters (3.35–5.19 mm). We observed a significant, negative, linear relationship between the baseline diameter and the shear rate 1-min AUC (slope = −2,497.02, r2 = 0.42, P < 0.001) in the RHtest, whereas no relationship was found between these variables in the CHtest (slope = −4.89, r2 = 0.0003, P = 0.91). In addition, we observed a significant linear relationship between the baseline diameter and the peak dilatory response in the RHtest (slope = −4.75, r2 = 0.64, P < 0.01), indicating that 64% of the variance between subjects in the peak dilatory response to the RHtest could be explained by differences in baseline brachial artery diameter (Fig. 7A). In contrast, only 14% of the variance was accounted for by baseline diameter in the CHtest, where shear rate was controlled (slope = −0.77, r2 = 0.14, P < 0.01) (Fig. 7B) The relationship of baseline diameter to the normalized (peak response/shear rate 1-min AUC) dilatory response in the RHtest was substantially weaker than before normalization and bore a much closer resemblance to the CHtest results (r2 = 0.185, P = 0.002) (Fig. 7C) Normalization of the RHtest to the peak stimulus provided comparable similar results (r2 = 0.343, P = 0.001) (data not shown).
Effect of shear rate control on trial-to-trial EDFMD response variability.
There was significantly less trial-to-trial shear rate variability (expressed as CVs calculated from the 1-min AUC data) in the CHtest vs. the RHtest (3.44 ± 2.43% vs. 6.45 ± 6.16%, P = 0.045). Day (1, 2, or 3) had no significant effect on the trial-to-trial shear rate variability (Fig. 6A). There was no significant difference in the trial-to-trial EDFMD variability between the RHtest vs. the CHtest (14.67 ± 15.03% vs. 8.32 ± 7.05%, P = 0.16) (Fig. 6B). There was also no effect of the testing day (1, 2, or 3) on the trial-to-trial EDFMD variability.
Effect of shear rate control on day-to-day EDFMD response variability.
There was significantly less day-to-day shear rate variability (expressed as CV calculated from the 1-min AUC data) in the CHtest vs. the RHtest (4.10 ± 2.37% vs. 12.83 ± 7.92%, P = 0.011) (Fig. 6C). Trial (1 or 2) had no significant effect on the day-to-day shear rate variability. Despite the reduction in day-to-day shear rate variability in the CHtest, there was no significant reduction in the day-to-day EDFMD variability compared with the RHtest (21.22 ± 16.90% for CHtest vs. 27.8 ± 14.37% for RHtest, P = 0.187) (Fig. 6D). Thus reductions in day-to-day shear rate variability with the CHtest did not result in reduced day-to-day EDFMD variability.
Differences in trial-to-trial and day-to-day stimulus and response variability.
The day-to-day shear rate variability in the RHtest was significantly larger than the trial-to-trial shear rate variability, whereas no increase was observed between trial-to-trial and day-to-day shear rate variability in the CHtest (6.44 ± 6.16 vs. 12.87 ± 7.92% for RHtest, P = 0.016; 3.44 ± 2.43 vs. 4.10 ± 2.37% for CHtest, P = 0.781). The day-to-day EDFMD variability was significantly greater than the trial-to-trial EDFMD variability for both tests (27.8 ± 14.37 vs. 6.45 ± 6.16% for RHtest, P = 0.050; 21.22 ± 16.9 vs. 3.44 ± 2.43% for CHtest, P = 0.001).
Subject's average heart rate was significantly greater during the CHtest vs. the RHtest (62.69 ± 8.32 vs. 57.99 ± 7.59 beats/min, P = 0.018). There was no effect of day or trial. Heart rate did not change over the postrelease period in either test.
Subject's average blood pressure was not significantly different in the RHtest vs. the CHtest (68.10 ± 4.69 vs. 66.29 ± 9.28 mmHg, P = 0.976). Baseline blood pressure was also not significantly different from blood pressure during cuff/piston release (68.33 ± 7.20 vs. 67.19 ± 6.88 mmHg, P = 0.246).
This study sought to develop a noninvasive technique (CHtest) that would allow controlled changes in brachial artery shear rate and the ability to assess the impact of this CHtest vs. the traditional RHtest on the EDFMD response. The primary, novel findings were fourfold. First, combined forearm heating and arterial compression provided an effective new approach for achieving a shear rate target rapidly and then successfully maintaining that target. Second, targeting uniform shear rates across subjects virtually eliminated the between-subject baseline diameter difference effect on the magnitude of the EDFMD response observed with the traditional RHtest. Third, shear rate stimulus variability contributed significantly to the between-subject but not to the within-subject response variability observed with the traditional RHtest. Fourth, normalization of the EDFMD response in the RHtest to either the peak or the 60-s AUC of the reactive hyperemic stimulus minimized the impact of stimulus variability on the EDFMD response as effectively as actual stimulus control.
CHtest: is shear stress the sole stimulus for observed brachial artery dilation?
Given that the endothelium is capable of conducting a vasodilatory signal up the vascular tree (2, 9), we were concerned that the downstream heat and arterial compression, which are a part of the CHtest, might produce vasodilation in the brachial artery at the upstream site of measurement. The absence of brachial artery vasodilation with either compression or heat in the CHtest confirmed that these interventions did not provide a vasodilatory stimulus. This agrees with the work of Mullen et al. (28) and Joannides et al. (24), who found that hand warming had no effect on brachial artery diameter if flow was not allowed to rise. Thus we believe that the stimulus for brachial artery dilation with the CHtest is solely the flow-mediated increase in shear stress at the upstream vessel imaging site.
Impact of stimulus duration.
The RHtest stimulus was large but transient and elicited a 6.33 ± 3.12% change in diameter. The CHtest produced a smaller, sustained stimulus that elicited a 3.00 ± 1.05% change in diameter. Other studies that have used both hand warming and reactive hyperemia have produced smaller stimuli with warming but observed larger responses (24, 28). The reason for this discrepancy is presently unknown.
The EDFMD response to the RHtest is thought to be mediated by the release of nitric oxide, as dilation to this brief stimulus is almost completely abolished in the presence of the nitric oxide synthase blocker N-monomethyl-l-arginine (25, 28). However, Mullen et al. (28) found that peak dilation in response to more sustained shear stress stimuli evoked by either 15 min of wrist occlusion, hand warming, or acetylcholine infusion was unaffected by N-monomethyl-l-arginine infusion or cyclooxygenase inhibition. Thus the mechanism(s) for dilation in response to a sustained shear stress stimulus remains to be determined.
This indicates that the peak dilatory response may be produced via different mechanisms, depending on the duration of the stimulus. Because the CHtest can provide a sustained, controlled stimulus increase, examining changes in the time course and time delay of the diameter adaptation under conditions where putative EDFMD dilatory mechanisms are blocked should provide critical new insight into the contribution of mechanisms to the initial vs. sustained response.
Effect of shear rate stimulus control on the between-subject EDFMD response variability.
Blood flow velocity through a conduit artery is a function of downstream vascular conductance and conduit artery cross-sectional area. Therefore, between-subject differences in 1) baseline diameter and 2) forearm vascular conductance after forearm occlusion are responsible for the large between-subject shear rate variability observed in the RHtest. The impact of baseline diameter on the magnitude of the brachial artery shear rate is evidenced in this study by the strong negative linear relationship between baseline diameter vs. 1-min shear rate AUC. In contrast, there was a strong positive relationship between 1-min shear rate AUC and EDFMD (Fig. 5A). The shear stress dependency of EDFMD has been observed by others (27, 31) and is to be expected, given that shear stress is recognized as a stimulus for EDFMD (11). The strong negative relationship between baseline brachial artery diameter and EDFMD (Fig. 7A) illustrates that brachial artery baseline diameter affected EDFMD via its effect on shear rate in the RHtest, such that small diameters experienced a larger shear rate and thus exhibited larger responses. Thus, in agreement with others (27, 31), we conclude that the shear rate variability across subjects due to between-subject baseline diameter differences contributed substantially to the EDFMD variability between subjects.
In contrast to the RHtest, the shear rate was well controlled across the range of subject baseline diameters in the CHtest (Fig. 3A), where the target velocity was calculated to offset differences in brachial artery diameter and artery compression controlled for differences in downstream conductance. Thus the CHtest demonstrated only a weak relationship between vessel diameter and EDFMD because the shear rate was controlled within a narrow range for all subjects (Fig. 7B).
These data have identified a major weakness in using unnormalized reactive hyperemia-elicited EDFMD responses as a tool for comparing EDFMD responses between groups differing in arterial cross-sectional area [e.g., gender comparisons (24, 27), spinal cord vs. control, children vs. adults]. This weakness is virtually eliminated with the new CHtest in which arterial mean blood velocity can be controlled.
Importantly, in a more recent study of EDFMD that used reactive hyperemia (27), the impact of stimulus variability on response variability was acknowledged and the dilation response normalized to the degree of stimulus created by reactive hyperemia. However, although it has been acknowledged that the stimulus is important, it is presently unknown what aspect of the reactive hyperemia profile (peak, duration, AUC) best represents the stimulus for EDFMD (23). When the EDFMD responses were normalized to the AUC in the first 60-s postcuff release, the RHtest and the CHtest exhibited a similar between-subject variability, indicating that normalization in this fashion eliminates the impact of stimulus variability on response variability to the same degree as actually controlling the stimulus. However, normalization of the RHtest dilation response to the peak shear rate stimulus was similarly effective. In addition, normalization of the RHtest peak dilation response to the shear rate 1-min AUC or the peak stimulus eliminated the majority of the relationship between baseline diameter and dilation response in the RHtest, creating a relationship similar to that seen in the CHtest (Fig. 7C). Thus normalization removes the impact of baseline diameter on response variability with comparable efficacy to shear rate stimulus targeting. This study thus provides support for normalizing the dilation in response to a 5-min occlusion-induced reactive hyperemia stimulus using either the 60-s AUC or the peak shear rate stimulus. This experiment provides the first direct confirmation of the validity of EDFMD response normalization.
Effects of stimulus control on the within-subject response variability.
The CHtest had a lower within-subject shear rate variability than the RHtest on both a day-to-day and a trial-to-trial basis. Furthermore, the RHtest but not the CHtest showed increases in shear rate variability as a function of time (day-to-day) between tests. Because baseline brachial artery diameter did not change across days, the shear rate variability across days in the RHtest must be due to differences in the downstream vascular conductance response to forearm occlusion, a factor that is controlled for in the CHtest.
The reduction in trial-to-trial shear rate variability in the CHtest vs. the RHtest did not result in a significant reduction in within-subject EDFMD trial-to-trial variability compared with the RHtest. Furthermore, despite the unchanged shear rate variability in the CHtest from trial-to-trial to day-to-day, both tests had increased EDFMD variability between days vs. between trials. This could be because, on a within-subject level, shear rate variability does not contribute significantly to day-to-day EDFMD variability. Instead, day-to-day EDFMD variability may be dominated by physiological factors (20).
The relatively small amount of dilation observed with the CHtest (3.00 ± 1.05%) is a potential limitation of this study. However, given that the repeatability of our baseline diameter measurements indicates that there was minimal experimenter error and given that many of the EDFMD responses in the CHtest were in the range observed with the RHtest (RHtest: 2–12.24% vs. CHtest: 1.21–5.28%), we believe that this limitation is unlikely to have been a confound of the main finding in this study. Nevertheless, future studies would benefit from the use of a larger shear rate stimulus, as demonstrated in the full release trials. We chose the 50 s−1 shear rate because initial pilot testing indicated that it was the highest shear rate that we would likely achieve in all subjects. However, the subgroup of six individuals who experienced full piston release achieved much greater shear rates and, in turn, a much greater degree of dilation. Thus future studies could take advantage of a higher target shear rate.
In this study, we used shear rate (mean blood velocity/vessel diameter) as an estimate of shear stress and found that shear rate control resulted in a substantial decrease in between-subject EDFMD response variability. In reality however, shear stress is also affected by blood viscosity (15), and differences in viscosity within and between subjects contribute to stimulus variability. Future studies that incorporate viscosity measures into the target velocity equation [e.g., target velocity = (desired shear rate × vessel diameter)/blood viscosity] should improve assessment of the stimulus-to-response relationship of EDFMD with the new CHtest.
Sympathetic activation has been shown to blunt the EDFMD response (19), and thus the elevated heart rate in the CHtest was a concern. However, given that blood pressure was constant throughout both tests, it is likely that the elevated heart rate was due to parasympathetic withdrawal rather than to sympathetic activation and thus did not affect our diameter results.
We did not account for the changing diameter during arterial compression release. The changing diameter without blood flow velocity adjustment during this period would necessarily result in a reduction in the shear rate achieved by the end of the release. The average reduction in shear rate at the end of the 5-min release period was only 1.5 s−1 and therefore is unlikely to have had an impact on the results. Thus our treatment of shear stress as “constant” during the 50 s−1 step was not problematic.
It should also be noted that the technique reported in this study only allows the investigation of peripheral conduit arteries; although it has been demonstrated that their endothelial-dependent dilation is well correlated to that of coronary arteries (1, 33), the results from this technique should not be generalized to resistance arterioles.
EDFMD response variability arises from three sources: measurement error, physiological variability in endothelial responsiveness, and stimulus variability. Despite shear rate control with the CHtest and the attempt to control for immediate history (food intake, exercise, time of day of testing) across days, a substantial degree of day-to-day EDFMD variability (increased compared with trial-to-trial) was observed. This is an important observation because it indicates the likelihood of a substantial involvement of physiological variability involved in the EDFMD response, as suggested by Hijmering et al. (20). The CHtest's potential to “isolate” physiological variability as a source of response variability could be applied in future studies to assess the impact of specific variables thought to influence endothelial responsiveness.
Although control of EDFMD stimulus parameters (rate of increase and decrease in magnitude) has been possible in vitro and in situ (5, 26), the CHtest represents the first technique capable of such control in vivo in humans. Thus this technique may provide the opportunity for comprehensive investigation into the nature of the relationship between shear stress and EDFMD in human conduit arteries and the mechanisms involved. This can include 1) defining the dose-response relationship of shear stress: EDFMD; 2) identifying the threshold stimulus required to elicit EDFMD; 3) characterizing the dynamic response characteristics of EDFMD: time constant, time delay, and dynamic linearity vs. nonlinearity (Fig. 4B); 4) understanding the role of the rate of change in shear stress on EDFMD; 5) linking the contribution of different EDFMD mechanisms to the characteristics of the shear stress stimulus in human conduit arteries.
A forearm heating and arterial compression combination provides an effective new approach for achieving and maintaining a shear rate target. We make these conclusions: 1) shear rate stimulus variability contributes substantially to between-subject but not to within-subject response variability in the traditional RHtest, 2) targeting uniform shear rates across subjects virtually eliminates the between-subject baseline diameter difference effect on the magnitude of the EDFMD response observed with the traditional RHtest, 3) normalization of EDFMD responses to either the peak or the 60-s AUC of the reactive hyperemic stimulus effectively eliminates the impact of stimulus variability on response variability, and 4) normalization of the EDFMD response in the RHtest to either the peak or the 60-s AUC of the reactive hyperemic stimulus eliminates the impact of stimulus variability on the EDFMD response as well as actual stimulus control.
This study was supported by an operating grant from the Natural Sciences and Engineering Research Council of Canada and infrastructure grants from the Canada Foundation for Innovation and the Ontario Innovation Trust. K. E. Pyke was supported by an Ontario Graduate Scholarship.
We thank all of the study participants for their time and patience.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2004 the American Physiological Society