Changes in internal carotid and vertebral arterial wall stiffness with head movement can be detected with shear wave elastography

ABSTRACT

Background: Safe practice is important for patients with neck pain, with the potential for injury to cervical arteries. Cervical manipulation or end range techniques/positions may place considerable strain on the arteries. Altered integrity of the arterial wall may render them more susceptible to minor trauma, particularly in the upper cervical region. Screening of blood flow velocity is limited for predicting those at risk. Examining properties of the cervical arterial wall (stiffness characteristics) and their response to head movement may provide an alternate measure of arterial susceptibility.

Objectives: To investigate whether shear wave ultrasound elastography can detect any changes in internal carotid (ICA) and vertebral (VA) arterial wall stiffness in neutral compared with contralateral head rotation.

Design: Observational study

Methods: Shear wave ultrasound elastography was used to measure the stiffness of the ICA and VA. Shear wave velocity (m/s), indicative of arterial stiffness, was measured in both arteries proximally (C3–4) and distally (C1–2) in neutral and contralateral head rotation as were intimal thickness (mm) and flow velocity (cm/s).

Results: Thirty participants (20–62 years) were successfully imaged. The VA was stiffer than ICA and it became significantly stiffer in contralateral rotation (p = 0.05). The ICA became significantly less stiff (p = 0.01). Effects were more apparent at C1–2 but significant in the ICA only (p = 0.03). Flow velocity and intimal thickness were unchanged in rotation.

Conclusions: Changes in VA and ICA arterial wall stiffness can be measured with shear wave ultrasound elastography. This measure may ultimately help identify arteries with greater vulnerability to rotational stresses.

Introduction

Safe practice is an important consideration when treating patients with neck pain because of the potential for adverse neurovascular effects caused by increased strain on vulnerable arterial structures. A rare but potentially serious consequence of increased stress on the arteries is cervical arterial dissection (CAD), which may lead to stroke. This is a catastrophic complication of neck manipulation and occurs most commonly in young to middle-aged people [1]. The pathophysiology of CAD typically involves altered integrity of the cervical arterial walls. Increased arterial stiffness has been proposed as a potential risk factor for CAD. Imaging and modeling studies indicate that people with CAD may have up to 58% stiffer cervical arteries than healthy individuals [2,3]. This may render them more susceptible to minor trauma or neck strain [4,5], particularly with head movement where the arteries may be particularly vulnerable [6].

Biomechanical studies, mainly cadaveric or animal, have demonstrated that end of range cervical movements (e.g. full rotation) are potentially risky movements for the cervical arteries as they place considerable stress on the artery [7,8]. It is possible that higher stiffness and greater change in stiffness in head rotation might indicate greater stress on the artery, which could represent a risk factor for neck manipulation. This hypothesis has not been tested in vivo and at this time, it is unknown how cervical movements could influence cervical arterial stiffness in vivo.

Recently a noninvasive and relatively inexpensive procedure, shear wave ultrasound elastography, was used to measure internal carotid (ICA) and vertebral arterial (VA) stiffness in vivo in a neutral head position. The measure provides an estimation of local arterial stiffness [9] and was shown to have good intra-tester reliability, ICA (ICC [2,1] 0.8, CI 0.5 to 0.9) and VA (ICC 0.8, CI 0.4 to 0.9) and low measurement error (ICA 0.2, VA 0.3 m/s) [10]. However, it is not known whether elastography can also detect changes in arterial stiffness with head movement and what the nature of any changes are. Importantly, conventional medical screening for cervical arterial integrity, using ultrasound-based methods, includes measurement of blood flow velocity and intima-medial thickness, to detect vessel stenosis and plaque formation but such screening is limited in detecting more subtle alterations in arterial wall integrity [11].

Therefore, the aim of this preliminary study was to determine whether elastography can detect a change in stiffness in the VA and ICA walls with contralateral cervical rotation and if so, provide preliminary insight into the nature of any change. A secondary aim was to document conventional measures of arterial compliance (blood flow velocity and intimal wall thickness) to determine if they paralleled any changes in stiffness in head rotation.

Methods

Design

This was an observational study undertaken at the University of Queensland between April 2015 to April 2016. Institutional ethical approval was obtained; no. 2015000326 and all participants provided their written informed consent.

Participants

Healthy adults over 18 years were recruited into the study from the university community via advertisement and personal contact. Participants were suitable for inclusion if they did not currently have headaches, neck pain or restriction of neck movement and had a Body Mass Index (BMI) within normal limits (18.5–24.9 kg/m²) [12]. Exclusion criteria included any cardiovascular risk factors (smoking, hypertension, oral contraception, hypercholesterolemia), which could potentially confound ultrasound findings; any connective tissue disease with known vascular component (e.g. Erhler’s Danlos Syndrome), which might influence arterial stiffness; recent major illness, surgery or neurological conditions. Participants were also excluded if they exhibited possible symptoms of vertebrobasilar insufficiency (VBI) such as dizziness/light-headedness, diplopia, nausea, and signs of nystagmus with positional testing [13].

Demographic and physical characteristics

Demographic details, general health and past medical history were collected using a standard form to ensure inclusion criteria were met and no reasons for exclusion were identified. Self-reported height and weight measures were collected and BMI was calculated. Active cervical rotation was assessed by a physical therapist to ensure full range pain-free movement. VBI testing was performed according to a standard protocol [13], sustaining the participant’s head at end range rotation, followed by extension for 10 s, to check for any symptoms indicative of VBI, which might limit participation. Seated blood pressure was measured on the right arm with an automated sphygmomanometer (OMRON, Hoofddorp, The Netherlands), and was repeated in supine to identify any postural effects on blood pressure, which might have confounded results.

Participants attended for a single, 45-min ultrasound session undertaken by an ultrasonographer with 25 years of experience. Images were acquired using the Siemens Acuson S3000 Virtual Touch^TM Imaging Quantification system (VTIQ) (Erlangen, Germany) using a 9 MHz (9L4) linear transducer. All measures were conducted in a quiet, darkened room.

Arterial characteristics

The primary measure was arterial stiffness (m/s) measured with shear wave elastography in neutral and contralateral head rotation. We used shear wave velocity as a measure of stiffness rather than shear modulus to overcome some of the limitations posed by anisotropy (directional variation in tissue stiffness) and wave dispersion (shear waves being reflected off the vessel wall boundaries) in vessels [14,15]. Secondary conventional measures of arterial compliance were blood flow velocity (Peak systolic [PS] and End diastolic [ED] velocity) (cm/s) and intimal wall thickness (mm) of the far wall of the artery, which was measured using calipers pre-programmed in the ultrasound system.

Image processing

The VTIQ system is optimized for vascular applications and provides quantitative measures of shear wave velocity (m/s) and qualitative color-coded maps (elastograms) where higher velocities, indicative of stiffer tissues are displayed in red, and slower, less stiff tissues in blue. Data is formed by a pulse sequence made up of up to 256 beam lines for each RF pulse. A ‘shear wave quality mode’ identifies whether the shear wave is of sufficient magnitude with an adequate signal-to-noise ratio to accurately estimate velocity within the region of interest (ROI); values >5–10% variation from mean values are disregarded [16]. This feature is important particularly when imaging small vessels to ensure accurate velocity measures, as the ROI may include part of the vessel lumen, which could confound results. Higher velocity values correspond to a stiffer arterial wall [17].

Procedure

The imaging procedure has been described previously [10]. Briefly, participants laid supine with their head supported in a neutral position and the right side vessels were imaged with the probe positioned longitudinally. Estimation of stiffness was based on the velocity of the shear-wave propagating longitudinally along the far arterial wall. Using the far wall accessed the target ROI through a fluid interface to limit reflection at tissue interfaces. We used one wall to indicate the stiffness of the whole artery. Both ICA and VA were measured at a proximal site (C3–4 level just distal to the carotid bulb-standard position) and a distal site at the atlanto-axial level (region more vulnerable to dissection). The experimental set-up is shown in Figure 1.

Figure 1.

(a and b) Experimental set-up.

The carotid bulb was located using B-mode ultrasound and the right proximal ICA identified immediately distal at the level of C3–4. Identification was confirmed with the use of Doppler ultrasound using a standard protocol with the transducer positioned longitudinally and at 60° to the vessel. The ICA was confirmed by its distinctive low-resistance waveform [18]. Shear wave velocity mode was then selected. The focal depth was adjusted to position a single voxel ROI (one intersegmental level) over the intima of the vessel wall level with C3–4 (Figure 2(a, b)). A radiofrequency pulse was applied at 90° to the vessel wall. Shear wave velocity (m/s) was recorded from three voxels (Figure 2(a,b)) positioned along part of the arterial wall with a highest quality image (established by the ‘quality mode’ function). The average of the 3 voxels was used in the analysis. Cardiac gating was not used as averaged velocity measures accounted for any tissue variation and the cervical arteries are low-resistance vessels so cardiac cycle variation should be small [15]. The transducer was then moved as far distally as possible along the artery to the C1–2 level and measures were repeated using the same protocol.

Figure 2.

(a and b) Elastogram showing distal ICA in neutral and contralateral rotation. Yellow boxes show voxels positioned along the far arterial wall from which shear wave velocities are measured.

For the VA, the transducer was angled posteriorly, maintaining a 90° angle, to identify the vessel between the cervical transverse processes. The artery was confirmed with Doppler ultrasound and blood flow velocity measures recorded. Shear wave velocity measures (average of 3 voxels) were obtained from the VA wall at C3–4 and C1–2 levels.

The participants then turned their head as far as possible to the contralateral (left) side. Range of rotation (degrees) was measured with an inclinometer (Baseline Bubble Inclinometer, New York, USA) placed on the mid-point of the forehead. Measurements of shear wave velocity, blood flow velocity and intimal thickness were then repeated at the proximal and distal sites of the ICA and VA in the rotated position (Figure 3(a,b)).

Figure 3.

(a and b) Elastogram showing distal VA in neutral and contralateral rotation.

Statistical analysis

Descriptive statistics were calculated for all outcomes. Measures of ICA and VA stiffness, flow velocity and intimal thickness in the neutral and contralateral rotation positions were tested for normality, and outcomes reported as means (SD). For each artery, the mean difference between neutral and rotated positions was analyzed for the whole artery (combined proximal and distal measures) and proximal and distal sites. Percentage change between positions for both sites was also calculated. Any differences in stiffness, blood flow velocity and wall thickness between head positions (neutral and contralateral rotation) were explored using a repeated-measures analysis of covariance (ANCOVA), with age as a covariate. Mean differences were reported with 95% confidence intervals. The significance level was set at p ≤ 0.05. Data were analyzed using IBM SPSS for Windows (release 25).

Results

Thirty participants (17 males, 13 female) mean age 29.8 (± 12.8), range 20–62 years entered the study. Blood pressure remained stable between positions. All participants were in good health with no major CV risk factors. All participants had full range, pain-free neck rotation and achieved >60 degrees of contralateral rotation during testing. No participants were excluded due to the inability to consistently measure shear wave velocity.

Arterial characteristics

The ICA and VA could be identified at both proximal and distal sites in neutral and contralateral rotation. Measures of arterial stiffness are shown in Table 1. Mean ICA stiffness in neutral was less than VA neutral stiffness. ICA stiffness was significantly reduced in contralateral rotation (p = 0.01) whereas the VA became significantly stiffer in rotation (p = 0.05). Stiffness changes were more pronounced in the distal region. For both arteries mean change was more marked in the distal (C1–2) segments (ICA decreased 12.1% distal, 7.2% proximal, VA increased 10.5% distal, 5.8% proximal), but this reached significance only in the ICA p = 0.03.

Table 1.

Mean (SD) arterial wall stiffness (m/s) in internal carotid (ICA) and vertebral artery (VA) in neutral and contralateral rotation positions and mean (95% CI, p value adjusted for age) difference, % difference, effect size of difference between positions and observed power.

	Positions		Difference between positions
Artery	Neutral	Rotation	Neutral minus rotation	% difference	Effect size	Power
ICA
Mean	2.4 (0.5)	2.1 (0.5)	0.2 (CI 0.01 to 0.4, 0.01)	8	0.11	0.75
Proximal	2.4 (0.5)	2.2 (0.5)	0.2 (CI 0.1 to 0.4, 0.14)	8	0.08	0.31
Distal	2.4 (0.5)	2.1 (0.5)	0.3 (CI 0.0 to 0.5, 0.04)	13	0.15	0.57
VA
Mean	3.7 (0.9)	4.0 (0.6)	−0. 3 (CI −0.6 to 0.0, 0.05)	8	0.07	0.51
Proximal	3.8 (0.9)	4.1 (0.9)	−0.2 (CI −0.7 to 0.2, 0.32)	6	0.04	0.17
Distal	3.5 (0.9)	3.9 (0.9)	−0.4 (CI −0.8 to 0.0, 0.07)	12	0.11	0.43

CI = confidence intervals; SD = standard deviation, negative difference values represent an increase in arterial stiffness.

Mean blood flow velocities for ICA and VA are shown in Table 2. There was no significant difference in these properties between neutral and rotated positions.

Table 2.

Mean (SD) blood flow velocity (cm/s) in internal carotid (ICA) and vertebral artery (VA) in neutral and contralateral rotation positions and mean (95%CI, p value) difference between positions.

		Positions		Difference between positions
Artery		Neutral	Rotation	Neutral minus rotation
ICA
Proximal	PS	74 (21)	79 (25)	−5 (−11 to 2, 0.13)
	ED	25 (10)	26 (8)	−1 (−4 to 3, 0.71)
Distal	PS	76 (16)	73 (16)	2 (−3 to 7, 0.36)
	ED	29 (7)	26 (12)	3 (−1 to 8, 0.16)
VA
Proximal	PS	55 (12)	52 (15)	3 (−2 to 8, 0.23)
	ED	16 (5)	15 (8)	1 (CI −2 to 4, 0.51)
Distal	PS	46 (12)	45 (11)	1 (−4 to 6, 0.61)
	ED	15 (4)	15 (5)	−1 (−2 to 1, 0.56)

PS = peak systolic, ED = end diastolic; CI = confidence intervals; SD = standard deviation; ANCOVA = repeated measures analysis of variance; Negative difference values represent an increase in blood flow velocity.

Intimal wall thickness did not significantly change between head positions for ICA and VA at either proximal or distal sites (Table 3).

Table 3.

Mean (SD) intimal wall thickness (mm) in internal carotid (ICA) and vertebral artery (VA) in neutral and contralateral rotation positions and mean (95%CI, p value) difference between positions.

	Positions		Difference between positions
Artery	Neutral	Rotation	Neutral minus rotation
ICA
Proximal	0.3 (0.1)	0.3 (0.1)	0.1 (0.0 to 0.1, 0.60)
Distal	0.3 (0.1)	0.3 (0.1)	0.0 (0.0 to 0.0, 0.85)
VA
Proximal	0.2 (0.1)	0.2 (0.1)	0.0 (0.0 to 0.0, 0.65)
Distal	0.2 (0.1)	0.2 (0.1)	0.0 (−0.1 to 0.0, 0.12)

CI = confidence intervals; SD = standard deviation; Negative difference values represent an increase in intimal wall thickness.

Discussion

The study demonstrated that is possible to detect changes in ICA and VA stiffness on contralateral head rotation in healthy individuals, using a novel measurement approach for these arteries, shear wave elastography. Head rotation is the movement most commonly associated with adverse neurovascular events with manipulative therapy and is used in screening protocols to determine those at risk [13]. The study found that changes in stiffness (shear wave velocity) could be measured between neutral and contralateral head rotation, with stiffness increasing in the VA and decreasing in the ICA. Changes were more apparent distally (C1–2) in both arteries consistent with their greater vulnerability in this region with head rotation [19]. The findings of this study suggest that shear wave elastography is a modality worth pursuing in future research, particularly in the vertebral artery, as it has the potential to provide more specific information about the integrity of the arterial wall and the effects of biomechanical stress, than currently available measures of arterial compliance. Validation in larger groups and clinical populations will be needed but may ultimately help inform risk assessment in manipulative therapy.

Arterial characteristics

Shear wave velocity measures in this study demonstrated that the VA became stiffer in contralateral rotation suggesting increased stress on the artery imposed by this position. This is consistent with the findings of cadaveric studies using ultrasound measurement [8,20], which found axial rotation to place the greatest strain on the VA. In this study changes in stiffness were more apparent in the distal C1–2 segment (10.5% compared with 5.8% proximally) though differences did not reach significance. The magnitude of change will need further confirmation in studies with larger numbers of participants. The arteries will also need to be measured bilaterally to account for any effects of vessel dominance. The greater stiffness in the VA in the C1–2 region might reflect it being more tortuous, or more angulated and becoming more stretched with contralateral rotation [21–23]. It does point to the increased vulnerability of the VA at the C1–2 segment and supports the need for caution with the use of end-range rotation positions during therapeutic procedures directed to this region.

The ICA wall, in contrast, had lower stiffness in both the neutral and rotated positions and became less stiff in contralateral rotation. Reducing stiffness may be due to its more anterior anatomical location, which brings the artery closer to the mid-line in contralateral rotation. This suggests that the ICA is less vulnerable in contralateral rotation. Future research should investigate the changes in stiffness in other head positions such as ipsilateral rotation and extension.

While changes occurred in arterial stiffness, blood flow velocity showed no significant changes in rotation in either ICA or VA. There is little agreement between studies on changes in blood flow in different head positions [24,25], reinforcing the limited value of this type of investigation to determine the risk of injury to the artery. VA blood flow studies at the C1–2 level are particularly challenging due to the angulation of the artery at this level, limiting the validity of their findings [26]. Similarly, the thickness of the arterial wall showed no change in rotation in this study. However, since the arterial walls are extremely thin (0.1–0.3 mm) particularly the VA, small changes may be difficult to detect. The measure is usually used to detect plaque formation so more subtle changes may be difficult to detect.

The poor sensitivity of conventional arterial characteristics such as flow velocity to examine arterial integrity reinforces the need for a new measure. Measuring properties of the arterial wall may provide a more clinically useful measure to estimate structural integrity. Further investigation of elastography imaging of the cervical arteries, particularly the more clinically relevant VA is justified to determine if it could be a measure in a screening tool to evaluate the risk of dissection.

Strengths and limitations

The study demonstrated that shear wave elastography can measure the effect of cervical rotation on the proximal and distal ICA and VA wall stiffness, as well as the whole artery and is a novel method investigating the effect of head rotation in vivo. Importantly, we were able to track the arteries up into the atlanto-axial region and obtain high-quality measures in the contralateral rotation positions, which is of relevance to cervical manipulative therapy. In addition, including the whole artery data provides a general measure of structural integrity. This might be relevant particularly when comparing between healthy and clinical groups and may be ultimately what determines arterial vulnerability.

Ultrasound is a relatively cheap, noninvasive imaging modality, which is safe and well tolerated by patients. While in the short term, shear wave ultrasound elastography may not be feasible for use by physical therapists, referral to an ultrasonographer is eminently feasible. With the development of smaller more portable units, it is feasible that shear wave ultrasound elastography could, with appropriate training, become a diagnostic modality in the toolkit of physical therapists.

It should be noted that the ROI for the shear wave measurement was somewhat larger than the intimal layer of the artery and extended slightly into the vessel lumen, which could have impacted measurement accuracy. However, the VTIQ system quality mode cancels out any signal variation >5–10%, which ensures that all measures obtained could be expected to reflect true values. To address some of the reported limitations of shear wave, we used the average of 3 velocity measures along the wall at each site to minimize the effects of anisotropy. The system was not cardiac gated so variations due to cardiac cycle were possible. However, the ICA and VA are low resistance vessels and Matuski (2104) showed that even the larger common carotid only varied by 8% during the cardiac cycle. Any effect of the cardiac cycle might be expected to be much less in the progressively smaller ICA and VA, and within measurement error.

Future research

Further research is now needed to develop the measure of shear wave elastography to determine its clinical value. Investigating vessels bilaterally for side to side differences will be important as the VA often shows left side dominance. Larger groups need to be examined to define normal cut off levels in healthy individuals of varying age. The effect of other head positions such as ipsilateral; rotation, extension and lateral flexion likewise needs to be determined. Once more is known of the measure, clinical groups, for example, those with migraine who are thought to have increased general cervical arterial stiffness and increased risk of CAD, should be investigated to establish the clinical utility of the measure. In this case, whether the whole artery measure provides information which is distinct from that obtained from individual proximal and distal measures will need to be determined. Ultimately, elastography could be trialed in groups of patients with CAD to determine whether differences in the behavior of arterial stiffness in comparison to age-matched healthy controls can be detected.

Conclusions

Shear wave elastography can detect changes in VA and ICA wall stiffness with contralateral head rotation. Stiffness increases in the VA suggesting increased stress on the artery in this position. The ICA in contrast became less stiff. Similar changes were not reflected in either blood flow velocity or wall thickness. Further investigation is warranted to determine the value of elastography to provide more clinically relevant information about arterial integrity, which could ultimately inform risk assessment in the cervical spine.

Biographies

Dr Lucy Caroline Thomas is an academic at the University of Queensland and a Titled Musculoskeletal Physical Therapist. She holds an honorary appointment at the Royal Brisbane and Women’s Hospital. She teaches musculoskeletal physical therapy to entry level and postgraduate students, for management of cervical spine disorders. Her research focus is on safe practice in the cervical spine, development of a screening tool for early recognition of cervical arterial dissection and she was the lead author on the Australian Physiotherapy Association 'Clinical guide to safe manual therapy practice in the cervical spine'. She presents regularly at national and international conferences and workshops.

Kalos Chan is a Physical Therapist in private practice in a Brisbane, Australia. This study was completed as part of his honours project at the University of Queensland.

Gail Durbridge is a Registered Radiographer and ultrasonographer and is Head of Teaching and Learning at the Centre for Advanced Imaging at the University of Queensland.

Funding Statement

The study was funded by an internal grant from the School of Health and Rehabilitation Sciences at the University of Queensland.

Disclosure statement

No potential conflict of interest was reported by the authors.

Ethical approval

No. 2015000326.