Growing evidence suggests that increased intracranial pulsatility is associated with cerebral small vessel disease (SVD). We systematically reviewed papers that assessed intracranial pulsatility in subjects with SVD. We included 27 cross-sectional studies (n=3356): 20 used Doppler ultrasound and 7 used phase-contrast MRI. Most studies measured pulsatility in the internal carotid or middle cerebral arteries (ICA, MCA), whereas few assessed veins or cerebrospinal fluid (CSF). Methods to reduce bias and risk factor adjustment were poorly reported. Substantial variation between studies in assessment of SVD and of pulsatility indices precluded a formal meta-analysis. Eight studies compared pulsatility by SVD severity (n=26–159, median = 74.5): arterial pulsatility index was generally higher in more severe SVD (e.g. MCA: standardized mean difference = 3.24, 95% confidence interval [2.40, 4.07]), although most did not match for age. Seventeen studies (n=9–700; median = 110) performed regression or correlation analysis, of which most showed that increased pulsatility was associated with SVD after adjustment for age. In conclusion, most studies support a cross-sectional association between higher pulsatility in large intracranial arteries and SVD. Future studies should minimize bias, adjust for potential confounders, include pulsatility in veins and CSF, and examine longitudinal relationship between pulsatility and SVD. Agreement on reliable measures of intracranial pulsatility would be helpful.

Introduction

Cerebral small vessel disease (SVD) is responsible for up to 45% of dementia and approximately 20% of all stroke worldwide [1,2]. SVD is mainly diagnosed on brain imaging, based on various features such as white matter hyperintensities (WMH), perivascular space (PVS), microbleeds, lacunes, and recent small subcortical (lacunar) infarcts. Although each individual imaging feature might represent different underlying tissue changes, evidence from autopsy and clinical studies indicate that these features are all related to pathologies in brain small vessels [2]. For example, though white matter changes could be seen in other conditions such as multiple sclerosis, WMH of presumed vascular origin are the most typical and known SVD changes. They are usually symmetrically and bilaterally distributed in subcortical areas, and various theories have been proposed to explain the small vessel pathology and how this affects the brain [2].

Ageing and hypertension are important risk factors for SVD [3], both of which are associated with loss of elasticity in the arterial walls. Growing evidence has shown a relationship between increased aortic pulsatility and lacunar infarcts or WMH volume [4,5]. It is hypothesized that the stiffened vessel walls would be less able to dampen the systemic arterial pressure pulse via the Windkessel effect, leading to high pulsatility being transmitted into the brain and causing or exacerbating small vessel damage, such as endothelial dysfunction and blood–brain barrier (BBB) impairment [6].

Cerebrovascular pulsatility is also an important driving force of the cerebrospinal fluid (CSF)–interstitial fluid (ISF) exchange via the glymphatic system in the perivascular spaces, which is essential for clearing metabolic and other waste products from the brain [7,8]. Altered pulsatility and abnormality in the PVS might cause or accelerate glymphatic dysfunction and aggregation of waste products such as Aβ, other proteins or cell debris, which is related to age-related neurodegenerative diseases and SVD [8]. Thus, for several reasons, increased intracranial pulsatility might be an important underlying mechanism of SVD. However, so far very few clinical studies have assessed pulsatility directly in the intracranial vessels [9,10].

Doppler ultrasound and magnetic resonance imaging (MRI) are two main techniques that could measure intracranial pulsatility. Doppler ultrasound measures real-time blood flow velocity in the large arteries, mainly internal carotid arteries (ICA) and middle cerebral arteries (MCA). Pulsatility is generally calculated using Gosling’s equation (pulsatility index, PI = (peak systolic velocity − peak diastolic velocity)/mean velocity)). Phase-contrast MRI can quantify fluid velocity and flow in the intracranial arteries, veins, and CSF. It has mostly been used in disorders such as hydrocephalus that present with abnormal CSF dynamics [11] with few studies in SVD. Some MRI studies used a similar equation (flow was used instead of velocity) to calculate the pulsatility in the vessels, but indices reported for CSF pulsatility vary [9,10]. It is unclear which measures are most relevant and reliable for quantifying intracranial pulsatility in patients with SVD.

Furthermore it is unclear if high pulsatility and SVD were both simply the result of exposure to vascular risk factors, since some studies did not adjust for age or important risk factors when assessing the relationship between intracranial pulsatility and SVD [9]. Also, WMH are commonly used by studies to represent SVD burden [1214], and it is unknown whether high pulsatility was also related to other features of SVD, such as lacunar stroke or PVS.

In order to provide a complete summary of all knowledge to date on intracranial pulsatility in SVD patients, we systematically reviewed papers that measured cerebral pulsatility in SVD with a view to performing a meta-analysis. We aimed to determine if there was an association between intracranial pulsatility and SVD, the magnitude of any association, whether pulsatility predicted SVD progression longitudinally, and if any individual SVD features were more strongly associated with pulsatility than others.

Methods

We performed this review according to the MOOSE guidelines for meta-analysis of observational studies [15]. We conducted a literature search of Ovid MEDLINE and Embase from 1946 up to April 2017 using the Ovid Web Gateway. We searched terms related to pulsatility and SVD in all contents of the papers using the strategy: “Pulsatility” or “Resistance” or “Velocity” or “cerebrospinal fluid pulsatility” or “Phase-contrast MRI” and “Cerebral small vessel disease” or “White matter hyperintensities” or “Leukoaraiosis”. English and non-English literatures were sought. Additional records were identified by hand from relevant reviews, primary papers, and from the authors’ publication lists. We defined SVD features according to the STRIVE Guideline [16].

Eligibility criteria

We included papers that reported primary results of studies that fulfilled all the following criteria: (1) recruited participants with SVD features; (2) assessed resting-state intracranial pulsatility (including ICA) using any imaging technique; (3) assessed the relationship between intracranial pulsatility and SVD. We only included studies that focused on sporadic SVD in this review as the mechanisms for hereditary SVDs might differ. We excluded review papers, abstracts, and papers that used pharmacological, CO2 or other stimulus without providing pre-stimulus data.

Data extraction

We extracted data on participant characteristics, study design, MRI or Doppler technicalities, location, and type of pulsatility indices assessed. For studies that compared pulsatility between groups and reported means and standard deviations, we extracted the results of the comparisons from text or tables where available, or from graphs where necessary; for those that performed association analysis such as correlation or regression models, we extracted the statistical methods, coefficients, P values, and confounding factors or co-variates (if any) that were adjusted for. For studies that performed both non-adjusted and adjusted analysis, we only included the adjusted results. We assessed the study quality using a checklist that includes factors such as study population and methods for bias controlling (Supplementary Table S1).

Statistical analysis

We displayed the results of comparisons between groups using forest plots in the Cochrane Collaboration’s Review Manager (Revman Version 5.3) and used standardised mean difference (SMD) to represent the difference between groups. We did not include data from studies that did not provide standard deviations (s.d.), because the SMD is inestimable without s.d.. In this review, we were not able to perform robust meta-analyses because of the limited data and the large heterogeneity in study population and pulsatility measures. Therefore, in the forest plot we did not calculate overall effect sizes. For studies that divided patients into more than two grades of WMH severity, we combined the means and s.d.s of PI from patients who had similar severity of SVD to create a single pair-wise comparison using formulae from the Cochrane handbook (Supplementary Equation S1) [17]. For example, we created new groups “mild WMHs” by combining groups of Fazekas score 0 to 1, and “severe WMHs” from Fazekas score 2 to 3.

Results

The search strategy identified 518 papers, of which 48 were potentially eligible and 27 were ultimately selected for further review (Figure 1). We excluded reviews (3), drug trials without providing baseline information on pulsatility and SVD (2), hereditary SVD study or non-relevant analysis (8), studies that used stimulus only (2) or did not assess intracranial pulsatility (6). One MRI study compared CSF pulsatility between patients with late-life depression and age-matched healthy volunteers, but we included the results in the forest plot because patients had higher WMH burden than the healthy volunteers and late-onset depression is associated with SVD [18].

PRISMA flow diagram of literature search and its results

Characteristics of included studies

Twenty-seven studies (n=3356) were included, of which 20 used ultrasound [12,14,1936] and 7 used MRI to measure pulsatility [9,10,13,18,3739], 26/27 were cross-sectional and one was a clinical trial of cilostazol [35] from which we only extracted the data obtained at baseline.

Doppler ultrasound studies

Table 1 summarises patient characteristics from 20 ultrasound studies (n=2935) with sample sizes ranging from 9 to 700 (median = 107). Fourteen used WMH (if MRI used for structural brain imaging) or white matter hypoattenuation (if CT used) to represent SVD burden. Two studies included patients with stroke: one compared pulsatility between lacunar and non-lacunar ischaemic stroke [25], and one compared pulsatility between patients with lacunar stroke (according to Trial of ORG 1072 in Acute Stroke Treatment (TOAST) classification) and age- and sex-matched healthy controls [32]. One study included patients with multi-infarct dementia [22] and one recruited patients with cerebral amyloid angiopathy (CAA) [31] (Table 1).

Table 1
Patient characteristics of ultrasound studies
First author Year Subjects/Disease Sample size Group Age ± s.d.(years) Vessels of interest 
Jordi Sanahuja [302016 Type 2 diabetes 202 Higher SVD score (n=21)  MCA 
    Lower SVD score (n=181)   
Oscar H. Del Brutto [332015 Community 70 Mild WMH (n=42) 72.2 ± 5.5 MCA, vertebral artery 
     77.4 ± 7.34  
Abbas Ghorbani [342015 Patients who had SVD on imaging 104 Patients who had WMH or lacunar infarct on MRI 68.4 MCA 
    Patients who had no lesion on MRI 62.6  
Monika Turk [212014 Leukoraiosis 96 Leukoaraiosis (n=52) 54.9 ± 8.3 ICA 
    Age and sex-matched healthy volunteers (n=44) 52.39 ± 7.34  
Benjamin S. Aribisala [122014 Lothian birth cohort 1936 700 – 70 ICA 
Redouane Ternifi [202014 Healthy volunteers – 70.56 ± 5.46 Brain tissue displacement 
Joao Sargento-Freitas [292014 Patients who visited neurosonology lab 439 – 63.47 ± 14.94 MCA 
Sushmita Purkayastha [282014 Patients with vascular risk factors 48 – 75±7 MCA 
Sang Won Han [352014 Lacunar stroke 130 – 64.7 MCA 
Alastain J.S. Webb [272012 TIA or minor stroke 110 Fazekas 3 (n=25) 74.9 ± 7.9 MCA 
    Fazekas 2 (n=24) 68.5 ± 11  
    Fazekas 1 (n=21) 66.5 ± 12  
    Fazekas 0 (n=30) 53 ± 15  
Vincent Mok [262012 Community 205 With severe WMH 74 ± 6 MCA 
    without severe WMH 69 ± 6  
Ioannis Heliopoulos [142012 Hypertensive patients 52 – 71.4 ± 4.5 MCA 
Kerstin Bettermann [362012 Patients with WMH 26 Patients with WMH 63.5 ± 11.25 MCA 
    Control without WMH 55.07 ± 7.91  
Iria Rodriguez [252010 Ischaemic stroke 186 Lacunar (n=35) 69.7 ± 10.8 MCA 
    Non-lacunar (n=151) 71.6 ± 8.1  
Tomotaka Tanaka [192009 Diabetic patients 122 Hypertensive (n=43) 66.9 ± 9.8 ICA 
    Non-hypertensive (n=79) 62.0 ± 11.0  
Eric E. Smith [312008 CAA 20 CAA (n=11) 73.5 ± 7.4 Basilar artery 
    Healthy volunteers (n=9) 70.9 ± 7.9  
Chelsea S. Kidwell [232001 Retrospective review in patients who had both TCD and MRI 55 – 62 (range 28–98) MCA 
Rosa M. Sánchez-Pérez [242003 Patients (>60 years) who visited neurological department for minor symptoms 116 – 74.44 ± 6.35 MCA 
Masahiko Hiroki [322002 Stroke 167 Small vessel disease (TOAST) (n=103) 70.9 ± 9.0 Central retinal artery 
    Age and sex-matched controls (n=64) 69.7 ± 8.8  
Stefan Biedert [221995 Dementia 78 Multi-infarct dementia (n=19) Age range 60–69 for all MCA, basilar artery 
    AD (n=23)   
    Age-matched healthy volunteers (n=36)   
First author Year Subjects/Disease Sample size Group Age ± s.d.(years) Vessels of interest 
Jordi Sanahuja [302016 Type 2 diabetes 202 Higher SVD score (n=21)  MCA 
    Lower SVD score (n=181)   
Oscar H. Del Brutto [332015 Community 70 Mild WMH (n=42) 72.2 ± 5.5 MCA, vertebral artery 
     77.4 ± 7.34  
Abbas Ghorbani [342015 Patients who had SVD on imaging 104 Patients who had WMH or lacunar infarct on MRI 68.4 MCA 
    Patients who had no lesion on MRI 62.6  
Monika Turk [212014 Leukoraiosis 96 Leukoaraiosis (n=52) 54.9 ± 8.3 ICA 
    Age and sex-matched healthy volunteers (n=44) 52.39 ± 7.34  
Benjamin S. Aribisala [122014 Lothian birth cohort 1936 700 – 70 ICA 
Redouane Ternifi [202014 Healthy volunteers – 70.56 ± 5.46 Brain tissue displacement 
Joao Sargento-Freitas [292014 Patients who visited neurosonology lab 439 – 63.47 ± 14.94 MCA 
Sushmita Purkayastha [282014 Patients with vascular risk factors 48 – 75±7 MCA 
Sang Won Han [352014 Lacunar stroke 130 – 64.7 MCA 
Alastain J.S. Webb [272012 TIA or minor stroke 110 Fazekas 3 (n=25) 74.9 ± 7.9 MCA 
    Fazekas 2 (n=24) 68.5 ± 11  
    Fazekas 1 (n=21) 66.5 ± 12  
    Fazekas 0 (n=30) 53 ± 15  
Vincent Mok [262012 Community 205 With severe WMH 74 ± 6 MCA 
    without severe WMH 69 ± 6  
Ioannis Heliopoulos [142012 Hypertensive patients 52 – 71.4 ± 4.5 MCA 
Kerstin Bettermann [362012 Patients with WMH 26 Patients with WMH 63.5 ± 11.25 MCA 
    Control without WMH 55.07 ± 7.91  
Iria Rodriguez [252010 Ischaemic stroke 186 Lacunar (n=35) 69.7 ± 10.8 MCA 
    Non-lacunar (n=151) 71.6 ± 8.1  
Tomotaka Tanaka [192009 Diabetic patients 122 Hypertensive (n=43) 66.9 ± 9.8 ICA 
    Non-hypertensive (n=79) 62.0 ± 11.0  
Eric E. Smith [312008 CAA 20 CAA (n=11) 73.5 ± 7.4 Basilar artery 
    Healthy volunteers (n=9) 70.9 ± 7.9  
Chelsea S. Kidwell [232001 Retrospective review in patients who had both TCD and MRI 55 – 62 (range 28–98) MCA 
Rosa M. Sánchez-Pérez [242003 Patients (>60 years) who visited neurological department for minor symptoms 116 – 74.44 ± 6.35 MCA 
Masahiko Hiroki [322002 Stroke 167 Small vessel disease (TOAST) (n=103) 70.9 ± 9.0 Central retinal artery 
    Age and sex-matched controls (n=64) 69.7 ± 8.8  
Stefan Biedert [221995 Dementia 78 Multi-infarct dementia (n=19) Age range 60–69 for all MCA, basilar artery 
    AD (n=23)   
    Age-matched healthy volunteers (n=36)   

Abbreviations: AD, Alzheimer's disease; CAA, cerebral amyloid angiopathy; ICA, internal carotid artery; MCA, middle cerebral artery; MRI, magnetic resonance imaging; SVD, small vessel disease; TCD, transcranial Doppler; TIA, transient ischaemic attack; WMH: white matter hyperintensities.

Three studies assessed pulsatility in the ICA (cervical), seventeen in the MCA, one in the basilar artery, one in the posterior cerebral artery (PCA), and one in the central retinal artery. Most studies calculated PI using Gosling’s equation. One study measured brain tissue pulsatile movement [20].

MRI studies

Patient characteristics and scanner information of seven MRI studies (n=421, range 35–101, median = 51) are shown in Table 2. Three studies recruited patients with dementia or cognitive impairment, including idiopathic dementia [9], Alzheimer’s disease [13], vascular dementia [13], and mild cognitive impairment [37]. Only one reported the diagnosis criteria for Alzheimer’s disease and vascular dementia [13]. One study recruited patients with late onset major depressive disorder [18]. Three studies recruited healthy volunteers (age range: 43–82, 18–75, and 62–82 years) [10,38,39] (Table 2).

Table 2
Patient characteristics and scanner information of MRI studies
First author Year Subjects/ Disease Sample size Groups Age (years) Scanner and sequences TE/TR (ms) Venc Time points in a cardiac cycle 
Clive B. Beggs [102016 Healthy volunteers 101 – 44.7 ± 17.8 (range 18–75) GE 3 Tesla; Retrospectively peripheral pulse-gated 2-D phase contrast cine sequences 7.9/40 20 cm/s aqueduct 32 
Anders Wåhlin [392014 Healthy volunteers 37 – 71 ± 6 Philips 3 Tesla; Retrospectively peripheral pulse or ECG-gated 2-D phase contrast cine sequences 6–11/10–16 70 cm/s for arteries and 7 cm/s for CSF 32 
Todd A.D. Jolly [382013 Healthy volunteers 35 – 65.67 ± 9.31 (range 43–82) Siemens 3 Tesla; Retrospectively ECG-gated 2-D phase contrast cine sequences 6.9/26.5 Arteries 75 cm/s, veins 40 cm/s, CSF 22 cm/s Not mentioned 
Marie C. Henry-Feugeas [372009 Mild cognitive impairment 71 Significant WHM (n=42) 74 ± 5 GE 1.5 Tesla; Retrospectively ECG-gated 2-D phase contrast cine sequences 7.6–9.9/20–23 Not mentioned 16 
    No significant WHM (n=29) 69 ± 5     
Grant A. Bateman [132008 Senile dementia 48 AD (n=12) 76 ± 4 Siemens 1.5 Tesla; Retrospectively ECG-gated 2-D phase contrast cine sequences 7/29 Arteries 75 cm/s, veins 40 cm/s, CSF 10 cm/s Not mentioned 
    Vascular dementia (n=12) 70 ± 11     
    Normal aging (n=12) 70 ± 5     
    Normal young (n=12) 25 ± 12     
Josephine H. Naish [182006 Late onset major depressive disorder 51 Responders to anti-depressant monotherapy (n=21) 71.0 ± 6.54 Philips 1.5 Tesla; Retrospectively peripheral pulse-gated 2-D phase contrast cine sequences 7–8.2/14–15 ms Not mentioned 16 
    Non-responders (n=8) 75.2 ± 6.18     
    Age-matched control (n=22) 72.9 ± 5.38     
    Young (n=19) 27.5 ± 4.4     
Grant A. Bateman [92002 Idiopathic dementia 78 Fazekas 0 (n=10) 71 ± 8 1.5 Tesla; Retrospectively ECG-gated 2-D phase contrast cine sequences 7/29 ms Not mentioned Not mentioned 
    Fazekas 1 (n=18) 75 ± 8     
    Fazekas 2 (n=24) 76 ± 7     
    Fazekas 3 (n=8) 77 ± 8     
    Healthy controls (n=18) 42 ± 17     
First author Year Subjects/ Disease Sample size Groups Age (years) Scanner and sequences TE/TR (ms) Venc Time points in a cardiac cycle 
Clive B. Beggs [102016 Healthy volunteers 101 – 44.7 ± 17.8 (range 18–75) GE 3 Tesla; Retrospectively peripheral pulse-gated 2-D phase contrast cine sequences 7.9/40 20 cm/s aqueduct 32 
Anders Wåhlin [392014 Healthy volunteers 37 – 71 ± 6 Philips 3 Tesla; Retrospectively peripheral pulse or ECG-gated 2-D phase contrast cine sequences 6–11/10–16 70 cm/s for arteries and 7 cm/s for CSF 32 
Todd A.D. Jolly [382013 Healthy volunteers 35 – 65.67 ± 9.31 (range 43–82) Siemens 3 Tesla; Retrospectively ECG-gated 2-D phase contrast cine sequences 6.9/26.5 Arteries 75 cm/s, veins 40 cm/s, CSF 22 cm/s Not mentioned 
Marie C. Henry-Feugeas [372009 Mild cognitive impairment 71 Significant WHM (n=42) 74 ± 5 GE 1.5 Tesla; Retrospectively ECG-gated 2-D phase contrast cine sequences 7.6–9.9/20–23 Not mentioned 16 
    No significant WHM (n=29) 69 ± 5     
Grant A. Bateman [132008 Senile dementia 48 AD (n=12) 76 ± 4 Siemens 1.5 Tesla; Retrospectively ECG-gated 2-D phase contrast cine sequences 7/29 Arteries 75 cm/s, veins 40 cm/s, CSF 10 cm/s Not mentioned 
    Vascular dementia (n=12) 70 ± 11     
    Normal aging (n=12) 70 ± 5     
    Normal young (n=12) 25 ± 12     
Josephine H. Naish [182006 Late onset major depressive disorder 51 Responders to anti-depressant monotherapy (n=21) 71.0 ± 6.54 Philips 1.5 Tesla; Retrospectively peripheral pulse-gated 2-D phase contrast cine sequences 7–8.2/14–15 ms Not mentioned 16 
    Non-responders (n=8) 75.2 ± 6.18     
    Age-matched control (n=22) 72.9 ± 5.38     
    Young (n=19) 27.5 ± 4.4     
Grant A. Bateman [92002 Idiopathic dementia 78 Fazekas 0 (n=10) 71 ± 8 1.5 Tesla; Retrospectively ECG-gated 2-D phase contrast cine sequences 7/29 ms Not mentioned Not mentioned 
    Fazekas 1 (n=18) 75 ± 8     
    Fazekas 2 (n=24) 76 ± 7     
    Fazekas 3 (n=8) 77 ± 8     
    Healthy controls (n=18) 42 ± 17     

Abbreviations: AD, Alzheimer's disease; CSF, cerebrospinal fluid; ECG, electrocardiogram; TE, Echo time; TR, repetition time; Venc, velocity encoding parameter; WMH, white matter hyperintensities.

All seven studies used phase-contrast MRI. Four studies used 1.5 Tesla and three used 3 Tesla scanners. All studies used retrospective gated phase-contrast MRI: four used peripheral pulse- and two used electrocardiogram (ECG)-gating; four studies reported the numbers of time points measured per cardiac cycle, among which three used 32 whereas one used 16 time points (Table 2).

There were large differences between studies as to where flow was measured. Figure 2 shows the regions of interests used in these studies. Six included the ICA(s) [9,10,13,3739]; in posterior circulation five chose the basilar artery [9,10,13,37,38] and one selected the vertebral arteries [39]; four studied pulsatility in the venous sinuses [9,13,37,38]; six measured CSF flow, among which one selected the cervical subarachnoid spaces [39], one selected tentorial incisura [13], and four measured the cerebral aqueduct [10,18,37,38]. In total, three studies measured flow or pulsatility across the cerebral arteries, veins, and the CSF system concurrently [13,18,37] (Figure 2).

Regions of interest for flow measurement used in phase-contrast MRI studies included in this review.
Figure 2
Regions of interest for flow measurement used in phase-contrast MRI studies included in this review.

*Anders Wåhlin et al. [39] selected the vertebral arteries instead of the basilar artery.

Figure 2
Regions of interest for flow measurement used in phase-contrast MRI studies included in this review.

*Anders Wåhlin et al. [39] selected the vertebral arteries instead of the basilar artery.

Quality assessment

The average quality score of 27 studies was 6.85/10. 26/27 studies were prospective. However, seven studies did not report participants’ demographic information or general health condition. Only about half of the studies adjusted or matched for risk factors (15/27) or reported the expertise of observers of structural MRI (15/27). Less than half of the studies reported dropout (12/27) or use of blinding in structural image analysis (11/27) (Supplementary Figure S1).

Result of comparisons of pulsatility measures

Doppler ultrasound studies

Three studies (n=356, median = 96) compared arterial PI between patients with different WMH severities. One study examined the ICA, one chose MCA, and one chose both MCA and basilar artery. PI was generally higher in patients with more severe WMH in the ICA [21], MCA [22,26], and the basilar artery [22] (e.g. in MCA: SMD = 3.24, 95% confidence interval (CI) [2.4, 4.07]), although the result for ICA did not reach statistical significance (SMD = 0.38, 95% CI [−0.02, 0.79] [21]). 2/3 of the studies were age-matched [21,22], one of which also matched for gender [21] (Figure 3A).

Forest plots of studies that compared pulsatility (using Doppler ultrasound or MRI) between SVD and control groups; *indicates studies that matched for age
Figure 3
Forest plots of studies that compared pulsatility (using Doppler ultrasound or MRI) between SVD and control groups; *indicates studies that matched for age

(AC): Pulsatility was measured by Doppler ultrasound. (A) Comparison of vascular pulsatility between patients with severe and mild white matter hyperintensities (WMH). (B) Comparison of central retinal artery pulsatility between patients with SVD-stroke (TOAST classification) and healthy volunteers. (C) Comparison of posterior cerebral artery pulsatility between patients with cerebral amyloid angiopathy (CAA) and healthy volunteers. (D and E): Pulsatility was measured by phase-contrast MRI. (D) Comparison of vascular or CSF pulsatility between patients with severe and mild WMH. (E) Comparison of flow waveform peak delays between patients with severe and mild WMH. It is worth noting that, in forest plot (E), shorter peak delay is suggested by the authors to represent higher intracranial stiffness.

Figure 3
Forest plots of studies that compared pulsatility (using Doppler ultrasound or MRI) between SVD and control groups; *indicates studies that matched for age

(AC): Pulsatility was measured by Doppler ultrasound. (A) Comparison of vascular pulsatility between patients with severe and mild white matter hyperintensities (WMH). (B) Comparison of central retinal artery pulsatility between patients with SVD-stroke (TOAST classification) and healthy volunteers. (C) Comparison of posterior cerebral artery pulsatility between patients with cerebral amyloid angiopathy (CAA) and healthy volunteers. (D and E): Pulsatility was measured by phase-contrast MRI. (D) Comparison of vascular or CSF pulsatility between patients with severe and mild WMH. (E) Comparison of flow waveform peak delays between patients with severe and mild WMH. It is worth noting that, in forest plot (E), shorter peak delay is suggested by the authors to represent higher intracranial stiffness.

Two studies looked at other SVD features. One study (n=167) showed that patients with lacunar stroke (TOAST) had higher PI in the central retinal artery compared with age- and sex-matched heathy controls (SMD = 0.35, 95% CI [0.03, 0.66]) [32] (Figure 3B). One study of CAA (n=20) found that patients with CAA had a significantly higher PI in PCA than non-age-matched healthy elderly controls (SMD = 1.07, 95% CI [0.12, 2.03]) [31] (Figure 3C).

MRI studies

Three phase-contrast MRI studies (n=124, median = 50) performed comparisons of cerebrovascular or CSF pulsatility between patients with different WMH severities. None of these studies corrected for age. The indices for pulsatility included PI, stroke volume, and delay between waveform peaks. The trend in all the comparisons is that higher arterial or venous PI (e.g. arterial PI: SMD = 0.93, 95% CI [0.40, 1.47] [9]), larger arterial or venous or CSF stroke volume (e.g. CSF stroke volume: SMD = 1.58, 95% CI [0.64, 2.52] [13]) was associated with more WMH, although some results did not reach a statistical significance (e.g. venous PI: SMD = 0.18, 95% CI [−0.33, 0.69] [9]) (Figure 3D).

Two studies (n=110) calculated the delay between waveform systolic peaks. There was no significant difference in arterial-venous (n=60, SMD = 0.95% CI [−0.51, 0.51]) [13] or arterial-aqueduct peak delays (n=51, SMD = 0.49, 95% CI [−0.07, 1.06]) [18] between different severities of WMH (Figure 3E).

Results of studies that performed regression or correlation analysis

Table 3 summarises studies that performed regression or correlation analysis, including 13 Doppler ultrasound (n=9–700, median = 116) and three MRI studies (n=35–100, median = 37). 14/16 studies adjusted for co-variates, of which 12 included age. No studies adjusted for blood pressure although five considered history of hypertension. Eight studies used visual rating scores to assess SVD burden, whereas the other eight measured WMH or brain volume, among which one used manual ROIs [38], one used semi-automated [10] and six used automated masks. Only 2/6 of the papers using automated masks reported that the masks were manually checked [35,39].

Table 3
Results of correlation analyses in MRI and ultrasound studies
First author Year Sample size Correlation or regression models Variables (yxCoefficient P value Adjustment for confounding factors 
Ultrasound studies 
ICA 
Benjamin S. Aribisala [122014 700 Multiple variable regression WMH volume ∼ ICA PI β = 0.09 0.016 Age, sex, ICV, HBp history 
Tomotaka Tanaka [192009 122 Multivariate regression analysis WMH volume ∼ ICA PI not shown >0.05 Age 
MCA 
Oscar H. Del Brutto [332015 70 Generalised linear model WMH severity ∼ MCA PI β = 0.065, 95% CI[−0.084, 0.177 ] 0.474 Age, sex, the cardiovascular health status 
    WMH severity ∼ vertebral artery β = 0.066, 95% CI [−0.024, 0.156] 0.146  
Joao Sargento-Freitas [292014 439 Multiple ordinal regression DWMH score ∼ MCA PI OR 17.994, 95% CI [6.875, 47.11] <0.001 Age, sex, HBp history, DM, smoking, dyslipidaemia, coronariopathy, heart failure, obesity, peripheral artery disease, alcoholism, IMT 
    PVH score ∼ MCA PI OR 5.739, 95% CI [2.288, 14.397] <0.001  
    Basal ganglia score ∼ MCA PI OR 11.844, 95% CI [4.486, 31.268] <0.001  
Sang Won Han [352014 130 Pearson correlation WMH volume ∼ MCA PI R=0.195 0.026 No 
Sushmita Purkayastha [282014 48 Multivariable linear regression WMH/ICV ∼ MCA PI OR 1.25, 95% CI [0.14, 2.09] <0.01 Age, sex, race, DM, HBp 
Alastain J. S. Webb [272012 110 Ordinal regression WMH score ∼ MCA PI β = 4.33 P=0.037 Age, sex, physiology 
Vincent Mok [262012 159* Multiple logistic regression Severe WMH (vs without severe WMH) ∼ MCA PI OR = 1.33, 95% CI [1.04,1.70] per 0.1 increase in PI <0.01 Age 
Ioannis Heliopoulos [142012 52 Multivariable regression WMH score ∼ MCA PI β = 0.262 P=0.025 Age, sex, BMI, HBp, DM, hyperlipidaemia, smoking 
Iria Rodriguez [252010 186 Multivariate logistic analysis Lacunar infarct (vs non-lacunar) ∼ MCA PI OR = 8.13, 95% CI [1.17, 56.27] 0.034 Previous ranking score, WMH, retinal microangiopathy 
Rosa M. Sánchez-Pérez [242003 116 Multiple linear regression Leukoaraiosis severity score ∼ MCA PI β = −0.108 0.353 Age, sex, vascular risk factors, cognitive performance, blood flow velocity in MCA 
Chelsea S. Kidwell [232001 55 Multivariate regression WMH score ∼ MCA PI 0.71 P<0.05 Age, sex, HBp, coronary artery disease 
Brain tissue pulsatility 
Redouane Ternifi [202014 Non-parametric spearman correlation WMH volume ∼ Max BTD ρ = −0.86 <0.001 No 
    WMH volume ∼ Mean BTD ρ = −0.72 <0.001 No 
MRI studies 
Clive B. Beggs [102016 101 Multiple linear regression total WMH volume ∼ CSF peak negative velocity β = −124.903 P=0.041 Age 
Anders Wåhlin [392014 37 Ordinary linear regression Total brain volume ∼ arterial PI β = −0.42 P<0.01 Age, ICV, arterial net flow 
    Total brain volume ∼ CSF flow volume pulsatility β = −0.44 P<0.01 Age, ICV, arterial net flow 
Todd A.D Jolly [382013 35 Partial correlation WMH volume ∼ pulse wave amplitude in arteries or venous sinuses not shown P>0.05 Age 
    WMH volume ∼ pulse width in arteries or sinuses not shown P>0.05 Age 
    WMH volume ∼ PI in arteries or venous sinuses not shown P>0.05 Age 
First author Year Sample size Correlation or regression models Variables (yxCoefficient P value Adjustment for confounding factors 
Ultrasound studies 
ICA 
Benjamin S. Aribisala [122014 700 Multiple variable regression WMH volume ∼ ICA PI β = 0.09 0.016 Age, sex, ICV, HBp history 
Tomotaka Tanaka [192009 122 Multivariate regression analysis WMH volume ∼ ICA PI not shown >0.05 Age 
MCA 
Oscar H. Del Brutto [332015 70 Generalised linear model WMH severity ∼ MCA PI β = 0.065, 95% CI[−0.084, 0.177 ] 0.474 Age, sex, the cardiovascular health status 
    WMH severity ∼ vertebral artery β = 0.066, 95% CI [−0.024, 0.156] 0.146  
Joao Sargento-Freitas [292014 439 Multiple ordinal regression DWMH score ∼ MCA PI OR 17.994, 95% CI [6.875, 47.11] <0.001 Age, sex, HBp history, DM, smoking, dyslipidaemia, coronariopathy, heart failure, obesity, peripheral artery disease, alcoholism, IMT 
    PVH score ∼ MCA PI OR 5.739, 95% CI [2.288, 14.397] <0.001  
    Basal ganglia score ∼ MCA PI OR 11.844, 95% CI [4.486, 31.268] <0.001  
Sang Won Han [352014 130 Pearson correlation WMH volume ∼ MCA PI R=0.195 0.026 No 
Sushmita Purkayastha [282014 48 Multivariable linear regression WMH/ICV ∼ MCA PI OR 1.25, 95% CI [0.14, 2.09] <0.01 Age, sex, race, DM, HBp 
Alastain J. S. Webb [272012 110 Ordinal regression WMH score ∼ MCA PI β = 4.33 P=0.037 Age, sex, physiology 
Vincent Mok [262012 159* Multiple logistic regression Severe WMH (vs without severe WMH) ∼ MCA PI OR = 1.33, 95% CI [1.04,1.70] per 0.1 increase in PI <0.01 Age 
Ioannis Heliopoulos [142012 52 Multivariable regression WMH score ∼ MCA PI β = 0.262 P=0.025 Age, sex, BMI, HBp, DM, hyperlipidaemia, smoking 
Iria Rodriguez [252010 186 Multivariate logistic analysis Lacunar infarct (vs non-lacunar) ∼ MCA PI OR = 8.13, 95% CI [1.17, 56.27] 0.034 Previous ranking score, WMH, retinal microangiopathy 
Rosa M. Sánchez-Pérez [242003 116 Multiple linear regression Leukoaraiosis severity score ∼ MCA PI β = −0.108 0.353 Age, sex, vascular risk factors, cognitive performance, blood flow velocity in MCA 
Chelsea S. Kidwell [232001 55 Multivariate regression WMH score ∼ MCA PI 0.71 P<0.05 Age, sex, HBp, coronary artery disease 
Brain tissue pulsatility 
Redouane Ternifi [202014 Non-parametric spearman correlation WMH volume ∼ Max BTD ρ = −0.86 <0.001 No 
    WMH volume ∼ Mean BTD ρ = −0.72 <0.001 No 
MRI studies 
Clive B. Beggs [102016 101 Multiple linear regression total WMH volume ∼ CSF peak negative velocity β = −124.903 P=0.041 Age 
Anders Wåhlin [392014 37 Ordinary linear regression Total brain volume ∼ arterial PI β = −0.42 P<0.01 Age, ICV, arterial net flow 
    Total brain volume ∼ CSF flow volume pulsatility β = −0.44 P<0.01 Age, ICV, arterial net flow 
Todd A.D Jolly [382013 35 Partial correlation WMH volume ∼ pulse wave amplitude in arteries or venous sinuses not shown P>0.05 Age 
    WMH volume ∼ pulse width in arteries or sinuses not shown P>0.05 Age 
    WMH volume ∼ PI in arteries or venous sinuses not shown P>0.05 Age 

Abbreviations: BMI, body mass index; BTD, brain tissue displacement; CI, confidence interval; CSF, cerebrospinal fluid; DM, diabetes mellitus; DWI, diffusion-weighted imaging; DWMH, deep white matter hyperintensities; HBp, hypertension; ICA, internal carotid artery; ICV, intracranial volume; IMT, intima-media thickness; MCA, middle cerebral artery; OR, odds ratio; PI, pulsatility index; PVH, periventricular white matter hyperintensities; SD, standard deviation; WMH, white matter hyperintensities.

*

The original sample size of this study [26] was 205 but only 159 participants were included in the analysis.

Doppler ultrasound studies

Two studies measured ICA and eleven measured MCA. Most studies (apart from two [19,24]) reported a significant association between increased ICA or MCA PI and more WMH after adjustment for age. One (n=186) found that higher MCA PI was predictive of having a lacunar infarct (vs other types of infarct) [25]. One paper (n=159) mentioned that they did not find significant associations between MCA PI and microbleeds or lacunes, although no detailed information was provided [26].

One study (n=9) found that higher brain tissue displacement, which was used for representing brain tissue pulsatility, was significantly correlated with larger WMH volume, however it did not adjust for any co-variates [20] (Table 3).

MRI studies

All three MRI studies adjusted for age. Two assessed WMH volume showing that increased WMH volume was significantly associated with higher CSF systolic peak velocity in one study (n=101, β = − 124.903, P=0.041) [10], but not with arterial or venous pulsatility (pulse amplitude, pulse width or PI) in the other (n=35) [38].

One study (n=37) found that increased arterial PI and cervical CSF pulsatility were associated with smaller brain volume in healthy volunteers [39] (Table 3).

Discussion

We identified 27 studies that assessed intracranial pulsatility in relation to SVD features including 3356 subjects. Most studies found a significant association between increased intracranial pulsatility and SVD. However, these studies showed considerable heterogeneity with regard to participants’ clinical characteristics, adjustment for co-variates, image acquisitions and processing, vessels or regions of interest studied, and pulsatility measures used. About half of the studies gave little detail on control of bias, such as use of blinding. We were not able to perform a formal meta-analyse due to the substantial heterogeneities and limited data, although we were able to calculate summary statistics for WMH and pulsatility for some studies.

The limitations of the literature include that SVD features differed or were assessed differently across studies. Most studies used WMH volume or semi-quantitative score to represent SVD burden. Only half of the papers reported the expertise of the observers doing the SVD rating. Semi-quantitative scales and volumes of WMH were shown to be closely correlated and nearly equivalent in estimating WMH burden [40,41]. However, volumetric methods might be more sensitive in detecting subtle WMH differences [42] and may require smaller sample sizes in longitudinal studies [43], but are more difficult to undertake and require more resources than visual rating in a large-scale study with follow-ups. Many studies used automated approaches to measure WMH volume but very few reported whether the WMH masks were manually checked, but failure to check increases errors. So far there is no automated method that can identify WMH without any manual input [44]. Two studies used recent small subcortical (lacunar) ischaemic stroke as the SVD feature: one referred to subcortical infarct on imaging whereas one used the definition of “small vessel disease” in the TOAST classification. Although there is overlap between the two definitions, they are different: “small vessel disease” in TOAST involves clinical features and consideration of vascular risk factors and does not necessarily need imaging evidence [45]. Studies that assessed pulsatility in relation to other SVD features including lacunes [26], microbleeds [26], or atrophy [39] were scant and lacked details of analysis. So far no clinical studies have reported the relationship between cerebrovascular pulsatility and PVS visibility. Enlarged PVS is one of the most consistent imaging findings of SVD, but the underlying pathophysiology remains unknown. Increasing evidence suggest that altered cerebrovascular pulsatility might affect the ISF–CSF exchange and impede the clearance of toxic solutes through the perivascular glymphatic system, which might possibly result in enlarged PVS [7,46]. Thus future SVD studies should consider including PVS especially when investigating the role of cerebrovascular pulsatility in SVD.

A third of papers provided little detail on patients’ demographic or health characteristics. Age and blood pressure are thought to influence intracranial pulsatility [47,48] and are also important risk factors of SVD [2] and should be adjusted for in relevant studies. Most studies that performed correlation or regression analysis have adjusted for age, but in studies that performed comparisons of WMH, patients with more severe WMH were significantly older than those with mild WMH [9,27,36,37]. Blood pressure also changes with age, but very few studies accounted for blood pressure in their analyses although some included history of hypertension as a co-variate.

Indices used to represent pulsatility also varied. Most studies focused on the ICA or MCA. When calculating vascular PI, both MRI and ultrasound studies applied Gosling’s equation. However, some MRI scanners only collected flow values at 16 time points in the cardiac cycle [18,37], meaning that low temporal resolution might affect the accuracy of the PI value as the peak flow might have been missed. In studies that used Doppler ultrasound, another source of variability is inherent when using the technique, including the dependency on the skills of ultrasound technicians and the positioning of the probe. Although the reliability of Gosling’s PI in representing vascular resistance has been questioned [49], there is evidence that ICA and MCA PI are well correlated with cerebrovascular reactivity measured using CO2 stimulus or invasive monitoring [50,51]. One study used a novel ultrasound technique to measure brain tissue pulsatile movement, but it only had nine participants and the validity still needs to be tested [20].

Apart from including large arteries, four MRI studies also considered pulsatility in veins and CSF. As the volume inside the cranium is fixed, the venous system and CSF are also important components in compensating for arterial pressure [52]. Two studies calculated venous PI [9,38] and one measured venous stroke volume [13]. CSF pulsatility assessment varied in terms of both locations and indices. There is so far no accepted definition of pulsatility in CSF, although stroke volume seemed to be the most used. Future studies need to test the reliability of different measures and also select ones that provide more relevant measurements about pulsatility. In addition, one MRI study measured the delay between arterial peak and venous sinus peak [13] whereas another looked at the delay between arterial blood and aqueduct CSF pulsations [18]. These two measures also might be non-comparable, as it is suggested that CSF pulse through the aqueduct is associated with capillary expansion whereas the venous pulse at neck level relates more directly to the arterial expansion [52]. So far no studies have measured pulsatility in vessels more downstream than MCA or PCA due to the limitations of the methodology, such as the spatial resolution or the sequences of the MRI scanner. Future SVD studies could consider techniques such as 4D phase-contrast MRI or 7-T MRI which enables flow assessment in multiple vessels including perforating arteries [53,54], or blood-oxygenation-level-dependent MRI or ultra-fast magnetic resonance encephalography which could measure pulsatility in brain tissues [55,56].

Despite these heterogeneities, in general, most cross-sectional studies found that arterial or venous pulsatility was associated with worse SVD, although the relationship could be confounded by risk factors, particularly age and blood pressure. For ICA, one community based-study (n=700) with all participants aged 70 years that adjusted for age and other medical co-variates, found increased ICA PI to be independently associated with larger WMH volume [12], whereas in another study of diabetic patients (n=122) the significance of the association disappeared after adjustment for age [19]. The relationship between increased MCA PI and WMH or lacunar infarct was found in most studies after adjustment for confounding factors. We are unable to draw conclusions on the relationship between CSF pulsatility and WMH because of different indices and locations used by each study, but the trend seems to be that larger CSF stroke volume is related to more WMH. It was also impossible to conclude if any specific SVD features are more associated with pulsatility due to the very limited data on any features other than WMH.

This is the first paper to comprehensively summarize studies that measured intracranial pulsatility in relation to SVD. The strengths included a systematic search including papers in non-English languages and a careful assessment of all included studies. However, we were not able to perform a meta-analysis due to many sources of heterogeneity, or to pool the results of association analyses as regression analyses were performed differently in each study.

In conclusion, most of the data support a cross-sectional association between SVD and higher pulsatility in large intracranial arteries such as MCA and ICA, although whether a specific SVD feature is more related to high intracranial pulsatility remains unknown, and it is not known if high pulsatility leads to more WMH or the opposite. Therefore, methodologically robust longitudinal studies are required to help establish cause and effect, such as measuring both WMH and intracranial pulsatility at baseline and follow-up. The sample size and follow-up duration of the longitudinal study might partly depend on the baseline WMH burden and which method of WMH estimation is used (visual rating or volumetric quantification), since baseline WMH burden is an important predictor for WMH progression and should be accounted for in the study design and analysis [57]. Ultimately, randomized clinical trials of agents to reduce vessel stiffness will be required to determine if reduction in pulsatility (as a measure of stiffness) prevents SVD progression. Doppler ultrasound might be affordable and easy to use when measuring pulsatility, but it has only limited access to individual vessels and requires experienced technicians. MRI techniques enable assessment of pulsatility in multiple and smaller vessels and in different types of brain tissues, which therefore should be encouraged in future studies. Agreement on reliable measures of intracranial pulsatility is also needed to allow for better comparison between studies, especially for CSF pulsatility. Future studies should clearly define participants’ clinical features, use blinding, improve expertise in SVD assessment, and adjust for relevant co-variates.

Clinical perspectives

  • Increasing evidence suggest that high intracranial pulsatility might be an underlying mechanism of small vessel disease (SVD).

  • Most studies support a cross-sectional association between higher pulsatility in large intracranial arteries and white matter hyperintensities, but there are substantial variations between studies in pulsatility indices, and there is lack of longitudinal data and studies on other important SVD features such as perivascular spaces.

  • Our results suggest that increased intracranial pulsatility might play an important role in the pathophysiology of SVD. However, future studies should minimize bias, adjust for potential confounders, include pulsatility in veins and CSF, and examine longitudinal relationship between pulsatility and SVD.

Author Contribution

Y.S. conceived the idea of the study and did the data search, extraction, and statistical analyses. M.J.T. cross-checked the search strategy. Y.S. drafted the report which was then revised by M.J.T., I.M., and J.M.W. All authors approved the manuscript.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Y.S. is funded by China Scholarships Council. M.J.T is funded by NHS Lothian Research and Development Office. We acknowledge support from the Scottish Funding Council and Chief Scientist Office through the Scottish Imaging Network A Platform for Scientific Excellence (SINAPSE), the Medical Research Council through the Centre for Cognitive Ageing and Cognitive Epidemiology (CCACE), the Fondation Leducq Network for the Study of Perivascular Spaces in Small Vessel Disease (ref. no. 16 CVD 05) and European Union Horizon 2020, PHC-03-15, project No 666881, ‘SVDs@Target’. The work is that of the authors and does not reflect the views of the funders.

Competing interests

The authors declare that there are no competing interests associated with the manuscript.

Abbreviations

     
  • CAA

    cerebral amyloid angiopathy

  •  
  • CI

    confidence interval

  •  
  • CSF

    cerebrospinal fluid

  •  
  • CT

    computed tomography

  •  
  • ECG

    electrocardiogram

  •  
  • ICA

    internal carotid artery

  •  
  • MCA

    middle cerebral artery

  •  
  • MOOSE

    meta-analysis of observational studies in epidemiology

  •  
  • MRI

    magnetic resonance imaging

  •  
  • PI

    pulsatility index

  •  
  • PVS

    perivascular space

  •  
  • SMD

    standardised mean difference

  •  
  • STRIVE

    Standards for Reporting Vascular changes on nEuroimaging

  •  
  • SVD

    small vessel disease

  •  
  • TOAST

    according to Trial of ORG 1072 in Acute Stroke Treatment

  •  
  • WMH

    white matter hyperintensities

References

References
1
Gorelick
P.B.
,
Scuteri
A.
,
Black
S.E.
,
Decarli
C.
,
Greenberg
S.M.
,
Iadecola
C.
et al. 
(
2011
)
Vascular contributions to cognitive impairment and dementia: a statement for healthcare professionals from the american heart association/american stroke association
.
Stroke
42
,
2672
2713
[PubMed]
2
Wardlaw
J.M.
,
Smith
C.
and
Dichgans
M.
(
2013
)
Mechanisms of sporadic cerebral small vessel disease: insights from neuroimaging
.
Lancet Neurol.
12
,
483
497
[PubMed]
3
Shi
Y.
and
Wardlaw
J.
(
2016
)
Update on cerebral small vessel disease: a dynamic whole-brain disease
.
Stroke Vasc. Neurol.
2
,
83
92
4
Mitchell
G.F.
,
van Buchem
M.A.
,
Sigurdsson
S.
,
Gotal
J.D.
,
Jonsdottir
M.K.
,
Kjartansson
O.
et al. 
(
2011
)
Arterial stiffness, pressure and flow pulsatility and brain structure and function: the Age, Gene/Environment Susceptibility–Reykjavik study
.
Brain
134
,
3398
3407
[PubMed]
5
Poels
M.M.
,
Zaccai
K.
,
Verwoert
G.C.
,
Vernooij
M.W.
,
Hofman
A.
,
van der Lugt
A.
et al. 
(
2012
)
Arterial stiffness and cerebral small vessel disease: the Rotterdam Scan Study
.
Stroke
43
,
2637
2642
[PubMed]
6
Mitchell
G.F.
(
2008
)
Effects of central arterial aging on the structure and function of the peripheral vasculature: implications for end-organ damage
.
J. Appl. Physiol. (1985)
105
,
1652
1660
[PubMed]
7
Iliff
J.J.
,
Wang
M.
,
Zeppenfeld
D.M.
,
Venkataraman
A.
,
Plog
B.A.
,
Liao
Y.
et al. 
(
2013
)
Cerebral arterial pulsation drives paravascular CSF-interstitial fluid exchange in the murine brain
.
J. Neurosci.
33
,
18190
18199
[PubMed]
8
Mestre
H.
,
Kostrikov
S.
,
Mehta
R.I.
and
Nedergaard
M.
(
2017
)
Perivascular spaces, glymphatic dysfunction, and small vessel disease
.
Clin. Sci.
131
,
2257
2274
[PubMed]
9
Bateman
G.A.
(
2002
)
Pulse-wave encephalopathy: a comparative study of the hydrodynamics of leukoaraiosis and normal-pressure hydrocephalus
.
Neuroradiology
44
,
740
748
[PubMed]
10
Beggs
C.B.
,
Magnano
C.
,
Shepherd
S.J.
,
Belov
P.
,
Ramasamy
D.P.
,
Hagemeier
J.
et al. 
(
2016
)
Dirty-appearing white matter in the brain is associated with altered cerebrospinal fluid pulsatility and hypertension in individuals without neurologic disease
.
J. Neuroimaging
26
,
136
143
[PubMed]
11
Battal
B.
,
Kocaoglu
M.
,
Bulakbasi
N.
,
Husmen
G.
,
Tuba Sanal
H.
and
Tayfun
C.
(
2011
)
Cerebrospinal fluid flow imaging by using phase-contrast MR technique
.
Br. J. Radiol.
84
,
758
765
[PubMed]
12
Aribisala
B.S.
,
Morris
Z.
,
Eadie
E.
,
Thomas
A.
,
Gow
A.
,
Valdes Hernandez
M.C.
et al. 
(
2014
)
Blood pressure, internal carotid artery flow parameters, and age-related white matter hyperintensities
.
Hypertension
63
,
1011
1018
[PubMed]
13
Bateman
G.A.
,
Levi
C.R.
,
Schofield
P.
,
Wang
Y.
and
Lovett
E.C.
(
2008
)
The venous manifestations of pulse wave encephalopathy: windkessel dysfunction in normal aging and senile dementia
.
Neuroradiology
50
,
491
497
[PubMed]
14
Heliopoulos
I.
,
Artemis
D.
,
Vadikolias
K.
,
Tripsianis
G.
,
Piperidou
C.
and
Tsivgoulis
G.
(
2012
)
Association of ultrasonographic parameters with subclinical white-matter hyperintensities in hypertensive patients
.
Cardiovasc. Psychiatry Neurol.
,
2012
,
616572
[PubMed]
15
Stroup
D.F.
,
Berlin
J.A.
,
Morton
S.C.
,
Olkin
I.
,
Williamson
G.D.
,
Rennie
D.
et al. 
(
2000
)
Meta-analysis of observational studies in epidemiology: a proposal for reporting. Meta-analysis Of Observational Studies in Epidemiology (MOOSE) group
.
JAMA
283
,
2008
2012
[PubMed]
16
Wardlaw
J.M.
,
Smith
E.E.
,
Biessels
G.J.
,
Cordonnier
C.
,
Fazekas
F.
,
Frayne
R.
et al. 
(
2013
)
Neuroimaging standards for research into small vessel disease and its contribution to ageing and neurodegeneration
.
Lancet Neurol.
12
,
822
838
[PubMed]
17
Higgins
J.
and
Green
S.
(eds). (
2011
)
Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0 [updated March 2011]
,
The Cochrane Collaboration
,
18
Naish
J.H.
,
Baldwin
R.C.
,
Patankar
T.
,
Jeffries
S.
,
Burns
A.S.
,
Taylor
C.J.
et al. 
(
2006
)
Abnormalities of CSF flow patterns in the cerebral aqueduct in treatment-resistant late-life depression: a potential biomarker of microvascular angiopathy
.
Magn. Reson. Med.
56
,
509
516
[PubMed]
19
Tanaka
T.
,
Shimizu
T.
and
Fukuhara
T.
(
2009
)
The relationship between leukoaraiosis volume and parameters of carotid artery duplex ultrasonographic scanning in asymptomatic diabetic patients
.
Comput. Med. Imaging Graph.
33
,
489
493
[PubMed]
20
Ternifi
R.
,
Cazals
X.
,
Desmidt
T.
,
Andersson
F.
,
Camus
V.
,
Cottier
J.P.
et al. 
(
2014
)
Ultrasound measurements of brain tissue pulsatility correlate with the volume of MRI white-matter hyperintensity
.
J. Cereb. Blood Flow Metab.
34
,
942
944
[PubMed]
21
Turk
M.
,
Zupan
M.
,
Zaletel
M.
,
Zvan
B.
and
Oblak
J.P.
(
2015
)
Carotid arterial hemodynamic in ischemic levkoaraiosis suggests hypoperfusion mechanism
.
Eur. Neurol.
73
,
310
315
[PubMed]
22
Biedert
S.
,
Forstl
H.
and
Hewer
W.
(
1995
)
Multiinfarct dementia vs Alzheimer’s disease: Sonographic criteria
.
Angiology
46
,
129
135
[PubMed]
23
Kidwell
C.S.
,
El-Saden
S.
,
Livshits
Z.
,
Martin
N.A.
,
Glenn
T.C.
and
Saver
J.L.
(
2001
)
Transcranial doppler pulsatility indices as a measure of diffuse small-vessel disease
.
J. Neuroimaging
11
,
229
235
[PubMed]
24
Sanchez-Perez
R.M.
,
Hernandez-Lorido
R.
,
Castano
P.
,
Diaz-Marin
C.
,
Carneado-Ruiz
J.
and
Molto-Jorda
J.
(
2003
)
Evaluation of hemodynamic parameters by transcranial Doppler in patients with leukoaraiosis. [Spanish]
.
Rev. Neurol.
37
,
301
311
[PubMed]
25
Rodriguez
I.
,
Lema
I.
,
Blanco
M.
,
Rodriguez-Yanez
M.
,
Leira
R.
and
Castillo
J.
(
2010
)
Vascular retinal, neuroimaging and ultrasonographic markers of lacunar infarcts
.
Int. J. Stroke
5
,
360
366
[PubMed]
26
Mok
V.
,
Ding
D.
,
Fu
J.
,
Xiong
Y.
,
Chu
W.W.C.
,
Wang
D.
et al. 
(
2012
)
Transcranial doppler ultrasound for screening cerebral small vessel disease: A community study
.
Stroke
43
,
2791
2793
[PubMed]
27
Webb
A.J.S.
,
Simoni
M.
,
Mazzucco
S.
,
Kuker
W.
,
Schulz
U.
and
Rothwell
P.M.
(
2012
)
Increased cerebral arterial pulsatility in patients with leukoaraiosis: arterial stiffness enhances transmission of aortic pulsatility
.
Stroke
43
,
2631
2636
[PubMed]
28
Purkayastha
S.
,
Fadar
O.
,
Mehregan
A.
,
Salat
D.H.
,
Moscufo
N.
,
Meier
D.S.
et al. 
(
2014
)
Impaired cerebrovascular hemodynamics are associated with cerebral white matter damage
.
J. Cereb. Blood Flow Metab.
34
,
228
234
[PubMed]
29
Sargento-Freitas
J.
,
Felix-Morais
R.
,
Ribeiro
J.
,
Gouveia
A.
,
Nunes
C.
,
Duque
C.
et al. 
(
2014
)
Different locations but common associations in subcortical hypodensities of presumed vascular origin: Cross-sectional study on clinical and neurosonologic correlates
.
BMC Neurology
14
,
24
,
30
Sanahuja
J.
,
Alonso
N.
,
Diez
J.
,
Ortega
E.
,
Rubinat
E.
,
Traveset
A.
et al. 
(
2016
)
Increased burden of cerebral small vessel disease in patients with type 2 diabetes and retinopathy
.
Diabetes Care
39
,
1614
1620
[PubMed]
31
Smith
E.E.
,
Vijayappa
M.
,
Lima
F.
,
Delgado
P.
,
Wendell
L.
,
Rosand
J.
et al. 
(
2008
)
Impaired visual evoked flow velocity response in cerebral amyloid angiopathy
.
Neurology
71
,
1424
1430
[PubMed]
32
Hiroki
M.
,
Miyashita
K.
,
Yoshida
H.
,
Hirai
S.
and
Fukuyama
H.
(
2003
)
Central retinal artery Doppler flow parameters reflect the severity of cerebral small-vessel disease
.
Stroke
34
,
e92
4
33
Del Brutto
O.H.
,
Mera
R.M.
,
De La Luz Andrade
M.
,
Castillo
P.R.
,
Zambrano
M.
and
Nader
J.A.
(
2015
)
Disappointing reliability of pulsatility indices to identify candidates for magnetic resonance imaging screening in population-based studies assessing prevalence of cerebral small vessel disease
.
J. Neurosci. Rural Prac.
6
,
336
338
[PubMed]
34
Ghorbani
A.
,
Ahmadi
M.J.
and
Shemshaki
H.
(
2015
)
The value of transcranial Doppler derived pulsatility index for diagnosing cerebral small-vessel disease
.
Adv. Biomed. Res.
4
,
54
[PubMed]
35
Han
S.W.
,
Song
T.J.
,
Bushnell
C.D.
,
Lee
S.S.
,
Kim
S.H.
,
Lee
J.H.
et al. 
(
2014
)
Cilostazol decreases cerebral arterial pulsatility in patients with mild white matter hyperintensities: Subgroup analysis from the effect of cilostazol in acute lacunar infarction based on pulsatility index of transcranial doppler (ECLIPse) study
.
Cerebrovasc. Dis.
38
,
197
203
[PubMed]
36
Bettermann
K.
,
Slocomb
J.E.
,
Shivkumar
V.
and
Lott
MEJ
(
2012
)
Retinal vasoreactivity as a marker for chronic ischemic white matter disease?
J. Neurol. Sci.
322
,
206
210
[PubMed]
37
Henry-Feugeas
M.C.
,
Roy
C.
,
Baron
G.
and
Schouman-Claeys
E.
(
2009
)
Leukoaraiosis and pulse-wave encephalopathy: Observations with phase-contrast MRI in mild cognitive impairment
.
J. Neuroradiology
36
,
212
218
38
Jolly
TAD
,
Bateman
G.A.
,
Levi
C.R.
,
Parsons
M.W.
,
Michie
P.T.
and
Karayanidis
F.
(
2013
)
Early detection of microstructural white matter changes associated with arterial pulsatility
.
Front. Human Neurosci.
,
7
,
782
[PubMed]
39
Wahlin
A.
,
Ambarki
K.
,
Birgander
R.
,
Malm
J.
and
Eklund
A.
(
2014
)
Intracranial pulsatility is associated with regional brain volume in elderly individuals
.
Neurobiol. Aging
35
,
365
372
[PubMed]
40
Valdes Hernandez Mdel
C.
,
Morris
Z.
,
Dickie
D.A.
,
Royle
N.A.
,
Munoz Maniega
S.
,
Aribisala
B.S.
et al. 
(
2013
)
Close correlation between quantitative and qualitative assessments of white matter lesions
.
Neuroepidemiology
40
,
13
22
[PubMed]
41
Oboudiyat
C.
,
Garderner
H.
,
Marquez
C.
,
Elkind
M.
,
Sacco
R.
,
DeCarli
C.
et al. 
(
2014
)
Comparing Semi-quantitative and Volumetric Measurements of MRI White Matter Hyperintensities: The Northern Manhattan Study
.
Neurology
82
,
Supplement S62.007
42
van Straaten
E.C.
,
Fazekas
F.
,
Rostrup
E.
,
Scheltens
P.
,
Schmidt
R.
,
Pantoni
L.
et al. 
(
2006
)
Impact of white matter hyperintensities scoring method on correlations with clinical data: the LADIS study
.
Stroke
37
,
836
840
[PubMed]
43
Chappell
F.M.
,
Del Carmen Valdes Hernandez
M.
,
Makin
S.D.
,
Shuler
K.
,
Sakka
E.
,
Dennis
M.S.
et al. 
(
2017
)
Sample size considerations for trials using cerebral white matter hyperintensity progression as an intermediate outcome at 1 year after mild stroke: results of a prospective cohort study
.
Trials
18
,
78
[PubMed]
44
Wardlaw
J.M.
,
Valdes Hernandez
M.C.
and
Munoz-Maniega
S.
(
2015
)
What are white matter hyperintensities made of? Relevance to vascular cognitive impairment
J. Am. Heart Assoc.
4
,
001140
[PubMed]
45
Adams
H.P.
Jr
,
Bendixen
B.H.
,
Kappelle
L.J.
,
Biller
J.
,
Love
B.B.
,
Gordon
D.L.
et al. 
(
1993
)
Classification of subtype of acute ischemic stroke. Definitions for use in a multicenter clinical trial. TOAST. Trial of Org 10172 in Acute Stroke Treatment
.
Stroke
24
,
35
41
[PubMed]
46
Weller
R.O.
,
Djuanda
E.
,
Yow
H.Y.
and
Carare
R.O.
(
2009
)
Lymphatic drainage of the brain and the pathophysiology of neurological disease
.
Acta Neuropathol.
117
,
1
14
[PubMed]
47
Stoquart-ElSankari
S.
,
Baledent
O.
,
Gondry-Jouet
C.
,
Makki
M.
,
Godefroy
O.
and
Meyer
M.E.
(
2007
)
Aging effects on cerebral blood and cerebrospinal fluid flows
.
J. Cereb. Blood Flow Metab.
27
,
1563
1572
[PubMed]
48
Wagshul
M.E.
,
Eide
P.K.
and
Madsen
J.R.
(
2011
)
The pulsating brain: a review of experimental and clinical studies of intracranial pulsatility
.
Fluids Barriers CNS
8
,
5
[PubMed]
49
Michel
E.
and
Zernikow
B.
(
1998
)
Gosling’s Doppler pulsatility index revisited
.
Ultrasound Med. Biol.
24
,
597
599
[PubMed]
50
Czosnyka
M.
,
Richards
H.K.
,
Whitehouse
H.E.
and
Pickard
J.D.
(
1996
)
Relationship between transcranial Doppler-determined pulsatility index and cerebrovascular resistance: an experimental study
.
J. Neurosurg.
84
,
79
84
[PubMed]
51
Beasley
M.G.
,
Blau
J.N.
and
Gosling
R.G.
(
1979
)
Changes in internal carotid artery flow velocities with cerebral vasodilation and constriction
.
Stroke
10
,
331
335
[PubMed]
52
Stivaros
S.M.
and
Jackson
A.
(
2007
)
Changing concepts of cerebrospinal fluid hydrodynamics: role of phase-contrast magnetic resonance imaging and implications for cerebral microvascular disease
.
Neurotherapeutics
4
,
511
522
[PubMed]
53
Rivera-Rivera
L.A.
,
Turski
P.
,
Johnson
K.M.
,
Hoffman
C.
,
Berman
S.E.
,
Kilgas
P.
et al. 
(
2016
)
4D flow MRI for intracranial hemodynamics assessment in Alzheimer’s disease
.
J. Cereb. Blood Flow Metab.
36
,
1718
1730
[PubMed]
54
Bouvy
W.H.
,
Geurts
L.J.
,
Kuijf
H.J.
,
Luijten
P.R.
,
Kappelle
L.J.
,
Biessels
G.J.
et al. 
(
2016
)
Assessment of blood flow velocity and pulsatility in cerebral perforating arteries with 7-T quantitative flow MRI
.
NMR Biomed.
29
,
1295
1304
[PubMed]
55
Kiviniemi
V.
,
Wang
X.
,
Korhonen
V.
,
Keinanen
T.
,
Tuovinen
T.
,
Autio
J.
et al. 
(
2016
)
Ultra-fast magnetic resonance encephalography of physiological brain activity - Glymphatic pulsation mechanisms?
J. Cereb. Blood Flow Metab.
36
,
1033
1045
[PubMed]
56
Makedonov
I.
,
Black
S.E.
and
Macintosh
B.J.
(
2013
)
BOLD fMRI in the white matter as a marker of aging and small vessel disease
.
PLoS One
8
,
e67652
[PubMed]
57
Gouw
A.A.
,
van der Flier
W.M.
,
Fazekas
F.
,
van Straaten
E.C.
,
Pantoni
L.
,
Poggesi
A.
et al. 
(
2008
)
Progression of white matter hyperintensities and incidence of new lacunes over a 3-year period: the Leukoaraiosis and Disability study
.
Stroke
39
,
1414
1420
[PubMed]

Supplementary data