Received: 27 March 2018 Revised: 9 August 2018 Accepted: 6 September 2018

DOI: 10.1002/jcu.22645

RESEARCH ARTICLE

Cerebral autoregulation in hemorrhagic stroke: A systematic

review and meta-analysis of transcranial Doppler
ultrasonography studies
Jatinder S. Minhas MRes, MRCP1 | Ronney B. Panerai PhD1,2 |
George Ghaly MBChB, MRCP3 | Pip Divall MSc3 | Thompson G. Robinson MD, FRCP1,2

1
Cerebral Haemodynamics in Ageing and
Stroke Medicine (CHiASM) Research Group, Abstract
Department of Cardiovascular Sciences, Purpose: International guidelines advocate intensive blood pressure (BP) lowering within 6 hours
University of Leicester, Leicester, United
of acute intracerebral hemorrhage (ICH) to a target systolic BP of 130-140 mm Hg, though more
Kingdom
2
intensive lowering may be associated with adverse outcome. Observational studies suggest
National Institute for Health Research
Leicester Biomedical Research Centre, impaired cerebral autoregulation (CA) following ICH. Transcranial Doppler ultrasonography
University of Leicester, Leicester, United (TCD), alongside continuous BP monitoring, provides a noninvasive bedside investigation that
Kingdom offers detailed perspectives on physiological perturbations post-acute ICH. This systematic
3
Department of Stroke Medicine, University review and meta-analysis focuses on all TCD studies of CA in ICH.
Hospitals of Leicester NHS Trust, Leicester,
Methods: MEDLINE, EMBASE, and CENTRAL were searched for studies of hemorrhagic stroke
United Kingdom
Correspondence
and blood flow measurement.
Jatinder S. Minhas, MRes, MRCP, Cerebral Results: Eight studies met inclusion criteria (293 ICH patients); CA was impaired up to 12-days
Haemodynamics in Ageing and Stroke post-acute ICH. Impaired CA was evidenced by reduced transfer function analysis phase and
Medicine (CHiASM) Research Group,
higher mean flow correlation values: these were associated with worsened clinical parameters
Department of Cardiovascular Sciences,
University of Leicester, Leicester, LE2 7LX, including ICH-volume and Glasgow Coma Scale. Meta-analysis of CBV demonstrated that, com-
United Kingdom. pared to controls, mean CBV was significantly lower in the ipsilateral (49.7 vs 64.8 cm s−1,
Email: jm591@le.ac.uk
Z = 4.26, P < .0001) and contralateral hemispheres following ICH (51.5 vs 64.8 cm s−1,
Funding information
Z = 3.44, P = .0006).
National Institute for Health Research, Grant/
Award Number: Senior Investigator Award; Conclusion: Lower mean CBV in combination with impaired CA may have implications for more
Dunhill Medical Trust Research Training intensive BP lowering and warrants further studies examining such strategies on cerebral blood
Fellowship, Grant/Award Number: flow and its regulatory mechanisms.
RTF97/0117

KEYWORDS

cerebral blood velocity, cerebral hemodynamics, hemorrhagic stroke, meta-analysis,

transcranial Doppler ultrasonography

1 | B A C KG RO U N D A N D P U R P O S E morbidity.1 Blood pressure (BP) control remains the foremost

Spontaneous acute intracerebral hemorrhage (ICH) is associated with
devastating consequences, as evidenced by both high mortality and
Abbreviations: ARI, autoregulation index; BP, blood pressure; CA, cerebral auto-
regulation; CBV, cerebral blood velocity; CO2, carbon dioxide; CrCP, critical
closing pressure; CVMR, cerebral vasomotor reactivity; dCA, dynamic cerebral
autoregulation; ECG, electrocardiogram; EtCO2, end-tidal CO2; HR, heart rate;
ICH, intracerebral hemorrhage; MABP, mean arterial blood pressure; MCA, mid-
dle cerebral artery; Mx, mean flow correlation index; PaCO2, partial pressure
carbon dioxide; PI, pulsatility index; RAP, resistance-area product; SD, standard
deviation; TCD, transcranial Doppler; TFA, transfer function analysis.
approach to managing acute ICH.2 However, controversy exists as to
whether intensive BP lowering in acute ICH risks cerebral ischemia.3
This is particularly the case if cerebral blood velocity (CBV) control
mechanisms are altered by chronic hypertension or brain injury. Cere-
bral autoregulation (CA) provides a protective mechanism for the brain
parenchyma from extremes of CBV change, with the risk of hyperper-
fusion or hypoperfusion in response to systemic BP changes.4 A key
limitation of large-scale randomized controlled trials has been the abil-
ity to provide mechanistic insight into CBV during the acute phase of
hemorrhagic stroke.

14 © 2018 Wiley Periodicals, Inc. wileyonlinelibrary.com/journal/jcu J Clin Ultrasound. 2019;47:14–21.

MINHAS ET AL.
Importantly, focused metabolic studies examining BP reduction and
risk of perihematomal hypoperfusion have not supported the hypothesis
that BP reduction is associated with perihematomal ischemia, confirming
the safety of early intensive BP treatment in acute ICH.5 These findings
concur with the intensive BP reduction in acute cerebral hemorrhage trial
(INTERACT-2), which did not associate rapid BP reduction with worse
clinical outcomes.6 The INTERACT-2 study randomized patients to
intensive- (target systolic BP < 140 mm Hg) or guideline-based (systolic
BP < 180 mm Hg) BP management. The antihypertensive treatment of
acute cerebral hemorrhage II (ATACH-II) trial used intravenous nicardi-
pine to test the superiority of very intensive reduction (target systolic BP
of 100-130 mm Hg) vs standard treatment (target systolic BP
7 effects of impaired CA and CBV changes on neurological outcome were
140-179 mm Hg). The key difference in the BP targets achieved
between the two studies is demonstrated by the early profile of systolic
BP in the standard-treatment group in ATACH-II being similar to values
observed in the intensive-treatment group in INTERACT-2. 6,7
Primary
outcome measures of death and disability showed no reduction with BP
lowering for both studies, but ATACH-II did not confirm the improved
functional outcome found in INTERACT-2.6,7 However, an important lim-
itation with both studies was the lack of inclusion of individuals with low
Glasgow Coma Score and large intraparenchymal hematomas, thus pro-
viding little data on safety of intensive lowering of systolic BP in individ-
uals with the highest intracranial pressures where CA is likely to be
further impaired.6,7
Transcranial Doppler ultrasonography (TCD) offers repeatable
noninvasive bedside investigation with high temporal resolution of
the steady-state (static) and dynamic components of CA.4 Dynamic
CA (dCA) responds to instantaneous (over seconds) BP and associated
cerebral perfusion pressure changes, and prior studies have demon-
strated impairment in acute ischemic stroke.8,9 dCA is often described
using transfer function analysis (TFA) parameters10; including phase
(reflecting speed of autoregulatory response) and gain (associated
with damping characteristics of dCA). Furthermore, TCD can provide
information regarding the time-course of autoregulation parameters
in acute ICH. Unfortunately, despite several TCD studies to date,
there remains a lack of consensus as to whether CA is impaired in ICH
and whether such impairment relates directly to clinical outcomes.
Furthermore, there has been no amalgamation of the evidence of the
relationship of CA impairment with time post ICH.
TCD studies are highlighted in this review as they offer the advan-
tage of understanding beat-to-beat dynamics of the relationship between
cerebral pressure-flow with attention to fast and slow responses, thus
providing mechanistic insights beyond comparable alternative imaging
methods. Therefore, this review will focus on TCD studies of CA in the
setting of documented acute ICH, and will report impairments in CA and
the natural history of CBV changes following acute ICH.
2 | MATERIAL AND METHODS
2.1 | Study identification
Cochrane Collaboration methodology for meta-analysis reviews modi-
fied for observational studies (www.equator-network.org) was used.
15
2.2 | Search strategy
Studies were identified with a search strategy across three English lan-
guage databases (MEDLINE, EMBASE, and CENTRAL) between 1966
and December 2017 accommodating different MeSH terms or subcate-
gories available on each database (Supporting Information Table 1). Bibli-
ographies of selected articles were screened for additional relevant
articles.
2.3 | Inclusion and exclusion criteria
TCD studies of human CA after ICH were included. Eligibility was
assessed by reading abstracts, and, if necessary, whole articles. The
assessed. Excluded were case reports, non-English language articles, pos-
terior territory stroke studies, and studies with ultrasound contrast agent
injection, as well as other perfusion-based studies with CT or MRI.
2.4 | Data extraction
The following data were extracted: (1) number of patients and con-
trols; (2) sex; (3) age; (4) acute (<72 hours) vs chronic phase assess-
ment (>72 hours); (5) method of data analysis; (6) neurological
outcome; (7) cerebral blood velocity (CBV); (8) partial pressure of car-
bon dioxide (PaCO2); (9) BP; (10) hematoma assessment method and
size; and (11) main conclusions.
Study quality was assessed using a checklist proposed previously
for authors, editors, and reviewers of meta-analyses of observational
studies modified for specific use in TCD autoregulation observational
studies (Supporting Information Tables S2 and S3)8 and based on the
meta-analysis of observational studies in epidemiology: a proposal for
reporting (MOOSE) guidelines.11 Two independent reviewers
(Jatinder S. Minhas and George Ghaly) undertook the methodological
quality screening and data extraction of the included studies.
For the meta-analyses, the variables were compared in acute ICH
patients vs control subjects. The software used was Review Manager
5.3 (RevMan 5) provided by the Cochrane Library. The analysis was
performed using the fixed effects model; weighted mean difference
was used for measurement data and 95% confidence intervals
(CI) were used as the effect indicator for dichotomous variables. The
heterogeneity assumption was checked by the χ2-based Q test.
3 | RE SU LT S
3.1 | General characteristics
Eight-hundred and thirteen publications met the search criteria and
were evaluated (Figure 1). The commonest reasons for exclusion were
alternative imaging modality used (eg, CT, MRI, and NIRS), animal
studies, subarachnoid hemorrhage patients, and neonatal hemorrhage.
Two further studies were excluded after initially meeting the eligibility
criteria, as it was noted that PET and brain tissue oxygenation mea-
surement, rather than TCD, were used.5,12 Overall, the eligibility cri-
teria were met by seven controlled and one observational study
(Table 1). Two studies used the same dataset,13,14 but both were
16
included due to the different methods adopted for assessment of
CA. Median score on the quality checklist was 10, range 9-11
(Supporting Information Table S3). This demonstrated areas of consis-
tent incomplete reporting of previously published minimum accept-
able methodological criteria for observational studies.8
Patient numbers ranged from 12 to 114. All studies included
patients in the “acute” (<72 hours) range, though five studies13–17 also
had measurements beyond 72 hours (chronic phase). Three studies
14,16,17
provided measurements beyond 5 days, with the longest mea-
surement interval from index to follow-up being 30 days.16 In all stud-
ies, except two, neurological outcome data were provided using the
National Institute Health Stroke Scale (NIHSS), with scores ranging
from 5 to 17. All studies included an ABC/2 assessment of hematoma
volume with a range from 14.7 to 51.7 cm3. Two studies reported
radiological outcome measures with correlation of blood flow with CT
data and assessment of vasospasm.15,18 Four studies detailed end-
tidal carbon dioxide (EtCO2) values for acute ICH patients. These ran-
ged from 34.3 to 34.9 mm Hg in the acute phase (<72 hours)13,14,16
and 34.7 to 35.2 mmHg in the chronic (>72 hours) phase.13,14,16,17
Studies provided varying detail for BP recordings, with three providing
no data, three providing mean arterial pressure (MAP), and one pro-
viding systolic readings only.
3.2 | Cerebral hemodynamic parameters
Six of the seven studies provided CBV for each hemisphere
separately,13,14,16–19 with three providing comparable healthy control
FIGURE 1 CONSORT diagram detailing search results
MINHAS ET AL.
CBV values.16–18 All eight studies assessed13–20 cerebral hemody-
namic measurements at rest. In one study, providing CBV data up to
30 days following ICH onset, it was reported that CBV in both the
affected and unaffected hemispheres failed to return to control sub-
ject values at all assessed time-points until Day 30.16
dCA assessed by TFA and mean flow correlation index (Mx), was
impaired in the acute (<72 hours) period in most,13,16,18,20 but not all
studies.14 However, this latter study reported that a “secondary
decline” in dCA more than 5 days post-ICH onset was associated with
clinical features likely to be associated with poor prognosis, including
hematoma volume and GCS.14 A further study found an increase in
pulsatility index (PI) was associated with larger hematoma volumes,
possibly reflecting increased mass effect.19
3.3 | Meta-analysis of CBV data
There was significant heterogeneity in study design and presentation
of basic physiological measurements. Therefore, a quantitative meta-
analysis was only possible for two acute studies (<72 hours) of CBV
data only,16,18 which, compared to controls, demonstrated signifi-
cantly lower mean CBV in the ipsilateral (49.7 vs 64.8 cm s−1,
Z = 4.26, P < .0001) and contralateral hemispheres following ICH
(51.5 vs 64.8 cm s−1, Z = 3.44, P = .0006) (Figure 2A,B). Unfortu-
nately, the studies did not provide data on hand dominance. A more
recent 2017 study my Ma et al was excluded from meta-analysis as
the cohort were of chronic ICH patients assessed at 4-6 days post
TABLE 1 Overview of transcranial Doppler studies with measurements in acute (<72 hours) and chronic (>72 hours) ICH
Number of ICH Acute (<72 h) vs
patients and Age in years chronic phase Method Hematoma assess
MINHAS ET AL.

controls (>72 h) of data Neurological method and size Main conclusions and
−1
Study (male: female) SD (control) assessment analysis severity CBV values (cm s ) EtCO2 values (mm Hg) BP values (systolic SD) (cm3  SD) results
Marti-Fabregas 51 patients (30:21) 66.2  12.4 12 h of onset PI NIHSS 17 Affected mean velocity Nil Nil ABC/2 51.7  70.4 TCD parameters, including
(2005) (acute) 35.6  16.5, unaffected pulsatility index (PI), were
39.6  18.0 correlated with CT data
in the acute stage of ICH
Reinhard 26 patients (21:5), 65  11, Acute D1 and Mx NIHSS 12  Mean MCA CBV(cm s−1) EtCO2(mm Hg)34.9 Mean MAP Finapres(mm ABC/2 26  22 CA preserved in acute ICH,
(2010) 55 controls (64  8) chronic D3 7, GCS Ipsilateral 43.6 (n = 26; (n = 26; SE = 0.9) 34.3 Hg)90.9 (n = 26; but a secondary decline
(44:11) and D5 13  2 SE = 3.4) 55.8 (n = 21; (n = 21; SE = 0.98; SE = 4.26) 85.5 ipsilateral occurred. This
SE = 3.6; P = .0020) 53.6 P = .5322) 35.1 (n = 21; SE = 4.57; decline was associated
(n = 22; SE = 3.6; (n = 22; SE = 0.96; P = .2220) 84.4 with poor clinical status
P = .0091). Contralateral P = .8815) (n = 22; SE = 4.51;
47.7 (n = 26, SE 3.5), 56.6 P = .1337)
(n = 21, SE 3.7, P = .015),
57.9 (n = 22, SE 3.7,
P = .0051)
Kiphuth (2011) 61 ICH patients ICH 62, IVH 70 With 24 h N/A N/A N/A N/A N/A ABC/2 ICH 25, IVH Cerebral vasospasm with
(37:24), (acute), 5-7 d 41 secondary infarction may
53 IVH patients (chronic) and occur in patients with
(28:25) 7-9 d spontaneous IVH
(chronic)
Nakagawa 21 patients (8:13), 66  15 <72 h (mean TFA Nil Ipsilateral 49.8  22.5/ Nil 77  15 (89  21) ABC/2 36  27 dCA may be less effective in
(2011) 23 controls (8:15) (65  9) time TCD contralateral 50.6  24.2 patients with
34  14) (control 66.6  18.9) spontaneous lobar or
basal ganglia ICH
compared to healthy
controls
Oeinck (2013) 26 patients (21:5), 65  11, Acute D1 and TFA NIHSS 12  Mean MCA CBV(cm s−1) EtCO2(mm Hg)34.9 Mean MAP Finapres(mm ABC/2 26  22 Dynamic temporal
55 controls (64  8) chronic D3 7, GCS Ipsilateral 43.6 (n = 26; (n = 26; SE = 0.9) 34.3 Hg)90.9 (n = 26; characteristics of CA
(44:11) and D5 13  2 SE = 3.4) 55.8 (n = 21; (n = 21; SE = 0.98; SE = 4.26) 85.5 (phase) are not generally
SE = 3.6; P = .0020) 53.6 P = .5322) 35.1 (n = 21; SE = 4.57; altered in acute ICH.
(n = 22; SE = 3.6; (n = 22; SE = 0.96; P = .2220) 84.4 Poorer individual phase
P = .0091). Contralateral P = .8815) (n = 22; SE = 4.51; values are, however,
47.7 (n = 26, SE 3.5), 56.6 P = .1337) associated with larger
(n = 21, SE 3.7, P = .015), ICH volume, lower blood
57.9 (n = 22, SE 3.7, pressure, and worsened
P = .0051) outcome
Ma (2016) 43 patients (31:13), 53.7  10.0 Acute (D1-2), TFA Admission D1-2 (49.6  19.1 affected, EtCO2(mm Hg) D1-2 (122.5  15.1), ABC/2 15.3  10.6 dCA is bilaterally impaired
30 controls (21:9) (52.3  8.1) chronic (4-6, (7.4  5.0), 52.3  20.6 unaffected), D1-2 (34.3  3.7), D4-6 D4-6 (117.2  14.1), lasting at least 10 to 12 d
10-12 and discharge D4-6 (56.0  17.2 (34.9  2.8), D10-12 D10-12 and recovers within a
D30) (5.1  4.5) affected, 58.4  16.7 (34.8  2.2), D30 (113.2  11.9), D30 month
unaffected), D10-12 (35.2  2.4) control (107.6  12.1) control
(60.0  14.5 affected, 35.1  2.5 92.1  11.6
61.2  16.1 unaffected),
D30 (59.3  14.1
affected, 62.4  13.8
unaffected) control
62.9  13.0
Lee (2017) 12 patients (11:1), 59.3  13.1, Within 3 d Mx Nil N/A N/A N/A ABC/2 21.7  28.6 Acute ICH may impair CA
7 controls (6:1) (61  9.2) (<72 h) and increase CBV
variability ipsilateral to
the hematoma

(Continues)
17
18 MINHAS ET AL.

stroke.17 However, this study demonstrated lower CBV values and

National Institutes of Health Stroke Scale; GCS, Glasgow Coma Score; TFA, Transfer Function Analysis; CA, cerebral autoregulation; ICH, intracerebral hemorrhage; IVH, intraventricular hemorrhage; SD, standard devia-
CBV, cerebral blood velocity; BP, blood pressure; CO2, carbon dioxide; PI, pulsatility index; MCA, middle cerebral artery; TCD, transcranial Doppler; MAP, mean arterial pressure; EtCO2, end-tidal carbon dioxide; NIHSS,
dCA is bilaterally disturbed

ipsilateral to hematoma
EtCO2 in ICH patients compared to controls.

independently predict
and larger hematoma
Main conclusions and

volume is likely to

poorer CA status
4 | DISCUSSION
results

This systematic review of cerebral hemodynamics following acute ICH

demonstrates impaired dCA, as assessed by TFA and Mx parameters,
ABC/2 14.7  10.2
Hematoma assess

in both ipsilateral and contralateral hemispheres compared to control

method and size

subjects. This impairment seems to persist for up to 12 days following

(cm3  SD)

ICH onset, though this conclusion is limited by the fact that only two
studies investigated changes in the “chronic” phase beyond 72 hours.
Importantly, worsening CA parameters was associated with clinical
BP values (systolic SD)

features likely to be associated with poor prognosis, including hema-

Control 92.1  11.6

toma volume and GCS score. The hypothesized relationship between

(118.6  15.2)

autoregulation, cerebral blood flow, and clinical outcome is summa-

rized in Figure 3. This concurs with similar findings in ischemic stroke,
D4-6

where worsening CA is associated with adverse prognostic features,

such as hemorrhagic transformation and cerebral edema.21
EtCO2 values (mm Hg)
EtCO2 (mm Hg) D4-6

There are several limitations with the studies included in this

Control 35.1  2.5

systematic review. First, studies were mainly limited to patients

(34.7  3.3)

with mild-to-moderate stroke severity, as evidenced by neurologi-

cal impairment (NIHSS 5-12) and GCS >8. Second, these studies
used BP instead of cerebral perfusion pressure to calculate dCA
tion. All studies examined the MCA and assessments within the acute and subacute phase refer to a single TCD assessment.

parameters, thus neglecting the potential contribution of intracra-

D4-6 (53.6  17.1 affected,
57.2  14.5 unaffected)

nial pressure. Third, no study formally assessed the effects of BP

CBV values (cm s )

lowering therapy on dCA, so there are limited data on the impact

Control 62.9  13.0
−1

of intensive BP lowering on cerebral hemodynamics. Fourth, the

review is limited to data reported in the selected articles and
does not examine individual patient data in the quantitative or
qualitative analyses. Lastly, this review is limited to TCD studies
(as opposed to CT, MRI, or NIRS) as they provide a robust assess-
NIHSS 9 (6-16)
of data Neurological

ment of beat-to-beat, dynamic central and peripheral hemody-

analysis severity

namic changes associated with the physiological perturbations

precipitated by acute ICH. This limits generalizability of the find-
Method

ings to other modalities though does provide important mechanis-

TFA

tic insights nonetheless. Therefore, a mechanistic understanding

from a cerebral hemodynamic perspective of different targets for
Acute (<72 h) vs
chronic phase

intensive BP lowering is lacking. Finally, the included studies gen-

assessment
4–6 d after
onset

erally suffer from methodological heterogeneity, lack of control

(>72 h)

subjects, and relatively small numbers. Nonetheless, the homoge-

neity of the evaluated pathology (supratentorial hemorrhage of
54.3  11.1.
(52.3  8.1)
SD (control)

similar stroke severity) should be regarded as a strength. Overall,

Age in years

the methodological quality of each study was good and assessed

Patients

using objective criteria. Therefore, there are important findings

with respect to key hemodynamic measures.
30 controls (21:9)
53 patients (40:13),
Number of ICH

(male: female)

4.1 | Cerebral blood velocity

patients and
controls
(Continued)

A limited meta-analysis demonstrated lower CBV values in ICH

patients in both the ipsilateral and contralateral hemispheres com-
pared to control subjects, though overall little data exist for CBV
Ma (2017)

changes following acute ICH. By contrast, studies in acute ischemic

TABLE 1

Study

stroke have shown similar lower CBV values for ipsilateral (43.5 cm
s−1) and contralateral (41.1 cm s−1) hemispheres compared to control
MINHAS ET AL. 19

FIGURE 2 Meta-analysis of CBV (cms−1) in acute ICH (<72 hours). A, Mean ipsilateral transcranial Doppler ultrasonographic measurements
(cm s ) compared to mean control measurements and B, Mean contralateral measurements (cm s−1) compared to mean control measurements.
−1

Control values are average values (cm s−1) from ipsilateral and contralateral measurements

subjects (49.6 cm s−1).22 Previous TCD work comparing flow veloci-
ties in small (<25 mL) vs large (>25 mL) ICH volumes measured using
computer tomography (CT), demonstrated that large ICH has consis-
−1 23
tently depressed ipsilateral mean velocities (35  13 cm s ).
4.2 | Pulsatility index
Previous studies have shown increased intracranial pressure and
decreased cerebral perfusion pressure lead to characteristic changes
in TCD waveforms—decrease in diastolic velocity and increase in PI.19
Importantly, Marti-Fabregas et al19 confirmed that PI values obtained
using TCD correlated well with CT signs associated with mass effect,
volume of hematoma and volume of surrounding edema, total volume,
midline shift and intraventricular extension. This correlation was par-
ticularly evident when PI values were significantly raised. Interestingly,
these findings highlight a role for TCD as an indirect assessment
method of intracranial hypertension and mass effect.
4.3 | Mean flow correlation index
In addition to dynamic indices of autoregulation, the time correlation
index known as the Mx, which is based on correlation coefficients
between BP and CBV, was also studied. Two studies assessed Mx in
acute ICH.14,20 Interestingly, the first of the two studies provided neg-
ative findings in the acute phase of ICH (<72 hours), demonstrating
no difference between Mx values in patients and controls up to Day
5. However, higher Mx values (ie, poorer autoregulation) were associ-
ated with a lower GCS, ventricular extension and lower cerebral per-
fusion pressure, ultimately leading to a poorer clinical picture at
90 days. These findings suggest some preservation of CA in acute
ICH, and certainly conflict with findings of studies with TFA and PI
outcome measures. The second more recent study assessed variability
of CBV using nonlinear entropy analyses.20 Ultimately, the authors
drew comparisons between Mx and nonlinear entropy analyses to
assess sensitivity to CA changes. The data suggested that acute ICH
does impair CA and increased CBV variability exists within the ipsilat-
eral hemisphere.20
4.4 | Transfer function analysis
Nakagawa et al18 assessed patients with lobar or basal ganglia hemor-
rhage early (<72 hours) post ICH and compared to controls. The
patients had higher gains across a wide range of frequency ranges
compared to controls. This suggested that dCA is impaired in the early

FIGURE 3 Hypothesized influences of cerebral autoregulation and cerebral blood flow on the outcome of acute intracerebral hemorrhage
20
days after ICH. However, data presented by Oeinck et al13 (same
14
dataset as Reinhard et al showed that phase was not altered in
acute ICH but (similar to Mx) there were prognostic associations. This
included poorer phase being associated with larger ICH volume, lower
BP and worsened outcome. Interestingly, gain was generally higher in
acute ICH but not associated with clinical factors for which phase has
shown an association. The largest TFA-based study in patients with
supratentorial ICH included a robust set of serial measurements (Days
1-2, 4-6, 10-12, and Day 30).16 Phase difference was lower bilaterally
and, therefore, possibly impaired up to Days 10-12 with a significant
recovery by Day 30.16 Importantly, phase difference values were
associated with clinical status in the acute stage, and impaired phase
difference at 4–6 days post-ICH onset was an independent predictor
16
for clinical outcome. More recently, the bilateral disturbance of dCA
was confirmed at the 4- to 6-day period post ICH with larger hema-
toma volume being independently predictive of poorer CA status ipsi-
lateral to the hematoma.
4.5 | Impaired autoregulation
This review highlights the complex interplay that exists between dif-
ferent markers of autoregulatory impairment. Crucially, there exists a
variation of approaches to assessment of autoregulatory impairment
leading to difficulties in interpreting the overall picture due to inher-
ent heterogeneity. However, we can summarize that higher gain,
higher Mx, and lower phase difference are all suggestive of poorer CA
(Figure 3), and provide a platform for future work examining CA
impairment and consideration for potential interventions targeting CA
improvement.
4.6 | Future studies
Future mechanistic work on blood flow is necessary to examine the
nature of the lower CBV values in acute ICH. A dual modality
approach of TCD and arterial spin labeling may provide more informa-
tion. Significant heterogeneity in outcome measures prevented any
further meta-analyses on CA measures or other physiological parame-
ters. Hence, basic physiological data are required in future TCD work
to ensure that more robust intra-study comparisons can be made, par-
ticularly baseline EtCO2 values, CBV values for all time points includ-
ing any controls used. Above all, future studies should include
assessment of dCA and its time course following interventions to
reduce BP.
5 | CO NC LUSIO N
There is limited evidence to suggest that dCA is impaired up to
12 days post ICH. There is evidence for an association between
reduced TFA phase and Mx values, and worsened clinical features,
including hematoma volume and Glasgow Coma Scale values. Further-
more, lower bilateral CBV values as compared to controls provide a
rationale for further studies examining the impact of intensive BP low-
ering strategies on cerebral blood flow and its regulatory mechanisms.
MINHAS ET AL.
AUTHOR CONT RIB UTI ON S
All authors read and approved the final manuscript.
Study design and suggestions on revising the article: Jatinder
S. Minhas, Ronney B. Panerai, and Thompson G. Robinson.
Searched and selected the studies: Jatinder S. Minhas, George
Ghaly, and Pip Divall.
Data analysis; drafted and revised the article: Jatinder S. Minhas.
Preparation of figures: Jatinder S. Minhas.
CONFLIC T OF INT E RE ST
All authors certify that they have no conflict of interest in the subject
matter or materials discussed in this manuscript.
ET HICS APPROVAL AND CONSENT TO PARTICIPATE
Not applicable.
AVAILABI LITY OF DATA AND MATERIALS
All data generated or analyzed during this study are included in this
published article.
COMPET ING INTERESTS
The authors declare that they have no competing interest.
ORCID
Jatinder S. Minhas https://orcid.org/0000-0002-0576-9105
RE FE RE NC ES
1. van Asch CJ, Luitse MJ, Rinkel GJ, van der Tweel I, Algra A, Klijn CJ.
Incidence, case fatality, and functional outcome of intracerebral hae-
morrhage over time, according to age, sex, and ethnic origin: a system-
atic review and meta-analysis. Lancet Neurol. 2010;9:167-176.
2. Manning L, Hirakawa Y, Arima H, et al. Blood pressure variability and
outcome after acute intracerebral haemorrhage: a post-hoc analysis of
INTERACT2, a randomised controlled trial. Lancet Neurol. 2014;13:
364-373.
3. Anderson CS, Selim MH, Molina CA, Qureshi AI. Intensive blood pres-
sure lowering in Intracerebral hemorrhage. Stroke. 2017;48:
2034-2037.
4. Panerai RB. Assessment of cerebral pressure autoregulation in
humans--a review of measurement methods. Physiol Meas. 1998;19:
305-338.
5. Diedler J, Karpel-Massler G, Sykora M, et al. Autoregulation and brain
metabolism in the perihematomal region of spontaneous intracerebral
hemorrhage: an observational pilot study. J Neurol Sci. 2010;295:
16-22.
6. Anderson CS, Heeley E, Huang Y, et al. Rapid blood-pressure lowering
in patients with acute intracerebral hemorrhage. N Engl J Med. 2013;
368:2355-2365.
7. Qureshi AI, Palesch YY, Suarez JI. Intensive blood-pressure lowering in
cerebral hemorrhage. N Engl J Med. 2016;375:e48.
8. Aries MJ, Elting JW, De Keyser J, Kremer BP, Vroomen PC. Cerebral
autoregulation in stroke: a review of transcranial Doppler studies.
Stroke. 2010;41:2697-2704.
9. Xiong L, Tian G, Lin W, et al. Is dynamic cerebral autoregulation bilat-
erally impaired after unilateral acute ischemic stroke? J Stroke Cerebro-
vasc Dis. 2017;26:1081-1087.
MINHAS ET AL.
10. Claassen JA, Meel-van den Abeelen AS, Simpson DM, Panerai RB,
international Cerebral Autoregulation ResearchNetwork (CARNet).
Transfer function analysis of dynamic cerebral autoregulation: a white
paper from the international cerebral autoregulation research net-
work. J Cereb Blood Flow Metab. 2016;36:665-680.
11. Stroup DF, Berlin JA, Morton SC, et al. Meta-analysis of observational
studies in epidemiology: a proposal for reporting. Meta-analysis of
observational studies in epidemiology (MOOSE) group. JAMA. 2000;
283:2008-2012.
12. Powers WJ, Zazulia AR, Videen TO, et al. Autoregulation of cerebral
blood flow surrounding acute (6 to 22 hours) intracerebral hemor-
rhage. Neurology. 2001;57:18-24.
13. Oeinck M, Neunhoeffer F, Buttler KJ, et al. Dynamic cerebral autoregula-
tion in acute intracerebral hemorrhage. Stroke. 2013;44:2722-2728.
14. Reinhard M, Neunhoeffer F, Gerds TA, et al. Secondary decline of
cerebral autoregulation is associated with worse outcome after intra-
cerebral hemorrhage. Intensive Care Med. 2010;36:264-271.
15. Kiphuth IC, Huttner HB, Breuer L, Engelhorn T, Schwab S,
Köhrmann M. Vasospasm in intracerebral hemorrhage with ventricular
involvement: a prospective pilot transcranial Doppler sonography
study. Cerebrovasc Dis. 2011;32:420-425.
16. Ma H, Guo ZN, Liu J, Xing Y, Zhao R, Yang Y. Temporal course of
dynamic cerebral autoregulation in patients with Intracerebral hemor-
rhage. Stroke. 2016;47:674.
17. Ma H, Guo ZN, Sun X, et al. Hematoma volume is a predictive factor
of disturbed autoregulation after spontaneous intracerebral hemor-
rhage. J Neurol Sci. 2017;382:96-100.
18. Nakagawa K, Serrador JM, LaRose SL, Sorond FA. Dynamic cerebral
autoregulation after intracerebral hemorrhage: a case-control study.
BMC Neurol. 2011;11:108.
21
19. Marti-Fabregas J, Belvis R, Guardia E, Cocho D, Martí-Vilalta JL. Rela-
tionship between transcranial Doppler and CT data in acute intracere-
bral hemorrhage. AJNR Am J Neuroradiol. 2005;26:113.
20. Lee Y, Tang S, Jeng J, Shieh J. Nonlinear analyses applied in cerebral
autoregulation and blood flow changes in patients with acute intrace-
rebral hemorrhage. Biomed Signal Process Control. 2017;31:102-107.
21. Castro P, Azevedo E, Serrador J, Rocha I, Sorond F. Hemorrhagic
transformation and cerebral edema in acute ischemic stroke: link to
cerebral autoregulation. J Neurol Sci. 2017;372:256-261.
22. Salinet AS, Robinson TG, Panerai RB. Effects of cerebral ischemia on
human neurovascular coupling, CO2 reactivity, and dynamic cerebral
autoregulation. J Appl Physiol. 2015;118:170-177.
23. Mayer SA, Thomas CE, Diamond BE. Asymmetry of intracranial hemo-
dynamics as an indicator of mass effect in acute intracerebral hemor-
rhage. A transcranial Doppler study. Stroke. 1996;27:1788-1792.
SUPPOR TI NG I NFORMATION
Additional supporting information may be found online in the Sup-
porting Information section at the end of the article.
How to cite this article: Minhas JS, Panerai RB, Ghaly G,
Divall P, Robinson TG. Cerebral autoregulation in hemorrhagic
stroke: A systematic review and meta-analysis of transcranial
Doppler ultrasonography studies. J Clin Ultrasound. 2019;47:
14–21. https://doi.org/10.1002/jcu.22645