Clinical Neurophysiology 132 (2021) 946–952

Contents lists available at ScienceDirect

Clinical Neurophysiology
journal homepage: www.elsevier.com/locate/clinph

Monitoring of patients with brainstem hemorrhage: A simultaneous

study of quantitative electroencephalography and transcranial Doppler
Ying Chen, Lijuan Wang, Jie Zhang, Sibo Wang, Yajie Qi, Jie Cao, Yingqi Xing ⇑
Department of Neurology, The First Hospital of Jilin University, Changchun, Jilin 130021, China

a r t i c l e i n f o h i g h l i g h t s

Article history:
 Quantitative electroencephalography (QEEG) predicts 90-day mortality in patients with acute severe
Accepted 9 December 2020
Available online 05 February 2021
brainstem hemorrhage (ASBH).
 Transcranial Doppler did not provide prognostic information for patients with ASBH.
 The simultaneous acquisition of EEG and TCD conferred no benefit for mortality prognosis in patients
Keywords:
Electroencephalography with ASBH.
Transcranial Doppler
Brain stem hemorrhage
Neurophysiological monitoring a b s t r a c t
Prognosis
Objective: To explore whether quantitative electroencephalography (QEEG) and transcranial Doppler
(TCD) can be used to evaluate patients with acute severe brainstem hemorrhage (ASBH).
Methods: We prospectively enrolled patients with ASBH and assessed their mortality at the 90-day
follow-up. The patients’ demographic data, serological data, and clinical factors were recorded.
Quantitative brain function monitoring was performed using a TCD-QEEG recording system attached
to the patient’s bedside.
Results: Thirty-one patients (55.3 ± 10.6 years; 17 men) were studied. Mortality at 90 days was at 61.3%.
There was no significant difference in TCD-related parameters between the survival group and the death
group (p > 0.05). Among the QEEG-related indexes, only the (delta + theta)/(alpha + beta) ratio (DTABR)
(odds ratio 11.555, 95%confidence interval 1.413–94.503, p = 0.022) was an independent predictor of
clinical outcome; the area under the ROC curve of DTABR was 0.921, cut-off point was 3.88, sensitivity
was 79%, and specificity was 100%.
Conclusions: In patients with ASBH, QEEG can effectively inform the clinical prognosis regarding 90-day
mortality, while TCD cannot.
Significance: QEEG shows promise for informing the mortality prognosis of patients with ASBH.
Ó 2021 Published by Elsevier B.V. on behalf of International Federation of Clinical Neurophysiology.

1. Introduction all-cause mortality rate for BH is roughly 50% (Behrouz, 2018;

Del Brutto and Campos, 2004; Van Asch et al., 2010). Studies of
Brainstem hemorrhage (BH) is a subtype of spontaneous BH with a large sample size are rare due to its low incidence. Fur-
intracerebral hemorrhage (ICH) with poor prognosis; the overall thermore, patients with severe BH who are in critical condition are
unable to move; therefore, brain function monitoring and progno-
sis evaluation can be challenging in these patients (Rohde et al.,
Abbreviations: QEEG, quantitative electroencephalography; TCD, Transcranial
2007; Tao et al., 2016). Invasive brainstem monitoring is difficult
Doppler; BH, brainstem hemorrhage; ASBH, acute severe brainstem hemorrhage;
CBF, Cerebral blood flow; ICH, Intracerebral hemorrhage; INR, International and can lead to nervous system complications due to patient posi-
Normalized Ratio; GCS, Glasgow Coma Scale; PI, Pulsatility index; BSI, Brain tioning and the complexity of the posterior fossa’s structure such
symmetry index; DAR, Delta/alpha ratio; DTABR, (Delta + theta)/(alpha + beta)ratio; as its narrow space (Vanaclocha et al., 2017). A noninvasive, conve-
RAP, Relative alpha power; RBP, Relative beta power; RDP, Relative delta power; nient, and real-time evaluation method of bedside prognosis is
RTP, Relative theta power.
⇑ Corresponding author. Fax: +86 431 88782378. necessary. Accurate prognosis prediction is the key to transmitting
E-mail addresses: 187080119@163.com (Y. Chen), wangljdr@163.com (L. Wang), accurate information to family members and establishing reason-
ZhangJ_0106@163.com (J. Zhang), wangsibo92@163.com (S. Wang), qi_yaj@163. able treatment methods.
com (Y. Qi), caojie_lily@sina.com (J. Cao), xingyq2009@sina.com (Y. Xing).

https://doi.org/10.1016/j.clinph.2020.12.026
1388-2457/Ó 2021 Published by Elsevier B.V. on behalf of International Federation of Clinical Neurophysiology.
Y. Chen, L. Wang, J. Zhang et al. Clinical Neurophysiology 132 (2021) 946–952

Transcranial Doppler (TCD) is one type of bedside brain- there were numerous clinical variations in the posterior cerebral
function monitoring technology that can dynamically monitor artery. Therefore, the middle cerebral artery was selected as the
changes in cerebral hemodynamics and noninvasively evaluate monitoring artery in this study, and the monitoring depth was
the level of intracranial pressure (Cardim et al., 2017; Hassler 50–60 mm.
et al., 1988; Rasulo et al., 2008). In addition, TCD can be used to The simultaneous EEG monitoring electrodes were placed
evaluate hemodynamic changes during posterior-fossa surgery according to the international 10–20 system, using 16 channels
and to assess posterior-fossa circulatory arrest (Hüttemann et al., (Fp1, Fp2, F3, F4, F7, F8, C3, C4, T3, T4, T5, T6, P3, P4, O1, O2) with
2000; Sharma et al., 2010). Therefore, we attempted to use TCD the frontal central zero (Fz) electrode pole used for grounding and
to monitor changes in acute severe brainstem hemorrhage (ASBH) Cz, A1, and A2 used as references (Jiang et al., 2019). The back-
cerebral blood flow to determine whether it can provide favorable ground EEG signal acquisition parameters were set as follows: time
information for prognostic evaluation. baseline, 30 mm/s; sensitivity, 10 lv/mm; impedance maintained
Electroencephalography (EEG) can reflect the changes in the below 10 KX; sampling frequency, 250 Hz; data filtering (high
electrical activity of neurons in real time. It is generally accepted pass, 0.3 Hz; low pass, 30 Hz).
that EEG can only reflect the changes in perfusion and metabolism The specific monitoring time took place at 9:00 a.m. or 2:00 p.
of supratentorial brain tissue after ischemia but is not sensitive to m. on the 2nd day after admission. Each patient was monitored
the changes in infratentorial brain tissue after ischemia. Single- only once. After the acquisition of a clear and stable signal, record-
photon emission computed tomography (SPECT) and magnetic res- ing was obtained for longer than 30 min, and the data were stored
onance imaging (MRI) studies on brainstem and cerebellar infarc- for further analysis.
tion have shown that infratentorial lesions could lead to changes
in cerebral blood flow (CBF) and metabolism of the contralateral 2.3. Clinical data and data analysis
supratentorial distal brain tissue due to damage of the main con-
necting nerve fibers (Fazekas et al., 1993; Rousseaux and All patients received routine monitoring of vital signs and
Steinling, 1999). Sheorajpanday et al. (2011) found for the first intensive nursing care in the Neurological Intensive Care Unit.
time that QEEG changes could provide early detection of ischemia We recorded demographic data, stroke risk factors, admission
due to posterior circulation cerebral infarction. The above findings GCS score, time from ICH onset until the time of monitoring,
suggest that QEEG can reflect the degree of ischemic necrosis of the admission laboratory tests, and hematoma volume (for the calcula-
infratentorial brain tissue. tion of hematoma volume, please refer to previous studies) (Chen
Our previous studies have shown that the combination of QEEG et al., 2018; Huttner et al., 2006), and the clinical outcome was
and TCD is informative for outcome prognoses in supratentorial assessed using the five-Point Glasgow Outcome Scale score 90 days
acute severe cerebral hemorrhage. In this study, we also attempted after ictus.
to simultaneously use QEEG and TCD to monitor brain function Blinded analysis was performed for each patient. All clearly
changes in patients with ASBH to determine whether they can help readable TCD waveforms were used in the calculations. Analytical
predict the outcome at the 90-day follow-up. variables included systolic velocity (Vs), diastolic velocity (Vd),
mean velocity (Vm) and pulsatility index (PI). Vm was calculated
using the following equation: Vm = (Vs - Vd) / 3 + Vd and PI was
2. Patients and methods
calculated using the following equation: PI = (Vs - Vd) / Vm.
Offline quantitative EEG analysis was performed using MATLAB
2.1. Study population and design
(MathWorks, Natick, MA, USA). EEG artifact interpretation was
performed independently by two physicians qualified in EEG inter-
In this prospective cohort study, we enrolled consecutive coma-
pretation. Artifact-free EEG segments in 5 min were analyzed, and
tose patients with ASBH from the intensive care unit of the Depart-
the spectral power was calculated using fast Fourier transform for
ment of Neurology, First Hospital of Jilin University, from June
each electrode. Power spectral density was calculated using
2017 to June 2019. The patients’ immediate family members
Welch’s averaged, modified periodogram spectral estimation
signed written informed consent before the beginning of the study.
method with a 2-s Hamming window and 50% overlap; the fre-
The study was approved by the Ethics Committee of the First
quency resolution was 0.5 Hz. From the resulting power spectra,
Hospital of Jilin University, China. We included patients with (1)
the absolute band power was summed across the delta (1–3 Hz),
first spontaneous BH; (2) admission time  72 h after onset; (3)
theta (4–7 Hz, RTP), alpha (8–13 Hz), and beta (14–30 Hz) bands
Glasgow Coma Scale (GCS) score of  8 on admission. Of these
(inclusive). Relative band power values for each resulting fre-
patients, we excluded those with (1) supratentorial ICH; (2) BH
quency band were computed as the ratio of the summed absolute
secondary to traumatic brain injury, cerebral infarction, vascular
band power to the total summed power across the 1–30 Hz range;
malformation, and venous thrombosis; (3) deficient temporal
the derived relative delta power (RDP), relative theta power (RTP),
acoustic bone window; (4) intracranial and extracranial major vas-
relative alpha power (RAP), relative beta power (RBP), delta/alpha
cular stenosis/occlusion; (5) previous ischemic or hemorrhagic
ratio (DAR), and (delta + theta)/(alpha + beta) ratio (DTABR) were
cerebrovascular disease; (6) pathological state affecting EEG activ-
computed as the ratio of absolute power for the respective fre-
ity and use of drugs affecting the central nervous system.
quency bands of interest. The brain symmetry index (BSI) was cal-
culated using a method delineated in previous studies (Chen et al.,
2.2. Monitoring 2018; Van Putten and Tavy, 2004).

TCD-QEEG synchronous joint monitoring was performed with a 2.4. Statistical analyses
bedside neuromonitor (NSD-8100; Delica, Shenzhen, China). Dur-
ing the monitoring period, the patient was in the supine position All statistical analyses were performed using SPSS version 17.0
and the head of the bed was raised by 15–30°. (SPSS Inc., Chicago, IL, USA) and MedCalc version 19.0.7 (MedCalc
TCD was performed using 2-MHz pulsed-wave Doppler probes Software; Mariakerke, Belgium). Normally distributed variables
fixed to each temporal window with a helmet. Preliminary exper- are described by mean and SD, and non-normally distributed vari-
iments showed that the changes in the middle cerebral artery and ables are described by median and interquartile range (IQR). Cate-
posterior cerebral artery in BH were essentially the same; however, gorical variables are presented as percentages. Student’s t-tests
947
Y. Chen, L. Wang, J. Zhang et al. Clinical Neurophysiology 132 (2021) 946–952

and median two-sample tests were used for normally distributed 3.2. TCD and QEEG parameters
variables. Nonparametric Wilcoxon (Kruskal–Wallis) analysis of
variance was used for non-normally distributed variables. The Fig. 1 shows the CT and TCD and QEEG findings of representa-
comparison of categorical variables was performed with the chi- tive patients with BH and normal controls. There was no significant
squared test. The correlation coefficients between variables and difference in TCD-related indicators among survivors and non-
90-day mortality were calculated using Spearman’s rank-order survivors, including the Vs, Vd, Vm, and PI (p > 0.05) (Table 2).
correlation analysis. The variables that showed a significant statis- QEEG-related indicators, compared between survivors and non-
tical difference (p  0.05) were selected from the above single fac- survivors, showed that higher RDP (p < 0.0001), lower RAP
tor analysis, and the independent predictors were screened by (p < 0.0001), lower RBP (p < 0.0001), higher DAR (p < 0.0001),
logistic regression analysis with 90-day mortality as the dependent and higher DTABR (p = 0.001) were significantly associated with
variable. The sensitivity, specificity, and area under the curve of mortality in patients with ASBH. (Table 2). There were no signifi-
different predictors were further analyzed by receiver operating cant differences in RTP and BSI between the two groups
characteristic (ROC) curves. The ROC curves were compared using (p > 0.05) (Table 2).
DeLong’s test. All tests performed were two-sided, and p < 0.05 was
considered statistically significant.
3.3. Correlation analyses

3. Results The outcomes of the correlation analyses are summarized in

Table 3. Higher GCS score (rho = 0.635, p < 0.0001), higher white
3.1. Baseline characteristics and clinical outcomes blood cell count (rho = 0.422, p = 0.018), and larger hematoma vol-
ume (rho = 0.638, p < 0.0001) in clinical variables were signifi-
Forty-six consecutive patients with severe BH were enrolled cantly correlated with prognosis. Regarding the QEEG variables,
over a 2-year period. Patients with the following were excluded: higher RDP (rho = 0.711, p < 0.0001), lower RAP (rho = 0.703,
secondary BH (n = 5), previous cerebrovascular diseases (n = 3), p < 0.0001), lower RBP (rho = 0.615, p < 0.0001), higher DAR
deficient temporal window (n = 3), signal artifacts (n = 2), and loss (rho = 0.696, p < 0.0001), and higher DTABR (rho = 0.711,
to follow-up (n = 2). Finally, we enrolled 31 patients, of which 19 p < 0.0001) were all significantly correlated with 90-day mortality.
(61.3%) died during the 90-day follow-up period. The median age No TCD variable was related to prognosis. Most of the results of the
was 55.3 ± 10.6 years, the maximum age was 82 years, the mini- correlation analyses and t-tests were generally consistent.
mum age was 41 years, and 17 (54.8%) of the patients were male.
The first TCD-QEEG was performed at a mean of 30.0 ± 19.6 h after
the onset of symptoms. No statistically significant differences 3.4. Regression analysis and ROC curves
between survivors and non-survivors were noted for the clinical
baseline data, including age, sex, risk factors, blood pressure, serum All variables with p  0.05 in the t-tests and correlation analy-
glucose, serum potassium, calcium, platelet count, activated partial ses were entered into the logistic regression model with the mor-
thromboplastin time, international normalized ratio, hematoma tality at 90 days as the dependent variable. The results showed that
location, and the time between the first TCD-QEEG monitoring after adjusting for GCS score, hematoma volume, and other QEEG
after the onset of symptoms. Higher GCS score (p < 0.0001), higher variables, DTABR (OR, 11.555; 95%CI, 1.413–94.503; p = 0.022)
white blood cell count (p = 0.027), and larger hematoma volume remained an independent predictor of 90-day mortality in patients
(p < 0.0001) were significantly correlated with mortality (Table 1). with ASBH.

Table 1
Baseline characteristics.

Characteristics All patients (n = 31) Survivors (n = 12) Non-survivors (n = 19) p value

Demographics
Age, y, (SD) 55.3 (10.6) 54.5 (11.6) 55.8 (10.2) 0.74
Male, n, (%) 17 (54.8) 7 (58.3) 10 (52.6) 1.00
Risk factors, n, (%)
Hypertension 25 (80.6) 9 (75.0) 16 (84.2) 0.653
Diabetes mellitus 2 (6.5) 1 (8.3) 1 (5.2) 1.00
Hyperlipidemia 13 (41.9) 8 (66.7) 5 (26.3) 0.06
Coronary heart disease 4 (12.9) 3 (25.0) 1 (5.3) 0.27
Smoking 15 (48.3) 6 (50.0) 9 (47.3) 1.00
Excessive drinking 8 (25.8) 5 (41.7) 3 (15.8) 0.21
Time from ICH onset to monitor, h, (SD) 30.0 (19.6) 31.1 (24.4) 29.3 (16.5) 0.81
GCS score, (SD) 5.2 (1.7) 6.6 (1.2) 4.4 (1.5) 0.000
SBP (mmHg), (SD) 168.0 (34.0) 174.4 (30.6) 164.5 (36.2) 0.42
DBP (mmHg), (SD) 86.3 (14.0) 89.8 (15.3) 84.1 (13.2) 0.29
WBC, 109/L, (SD) 12.4 (3.7) 10.6 (3.0) 13.6 (3.6) 0.027
Platelet, 109/L, (SD) 233.0 (94.1) 251.1 (57.4) 222.1 (111.3) 0.41
APTT, S, (SD) 31.3 (4.5) 31.2 (1.8) 31.4 (5.6) 0.82
INR, (IQR) 1.01 (0.91–1.00) 0.95 (0.90–0.99) 1.05 (0.92–1.01) 0.41
Glucose (mmol/L), (SD) 7.2 (2.3) 6.9 (2.9) 7.4 (1.9) 0.49
Potassium (mmol/L), (SD) 3.7 (0.4) 3.7 (0.3) 3.6 (0.4) 0.55
Calcium (mmol/L), (SD) 139.6 (6.4) 138.9 (3.8) 139.9 (7.7) 0.67
Sodium (mmol/L), (SD) 2.1 (0.1) 2.2 (0.1) 2.1(0.1) 0.32
Hematoma volume, cm3, (SD) 7.0 (3.6) 4.2 (1.8) 8.7 (3.4) 0.000
Intraventricular hemorrhage, n (%) 8 (25.8) 2 (16.7) 6 (31.6) 0.43

IQR, interquartile range; GCS, Glasgow Coma Scale; SBP, systolic blood pressure; DBP, diastolic blood pressure; WBC, white blood cell count; APTT, activated partial
thromboplastin time; INR, International Normalized Ratio.

948
Y. Chen, L. Wang, J. Zhang et al. Clinical Neurophysiology 132 (2021) 946–952

Fig. 1. Examples of representative patients. A is a non-survivor; QEEG shows that the slower delta frequency band significantly increased, and the faster alpha and beta
frequency bands significantly decreased. Moreover, DAR, DTABR, and BSI also increased. TCD shows that the velocity and the PI of the bilateral hemispheres did not
significantly change. B is a survivor; QEEG shows that the slower delta frequency band increased, and the faster alpha frequency band decreased; DAR and DTABR also
increased, and TCD was similar to that in patient A. C is a healthy control; QEEG and TCD appear normal. DAR, delta/alpha ratio; DTABR, (delta + theta)/(alpha + beta) ratio;
BSI, brain symmetry index; VS, systolic flow velocity; VM, mean flow velocity; VD, diastolic flow velocity; PI, pulsatility Index; TCD, transcranial Doppler; QEEG, quantitative
electroencephalography.

Table 2
Transcranial Doppler and quantitative electroencephalography parameters.

All patients (n = 31) Non-survivors (n = 19) Survivors (n = 12) p value

TCD Parameters
VS, cm/s, (SD) 96.2 (28.3) 98.9 (30.7) 91.8 (24.7) 0.505
VM, cm/s, (SD) 59.5 (19.2) 61.4(21.3) 56.8 (15.7) 0.522
VD, cm/s, (SD) 41.5 (15.3) 42.7 (16.9) 39.6 (12.7) 0.602
PI, (SD) 0.94 (0.17) 0.95 (0.16) 0.92 (0.18) 0.643
QEEG Parameters
RDP, %, (SD) 69.9 (9.0) 74.6 (4.9) 62.2 (8.8) 0.000
RTP, %, (SD) 8.4 (1.2) 8.3 (1.4) 8.5 (1.0) 0.677
RAP, %, (SD) 13.6 (6.2) 10.4 (2.6) 18.6 (6.9) 0.000
RBP, %, (SD) 8.2 (3.3) 6.7 (3.1) 10.6 (1.9) 0.000
DAR, (SD) 6.5 (3.2) 8.0 (3.0) 4.2 (1.7) 0.000
DTABR, (SD) 4.6 (2.7) 5.7 (2.8) 2.8 (0.9) 0.001
BSI, (SD) 0.39 (0.08) 0.41 (0.09) 0.36 (0.04) 0.077

TCD, transcranial Doppler; QEEG, quantitative electroencephalography; VS, systolic flow velocity; VM, mean flow velocity; VD, diastolic flow velocity; PI, pulsatility Index;
RDP, relative delta power; RTP, relative theta power; RAP, relative alpha power; RBP, relative beta power; DAR, delta/alpha ratio; DTABR, delta + theta)/(alpha + beta)ratio;
BSI, brain symmetry index.

We established ROC curve models for DTABR and clinical vari- (95%CI, 0.709–0.967), of the GCS score was 0.868 (95%CI, 0.699–
ables significantly related to 90-day mortality in the univariate 0.962), and of the white blood cell count was 0.750 (95%CI,
analysis; the first model was composed of DTABR, the second 0.563–0.887) (Fig. 2).
model was composed of hematoma volume, the third model was After the ROC curve comparison, it was found that DTABR,
composed of white blood cell count, and the fourth model was hematoma volume, GCS, and white blood cell count did not signif-
composed of the GCS score. All models predicted 90-day mortality icantly differ (all p > 0.05). For patients with ASBH, DTABR had sim-
in patients with ASBH. The area under the ROC curve of DTABR was ilar predictive performance with hematoma volume, GCS, and
0.921 (95%CI, 0.766–0.987), the DTABR cut-off point for the identi- white blood cell count, providing a good predictive effect on clini-
fication of 90-day mortality in patients with ASBH was 3.88, which cal prognosis. Hematoma volume, GCS, and white blood cell count
yielded a sensitivity of 79% and a specificity of 100% (Fig. 2). The were combined with DTABR to establish the joint prediction
area under the ROC curve of the hematoma volume was 0.877 model. The area under the curve after the combination remained

949
Y. Chen, L. Wang, J. Zhang et al. Clinical Neurophysiology 132 (2021) 946–952

Table 3 In previous studies, the prognosis of infratentorial ICH was

Spearman correlation between clinical variables, TCD parameters, QEEG parameters, mostly assessed by various scales based on clinical data. The GCS
and prognosis.
score, hematoma volume, and leukocyte count were the most com-
Spearman’s rho p value mon independent predictor of mortality in patients with brainstem
Clinical Variables and cerebellar hemorrhage (Huang et al., 2017; Pong et al., 2012;
GCS score 0.635 0.000 Takeuchi et al., 2013). In agreement with previous studies, our
WBC 0.422 0.018 results showed that the GCS score, hematoma volume, and white
Platelet count 0.267 0.147
Hematoma volume 0.638 0.000
blood cell count were correlated with 90-day mortality. However,
TCD Parameters when the above-mentioned clinical variables were entered into
VS 0.119 0.525 the multiple regression model, they were not independent predic-
VM 0.089 0.635 tors; the reason may be that the patients in this study were all sev-
VD 0.130 0.487
ere comatose patients with GCS  8; thus, the GCS score and
PI 0.104 0.579
QEEG Parameters hematoma volume no longer had obvious advantages in the prog-
RDP 0.711 0.000 nosis evaluation of patients with ASBH.
RTP 0.107 0.565 It is generally believed that EEG is more sensitive to supratento-
RAP 0.703 0.000 rial ischemia than to infratentorial ischemia. EEG primarily reflects
RBP 0.615 0.000
the activity of neurons in the cerebral cortex (Niedermeyer and
DAR 0.696 0.000
DTABR 0.711 0.000 Lopes da Silva, 2005; Nunez and Srinivasan, 2006), which is why
BSI 0.319 0.081 it may not be effectively or directly sensitive to brainstem activity.
For patients with supratentorial ischemia, previous studies have
TCD, transcranial Doppler; QEEG, quantitative electroencephalography; GCS, Glas-
gow Coma Scale; WBC, white blood cell count; VS, systolic flow velocity; VM, mean reported that QEEG can quantitatively reflect the changes in brain
flow velocity; VD, diastolic flow velocity; PI, pulsatility index; RDP, relative delta function and predict the prognosis (Cuspineda et al., 2007;
power; RTP, relative theta power; RAP, relative alpha power; RBP, relative beta Finnigan et al., 2007; Leon-Carrion et al., 2009), but changes in
power; DAR, delta/alpha ratio; DTABR, (delta + theta)/(alpha + beta)ratio; BSI, brain
QEEG after infratentorial cerebral hemorrhage have not been
symmetry index.
reported. QEEG can analogize the function of SPECT and PET func-
tional imaging to a certain extent, quantitatively reflecting CBF cou-
pled with the cerebral metabolic rate for oxygen (Nagata et al.,
1989; Zazulia et al., 2001). Previous SPECT studies have confirmed
that infratentorial cerebral infarction can lead to supratentorial
remote ischemic changes. Remote perfusion changes after pure
brainstem infarction using SPECT may be seen both infratentorially
and supratentorially due to damage to the corticopontocerebellar
tract, depending on the lesion site rather than on the neurological
deficit (Fazekas et al. 1993). HPMAO-SPECT and MRI studies further
showed that besides corticopontocerebellar tract damage, the
cerebellar-thalamic-cortical tract may also be involved
(Rousseaux and Steinling, 1999). Brainstem and cerebellar infarc-
tion may affect regional CBF and metabolic activity in the cerebel-
lum and cerebral hemispheres, even leading to severe cerebral
lesions (Rousseaux and Steinling, 1999), showing that brainstem
and cerebellum infarction can lead to supratentorial cerebral ische-
mia caused by injury of the connecting nerve fibers. We speculate
that brainstem hemorrhage may cause similar changes.
Severe impairment of the ascending activation system of the
brainstem reticular formation in patients with ASBH leads to coma.
The anatomical basis of the EEG response depends on the integrity
of the brainstem reticular system and the integration of the thala-
Fig. 2. Receiver operating characteristic (ROC) curves to predict outcome in this mus and cortex (Procaccio et al., 2001). Data obtained from animal
cohort of four models. DTABR (AUROC 0.921 [0.766 to 0.987]); Hematoma Volume studies suggest that near-normal EEG activity (alpha, beta-, and
(AUROC 0.877 [0.709 to 0.967]); GCS (AUROC 0.868 [0.699 to 0.962]); WBC (AUROC theta) is unlikely to occur if significant amounts of tegmental retic-
0.750 [0.563 to 0.887]). DTABR is the independent predictors of QEEG. AUROC, area
ular neurons at the pontine mesencephalic junction are destroyed
under the receiver operating curve; GCS, Glasgow Coma Scale; WBC, white blood
cell count; QEEG, quantitative electroencephalography. (Walter et al., 2018). In a well-documented case of brainstem
hematoma with unilateral involvement of the mesencephalic
tegmentum, EEG acquired 7 h after coma onset showed a slow
0.921, which was the same as that in the DTABR single model. The
alpha activity, while the EEG at the 7th day was dominated by gen-
combined model of clinical factors and DTABR did not increase the
eralized delta activity (Steudel et al., 1979). It can be inferred that
prediction efficiency.
reactive inflammation/edema near the hematoma also affects the
function of the contralateral mesencephalic tegmentum, resulting
4. Discussion in the slowing of EEG activity over time.
In this study, the results of QEEG monitoring in patients with
In this study, QEEG and TCD acquired simultaneously were ASBH showed that, similar to supratentorial cerebral ischemia,
firstly analyzed in relation to their abilities to predict 90-day mor- the slower-frequency delta band power significantly increased,
tality in patients with ASBH. The QEEG (delta + theta)/(alpha + be while the faster-frequency alpha and beta band powers signifi-
ta) power ratio (DTABR) was found to be an independent predictor cantly decreased with the decrease of cerebral blood flow. RDP,
of mortality, with an area under the curve of 0.921. In contrast, no RAP, RBP, DAR, and DTABR were significantly correlated with 90-
TCD parameters were found to be independent predictors. day mortality. Multiple logistic regression analysis showed that
950
Y. Chen, L. Wang, J. Zhang et al. Clinical Neurophysiology 132 (2021) 946–952

after adjusting for GCS score, hematoma volume, and other QEEG ysis and interpretation. Yajie Qi, Jie Cao contributed to analysis and
variables, DTABR remained an independent predictor, and its area interpretation of the data and revision of the manuscript. All
under the ROC curve was 0.921. Sheorajpanday et al. (2011) used authors gave final approval of the version to be published.
QEEG to monitor posterior circulation infarction and found that a
pairwise derived BSI (pdBSI) > 0.24 was 100% sensitive for the pres- Funding
ence of a recent ischemic lesion, but all QEEG parameters including
DTABR could not predict short-term prognosis 7 days after onset. This work was supported by the National Natural Science Foun-
The difference between this study’s and Sheorajpanday’s conclu- dation of China [grant numbers 81971620]; Natural Science Foun-
sion may be attributable to their patients’ mild symptoms and small dation of Jilin Science and Technology Department [grant number
volume of infratentorial infarction, or maybe, some of their patients 20200201522JC]; Youth Development Fund of The First Hospital
might have had posterior circulation infarction restricted to the of Jilin University [grant number JDYY92018011].
cerebral cortex (e.g., the occipital cortex), thus being supratentorial
and not infratentorial, and obscuring the distant effect. In our study,
Declaration of Competing Interest
all patients had acute severe coma with a large hematoma, an obvi-
ous ischemic injury of local brain tissue, and extensive destruction
The authors declare that they have no known competing finan-
of nerve fibers; therefore, the QEEG parameters were significantly
cial interests or personal relationships that could have appeared
correlated with prognosis, showing good predictive effectiveness.
to influence the work reported in this paper.
We calculated the area under the ROC curve after combining DTABR
with the GCS score, hematoma volume, and white blood cell count.
Acknowledgements
The results showed that the area under the ROC curve remained
0.921, suggesting that the ROC curve prediction model of DTABR
The authors thank Qiang Zhang and Ying Wang for providing
was stable; adding other variables did not increase the predictive
technical assistance with TCD-QEEG monitoring.
accuracy, over and above that of DTABR alone.
TCD can reflect changes in cerebral hemodynamics in real time
and reflect the increase in intracranial pressure to a certain extent. References
Our previous studies on supratentorial intracerebral hemorrhage Behrouz R. Prognostic factors in pontine haemorrhage: A systematic review. Eur
confirmed that increased PI in the unaffected hemisphere is an Stroke J 2018;3:101–9. https://doi.org/10.1177/2396987317752729.
independent prognostic factor (Chen et al., 2018). The changes in Cardim D, Schmidt B, Robba C, Donnelly J, Puppo C, Czosnyka M, et al. Transcranial
Doppler monitoring of intracranial pressure plateau waves. Neurocrit Care
intracranial pressure under the tentorium are complex, and there
2017;26:330–8. https://doi.org/10.1007/s12028-016-0356-5.
may be gradient changes (Slavin and Misra, 2003). In this study, Chen Y, Xu W, Wang L, Yin X, Cao J, Deng F, et al. Transcranial Doppler combined
our results showed that for patients with ASBH, bilateral middle with quantitative EEG brain function monitoring and outcome prediction in
cerebral artery velocity and pulsatility index (PI) remained patients with severe acute intracerebral hemorrhage. Crit Care 2018;22:36.
https://doi.org/10.1186/s13054-018-1951-y.
unchanged; there was no difference between the survival group Cuspineda E, Machado C, Galán L, Aubert E, Alvarez MA, Llopis F, et al. QEEG
and deceased group; therefore, these TCD variables were not useful prognostic value in acute stroke. Clin EEG Neurosci 2007;38:155–60. https://
for mortality prognosis. doi.org/10.1177/155005940703800312.
Del Brutto OH, Campos X. Validation of intracerebral hemorrhage scores for patients
We previously used QEEG combined with TCD to predict the with pontine hemorrhage. Neurology 2004;62:515–6. https://doi.org/10.1212/
outcome of supratentorial intracerebral hemorrhage, and it WNL.62.3.515.
showed the best prediction effect. However, in this study, QEEG Fazekas F, Payer F, Valetitsch H, Schmidt R, Flooh E. Brain stem infarction and
diaschisis. A SPECT cerebral perfusion study. Stroke 1993;24:1162–6. https://
and TCD did not show similar results for ASBH prognosis because doi.org/10.1161/01.STR.24.8.1162.
the TCD variables could not provide prognostic information. There Finnigan SP, Walsh M, Rose SE, Chalk JB. Quantitative EEG indices of sub-acute
were some limitations in this study. First, the study was conducted ischaemic stroke correlate with clinical outcomes. Clin Neurophysiol
2007;118:2525–32. https://doi.org/10.1016/j.clinph.2007.07.021.
at the neurology intensive care unit of the Affiliated Hospital of the Jiang M, Su Y, Liu G, Chen W, Gao D. Predicting the non-survival outcome of large
university, which comprises a single center. In addition, the inci- hemispheric infarction patients via quantitative electroencephalography:
dence of BH was low, and the sample size was relatively small; Superiority to visual electroencephalography and the Glasgow Coma Scale.
Neurosci Lett 2019;27:88–92. https://doi.org/10.1016/j.neulet.2019.05.007.
the large dispersion of each patient’s results led to a wide confi-
Hassler W, Steinmetz H, Gawlowski J. Transcranial Doppler ultrasonography in
dence interval for DTABR. Furthermore, we included patients raised intracranial pressure and in intracranial circulatory arrest. J Neurosurg
within 72 h from onset, and no ultra-early and dynamic monitor- 1988;68:745–51. https://doi.org/10.3171/jns.1988.68.5.0745.
ing was performed. Huang K, Ji Z, Sun L, Gao X, Lin S, Liu T, et al. Development and validation of a
grading scale for primary pontine hemorrhage. Stroke 2017;48:63–9. https://
doi.org/10.1161/STROKEAHA.116.015326.
Hüttemann E, Schelenz C, Sakka SG, Reinhart K. Atropine test and circulatory arrest
5. Conclusion in the fossa posterior assessed by transcranial Doppler. Intensive Care Med
2000;26:422–5. https://doi.org/10.1007/s001340051176.
This was the first study to simultaneously acquire EEG and TCD Huttner HB, Steiner T, Hartmann M, Köhrmann M, Juettler E, Mueller S, et al.
Comparison of ABC/2 estimation technique to computer-assisted planimetric
and investigate the abilities of the resulting measures to inform the analysis in warfarin-related intracerebral parenchymal hemorrhage. Stroke
mortality prognosis in ASBH. A QEEG variable (DTABR) was identi- 2006;37:404–8. https://doi.org/10.1161/01.STR.0000198806.67472.5c.
fied as an independent predictor of 90-day mortality, whereas no Leon-Carrion J, Martin-Rodriguez JF, Damas-Lopez J, Barroso y Martin JM,
Dominguez-Morales MR. Delta-alpha ratio correlates with level of recovery
TCD measures were identified as predictors. Additional variables, after neurorehabilitation in patients with acquired brain injury. Clin
such as the GCS score, did not improve the predictive accuracy of Neurophysiol 2009;120:1039–45. https://doi.org/10.1016/j.clinph.2009.01.021.
DTABR. Nagata K, Tagawa K, Hiroi S, Shishido F, Uemura K. Electroencephalographic
correlates of blood flow and oxygen metabolism provided by positron emission
tomography in patients with cerebral infarction. Electroencephalogr Clin
Neurophysiol 1989;72:16–30. https://doi.org/10.1016/0013-4694(89)90027-8.
Author contributions
Niedermeyer E, Lopes da Silva F. Electroencephalography: basic principles, clinical
applications, and related fields. Philadelphia: Lippincott Williams & Wilkins; 2005.
Ying Chen and Yingqi Xing contributed to study conception and Nunez PL, Srinivasan R. Electric fields of the brain: the neurophysics of EEG. New
design, data collection, analysis and interpretation, and drafting of York: Oxford University Press; 2006.
Pong V, Chan KH, Chong BH, Lui WM, Leung GK, Tse HF, et al. Long-term outcome
the manuscript. Lijuan Wang, Jie Zhang, and Sibo Wang con- and prognostic factors after spontaneous cerebellar hemorrhage. Cerebellum
tributed to study conception and design, data collection, and anal- 2012;11:939–45. https://doi.org/10.1007/s12311-012-0371-9.

951
Y. Chen, L. Wang, J. Zhang et al. Clinical Neurophysiology 132 (2021) 946–952

Procaccio F, Polo A, Lanteri P, Sala F. Electrophysiologic monitoring in Takeuchi S, Suzuki G, Takasato Y, Masaoka H, Hayakawa T, Otani N, et al. Prognostic
neurointensive Care. Curr Opin Crit Care 2001;7:74–80. https://doi.org/ factors in patients with primary brainstem hemorrhage. Clin Neurol Neurosurg
10.1097/00075198-200104000-00004. 2013;115:732–5. https://doi.org/10.1016/j.clineuro.2012.08.022.
Rasulo FA, De Peri E, Lavinio A. Transcranial Doppler ultrasonography in intensive Tao C, Zhang R, Hu X, Song L, Wang C, Gao F, et al. A novel brainstem hemorrhage
care. Eur J Anaesthesiol Suppl 2008;42:167–73. https://doi.org/10.1017/ model by autologous blood infusion in rat: white matter injury, magnetic
S0265021507003341. resonance imaging, and neurobehavioral features. J Stroke Cerebrovasc Dis
Rohde V, Berns E, Rohde I, Gilsbach JM, Ryang YM. Experiences in the management 2016;25:1102–9. https://doi.org/10.1016/j.jstrokecerebrovasdis.2016.01.025.
of brainstem hematomas. Neurosurg Rev 2007;30:219–23. https://doi.org/ Vanaclocha V, Sáiz-Sapena N, Rivera-Paz M, Herrera JM, Ortiz-Criado JM, Verdu-
10.1007/s10143-007-0081-9. López F, et al. Can we safely monitor posterior fossa intracranial pressure? A
Rousseaux M, Steinling M. Remote regional cerebral blood flow consequences of cadaveric study. Br J Neurosurg 2017;31:557–63. https://doi.org/10.1080/
focused infarcts of the medulla, pons and cerebellum. J Nucl Med 02688697.2017.1332336.
1999;40:721–9. Van Asch CJ, Luitse MJ, Rinkel GJ, Van der Tweel I, Algra A, Klijn CJ. Incidence, case
Sharma D, Ellenbogen RG, Vavilala MS. Use of transcranial Doppler ultrasonography fatality, and functional outcome of intracerebral haemorrhage over time,
and jugular oximetry to optimize hemodynamics during pediatric posterior according to age, sex, and ethnic origin: a systematic review and meta-
fossa craniotomy. J Clin Neurosci 2010;17:1583–4. https://doi.org/10.1016/j. analysis. Lancet Neurol 2010;9:167–76. https://doi.org/10.1016/S1474-4422
jocn.2010.04.009. (09)70340-0.
Sheorajpanday RV, Nagels G, Weeren AJ, De Deyn PP. Quantitative EEG in ischemic Van Putten MJ, Tavy DL. Continuous quantitative EEG monitoring in hemispheric
stroke: correlation with infarct volume and functional status in posterior stroke patients using the brain symmetry index. Stroke 2004;35:2489–92.
circulation and lacunar syndromes. Clin Neurophysiol 2011;122:884–90. https://doi.org/10.1161/01.STR.0000144649.49861.1d.
https://doi.org/10.1016/j.clinph.2010.08.020. Walter U, Fernández-Torre JL, Kirschstein T, Laureys S. When is ‘‘brainstem death”
Slavin KV, Misra M. Infratentorial intracranial pressure monitoring in neurosurgical brain death? The case for ancillary testing in primary infratentorial brain lesion.
intensive care unit. Neurol Res 2003;25:880–4. https://doi.org/10.1179/ Clin Neurophysiol 2018;129:2451–65. https://doi.org/10.1016/
016164103771954014. j.clinph.2018.08.009.
Steudel WI, Krüger J, Grau H. Zur Alpha- und Spindel-Aktivität bei komatösen Zazulia AR, Diringer MN, Videen TO, Adams RE, Yundt K, Aiyagari V, et al.
Patientennach einer Schädel-Hirn-Verletzung unter besonderer Hypoperfusion without ischemia surrounding acute intracerebral hemorrhage. J
Berücksichtigung der Computertomographie. EEG EMG Z Elektroenzephalogr Cereb Blood Flow Metab 2001;21:804–10. https://doi.org/10.1097/00004647-
Elektromyogr Verwandte Geb 1979;10:143–7. 200107000-00005.

952