ArticlePDF Available

Heart rate variability for assessing comatose patients with different Glasgow Coma Scale scores

Authors:
  • The National Institute of Neurology and Neurosurgery, Havana, Cuba
  • Institute of Neurology and Neurosurgery Havana Cuba

Abstract and Figures

Objective: To assess the autonomic nervous system (ANS) in coma by heart rate variability (HRV). Methods: Sixteen comatose patients and 22 normal subjects with comparable ages and genders were studied. Patients were classified in two subgroups according to the Glasgow Coma Scale (GCS). Time, frequency, and informational HRV domain indices were calculated. Results: A notable reduction of HRV was found in patients. Regarding the time domain indices, the triangular index, and the Delta_RRs, were significantly reduced in the subgroup with GCS=3. Absolute power for the whole frequency spectrum decreased whenever GCS scores were lower. A significant decrement was found for absolute power of the VLF and LF bands in the subgroup of GCS=3, and although it was lower for the HF band in these patients, those changes were not statistically significantly different. The LF/HF ratio and the Shannon´s entropy indices were significantly reduced in the subgroup with GCS=3. Our results are discussed regarding the progressive dysfunction the ANS networks when coma deepens. Conclusions: The HRV procedure is a powerful tool to assess the ANS in comatose patients. Significance: HRV is a minimally invasive, low-cost methodology, suitable for assessing the ANS in coma.
Content may be subject to copyright.
Heart rate variability for assessing comatose patients with different Glasgow
Coma Scale scores
Yazmina Machado-Ferrer
a
, Mario Estévez
b
, Calixto Machado
b,
, Adrián Hernández-Cruz
a
,
Frederick R. Carrick
c
, Gerry Leisman
d,e,f
, Robert Melillo
d,e,f
, Phillip DeFina
g
, Mauricio Chinchilla
h
,
Yanín Machado
h
a
‘‘Manuel Fajardo Rivero’’ Hospital, Havana, Cuba
b
Institute of Neurology and Neurosurgery, Department of Clinical Neurophysiology, Havana, Cuba
c
F.R. Carrick Institute for Graduate Studies, Cape Canaveral, FL, USA
d
The National Institute for Brain and Rehabilitation Sciences, Nazareth, Israel
e
F.R. Carrick Institute for Clinical Ergonomics, Rehabilitation, and Applied Neurosciences (CERAN), NY, USA
f
Nazareth Academic Institute, Nazareth, Israel
g
International Brain Research Foundation, NY, USA
h
Hermanos Ameijeiras Hospital, Service of Neurology, Havana, Cuba
article info
Article history:
Accepted 3 September 2012
Available online xxxx
Keywords:
Heart rate variability
Coma
Comatose patients
Autonomic nervous system
highlights
We study comatose patients classified into two subgroups according to the GCS, and compared with a
normal control group, by calculating HRV indices in time, frequency and informational domains.
In general, a notable reduction of HRV variability was found in patients, compared to normal subjects,
and significant statistical differences between the subgroups of patients, classified according to GCS
scores, were found for HRV indices in all domains.
Our results are discussed regarding the progressive dysfunction the ANS networks when coma deepens.
abstract
Objective: To assess the autonomic nervous system (ANS) in coma by heart rate variability (HRV).
Methods: Sixteen comatose patients and 22 normal subjects with comparable ages and genders were
studied. Patients were classified in two subgroups according to the Glasgow Coma Scale (GCS). Time, fre-
quency, and informational HRV domain indices were calculated.
Results: A notable reduction of HRV was found in patients. Regarding the time domain indices, the trian-
gular index, and the Delta_RRs, were significantly reduced in the subgroup with GCS = 3. Absolute power
for the whole frequency spectrum decreased whenever GCS scores were lower. A significant decrement
was found for absolute power of the VLF and LF bands in the subgroup of GCS = 3, and although it was
lower for the HF band in these patients, those changes were not statistically significantly different. The
LF/HF ratio and the Shannon
´s entropy indices were significantly reduced in the subgroup with GCS = 3.
Our results are discussed regarding the progressive dysfunction the ANS networks when coma deepens.
Conclusions: The HRV procedure is a powerful tool to assess the ANS in comatose patients.
Significance: HRV is a minimally invasive, low-cost methodology, suitable for assessing the ANS in coma.
Ó2012 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights
reserved.
1. Introduction
Several papers have appeared in the last few decades on the
assessment of heart rate variability (HRV) in comatose patients
and in brain death (BD) (Biswas et al., 2000; Conci et al., 2001;
Cooke et al., 2006; Freitas et al., 1996; Gujjar et al., 2004; Kahr-
aman et al., 2010; Leipzig and Lowensohn, 1986; Li et al., 2003;
Machado et al., 2005; Mejia and Pollack, 1995; Rapenne et al.,
2000; Ryan et al., 2011; Shimomura et al., 1991; Su et al., 2005;
Vakilian et al., 2011). Schwarz et al. (1987) using a time domain
analysis found that the distribution of HRV and of heart rate
showed a discrimination of the group of comatose patients from
brain dead subjects and healthy subjects.
1388-2457/$36.00 Ó2012 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.clinph.2012.09.008
Corresponding author. Address: Institute of Neurology and Neurosurgery,
Department of Clinical Neurophysiology, 29 y D, Vedado, La Habana 10400, Cuba.
Tel./fax: +837-834 5578.
E-mail address: braind@infomed.sld.cu (C. Machado).
Clinical Neurophysiology xxx (2012) xxx–xxx
Contents lists available at SciVerse ScienceDirect
Clinical Neurophysiology
journal homepage: www.elsevier.com/locate/clinph
Please cite this article in press as: Machado-Ferrer Y et al. . Clin Neurophysiol (2012), http://dx.doi.org/10.1016/j.clinph.2012.09.008
Biswas et al. (2000) found that children with a Glasgow Coma
Scale (GCS) score of 3–4 had a lower low-frequency/high-fre-
quency (LF/HF) ratio compared with those who had a GCS score
of 5–8, and that patients who progressed to BD had a markedly
lower LF/HF ratio, with a significant decrease after the first 4 h of
hospitalisation. Other patients with more favourable outcomes
had significantly higher LF/HF ratios. These authors concluded that
the LF/HF ratio may be helpful not only in identifying those pa-
tients who will progress to BD but also in predicting which patients
will have favourable outcome.
Su et al. (2005) studied sympathetic and parasympathetic indi-
ces after severe head trauma, correlating HRV and GCS. However,
they erred in stating that there had been no previous analysis of
HRV in comatose patients grouped according to GCS (Machado
et al., 2005). These authors correlated the parameters derived from
spectral analysis of HRV with the GCS in five groups of patients
with brain damage of various severities. They reported that an in-
crease in severity of the brain-stem damage was accompanied with
an augmentation of a sympathetic and a decrement of parasympa-
thetic drives, as indicated by increasing of the LF band expressed in
normalised units (%) and the ratio LF/HF and decreasing the HF
band, respectively. Both LF and HF indices were nearly abolished
in BD.
In early 1990s, we ran a protocol assessing comatose patients
with different GCS scores, some progressing to BD (García et al.,
1995). Thirty-five comatose patients with acute stroke were stud-
ied serially. We found that GCS scores increased or decreased rang-
ing from 3 to 15, with either improvement or deterioration of
coma. Patients were divided into four groups (GCS from 10 to 15;
GCS from 7 to 9; GCS from 3 to 7; and BD). HRV was calculated
using the variation coefficient (VC) calculated as VC = (SD/
M) 100, where SD is the standard deviation of R–Rs in the sample
and M the mean value for the time series of R–R intervals. Forty
healthy subjects were also studied. We found that HRV did not pro-
gressively decrease when GCS scores diminished from 15 to 3 fol-
lowing a linear relationship, according to patients’ evolution. HRV
mean values and dispersion were higher in the group with GCS
score ranging from 7 to 9 (García et al., 1995; Machado et al.,
2005). A GCS score of about 8 has been related to a diencephalic le-
vel of consciousness impairment (Plum and Posner, 1980).
Nonetheless, in that study (García et al., 1995) we only analysed
HRV indices in the time domain, and hence in this article we study
comatose patients classified into two subgroups according to the
GCS scores, compared with healthy subjects, assessing the auto-
nomic nervous system (ANS) by calculating HRV indices in time,
frequency and informational domains.
2. Methods
2.1. Subjects
A total of consecutive 21 comatose patients admitted in the
intensive care unit of ‘‘Manuel Fajardo Rivero’’ Hospital (Havana)
were initially enrolled in this study. Nonetheless, five patients
were excluded according to the following criteria: patients with
diabetes mellitus; cardiac arrhythmia; electrocardiographic signs
of ischaemia; and antecedents of long-term use of drugs such as
hypnotics, autonomic stimulants or alpha- or beta-blockers that
may affect the autonomic system. It was not possible to elude
those cases in which the use of inotropics and vasoconstrictors,
to stabilise the arterial pressure was necessary. Hence, the final
sample included 16 patients, nine females (56.25%) and seven
males (43.75%), with an average age of 72.13 years and an SD of
10.94 years. Patients were also divided into two subgroups accord-
ing to the GCS score: subgroup with GCS = 3, and subgroup with
the GCS score from 4 to 8. A group of 22 healthy subjects, 12 fe-
males (54.5%) and 10 males (45.5%), with a mean age of 71.20
and an SD of 9.8 years, was considered the control group. Informed
written consent was obtained from patients’ relatives and normal
volunteers. The Ethical Committees of the Institute of Neurology
and Neurosurgery, and of the ‘‘Manuel Fajardo Rivero’’ Hospital,
approved the study.
All patients were neurologically examined and the GCS was ap-
plied to all cases. Only patients with a GCS score of 8 or less were
enrolled in this study. A computed tomography (CT) scan of the
brain was performed in all cases. Brain-dead patients were not in-
cluded in the final sample, according to the Cuban Criteria for BD
diagnosis (Machado, 2003, 2005, 2010; Machado et al., 2004).
2.2. ECG recording
Electrocardiography (ECG) was recorded with the MEDICID-05
with disposable electrodes placed on the chest in positions CM
2
and V
5
and using a sampling frequency of 200 Hz. Filters were
set for a band spectrum of 0.5–50 Hz. The ECG was recorded in
every session for 30 min.
2.3. ECG analysis
Bipolar ECG recordings were offline exported as ASCII files to a
software tool developed by our staff written in Delphi version 7.0
(MultiTools version 3.1.2, 2009–2012), for visual inspection, detec-
tion of the ‘R’ wave peaks and enabling manual editing. Accurate ‘R’
peak automatic detections obtained with the software’s algorithms
were always visually checked and properly corrected, when it was
necessary, by a member of our staff. To transform the ordinal
sequential R–R interval (RRI) series into proper temporal series,
the original RRI sequences were interpolated using the piecewise
cubic Hermite interpolation method obtaining a sampling period
of 205.078 ms. The entire consecutive sequences of RRI obtained
using this procedure were stored digitally for further instrumental
analysis.
2.4. Pre-processing of RRI sequences
The duration of the RRI temporal series used in this study was of
420 s. The recommended duration (Task Force, 1996) has been
300 s, but to study the very-low-frequency (VLF) spectral band, it
was necessary to increase this duration. We decided at last to
use 420 s considering that it allowed us measuring nine valid dis-
crete spectral frequencies (DFi) from the spectra, after discarding
the first 10 DFi, also a mandatory requisite stated by this Task
Force (1996). These nine DFi were a reasonable amount of values
for assessing a particular recently identified sub-band of the VLF
range, from 0.027344 to 0.03125, particularly sensitive to predict
the risk of occurrence of cardiac disturbances, such as ventricular
tachyarrhythmia (Bilgin et al., 2010).
The first 2 min in the RRI series were discarded, and the subse-
quent 420 s were submitted to a stability control test of the fre-
quency pattern of those segments, using a time–frequency
method, as will be later described. If the selected segment did
not show the necessary spectral stability, the commencement of
the sample was displaced for periods of 30 s, and was newly tested.
Only when the required stability was achieved, the segments were
accepted for subsequent processing. In the group of patients it was
not necessary to use alternative segments, but in the control group
it was needed in three subjects, in whom displacements were re-
quired to relocate the optimal segments from 30 to 90 s. Stability
was defined when power and frequency indices did not show
ostensible changes for at least 80% of the whole analysed time of
420 s.
2Y. Machado-Ferrer et al. / Clinical Neurophysiology xxx (2012) xxx–xxx
Please cite this article in press as: Machado-Ferrer Y et al. . Clin Neurophysiol (2012), http://dx.doi.org/10.1016/j.clinph.2012.09.008
The temporal series finally selected for the spectral analysis
were submitted to a pre-processing sequence of actions, as has
been suggested by several authors(de Souza Neto et al., 2007),
including (a) subtraction of the mean value from all the items in
the RRI series included in the window to diminish the effect of
the DC value and reduce the spectral zero-frequency; (b) applica-
tion of a median filter to replace outliers or abnormal values; (c)
standard linear detrending to avoid any possible drifts in the RRI
series; and (d) high-pass digital filtering (low cut-off frequency
0.02 Hz) using a sixth-order Butterworth infinite impulse response
filter, but using a zero-phase shifting algorithm to avoid the distor-
tion of original phase components.
2.5. Calculations of HRV indices
2.5.1. Time domain HRV indices
HRV indices calculated in the time domain were also calculated
as detailed elsewhere (Montes-Brown et al., 2010, 2012). Standard
recommended HRV indices were considered: the mean RRI value
(M), the SD and the square root of the mean squared differences
of successive RRI intervals. It was also calculated the mode (Mo)
for a bin of 5 ms, the amplitude of the mode (AMo), as the number
of RRI in the series with values equal to the mode expressed as per-
cent of the total number of RRI, the delta range (Delta_RRs), as the
difference between the maximal RRI value and the minimal RRI va-
lue observed in the series, the stress index and the triangular index
(Task Force, 1996).
2.5.2. Frequency domain HRV indices
Pre-processed temporal RRI series were submitted to spectral
analysis using the Welch periodogram method using a Hann win-
dow with 2048 samples and the fast Fourier transform (FFT) algo-
rithm. The spectral resolution was 1/420 = 0.00238095 Hz.
Selected processing parameters allowed the study of spectral fre-
quencies from 0.02 to 0.4 Hz, including the VLF band from 0.023
to 0.04 Hz, the LF band from 0.04 to 0.085 Hz, the mid-frequency
(MF) band from 0.085 to 0.15 Hz and the HF band from 0.15 to
0.40 Hz. Absolute and normalised (%) values were considered for
each spectral band. The ratio LF/HF was also calculated.
2.5.3. Informational domain HRV indices
Shannon’s entropy (ShaEn), an informational domain HRV in-
dex, following the new classification proposed by some authors
(Bravi et al., 2011), was also calculated for every RRI interpolated
temporal series. The classic expression used for developing the
algorithm was
H¼
X
N
i¼1
p
i
log
2
p
i
ð1Þ
where P
i
is the probability of every possible value of the duration of
an RRI, and Nis the total quantity of samples. If the events observed
would be all equiprobable, then the maximal ShaEn would be 11
(log
2
2048), and if all the RRIs would have the same value, ShaEn
should be 0 ((1) (2048/2048) (log
2
2048/2048) = 0) (Clark
et al., 2012; Takahashi et al., 2012; Viola et al., 2011).
2.5.4. HRV time-varying spectral analysis
To assess the stability of the spectral indices in every selected
window of the RRI time series of 420 s a time-varying spectral
method using one of the most used methods was carried out: the
short time Fourier transform (STFT) (Elsenbruch et al., 2000). It
can be considered as one of the most conventional and intuitive,
among the different methods used for time–frequency analysis of
temporal series (Mainardi, 2009; Martinmaki et al., 2006;
Martinmaki and Rusko, 2008; Seely and Macklem, 2004). A sliding
window of 256 s with displacement steps of 2 s was used, yielding
83 consecutive spectra, which confirm a compressed spectral ma-
trix that can be quantitatively analysed and graphically presented
using 3-D surface spectrograms. Only the segments which verified
stability over time were selected for the spectral analysis.
All digital signal data processing in this study was carried out
using standard custom-tailored programs developed by our staff
with MATLAB (MathWorks Inc. Version 7.10.0.499, R2010a).
2.6. Statistical analysis
Comparisons between the control group and the group of pa-
tients were achieved using the Mann–Whitney test. Differences
were considered statistically significant for p< 0.05. The Spearman
rank correlation test was used to analyse relationships between
different HRV variables and the GCS items. The package STATISTI-
CA, version 8.0 (Statsoft Inc. 2007), was used for statistical analysis.
3. Results
Brain injury aetiology was intracerebral haemorrhage in eight
patients, ischaemic cerebral infarct in four, post-anoxic encepha-
lopathy in two and metabolic encephalopathy in two cases. The
subgroup with GCS = 3 included six patients, and the subgroup
with GCS score ranging from 4 to 8, 10 cases (three cases with 4,
three with 6, one with 7 and three with GCS = 8).
In order to stabilise the arterial pressure in patients, dopamine
infusion (5.5
l
gkg
–1
min
–1
) was used in the six patients with
GCS = 3, and in two cases from the subgroup with GCS from 4 to
8 (two patients with GCS = 4).
Heart rate (beats per minute) was significantly higher in the pa-
tients, compared to the control group (Table 1). In our subgroup of
patients with GCS = 3, we found two cases with heart rate greater
than 110 beats/min, meanwhile the rest (four patients) showed a
heart rate lower than 65 beats/min. These four patients also pre-
sented arterial blood pressure less than 100/60 mmHg.
The results of statistical comparisons between patients and the
control group for HRV indices calculated for time, frequency and
informational domains are shown in Table 1. HRV showed an over-
all decrement in these cases, with reduced PSD values for VLF and
LF bands, and a significant decrement of the LH/HF ratio. Time-
domain indices of HRV, such as SD and Delta_RRs, expressing
global variability, and others showing short-time variability
(RMSSD, pNN50 and pMean2%), were also statistically significantly
lower in the group of patients. It was interesting to note that the
Shannon’s entropy index was significantly lower in patients. Abso-
lute PSD for all frequency bands were also significantly reduced in
the group of patients. Meanwhile, relative PSD only showed statis-
tical differences between control subjects and patients for P_MF
(significantly decreased in patients), and for P_HF (significantly
incremented in patients). The LF/HF ratio was significantly dimin-
ished in the group of patients.
In Table 2, the same HRV indices from Table 1 are shown, but
now comparing both subgroups of patients, classified according
to the GCS scores. Regarding the HRV indices in the time domain,
Delta_RRs, and the Triangular indices were significantly decreased
in the subgroup with GCS = 3. Overall, absolute power for the
whole frequency spectrum decreased whenever GCS scores were
lower. Nonetheless, significant decrements were only found for
absolute PSD of the VLF and LF bands in the subgroup with
GCS = 3. The LF/HF ratio was significantly lower in the subgroup
with GCS = 3. Concerning the normalised relative power indices,
nu_VLF and nu_MF were significantly lower, and the nu_HF was
significantly higher in the subgroup with GCS = 3. Shannon’s entro-
py index was significantly lower in the subgroup with GCS = 3.
Y. Machado-Ferrer et al. / Clinical Neurophysiology xxx (2012) xxx–xxx 3
Please cite this article in press as: Machado-Ferrer Y et al. . Clin Neurophysiol (2012), http://dx.doi.org/10.1016/j.clinph.2012.09.008
In Table 3, the Spearman correlation analysis between HRV
indices values and the GCS items (eyes, verbal, motor responses
and total score) in the group of patients is shown. For ‘eyes re-
sponse’ no significant correlations were found. Regarding verbal
response, a positive significant correlation was found for indices
RMSSD, and P_HF (classic parasympathetic indices). Concerning
the motor response, a positive significant correlation was found
for the TriI, Shannon’s entropy, P_VLF, and P_Tot (classic sympa-
thetic indices) and P_HF. The total GCS score showed a positive sig-
nificant correlation with P_VLF, P_Tot and LF/HF, suggesting a main
sympathetic link between the total score of GCS and the sympa-
thetic branch of the ANS.
The grand average of non-dimensional PSD, calculated as the
geometric means of every discrete frequency, is compared between
the control group (upper panel) and patients (bottom panel) in
Fig. 1. A remarkable significant PSD reduction (almost 10 times
lower) was found in the group of patients.
In Fig. 2 power spectra from the six patients with GCS 3 (A), from
the 10 patients with GCS with values from 4 to 8 (B), and from the 22
healthy control subjects (C) are shown. The spectral range included
frequencies from 0.026 to 0.4 Hz, with a sample resolution of
0.00238095 Hz. The obvious PSD decrement of the whole spectrum,
comparing patients (A and B) with healthy subjects (C), is clearly
noted. Moreover, PSD values in the subgroup of patients with
GCS = 3 (A) are reduced for the whole spectrum, compared with
the subgroup of patients with GCS score ranging from 4 to 8 (B).
In Fig. 3 compressed spectral matrices (CSMs) from a normal
subject (panels A and A
0
) and from a patient with GCS = 8 (panels
Table 1
Comparative analysis between the group of patients and the control group for time, frequency and informational HRV domains indices. O/N, order number of items in the table;
results of items from 9 to 15 are expressed as mean [Quartile range] (Min value and Max values). SD, standard deviation of R–R sequence; Delta_RRs, difference between R–R
maximum and R–R minimum; RMSSD, root mean square of squared differences of R–R series; pNN50, proportion of R–R intervals exceeding the precedents by 50 ms; pMean2%,
proportion of R–R intervals exceeding the precedents for at least 2% from the mean value of the R–R series; StrI, stress neurovegetative index; Ln, natural logarithm; nu,
normalised units; nd, non-dimensional units; cu, conventional units; SEM, standard error of the mean; U, value of the Tscore of Mann Whitney
´s test; p, associated probability to
the value of Tscores. Bold values indicate significant statistical differences.
O/N Indices Control group (N= 22) Patients (N= 16) Up
1 Heart rate (beat/min) (M ± sem) 70.1 ± 1.63 86.97 ± 5.87 96.0 0.018
2 Mean of RRs (ms) (M ± sem) 866.6 ± 21.29 737.6 ± 48.7 96.0 0.018
3 Delta_RRs (ms) 232.77 ± 63.3 69.48 ± 50.55 9.0 0.000...
4 SD (ms) (M ± sem) 37.2 ± 2.31 13.3 ± 3.32 35.0 0.000.. .
5 RMSSD (ms) (M ± sem) 22.2 ± 1.37 7.6 ± 1.41 29.0 0.000...
6 pNN50 (%) (M ± sem) 3.6 ± 0.90 0.00 ± 1.19 80.0 0.000...
7 pMean2% (%) (M ± sem) 39.8 ± 2.77 13.0 ± 62.29 63.0 0.000.. .
8 StrI (cu) (M ± sem) 8.294 ± 1.01 908.9 ± 351.80 18.0 0.000...
9 TriI (cu) (M ± sem) 38.79 ± 2.71 4.47 ± 0.77 0.0 0.000...
10 Shannon’s entropy (cu) 4.743 [0.62] (3.54–5.44) 3.038 [2.25] (0.903/4.31) 38.0 0.000...
11 Ln (P_MBF) ms
2
10.6490 [1.29] (9.22–11.68) 7.7233 [3.54] (0.26–9.67) 5.0 0.000...
12 Ln (P_BF) ms
2
11.2950 [1.40] (9.48–12.91) 7.7559 [3.84] (2.18–10.87) 12.0 0.000...
13 Ln (P_MF) ms
2
10.8695 [2.00] (8.57–13.05) 7.5288 [3.71] (2.33–10.10) 13.0 0.000...
14 Ln (P_AF) ms
2
11.0190 [1.22] (9.68–12.20) 8.2469 [2.96] (5.36–11.07) 27.0 0.000.. .
15 Ln (P_Tot) ms
2
12.6975 [1.11] (11.08–13.93) 9.9066 [3.35] (5.60–11.90) 9.0 0.000...
16 LF/HF (nd) 2.059 [3.24] (0.59–8.82) 1.064 [1.20] (0.02–4.69) 95.0 0.017
17 nu_VLF (%) (M ± sem) 18.423 ± 8.26 14.601 ± 13.73 120.0 0.098
18 nu_LF (%) (M ± sem) 21.893 ± 10.43 23.186 ± 14.84 126.0 0.139
19 nu_MF (%) (M ± sem) 31.482 ± 10.28 12.838 ± 6.94 100.0 0.024
20 nu_HF (%) (M ± sem) 28.198 ± 14.59 49.323 ± 27.68 97.0 0.019
Table 2
Comparative analysis between the subgroups of patients with GCS = 3, and the subgroup with GCS score ranging from 4 to 8, for time, frequency and informational HRV domains
indices. O/N, order number of items in the table; results of items from 9 to 15 are expressed as mean [Quartile range] (Min value and Max values). Delta_RRs, difference between
R–R maximum and R–R minimum; SD, standard deviation of R–R sequence; RMSSD, root mean square of squared differences of R–R series; pNN50, proportion of R–R intervals
exceeding the precedents by 50 ms; pMean2%, proportion of R–R intervals exceeding the precedents for at least 2% from the mean value of the R–R series; StrI, stress
neurovegetative index; Ln, natural logarithm; nu, normalised units; nd, non-dimensional units; cu, conventional units; M, mean value; SEM, standard error of the mean; U, value
of the Tscore of Mann Whitney
´s test; p, associated probability to the value of Tscores; GCS, Glasgow Coma Scale. Bold values indicate significant statistical differences.
Indices GCS = 3 (N= 6) GCS from 4 to 8 (N= 10) Up
1 Heart rate (beat/min) (M ± sem) 77.7 ± 11.0 92.5 ± 6.5 17 0.16
2 Mean RRs (ms) (M ± sem) 834.8 ± 90.6 679.3 ± 50.7 17 0.16
3 Delta_RRs (ms) 51.94 ± 56.9 80.00 ± 46.2 5 0.007
4 SD (ms) (M ± sem) 15.71 ± 8.0 15.85 ± 2.86 22 0.38
5 RMSSD (ms) (M ± sem) 10.29 ± 2.5 9.95 ± 1.78 26 0.66
6 pNN50 (%) (M ± sem) 4.16 ± 3.05 0.31 ± 0.21 22 0.38
7 pMean2% (%) (M ± sem) 180.4 ± 166.0 16.7 ± 6.87 24 0.51
8 SrtI (cu) (M ± sem) 1729.7 ± 787.7 416.5 ± 226.5 24 0.51
9 TriI (cu) (M ± sem) 3.1 ± 1.23 6.04 ± 0.85 11 0.04
10 Shannon’s entropy (cu) 1.497 [0.71] (0.90–2.53) 3.963 [0.81] (2.92–5.44) 2 0.01
11 Ln (P_VLF) ms
2
3.675 [5.98] (0.27–8.88) 8.176 [1.60] (6.16–9.77) 7 0.01
12 Ln (P_LF) ms
2
4.246 [4.98] (2.18–9.93) 8.679 [2.18] (6.31–10.84) 11 0.04
13 Ln (P_MF) ms
2
4.845 [4.55] (2.33–10.1) 7.902 [2.57] (4.91–9.71) 17 0.15
14 Ln (P_HF) ms
2
6.657 [3.97] (5.36–10.63) 8.741 [1.94] (6.81–11.07) 16 0.13
15 Ln (P_Tot) ms
2
6.761 [3.43] (5.61–11.4) 10.157 [1.68] (7.69–11.90) 13 0.06
16 LF/HF (nd) 0.262 [0.94] (0.07–1.09) 1.27 [2.32] (0.11–4.69) 4 0.004
17 nu_VLF (%) (M ± sem) 4.57 ± 1.87 20.62 ± 4.42 6 0.009
18 nu_BF (%) (M ± sem) 11.26 ± 4.23 30.34 ± 3.95 25 0.59
19 nu_MF (%) (M ± sem) 7.859 ± 3.62 12.02 ± 1.86 7 0.01
20 nu_HF (%) (M ± sem) 69.83 ± 8.79 37.02 ± 7.56 12 0.05
4Y. Machado-Ferrer et al. / Clinical Neurophysiology xxx (2012) xxx–xxx
Please cite this article in press as: Machado-Ferrer Y et al. . Clin Neurophysiol (2012), http://dx.doi.org/10.1016/j.clinph.2012.09.008
B and B
0
) are shown. In panel A, the whole frequency spectrum
(0.04–0.4 Hz) is considered. A high PSD of about 12,000 ms
2
limits
any possibility to study the HF range. Nonetheless, in panel A
0
when the frequency range from 0.15 to 0.40 Hz is only analysed,
PSD fluctuations within the HF range can be determined. Consider-
ing the PSD scale, the PSD maximum value is 1500 ms
2
, and hence
it is easy to understand that PSD values of about 12,000 ms
2
over-
lap any analysis of the HF band, when the whole HF range is as-
sessed. In panels B and B
0
, the same analysis is shown for a
patient with GCS = 8, considering the whole and the HF ranges,
respectively. PSD values within the HF band are better visualised
in panel B
0
.
CSMs from the same normal subject of Fig. 3 are shown in Fig. 4
(panels A and A
0
). CSMs for the whole frequency (panel C), and the
HF (panel C
0
) ranges, corresponding to a patient with GCS = 3, are
also presented. This Figure is very illustrative to demonstrate the
PSD reduction in P_LF and P_MF values, with a relative preserva-
tion of PSD values within the 0.15–0.40-Hz frequency range in
the patient with GCS = 3, but with very much lower values, com-
pared to the healthy subject.
4. Discussion
To discuss our results it is necessary to exclude the possible ef-
fect of drugs used in patients. We already remarked that we care-
fully discarded those cases with a personal history of long-term use
of drugs such as hypnotics, autonomic stimulants or blockers that
may affect the autonomic system. Nonetheless, our patients with
lowest GCS were under dopamine therapy.
Dopamine effect is dose dependent (Dandamudi and Chen,
2011; Elkayam et al., 2008; Tian, 2012). Low doses (from 2 to
5
l
gkg
–1
min
–1
) induce an interaction with D1 dopaminergic
receptors located in kidneys, mesenteric plexus and coronary arter-
ies, dilating blood vessels, thus increasing overall renal perfusion.
Nonetheless, with the dose we used in our patients (5.5
l
gkg
1
min
–1
), it rouses a sympathetic predominance, with a cardiac ino-
tropic and positive chronotropic effect, raising the heart rate, and
increasing the blood pressure (mainly systolic, and less affecting
diastolic blood pressure), and producing a decrement of the global
HRV. This physiological effect on HRV indices resembles that re-
lated to physical stress in healthy subjects, producing a clear sym-
pathetic predominance, with an increment of LF and VLF, and the
LH/HF indices (Chen et al., 2011; Hottenrott et al., 2006; Ng et al.,
2009; Simoes et al., 2010). Our six cases with GCS = 3, and the other
2 with GCS = 4, on the contrary, showed significant lower values for
P_VLF and P_LF, with a significant decrement of the LH/HF index.
Moreover, the heart rate and arterial blood pressure values were
not in the range of those produced by the effect of dopamine (Tian,
2012). Hence, our findings were related to the dysfunction of the
ANS in deep coma, and not to the dopamine effect.
In our cases, we found in our cases a significant increment of
heart rate compared to healthy controls. The significant increment
of the heart rate in comatose patients may be considered as a con-
sequence of the diminished control of the vagal centres in the brain
stem over the pacemaker cells of the sinusal node, and also due to
an imbalance produced by the relative increment between the
sympathetic influences, represented in the PSD of the VLF, LF and
MF spectral bands, and the reduction in the vagal PSD in the HF
band (Biswas et al., 2000; Baillard et al., 2002; García et al.,
1995; Gujjar et al., 2004; Machado et al., 2005). It is clear that
the heart frequency represents the integral effect of many factors
(neural, hormonal, humoral and environmental) impending over
the cells of the auricular node (Montes-Brown et al., 2010, 2012).
However, in these patients, where the majority of the independent
variables were carefully controlled, the main factor must be as-
cribed to the ANS control (Baillard et al., 2002; García et al.,
1995; Machado et al., 2005; Montes-Brown et al., 2010, 2012). In
our subgroup of patients with GCS = 3, we found two cases with
Fig. 1. The geometric mean values obtained from the individual spectra of the
control group (upper panel), and for the group of patients (lower panel), are shown.
The geometric mean values, representing the estimation of the power spectral
density of the RRI intervals of both groups, are plotted with the corresponding plus
and minus one standard error of the mean (SEM), in order to provide a graphical
and a statistical measure of the mean values variability.
Table 3
Results obtained from the Spearman correlation among the HRV indices and the GCS
items in the group of patients. SD, standard deviation; Delta_RRs, difference between
R–R maximum and R–R minimum; RMSSD, root mean square of successive squared
differences; pNN50, proportion of R–R intervals exceeding more than 50 ms from
precedent ones; pMean2%, proportion of R–R intervals exceeding precedent ones for
more than 2% from the mean value in the sequence; StrI, neurovegetative stress
index; TriI, triangular index; ShaEn, Shannon
´s entropy; P_VLF, absolute PSD in the VLF
spectral band; P_LF, P_MF, and P_HF, idem for the corresponding spectral bands;
LF/HF, ratio between the absolute PSD of the LF and the HF bands; nu_VLF, values of
the PSD in the VLF band expressed as normalised units (%); nu_LF, nu_MF, and nu_HF,
idem for the corresponding spectral bands. Highlighted values indicate significant
Spearman correlation indices for p< 0.05; GCS, Glasgow Coma Scale; GCS_Total, total
GCS score. Bold values indicate significant statistical differences.
Indices Eyes response Verbal response Motor response GCS_Total
Heart rate 0.28 0.12 0.33 0.34
Mean R-Rs 0.28 0.12 0.33 0.34
Delta_RRs 0.28 0.29 0.38 0.43
SD 0.04 0.33 0.39 0.28
RMSSD 0.15 0.57 0.23 0.25
pNN50 0.23 0.36 0.06 0.15
pMean2% 0.10 0.45 0.03 0.02
StrI 0.08 0.16 0.17 0.20
TriI 0.09 0.24 0.53 0.47
ShaEn 0.02 0.25 0.55 0.45
P_VLF 0.30 0.24 0.60 0.61
P_LF 0.10 0.16 0.48 0.44
P_MF 0.01 0.16 0.38 0.30
P_HF 0.15 0.53 0.51 0.45
P_Tot 0.20 0.45 0.59 0.53
LF/HF 0.43 0.16 0.37 0.51
nu_VLF 0.01 0.08 0.49 0.41
nu_LF 0.39 0.41 0.18 0.25
nu_MF 0.15 0.37 0.33 0.36
nu_HF 0.02 0.33 0.28 0.25
Y. Machado-Ferrer et al. / Clinical Neurophysiology xxx (2012) xxx–xxx 5
Please cite this article in press as: Machado-Ferrer Y et al. . Clin Neurophysiol (2012), http://dx.doi.org/10.1016/j.clinph.2012.09.008
heart rate greater than 110 beats/min, meanwhile the rest showed
a heart rate lower than 65 beats/min. These findings deserve spe-
cial attention for future research.
In general, time, frequency and informational domain indices
pointed out a notable reduction of HRV in comatose patients, com-
pared to normal subjects. Nonetheless, the most important find-
ings in this study came from the comparison of comatose
patients, classified into two subgroups according to the GCS scores.
Regarding the time domain, the triangular index and the Delta_RRs
were significantly reduced in the subgroup with GCS = 3, indicating
the potential efficacy of this indices for future studies of comatose
patients.
Concerning the differences between both subgroups for HRV
indices in the frequency domain, altogether, in our patients the
PSD for the whole frequency spectrum decreased whenever GCS
scores were lower. A significant decrement was found for both
P_VLF and P_LF in the subgroup of GCS = 3, meanwhile although
P_HF was lower in these patients, those changes were not statisti-
cally different, showing a remaining activity in the HF band, com-
pared with the LF range indices. Then a question arises to explain
why the slow bands tend to disappear when coma deepens with
GCS = 3, remaining an activity within the HF range.
To understand these findings, it is necessary to remark on new
functional areas of the ANS, recently described by Goodchild and
Moon (2009). These authors demonstrated the existence of vaso-
motor sympathetic nuclei in the brainstem and cervical spine:
the rostral ventrolateral medulla, the caudal ventrolateral medulla,
the giganto cellularis depressor area, the caudal ventrolateral med-
ullary vasodepressor area, the caudal pressor area, the intermedi-
ate pressor area and other neuronal structures extending from
the inferior pole of the inferior olive to the C1–C2 cervical spinal
segments. Macefield and Henderson (2010) have shown the pres-
ence of almost all of these structures, using fMRI and recording
muscle sympathetic neural activity.
Then, the loss of absolute power of the LF band, when coma
deepens until a GCS = 3, might be explained by the destruction of
brain-stem sympathetic structures, mainly those related to the ros-
tral ventrolateral medulla oblongata area. Regarding the VLF band,
there are still discussions about its origin, probably related to the
renin–angiotensin system, to the thermoregulatory function mech-
anisms, and to central sympathetic activity (Takabatake et al.,
2001). Nonetheless, other authors (Goodchild and Moon, 2009;
Llewellyn-Smith, 2009; Macefield and Henderson, 2010) have af-
firmed that the VLF reflects a functional activity integrated in the
lower part of the medulla, and the first two segments of the spine.
Hence, this area is particularly involved in an adrenergic control of
both vasomotor tone and heart rate, and then it should be particu-
larly affected during the clinical evolution of coma, when the cau-
dal region of medulla oblongata is progressively destroyed. We
recently studied a brain-dead patient by HRV, and found the per-
manence of VLF waves during 10 min, vanishing after that moment
(Machado, 2011). We considered those VLF waves as a probable
correlate of residual vasomotor activity after BD.
Concerning the parasympathetic neural control of the heart
rate, the so-called polyvagal theory (Porges, 2007), and other evi-
dences point to a predominance of the neurons in the NA as the
main source of the chronotropic cardiac activity, while the neurons
of VDN innervate mainly the gastro-intestinal tractus (Berntson
et al., 2007; Philbin et al., 2010). Nonetheless, these parasympa-
thetic nuclei might be also destroyed when the medulla oblongata
Fig. 2. Power spectra represented as ribbons in 3-D diagrams, obtained from the 6 patients with GCS = 3 (A), from the 10 patients with GCS score 4 to 8 (B), and from the 22
healthy control subjects (C), are shown. The spectral range was 0.026–0.4 Hz, with a frequency resolution of 0.00238095 Hz.
6Y. Machado-Ferrer et al. / Clinical Neurophysiology xxx (2012) xxx–xxx
Please cite this article in press as: Machado-Ferrer Y et al. . Clin Neurophysiol (2012), http://dx.doi.org/10.1016/j.clinph.2012.09.008
function is gradually destroyed when coma becomes deeper. None-
theless, the HF band is closely related to respiratory mechanisms,
arising in pulmonary receptors, and others, which send informa-
tion to respiratory centres, are closely anatomically related to the
above-mentioned parasympathetic nuclei in the brainstem (Gujjar
et al., 2004; Goodchild and Moon, 2009; Porges, 2007).
Hence, we consider that the remaining but reduced absolute
power within the HF band in our cases might be explained by
the effect of mechanical ventilation. An useful method in future re-
search for testing if this remaining activity in the HF band is really
due to the effect of mechanical ventilation would be to change the
frequency of the ventilator during a short period of time. For exam-
ple, changing the frequency of ventilator from 15 per min (the peak
frequency would appear at 0.25 Hz) to 18 per min (the peak would
appear at 0.3 Hz) would allow to confirm that the remaining activ-
ity is certainly due to the effect of the mechanical ventilation. It
could be also a previous step before applying the apnoea test for
BD diagnosis.
Our findings are similar to those reported by other authors
(Biswas et al., 2000; Ryan et al., 2011; Su et al., 2005;), who have
found a progressive disappearance of different frequency bands
when clinical evolution in comatose patients deteriorates, despite
that the aetiology of brain injury in our patients is different,
because we did not include traumatic brain-injured patients.
Su et al. (2005) affirmed that the absolute power of the LF and
HF bands decreased stepwise when GCS was diminishing until a
GCS = 4. These authors could not technically consider the VLF band,
because they only studied a time segment of 288 s. Ryan et al.
(2011) demonstrated that the VLF band is an independent predic-
tor of mortality and morbidity in haemodynamically stable trau-
matic brain-injury patients. Gujjar et al. (2004) found that the LF
and VLF spectral PSD correlated with mortality. A significant
reduction in LF power has been associated with traumatic brain in-
jury, BD and sympathetic blockade (Goldstein et al., 1998;
Fathizadeh et al., 2004).
We found in our cases a significant decrement of the LF/HF ratio
in the subgroup of GCS = 3, in relation to the subgroup with GCS
ranging from 4 to 8. This finding is related to our previous discus-
sion about the progressive loss of P_LF, meanwhile some remaining
activity persists in the HF band. Several authors have affirmed that
the LF/HF index mirrors the sympathovagal balance (Appel et al.,
1989; Montano et al., 1994; Sands et al., 1989), and others affirm
that it reflects sympathetic modulations (Malliani et al., 1991;
Malliani and Pagani, 1991; Montano et al., 1994). Biswas et al.
(2000) reported that their patients with a GCS of 3 or 4 was associ-
ated with a lower LF/HF. Ryan et al. (2011) found that decrements in
VLF and LF/HF distribution were the only indices significantly
correlated with mortality in traumatic brain-injury patients.
Fig. 3. In this figure compressed spectral matrices (CSM) from a normal subject (panels A and A
0
) and from a patient with GCS = 8 (panels B and B
0
), are shown. The power
spectral density (PSD) in the HF range is not clearly visualised in the panel A, because PSD in the very low (VLF) and low frequency ranges (LF), from 0.026 to 0.15 Hz, is very
high, and overlaps the high frequency (HF) range components. Nonetheless, in panel A
0
when the frequency range from 0.15 to 0.40 Hz is only analysed, PSD fluctuations
within the HF range can be determined, allowing to assess HF frequencies stability during the whole analysis time in this subject. The CSM in the panel B corresponds to a
comatose patient with GCS = 8 and LF and HF components are clearly detected, because PSD within the LF range is not so high as in the control healthy subject, presented in
panel A. Nevertheless, the diagram in B
0
allows a better analysis of the HF range in this patient.
Y. Machado-Ferrer et al. / Clinical Neurophysiology xxx (2012) xxx–xxx 7
Please cite this article in press as: Machado-Ferrer Y et al. . Clin Neurophysiol (2012), http://dx.doi.org/10.1016/j.clinph.2012.09.008
We used in this study an informational domain HRV index, the
Shannon’s entropy and two other time-domain indices: the trian-
gular index and the Delta_RRs These indices, assessing an overall
HRV, also showed a significant positive correlation with the GCS,
and were significantly reduced in those patients with GCS = 3.
Then, not only the LF/HF ratio should be considered studying
comatose patients, because the Shannon’s entropy, the triangular
index, and the Delta_RRs, with known physiological relationship,
were also very useful for differentiating our two subgroups of
comatose patients.
Nonetheless, for future studies including patients with GCS = 3,
showing a very reduced HRV, sampling frequency should be no less
than 1000 Hz, which methodologically might give the opportunity
of detecting very reduced variability of HRV indices.
Therefore, we conclude that HRV is a minimally invasive, low-
cost methodology particularly suitable for assessing the ANS in
coma (García et al., 1995; Machado et al., 2005; Ryan et al., 2011).
References
Appel ML, Berger RD, Saul JP, Smith JM, Cohen RJ. Beat to beat variability in
cardiovascular variables: noise or music? J Am Coll Cardiol 1989;14:1139–48.
Baillard C, Vivien B, Mansier P, Mangin L, Jasson S, Riou B, Swynghedauw B. Brain
death assessment using instant spectral analysis of heart rate variability. Crit
Care Med 2002;30:306–10.
Berntson GG, Cacioppo JT, Grossman P. Whither vagal tone. Biol Psychol
2007;74:295–300.
Bilgin S, Colak OH, Polat O, Koklukaya E. Determination of a new VLF band in HRV
for ventricular tachyarrhythmia patients. J Med Syst 2010;34:155–60.
Biswas AK, Scott WA, Sommerauer JF, Luckett PM. Heart rate variability after acute
traumatic brain injury in children. Crit Care Med 2000;28:3907–12.
Bravi A, Longtin A, Seely AJ. Review and classification of variability analysis
techniques with clinical applications. Biomed Eng Online 2011;10:90.
Chen JL, Yeh DP, Lee JP, Chen CY, Huang CY, Lee SD, et al. Parasympathetic nervous
activity mirrors recovery status in weightlifting performance after training. J
Strength Cond Res 2011;25:1546–52.
Clark MT, Rusin CG, Hudson JL, Lee H, Delos JB, Guin LE, et al. Breath-by-breath
analysis of cardiorespiratory interaction for quantifying developmental
maturity in premature infants. J Appl Physiol 2012;112:859–67.
Conci F, Di RM, Castiglioni P. Blood pressure and heart rate variability and
baroreflex sensitivity before and after brain death. J Neurol Neurosurg
Psychiatry 2001;71:621–31.
Cooke WH, Salinas J, Convertino VA, Ludwig DA, Hinds D, Duke JH, et al. Heart rate
variability and its association with mortality in prehospital trauma patients. J
Trauma 2006;60:363–70.
Dandamudi S, Chen HH. Evolving treatment strategies for management of
cardiorenal syndrome. Curr Treat Options Cardiovasc Med 2011;13:556–69.
de Souza Neto EP, Abry P, Loiseau P, Cejka JC, Custaud MA, Frutoso J, et al. Empirical
mode decomposition to assess cardiovascular autonomic control in rats.
Fundam Clin Pharmacol 2007;21:481–96.
Elkayam U, Ng TM, Hatamizadeh P, Janmohamed M, Mehra A. Renal vasodilatory
action of dopamine in patients with heart failure: magnitude of effect and site
of action. Circulation 2008;117:200–2005.
Elsenbruch S, Wang Z, Orr WC, Chen JD. Time–frequency analysis of heart
rate variability using short-time fourier analysis. Physiol Meas 2000;21:
229–40.
Fathizadeh P, Shoemaker WC, Wo CC, Colombo J. Autonomic activity in trauma
patients based on variability of heart rate and respiratory rate. Crit Care Med
2004;32:1300–5.
Freitas J, Puig J, Rocha AP, Lago P, Teixeira J, Carvalho MJ, et al. Heart rate variability
in brain death. Clin Auton Res 1996;6:141–6.
García OD, Machado C, Román JM, Cabrera A, Díaz-Comas L, Grave de Peralta R. Heart
rate variability in coma and brain death. In: Machado C, editor. Proceedings of
the second international symposium on brain death. Amsterdam: Elsevier; 1995.
p. 191–200.
Goldstein B, Toweill D, Lai S, Sonnenthal K, Kimberly B. Uncoupling of the
autonomic and cardiovascular systems in acute brain injury. Am J Physiol
1998;275:R1287–92.
Fig. 4. In A, and in A
0
, the compressed spectral matrices (CSM) of the same control subject of the Fig. 3 are shown. The CSM from a patient with GCS = 3, corresponding to the
whole spectrum range (panel C), and to the high frequency (HF) range (from 0.15 to 0.40 Hz), are presented. CSM from the patient (C and C
0
) only show components in the HF
range, with a very low power density value. Nonetheless, the estimated spectra were stable regarding frequency and PSD for the entire period of 1200 s of ECG recording, so
assuring the righteous decision of using segments of 420 s from the whole RRI series.
8Y. Machado-Ferrer et al. / Clinical Neurophysiology xxx (2012) xxx–xxx
Please cite this article in press as: Machado-Ferrer Y et al. . Clin Neurophysiol (2012), http://dx.doi.org/10.1016/j.clinph.2012.09.008
Goodchild AK, Moon EA. Maps of cardiovascular and respiratory regions of rat
ventral medulla: focus on the caudal medulla. J Chem Neuroanat
2009;38:209–21.
Gujjar AR, Sathyaprabha TN, Nagaraja D, Thennarasu K, Pradhan N. Heart rate
variability and outcome in acute severe stroke: role of power spectral analysis.
Neurocrit Care 2004;1:347–53.
Kahraman S, Dutton RP, Hu P, Stansbury L, Xiao Y, Stein DM, et al. Heart rate and
pulse pressure variability are associated with intractable intracranial
hypertension after severe traumatic brain injury. J Neurosurg Anesthesiol
2010;22:296–302.
Leipzig TJ, Lowensohn RI. Heart rate variability in neurosurgical patients.
Neurosurgery 1986;19:356–62.
Li SJ, Su YY, Liu M. Study on early heart rate variability in patients with severe acute
cerebral vascular disease. Zhongguo Wei Zhong Bing Ji Jiu Yi Xue
2003;15:546–9.
Llewellyn-Smith IJ. Anatomy of synaptic circuits controlling the activity of
sympathetic preganglionic neurons. J Chem Neuroanat 2009;38:231–9.
Hottenrott K, Hoos O, Esperer HD. Heart rate variability and physical exercise.
Current status. Herz 2006;31:544–52.
Macefield VG, Henderson LA. Real-time imaging of the medullary circuitry involved
in the generation of spontaneous muscle sympathetic nerve activity in awake
subjects. Hum Brain Mapp 2010;31:539–49.
Machado C. Comisión Nacional para la Determinación y Certificación de la Muerte.
[Resolution for the determination and certification of death in Cuba]. Rev
Neurol 2003;36:763–70.
Machado C, Abeledo M, Alvarez C, Aroche RM, Barrios I, Lasanta AM, et al. Cuba has
passed a law for the determination and certification of death. Adv Exp Med Biol
2004;550:139–42.
Machado C, Garcia OD, Gutierrez J, Portela L, Garcia MC. Heart rate variability in
comatose and brain-dead patients. Clin Neurophysiol 2005;116:2859–60.
Machado C. Determination of death. Acta Anaesthesiol Scand 2005;49:592–3.
Machado C. Describing life to define death: a Cuban perspective. MEDICC Rev
2010;12:40.
Machado C. Blood pressure patterns after brain death [electronic response to Fugate
et al.]. Neurology <http://www.neurology.org/content/77/4/399.full/reply#
neurology_el_43037>; 2011 [accessed 29.03.2012].
Mainardi LT. On the quantification of heart rate variability spectral parameters
using time–frequency and time-varying methods. Philos Trans A Math Phys Eng
Sci 2009;367:255–75.
Malliani A, Pagani M. Spectral analysis of cardiovascular variabilities in the
assessment of sympathetic cardiac regulation in heart failure. Pharmacol Res
1991;24(Suppl 1):43–53.
Malliani A, Pagani M, Lombardi F. Neurovegetative regulation and cardiovascular
diseases. Ann Ital Med Int 1991;6:460–9.
Martinmaki K, Rusko H, Saalasti S, Kettunen J. Ability of short-time Fourier
transform method to detect transient changes in vagal effects on hearts: a
pharmacological blocking study. Am J Physiol Heart Circ Physiol
2006;290:H2582–9.
Martinmaki K, Rusko H. Time–frequency analysis of heart rate variability during
immediate recovery from low and high intensity exercise. Eur J Appl Physiol
2008;102:353–60.
Mejia RE, Pollack MM. Variability in brain death determination practices in children.
JAMA 1995;274:550–3.
Montano N, Ruscone TG, Porta A, Lombardi F, Pagani M, Malliani A. Power spectrum
analysis of heart rate variability to assess the changes in sympathovagal balance
during graded orthostatic tilt. Circulation 1994;90:1826–31.
Montes-Brown J, Machado A, Estevez M, Carricarte C, Velazquez-Perez L. Autonomic
dysfunction in presymptomatic spinocerebellar ataxia type-2. Acta Neurol
Scand 2012;125:24–9.
Montes-Brown J, Sanchez-Cruz G, Garcia AM, Baez ME, Velazquez-Perez L. Heart
rate variability in type 2 spinocerebellar ataxia. Acta Neurol Scand
2010;122:329–35.
Ng J, Sundaram S, Kadish AH, Goldberger JJ. Autonomic effects on the spectral
analysis of heart rate variability after exercise. Am J Physiol Heart Circ Physiol
2009;297:H1421–8.
Philbin KE, Bateman RJ, Mendelowitz D. Clonidine, an alpha2-receptor agonist,
diminishes GABAergic neurotransmission to cardiac vagal neurons in the
nucleus ambiguus. Brain Res 2010;1347:65–70.
Porges SW. A phylogenetic journey through the vague and ambiguous Xth cranial
nerve: a commentary on contemporary heart rate variability research. Biol
Psychol 2007;74:301–7.
Plum F, Posner JB. The diagnosis of stupor and coma. Philadelphia: FA Davis; 1980.
Rapenne T, Moreau D, Lenfant F, Boggio V, Cottin Y, Freysz M. Could heart rate
variability analysis become an early predictor of imminent brain death? A pilot
study. Anesth Analg 2000;91:329–36.
Ryan ML, Ogilvie MP, Pereira BM, Gomez-Rodriguez JC, Manning RJ, Vargas PA, et al.
Heart rate variability is an independent predictor of morbidity and mortality in
hemodynamically stable trauma patients. J Trauma 2011;70:1371–80.
Sands KE, Appel ML, Lilly LS, Schoen FJ, Mudge Jr GH, Cohen RJ. Power spectrum
analysis of heart rate variability in human cardiac transplant recipients.
Circulation 1989;79:76–82.
Schwarz G, Pfurtscheller G, Litscher G, List WF. Quantification of autonomic activity
in the brainstem in normal, comatose and brain dead subjects using heart rate
variability. Funct Neurol 1987;2:149–54.
Seely AJ, Macklem PT. Complex systems and the technology of variability analysis.
Crit Care 2004;8:R367–84.
Shimomura C, Matsuzaka T, Koide E, Kinoshita S, Ono Y, Tsuji Y, et al. Spectral
analysis of heart rate variability in the dysfunction of the brainstem. No To
Hattatsu 1991;23:26–31.
Simoes RP, Mendes RG, Castello V, Machado HG, Almeida LB, Baldissera V, et al.
Heart-rate variability and blood-lactate threshold interaction during
progressive resistance exercise in healthy older men. J Strength Cond Res
2010;24:1313–20.
Su CF, Kuo TB, Kuo JS, Lai HY, Chen HI. Sympathetic and parasympathetic activities
evaluated by heart-rate variability in head injury of various severities. Clin
Neurophysiol 2005;116:1273–9.
Takabatake N, Nakamura H, Minamihaba O, Inage M, Inoue S, Kagaya S, et al. A
novel pathophysiologic phenomenon in cachexic patients with chronic
obstructive pulmonary disease: the relationship between the circadian
rhythm of circulating leptin and the very low-frequency component of heart
rate variability. Am J Respir Crit Care Med 2001;163:1314–9.
Takahashi AC, Porta A, Melo RC, Quiterio RJ, da Silva E, Borghi-Silva A, Tobaldini E,
Montano N, Catai AM. Aging reduces complexity of heart rate variability
assessed by conditional entropy and symbolic analysis. Intern Emerg Med
2012;7:229–35.
Task Force of the European Society of Cardiology and the North American Society of
Pacing and Electrophysiology. Heart rate variability. Standards of measurement,
physiological interpretation, and clinical use. Eur Heart J 1996;17:354–81.
Tian SR. The debate of dopamine’s clinical application. Zhongguo Wei Zhong Bing Ji
Jiu Yi Xue 2012;24:451–2.
Vakilian AR, Iranmanesh F, Nadimi AE, Kahnali JA. Heart rate variability and QT
dispersion study in brain death patients and comatose patients with normal
brainstem function. J Coll Physicians Surg Pak 2011;21:130–3.
Viola AU, Tobaldini E, Chellappa SL, Casali KR, Porta A, Montano N. Short-term
complexity of cardiac autonomic control during sleep: REM as a potential risk
factor for cardiovascular system in aging. PLoS One 2011;6:e19002.
Y. Machado-Ferrer et al. / Clinical Neurophysiology xxx (2012) xxx–xxx 9
Please cite this article in press as: Machado-Ferrer Y et al. . Clin Neurophysiol (2012), http://dx.doi.org/10.1016/j.clinph.2012.09.008
... Estimates are generated automatically to reduce the requirement for human supervision [7][8][9]. Therefore Electrocardiogram (ECG) and Electroencephalogram (EEG) signals are frequently used in the ICU [6,10,11]. ...
... Multiple Heart Rate Variability (HRV) indices measured in the temporal or frequency domains have been examined in coma patients with varied aims. These aims are to evaluate brain death [14][15][16], differentiation of autonomic nerve activity in different stages of coma [17], calculation and definition of HRV indices of patients in a coma [18], correlation of HRV parameters with Glasgow Coma Score (GCS) [11,[19][20][21], the relationship between HRV indices and mortality in patients who are comatose with Traumatic Brain Injury (TBI) [22][23][24][25]. ...
... Machado-Ferrer et al. [11] utilized an informational domain HRV index, Shannon's entropy, and two time-domain indices: the triangle index and Delta RRs. These indices, which evaluate total HRV, also showed a strong positive association with the GCS and were considerably lower in patients with GCS = 3. ...
Article
The Glasgow coma score (GCS) is the most commonly used scale to measure the depth of a coma. Determining the GCS score depends on the clinician's experience and cannot be done fully on sedated and intubated patients. In addition, the success of the scale is being questioned due to differences in reliability and performance among clinicians. Therefore, this study aimed to evaluate GCS with an objective approach using Electrocardiography (ECG) and Electrooculography (EOG) signals based on numerical results. In this study, ECG and EOG signals were obtained simultaneously with the recording scenario in which the family and the nurse play an active role as the emotional stimulus, with a new approach. Using the features extracted from physiological signals, the effects of the family and the nurse on unconscious patients and their level of consciousness were evaluated with statistical analysis methods. In addition, classification studies were performed and the success of physiological signals in classifying the GCS was compared. In the classification of consciousness levels, 88.64% accuracy with ECG signals and 73.70% accuracy with EOG signals were obtained. In addition, our results showed that although the patients were unconscious, they were aware of the emotional stimuli applied by the family/nurse. Thus, in this study, a novel approach was proposed that, the GCS of coma patients could be determined with the analysis of EOG and ECG signals obtained with tactile and auditory stimuli.
... LF, HF, and LF/HF have been reported to be prognostic predictors in TBI (n = 19) [28,40]. The VLF and LF components have been reported to be lower in comatose patients (n = 16) after TBI than in healthy controls [41]. These results could be explained by direct physical damage or the indirect effect of increased intracranial pressure on the control centers of the ANS, including the insula, cingulate gyrus, amygdala, hypothalamus, and brainstem. ...
Article
Full-text available
Measurement of heart rate variability can reveal autonomic nervous system function. Changes in heart rate variability can be associated with disease severity, risk of complications, and prognosis. We aimed to investigate the prognostic value of heart rate variability measurements in patients with moderate-to-severe traumatic brain injury after decompression surgery. We conducted a prospective study of 80 patients with traumatic brain injury after decompression surgery using a noninvasive electrocardiography device for data collection. Assessment of heart rate variability parameters included the time and frequency domains. The correlations between heart rate variability parameters and one-year mortality and functional outcomes were analyzed. Time domain measures of heart rate variability, using the standard deviation of the RR intervals and the square root of the mean squared differences of successive RR intervals, were statistically significantly lower in the group of patients with unfavorable outcomes and those that died. In frequency domain analysis, very low-frequency and total power were significantly higher in patients with favorable functional outcomes. High-frequency, low-frequency, and total power were statistically significantly higher in patients who survived for more than one year. Multivariate analysis using a model combining age and the Glasgow Coma Scale score with variables derived from heart rate variability substantially improved the prognostic value for predicting long-term outcome. These findings reinforced the concept that traumatic brain injury impacts the brain-heart axis and cardiac autonomic modulation even after decompression surgery, and variables derived from heart rate variability may be useful predictors of outcome.
... Our group has used the HRV methodology to explore the autonomic nervous system in comatose and brain-dead patients (2,4,5). However, HRV indices have shown to be under the influence of multiple demographic and physiologic variables (age, sex, heart rate, blood pressure, and respiratory rate) becoming confounding factors (CFs) that need to be considered in these studies (3). ...
Article
Full-text available
In a recently published issue of Pediatric Critical Care Medicine, Piantino et al (1) demonstrated that in children with brain injury, lower heart rate variability (HRV) is an early indicator of autonomic system failure, and it predicts progression to brain death. But we propose to consider some methodological and standardized procedures that we have recently discussed (2, 3).Our group has used the HRV methodology to explore the autonomic nervous system in comatose and brain-dead patients (2, 4, 5). However, HRV indices have shown to be under the influence of multiple demographic and physiologic variables (age, sex, heart rate, blood pressure, and respiratory rate) becoming confounding factors (CFs) that need to be con-sidered in these studies (3). In addition, in patients in coma, the presence of non-Gaussian, nonlinear, and nonstationary processes associated with the multiple mechanisms regulating the autonomic control of the cardiovascular system cannot be ruled out. The traditional approach using the Fourier analysis for the study of spectral components of HRV has been successfully and widely used since its introduction, but strictly speaking, this method is limited to the study of linear systems. One of the well-known methods introduced to analyze nonlinear and nonstationary biomedical signals is the Hilbert-Huang transform (HHT) (2).We studied comatose patients classified into two subgroups according to their Glasgow Coma Scale (GCS): 23 patients with GCS from 6 to 8, and 24 patients in coma with GCS from 3 to 5. A group of 33 healthy participants with ages and sex in the same ranges of the group of patients in coma was considered as the control group. Conventionally, in the study by Estévez-Báez et al (2), the groups were referred to as “Control,” “Glas68,” and “Glas35”.We used a statistical method for the adjustment of HRV in-dices to the effects of CFs and correct them when necessary. We also applied the HHT, considering that HRV indices in coma are nonlinear and nonstationary biologic signals. Our most important result was the reduction of HRV observed in coma-tose patients, which showed a progressive trend associated with deepening of coma, assessed by the GCS.The grand averages of the Hilbert marginal spectra showed a marked reduction of the power spectral density for the 2 group of patients in coma, which was even more intense in the Glas35 group. Second, the peaks of the HRV spectra were progressively not only reduced in magnitude, but almost disappeared in the Glas35 group, particularly for the high-frequency (HF) and low-frequency bands. Third, the appearance of a significant peak in the very HF (VHF) range (0.4–0.6 Hz), particularly evident in the Glas68 group, and only slightly evident in the Glas35 group. This was the first-ever report of the appearance of VHF peaks in coma.HRV indices showed that they can efficiently predict mortality in comatose patients, particularly in the VHF range, which was shown for the first time in the study by Estévez-Báez (2). This deserves further studies to fully define its pathophysiologic meaning. Methodologic and standardized procedures described in this study should be considered in future investigations.
... BD has been characterized by the loss of all HRV power. [51][52][53] With my group, I recently reported a brain-dead case in which the VLF oscillations were the last to vanish, possibly related to residual sympathetic vasomotor activity that progressively disappeared due to the extension of necrosis affecting the nervous centers of the lower part of the medulla and the first 2-3 cervical spine segments. Therefore, the preservation of HRV bands in this patient was a clear demonstration of persistent medullary autonomic activity within vagal and other autonomic central nuclei. ...
Article
Full-text available
In this paper, I review the case of Jahi McMath, who was diagnosed with brain death (BD). Nonetheless, ancillary tests performed nine months after the initial brain insult showed conservation of intracranial structures, EEG activity, and autonomic reactivity to the “Mother Talks” stimulus. She was clinically in an unarousable and unresponsive state, without evidence of self-awareness or awareness of the environment. However, the total absence of brainstem reflexes and partial responsiveness rejected the possibility of a coma. Jahi did not have uws because she was not in a wakefulness state and showed partial responsiveness. She could not be classified as a LIS patient either because LIS patients are wakeful and aware, and although quadriplegic, they fully or partially preserve brainstem reflexes, vertical eye movements or blinking, and respire on their own. She was not in an MCS because she did not preserve arousal and preserved awareness only partially. The CRS-R resulted in a very low score, incompatible with MCS patients. mcs patients fully or partially preserve brainstem reflexes and usually breathe on their own. MCS has always been described as a transitional state between a coma and UWS but never reported in a patient with all clinical BD findings. This case does not contradict the concept of BD but brings again the need to use ancillary tests in BD up for discussion. I concluded that Jahi represented a new disorder of consciousness, non-previously described, which I have termed “reponsive unawakefulness syndrome” (RUS).
Chapter
Full-text available
Forty normal subjects, 21 males, between 17 and 75 years old, were studied to document our normative data. Thirty-five comatose patients, 18 males; among 29 and 91 years old, mean age 56; with acute stroke (less than 72 hours' evolution) were serially studied by HRV and GCS11. All of them had at least once, GCS scoring equal or less than 8 points. They were followed up in a cerebro-vascular unit. HRV measurement and clinical assessment were performed one after the other. Etiologies were as follows: infarction in 11 patients, hemorrhage in 24 (subarachnoid hemorrhage in 9, cerebral hemorrhage in 15). BD was diagnosed in 14 patients, according to the criteria12 of the Institute of Neurology and Neurosurgery in Havana. Statistically significant non-lineal correlation was observed between changes in HRV and different levels of brain functional impairment as measured by the Glasgow Coma Scale. A particularly remarkable, not previously reported observation is a tendency of HRV to increase when supratentorial brain damage causes dysfunction at the diencephalic level. This could relate to the functional suppression of the hypothalamic defense area with brainstem integrity. A very low heart rate variability in all brain dead patients is in agreement with the complete interruption of the autonomic cardio-vascular pathways. Heart rate variability measurements can be an adjunct in the determination of brain death in the context of clinical criteria and other confirmatory tests.
Article
Full-text available
‘Physicians are the only professionals authorized to diagnose and certify death according to a norm established by the Ministry of Public Health’ (5). According to the Cuban Parliament, a resolution is a law that legalizes a working standard within any Ministry and that can be signed and changed by the Minister in function (4, 5). Hence, our lawyers left to physicians the responsibility of presenting norms related to human death. Therefore, the Ministry of Public Health needed to respond to the present Civil Code (5) by writing a resolution on this subject. As the Civil Code did not require a definition, the Commission followed this norm and presented not a concept, but ways to diagnose death. Hence, the Commission enumerated three possible situations for diagnosing death (3, 4): 1) Outside the intensive care environment (without life support) physicians apply the cardio-circulatory and respiratory criteria. 2) In forensic medicine circumstances physicians utilize cadaveric signs. (They do not even need a stethoscope.) 3) In the intensive care environment (with life support) when cardio-circulatory and/or respiratory arrest occurs physicians utilize the cardio-circulatory and respiratory criteria. When physicians suspect an irreversible loss of brain functions in a heart-beating and ventilatory supported case, BD diagnostic criteria are applied. The diagnosis of death was based on the finding of any of the Signs of Death: I Irreversible loss of respiratory function. II Irreversible loss of cardio-circulatory functions. III Algor mortis (postmortem coldness). IV Livor mortis (postmortem lividity). V Rigor mortis (postmortem rigidity). VI Cadaveric spasm. VII Loss of muscle contractions. VIII Putrefaction. IX Irreversible loss of brain functions. Signs I and II correspond to the classical respiratory and cardio-circulatory functions. Signs III to VIII are related to forensic circumstances. Sign IX corresponds to the BD diagnosis. This method of diagnosing death, based on finding any of the signs of death, was not related to the concept that there are different types of death. The irreversible loss of cardio-circulatory and respiratory functions can only cause death when ischemia and anoxia are prolonged enough to produce an irreversible destruction of the brain. According to the Commission there is only one kind of death, based on the irreversible loss of brain functions (3—10). This Cuban law did not even mention the term ‘transplants’. It is clear the human beings die regardless bodies would be useful or not for transplantation (3, 4). C. Machado
Article
Full-text available
Heart rate variability (HRV) has long been used in risk stratification for sudden cardiac death and diabetic autonomic neuropathy. In recent years, both time and frequency domain indices of HRV also gained increasing interest in sports and training sciences. In these fields, HRV is currently used for the noninvasive assessment of autonomic changes associated with short-term and long-term endurance exercise training in both leisure sports activity and high-performance training. Furthermore, HRV is being investigated as a diagnostic marker of overreaching and overtraining. A large body of evidence shows that, in healthy subjects and cardiovascular patients of all ages (up to an age of 70 years), regular aerobic training usually results in a significant improvement of overall as well as instantaneous HRV. These changes, which are accompanied by significant reductions in heart rates both at rest and during submaximal exercise, reflect an increase in autonomic efferent activity and a shift in favor of enhanced vagal modulation of the cardiac rhythm. Regular aerobic training of moderate volume and intensity over a minimum period of 3 months seems to be necessary to ensure these effects, which might be associated with a prognostic benefit regarding overall mortality. At present, available data does not allow for final conclusions with respect to the usefulness of traditional HRV indices in assessing an individual’s exercise performance and monitoring training load. The discrepant results published so far are due to several factors including insufficient study size and design, and different HRV methods. Large-sized and prospectively designed studies are necessary for clarification. It also remains to be seen, whether the traditional HRV indices prove useful in the diagnosis of overreaching and overtraining. Preliminary results, though promising, need to be confirmed in larger cohorts. A basic problem in HRV analysis is nonstationarity of the heart rate signal, which holds particularly true for exercise conditions. Whether, in these conditions, more robust nonlinear HRV methods offer a benefit has to be established in further work.
Chapter
During the last several decades physicians and the community have needed urgent changes in the legal codes for accepting brain death (BD) as death, to obtain organs from heart-beating donors. The “dead donor rule” requires that donors must be first declared dead.1 For this reason, most codes legalizing BD are usually sections of transplant laws.2 Thus, a conceptual and practical controversy emerged: if brain-dead cases were not useful as organ donors, they were usually kept on life support until cardiac arrest occurred.2–4
Article
Objective. —To investigate variability in practices for determining brain death and organ procurement results in pediatric intensive care units (PICUs).Design. —Prospective cohort study.Setting. —Pediatric ICUs.Patients. —Children undergoing brain death evaluations selected from 5415 consecutive PICU admissions.Main Outcome Measures. —Data from children undergoing brain death evaluations including number of coma examinations, number and duration of apnea tests, Pco2 measurements at the end of the apnea test, ancillary tests used to confirm brain death, organ procurement, and reasons for nonprocurement.Results. —A total of 93 (37%) of 248 deaths were brain deaths. Compared with the other deaths, children who were classified as brain dead were sicker on admission (mean Pediatric Risk of Mortality [PRISM] score±SD: 31±11 vs 23±12, P<.001; pre-ICU cardiopulmonary resuscitation: 72% vs 40%, P<.001), and had more traumatic injuries (42% vs 12%, P<.001). Variability in apnea testing included lack of apnea testing in 23 patients (25%) and controversial apnea testing practices in 20 patients (22%). Three patients (3%) had brain death evaluations within hours of discontinuing barbiturate infusions, and four of 30 patients younger than 1 year did not have a confirmatory test. Solid organ procurement was successful in 32%. Reasons for nonprocurement included parental refusal (12%), disease state (12%), and medical examiner's case (22%).Conclusions. —Substantial variability exists in the criteria used by clinicians for the diagnosis of brain death. Some practices are contradictory to the Guidelines for the Determination of Brain Death in Children and to recommendations for apnea testing. Organ procurement could be improved by increased medical examiner cooperation.(JAMA. 1995;274:550-553)
Article
Cachexic patients with chronic obstructive pulmonary disease (COPD) show abnormalities of the autonomic nervous system (ANS), neuroendocrine function, and energy expenditure. Leptin has been implicated in the regulation of ANS, neuroendocine function, and thermogenesis in humans. We assessed the physiologic significance of the circadian rhythm of circulating leptin using power spectrum analysis of heart rate variability (HRV) in nine cachexic male patients with COPD, eight noncachexic patients with COPD, and seven healthy control subjects. A diurnal pattern of 24-h leptin levels was present in both the control subjects (analysis of variance [ANOVA]; F = 7.80, p < 0.0001) and noncachexic COPD patients (F = 9.29, p < 0.0001), but was strikingly absent in the cachexic COPD patients (F = 2.09, p = NS). Analysis of HRV demonstrated that the diurnal rhythm of 24-h very low frequency (VLF; 0.003 to 0.04 Hz) showed significantly identical fluctuations with those of 24-h leptin levels, in all of the three groups (r = 0.388, p < 0.0001). Because VLF has been considered to reflect neuroendocrine and thermoregulatory influences, these data may suggest that the loss of circadian rhythm of circulating leptin has clinical importance in the pathophysiologic features in cachexic patients with COPD.
Article
Dopamine (DA) is a hormone in the catecholamine family. It is produced in the brain as a neurotransmitter and is responsible for mood changes including personality, love and euphoria in drug addicts. DA was first synthesized in 1910 by the British scientists George Barger and James Ewens. In 1958, DA's function as a neurotransmitter was first recognized by Sweden scientists Arvid Carlsson and Nils-Ake Hillarp. For this discovery, Carlsson was awarded the 2000 Nobel Prize in Physiology or Medicine. However, DA rarely is used as a neurotransmitter clinically; rather its vasopressor effects have been widely applied to different clinical scenarios. When it is given as peripheral intravenous infusion, DA activates DA receptor, and α and β receptors for the treatment of shock. The following information comes from pharmacology textbook in my medical school 30 years ago. DA pharmacology:(1)Cardiac: activates β1 receptors in the heart, increases myocardial contractility, and cardiac output. (2)Blood vessels and blood pressure: activates a receptor in blood vessel and DA receptor with minimum effect on β2 receptor. (3)Kidney: dilates kidney blood vessel to increase kidney blood flow, and therefore increases glomerular filtration rate. Also DA can increase sodium excretion and urine output without significant kidney hemodynamic change, which means that DA has direct effect on kidney tubule system. The textbook also mentions that DA's effects are dose dependant and also depend on the distribution of receptors of target organs. At low dose (intravenous infusion rate at 2 μg×kg(-1)×min(-1)), DA increases myocardial contractility, selectively constricts blood vessels of skin and skeleton muscle, dilates kidney, splanchnic, and coronary blood vessels by activating DA receptor with minimum blood pressure change. Finally textbook states, "application: shock and combining with diuretics, dopamine can be used for acute renal failure". There is my handwriting on this page where I religiously recorded what my professor told us in the classroom: "use it in patients with cardiac and renal dysfunction." This was my understanding of DA at that time. Kidney is one of the most important organs in our body and many diseases such as hypertension, diabetes and autoimmune diseases can damage kidney. Protecting kidney is a very important goal during treatment planning. With the advanced search, we further evaluate DA's clinical effects. During my residency training, I have to study updated textbooks, join journal club discussions and attend different conferences. One day at a lecture given by nephrologist, we were told that we still couldn't find the "magic kidney protection" medication. After years of research, studies show that dopamine doesn't provide benefit to kidney dysfunction. This new knowledge changed my view of dopamine. In 2000, Lancet published a study from Australia and New Zealand. This is double-blind, randomized, placebo-controlled trial for 328 patients admitted to 23 intensive care units (ICUs). Patients have at least one indicator of early renal dysfunction [urine output averaging 0.5 ml×kg(-1)×h(-1) for 4 hours or longer, serum creatinine (SCr) concentration more than 150 μmol/L without premorbid renal dysfunction, or an increase in Scr concentration of more than 80 μmol/L in less than 24 hours without a creatinine kinase level more than 5000 U/L or myoglobin in the urine]. Patients received either DA infused at a rate of 2 μg×kg(-1)×min(-1) or an identical amount of placebo administered through a central venous catheter. The primary outcome was peak SCr level during the study. Secondary outcomes included reason for cessation of trial infusion, development of cardiac arrhythmias, duration of mechanical ventilation, length of ICU stay and hospital stay, peak plasma urea concentration during study infusion, change in SCr and urea concentration from baseline to peak value, hourly urine output at predetermined times, number of patients requiring renal replacement therapy, number of patients whose serum creatinine concentrations exceeded 300 μmol/L, and survival to ICU and to hospital discharge. Study shows that no significant differences were found between DA and placebo in any primary or secondary outcome measures. The conclusion is: "renal-dose" DA (2 μg×kg(-1)×min(-1)) does not appear to confer any benefit to critically ill patients at risk for renal failure. Another meta-analysis in 2001 shows the use of low-dose DA for the treatment or prevention of acute renal failure cannot be justified on the basis of available evidence and should be eliminated from routine clinical use. Despite the fact that DA has been proven to provide no renal protection, my professor told me that we could use DA to treat shock. Do we have new studies to prove or disprove this concept? In 2006, Critical Care Medicine published SOAP Study (sepsis occurrence in acutely ill patients study) which includes 3147 patients with shock from 196 ICUs in Europe. This is a cohort, multiple-center, observation study. Patients were followed up until death, until hospital discharge, or for 60 days. Of 3147 patients, 1058 (33.6%) had shock at any time; 462 (14.7%) had septic shock. The intensive care unit mortality rate for shock was 38.3% and 47.4% for septic shock. Of patients in shock, 375 (35.4%) received DA (DA group) and 683 (64.6%) never received DA. Conclusion of this study suggests that DA administration may be associated with increased mortality rates in shock. Again, in 2010, The New England Journal of Medicine published "Comparison of DA and norepinephrine (NE) in the treatment of shock". In this multicenter, randomized trial, researchers assigned 1679 patients with shock into 2 groups (858 patients receive DA and 821 patients receive NE as first line vasopressor). When the blood pressure could not be maintained with the dose of 20 μg×kg(-1)×min(-1) for DA or a dose of 0.19 μg×kg(-1)×min(-1) for NE, open-label norepinephrine, epinephrine, or vasopressin could be added. The primary outcome was the rate of death at 28 days after randomization; secondary end points included the number of days without need for organ support and the occurrence of adverse events. Conclusions show that the death rate is about the same in 2 groups; however, DA group has greater number of adverse events that include arrhythmia(P<0.001), open-label vasopressors(P=0.007), skin ischemia (P=0.09). Let's look at a meta-analysis published this year to compare DA versus NE in the treatment of septic shock. This study shows that DA is associated with more death and higher incidence of arrhythmia compared to NE. Historically norepinephrine was the first vasopressor used to treat shock years ago. However, due to lack of intravascular volume resuscitation at that time, patients with shock show the signs of worsened tissue perfusion after NE administration because it has stronger vasocontrictive effect. DA is much milder than NE with greater inotropic activity; many physicians accept it as vasopressor of choice. Now time is different; fluid resuscitation is the first-line therapeutic strategy to treat shock before vasopressor application. Many recent studies prove that NE is better drug to treat the septic shock than DA. Personally I rarely use DA nowadays and I know DA is not a common drug used in ICU setting in USA. However, DA has been long used by medical personnel; many doctors are familiar with the medication and feel comfortable for its application, therefore, it becomes part of their routine treatment. SOAP Study shows that dopamine was used more in community than in university or city hospitals (43.6%, 36.3% and 29.9%, respectively, P=0.016). As Dr. David Bracco pointed out in his editorial paper, one French survey showed that in selected clinical situations, the choice of catecholamine is based on personal and cultural preferences, not evidence based. There is some evidence that some community hospital physicians are afraid of NE and believe in DA because DA, "a little bit β and α, as inotrope or vasopressor, may do the job". With the new strategy of patient management, maybe we need to educate doctors to change their practice according to evidence. Maybe it is time to abandon dopamine as the first-line vasopressor to treat the patients with shock and "low-dose DA" as the treatment to prevent or treat renal dysfunction. The controversy of DA's clinical application has been going on for years. Some doctors call DA as "silent killer" while others believe that DA is an "obsolete" medication to treat shock. With further research, maybe we will all come to consensus to this topic and decide the fate of DA. silent killer" while others believe that DA is an "obsolete" medication to treat shock. With further research, maybe we will all come to consensus to this topic and decide the fate of DA.
Article
To understand the central neural processes involved in blood pressure regulation we recorded muscle sympathetic nerve activity (MSNA) via a tungsten microelectrode in the common peroneal nerve while performing functional magnetic resonance imaging (fMRI) of the brainstem at 3T. Blood oxygen level dependent (BOLD) changes in signal intensity were measured over 4 s every 8 s (200) volumes; MSNA was recorded during the previous 4 s epoch, which takes into account peripheral conduction delays along unmyelinated axons and neurovascular coupling delays. Analysis of temporal coupling between BOLD signal intensity and nerve signal intensity revealed sites in which the two signals covaried, but only in the medulla. Because scans were conducted in a caudorostral direction, we could constrain the analysis to the medulla by only examining the first 1 s of the fMRI and nerve signals. Increases in MSNA were associated with robust bilateral increases in signal intensity in the dorsolateral region of the medulla that corresponds to the human equivalent of the rostal ventrolateral medulla (RVLM). Reciprocal decreases in signal intensity occurred in the regions of the nucleus tractus solitarius (NTS) and caudal ventrolateral medulla (CVLM). Group analysis also revealed increases in signal intensity in the caudal pressor area (CPA), medullary raphé (MR), and dorsal motor nucleus of the vagus (DMX). We have shown for the first time that this combined approach of recording sympathetic neural activity and fMRI provides real-time imaging of the neural processes responsible for the generation of sympathetic nerve activity in awake human subjects. Hum Brain Mapp, 2010. © 2009 Wiley-Liss, Inc.