1. science
2. medicine
3. surgery

British Journal of Anaesthesia 1998; 81: 333–337
Respiratory sinus arrhythmia and clinical signs of anaesthesia in
children
C. M. BLUES AND C. J. D. POMFRETT
Summary
We have investigated changes in respiratory
sinus arrhythmia (RSA) and compared these with
clinical signs of anaesthesia in children. Children
aged 3–10 yr were anaesthetized by gaseous
induction with halothane and nitrous oxide.
Multiple heart rate variability (HRV) spectra were
obtained by power spectral analysis of continuous epochs of time from before introduction of
halothane (baseline) until the pupils were central
and fixed (stage 3). Measurement of RSA was
performed by integration of the area under the
spectral curve within the range of the respiratory
frequency 0.15 Hz. In all patients RSA decreased
continuously during induction unless stimulation
occurred with insertion of an airway. Values of
RSA were compared at three times: baseline, loss
of pharyngeal tone and stage 3. The decrease
in RSA from baseline to loss of pharyngeal tone
and from loss of pharyngeal tone to stage 3 was
significant (P:0.003 and P:0.018, respectively).
These results show that RSA can be related to the
clinical signs of anaesthesia and has potential as a
measure of depth of anaesthesia in children. (Br.
J. Anaesth. 1998; 81: 333–337).
Keywords: heart, heart rate; measurement techniques, spectral analysis; anaesthesia, paediatrics; anaesthesia, depth
In health, heart rate varies constantly. These beat-tobeat changes in instantaneous heart rate are termed
heart rate variability (HRV). Power spectral analysis
has facilitated the study of the autonomic control
systems that modulate HRV.
Two spectral components of HRV have been
observed: a low frequency component mediated by
both the parasympathetic and sympathetic systems
and a high frequency component centred at the frequency of ventilation that is thought to be exclusively
under parasympathetic control.1–4 High frequency
oscillation of heart rate with respiration is termed
respiratory sinus arrhythmia (RSA) and is thought to
be a selective index of vagal control of the heart5 and
a window on the autonomic activity processed in the
nucleus solitarius of the brainstem.6
Since the suggestion that RSA may offer a method
of measuring depth of anaesthesia,7 several groups
have studied the effects of general anaesthetic agents
on RSA.8–14 All of these studies measured RSA before
and after induction of anaesthesia; the changes in
RSA during the progressive stages of induction of
anaesthesia have not been investigated.
The objectives of this study were to examine the
changes in RSA during gaseous induction and to
relate these changes to clinical signs during the progression through the stages of anaesthesia in children
aged 3–10 yr. The higher spontaneous rate of respiration in this age group produces a distinct high frequency component that is needed for measurement
of RSA.
Patients and methods
We studied 13 ASA I patients, aged 3–10 yr, undergoing elective ENT surgery. After approval of the
local Ethics Committee, informed written consent
was obtained from the parents. No patient had a history of autonomic dysfunction or was receiving any
concurrent medication.
Children were premedicated with oral midazolam
0.5 mg kg91. On arrival in the anaesthetic room, monitoring leads were attached and the patient preoxygenated for as long as possible to provide a baseline
measurement. Nitrous oxide was administered followed by as small increments of halothane as the
clinical situation would permit until the pupils
became small and central, and the eyes immobile
(stage 3). The airway was secured as appropriate for
the procedure and anaesthesia was maintained with
oxygen, nitrous oxide and halothane.
Data were collected continually. Ventilatory flow
was recorded using a Magtrak turbine (Ferraris
Medical Ltd) inserted at the proximal end of the
breathing system (T-piece). The Magtrak produced
a flow-related pulse train during inspiration.
Electrocardiogram (ECG) data were collected using
standard ECG leads and recorded using a software
controlled amplifier (CED 1902);12-bit digitized
ECG and respiratory pulse trains were digitized at
1 kHz using a laboratory interface and commercial
software (CED 1401 with Spike 2 software). From
the parent and child’s perspective, only routine
monitors came into contact with the child; they were
prepared for the presence of the computer in the
anaesthetic room. End-tidal values of nitrous oxide,
halothane and carbon dioxide were recorded
together with MAC values, as calculated by a Datex
Capnomac. The times of all events during induction
were recorded using an event marker and the nature
of the event was documented.
C. M. BLUES, BSC, MB, CHB, FRCA, C. J. D. POMFRETT, BSC, PHD,
Department of Anaesthesia, University of Manchester,
Manchester Royal Infirmary, Manchester M13 9WL. Accepted for
publication: April 28, 1998.
334
British Journal of Anaesthesia
Table 1 Values of heart rate (HR) and respiratory frequency (f) recorded at the three defined points
during induction. Group values are shown as median (range). Statistical differences using the two-tailed
Wilcoxon matched pairs signed ranks test: **P:0.01 compared with baseline; ††P:0.01 compared with
loss of pharyngeal tone
Baseline
Loss of pharyngeal tone
Stage 3
Patient
No.
HR
(beat min91)
f (bpm)
HR
(beat min91)
f (bpm)
HR
(beat min91)
f (bpm)
1
2
3
4
5
6
7
8
9
10
11
12
Median
151
97
95
94
91
96
91
108
142
92
85
106
95 (66)
24
20
21
21
21
21
25
21
21
21
22
21
21 (5)
130
116
159
See text
106
101
110
111
139
99
92
116
111 (66)
31
36
33
See text
21
36
45
31
36
32
36
31
33 (24)**
130
113
160
97
107
99
100
92
138
115
74
130
110 (86)
32
45
45
36
30
42
45
31
42
39
40
51
41 (21)** ††
DATA ANALYSIS
Each R wave of the ECG was extracted from the data
as an event using an algorithm for peak detection.
Correct identification of every R wave was verified
manually before analysis. A waveform was created by
smoothing this event data over 1-s intervals using a
raised cosine function and a sampling frequency of
4 Hz.15 The result was a smoothed heart rate frequency
curve (tachygram) representing equi-distant data
suitable for spectral analysis. The same procedure was
performed on inspiratory event times calculated from
the Magtrak pulse trains, with a sampling frequency
of 1 Hz. This gave an inspiratory tachygram with which
to determine instantaneous inspiratory frequency. To
remove slow trends in the data, a high-pass filter was
applied to the re-sampled HRV tachygram series
removing frequencies less than 0.056 Hz. A low-pass
filter at 1 Hz was also applied to exclude components
greater than the Nyquist frequency.9
Spectral analysis was then performed on 128-s
segments of the filtered heart rate signal.9 16 Starting
at time 0, the analysis was repeated at 60-s intervals,
thereby overlapping the epochs by 68 s. A 256-point
fast Fourier transformation was applied to the data.
The power spectra describe HRV power as a function of frequency. HRV power within a frequency
range was obtained by integration of the area under
the power spectral curve. The spectrum was not normalized.17 18–22
Since the rate of spontaneous respiration varied, it
was not appropriate to choose one band width for
measurement of RSA. The high frequency component is centred around the frequency of respiration
(f). The magnitude of RSA was thus measured as
the power of the high frequency component of the
variability spectrum that lay within the band width
(f)<0.15 Hz.19
STATISTICAL ANALYSIS
No presumption of normality was made. RSA measured at three clinical stages was assessed using the
two-tailed Wilcoxon matched pairs signed rank test.
P:0.05 was considered statistically significant. The
stages were: (I) before introduction of halothane
(baseline); (2) loss of pharyngeal tone; and (3) return
of the pupils to a central fixed position (stage 3). Loss
of pharyngeal tone was taken as a stage as it was
clearly identifiable without the need for stimulation
of a reflex.
Results
One child was excluded because of a technical fault
in data collection. We studied 12 children, aged 3–10
yr. A discernible high frequency peak corresponding
to RSA was observed in all HRV spectra. The
frequency of this peak increased as respiratory frequency increased during administration of halothane
(table 1). There were no significant changes in heart
rate.
The size of the RSA peak decreased markedly in all
12 children as they progressed through the stages of
anaesthesia. The characteristic pattern of this decline
is shown in figure 1, in which consecutive HRV spectra obtained from the overlapping epochs of time are
shown for a single child (patient No. 3, table 2). Five
epochs (1–3, 5, 6) have been taken from this plot and
displayed sequentially in figure 2. Each HRV spectrum shows the RSA present at a specific clinical
stage of anaesthesia. The HRV spectrum shown for
the epoch at which stage 3 of anaesthesia is reached
(fig. 2E) has been plotted on a magnified scale (x10).
A second complete plot, from patient No. 8, is
shown in figure 3. In this plot a high level of low frequency activity can be seen. RSA increases initially as
halothane is commenced and then decreases in the
characteristic manner. The increase in RSA seen in
epoch 6 corresponds to insertion of a Guedel airway.
This pattern of decline in RSA occurred in all 12
children. Despite this, grouping such continuous
data for the purpose of statistical analysis is difficult
as the speed of induction could not be standardized
in the clinical situation. We took three clinical stages
that were clearly identifiable in each of the children
and compared the magnitude of the RSA (table 2).
There was wide variation in absolute baseline values.
One patient was excluded from analysis of loss of
pharyngeal tone as early insertion of an airway precluded its identification. In four patients, insertion of
an airway unavoidably caused stimulation during the
measurement period of stage 3 RSA (patient Nos 1,
Respiratory sinus arrhythmia and clinical signs of anaesthesia
Figure 1 Three-dimensional plot of multiple heart rate variability
(HRV) spectra obtained from consecutive overlapping epochs of
time during induction of anaesthesia in patient No. 3.
2, 7 and 8) and therefore these patients are not
included in the statistical analysis.
RSA decreased significantly between baseline
measurement and loss of pharyngeal tone (P:0.003)
and between loss of pharyngeal tone and stage 3
anaesthesia (P:0.018). When the values were considered as percentage change from baseline, the median
value at loss of pharyngeal tone was 11.4% and at
stage 3, 3.39%.
Discussion
We have demonstrated, for the first time, a continuous decrease in RSA during induction of anaesthesia,
in the absence of stimulation, which can be related to
the clinical signs of anaesthesia first described by
Guedel in l937.
The study of RSA in children confers an important
methodological advantage. All quantitative measures
335
of HRV are based on the premise of separable
frequency ranges with no overlap. The high rate of
respiration in children generates a high frequency
component that can be separated clearly from the
low frequency component.3 In adults, this clarity can
only be achieved by breathing in time to a
metronome. By taking a band width for integration
centred about the frequency of respiration of the
monitored individual, we were able to minimize the
loss of data falling outside the band width.
In 1935 it was first suggested that RSA may
provide a quantitative measure for monitoring the
level of anaesthesia when it was shown that RSA
decreased in dogs after induction with ether20 and
halothane.21 With the advent of microcomputers,
interest in changes in RSA during anaesthesia was,
renewed. In 1985, Donchin, Feld and Porges, using
measurement of R-R intervals, showed a decrease in
RSA in 10 female patients after induction of anaesthesia with isoflurane.7 Since then, several other studies
have reported decreases in RSA after administration of
anaesthetic agents.8–14 Seven of these eight studies
used i.v. induction of anaesthesia.7 9–14 One study used
gaseous induction with isoflurane but no measurements were made during induction.8 Use of an i.v.
induction agent precludes the study of RSA during
induction of anaesthesia and introduces the problem
of the effect of the induction agent itself on the autonomic nervous system.22 Volatile agents also affect the
autonomic nervous system, but the known cardiac
properties of halothane would suggest a shift in the
sympatho-vagal balance towards vagal predominance. Such an action would lead to an increase in
the proportion of power within the high frequency
band, the opposite effect to the decrease in RSA
shown in this study. Seven of the studies compared
pre-induction RSA in spontaneously breathing
patients with RSA measured during controlled ventilation.7 8 10–14 RSA is known to increase with controlled
breathing despite no changes in variables of respiration or depth of anaesthesia.4 All of these studies
compared RSA measured before and at sometime
after induction. Changes in RSA during induction of
anaesthesia in spontaneously breathing patients have
not been investigated previously.
The only studies of RSA in children have been in
Table 2 Values for respiratory sinus arrhythmia (% of baseline) measured at the three defined points
during induction. Group values of absolute RSA and % of baseline are shown as median (range) and
exclude epochs where stimulation occurred during the measurement period (labelled a). **P:0.01
compared with baseline (n:11); †P:0.05 compared with baseline (n:8); ‡P:0.05 compared with loss of
pharyngeal tone (n:7)
Patient No.
(age in years)
Baseline
RSA (% of baseline)
Loss of pharyngeal tone
RSA (% of baseline)
Stage 3
RSA (% of baseline)
1 (5)
2 (5)
3 (5)
4 (6)
5 (7)
6 (6)
7 (5)
8 (4)
9 (3)
10 (10)
11 (8)
12 (5)
Median RSA
Median % baseline
6.31E-06 (100)
2.13E-04 (100)
2.95E-05 (100)
9.13E-05 (100)
1.57E-04 (100)
3.79E-05 (100)
4.47E-04 (100)
2.85E-05 (100)
1.73E-06 (100)
8.93E-05 (100)
3.13E-05 (100)
5.7E-05 (100)
4.74E-05 (4.45E-04)
100 (0)
1.01E-06 (16)
5.78E-07 (0.3)
3.40E-07 (1.1)
See text
1.45E-05 (9.2)
4.34E-06 (11.4)
3.84E-07 (0.08)
3.42E-06 (12)
7.76E-07 (44)
9.26E-06 (10)
5.07E-06 (16)
7.35E-06 (12.9)
3.42E-06 (1.41E-05)**
11.4 (44.7)
1.13E-06 (17)a
8.25E-07 (0.4)a
3.96E-08 (0.13)
3.48E-06 (3.8)
4.65E-06 (2.97)
1.50E-06 (3.94)
1.88E-06 (0.42)a
5.43E-06 (19)a
1.39E-07 (8)
2.78E-07 (0.3)
4.42E-06 (14)
1.9E-08 (0.03)
8.9E-07 (4.63E-06)†‡
3.39 (14)
336
British Journal of Anaesthesia
Figure 3 Three-dimensional plot of multiple heart rate variability
(HRV) spectra obtained from consecutive overlapping epochs of
time during induction of anaesthesia in patient No. 8.
Figure 2 Heart rate variability (HRV) spectra during successive
clinical stages of anaesthesia. A: Epoch 1, before introduction of
halothane, respiratory frequency (f):0.47 Hz. B: Epoch 2, after
introduction of halothane during the excitement phase, f:0.47 Hz.
C: Epoch 3, loss of eyelash reflex, f:0.51 Hz. D: Epoch 5, loss of
pharyngeal reflex, f:0.56 Hz. E: Epoch 6, magnified view of RSA
at stage 3, f:0.56 Hz.
neonates, investigating RSA as a tool for the differentiation of groups of neonates at risk from sudden
infant death syndrome and in the assessment of neurological impairment after asphyxia.24–27 RSA changes
during anaesthesia in children have not been studied
previously. Few studies have attempted to investigate
any method of assessing anaesthetic adequacy in this
age group. O’Kelly, Smith and Pilkinton examined
the auditory evoked response but found the method
unreliable because of the size of the patient and the
consequent interaction between the signal frequency
of the EEG and ECG.23
Studies in neonates demonstrate that this population can be divided into two distinct groups representing significantly different patterns for HRV.19 28
One group has the majority of power concentrated in
the high frequency range while in the second group,
low frequency activity is dominant. Our study
showed that the contribution of RSA to total power
was also a variable feature in older children; some
children had considerably more low frequency activity than others. This variation between children in the
balance between sympathetic and parasympathetic
control of the nervous system makes comparison of
absolute values of RSA between individuals of little
use in the study of the effects of anaesthesia.
However, changes in RSA can be used to examine
variations in vagal tone of an individual in different
clinical situations. We followed each child through
induction and compared the changes in RSA with
baseline measurements. This was expressed as percentage change from baseline values.
When using RSA as an index of central vagal tone,
it must always be remembered that this is a cardiac
measure derived from the ECG. It reflects only the
final vagal influences on heart rate. RSA represents
not only the central levels of vagal activity that we
wish to measure, but the working of the efferent
pathways conducting the neural impulses, the traffic
at the cholinergic receptor sites in the sinus node and
the end organ itself, the heart.29–31 When studying the
effect of anaesthesia on brainstem autonomic activity, the effect of anaesthetic agents on the rest of the
system cannot be ignored. We minimized this possible error by using only three agents (midazolam,
halothane and nitrous oxide).
RSA offers considerable promise as a non-invasive
and sensitive index of depth of anaesthesia in children. In five children a Guedel airway was inserted.
In each case the stimulus was accompanied by an
increase in power in the high frequency range and in
the ratio of this component to the total power of the
spectrum. In two patients, when an i.v. cannula was
Respiratory sinus arrhythmia and clinical signs of anaesthesia
inserted after stage 3 had been reached, an increase
in RSA was seen.
Future studies should investigate the relationship
between changes in RSA and cognitive function
using tests of cognition and memory. Such studies
may need to be performed in an older age group.
We have shown that RSA is a measure of the stage
of anaesthetic depth in children but not as yet a
measure of the level of conscious awareness.
References
1. Askelrod S, Gordon D, Ubel FA, Shannon DC, Barger AC,
Cohen RJ. Power spectrum analysis of heart rate fluctuations:
a quantitative probe to beat to beat cardiovascular control.
Science 1981; 213: 220–222.
2. Askelrod S, Gordon D, Madwed JB, Snidman NC, Shannon
DC, Cohen RJ. Haemodynamic regulation: investigation by
spectral analysis. American Journal of Physiology 1985; 249:
M867–875.
3. Pomeranz B, Macaulay RJB, Caudill MA, Kutz I, Adam D,
Gordon D, Kilborn KM, Barger AC, Shannon DC, Cohen RJ,
Benson H. Assessment of autonomic function in humans by
heart rate spectral analysis. American Journal of Physiology
1985; 248: H151–153.
4. Pagani P, Lombardi F, Guzzetti S, Rimoldi O, Furlan R,
Pizzinelli P, Sandrone G, Malfatto G, Dell’Orto S, Piccaluga
E, Turiel M, Baselli G, Cerutti S, Malliani A. Power spectral
analysis of heart rate and arterial pressure variability as a
marker of sympatho-vagal interaction in man and conscious
dog. Circulation Research 1986; 59: 178–193.
5. Bernston GG, Cacioppo JT, Quigley KS. Respiratory sinus
arrhythmia: Autonomic origins, physiological mechanisms,
and psychophysiological implications. Psychophysiology 1993;
30: 183–196.
6. Porges SW. Respiratory sinus arrhythmia: physiological basis,
quantitative methods and clinical implications. In: Grossman P,
Janssen K, Vaitl D, eds. Cardiorespiratory and Cardiosomatic
Psychophysiology. New York: Plenum, 1986; 101–116.
7. Donchin Y, Feld JM, Porges SW. Respiratory sinus arrhythmia
during recovery from isoflurane-nitrous oxide anesthesia.
Anesthesia and Analgesia 1985; 64: 811–815.
8. Kato M, Komatsu T, Kimura T, Sugiyama F, Nakashima K,
Shimada Y. Spectral analysis of heart rate variability during
isoflurane anesthesia. Anesthesiology 1992; 77: 669–674.
9. Galletly DC, Corfiatis T, Westenberg AM, Robinson BJ. Heart
rate periodicities during induction of propofol-nitrous
oxide–isoflurane anaesthesia. British Journal of Anaesthesia
1992; 68: 360–364.
10. Komatsu T, Kimura T, Shimada Y. Continuous, on-line,
real-time spectral analysis of heart rate variations during
anaesthesia: Computing and Monitoring in Anesthesia and
Intensive Care. Berlin: Springer-Verlag, 1992; 335–338.
11. Pomfrett CJD, Barrie JR, Healy TEJ. Respiratory sinus
arrhythmia: an index of light anaesthesia. British Journal of
Anaesthesia 1993; 71: 212–217.
12. Galletly DC, Westenberg AM, Robinson BJ, Corfiatis T. Effect
of halothane, isoflurane and fentanyl on spectral components
of heart rate variability. British Journal of Anaesthesia 1994; 72:
177–180.
13. Pomfrett CJD, Sneyd JR, Barrie JR, Healy TEJ. Respiratory
sinus arrhythmia: Comparison with EEG indices during
isoflurane anaesthesia at 0.65 and 1.2 MAC. British Journal of
Anaesthesia 1994; 72: 397–402.
337
14. Huang H-H, Chan H-L, Lin P-L, Wu C-P, Huang C-H. Time
frequency spectral analysis of heart rate variability during
induction of anaesthesia. British Journal of Anaesthesia 1997;
79: 754–758.
15. Berger RD, Askelrod S, Gordon D, Cohen RJ. An efficient
algorithm for spectral analysis of heart rate variability. IEEE
Transactions on Biomedical Engineering 1986; 33: 900–904.
16. Latson TW, O’Flaherty D. Effects of surgical stimulation on
autonomic nervous function: assessment by changes in heart
rate variability. British Journal of Anaesthesia 1993; 70: 301.
17. Jaffe RS, Fung DL. Constructing a heart rate variability analysis system. Journal of Clinical Monitoring 1994; 10: 45–58.
18. Bianchi A, Bontempi B, Cerutti S, Gianoglia P, Comi G,
Natali Sora MG. Spectral analysis of heart rate variability
signal and respiration in diabetic subjects. Medical and
Biological Engineering and Computing 1990; 28: 205–211.
19. Baldzer K, Dykes FD, Jones A, Brogan M, Carrigan TA,
Giddens DP. Heart rate variability analysis in full term infants:
Spectral indices for study of neonatal cardiorespiratory
control. Pediatric Research 1989; 26: 188–195.
20. Samaan A. The effect of adrenaline, atropine and ether anaesthesia on the heart rate of normal dogs and the animal
deprived of different parts of the autonomic nervous system.
Archives Internationales de Pharmacodynamie et de Therapie
1935; 50: 101.
21. McGrady JD, Vallbona C, Hoffe HE. The effect of preanesthetic and anesthetic agents on the respiration-heart rate
response of dogs. American Journal of Veterinary Research 1965;
153: 256–256.
22. Scheffer GJ, Ten Voorde BJ, Karemaker JM, Ross HH,
Delange JJ. Effects of thiopentone, etomidate and propofol on
beat to beat cardiovascular signals in man. Anaesthesia 1993;
48: 849.
23. O’Kelly SW, Smith DC, Pilkinton SN. The auditory evoked
potential and paediatric anaesthesia. British Journal of
Anaesthesia 1995; 75: 428–130.
24. Kitney RI, Ong HG. An analysis of cardiorespiratory control
in babies and its relation to sudden infant death syndrome.
Automedica 1986; 5: 289–310.
25. Kluge KA, Harper RM, Schechtman VL, Wilson AJ, Hoffman
HJ, Southall DP. Spectral analysis assessment of respiratory
sinus arrhythmia in normal infants and infants who died of
sudden infant death syndrome. Paediatric, Research 1988; 24:
677–682.
26. Rothe M, Zweiner U, Eiselt M, Witte H, Zwacka G, Frenzel J.
Differentiation of healthy newborns and newborns-at-risk by
spectral analysis of heart rate fluctuations and respiratory
movements. Early Human Development 1987; 15: 349–363.
27. Witte H, Rother M. High frequency and low frequency heart
rate fluctuation analysis in new-borns—A review of possibilities and limitations. Basic Research in Cardiology 1992; 87:
193–204.
28. Thompson CR, Brown JS, Gee H, Taylor EW. Heart rate variability in healthy new-borns: the contribution of respiratory
sinus arrhythmia. Early Human Development 1993; 31:
217–228.
29. Katona PG, Lipson D, Dauchot PJ. Opposing peripheral and
central effects of atropine on parasympathetic cardiac control.
American Journal of Physiology 1977; 232: H146–151.
30. Hall GT, Potter EK. Attenuation of vagal action following
sympathetic stimulation is modulated by pre-junctional
(-2-adrenoreceptors in the dog). Journal of the Autonomic
Nervous System 1990; 30: 129–138.
31. Warner MR, Levy MN. Neuropeptide Y as a putative modulator of the vagal effects upon heart rate. Circulation Research
1989; 24: 567–574.