Download Low-frequency heart rate variability

Clinical Science (2001) 100, 43–46 (Printed in Great Britain)
Low-frequency heart rate variability:
reproducibility in cardiac transplant
recipients and normal subjects
S. W. LORD*, R. R. SENIOR*, M. DAS*, A. M. WHITTAM†, A. MURRAY†
and J. M. MCCOMB*
*Department of Cardiology, Freeman Hospital, Newcastle upon Tyne NE7 7DN, U.K., and †Regional Medical Physics
Department, Freeman Hospital, Newcastle upon Tyne NE7 7DN, U.K.
A
B
S
T
R
A
C
T
Heart rate variability is a measure of autonomic nervous influence on the heart. It has been
suggested that it could be used to detect autonomic reinnervation to the transplanted heart, but
the reproducibility of the measurement is unknown. In the present study, 21 cardiac transplant
recipients and 21 normal subjects were recruited. Three measurements of heart rate variability
were performed during the day : in the morning, in the early afternoon and in the late afternoon.
These tests were then repeated 1 week later and then again 1 week after that, making nine tests
in all. The within-subject S.D. was 0.49 log units in normal subjects and 0.79 log units in
transplant recipients. In both cases, this is about 15 % of the population range. There was
significant variation in heart rate variability between different times of day in both groups, and
from day to day in transplant recipients. It was concluded that the reproducibility of
measurements of heart rate variability is low, and that differences between measurements
performed at different times of day should be interpreted with caution.
INTRODUCTION
Heart rate variability (HRV) is thought to measure the
neural influences controlling the heart [1]. A change in
HRV would therefore be expected following division of
the cardiac nerves, and this has been demonstrated in
cardiac transplant recipients within 6 weeks of surgery
[2]. However, there is now evidence that sympathetic
nerves regrow to the sinus node in some cardiac transplant recipients, and this has been demonstrated using
several invasive techniques more than 2 years after
transplantation [2–4]. Hence it may be expected that
HRV will return in the months or years after transplantation. Consistent with this, it has been shown that
an increased low-frequency component of HRV measured from short (3 min) recordings relates to an increased heart rate response to the intracoronary injection
of tyramine, the ‘ gold standard ’ measure of sympathetic
reinnervation [5].
In normal subjects, the low-frequency content of HRV
relates to both sympathetic and parasympathetic activity,
preventing differentiation between the two. Bernardi and
colleagues [6] used β-blockade and atropine to demonstrate the absence of parasympathetic reinnervation in
transplant recipients, while providing evidence in favour
of sympathetic regrowth. Thus the low-frequency content of HRV is thought to be a measure mainly of
sympathetic effector function in cardiac transplant recipients.
In order to use low-frequency HRV as a reliable test of
sympathetic reinnervation, it is also necessary to measure
the reproducibility of the test, which has not been
investigated previously in cardiac transplant recipients.
Data concerning the reproducibility of the spectral
Key words : cardiac transplantation, heart rate variability, reproducibility.
Abbreviation : HRV, heart rate variability.
Correspondence : Dr S. W. Lord (e-mail l.a.baines!ncl.ac.uk).
# 2001 The Biochemical Society and the Medical Research Society
43
44
S. W. Lord and others
method in normal subjects are also limited, and refer
mainly to analysis of 24-h tapes. Thus the aim of the
present study was to determine, through serial testing,
whether low-frequency HRV determined from short
recordings is a reproducible measurement in both cardiac
transplant recipients and normal subjects.
METHODS
Subjects
A total of 21 cardiac transplant recipients were recruited
from subjects attending routine surveillance outpatient
clinics. In addition, 21 normal subjects were recruited
from hospital volunteers. All subjects were in sinus
rhythm. In transplant recipients, immunosuppression
and anti-hypertensive therapies were continued throughout. All normal subjects were free from clinical cardiac
disease and were not taking any medication.
Informed written consent was obtained from all
subjects, and the study was approved by the Joint Ethics
Committee of the University of Newcastle upon Tyne
and Newcastle Health Authority.
Study design
Subjects had three HRV measurements performed during
the day : in the morning (between 08.00 and 09.00 hours),
in the early afternoon (between 12.00 and 13.00 hours)
and in the late afternoon (between 15.00 and 16.00
hours). These tests were then repeated 1 week later, and
then again 1 week after that, making nine tests in all. Tests
on the three separate days were performed at the same
times of day.
Data collection and analysis
Measurement of HRV was performed in subjects following a previously described standard protocol [5]. A
10-min ECG recording was made from a chest lead
(standard limb lead II), with the subject instructed to
breathe in time with a signal from the investigator. This
was set at a rate of 10 breaths\min (0.167 Hz). The ECG
trace was recorded using commercial software (Lab
View ; National Instruments) on to a laptop personal
computer via an analogue-to-digital converter sampling
at 250 Hz.
Power spectral analysis of HRV was performed using
standard methods. R waves in the ECG trace were
detected automatically, and any incorrect detections or
ectopic beats were removed manually, with interpolated
beats inserted at the mid-point between the previous and
subsequent beats. The ECG trace was resampled at 4 Hz
using the method of Berger et al. [7], and an RR interval
spectrum was calculated from each of three 5-min
segments (which overlapped by 50 %) by multiplication
by the Hanning function and Fourier transformation.
# 2001 The Biochemical Society and the Medical Research Society
The final estimate was obtained by averaging the resulting
spectra. The low-frequency content was defined as the
area under the spectrum between 0.04 and 0.15 Hz.
Statistical analysis
HRV data are presented as the natural logarithm of the
spectral power, in units of ln (ms#). Data were analysed
only for subjects for whom all nine measurements were
completed.
ANOVA was used for comparisons of values obtained
from transplant recipients and those from normal subjects. Data were analysed considering variation among
subjects, variation among times of day, and variation
among days. Differences between different times of day
or between days were analysed, and are presented only
where ANOVA indicated a significant difference. The
results are presented as estimates of the S.D. for each
subject and the best estimate of within-subject S.D. for
each group. These are derived separately for the normal
and transplant groups by calculating the means of the
subject S.D.s. The coefficients of variation of the original
data are also presented.
RESULTS
Altogether, 18 of 21 transplant recipients and all 21
control subjects completed three sets of three measurements. The mean age of the normal subjects was 37 years,
and that of the cardiac transplant recipients was 55 years.
The latter were studied at a mean of 58 months (range
1–125 months) after transplantation.
Data from the transplant recipients and the control
subjects are presented in Table 1. Figure 1 shows the
mean and variance for each subject, and Figure 2 shows a
Bland–Altman plot of mean against S.D. for both groups
of subjects. The within-subject S.D. was 0.49 log units in
normal subjects and 0.79 log units in transplant recipients. In both cases this is between 14 % and 15 % of the
population range. Coefficients of variation differed between controls and transplant recipients, mainly because
of differences in the mean values of the measurement.
Inclusion of the data from the three transplant recipients
that did not complete all nine measurements tended to
increase the coefficient of variation (from 0.76 to 0.80).
Among control subjects, the between-day variance
was higher than expected by chance (F l 16.4, P
0.001), suggesting a possible difference from day to day.
Inspection of the data suggested that this difference was
due to reduced HRV measurements on day 1. This effect
was not, however, seen among transplant recipients.
There were significant differences between measurements
made at different times of day in both groups (controls,
F l 3.3, P 0.05 ; transplant recipients, F l 5.4,
P 0.01). There was no obvious explanation for this
effect among either transplant recipients or controls.
Reproducibility of heart rate variability
Table 1 Variability of low-frequency HRV in transplant recipients and control subjects
All data, except coefficients of variation, are in logarithmic units (so that an increment of 1 S.D. represents a factor of eS.D. in the original units, ms2 ; for an S.D. of
0.49, this factor is 1.58). NS, not significant.
Group
Controls
Transplant
recipients
Figure 1
Group
mean
Group
range
Within-subject
S.D.
Max. between-day
difference
Max. between-time
difference
Coefficient
of variation
6.62
0.87
3.3 (4.7 to 8.0)
5.6 (k2.1 to 3.5)
0.49
0.79
0.43
NS
0.20
0.48
0.45
0.76
Variability of low-frequency (LF) HRV in transplant recipients and normal subjects
Values are meansp1.96 S.D.
Figure 2
Relationship betweenmean and S.D. for all subjects
DISCUSSION
These data demonstrate that the low-frequency component of HRV is only moderately reproducible in cardiac
transplant recipients and normal subjects : the S.D. of an
individual measurement was much greater than the value
of 5 % of the population mean suggested in a consensus
document [8]. The wider range of values observed in
transplant recipients is probably due to the presence of
both subjects with and subjects without autonomic
reinnervation.
Studies in other groups, including normal subjects [9],
subjects with Type I diabetes [10], chronic angina
sufferers [11] and subjects with congestive heart failure
[12], have suggested that low-frequency HRV is a
reproducible measurement. Those studies dealt largely
with measurements from 24 h tapes, which were repeated
only once and at varying time intervals. Pitzalis and
colleagues [13] studied their subjects three times, but at
irregular intervals, and they reported the results as
correlation coefficients. They found that total power
and high-frequency power were not at all reproducible, and
that low-frequency power had an intraclass correlation
coefficient of between 0.60 and 0.77. These values are in
fact consistent with our data, although the authors do not
provide a direct estimate of the variance of an individual
measurement, which would be useful for the interpretation of the difference between two measurements.
Our data were collected under carefully controlled
conditions at regular predetermined time intervals. Three
measurements were taken at each time of day and on each
# 2001 The Biochemical Society and the Medical Research Society
45
46
S. W. Lord and others
day, meaning that systematic effects due to variations
between days or between times of day could be separated
using ANOVA. Similarly, because nine measurements
were made in each subject, variance within each subject
can be estimated accurately. Thus we believe that our
data represent the most accurate available estimate of the
reproducibility of low-frequency HRV in normal subjects, as well as the only available estimate in cardiac
transplant recipients.
It has been suggested that measures of HRV and the
related parameter, baroreflex sensitivity, could be used
to determine prognosis after myocardial infarction and to
identify subjects at high risk of sudden death [14]. HRV
can also be used to detect autonomic dysfunction, and
denervation and reinnervation after cardiac transplantation [5]. Our data suggest that the variability of these
measurements may cause significant overlap between
normal and abnormal ranges. Some, but not all, of the
variability is due to the time of day, and some may be due
to familiarity with the measurement ; both of these effects
are recognized with other autonomic function tests. It is
possible that the underlying state of the autonomic
nervous system is itself a moving target, so that attempts
to pin it down will always be limited by variation.
The lower HRV measured in normal subjects on day 1
compared with days 2 and 3 was probably related to a
training effect, although the exact mechanism is not
obvious. This effect would not be detected by studies
involving only two measurements. The high diurnal
variation in HRV seen in transplant recipients as compared with normal subjects may be related to known
abnormalities of diurnal variations in heart rate and
blood pressure, or to varying effects of anti-hypertensive
or immunosuppressive drugs. Given that there was no a
priori hypothesis, it is possible that the observed differences are due to chance alone.
We have analysed and presented our data as logarithms
so that it is normally distributed, allowing easy interpretation of the derived S.D. and variances. We used a
relatively low frequency of metronomic respiration
because this can be sustained comfortably for 10 min.
Higher frequencies of respiration produce progressively
falling tidal volume and\or falling carbon dioxide tension,
leading in our view to potential non-stationarity of the
data. We have analysed short segments of data rather than
24 h tapes because we intend to understand the shortterm control of the sinus node and to exclude influences
(such as circulating concentrations of hormones) that
have longer half-lives. Longer segments are an average of
short-term segments, and so can be expected to have a
much lower variance, as reported in the literature.
In conclusion, our data show that even under carefully
controlled conditions HRV is not highly reproducible,
either in transplant recipients or in normal subjects.
Differences between pairs of measurements must be
interpreted with caution.
REFERENCES
1 Pomeranz, B., Macaulay, R. J. B., Caudill, M.A. et al.
(1985) Assessment of autonomic function in humans by
heart rate spectral analysis. Am. J. Physiol. 248,
H151–H153
2 Wilson, R. F., Christensen, B. V., Olivari, M. T., Simon,
A., White, C. W. and Laxson, D. D. (1991) Evidence for
structural sympathetic reinnervation after orthotopic
cardiac transplantation in humans. Circulation 83,
1210–1220
3 Kaye, D. M., Esler, M., Kingwell, B., McPherson, G.,
Esmore, D. and Jennings, G. (1993) Functional and
neurochemical evidence for partial cardiac sympathetic
reinnervation after cardiac transplantation in humans.
Circulation 88, 1110–1118
4 Lord, S. W., Brady, S., Holt, N. D., Mitchell, L., Dark,
J. H. and McComb, J. M. (1996) Exercise response after
cardiac transplantation : correlation with sympathetic
reinnervation. Heart 75, 40–43
5 Lord, S. W., Clayton, R. H., Mitchell, L., Dark, J. H.,
Murray, A. and McComb, J. M. (1997) Sympathetic
reinnervation and heart rate variability after cardiac
transplantation. Heart 77, 532–538
6 Bernardi, L., Bianchini, B., Spadacini, G. et al. (1995)
Demonstrable cardiac reinnervation after human heart
transplantation by carotid baroreflex modulation of RR
interval. Circulation 92, 2895–2903
7 Berger, R. D., Akselrod, S., Gordon, D. and Cohen, R. J.
(1986) An efficient algorithm for spectral analysis of heart
rate variability. IEEE Trans. Biomed. Eng. 33, 900–904
8 Task Force of the European Society of Cardiology and the
North American Society of Pacing and Electrophysiology.
(1996) Heart rate variability. Standards of measurement,
physiological interpretation, and clinical use. Eur. Heart J.
17, 354–381
9 Dimier-David, L., Billon, N., Costagliola, D., Jaillon, P.
and Funck-Brentano, C. (1994) Reproducibility of noninvasive measurement and of short-term variability of
blood pressure and heart rate in healthy volunteers. Br. J.
Clin. Pharmacol. 38, 109–115
10 Burger, A. J., Charlamb, M., Weinrauch, L. A. and D’Elia,
J. A. (1997) Short- and long-term reproducibility of heart
rate variability in patients with long-standing type I
diabetes mellitus. Am. J. Cardiol. 80, 1198–1202
11 Kamalesh, M., Burger, A. J., Kumar, S. and Nesto, R.
(1995) Reproducibility of time and frequency domain
analysis of heart rate variability in patients with chronic
stable angina. Pacing Clin. Electrophysiol. 18, 1991–1994
12 Stein, P. K., Rich, M. W., Rottman, J. N. and Kleiger, R. E.
(1995) Stability of index of heart rate variability in patients
with congestive heart failure. Am. Heart J. 129, 975–981
13 Pitzalis, M. V., Mastropasqua, F., Massari, F. et al. (1996)
Short and long term reproducibility of time and frequency
domain heart rate variability measurements in normal
subjects. Cardiovasc. Res. 32, 226–233
14 La Rovere, M. T., Bigger, Jr, J. T., Marcus, F. I., Mortara,
A. and Schwartz, P. J. (1998) Baroreflex sensitivity and
heart-rate variability in prediction of total cardiac
mortality after myocardial infarction. ATRAMI
(Autonomic Tone and Reflexes After Myocardial
Infarction) Investigators. Lancet 351, 478–484
Received 20 April 2000/21 August 2000; accepted 29 September 2000
# 2001 The Biochemical Society and the Medical Research Society