Recommendations for the Practice of Clinical Neurophysiology:

Guidelines of the International Federation of Clinical Physiology (EEG Suppl. 52)

Editors: G. Deuschl and A. Eisen
q 1999 International Federation of Clinical Neurophysiology. All rights reserved.
Published by Elsevier Science B.V. 283

Chapter 7.2

Heart rate variability

R. Baron a,* and D.J. Ewing b

a
Department of Neurology, Christian-Albrechts-University Kiel, Niemannsweg 147, D-24105 Kiel (Germany)
b
The Scottish Of®ce Department of Health, St. Andrews House, Edinburgh, EH1 3DG (UK)

Method system, i.e. primary autonomic failure, also reduce

heart rate variability. More recent indications
Recording of heart rate over long periods using include primary cardiac disorders and the effects
ambulatory ECG monitoring reveals that the heart of antiarrhythmic drugs on heart rate.
rate varies continuously, mainly in¯uenced by the
cardiac sympathetic and parasympathetic innerva-
tion. Therefore, analysis of heart rate variation Physiological background
provides techniques for the investigation of cardiac
autonomic innervation. The R-R intervals (inverse Heart rate variation during normal activity (time-
of heart rate) yield detailed beat-to-beat informa- domain technique)
tion and its variation is the most useful non-inva- During normal daily activities a beat-by-beat
sive index of cardiac autonomic neuropathy (Ewing variation in heart rate (or change in the R-R interval
et al. 1981; Low et al. 1990; Ewing 1993). length) occurs that is partly determined by the
balance between the slowing effect of the autonomic
parasympathetic and the accelerating effect of the
Indication sympathetic innervation, as well as by humoral
mechanisms, and the intrinsic rhythmicity of the
Patients with disorders which might involve
cardiac pacemaker tissue. Physiological variations
damage of the autonomic nervous system should
of the heart rate include changes associated with
be tested. These include peripheral polyneuropa-
respiration (so-called sinus arrhythmia, 0.15±0.45
thies caused by diabetes mellitus, alcoholism,
Hz), as well as slower alterations associated with
chronic renal failure, AIDS, leprosy, the Guillain±
blood pressure ¯uctuations and barore¯ex mechan-
Barre syndrome and paraneoplastic neuropathies.
isms (0.1 Hz) and with hormonal changes and ther-
Furthermore, central lesions of autonomic centres
moregulation (mainly the renin-angiotensin system,
due to multiple sclerosis or syringomyelia and
0.05 Hz) and also very slow variations in response to
degenerative disorders of the autonomic nervous
day and night. When subjects are supine, parasym-
pathetic activity is most prominent with only
* Correspondence to: Priv.-Doz. Dr. med. Ralf Baron, minimal sympathetic activity.
Klinik fuÈr Neurologie, Christian-Albrechts-UniversitaÈt The time-domain technique is only one method of
Kiel, Niemannsweg 147, D-24105 Kiel (Germany). looking at heart rate variability during normal
284

activity. Alternatively the frequency-domain analy- cardiac output, while systemic vascular resistance
sis can be performed but requires much more is still elevated in response to the falling blood
complicated equipment. Therefore, this chapter pressure of phase II. This, in turn, produces a
will concentrate on the time-domain technique. re¯ex bradycardia and peripheral vasodilatation to
restore the circulatory haemodynamics to normal.
Heart rate variation in response to physiological The re¯ex pathways involved in the Valsalva
stimuli (cardiovascular re¯ex tests) response are complex. The changes in heart rate
are mainly mediated by parasympathetic nerves.
Heart response to deep breathing. During deep
respiration the R-R intervals vary in a sinusoidal
shape that lengthen during inspiration and shorten Technical requirements
during expiration. This variation is greatest at
The time-domain methods to assess resting heart
around 6 breaths per minute and is predominantly
rate variability requires a microprocessor to record
mediated by parasympathetic car-diac nerves. The
the ECG signal and the appropriate program to
response diminishes with age.
analyse the R-R interval signal available in most
Heart rate response to standing up. The heart rate commercial EMG machines. The cardiovascular
response to standing consists of an immediate and re¯ex tests can be performed very easily with
rapid increase, followed by a relative slowing to a minimal equipment. A conventional ECG machine
level that is usually more rapid than the supine heart and a pressure gauge attached to a mouthpiece by a
rate. In normal subjects the heart rate is maximal at tube are needed.
around the 15th beat after starting to stand up, and
the relative bradycardia is reached around the 30th Clinical protocols
beat. Parasympathetic pathways are mainly
involved, sympathetic to a lesser extent. The Heart rate variation during resting activity (time-
extent of the response diminishes with age. domain technique)
A number of different techniques are available to
Heart rate response to the Valsalva measure heart rate variation (Tables 1 and 2). With
manoeuvre. Four phases of the normal cardio-
the time-domain technique a statistical analysis is
vascular response can be distinguished during the
applied to a sequence of R-R intervals or R-R
Valsalva manoeuvre (blowing against resistance for
10±20 s). Phase I is the immediate onset of strain. TABLE 1
This induces a sudden increase in intrathoracic
HOW TO REPORT THE RESULTS
pressure, which is re¯ected by a brief rise in
blood pressure and often a re¯ex drop in heart Heart rate variation analysis
rate for 2±3 s. As the strain continues (phase II), Patient:
venous return is reduced, and this produces a Test Result Age related N/P
normal value
progressive fall in cardiac output and blood
pressure. This blood pressure fall results in a HR variation at rest
steadily increasing heart rate and peripheral CV or RSMMD
vasoconstriction. Phase III is the period
immediately following release of strain. The Cardiovascular re¯ex tests
Deep breathing (E-I)
release of intrathoracic pressure and consequent CV or RSMMD
increase in pulmonary venous capacitance leads (deep breathing)
to a further fall in cardiac output, decrease in Standing up (30:15 ratio)
blood pressure, and a re¯ex increase in heart rate Valsalva ratio
for usually 3 or 4 beats. Phase IV consists of a
Interpretation Normal Pathologic
rebound hypertension caused by the increased
 285

TABLE 2

NORMAL VALUES a

Age (years)

20 25 30 35 40 45 50 55 60 65

CV (rest) 3.43 3.15 2.90 2.66 2.45 2.25 2.07 1.91 1.75 1.61
RSMMD (rest) 16.39 14.54 12.89 11.43 10.13 8.98 7.96 7.06 6.26 5.55

E-I (DB) 136.0 127.4 119.4 111.9 104.8 98.2 92.0 86.2 80.8 75.7
CV (DB) 4.79 4.47 4.18 3.91 3.65 3.41 3.19 2.98 2.78 2.60
RSMMD (DB) 19.27 17.71 16.27 14.95 13.74 12.63 11.60 10.66 9.80 9.01
30:15 1.15 1.14 1.12 1.11 1.10 1.09 1.08 1.07 1.07 1.06
VR 1.22 1.22 1.21 1.20 1.19 1.19 1.18 1.17 1.17 1.16
a
Lower normal limits for age-dependent tests relating to control of heart rate in 120 healthy subjects aged 15±67 years. Values given are the
2.3 centiles. CV in %, RMSSD in ms, E-I in ms; DB, deep breathing. From Ziegler et al. (1992).

interval differences to measure the variation around breathing cycles gives the maximum-minimum (E-
the mean. The subject rests for at least for 5 min in a I) R-R interval ratio.
supine position during which the heart rate is If the subject is able to perform the deep
recorded continuously using a computer to detect breathing task for about 2 min and computer-
the R wave and save the R-R intervals. The based equipment is available, CV and RMSSD
computer program calculates the standard deviation values can also be calculated according to the
(SD) of the given R-R interval sequence and the descriptions above.
coef®cient of variation (CV) around the mean R-
R interval for that sequence (CV ˆ SD=mean £
100). An alternative approach is to calculate the Standing up
square root of successive R-R interval differences The subject is asked to lie quietly on a couch
(RMSSD). A theoretical advantage of this approach under resting conditions and then to stand up
is that it is independent of the prevailing heart rate. unaided as quickly as practicable and remain
standing quietly for 1 min. The longest R-R interval
around the 30th beat after standing divided by the
Heart rate variation in response to standardized shortest R-R interval around the 15th beat is the
stimuli (cardiovascular re¯ex tests) 30:15 ratio.
Deep breathing
During the test the subject sits quietly and
breathes deeply and evenly at 6 breaths/min (5 s Valsalva manoeuvre
in and 5 s out). The heart rate is continuously The test is performed by asking the subject to sit
recorded. Maximum minus minimum (E-I) R-R quietly and then blow into a mouthpiece attached to
interval: The difference between maximal and an aneroid pressure gauge at a pressure of 40
minimal R-R interval during each 10 s breathing mmHg and to hold the pressure for 15 s. The ratio
cycle is measured and the mean of the differences of the longest R-R interval shortly after the
during 3 successive breathing cycles gives the manoeuvre (within about 20 beats) to the shortest
maximum minus minimum (E-I) R-R interval. R-R interval during or immediately after the strain
Expiratory to inspiratory ratio (E/I): The maximum period (phase II or III) is measured. The result is
and minimum R-R interval during each 10s expressed as the Valsalva ratio (VR) which is taken
breathing cycle is measured. The mean of the as the mean ratio from three successive Valsalva
maximum-minimum ratio during three successive manoeuvres.
286

Applications for clinical practice longer recording periods the variation around the
mean becomes more inclusive of general auto-
Standard battery of tests nomic activity, e.g. humoral factors, as opposed
It is a false economy to use only one single test of to a speci®c index of sympathetic and parasympa-
heart rate variation as the repeatability of each test thetic innervation.
is not perfect. On the other hand several tests During the respiratory stimulus the subject has to
described above measure the same stimulus in be able to breathe deeply and evenly for a period of
different ways and give no further information. 2 min since otherwise false pathologic results may
Therefore, it is recommended to perform a standard occur. Close attention has therefore to be paid that
battery of tests and calculate several values the subject breathes deeply and evenly. Patients
described above (CV or RMSSD during rest and with obstructive or restrictive bronchial and lung
deep breathing, E-I, 30:15 ratio and Valsalva disease are often unable to perform this respiratory
ratio, see Tables 1 and 2). The heart rate variation task.
should be interpreted as pathologic if more than During the `standing up' task a misleading 30:15
50% of the values are abnormal. However, heart ratio may be produced, either if the timing is
rate variation mainly indicates abnormalities that counted only from when the subject completes the
affect the innervation of the heart (parasympathetic standing up, or if the 15th and 30th beats are
cardiac dysfunction). Results should be correlated adhered to very rigidly, without allowing for slight
with other tests of autonomic function which variations in the speed of the response among indi-
measure different parts of the autonomic nervous viduals. Patients with orthostatic problems due to
system (e.g. blood pressure responses to standing, autonomic neuropathy might be unable to perform
see Chapter 7.3; sympathetic sudomotor innerva- the test.
tion, see Chapter 7.1) in order to classify and quan- The Valsalva manoeuvre is contraindicated in
tify the full clinical picture of the autonomic subjects with obvious clinical cardiac failure, as
neuropathy. the barore¯exes are not activated because of the
already increased venous return. The blood pres-
Factors affecting the quality of the investigation, sure pattern re¯ects direct mechanical transmission
pitfalls and contraindications of the intrathoracic pressure, with no rebound over-
shoot of blood pressure and an unaltered heart rate.
Tests of heart rate variation can only be The test is also contraindicated in diabetic patients
performed if the subject's heart is in sinus rhythm. with proliferative retinopathy because of the possi-
Patients with arrhythmias or ectopic beats must be bility of provoking retinal haemorrhage.
excluded since otherwise false normal results might
be produced. Furthermore, occasional nonsinus, References
ectopic beats have to be carefully excluded from
the analysis. Computer programs to measure Ewing, D.J. Noninvasive evaluation of heart rate: the time domain.
successive R-R intervals represent a time-saving In: P.A. Low (Ed.), Clinical Autonomic Disorders. Little,
Brown, Boston, 1993: 297±314.
approach towards statistical analysis. However, Ewing, D.J., Campbell, I.W. and Clarke, B.F. Heart rate changes in
the program must be able to pick up only the R diabetes mellitus. Lancet, 1981, 1: 183±186.
wave and not a tall T wave. Ectopic beats, muscle Low, P.A., Offer-Gehrking, T.L., Proper, C.J. and Zimmerman, I.
The effect of aging on cardiac autonomic and postganglionic
and movement artefacts must be detected and
sudomotor function. Muscle Nerve, 1990, 13: 152±157.
excluded by the system. Ziegler, D., Laux, G., Dannehl, K., SpuÈler, M., MuÈhlen, H., Mayer,
Any medication that affects heart rate may in¯u- P. and Gries, F.A. Assessment of cardiovascular autonomic
ence the results of the tests. function: age-related normal ranges and reproducibility of spec-
tral analysis, vector analysis, and standard tests of heart rate
Analysis of resting heart rate variation includes variation and blood pressure responses. Diabetic Med., 1992,
the danger of overinterpretation of the results. With 9: 166±175.