Marc R Dweck BSc (Hons) MB ChB*, Christopher C Lang* BSc (Hons) MB ChB MD, James MM Neilson PhD**, Andrew D Flapan MB ChB MD*
*Department of Cardiology and **Department of Medical Physics, Royal Infirmary of Edinburgh, Little France Crescent, Edinburgh, United Kingdom
Address for Correspondence: Dr C Lang Department of Cardiology Royal Infirmary of Edinburgh Little France Crescent Edinburgh United Kingdom. E mail: chrislang@katamail.com
SMJ 2006 51(1): 57
Abstract
Background:
Methods:
18 healthy subjects (6 female, 12 male, median age 22) underwent four separate
24-hour ECG recordings using 2-channel Holter recorders. The protocol contained
five different arousal events: Natural Waking (woke naturally, then stood up);
Morning Alarm (woken by alarm in the morning, then stood up); Night Alarm (woken
by alarm during the night, then stood up); Morning Alarm-Remain Lying (woken by
alarm in the morning but remained supine) and Lying to Standing (stood up from a
supine position during the day). Holter recordings were analysed using a
commercial package for dynamic assessment of the QT-RR relationship.
Results:
In the twenty minutes after arousal no changes were seen in overall QT-RR
relationship in any of the groups. However, marked T-wave morphology changes,
including T wave inversion, were observed in all the arousal events.
Postural changes only accounted for a small proportion of change in T
wave morphology.
Conclusions:
In healthy subjects noxious arousal causes marked changes in the morphology of
the T wave. This may reflect abnormal adaptation of repolarisation to sudden
changes in heart rate and autonomic tone.
Keywords: Repolarisation, QT, T-wave, sudden death.
Introduction
Acute arousal such as a sudden fright or being woken from sleep can result in
cardiac arrhythmias in susceptible individuals such as those with ischaemic
heart disease or the congenital long-QT syndrome (1, 2). There is a well established excess in the incidence of
sudden cardiac death in the early hours of the morning (3, 4), which coincides with changes in autonomic tone, heart rate,
blood pressure and catecholamine levels and a greater likelihood of ischemia.
The QT interval as measured on the ECG reflects the duration of ventricular repolarisation and has an intimate relationship with heart rate, but is also influenced by changes in autonomic tone (which in turn also affects heart rate) and other factors (5,6,7,8). For this reason, the relationship between QT and heart rate can vary within individuals and also between individuals and groups of patients.
We
hypothesised that sudden arousal may acutely affect the cardiac repolarisation
characteristics, as measured by the dynamic relationship between the QT interval
and heart rate, and also by assessing T wave morphology.
Methods
Subjects
18
healthy volunteers (6 females, 12 males, median age 22) with no history of
cardiac problems, and on no medication were recruited and consent obtained. The study was carried out in accordance with local research
ethics guidelines.
Ambulatory
ECG recordings
A
two-channel “Tracker 2” 24h ECG recorder with bipolar thoracic leads (CM1
and CM5) (DelMar Reynolds Medical, Hertford, England) was attached to the
subjects. Skin preparation was conducted before attachment of the electrodes to
ensure a good contact, and a monitor was used to select optimal electrode
positions to provide a large T-wave and small P wave. Subjects were asked to
avoid alcohol and strenuous physical exercise during the study period.
Study
Design
In
order to ascertain the relative influence of changes in posture, sudden waking
from different stages of sleep and natural awakening (defined as spontaneous
awakening without an alarm call), a protocol was devised whereby each subject
underwent four 24h ECG recordings, forming five separate events for analysis.
The different events were as follows:
1. Lying to Standing
(L-S)
Subjects
lay still on the couch, at the time of Holter attachment, for 20 minutes, then
rose and stood up for 20minutes. This was to assess the postural component of
any changes in the other waking events.
2. Natural Waking (NW)
Subjects
woke naturally in the morning without being woken by an alarm, rose from bed and
stood up for 20minutes.
3. Morning Alarm (MA)
Subjects
were woken by a noxious alarm between 0700h and 0900h. They then rose from bed
and stood up for 20minutes.
4. Night Alarm (NA)
Subjects
were woken by a noxious alarm between 0300 and 0500h. They then rose from bed,
and completed a simple mental arithmetic test, lasting approximately 20minutes.
Subjects were therefore kept awake and forced to achieve full consciousness.
5. Morning Alarm- Remain
Lying (MA-L)
Subjects
were woken by a noxious alarm and lay still in bed for 20minutes before rising.
This was designed to assess the individual effect of sudden waking from sleep
without the influence of a change in posture.
QT
and T wave Analysis
Tapes were replayed on Pathfinder Analyser. RR and QT are both measured with the end of the T wave (TEND) being determined by the slope method. QT and RR data was analysed using VERDA software (DelMar Reynolds Medical Limited, Hertford, England) to give a continuous measurement of the QT-RR relationship. The method for continuous assessment of the QT-RR relationship has been described elsewhere (9). In brief, the QT and RR interval data are processed automatically by applying a two-component time delay to the RR data. This removes the effect of QT adaptation lag on QT-RR plots and therefore the hysteresis loops normally observed. Accurate continuous estimation of the underlying relationship is therefore possible. The data is modelled to a general exponential formula of the form QT0= QT (RR0/RR) J where QT0 is the intercept of the curve at an RR interval of 1000ms, and J is a variable exponent. The overall relationship between QT and RR can therefore be expressed at any point in time using the variables QTo and J. The software produces continuous output of QT, QTo, RR, Slope, J and correlation (r) from data averaged over a scrolling 3 minute time window.
T
wave amplitude was measured using on-screen calipers, and quoted as the
percentage change from the pre-event value in order to compensate for the
inter-subject variation in T-wave size. Subjects recorded the timing of the
events as those displayed on the Tracker, so that they corresponded to times on
the time-track of the recordings. Baseline pre-event values were taken during
the 30 minutes before arousal events. Post-event values were taken 20 minutes
after arousal events.
Statistics
All
statistical comparisons of continuous variables were made using paired T-tests.
Results
Heart
Rate Changes
There
was a biphasic response in HR change for each arousal event. During the first
twenty seconds there was a notable HR increase achieved over the space of four
or five beats which was termed ‘Instantaneous HR change’ (IHRC). After this
twenty second period the HR then reduced towards a new stable value, midway
between the initial change and the pre-event value and was termed ‘late HR
change’ (LHRC). (Figure 1).
The
IHRC was very similar in magnitude in the NW, MA and NA groups reaching their
peaks at around 100 beats/min (Table 1
and Figure 2). However, the
peaks reached in the L-S group and MA-L groups were lower at 95 beats/min and 87
beats/min respectively. The difference between both these peak values and the
night alarm peak value was statistically significant (p<0.005). In all groups
this early and rapid change was closely linked to the timing of the observed
T-wave changes (Figures 3, 4).
In all arousal events where the subject stood up, the heart recovered to between
70 and 80 beats/min. However in the MA-L group, where the patient remained
lying, the heart rate dropped back to 56 beats/min, within 10% of the pre-event
value.
Changes
in the QT/RR Relationship
None
of the arousal events produced a significant change in the QT/RR relationship
within the time period examined (i.e. 20 minutes after arousal). The QT
shortened with arousal, but in line with changes in HR as displayed by the lack
of significant change in the values of QT0and J. This finding
suggests that in healthy subjects with normal heart the QT-RR relationship is
stable under these circumstances. (Table
2).
T-Wave
Morphology Changes
The T-wave underwent a characteristic change in morphology after all of the arousal events, and was related to the abruptness and timing of wakening. In all event groups, the T-wave remained unchanged for up to 10 seconds, but then underwent a significant reduction in amplitude (Initial Flattening, IF). This lasted for between twenty and forty seconds, before the T-wave regained some of its amplitude (Late Recovery, LR) (Figures 3-4, Table 3).This change was observed in 96% of the arousal events conducted.
Initial
Flattening
These
changes varied in their intensity (see Table 4). In the L-S and MA-L groups they
were often subtle (T-wave reduced by 20% to 60% of its original height). In the
MA, NW and NA groups they were often more pronounced with the T-wave becoming
appreciably flatter (>60% change in height); appearing to flatten completely
(>80%) or indeed inverting (>100% loss of height). In only six arousal
events across all the groups was a flattening of less than 20% observed (No
change). T-wave flattening of over 60% was seen in all NA cases (with 4
inverting, and 6 flattening) (see Table 4). In the NA, MA, and NW groups the
mean T-wave amplitude dropped by more than 75% of its original value. Smaller
yet still significant changes were observed in the L-S and MA-L groups
(p<0.0005 and p<0.0001 respectively).
The
initial change in the NA group was significantly greater than in the NW group
(p=0.037) although not in the MA group (p=0.196).
Late
Recovery
In
all groups the T-wave was seen to partially regain its height after the initial
flattening. A recovery of at least 20% was observed in all subjects in all
groups except for three subjects in the L-S group, where the T wave remained
reduced in height after two minutes. In all groups, recovery of the height of
the T wave was not complete, being significantly lower than that of the T wave
prior to the arousal event. However, in the MA-L group it did return to 94% of
its original value.
Discussion
This
study examined the effect of different methods of arousal upon ventricular
repolarisation, as measured by short-term changes in the QT/RR relationship, and
quantitative assessment of T-wave morphology. Although no change was seen in the
QT dynamics a marked change in T-wave morphology was observed.
QT
Dynamics
We
hypothesised that sudden changes in autonomic tone and levels of circulating
catecholamines associated with arousal events would influence the QT/RR
relationship. This did not prove to be the case. This may be because the
greatest influence on heart rate in these young, healthy subjects was the
withdrawal of vagal tone rather than an increase in sympathetic activity. It is
thought that vagal activity has less effect on action potential duration and
therefore QT interval compared to sympathetic activity. Previous studies have
shown significant changes in QT dynamics during the first few hours after waking
(10), (however, the investigators relied on assessment of QT-RR slope averaged
over longer time periods, which will be affected by heart rate if the QT-RR
relationship is exponential). These were not detectable in this study in the
twenty minutes after arousal suggesting that they develop more gradually over
the morning. Lang et al (9) demonstrated that medium term (i.e. over 2 hours)
changes in the QT/RR relationship develop on rising in the morning in patients
with previous aborted sudden cardiac death, but these changes were gradual, and
were blunted in those patients receiving beta-blocker therapy. However, in the
control group, no changes were observed on waking. Their study appeared to
implicate the sympathetic nervous system in more gradual changes in the QT/RR
relationship in response to increased metabolic stress, but only in subjects
with cardiac disease. Our data would appear to underline the stable nature of
the QT/RR relationship in healthy individuals.
Heart
Rate Changes
Arousal-events
induced a biphasic response in HR: an initial sharp increase followed by a
decrease to a later stable value (Figure
1).
Initial
Abrupt Change
In
the NW, NA and MA groups this sharp increase reached its peak at 100 beats/min
over a period of four or five beats. Abrupt withdrawal of parasympathetic tone
is responsible for these rapid HR changes. Similar but smaller HR changes were
observed when the subjects stood up (L-S) or woke suddenly but stayed supine
(MA-L). These more subtle changes would have been the result of more gradual
changes in autonomic tone.
T-Wave
Morphology Changes
The
T-wave underwent a biphasic change in morphology after arousal-events: a delayed
decrease in amplitude, including complete inversion, followed by gradual
recovery (figure 3). Previously,
less well-characterised changes in T-wave amplitude have been reported during
exercise (11). There are two potential causes of T wave flattening. Firstly, a
physiological change in the transmural repolarisation gradient, and secondly a
postural change, which is caused by the change in position of the heart, and
therefore the T wave axis on the ECG. A change in the subject’s posture was
involved in four of five arousal events when subjects stood up from a lying
position. However, the same change in T-wave morphology was observed in the MA-L
group, when the subjects remained lying down.
This suggests that posture is not solely responsible for the observed T
wave changes. This could not be excluded in a previous study examining the
effect of arousal on the T wave (12). Furthermore, one would expect a postural
change to be immediate in response to standing (occurring as the heart adopts
its new position) and to persist as long as the new position is maintained.
However, in this study the T-wave remained unchanged for up to 10 seconds after
standing, and subsequently regained some of its height. This indicates that
changes in T-wave morphology must in part be due to changes in the pattern of
ventricular repolarisation. Rosenbaum et al suggested a mechanism whereby sudden
changes in heart rates could affect T-wave shape (13). They assessed the effect
of abrupt cycle-length shortening (within the physiological range) on the action
potential duration of individual myocytes, in a perfused guinea pig heart. The
study showed that the normal apex to base action potential duration (APD)
gradient of the ventricle was lost during these abrupt HR increases. The apex to
base APD gradient is thought to contribute to the generation of the T-wave on
the electrocardiogram (14), and narrowing of this distribution could therefore
result in reduced T-wave amplitude.
Abrupt
increases in heart rate were produced by the arousal events in this study and
occurred over a similar time period to the T-wave morphological changes (Figure
3). Furthermore, recovery in heart rate also mirrored recovery of T-wave
height in terms of timing, amplitude of recovery, and differences between groups
(e.g. more complete recovery of both variables in group MA-L when the subjects
remained lying during arousal). Sudden changes in HR, induced by the ANS (e.g.
by parasympathetic withdrawal) might therefore be the cause of the T-wave
morphology changes.
Variation
in the Magnitude of T-wave Flattening
The
range of T-wave flattening varied in magnitude, from subtle changes not readily
visible to the naked eye, to complete T-wave flattening and even T-wave
inversion. The changes were more marked after the most contrasting arousal
events (e.g. most marked with the sudden night-alarm), and this may be a
function of the differing levels of sympathetic activation they induced.
Considerable inter-subject variation existed, with some subjects tending to show
only subtle changes whilst others had flattened or inverted their T waves as
their typical response. This may have been a due to a difference in the
susceptibility of the subjects’ ventricles to the effects of the ANS or to
greater postural change in cardiac axis in some subjects.
Implications
and Future Work
Variation
in T-wave morphology has been linked to SCD, and may in fact provide a more
reliable guide than QT-interval (15). Further work is required to test this type
of protocol in patients at potential risk of ventricular arrhythmias, although
should only be performed in a hospital setting where resuscitation facilities
would be available.
This
study suggests that acute short-term changes in the QT/RR relationship do not
occur in healthy individuals in situations of acute physical or emotional
stress. Having established in this study that it is feasible to assess short
term changes in the QT/RR relationship, it would be informative to assess
patients with structural heart disease and ischaemic heart disease to see
whether the QT/RR relationship is as stable in this group.
Limitations
The
method of QT/RR analysis employed in this study compensates for the QT
adaptation lag, removing hysteresis from QT:RR plots. This method may therefore
obscure the apparent mismatch between QT and RR during a sudden increase in
heart rate (where the RR interval has suddenly shortened and the QT lags behind
and would be considered inappropriately long). While this mismatch may well
predispose to arrhythmia and is effectively ignored for the purposes of this
study we would still expect to detect any change in the QT:RR relationship
brought about by non-HR mediated influences on cardiac repolarisation.
Conclusions
Acute
and noxious arousal did not cause changes in short-term QT dynamics in this
study, but an alteration in ventricular repolarisation was evident in the form
of changes in T-wave morphology. These changes were most marked when subjects
were woken unexpectedly from deep sleep, even when allowances were made for
change in posture.
References
1. Wellens HJ, Vermeulen A, Durrer D. Ventricular Fibrillation Occurring on Arousal from Sleep by Auditory Stimuli. Circulation. 1972;46: 661-665.
2. Ali R, et al. Clinical and Genetic Variables Associated with Acute Arousals and Nonarousal-Related Cardiac Events Among Subjects with the Long QT Syndrome. Am J Cardiol. 2000;85: 457-461.
3. Willich SN, et al. Circadian Variation in the Incidence of Sudden Cardiac Death in the Framingham Heart Study Population. Am J Cardiol. 1987;60: 801-806.
4. D'Avila A, Wellens F, Andries E, Brugada P. At what time are implantable defibrillator shocks delivered? Evidence for individual circadian variance in sudden cardiac death. European Heart J. 1995;6: 1231-1233.
5. Fananapazir L, Bennet DH, Faragher EB. Contribution of heart rate to QT interval shortening during exercise. European Heart J 1983;4:265-271.
6. Boyett MR, Jewell BR. A study of the factors responsible for rate-dependent shortening of the action potential in mammalian ventricular muscle. J Physiol 1978;285:359-80
7. Stern S, Eisberg S. The effect of propanolol on the electrocardiogram of normal subjects. Am Heart J 1969;77:192-5
8. Rickard AF, Norman J. Relations between QT interval and heart rate. New design of physiological adaptive pacemaker. Br Heart J 1981;45:56-61
9. Lang C, Neilson JM, Flapan AD. Abnormal Repolarisation Dynamics are Seen During the 'High Risk' Waking Hours in Cardiac Arrest Survivors. Europace. 2001; 2 (Supp C): C4.
10. Yi G, Guo XH, Gallagher MM, Hnatkova K, Reardon M, Poloniecki J, Camm AJ, Malik M. Circadian Pattern of QT/RR Adaptation in patients with and without Sudden Cardiac Death after Myocardial infarction. Annals of Non-invasive Electrocardiology 1999; 4: 286-294.
11. Kitchin A, Neilson J. The T wave of the electrocardiogram during and after exercise in normal subjects. Cardiovascular Research 1972;6:143-149.
12. Toivonen L, Helenius K, Viitasalo M. Electrocardiographic Repolarisation During Stress from Awakening on Alarm Call. J Am Coll Cardiol 1997;30:774-9.
13. Rosenbaum D, Kaplan T, Kanai A, Jackson L, Garan H, Cohen R, Salama G. Ventricular action potentials, ventricular extracellular potentials and the ECG of the guinea pig. Circulation 1991;84:1333-1345.
14. Watanabe T, Rautaharju PM, McDonald TF. Ventricular action potentials, ventricular extracellular potentials and the ECG of the guinea pig. Circ Res 1985;57:362-73.
15. Zabel M, Malik M. Predictive value of T-wave morphology variables and QT dispersion for post-myocardial infarction risk assessment. J Electrocardiol 2001;34:27-35.