Download Circa- and ultradians in the occurrence of simple extrasystoles in

Acta Physiologica Hungarica, Volume 95 (3), pp. 321–329 (2008)
DOI: 10.1556/APhysiol.95.2008.3.4
Circa- and ultradians in the occurrence
of simple extrasystoles in healthy men
at lowland in the light of inferential statistics
M Mikulecký1, Š Kujaník2
1 Department
on Biometry and Statistics, Neuroendocrinology Letters, Stockholm-Bratislava,
Studenohorská 10, SK 841 03 Bratislava 47, Slovak Republic, E-mail: [email protected]
2 Department of Physiology, Faculty of Medicine, University of Pavol Jozef Šafárik,
Trieda SNP 1, SK 040 66 Košice, Slovak Republic
Received: 14 March 2007
Accepted: 12 December, 2007
Purpose. An exact description of circadian and ultradian rhythms of simple extrasystoles in healthy
subjects is lacking. Thirty seven healthy male subjects, aged 50–76, were randomly taken. Simple
extrasystoles from 24-hour Holter ECG were registered and calculated per each hour and processed by
cosinor regression. Their occurrence is formulated by 95% confidence and tolerance corridors for
supraventricular and ventricular extrasystoles.
Results. The number of extrasystoles is relatively low, dispersion of ventricular extrasystoles is
significantly higher than supraventricular ones. A significant increase of their frequency versus the
general zero trend straight line was found only around 9 a.m. In supraventricular ones a significant
increase around 6–7 a.m., 9 a.m., noon, and 4–5 p.m. and a significant depression around 1–2 a.m. and
10 p.m. was present. Only the 24-hour rhythm is significantly present (α=0.05) in ventricular beats
while in supraventricular ones also the period lengths of 12, 6, 4, 3.4 and 2.4 hours are significant. The
significant difference between supraventricular and ventricular extrasystoles exists in the 24-hour
amplitudes and 6-hour acrophases.
Conclusion. One circadian and several ultradian rhythms of simple extrasystoles are present in
healthy male subjects over 50 years of age. The 95% tolerance chronograms can be exploited in clinical
practice.
Keywords: healthy subjects, male, Holter ECG, extrasystoles, Cosinor regression, statistical
inference
Simple supraventricular and ventricular premature beats can be found in low numbers in
healthy subjects, especially the elderly ones (4, 9, 13, 15, 20, 22). The number of them
is slightly increasing with age (16, 19, 20). Nevertheless, there is a part of healthy
Corresponding author: Prof. Štefan Kujaník, MD, PhD
Phone/Fax: +241-55-6423763; E-mail: [email protected]
0231–424X/$ 20.00 © 2008 Akadémiai Kiadó, Budapest
322
M Mikulecký and Š Kujaník
subjects, particularly the younger ones, with no extrasystole per 24 hours. The
occurrence of extrasystoles in time shows the periodic fluctuations, i.e. biological
rhythms (6, 7, 11).
There are not many statistically exact studies of the circa- and particularly
ultradian profile of supraventricular and ventricular extrasystoles in healthy human
subjects in the literature. The authors usually use descriptive statistics with the emphasis
on the point estimates of means, SD, SE and p values. The present paper prefers the
prediction of 24-hour chronogram for the underlying population with the aid of 95%
confidence (for means) and tolerance (for individual measurements) of one-hour
frequencies of extrasystoles. This approach gives the possibility to serve as the norm for
future practical evaluation. Such an orientation of research towards practice is the
substance of the W. S. Gosset’s – Student’s revolutionary inductive thinking (12).
Materials and Methods
The ECG data of 37 healthy male subjects, aged 50–76 years, have been randomly
taken into this study from the clinical archive 1993–1996. They had been living in the
lowland region of Kosice (211 m above sea level). The demands of the Declaration of
Helsinki had been fully respected and the subjects had given informed consent to
participate. Women had not been included because of some significant
electrocardiological gender differences (5).
Our subjects had had no history of a chronic disease, had displayed normal
cardiopulmonary findings (blood pressure, ECG, examination by auscultation) and had
not taken any medication. A usual resting hospital regimen with sleeping period 22 p.m.
to 6 a.m. (CET) had been secured, any increased physical activity had been excluded.
The primary 24-hour Holter ECG monitoring (Premier IV Diagnostic Monitoring or
Memoport C Hellige) in the precordial DAJ leads had been evaluated by a cardiologist
who had not been included in statistical processing of the obtained primary data at that
time. The originally observed data (1,066 supraventricular and 1,131 ventricular
measurements, i.e. the sum of 2,197 measurements including zero values) are available
at the e-mail [email protected].
The values of simple supraventricular and ventricular extrasystoles were processed
by the Halberg cosinor regression (3, 17), testing the presence of 24-hour rhythm and its
1 st up to the 10th harmonics, i.e. 24 h, 12 h, 8 h, 6 h, 4.8 h, 4 h, 3.43 h, 3 h, 2.67 h and
2.4 hours.
The data were fitted to the regression equation
Yt = a0+Σ (ai.cos[(360°/τi).t+Φi]
Yt = point estimate of the dependent variable, i.e. of the mean hourly number of the
given kind of extrasystoles at the time t,
a0 = mesor value, i.e. the mean value of the sinusoids (rhythm adjusted mean
because the zero trend of data is supposed),
Acta Physiologica Hungarica 95, 2008
Extrasystoles in healthy male persons at lowland
323
ai =
τi =
t=
Φ1 =
amplitude of an i-th sinusoid (with i = 1, 2, 3 , 4, 5, 6, 7, 8, 9, 10),
period length in hours,
time in hours
acrophase, i.e. the difference between the time zero and that of the first acme
(peak), given as angle degrees where 360° correspondings with one period (it
is understood that its value is negative as commonly used for the clockwise
direction).
The results are presented as approximated regression curves showing the mean
estimate of the regression function, its 95% confidence (for mean) and 95% tolerance
(for separate individual measurements) corridors. The 95% confidence intervals are
given also for amplitudes and acrophases. The significance of deviations from the daily
mean in separate hours is visible in the graphs as nonoverlapping the mesor straight line
by the actual confidence corridor.
Statistical significance of the presence of separate periodic components was tested
by using 95% confidence intervals for the amplitudes: the non-overlapping of the null
value means a significant presence on the level α=0.05. The relation of the confidence
interval to the null value gives the exact idea about the effect size expected in future
practice. The F-test was used for mutual comparison of the total variances for either
kind of extrasystole. The level of statistical significance was set at α=0.05 (5%),
corresponding to the 95% level of confidence and tolerance. The 3 degrees of freedom,
i.e. 24 one-hour means minus 21 optimized parametres (the mesor, 10 amplitudes and
10 acrophases) were used.
Results
The primary Holter ECG evaluation had showed the relatively rare occurrence of simple
supraventricular or ventricular extrasystoles with very large intra- and interindividual
excursions as seen from the broad tolerance and confidence corridors. No
supraventricular extrasystoles within 24-hours had been found in 3 men (8.11%), no
ventricular ones in 9 men (24.32%). The maximum number of supraventricular beats
within 24-hours in one person had been 39, of ventricular ones 62. The total number of
supraventricular extrasystoles had been 527, of ventricular ones 458 in the whole
sample. The maximum number of supraventricular extrasystoles per hour and subject
had been between 2 and 6, that of ventricular ones between 2 and 9, the minimum
number had been zero. All extrasystoles had been asymptomatic. Abnormal findings
(atrial fibrillation or flutter, advanced AV block, sinus pauses, marked sinus bradycardia
under 40/min or complex ventricular arrhythmia) had never been seen.
The cosinor regression shows that the mesor values are significantly different from
zero in either kind of extrasystoles´ but higher for the hourly numbers of
supraventricular beats (0.58/h, with 95% confidence 0.56–0.60/h) than for ventricular
Acta Physiologica Hungarica 95, 2008
324
M Mikulecký and Š Kujaník
ones (0.52/h; 0.38–0.65/h). Nevertheless, the variance for ventricular extrasystoles is
more
than
30-times
higher
than
that
for
supraventricular
ones
(F=0.0412/0.0012=34.95334, P<<0.001).
The confidence interval does not overlap the abscissa for supraventricular
extrasystoles (Fig. 1), their significant presence is therefore documented during all 24
hours. The wide dispersion of ventricular extrasystoles results in such a wide corridor of
confidence that it overlaps the abscissa in all hours. The nonoverlapping time intervals
(significant presence of extrasystoles) are sparse, except the morning, noon and very
shortly late afternoon (Fig. 2). The 95% corridor for supraventricular extrasystoles is
completely merging into the ventricular ones (Fig. 3). Accordingly, there is no
statistically significant difference between the mean values of either kind.
Six 95% confidence intervals of amplitudes of supraventricular and one for
ventricular premature beats differ significantly from zero (Fig. 4), thus testifying to the
presence of the corresponding periodic components. There is a wide mutual overlapping
of the confidence intervals for the separate amplitudes between supraventricular and
ventricular extrasystoles. Accordingly, no statistically significant difference between
them, except the 24-hour amplitudes, was found.
Fig. 1. The 24-hour chronogram of supraventricular extrasystoles (SVEx/h) in 37 healthy male subjects. The
narrower corridor represents the 95% confidence, the wider one the 95% tolerance. Significant departures
from the mesor (M, horizontal line) are shadowed
Acta Physiologica Hungarica 95, 2008
Extrasystoles in healthy male persons at lowland
325
Fig. 2. Analogy of Fig. 1 for ventricular extrasystoles (Vex/h) in 37 healthy male subjects
Fig. 3. The 95% confidence corridors for either kind of extrasystoles in 37 healthy male subjects
Acta Physiologica Hungarica 95, 2008
326
M Mikulecký and Š Kujaník
Fig. 4. The amplitudes (A) of the mean hourly numbers of extrasystoles with 95% confidence intervals (bars,
means shown as the abscissa inside of them) for separate period lengths in hours (h). Significant amplitudes
are shown by heavy, nonsignificant by dashed lines. Significant difference between supraventricular (SV) and
ventricular (V) extrasystoles is shown by the asterisk
Table I shows the nonsignificant shifts of acrophases between ventricular and
supraventricular extrasystoles by 12 to 100 minutes. The significant 72-minutes’ shift in
6-hour rhythm of ventricular acrophase was found only.
Table I
Acrophases for the significant periodic components only in the healthy male subjects (n=37)
Type of
extrasystoles
SV
V
SV
SV
SV
SV
SV
Period length
(hours)
24
24
12
6
4
3.4
2.4
95% confidence (clock hours of the peaking times)
Mean
Lower bound
Upper bound
11:11
10:19
12:02
10:46
08:12
13:19
07:41
06:49
08:33
00:08
23:36
00:40
00:28
00:00
00:56
02:41
02:24
02:58
01:58
01:45
02:12
SV = supraventricular extrasystoles
V = ventricular extrasystoles
Only the times for the first peaks of each periodic component are given
Acta Physiologica Hungarica 95, 2008
Extrasystoles in healthy male persons at lowland
327
Discussion
No clinical examination used had been able to prove any cardiac pathology in our
subjects. Therefore, our sample may consist of “really healthy“ male subjects compared
to some other studies (e.g. 9, 13, 20, 22). In those studies were displayed also long
pauses, multiform extrasystoles, bigeminy or ventricular tachycardia in very old
persons. But no attempts to exclude silent myocardial ischaemia can be fully successful.
Only a few publications can be compared with the present one. A mere statistical
description of the given samples is usually presented by many authors. Moreover, only
the circadian rhythm is tested mostly in healthy subjects.
Kostis and coworkers (15, 16) had claimed 100 ventricular extrasystoles per 24
hours, i.e. 4.17 per hour as an upper limit of the normal values. According to this
criterion, all our subjects should be “overnormal“: nobody of them showed more than
62 ventricular extrasystoles per 24 hours, i.e. 2.58 per hour in average. Either of
extrasystole type had been in the range of 0 to 9 per hour what does not reach any
clinical importance. That should be so in typical healthy persons.
Several comparable studies in healthy subjects (9, 19, 20) or in apparently healthy
ones (4, 10, 14, 16, 22) have found both supraventricular and ventricular extrasystoles
similarly to the present study. Nevertheless, they had been processed without cosinor
regression up to the 10th harmonics of the circadian. One of the partially relevant papers
(21) found in very old subjects supraventricular extrasystoles peaking in the time
similar to our data.
Descriptively chronobiological analyses of extrasystoles are often mentioned in
connection with pathologic states. In patients with ischemic heart disease, the most
quiet interval of ventricular extrasystoles is between midnight and 2:00 (18). That is in
agreement with our healthy men shortly before midnight. The morning increase of
extrasystoles in many cardiovascular diseases after waking up starts suddenly in the
interval 6–7:00, the same way as in our subjects.
Exact scrutiny of the frequency of extrasystoles can be of critical importance in
clinical medicine. Frequent ventricular complexes can sometimes appear in apparently
healthy subjects (14, 22) as predictors of cardiovascular death or, in older subjects, as
predictors of acute myocardial infarction (22). Another study (10), however, did not
confirm such prediction.
The present statistical approach is based on W. S. Gosset’s (Student’s) revolution
in statistics (8) more than one century ago – from description (standard deviation of a
sample) towards induction (confidence for the corresponding population). It respects
the idea of J. H. Poincaré: science does foresee, and therefore it can be helpful in
practice. By the way, that had brought to Gosset the triumph in practice – to brew by
Guinness the world’s best beer.
The results obtained in the present contribution should depend upon the diurnal
variation in electrophysiological properties of the heart (1). The early morning increase
in sympathetic activity, catecholamine levels, blood pressure, the shortest ventricular
refractory periods are present already in healthy subjects. They can decrease the
Acta Physiologica Hungarica 95, 2008
328
M Mikulecký and Š Kujaník
electrical stability of the heart and support the origin of predominantly ventricular
extrasystoles. For example, the QT interval duration (a parameter measured frequently
for estimation of the electrical stability of the heart, being shorter for a lower stability)
is shorter during waking hours (2) because of longer cardiac cycle or RR interval during
sleep. Waking hours are the time of higher extrasystole occurrence in our healthy
persons, too.
Acknowledgements
The paper is supported by grant No. 1/0239/08 from the Slovak grant agency VEGA. We are grateful to
Assoc. Prof. Marian Snincak, MD, PhD, Juraj Podracky, MD, PhD, Marian Palinsky, MD, PhD, and Prof.
Juraj Koval, MD, PhD, for help in obtaining the primary ECG data.
REFERENCES
1. Behrens S, Franz MR: Circadian variation of arrhythmic events, electrophysiological properties, and the
autonomic nervous system. Eur. Heart J. 22, 2144–2146 (2001)
2. Bexton RS, Vallin HO, Camm AJ: Diurnal variation of the Q-T interval – influence of the autonomic
nervous system. Br. Heart J. 55, 253–258 (1986)
3. Bingham Ch, Arbogast B, Cornélissen GG, Lee JK, Halberg F: Inferential statistical methods for
estimating and comparing cosinor parameters. Chronobiologia 9, 397–439 (1982)
4. Bjerregaard P: Premature beats in healthy subjects 40–79 years of age. Eur. Heart J. 3, 493–503 (1982)
5. Bonnemeier H, Richardt G, Potratz J, Wiegand UK, Brandes A, Kluge N, Katus HA: Circadian profile of
cardiac autonomic nervous modulation in healthy subjects: differing effects of aging and gender on heart
rate variability. J. Cardiovasc. Electrophysiol. 14, 791–799 (2003)
6. Clarke JM, Hamer J, Shelton JR, Taylor S, Venning GR: The rhythm of the normal human heart. Lancet
1, 508–512 (1976)
7. de Leonardis V, De Scalzi M, Vergassola R, Romano S, Becucci A, Cinelli P: Circadian variations of
heart rate and premature beats in healthy subjects and in patients with previous myocardial infarction.
Chronobiol. Int. 4, 283–289 (1987)
8. Fedor-Freybergh PG, Mikulecky M: From the descriptive towards inferential statistics. Hundred years
since conception of the Student’s t-distribution. Neuroendocrinol. Lett. 26, 167–171 (2005)
9. Fleg JL, Kennedy HL: Cardiac arrhythmias in a healthy elderly population. Detection by 24-hour
ambulatory electrocardiography. Chest 81, 302–307 (1982)
10. Fleg JL, Kennedy HL: Long term prognostic significance of ambulatory electrocardiographic findings in
apparently healthy subjects greater than or equal to 60 yers of age. Am. J. Cardiol. 70, 748–751 (1992)
11. Garcia A, Valdes M, Sanchez V, Soria F, Vicente T, Apellaniz G, Perez F, Hernandez A: An analysis of
the circadian rhythm of the heart rate and arrhythmias in healthy elderly subjects. Rev. Espan. Cardiol.
45, 232–237 (1992)
12. Gardner MJ, Altman DG: Confidence intervals rather than P values: estimation rather than hypothesis
testing. Br. Med. J. 292, 746–750 (1986)
13. Glasser SP, Clark PI, Applebaum HJ: Occurrence of frequent complex arrhythmias detected by
ambulatory monitoring. Chest 75, 565–568 (1979)
14. Kennedy HL, Underhill SJ: Frequent or complex ventricular ectopy in apparently healthy subjects: a
clinical study of 25 cases. Am. J. Cardiol. 38, 141–148 (1976)
15. Kostis JB, Moreyra AE, Natarajan N, Gotzoyannis S, Hosler M, McCrone K, Kuo PT: Ambulatory
electrocardiography: What is normal? (abstract). Am. J. Cardiol. 43, 420 (1979)
Acta Physiologica Hungarica 95, 2008
Extrasystoles in healthy male persons at lowland
329
16. Kostis JB, McCrone K, Moreyra AE, Gotzoyannis S, Aglitz NM, Natarajan N, Kuo PT: Premature
ventricular complexes in the absence of identifiable heart disease. Circulation 63, 1351–1356 (1981)
17. Kubacek L, Valach A, Mikulecky M (2002): Time series analysis with periodic components. Software
manual. ComTel, Bratislava.
18. Lown B, Tykocinski M, Garfein A, Brooks P: Sleep and ventricular premature beats. Circulation 48,
691–701 (1973)
19. Orth-Gomer K, Hogstedt C, Bodin L, Soderholm B: Frequency of extrasystoles in healthy male
employees. Br. Heart J. 55, 259–264 (1986)
20. Rasmussen V, Jensen G, Schnohr P, Hansen JF: Premature ventricular beats in healthy adult subjects 20
to 79 years of age. Eur. Heart J. 6, 335–341 (1985)
21. Rossi A, Sforza Ch, Carandente F: Heart rate and extrasystolic arrhythmias in active subjects aged over
90. A chronobiological study. Chronobiologia 13, 309–318 (1986)
22. Sajadieh A, Nielsen OW, Rasmussen V, Hein HO, Frederiksen BS, Davanlou M, Hansen JF: Ventricular
arrhythmias and risk of death and acute myocardial infarction in apparently healthy subjects of age > or =
55 years. Am. J. Cardiol. 97, 1351–1357 (2006)
Acta Physiologica Hungarica 95, 2008