Heart rate variability decreased by coronary artery surgery has no
prognostic value
Short title: Low HRV and survival after CABG
Authors: Goran Milicevic, Ljubica Forta, Marcel Majseca and Vinko Bakula.
Affiliations of the authors: General Hospital Sveti Duh, Zagreb, and aHospital for
Medical Rehabilitation, Krapinske Toplice, Croatia.
Previous presentations: The work was presented in a part at the 3rd International
Congress of the Mediterranean Society of Pacing and Electrophysiology, Lisbon 2003.
Support: The study was supported from the Project 0219163 of the Ministry of Science
and Technology of the Republic of Croatia (Diagnosis and Therapy of Decreased Heart
Rate Variability).
Potential conflict of interest: NONE
Correspondence and reprints:
Asst. Prof. Goran Milicevic,
Division of Cardiology, Clinical Hospital Sveti Duh,
Sveti Duh 64, HR-10000 Zagreb, Croatia
fax: +385 13712308,
ph. +385 913712213,
e-mail. goran.milicevic1@zg.hinet.hr
Wordcount: 3287
Abstract
Background: Decreased heart rate variability (HRV) may predict cardiac death after
myocardial infarction (MI). Coronary artery bypass grafting (CABG) strongly decreases
HRV, but improves survival. The aim of the study was to determine the prognostic
value of HRV decreased by coronary surgery.
Design and Methods: Four-year follow-up was performed in 175 consecutive patients
with HRV decreased by CABG (51) or MI (124). Mortality and secondary events rate
were analyzed. Decreased HRV, defined by the standard deviation of mean RR interval
(SDNN) < 100 ms, was detected by a routine 24-hour Holter ECG at admission to
stationary rehabilitation 3 weeks to 3 months after acute MI or CABG. Two groups did
not differ except by age; CABG patients were younger (56 vs 64 years, p<0.01), but this
did not influence differences in survival (NS).
Results: HRV was lower among CABG patients than among MI patients (SDNN =
6620 ms vs 7714 ms; p<0.001), but cumulative survival and event-free survival were
much better in the CABG group than in the MI group. During a 4620 months followup, there were 10% new events in the CABG and 43% in the MI group (p<0.001).
Mortality was 8% in the CABG and 33% in the MI group (log-rank=3.6; p<0.001).
Unlike in the MI group, HRV was not different between survivors and non-survivors in
the CABG group.
Conclusions: In contrast to the strong prognostic potential of HRV in patients with MI,
decreased HRV has no prognostic significance in patients who have undergone CABG
surgery.
Abstract wordcount: 248
2
Key words: prognosis, survival, heart rate variability, coronary artery bypass grafting,
myocardial infarction
3
Introduction
Decreased heart rate variability (HRV) has been found to be an important predictor of
cardiac death in patients with myocardial infarction [1]. Measurement of HRV had
prognostic value in unselected older population [2] and attempts were made to involve
HRV in a routine clinical practice [3, 4]. However, it seems that HRV may not be useful
in all coronary disease settings [5, 6].
Coronary artery bypass grafting improves survival [7], but simultaneously
decreases HRV [8-11] even more than myocardial infarction does [12]. This may seem
contradictory to someone who uses low HRV as a prognostic marker of an increased
risk of cardiac death. Clinical practice shows that cardiac surgery patients with
extremely decreased HRV often have an excellent outcome. Stein et al. [5] have found
no prognostic significance for HRV decreased by cardiac surgery in a retrospective
study, but there is very little prospective data.
Therefore, the aim of this study was to define the prognostic significance of
decreased HRV in cardiac bypass surgery patients. For that purpose, we analysed the
survival of patients with HRV decreased in association with coronary artery bypass
grafting and compared it with the survival of patients whose HRV was decreased in
association with myocardial infarction.
Methods
Patients
Survival was prospectively analysed in 175 consecutive stationary cardiac rehabilitation
patients who were in sinus rhythm and with no sinus node disease or second or thirddegree atrioventricular block, but whose routine made 24-hour Holter electrocardiogram
4
showed decreased HRV. They had suffered myocardial infarction (MI; 124 patients) or
undergone coronary artery bypass grafting (CABG; 51 patients) 3 weeks to 3 months
before. Electrocardiogram and enzyme testing verified the diagnosis of MI. Cardiac
surgery was performed by the use of an extracorporal circulation machine. All CABG
patients received 3 or more bypass grafts. Patients with perioperative infarction and
those with urgent revascularization following myocardial infarction were not included,
nor were not those older than 79 years or those with diseases limiting survival or
affecting HRV (cancer, stroke, insulin-dependent diabetes, thyroid disease).
Survival of CABG patients was compared with that of MI patients because of
each group had the same underlying (coronary artery) disease. There was no significant
difference (NS) between the myocardial infarction and CABG groups for gender (71%
male), incidence of hypertension (24%), diabetes (19%), left ventricular systolic
dysfunction (34%) defined by ejection fraction lower than 40%, and time elapsed from
their MI (1.60.8 months) or CABG (1.50.9 months). There were no differences in the
use of medication (beta adrenergic blockers, digoxine, amiodaron, sotalol, propafenon,
mexiletine, calcium channels antagonists, angiotensin-converting enzyme inhibitors,
nitrates and diuretics) either. The only significant difference recorded was age; CABG
patients were younger (5611 vs 649 years, p<0.01).
Follow-up
As approved by the Hospital Ethics Committee, a questionnaire on the present health
status was mailed to patients. If there was no response, the family physician was
contacted. This study focused on combined cardiovascular and cerebrovascular
mortality as the primary endpoint and on both fatal and nonfatal events (myocardial
infarction, unstable angina, stroke and transient ischemic attacks) as the secondary
5
endpoint. Fife of 129 MI patients enrolled in the study were excluded because of a noncardiovascular cause of death (cancer). Those who underwent cardiac revascularization
procedures (coronary angioplasty or bypass surgery) during the follow-up were
considered alive at the date of intervention and excluded from further analysis.
Measurements
HRV was calculated from 24-hour Holter electrocardiogram, by a commercial system
(Oxford Instruments). R-R intervals that included ectopic beats were excluded and
extrapolated by linear interpolation. Spectral analysis was computed using fast Fourier
transformation. Ten-minute epochs were repeatedly transformed and averaged
throughout the 24-hour period. Details were published elsewhere [4]. Standard deviation
of mean RR interval (SDNN) was used as a representative of overall HRV. The cut-off
between the normal and moderately diminished HRV was defined arbitrarily by the
value of SDNN lower then 100 ms [1]. Most of the variables proposed by the Task
Force on the Heart Rate Variability [3] were analyzed. Left ventricular ejection fraction
was determined by a Simpson rule, from apical four- and two-chamber view.
Statistics
Differences in patient characteristics were analysed by Fisher exact test for categorical
variables and by Student’s t-test for continuous variables. Most of the HRV variables
analysed best fit a logarithmic distribution [4], so median values are given for
comparison between MI and CABG groups in Table 1. Median values of logarithmically
transformed variables were compared by the Mann-Whitney U-test. Except in Table 1,
all possible differences in the SDNN were analysed by t-test, because SDNN was
normally distributed [4].
6
Significance of differences in the rate of events or death between the two groups
was verified by Fisher exact tests. Time to death and time to predefined cardiovascular
events was analysed by the Kaplan-Meier method and compared with the log-rank test.
The effects of sex, age, time elapsed from MI or CABG, concomitant diseases, left
ventricular ejection fraction and medication on difference in the survival was analyzed
by proportional hazard (Cox) regression. SPSS for Windows, version 7.5 was used.
Results
Overall HRV was lower among CABG patients than among MI patients (SDNN =
6620 ms vs 7714 ms; p<0.001), independent of positive or negative history of MI
prior to coronary artery surgery (SDNN = 6821 ms vs 6421 ms; NS). In contrast to
overall HRV, the values of other HRV variables (rMSSD, pNN50, LF, HF, LF/HF ratio)
did not differ between the two groups (Table 1).
During the follow-up period (4620 months, range 0.03 to 77), 45 patients died
and 13 survived secondary cardiovascular or cerebrovascular event. Seventeen patients,
all from the MI group, underwent revascularisation procedures and sustained no further
secondary events. Cumulative survival and cumulative event-free survival were much
better in the CABG group than in the MI group. New events occurred in five (10%)
CABG patients and in 53 (43%) MI patients (p<0.001). Four (8%) CABG patients and
41 (33%) MI patients died (p<0.001). Kaplan-Meier curves depicting difference in
cumulative survival are shown in Figure 1, and the log-rank test confirmed the same
difference (p<0.001) for the survival and for the event-free survival.
The paradoxical finding of better survival in the group with worse HRV would
indicate that HRV has no prognostic value in CABG group or has no prognostic value at
7
all. The latter is not true because survival was strongly related to HRV in MI group.
Deceased MI patients had lower SDNN than those who survived the follow-up period
(6815 ms vs 8212 ms; p<0.001). The same difference was found for the event-free
survival in the MI group (6914 ms vs 8312 ms; p<0.001). That was not the case in the
CABG group. The difference in HRV between survivors (SDNN 6721 ms) and nonsurvivors (SDNN 5920 ms) did not reach statistical significance among the CABG
patients. The same difference was found for the event-free survival (SDNN was 6720
ms in survivors and 5822 ms in non-survivors; difference NS), confirming that HRV
has no prognostic value only in CABG patients.
Beside HRV, the two groups also differed according to age, however, the age
difference did not influence differences in survival (2 value was 0.24 for survival and
1.18 for event-free survival; NS both).
Discussion
This study clearly documents that the usual association between depressed HRV
and increased mortality does not exist in coronary artery disease patients who have
undergone bypass surgery. The seemingly contradictory findings of a simultaneous
decrease in HRV and improvement in survival by CABG indicate that the low HRV is
not an invariable predictor of an adverse outcome. Conversely, the prognostic value of a
decreased HRV was confirmed once more for MI patients.
Our results are consistent with data reported retrospectively from the CAST trial
by Stein and colleagues [5] where post-CABG patients had markedly decreased HRV in
conjunction with improved short-term survival. They also concluded that "recognition
8
of this is necessary to prevent misclassification of risk". Similar opinion on the value of
decreased HRV as a prognostic marker in the absence of myocardial infarction can be
found elsewhere in the literature [6], but the explanation for this opinion was not given.
A non-lethal decrease of HRV in CABG patients can be explained by the
increase of perioperative sympathetic activity and by a mechanism of autonomic
denervation different from that accompanying heart infarction, i.e., myocardial necrosis.
The decrease in HRV following CABG is related to increased sympathetic activity and
to perioperative procedures [9], so our findings are actually consistent with what might
be expected. It seems that the heart, denervated by a surgeon knife and simultaneously
stimulated by revascularisation, has a great potential of autonomic reinnervation. There
is evidence that HRV decreased in such a manner can recover completely over a period
of time [10]. Thus, it appears that a non-pathologic, postoperatively decreased HRV has
no importance for the routine clinical work and postoperative prognosis during, at least,
the first months following cardiac surgery.
An incidental finding of the study was that there were no differences in HRV
indices which reflect specific components of overall autonomic modulation of heart rate
(rMSSD, pNN50, LF, HF) between the CABG and MI groups. That could be the
consequence of a smaller perioperative decrease in these segments than in overall HRV
[10]. The lack of difference in HF, pNN50 and rMSSD could equally be due to the fact
that among cardiac patients there is a high prevalence of non-respiratory sinus
arrhythmia which exaggerates HRV in some patients and does not reflect cardiac
autonomic modulation. However, it seems that central frequency of the HRV spectrum
shifts leftward in association with progression in ischemic heart disease [13], and
probably, because of that, long-term measures of sympathovagal balance do not reflect
9
clinical finding in cardiac patients with low overall HRV [14]. We hope that our
ongoing study of perioperative changes of HRV will shed more light upon this
discrepancy in the changes of HRV elements.
Although not the primary focus of this study, the mortality rate in the MI group
(30% in 3 years) seems to be very high. That fact does not surprise us because in our
country myocardial infarction patients who undergo stationary cardiac rehabilitation are
generally at higher risk, and because only those with low HRV were included. A similar
mortality rate has been found in other myocardial infarction patients with decreased
HRV [1].
The limitation of this study is that groups selected by low HRV do not represent
all patients with MI or CABG, but our intention was just to analyse predictive value of
decreased HRV in CABG patients, not in a general cardiac population. From a
methodological stand this may not be the optimal approach. It remains possible that low
HRV after CABG would have a predictive value when compared to patients with
normal or high HRV after CABG, but survival in CABG patients with low HRV was
too good to compare it with that in patients with normal HRV. Finally, it would be
interesting to define predictive value of decreased HRV in diabetic patients because
Stein and co-workers [5] found in the CAST sample low predictive value of HRV not
only for CABG, but for diabetic patients as well. Unfortunately, our sample was too
small to test this hypothesis.
The study protocol covered cardiovascular events, but the mode of cardiac death
was not analysed, what might be considered as a limitation of the study as well. The
reason for that was the small number of deceased patients (only four) in the CABG
group. Due to technical limitations, we obtain no measurements of baroreflex sensitivity
10
[15, 16] or of other indexes of autonomic nervous system activity that have been
associated with increased mortality risk [17]. It is possible that measurement of these
parameters would have added insights into the previously mentioned difference in the
decrease of overall HRV compared with other HRV parameters that reflect activity of
different components of the autonomic nervous system.
It must be noted that there were mean differences in HRV post-CABG in
survivors and non-survivors, even though these were not significant. This suggests that
under ordinary clinical circumstances SDNN would not be useful in risk stratifying postCABG patients, but it also suggests that with a sufficient statistical power, i.e., with a
large enough sample, SDNN might be significantly different in non-survivors. Finally, it
would be interesting to define time-course of HRV recovery [18] in CABG patients by
repeated HRV measurement, but the study design was directed strictly to the survival.
In conclusion, no matter how strong is the prognostic potential of HRV in
patients with myocardial infarction, low HRV can not be used as a prognostic marker of
increased risk of cardiac death in patients shortly after coronary artery surgery.
11
References:
1. Kleiger RE, Miller JP, Bigger JT, Moss AJ, and the Multicenter Post-Infarction
Research Group. Decreased heart rate variability and its association with increased
mortality after acute myocardial infarction. Am J Cardiol 1987; 59:256-262.
2. Tsuji H, Venditti FJ, Manders ES, Evans JC, Larson MG, Feldman CL. Reduced
heart variability and mortality risk in an elderly cohort. The Framingham Heart
Study. Circulation 1994; 90:878-883.
3. Task Force of The European Society of Cardiology and The North American Society
of Pacing and Electrophysiology. Heart rate variability. Standards of measurement,
physiological interpretation, and clinical use. Circulation 1996; 93:1043-1065.
4. Milicevic G, Lakusic N, Szirovitza L, Cerovec D, Majsec M. Different cut-off level
of pathological heart rate variability in different groups of cardiac patients. J
Cardiovasc Risk 2001; 8:93-102.
5. Stein PK, Domitrovich PP, Kleiger RE, Schechtman KB, Rottman JN. Clinical and
demographic determinants of heart rate variability in patients post myocardial
infarction: insights from the cardiac arrhythmia suppression trial (CAST). Clin
Cardiol 2000; 23:187-194.
6. Schwartz PJ. The autonomic nervous system and sudden death. Eur Heart J 1998;
19 (suppl F):F72-80.
7. Holmes DR, Davis KB, Mock MB, Fisher LD, Gersh BJ, Killip T, et al. The effect
of medical and surgical treatment on subsequent sudden cardiac death in patients
with coronary artery disease: a report from the Coronary Artery Surgery Study.
Circulation 1986; 73:1254-1263.
12
8. Niemela MJ, Airaksinen KEJ, Tahvanainen KUO, Linnaluoto MK, Takkunen JT.
Effect of coronary artery bypass grafting on cardiac parasympathetic nervous
function. Eur Heart J 1992; 13:932-935.
9. Hogue CW, Stein PK, Apostolidou I, Lappas DG, Kleiger RE. Alterations in
temporal patterns of heart rate variability after coronary artery bypass graft surgery.
Anesthesiology 1994; 81:1356-1364.
10. Kuo C-D, Chen G-Y, Lai S-T, Wang Y-Y, Shih C-C, Wang J-H. Sequential
changes in heart rate variability after coronary artery bypass grafting. Am J Cardiol
1999; 83:776-779.
11. Laitio TT, Huikuri HV, Kentala ESH, Mäkikallio TH, Jalonen JR, Helenius H, et al.
Correlation properties and complexity of perioperative RR-interval dynamics in
coronary artery bypass surgery patients. Anesthesiology 2000; 93:69-80.
12. Milicevic G, Istvanovic N, Majsec M. Coronary artery bypass grafting reduces heart
rate variability more than myocardial infarction does. In Bloch Thomsen PE (editor)
Europace 2001. Bologna: Monduzzi Editore; 2001. pp. 355-358.
13. Miličević G, Lakušić N, Majsec M. Disease progression shifts heart rate variability
spectra leftward. Europace 2002; 3 (suppl A):A202.
14. Milicevic G, Lakusic N, Majsec M. Long-term measures of sympathovagal balance
do not reflect clinical finding in cardiac patients with low overall heart rate
variability. Ann Noninvasive Electrocardiology 2000; 5:S20.
15. Sleight P, La Rovere MT, Mortara A, Pinna G, Maestri R, Leuzzi S, et al.
Physiology and pathophysiology of heart rate and blood pressure variability in
humans: is power spectral analysis largely an index of baroreflex gain? Clin Sci
(Lond) 1995; 88:103-109.
13
16. De Ferrari GM, Landolina M, Mantica M, Manfredini R, Schwartz PJ, Lotto A.
Baroreflex sensitivity, but not heart rate variability, is reduced in patients with lifetreating ventricular arrhythmias long after myocardial infarction. Am Heart J 1995;
130:473-480.
17. Ghuran A, Reid F, La Rovere MT, Schmidt G, Bigger JT Jr, Camm AJ, Schwartz
PJ, et al. Heart rate turbulence-based predictors of fatal and nonfatal cardiac arrest
(The autonomic tone and reflexes after myocardial infarction substudy). Am J
Cardiol 2002; 89:184-190.
18. Doulalas AD, Flather MD, Pipilis A, Campbell S, Studart F, Rizos IK, et al.
Evolutionary pattern and prognostic importance of heart rate variability during the
early phase of acute myocardial infarction. Int J Cardiol 2001; 77:169-179.
14
Table 1. Median HRV values and differences between logarithmically transformed
values (in ms, except pNN50 in percents)
RR
SDNN
SDANNi
SDNNi
rMSSD
pNN50
TP
ULF
VLF
LF
HF
ratio
MI
785
79
69
31
20
1.4
1116
120
722
138
68
2.1
CABG
727
68
60
26
19
1.1
722
81
359
114
67
1.4
P
.004
<.001
<.001
<.001
.4
.4
<.001
<.001
<.001
.08
.4
.2
Legend: RR, mean of R-R intervals for normal beats, SDNN - standard deviation of all normal
R-R intervals, SDANNi, standard deviation of the 5-minute means of R-R intervals, SDNNi,
mean of the 5-minute standard deviations of RR intervals, rMSSD, square root of the mean of
the squared successive differences in R-R intervals, pNN50, percentage of R-R intervals that
are at least 50 ms different from the previous interval, TP, total power (0.0-0.5 Hz), ULF, ultra
low frequency (< 0.0033 Hz), VLF, very low frequency frequency (0.0033-0.04 Hz), LF, low
frequency (0.04-0.15 Hz), HF, high frequency (0.15-0.40 Hz), ratio, low to high frequency
ratio, MI, myocardial infarction group, CABG, coronary artery bypass grafting group.
15
Figure 1. Survival curves (Kaplan-Meier) of patients with HRV decreased by coronary
artery surgery or myocardial infarction.
16