- Email: [email protected]
Heart rate variability in healthy children and in those with congenital heart disease both before and after operation
Heart Rate Variability in Healthy Children and in Those With Congenital Heart Disease Both Before and After Operation Narakesari P. Heragu,
MBBS,
and William A. Scott,
MD
There are no reports of standard measures of heart rate variability (HRV) in pediatric patients with heart disease. Time domain (standard deviation of all normal RR intervals [SDNN], standard deviation of all 5-minute mean RR intervals, average standard deviation of all 5-minute RR intervals, and frequency domain (total, low- [LF], and high-frequency [HF] power) measures of HRV were (1) obtained in 45 healthy children, (2) compared between 36 children with congenital heart disease and agematched controls, (3) compared before and after surgery, and (4) compared between age-matched postoperative patients staying <7 days (group I, n 5 16) and those staying longer (group II, n 5 16). In healthy children, SDNN increased rapidly during infancy and more gradually thereafter, while the LH/HF ratio decreased until preschool age, with a later increase into adolescence. Compared with controls, preoperative patients had decreased total (53 6 55 vs 84 6 75 beats/min2/
Hz, p 5 0.01) and HF (12 6 14 vs 29 6 46 beats/min2/ Hz, p 5 0.03) power despite having similar heart rates. In the immediate postoperative period, all measures of HRV were decreased from preoperative values. Groups I and II did not differ in mean RR interval or HRV preoperatively; however, postoperatively, HRV was decreased in group II when compared with group I (SDNN 53 6 17 vs 40 6 14 ms, p 5 0.01), although the mean RR interval remained comparable (499 6 81 vs 481 6 62 ms, p 5 0.3). It is concluded that (1) there are significant age-related changes in HRV in healthy children, (2) preoperatively, children with congenital heart disease have reduced total and HF power when compared with healthy controls, (3) HRV is further reduced postoperatively in all patients, and (4) prolonged postoperative hospitalization is associated with a greater reduction in HRV. Q1999 by Excerpta Medica, Inc. (Am J Cardiol 1999;83:1654 –1657)
ecreased heart rate variability (HRV) is an independent predictor of increased morbidity and D mortality in adults with various forms of heart disease
come. Our hypotheses were that (1) there are agerelated changes in HRV in children, (2) HRV is decreased in children with congenital heart disease, and (3) HRV provides additional information about patient status after congenital heart disease surgery.
including myocardial infarction, coronary heart disease, congestive heart failure, and chronic mitral regurgitation.1– 6 There have been very few reports of HRV trends throughout the pediatric age range. The techniques used have varied and normal ranges have not been established.7–10 Few studies involve children older than the neonatal age group, and the reported values for HRV have varied widely. There have been only 2 reports on HRV in pediatric heart disease, neither of which used currently standard analysis techniques. Gordon et al11 found that decreased HRV measured as a “respiratory quotient” ,2.0 was associated with increased mortality in critically ill pediatric cardiac patients. In another report based on 10 patients, sinus arrhythmia increased after surgery for atrial septal defect.12 The goals of this study were (1) to contribute to normative HRV data in healthy children, (2) to compare HRV in preoperative pediatric cardiac patients with that in normal children, and (3) to evaluate the association between HRV in the immediate postoperative period and measures of outFrom the Department of Pediatrics, Section on Cardiology, University of Texas Southwestern Medical Center, Dallas, Texas. Manuscript received October 15, 1998; revised manuscript received and accepted February 4, 1999. Address for reprints: William A. Scott, MD, Section on Cardiology, University of Texas Southwestern Medical Center, 5323 Harry Hines Boulevard, Dallas, Texas 75235-9063.
1654
©1999 by Excerpta Medica, Inc. All rights reserved.
METHODS The Institutional Review Board of the University of Texas Southwestern Medical Center in Dallas, Texas, approved the study protocol. Informed consent was obtained from the legal guardians of all subjects. HRV data were obtained from 45 healthy children, and 36 children with congenital heart disease, who underwent elective cardiac surgery at the Children’s Medical Center, in Dallas, Texas. Patients with arrhythmias or pacemakers were excluded. Clinical data including age, diagnoses, presence of significant congestive heart failure, preoperative medications, type of surgical procedure, and cardiopulmonary bypass time were collected for all patients from medical records. Postoperative patient data included duration of mechanical ventilation, inotropic support, intensive care, pleural drainage, and hospitalization. Electrocardiographic data were obtained with Zymed (model 1100-010, Camarillo, California) 3-channel ambulatory electrocardiographic monitors on patients and controls. Each patient had another recording at least once during the postoperative hospital stay. Among the patients, 32 who could be agematched were subdivided into those who had a satisfactory postoperative course (group I), defined as hos0002-9149/99/$–see front matter PII S0002-9149(99)00173-3
TABLE I Age-Related Changes in HRV Age (yr) and No. Mean RR (ms) SDNN (ms) 0–0.4 (8) 0.4–2 (8) 2–5 (10) 6–11 (11) 11–17 (8)
416 479 601 657 730
6 6 6 6 6
21* 59* 51* 42* 76*
45 70 109 117 136
6 6 6 6 6
7* 21* 28‡ 16 24
SDNN/RR 0.11 0.14 0.18 0.17 0.19
6 6 6 6 6
SDANN (ms) ASDNN (ms) LF (beats/min2) HF (beats/min2)
0.01* 36 0.03* 58 0.04 88 0.02 98 0.02 114
6 6 6 6 6
6* 16* 23‡ 17 25
25 34 61 59 70
6 6 6 6 6
3 17* 25 8* 10
19 12 10 14 14
6 6 6 6 6
20 10 4 15 18
8 22 64 17 12
6 6 6 6 6
†
7 34 70‡ 12 10
LF/HF 2.7 2.1 0.5 0.8 1.6
6 6 6 6 6
1.6† 2.3 0.5‡ 0.6 1.6
*p ,0.05 versus the immediately older age group; † p ,0.05 versus 2- to 5-year age group; ‡ p ,0.05 versus 11- to 17-year age group.
pital stay ,7 days, and those who had a suboptimal result (group II) and stayed $7 days. To avoid the influence of inotropic agents and artificial ventilation, the postoperative recordings were obtained after both were discontinued. The timing of the postoperative recording was later in group II than in group I (9.2 6 4.5 vs 2.5 6 1.3 days). Heart rate variability analysis: The electrocardiographic recordings were analyzed on the Zymed 1410 scanner. The raw electrocardiographic data was digitized at a sampling rate of 192 Hz. Abnormal beats, significant pauses, and areas of artifact were automatically (using a computerized algorithm) and later manually identified and rejected. Any recording with significant (.5% of recorded beats) arrhythmia or artifact, or with ,14 hours of data was excluded. The average duration of data acquired was 21.1 hours. A normal-normal (NN) interval data file was created after removing abnormal beats with linear interpolation. Time domain measures of HRV were calculated for the entire duration of each study, and included: standard deviation of all analyzed NN intervals (SDNN); standard deviation of the 5-minute averages of NN intervals (SDANN); the average of all 5-minute standard deviations of NN intervals (ASDNN); and SDNN/RR quotient of SDNN and mean RR interval over the period of recording. Frequency domain analysis was performed on 256-second segments of data selected at the time of lowest heart rate during the entire recording. A segment free of abnormal beats was chosen. The NN interval-time series was sampled at 4 cycles/s and fast-Fourier transform used to derive a power spectral density graph. HRV power at various predetermined frequency intervals was calculated as area under the curve. These included: total power 5 power of entire area under power spectral density graph, expressed as beats/ min2; low-frequency (LF) power 5 area under power spectral density graph in the frequency range of 0.04 to 0.15; high-frequency (HF) power 5 area under power spectral density graph in the frequency range of 0.15 to 0.40; and LF/HF ratio 5 ratio of LF to HF power. Statistical methods: Paired and unpaired Student’s t tests were used for normally distributed data and Mann-Whitney rank-sum for data that failed tests of normality; regression analysis was used to establish correlations. All values are expressed as mean 6 SD.
FIGURE 1. The mean RR interval versus SDNN in healthy children with a best fit regression line showing a strong linear correlation. Adjusted R2 5 0.85, p <0.01.
RESULTS
Healthy controls: Normative data from 45 healthy children, aged 3 weeks to 16 years are listed in Table I. SDNN increased rapidly in the first 2 years of life and later more gradually to reach adult values by adolescence. Similar changes were seen in SDANN and ASDNN. The time domain measures correlated well with each other (e.g., SDNN vs SDANN, R2 5 0.94, p ,0.001). SDNN had a strong correlation with mean RR intervals (Figure 1). Hence, the ratio SDNN/ mean RR interval remained constant over most of the pediatric age range and did not change significantly over the age of 2 years. Among frequency domain measures, there was an increase in the absolute values of HF power from early infancy to preschool age, with a later decrease into adolescence. LF power tended to change in the opposite direction, although these changes were not statistically significant. This resulted in a decrease in the LF/HF ratio from early infancy to the preschool age and a later increase into adolescence. Preoperative congenital heart disease: Among the 36 patients aged 2 weeks to 15 years, 19 had acyanotic heart disease, 18 of them with left-to-right shunts.
CONGENITAL HEART DISEASE/AGE– AND DISEASE–RELATED CHANGES
1655
TABLE II Comparison of HRV in Patient Subjects and AgeMatched Controls Variable Mean RR (ms) SDNN (ms) SDNN/RR SDANN (ms) ASDNN (ms) TP (beats/min2) LF (beats/min2) HF (beats/min2) LF/HF
Controls (n 5 36) 563 93 0.16 77 48 84 12 29 1.6
6 6 6 6 6 6 6 6 6
125 41 0.04 35 24 75 12 46 1.7
Preoperative (n 5 36) 571 90 0.15 79 42 53 11 12 1.7
6 6 6 6 6 6 6 6 6
102 42 0.05 38 21 55* 10 14* 1.6
Postoperative (n 5 36) 517 48 0.09 43 20 23 5 2 3.7
6 6 6 6 6 6 6 6 6
89† 22† 0.03† 20† 12† 27† 6† 3† 3.4†
p ,0.01), were on mechanical ventilation longer (mean 3.6 vs 0.8 days, p ,0.01) and had longer duration of pleural drainage (mean 9.2 vs 1.7 days, p ,0.001).
DISCUSSION
Healthy controls: Time domain measures of HRV increase with age throughout the pediatric age range, achieving adult values by adolescence.13 Our findings were similar to those of Massin and von Bernuth9 but at variance with Goto et al8 who reported that SDNN increased until age 9 and later decreased into adolescence. The differences in observations are most likely *p ,0.05 between controls and preoperative patients; related to shorter durations of recordings in the latter † p ,0.01 between preoperative and postoperative values. study. The finding of strong linear correlation of time domain measures with mean RR intervals is also similar to that of the report by Massin TABLE III Comparison of HRV Between Patient Groups and von Bernuth.9 As mean RR inPreoperative Postoperative terval varies significantly with age, Group I Group II Group I Group II use of a method of normalization for Variable (n 5 16) (n 5 16) (n 5 16) (n 5 16) RR interval would be useful to compare time domain values across difMean RR (ms) 515 6 104 509 6 110 499 6 81 481 6 62 SDNN (ms) 73 6 39 72 6 46 53 6 17 38 6 14* ferent ages. We found that the quoSDNN/RR 0.13 6 0.05 0.13 6 0.06 0.11 6 0.01 0.08 6 0.02* tient of SDNN and mean RR interSDANN (ms) 63 6 34 64 6 44 46 6 14 36 6 14* vals tended to remain stable across ASDNN (ms) 37 6 23 32 6 22 24 6 13 14 6 5* most of the pediatric age range. TP (beats/min2) 54 6 70 42 6 43 36 6 34 22 6 29 In the frequency domain our findLF (beats/min2) 8.9 6 7.4 8.4 6 8.2 6.9 6 7.4 4.3 6 6.4 HF (beats/min2) 8.7 6 11.4 6.5 6 8.7 2.9 6 3.7 1.2 6 1.6 ings were similar to those of Finley LF/HF 2.3 6 1.7 1.9 6 1.7 4.2 6 4.0 3.6 6 2.8 and Nugent10 where the LF/HF ratio decreased in early childhood, fol*p ,0.05 between groups I and II. lowed by a later increase. However, unlike their report, absolute values of LF and HF power tended to change Cyanotic lesions were present in 17: 4 with tetralogy in opposite directions in our subjects, with HF power of Fallot and 17 with complex lesions. Table II shows increasing in early childhood and showing a later the comparison with healthy age-matched controls. decline into adolescence. Again, these differences are Time domain measures of HRV were not significantly most likely related to the duration of electrocardiodifferent. Patients had significantly lower total and HF graphic recordings. Furlan et al14 derived frequency power, whereas LF power and LF/HF ratio were not domain values for HRV using an averaging technique spanning 6-hour segments of data, which could obsignificantly different. Postoperative congenital heart disease: HRV re- scure detailed information about autonomic modulacordings were obtained an average of 5.8 days after tion of HRV. This observation and concern about the surgery. For the entire group of 36 patients, the mean lack of “stationarity” of the signal partly underlie the RR interval, and all time and frequency domain mea- recommendations of the Task Force of the European sures of HRV were significantly lower in the postop- Society of Cardiology and the North American Socierative period than preoperatively (Table III). The ety of Pacing and Electrophysiology to preferably use magnitude of decrease was greater for HF than for LF of short-term recordings for these derivations.15 We selected stable segments of sinus rhythm at the lowest power, resulting in an increase in LF/HF ratio. There were no significant preoperative differences mean heart rate in an attempt to standardize condiin mean RR interval, time or frequency domain mea- tions. The continuous supine position or controlled sures of HRV between groups I and II. Postopera- ventilation, as is used in adults, was not possible in our tively, there was no significant difference in heart rate; younger population. Preoperative congenital heart disease: HF power however, group II had significantly decreased HRV as measured in the time domain. The differences in has been shown to correlate with parasympathetic SDNN persisted even when adjusted for the RR inter- modulation.6,16 –18 The finding of decreased HF power val (SDNN/RR quotient). Although all of the fre- in children with congenital heart disease before surquency domain measures were decreased in group II, gery suggests relative parasympathetic withdrawal. these did not reach statistical significance. Subjects in The observed decrease in total power was only partly group II had significantly longer cardiopulmonary by- contributed to by a decrease in HF power. This sugpass times (mean 92.5 vs 52.5 minutes, p 5 0.02) and gests that there could be alterations in other factors length of stay in intensive care (mean 9.4 vs 1.8 days, thought to contribute to HRV including sympathetic 1656 THE AMERICAN JOURNAL OF CARDIOLOGYT
VOL. 83
JUNE 15, 1999
tone and the renin-angiotensin and thermoregulatory systems.16 These findings are in agreement with the hypothesis that decreased HRV is a nonspecific marker of cardiovascular stress, reflecting an alteration in autonomic nervous system input to the heart.15 Postoperative congenital heart disease: The significant decrease in HRV and HF power in the immediate postoperative period suggests that further parasympathetic withdrawal is an important component of these changes. This could partly be the result of decreased sensitivity of the sinus node to autonomic nervous input immediately after surgery for congenital heart disease. None of these patients was taking b blockers or any specific antiarrhythmic medications, which are known to alter HRV. Reduced HRV, independent of heart rate, in subjects who had a prolonged postoperative course suggests that HRV provides additional information on the well being of such patients. Whereas all measures of HRV reflected this, only long-term differences in time domain measures reached statistical significance. It has been observed that the values of these measures of HRV increase with increasing duration of recording.15,19 These differences in time domain measures could reflect differences in the renin-angiotensin system.16 Importantly, pleural drainage was longer in subjects who had longer lengths of hospital stay. Higher renin and angiotensin levels have been found in postoperative pediatric cardiac patients who had longer duration of pleural drainage.20,21 The difference in timing of the postoperative analyses between the 2 groups was to ensure that all subjects were in a similar stable hemodynamic state and off all vasoactive medications. Because the recordings were done later in group II, it would be expected that the differences in HRV would be reduced. Hence, the observed differences in HRV between these 2 patient groups assumes more significance. In this study it was not possible to control for differences in diagnoses and type of surgery because of the small number of subjects in each diagnostic category; however, these groups were apparently homogenous from a heart rate control or autonomic standpoint, because there were no preoperative differences in heart rate or HRV. In the only other report on HRV in perioperative critically ill pediatric cardiac patients, Gordon et al11 showed that children who died were more likely to have a “lower respiratory quotient” i.e., a higher respiratory-to-lower frequency power ratio. This was interpreted as being due to a relative increase in parasympathetic tone and possibly sympathetic exhaustion.11 Data were gathered in 2 centers using different analytic methods, standard measures of HRV were not used, and the duration of recording varied from 1 to 27 hours. Significant (38% to 62%) and possibly important segments were excluded from analysis. Most of their patients were likely to have been receiving vasopressors and artificial ventilation, both of which
affect HRV. Our data represent a more complete evaluation of longer term stable recordings and thus can not be compared with data from that report. Acknowledgment: We are grateful to Cathy Tipton, RCDS, Mary Drury, RN, RNP, Susan Stollzman, RN, RNP, Claudio Ramaciotti, MD, and Lynn Mahony, MD, for their assistance.
1. Kleiger RE, Miller JP, Bigger JT, Moss AJ, and the Multicenter Post-Infarction Research Group. Decreased HRV and its association with increased mortality after acute myocardial infarction. Am J Cardiol 1987;59:256 –262. 2. Bigger JT, Fleiss JL, Steinman RC, Ronitzky LM, Kleiger RE, Rottman JN. Frequency domain measures of heart period variability and mortality after myocardial infarction. Circulation 1992;85:164 –171. 3. Rich MW, Saini JS, Kleiger RE, Carney RM, teVelde A, Freedland KE. Correlation of HRV with clinical and angiographic variables and late mortality after coronary angiography. Am J Cardiol 1988;62:714 –717. 4. Brouwer J, van Veldhuisen DJ, Man in’t Veld AJ, Haaksma J, Dijk A, Visser KR, Boomsma F, Dunselman PHJM, Lie KI, for the Dutch Ibopamine Multicenter Trial Study Group. Prognostic value of HRV during long-term follow-up in patients with mild to moderate heart failure. J Am Coll Cardiol 1996;28:1183– 1189. 5. Stein KM, Bores JS, Hochreites C, Okin PM, Herrold EM, Devereux RB, Kligfield P. Prognostic value and physiologic correlates of HRV in chronic severe mitral regurgitation. Circulation 1993;88:127–135. 6. Binkley PF, Nunziata E, Haas GJ, Nelson SD, Cody RJ. Parasympathetic withdrawal is an integral component of autonomic imbalance in congestive heart failure: demonstration in human subjects and verification in a paced canine model of ventricular failure. J Am Coll Cardiol 1991;18:464 – 472. 7. Schechtman VL, Harper RM, Kluge KA. Development of heart rate variation over the first six months of life in normal infants. Pediatr Res 1989;26:343–346. 8. Goto M, Nagashima M, Baba R, Nagano Y, Yokota M, Nishibata K, Tsuji A. Analysis of HRV demonstrates effects of development on vagal modulation of heart rate in healthy children. J Pediatr 1997;130:725–729. 9. Massin M, von Bernuth G. Normal ranges of HRV during infancy and childhood. Pediatr Cardiol 1997;18:297–302. 10. Finley J, Nugent ST. HRV in infants, children and young adults. J Auton Nerv Syst 1995;51:103–108. 11. Gordon D, Herrera VL, McAlpine L, Cohen RJ, Akselrod S, Lang P, Norwood W. Heart-rate spectral analysis: a noninvasive probe of cardiovascular regulation in critically ill children with heart disease. Pediatr Cardiol 1988;9: 69 –77. 12. Finley JP, Nugent ST, Hellenbrand W, Craig M, Gillis DA. Sinus arrhythmia in children with atrial septal defect: an analysis of HRV before and after surgical repair. Br Heart J 1989;61:280 –284. 13. Umetani K, Singer DH, McCraty R, Atkinson M. Twenty-four hour time domain HRV and heart rate: relations to age and gender over nine decades. J Am Coll Cardiol 1998;31:593– 601. 14. Furlan R, Guzetti S, Crivellaro W, Dassi S, Tinelli M, Baselli G, Cerutti S, Lombardi F, Pagani M, Malliani A. Continuous 24-hour assessment of neural regulation of systemic arterial pressure and RR variabilities in ambulant subjects. Circulation 1990;81:537–547. 15. Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology. HRV: standards of measurement, physiological interpretation, and clinical use. Circulation 1996;93:1043–1065. 16. Askelrod S, Gordon D, Ubel FA. Shannon DC, Barger AC, Cohen RJ. Power spectrum analysis of heart rate fluctuation: a quantitative probe of beat to beat cardiovascular control. Science 1981;213:220 –222. 17. Pagani M, Lombardi F, Guzzetti S, Rimoldi O, Furlan R, Tizzi Nelli P, Sandrone G, Malfatto G, Dell’Orto S, Piccaluga E, Turiel M, Baselli G, Cerutti S, Malliani A. Power spectral analysis of heart rate and blood pressure variabilities as a marker of sympathovagal interaction in man and conscious dog. Circ Res 1986;59:178 –193. 18. Malliani A, Pagani M, Lombardi F, Cerutti S. Cardiovascular neural regulation explored in the frequency domain. Circulation 1991;84:1482–1492. 19. Saul JP, Albrecht P, Berger RD, Cohen RJ. Analysis of long-term HRV: methods, 1/f scaling and implications. In: Computers in Cardiology 1987. Washington, DC: IEEE Comput Soc Press, 1988:419 – 422. 20. Mainwaring RD, Lamberti JJ, Moore JW, Billman GF, Nelson JC. Comparison of the hormonal response after bidirectional Glenn and Fontan procedures. Ann Thorac Surg 1994;57:59 – 64. 21. Mainwaring RD, Lamberti JJ, Carter TL Jr, Moore JW, Nelson JC. Renin, angiotensin II and the development of effusions following bidirectional Glenn and Fontan procedures. J Card Surg 1995;10:111–118.
CONGENITAL HEART DISEASE/AGE– AND DISEASE–RELATED CHANGES
1657
MBBS,
and William A. Scott,
MD
There are no reports of standard measures of heart rate variability (HRV) in pediatric patients with heart disease. Time domain (standard deviation of all normal RR intervals [SDNN], standard deviation of all 5-minute mean RR intervals, average standard deviation of all 5-minute RR intervals, and frequency domain (total, low- [LF], and high-frequency [HF] power) measures of HRV were (1) obtained in 45 healthy children, (2) compared between 36 children with congenital heart disease and agematched controls, (3) compared before and after surgery, and (4) compared between age-matched postoperative patients staying <7 days (group I, n 5 16) and those staying longer (group II, n 5 16). In healthy children, SDNN increased rapidly during infancy and more gradually thereafter, while the LH/HF ratio decreased until preschool age, with a later increase into adolescence. Compared with controls, preoperative patients had decreased total (53 6 55 vs 84 6 75 beats/min2/
Hz, p 5 0.01) and HF (12 6 14 vs 29 6 46 beats/min2/ Hz, p 5 0.03) power despite having similar heart rates. In the immediate postoperative period, all measures of HRV were decreased from preoperative values. Groups I and II did not differ in mean RR interval or HRV preoperatively; however, postoperatively, HRV was decreased in group II when compared with group I (SDNN 53 6 17 vs 40 6 14 ms, p 5 0.01), although the mean RR interval remained comparable (499 6 81 vs 481 6 62 ms, p 5 0.3). It is concluded that (1) there are significant age-related changes in HRV in healthy children, (2) preoperatively, children with congenital heart disease have reduced total and HF power when compared with healthy controls, (3) HRV is further reduced postoperatively in all patients, and (4) prolonged postoperative hospitalization is associated with a greater reduction in HRV. Q1999 by Excerpta Medica, Inc. (Am J Cardiol 1999;83:1654 –1657)
ecreased heart rate variability (HRV) is an independent predictor of increased morbidity and D mortality in adults with various forms of heart disease
come. Our hypotheses were that (1) there are agerelated changes in HRV in children, (2) HRV is decreased in children with congenital heart disease, and (3) HRV provides additional information about patient status after congenital heart disease surgery.
including myocardial infarction, coronary heart disease, congestive heart failure, and chronic mitral regurgitation.1– 6 There have been very few reports of HRV trends throughout the pediatric age range. The techniques used have varied and normal ranges have not been established.7–10 Few studies involve children older than the neonatal age group, and the reported values for HRV have varied widely. There have been only 2 reports on HRV in pediatric heart disease, neither of which used currently standard analysis techniques. Gordon et al11 found that decreased HRV measured as a “respiratory quotient” ,2.0 was associated with increased mortality in critically ill pediatric cardiac patients. In another report based on 10 patients, sinus arrhythmia increased after surgery for atrial septal defect.12 The goals of this study were (1) to contribute to normative HRV data in healthy children, (2) to compare HRV in preoperative pediatric cardiac patients with that in normal children, and (3) to evaluate the association between HRV in the immediate postoperative period and measures of outFrom the Department of Pediatrics, Section on Cardiology, University of Texas Southwestern Medical Center, Dallas, Texas. Manuscript received October 15, 1998; revised manuscript received and accepted February 4, 1999. Address for reprints: William A. Scott, MD, Section on Cardiology, University of Texas Southwestern Medical Center, 5323 Harry Hines Boulevard, Dallas, Texas 75235-9063.
1654
©1999 by Excerpta Medica, Inc. All rights reserved.
METHODS The Institutional Review Board of the University of Texas Southwestern Medical Center in Dallas, Texas, approved the study protocol. Informed consent was obtained from the legal guardians of all subjects. HRV data were obtained from 45 healthy children, and 36 children with congenital heart disease, who underwent elective cardiac surgery at the Children’s Medical Center, in Dallas, Texas. Patients with arrhythmias or pacemakers were excluded. Clinical data including age, diagnoses, presence of significant congestive heart failure, preoperative medications, type of surgical procedure, and cardiopulmonary bypass time were collected for all patients from medical records. Postoperative patient data included duration of mechanical ventilation, inotropic support, intensive care, pleural drainage, and hospitalization. Electrocardiographic data were obtained with Zymed (model 1100-010, Camarillo, California) 3-channel ambulatory electrocardiographic monitors on patients and controls. Each patient had another recording at least once during the postoperative hospital stay. Among the patients, 32 who could be agematched were subdivided into those who had a satisfactory postoperative course (group I), defined as hos0002-9149/99/$–see front matter PII S0002-9149(99)00173-3
TABLE I Age-Related Changes in HRV Age (yr) and No. Mean RR (ms) SDNN (ms) 0–0.4 (8) 0.4–2 (8) 2–5 (10) 6–11 (11) 11–17 (8)
416 479 601 657 730
6 6 6 6 6
21* 59* 51* 42* 76*
45 70 109 117 136
6 6 6 6 6
7* 21* 28‡ 16 24
SDNN/RR 0.11 0.14 0.18 0.17 0.19
6 6 6 6 6
SDANN (ms) ASDNN (ms) LF (beats/min2) HF (beats/min2)
0.01* 36 0.03* 58 0.04 88 0.02 98 0.02 114
6 6 6 6 6
6* 16* 23‡ 17 25
25 34 61 59 70
6 6 6 6 6
3 17* 25 8* 10
19 12 10 14 14
6 6 6 6 6
20 10 4 15 18
8 22 64 17 12
6 6 6 6 6
†
7 34 70‡ 12 10
LF/HF 2.7 2.1 0.5 0.8 1.6
6 6 6 6 6
1.6† 2.3 0.5‡ 0.6 1.6
*p ,0.05 versus the immediately older age group; † p ,0.05 versus 2- to 5-year age group; ‡ p ,0.05 versus 11- to 17-year age group.
pital stay ,7 days, and those who had a suboptimal result (group II) and stayed $7 days. To avoid the influence of inotropic agents and artificial ventilation, the postoperative recordings were obtained after both were discontinued. The timing of the postoperative recording was later in group II than in group I (9.2 6 4.5 vs 2.5 6 1.3 days). Heart rate variability analysis: The electrocardiographic recordings were analyzed on the Zymed 1410 scanner. The raw electrocardiographic data was digitized at a sampling rate of 192 Hz. Abnormal beats, significant pauses, and areas of artifact were automatically (using a computerized algorithm) and later manually identified and rejected. Any recording with significant (.5% of recorded beats) arrhythmia or artifact, or with ,14 hours of data was excluded. The average duration of data acquired was 21.1 hours. A normal-normal (NN) interval data file was created after removing abnormal beats with linear interpolation. Time domain measures of HRV were calculated for the entire duration of each study, and included: standard deviation of all analyzed NN intervals (SDNN); standard deviation of the 5-minute averages of NN intervals (SDANN); the average of all 5-minute standard deviations of NN intervals (ASDNN); and SDNN/RR quotient of SDNN and mean RR interval over the period of recording. Frequency domain analysis was performed on 256-second segments of data selected at the time of lowest heart rate during the entire recording. A segment free of abnormal beats was chosen. The NN interval-time series was sampled at 4 cycles/s and fast-Fourier transform used to derive a power spectral density graph. HRV power at various predetermined frequency intervals was calculated as area under the curve. These included: total power 5 power of entire area under power spectral density graph, expressed as beats/ min2; low-frequency (LF) power 5 area under power spectral density graph in the frequency range of 0.04 to 0.15; high-frequency (HF) power 5 area under power spectral density graph in the frequency range of 0.15 to 0.40; and LF/HF ratio 5 ratio of LF to HF power. Statistical methods: Paired and unpaired Student’s t tests were used for normally distributed data and Mann-Whitney rank-sum for data that failed tests of normality; regression analysis was used to establish correlations. All values are expressed as mean 6 SD.
FIGURE 1. The mean RR interval versus SDNN in healthy children with a best fit regression line showing a strong linear correlation. Adjusted R2 5 0.85, p <0.01.
RESULTS
Healthy controls: Normative data from 45 healthy children, aged 3 weeks to 16 years are listed in Table I. SDNN increased rapidly in the first 2 years of life and later more gradually to reach adult values by adolescence. Similar changes were seen in SDANN and ASDNN. The time domain measures correlated well with each other (e.g., SDNN vs SDANN, R2 5 0.94, p ,0.001). SDNN had a strong correlation with mean RR intervals (Figure 1). Hence, the ratio SDNN/ mean RR interval remained constant over most of the pediatric age range and did not change significantly over the age of 2 years. Among frequency domain measures, there was an increase in the absolute values of HF power from early infancy to preschool age, with a later decrease into adolescence. LF power tended to change in the opposite direction, although these changes were not statistically significant. This resulted in a decrease in the LF/HF ratio from early infancy to the preschool age and a later increase into adolescence. Preoperative congenital heart disease: Among the 36 patients aged 2 weeks to 15 years, 19 had acyanotic heart disease, 18 of them with left-to-right shunts.
CONGENITAL HEART DISEASE/AGE– AND DISEASE–RELATED CHANGES
1655
TABLE II Comparison of HRV in Patient Subjects and AgeMatched Controls Variable Mean RR (ms) SDNN (ms) SDNN/RR SDANN (ms) ASDNN (ms) TP (beats/min2) LF (beats/min2) HF (beats/min2) LF/HF
Controls (n 5 36) 563 93 0.16 77 48 84 12 29 1.6
6 6 6 6 6 6 6 6 6
125 41 0.04 35 24 75 12 46 1.7
Preoperative (n 5 36) 571 90 0.15 79 42 53 11 12 1.7
6 6 6 6 6 6 6 6 6
102 42 0.05 38 21 55* 10 14* 1.6
Postoperative (n 5 36) 517 48 0.09 43 20 23 5 2 3.7
6 6 6 6 6 6 6 6 6
89† 22† 0.03† 20† 12† 27† 6† 3† 3.4†
p ,0.01), were on mechanical ventilation longer (mean 3.6 vs 0.8 days, p ,0.01) and had longer duration of pleural drainage (mean 9.2 vs 1.7 days, p ,0.001).
DISCUSSION
Healthy controls: Time domain measures of HRV increase with age throughout the pediatric age range, achieving adult values by adolescence.13 Our findings were similar to those of Massin and von Bernuth9 but at variance with Goto et al8 who reported that SDNN increased until age 9 and later decreased into adolescence. The differences in observations are most likely *p ,0.05 between controls and preoperative patients; related to shorter durations of recordings in the latter † p ,0.01 between preoperative and postoperative values. study. The finding of strong linear correlation of time domain measures with mean RR intervals is also similar to that of the report by Massin TABLE III Comparison of HRV Between Patient Groups and von Bernuth.9 As mean RR inPreoperative Postoperative terval varies significantly with age, Group I Group II Group I Group II use of a method of normalization for Variable (n 5 16) (n 5 16) (n 5 16) (n 5 16) RR interval would be useful to compare time domain values across difMean RR (ms) 515 6 104 509 6 110 499 6 81 481 6 62 SDNN (ms) 73 6 39 72 6 46 53 6 17 38 6 14* ferent ages. We found that the quoSDNN/RR 0.13 6 0.05 0.13 6 0.06 0.11 6 0.01 0.08 6 0.02* tient of SDNN and mean RR interSDANN (ms) 63 6 34 64 6 44 46 6 14 36 6 14* vals tended to remain stable across ASDNN (ms) 37 6 23 32 6 22 24 6 13 14 6 5* most of the pediatric age range. TP (beats/min2) 54 6 70 42 6 43 36 6 34 22 6 29 In the frequency domain our findLF (beats/min2) 8.9 6 7.4 8.4 6 8.2 6.9 6 7.4 4.3 6 6.4 HF (beats/min2) 8.7 6 11.4 6.5 6 8.7 2.9 6 3.7 1.2 6 1.6 ings were similar to those of Finley LF/HF 2.3 6 1.7 1.9 6 1.7 4.2 6 4.0 3.6 6 2.8 and Nugent10 where the LF/HF ratio decreased in early childhood, fol*p ,0.05 between groups I and II. lowed by a later increase. However, unlike their report, absolute values of LF and HF power tended to change Cyanotic lesions were present in 17: 4 with tetralogy in opposite directions in our subjects, with HF power of Fallot and 17 with complex lesions. Table II shows increasing in early childhood and showing a later the comparison with healthy age-matched controls. decline into adolescence. Again, these differences are Time domain measures of HRV were not significantly most likely related to the duration of electrocardiodifferent. Patients had significantly lower total and HF graphic recordings. Furlan et al14 derived frequency power, whereas LF power and LF/HF ratio were not domain values for HRV using an averaging technique spanning 6-hour segments of data, which could obsignificantly different. Postoperative congenital heart disease: HRV re- scure detailed information about autonomic modulacordings were obtained an average of 5.8 days after tion of HRV. This observation and concern about the surgery. For the entire group of 36 patients, the mean lack of “stationarity” of the signal partly underlie the RR interval, and all time and frequency domain mea- recommendations of the Task Force of the European sures of HRV were significantly lower in the postop- Society of Cardiology and the North American Socierative period than preoperatively (Table III). The ety of Pacing and Electrophysiology to preferably use magnitude of decrease was greater for HF than for LF of short-term recordings for these derivations.15 We selected stable segments of sinus rhythm at the lowest power, resulting in an increase in LF/HF ratio. There were no significant preoperative differences mean heart rate in an attempt to standardize condiin mean RR interval, time or frequency domain mea- tions. The continuous supine position or controlled sures of HRV between groups I and II. Postopera- ventilation, as is used in adults, was not possible in our tively, there was no significant difference in heart rate; younger population. Preoperative congenital heart disease: HF power however, group II had significantly decreased HRV as measured in the time domain. The differences in has been shown to correlate with parasympathetic SDNN persisted even when adjusted for the RR inter- modulation.6,16 –18 The finding of decreased HF power val (SDNN/RR quotient). Although all of the fre- in children with congenital heart disease before surquency domain measures were decreased in group II, gery suggests relative parasympathetic withdrawal. these did not reach statistical significance. Subjects in The observed decrease in total power was only partly group II had significantly longer cardiopulmonary by- contributed to by a decrease in HF power. This sugpass times (mean 92.5 vs 52.5 minutes, p 5 0.02) and gests that there could be alterations in other factors length of stay in intensive care (mean 9.4 vs 1.8 days, thought to contribute to HRV including sympathetic 1656 THE AMERICAN JOURNAL OF CARDIOLOGYT
VOL. 83
JUNE 15, 1999
tone and the renin-angiotensin and thermoregulatory systems.16 These findings are in agreement with the hypothesis that decreased HRV is a nonspecific marker of cardiovascular stress, reflecting an alteration in autonomic nervous system input to the heart.15 Postoperative congenital heart disease: The significant decrease in HRV and HF power in the immediate postoperative period suggests that further parasympathetic withdrawal is an important component of these changes. This could partly be the result of decreased sensitivity of the sinus node to autonomic nervous input immediately after surgery for congenital heart disease. None of these patients was taking b blockers or any specific antiarrhythmic medications, which are known to alter HRV. Reduced HRV, independent of heart rate, in subjects who had a prolonged postoperative course suggests that HRV provides additional information on the well being of such patients. Whereas all measures of HRV reflected this, only long-term differences in time domain measures reached statistical significance. It has been observed that the values of these measures of HRV increase with increasing duration of recording.15,19 These differences in time domain measures could reflect differences in the renin-angiotensin system.16 Importantly, pleural drainage was longer in subjects who had longer lengths of hospital stay. Higher renin and angiotensin levels have been found in postoperative pediatric cardiac patients who had longer duration of pleural drainage.20,21 The difference in timing of the postoperative analyses between the 2 groups was to ensure that all subjects were in a similar stable hemodynamic state and off all vasoactive medications. Because the recordings were done later in group II, it would be expected that the differences in HRV would be reduced. Hence, the observed differences in HRV between these 2 patient groups assumes more significance. In this study it was not possible to control for differences in diagnoses and type of surgery because of the small number of subjects in each diagnostic category; however, these groups were apparently homogenous from a heart rate control or autonomic standpoint, because there were no preoperative differences in heart rate or HRV. In the only other report on HRV in perioperative critically ill pediatric cardiac patients, Gordon et al11 showed that children who died were more likely to have a “lower respiratory quotient” i.e., a higher respiratory-to-lower frequency power ratio. This was interpreted as being due to a relative increase in parasympathetic tone and possibly sympathetic exhaustion.11 Data were gathered in 2 centers using different analytic methods, standard measures of HRV were not used, and the duration of recording varied from 1 to 27 hours. Significant (38% to 62%) and possibly important segments were excluded from analysis. Most of their patients were likely to have been receiving vasopressors and artificial ventilation, both of which
affect HRV. Our data represent a more complete evaluation of longer term stable recordings and thus can not be compared with data from that report. Acknowledgment: We are grateful to Cathy Tipton, RCDS, Mary Drury, RN, RNP, Susan Stollzman, RN, RNP, Claudio Ramaciotti, MD, and Lynn Mahony, MD, for their assistance.
1. Kleiger RE, Miller JP, Bigger JT, Moss AJ, and the Multicenter Post-Infarction Research Group. Decreased HRV and its association with increased mortality after acute myocardial infarction. Am J Cardiol 1987;59:256 –262. 2. Bigger JT, Fleiss JL, Steinman RC, Ronitzky LM, Kleiger RE, Rottman JN. Frequency domain measures of heart period variability and mortality after myocardial infarction. Circulation 1992;85:164 –171. 3. Rich MW, Saini JS, Kleiger RE, Carney RM, teVelde A, Freedland KE. Correlation of HRV with clinical and angiographic variables and late mortality after coronary angiography. Am J Cardiol 1988;62:714 –717. 4. Brouwer J, van Veldhuisen DJ, Man in’t Veld AJ, Haaksma J, Dijk A, Visser KR, Boomsma F, Dunselman PHJM, Lie KI, for the Dutch Ibopamine Multicenter Trial Study Group. Prognostic value of HRV during long-term follow-up in patients with mild to moderate heart failure. J Am Coll Cardiol 1996;28:1183– 1189. 5. Stein KM, Bores JS, Hochreites C, Okin PM, Herrold EM, Devereux RB, Kligfield P. Prognostic value and physiologic correlates of HRV in chronic severe mitral regurgitation. Circulation 1993;88:127–135. 6. Binkley PF, Nunziata E, Haas GJ, Nelson SD, Cody RJ. Parasympathetic withdrawal is an integral component of autonomic imbalance in congestive heart failure: demonstration in human subjects and verification in a paced canine model of ventricular failure. J Am Coll Cardiol 1991;18:464 – 472. 7. Schechtman VL, Harper RM, Kluge KA. Development of heart rate variation over the first six months of life in normal infants. Pediatr Res 1989;26:343–346. 8. Goto M, Nagashima M, Baba R, Nagano Y, Yokota M, Nishibata K, Tsuji A. Analysis of HRV demonstrates effects of development on vagal modulation of heart rate in healthy children. J Pediatr 1997;130:725–729. 9. Massin M, von Bernuth G. Normal ranges of HRV during infancy and childhood. Pediatr Cardiol 1997;18:297–302. 10. Finley J, Nugent ST. HRV in infants, children and young adults. J Auton Nerv Syst 1995;51:103–108. 11. Gordon D, Herrera VL, McAlpine L, Cohen RJ, Akselrod S, Lang P, Norwood W. Heart-rate spectral analysis: a noninvasive probe of cardiovascular regulation in critically ill children with heart disease. Pediatr Cardiol 1988;9: 69 –77. 12. Finley JP, Nugent ST, Hellenbrand W, Craig M, Gillis DA. Sinus arrhythmia in children with atrial septal defect: an analysis of HRV before and after surgical repair. Br Heart J 1989;61:280 –284. 13. Umetani K, Singer DH, McCraty R, Atkinson M. Twenty-four hour time domain HRV and heart rate: relations to age and gender over nine decades. J Am Coll Cardiol 1998;31:593– 601. 14. Furlan R, Guzetti S, Crivellaro W, Dassi S, Tinelli M, Baselli G, Cerutti S, Lombardi F, Pagani M, Malliani A. Continuous 24-hour assessment of neural regulation of systemic arterial pressure and RR variabilities in ambulant subjects. Circulation 1990;81:537–547. 15. Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology. HRV: standards of measurement, physiological interpretation, and clinical use. Circulation 1996;93:1043–1065. 16. Askelrod S, Gordon D, Ubel FA. Shannon DC, Barger AC, Cohen RJ. Power spectrum analysis of heart rate fluctuation: a quantitative probe of beat to beat cardiovascular control. Science 1981;213:220 –222. 17. Pagani M, Lombardi F, Guzzetti S, Rimoldi O, Furlan R, Tizzi Nelli P, Sandrone G, Malfatto G, Dell’Orto S, Piccaluga E, Turiel M, Baselli G, Cerutti S, Malliani A. Power spectral analysis of heart rate and blood pressure variabilities as a marker of sympathovagal interaction in man and conscious dog. Circ Res 1986;59:178 –193. 18. Malliani A, Pagani M, Lombardi F, Cerutti S. Cardiovascular neural regulation explored in the frequency domain. Circulation 1991;84:1482–1492. 19. Saul JP, Albrecht P, Berger RD, Cohen RJ. Analysis of long-term HRV: methods, 1/f scaling and implications. In: Computers in Cardiology 1987. Washington, DC: IEEE Comput Soc Press, 1988:419 – 422. 20. Mainwaring RD, Lamberti JJ, Moore JW, Billman GF, Nelson JC. Comparison of the hormonal response after bidirectional Glenn and Fontan procedures. Ann Thorac Surg 1994;57:59 – 64. 21. Mainwaring RD, Lamberti JJ, Carter TL Jr, Moore JW, Nelson JC. Renin, angiotensin II and the development of effusions following bidirectional Glenn and Fontan procedures. J Card Surg 1995;10:111–118.
CONGENITAL HEART DISEASE/AGE– AND DISEASE–RELATED CHANGES
1657