A. Rydberg1
Published online: 20 January 2004 Abstract Heart rate variability (HRV) can be used to study cardiovascular autonomic control. This study examines HRV in children with Fontan circulation and its change over time. Thirty-four children in two groups were examined. Group A consisted of 10 patients who had undergone total cavopulmonary connection. Group B consisted of 24 healthy children/adolescents matched for gender, height, and weight. To analyze HRV parameters, all examinations included echocardiography and 24-hour ambulatory electrocardiogram. Comparing all patients and controls, there were no significant differences in HRV parameters. Analyzing subgroups of patients younger than 10 years old, two of the HRV parameters were significantly different compared to controls. For the group of patients older than 10 years, eight of the HRV parameters were significantly reduced. Most significant differences were found regarding low-frequency range (p %26lt; 0.008) and high-frequency range p %26lt; 0.008. This study confirms the finding of earlier studies that patients with Fontan circulation have a reduced HRV, and our findings indicate that there is a progressive reduction of HRV over time.
Keywords Heart rate variability - Fontan circulation Heart rate variability (HRV) is regulated by sympathetic and parasympathetic activity, reflecting spontaneous changes in autonomic activity. Heart rate is not constant but fluctuates over time. Analysis of HRV is a useful method for studying cardiovascular autonomic control. In addition, HRV has been found to be impaired in cardiac diseases [18]. In 1998, Massin and von Bernuth [16] found that HRV was reduced in children with congenital heart disease; they suggested that the reduction is correlated with functional limitations assessed by clinical state. During the past few decades, new operative methods have improved the survivability of patients with complex cardiac anomalies. An increasing number of patients with a variety of functional single ventricles (including hypoplastic left heart syndrome) are now offered a modified Fontan operation as a definitive repair [5, 17]. Today, the most common surgical technique is to create a total cavopulmonary connection (TCPC) so that the systemic venous blood from the superior caval vein is connected directly to the pulmonary artery, and the venous blood from the inferior caval vein is connected to the pulmonary artery by a lateral tunnel in the right atrium [12]. In the past, the atriopulmonary connection was used more frequently [10]. The atriopulmonary connection includes atrial suturing, atrial distension, and elevation of atrial pressure梖actors all influencing sinus node function. Using a lateral tunnel close to the sinus node, the TCPC may lead to the development of sinus node dysfunction. K黵er et al. [11] showed sinus node abnormalities in 50% of Fontan patients, probably due to a continuous decrease in sinus node function over time. A high loss of sinus rhythm was found in a 2-year follow-up study of patients with TCPC [9]. Fishberger and others have shown that the presence of sinus node dysfunction in this group of patients indicates a risk for developing arrhythmia [4, 7, 13]. The underlying mechanism leading to the development arrhythmias is unclear, but it has been speculated that sinus node dysfunction with bradycardia predisposes to prolonged refractoriness and ectopic beats, providing an environment for atrial reentry [6]. Due to sinus node dysfunction and arrhythmia, among other reasons, a new surgical option with an extracardiac connection has been introduced [14]. In 1999, Butera et al. [2] showed that patients with TCPC had significantly reduced HRV and suggested that surgery alters the intracardiac nervous system. We previously observed this, but we also observed that the HRV decreases over time. This study prospectively examines HRV in patients with TCPC circulation and the change in HRV over time. Materials and Methods
Patients
Thirty-four children/adolescents in two groups were examined. Group A consisted of 10 patients who had undergone TCPC, and group B consisted of 24 children/adolescents with a structurally normal heart. Both groups were selected from the same region. All examinations in group A were performed according to local protocol for routine follow-up of patients with Fontan circulation. The study protocol for the two groups was approved by the ethics committee of the medical faculty and informed consent was obtained from children and parents for both groups. Group AThe mean age at the TCPC procedure was 84 months (range, 18?57; SD = 62), and the examinations were performed at a mean of 33 months (range, 13?3; SD = 21) after the final operation. The mean age at testing was 117 months (range, 31?10; SD = 69), and the mean body surface area was 1.085 m2 (range, 0.50?.80; SD = 0.45). There were four patients with tricuspid atresia, three with pulmonary atresia, one with double-inlet left ventricle and transposition of the great arteries, one with unbalanced atrioventricular septal defect, and one with transposition of the great arteries and a large ventricular septal defect (functional single ventricle). All patients had undergone a TCPC. Two patients were medicated with digoxin and three patients were medicated with ACE inhibitors. One patient was medicated with both digoxin and an ACE inhibitor. None of the patients were medicated with Group BThis group consisted of 24 healthy children, matched for gender, height, and weight to group A. Group B had 14 boys and 10 girls, ranging from 13 to 218 months of age. Methods
Echocardiographic transthoracic examinations were performed on all 34 children/youths in both groups. Global ventricular function was semiquantitatively assessed in all projections. The ventricular function was graded on a scale from 1 to 3, where 1 was evaluated as poor, 2 as fair, and 3 as good. Using the Sherpa or the Tracker III recorder (Reynolds Medical, Herts, UK), all 10 patients and all 24 healthy children/youths underwent standard 24-hour ambulatory electrocardiographic monitoring during daily activity. Four electrodes were attached: one in the V2 position, one in the V5 position, and two as references underneath the right clavicle. Electrocardiographic monitoring less than 21 hours was excluded. All recordings were analyzed using a PC-based Holter system (Aspect Holter System, Danica Biomedical, Borl鋘ge, Sweden) [8]. One of the two channels (V2 or V5) was analyzed. One analyst reviewed all the data. Fast Fourier transformation was used to perform the spectral analysis. The analysis was performed during a 24-hour period. Frequency Domain IndicesSpectral power was determined over three frequency regions (Table 1): very low frequency (VLF), 0.003?.04 Hz; low frequency (LF), 0.04?.15 Hz; and high frequency (HF) 0.15?.4 Hz. The total power (TP) was between 0.003 and 0.4 Hz. Table 1 Definition of frequency domain parameters, time domain parameters, and mean heart rate
Time Domain IndicesWe examined six time domain parameters: the standard deviation of all NN intervals (SDNN), the mean of the 5-minute standard deviations of all NN intervals (SDNN index), the standard deviation of the average NN intervals in 5-minute segments (SDANN), the number of interval differences of adjacent NN intervals %26gt;50 msec (NN50 count), and the proportion derived by dividing NN50 by the total number of NN intervals (pNN50). Mean Heart RateNN interval is the normal-to-normal interval; that is, all intervals between adjacent QRS complexes resulting from sinus node depolarization. The 24-hour mean NN was measured as the mean of all NN intervals during the analysis. Statistical Analysis
Data were analyzed using Statistica (Version 5.5, StatSoft). A linear regression model was used to study the relationship between the parameters and age or mean RR interval. Student Results
Echocardiography
Group ASystemic ventricular function was evaluated by echocardiography and determined to be good in six patients, fair to good in two patients, and fair in one patient. No differences were noted between patients younger and those older than 10 years. Group BAll healthy children/youths were examined with complete two-dimensional and Doppler echocardiography. All findings were normal including ventricle function. HRV Analysis
Figures 1,2,3,4 show regression lines for frequency domain variables and SDNNI as functions of mean interval, calculated from 24-hour registration. All variables are highly dependent on the mean RR interval, and a strong correlation was found between the mean RR interval and most HRV parameters (e.g., the SDNN index). Figures 1,2,3,4 illustrate the differences between patients younger and those older than 10 years. They show that the regression lines of the patients younger than 10 years did not differ from those of the controls; however, the regression line 
Figure 1 Regression lines for SDNNi as a function of mean RR interval for controls and patients.
Figure 2 Regression lines for VLF as a function of mean RR interval for controls and patients.
Figure 3 Regression lines for LF as a function of mean RR interval for controls and patients.
Figure 4 Regression lines for HF as a function of mean RR interval for controls and patients.
There were no significant differences between the patients in group A and the healthy control group B (Table 2). There was a tendency toward lower HRV and higher mean NN (indicating bradycardia) in the patient group. Table 2 HRV parameters, for the patients group (group A) and the control group (group B)
NS, not significant.
For subgroups of patients younger than 10 years, only two of the HRV parameters were significantly different compared to controls (Table 3). The younger patients showed a significant prolonged mean RR interval compared to controls (657 vs 590 msec; p = 0.039), indicating an age-adjusted relative bradycardia. The SDNN, estimating the overall HRV, seemed to be significantly increased (p = 0.039). Table 3 HRV parameters, for patients younger than 10 years and those older than 10 years with their matched controls
NS, not significant
In the group of patients older than 10 years, six of the nine HRV parameters during a 24-hour analysis were significantly reduced compared to those of healthy children/youths (p %26lt; 0.05) (Table 3). TP, LV power, HF power, SDNNi, NN50 count, and pNN50 were all reduced. Most significant differences were found with regard to frequency domain indices. In the LF range, group A had a mean of 463 msec2 and group B had a mean of 2003 msec2 (p %26lt; 0.008). In the HF range, the mean in group A was 265 msec2 versus 1358 msec2 for group B (p %26lt; 0.008). Discussion
Our study demonstrates a significant reduction in HRV in patients older than 10 years palliated with Fontan circulation. This difference was not found in the younger patient group compared to matched controls. This finding may indicate a progressive autonomic dysfunction in patients with Fontan circulation. Heart rate and rhythm are largely under the control of the autonomic nervous system and the vagal and sympathetic activity constantly interact. Many factors may reduce HRV, including sinus node dysfunction. In sinus node dysfunction, the reduction in HRV is due to a diminished responsiveness of sinus nodal cells to neural modulations. Many patients with Fontan circulation show sinus node abnormalities probably due to a continuous decrease in sinus node function over time [3]. With this background, it is of great interest that we found a significant reduction in HRV in patients older than 10 years. In 1997, Massin and von Bernuth [15] showed that age is an important determinant of HRV in normal subjects and also confirmed a progressive maturation of the autonomic nervous system during childhood. In our study, the group of healthy children and youths was smaller, but the findings (especially regarding frequency domain indices) were quite similar. In 1998, Massin and von Bernuth [16] published another HRV study that examined 258 children with congenital heart disease, including 32 patients after Fontan repair. They concluded that children with congenital heart disease have a reduced HRV. This study confirms our conclusion regarding the older group of patients. In 1999, Butera et al. [1] studied patients with Fontan-type circulation and found a significant reduction in HRV, particularly abnormal sympathovagal balance. They matched heart rate to a control group including 18 healthy children referred to the institution with a history of palpitations and who seemed to have sympathetic activation that influenced the HRV parameters. Our control group comprised matched, healthy children and youths with no complaints of palpitations or other signs of sympathetic drive. Limitations
In this study, the number of patients is small, but regarding the surgical technique the group is quite homogenous. Echocardiographic assessement of ventricular function is difficult even in hearts with normal anatomy. In univentricular hearts, the ventricular anatomy is complex, and simple geometric assumptions for calculating ventricular volumes are questionable. With such a limited number of patients with mostly good ventricular function, any conclusion about decreased HRV caused by decreased ventricular function would be only speculative. Massin and von Bernuth [16] concluded that reduced HRV in children with congenital heart disease was caused by functional limitation as clinically defined by the New York Heart Association. We did not evaluate functional classes in the patient group, but we could not find any difference between patients younger than 10 years and those older than 10 years regarding global assessed ventricular function. Conclusion
Our study shows reduced HRV in patients with Fontan circulation older than 10 years, which is in agreement with previous studies [2]. However, in patients younger than 10 years, there were only minor differences compared to controls. Since there is a strong positive correlation between HRV parameters and mean RR interval, one would expect lower HRV when the heart rate is higher. The fact that the patients younger than 10 years did not differ from controls (in contrast to patients older than 10 years, who clearly deviated from controls) indicates that patients with TCPC show a progressive reduction of HRV parameters over time. References | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
-blockers.
s t-test was used to compare group A and group B. Subgroup analysis was performed using the Mann朩hitney U-test. In all the statistic tests, the null hypothesis was rejected at the 5% level (p %26lt; 0.05).
