EVALUATION OF CARDIAC AUTONOMIC FUNCTION WITH HOLTER ECG MONITORING IN PATIENTS WITH SYSTEMIC SCLEROSIS

Purpose. Cardiac autonomic dysfunction is frequently encountered among patients with systemic sclerosis. Its presence correlates with potentially fatal ventricular arrhythmias, having a good positive predictive value for mortality. The aim of this paper was to identify cardiac autonomic dysfunction in patients with systemic sclerosis using Holter ECG monitoring and to assess the possible correlations between its presence and other disease characteristics. Material and methods. Forty nine patients with diffuse cutaneous and limited cutaneous scleroderma, diagnosed according to the American College of Rheumatology/EULAR 2013 criteria underwent ECG, transthoracic echocardiography, blood sample testing, chest X-ray and spirometry. Subsequently, all patients and a control group of 49 healthy subjects underwent Holter ECG monitoring with time and frequency domain heart rate variability (HRV) analysis. Results. Scleroderma patients had signifi cantly lower HRV values compared to controls: SDNN (123 ± 39.4 vs. 143.2 ± 32, p =0.001), SDANN (137.2 ± 34.9 vs. 127.9 ± 25.7, p=0.001), TI (-) (30.6 ± 9.6 vs. 38.1 ± 8.9, p=0.01) and TINN (758.8 ± 208 vs. 815.1 ± 138, p=0.04). The LF/HF ratio was signifi cantly higher among patients with diffuse scleroderma (1.18 ± 0.34 vs. 1.08 ± 0.43, p= 0.005). There was a positive correlation between the TI values, SDANN and the duration from the onset of Raynaud’s phenomenon (r=0.399, p=0.016) (r=0.419, p=0.011), between the SDANN values and systolic pulmonary arterial pressure (sPAP) values (0.032, p=0.034) and a negative correlation between the LF/HF values and the patients’ age (r=-0.442, p=0.001), the duration from the onset of non-Raynaud’s phenomenon (r=-0.395, p=0.034), echocardiographic value of sPAP(r=-0.330, p=0.035) and the total number of premature ventricular contractions (r=-0.0459, p=0.001). Conclusion. Patients with systemic sclerosis often have cardiac autonomic dysfunction, which can be diagnosed with Holter ECG monitoring.


INTRODUCTION
Scleroderma is a systemic connective tissue disease of unknown etiology characterized by excessive fi brosis of the skin, blood vessels and internal organs, including the heart. Cardiac involvement is frequently encountered among scleroderma patients, with a prevalence of clinical manifestations of 15-35% (1). Subclinical cardiac involvement is even more frequent (2), and cardiac autonomic dysfunction is an important part of it (3). This can be present from an early stage of the disease, preceding the onset of fi brosis (4), and is characterized by a reduction in the parasympathetic tone and an increase in the sympathetic tone, especially in the early stages (5). The diagnosis of cardiac autonomic dysfunction is frequently established in a late stage of the disease, due to the discrepancies between the mild clinical manifestations and the severity of cardiac involvement.
Previous studies that assessed cardiac autonomic function in scleroderma patients have shown that the presence of cardiac autonomic dysfunction correlates with the presence of potentially malignant ventricular arrhythmias (6)(7)(8), having a good positive predictive value for mortality (9). Therefore, an early diagnosis of cardiac autonomic dysfunction is crucial for guiding optimal treatment from an early stage of the disease.
One of the existing methods for the evaluation of cardiac autonomic function is the study of heart rate variability (HRV). The variations in heart rate are mediated by the cardiac autonomic nervous system; therefore the analysis of HRV will refl ect the infl uences of the autonomic nervous system upon the heart. A widely available, cost-effective, non-invasive method capable of identifying alterations in HRV is Holter ECG monitoring (10). Currently, there are only a few studies in the literature that evaluated HRV with Holter ECG monitoring for the assessment of cardiac autonomic function in patients with scleroderma (7,9,(11)(12)(13)(14)(15)(16)(17)(18). Although these studies suggested a correlation between alterations in HRV, heart rate turbulence and the presence of malignant ventricular arrhythmias, they were conducted in small sample-sized populations (between 12 and 68 patients, with an average of 36 patients); therefore, additional studies are needed to confi rm their main fi ndings.
The aim of this paper was to identify the presence of cardiac autonomic dysfunction among scleroderma patients using Holter ECG monitoring and to assess the possible correlations between its presence and other characteristics of the disease.

Patient population
Among the 134 patients, both men and women, diagnosed according to the American College of Rheumatology / EULAR 2013 criteria between November 2011 -July 2014 with either diffuse cutaneous or limited cutaneous scleroderma in the Rheumatology Clinic in Cluj-Napoca, 49 patients were included in the study. Exclusion criteria were: the presence of other rheumatologic diseases: rheumatoid arthritis, polymyositis, Sjogren's Syndrome, Still's disease, mixed connective tissue disease; cardiovascular diseases: ischemic heart disease, congestive heart failure, a left ventricular ejection fraction < 50%, systemic arterial hypertension, moderate and severe pulmonary hypertension, signifi cant valve disease, atrial fi brillation, atrial fl utter, atrial tachycardia, major conduction disorders such as second and third degree AV block, complete left or right bundle branch block; the uses of antiarrhythmic drugs: beta blockers, class I and III antiarrhythmic drugs, Digoxin; other diseases that may impair HRV: diabetes mellitus, neuropathies, renal dysfunction (serum creatinine level of > 1.5 mg/dl), signifi cant pulmonary disease (severe obstructive or restrictive lung disease) and thyroid dysfunction (uncontrolled hyperthyroidism of hypothyroidism).
The control group comprised patients without known cardiovascular disease, admitted between April 2012 and June 2014 to the Physical and Rehabilitation Medicine Department, Rehabilitation Hospital, Cluj-Napoca. These patients were selected from a group of 85 patients who underwent Holter ECG monitoring for arrhythmia detection. Patients with arrhythmias detected on Holter ECG monitoring (1 patient with paroxysmal supraventricular tachycardia and 1 patient with paroxysmal atrial fibrillation) and / or associated cardiovascular diseases were excluded from the study. A total of 49 age and sex-matched subjects were then selected to represent the control group.

History taking and clinical exam
Special attention was paid to the duration from the onset of Raynaud's phenomenon and of non-Raynaud's phenomenon, disease duration from the moment of diagnosis, the presence of symptoms: palpitations, exertional and resting dyspnea, vertigo, syncope; skin (Rodnan) score, telangiectasias, cutaneous calcinosis and digital ulcers.

lead ECG and 24 hour Holter ECG monitoring
The 12 lead ECGs were recorded using an Esaote P8000 electrocardiograph, with an ECG amplifi er sensitivity of 10 mm/mV, at a speed of 25 mm/s. All ECGs were carefully screened for the presence of supraventricular or ventricular arrhythmias, conduction disorders, QRS axis deviations, signs of atrial or ventricular hypertrophy, myocardial ischemia, low QRS voltage, prolonged QRS and QTc duration.
The 24 hour Holter ECG monitoring was performed using a 7-lead BTL CardioPoint H600 device, having a 2,000 Hz sampling frequency and 16 bits digital resolution. The studied parameters were: minimum, average and maximum heart rate while awake, asleep and during the whole 24 hour monitoring period; the presence of supraventricular arrhythmias: frequent premature atrial contractions (PAC), paroxysmal supraventricular tachycardias, atrial fi brillation, atrial fl utter; the presence of ventricular arrhythmias: frequent premature ventricular complexes (PVC), non-sustained and sustained ventricular tachycardias; the presence of paroxysmal conduction disorders; QT and corrected QT interval. For HRV analysis, data was automatically processed by the integrated software and was subsequently verifi ed and corrected when appropriate by a full time cardiologist in all cases.

Echocardiography
Transthoracic echocardiography was performed using an Esaote MyLab TM X-View 50 system, with a 7.5-10 MHz transducer. Standard image acquisitions were performed for all subjects. The assessed parameters were: chamber size and wall thickness, systolic function of the left ventricle (LV) and right ventricle (RV), diastolic function of the left ventricle, wall motion abnormalities, the presence of valve disease (stenosis and regurgitations), systolic, mean and diastolic pulmonary arterial pressure (sPAP, mPAP, dPAP) and the presence of pericardial effusion. Mild pulmonary hypertension was defi ned as sPAP of 35-49 mmHg, moderate pulmonary hypertension as sPAP between 50-69 mmHg and severe pulmonary hypertension as sPAP ≥ 70 mmHg.
Heart rate variability analysis Heart rate variability analysis was performed in both time and frequency domains.

Time domain parameters
The following parameters from the temporal domain were analyzed: • NN, which corresponds to the number of RR intervals between 2 sinus QRS complexes • HRV TI (Heart Rate Variability triangular Index), represents the integral of the density distribution (the number of all NN intervals) divided by the maximum of the density distribution. The measure is approximated by the value: (total number of NN intervals)/ (number of NN intervals in the modal bin) • TINN (Triangular interpolation of the NN), which represents the base of the HRV TI triangle and approximates the distribution of the NN intervals. The HRV triangular index (HRV TI) represents the baseline width of the distribution measured as a base of a triangle, approximating the NN interval distribution (the minimum square difference is used to fi nd such a triangle). • SDNN (Standard Deviation of NN), corresponding to the standard deviation of the NN intervals (in milliseconds) • SDANN (Standard Deviation of Averages of all Normal intervals), corresponding to the standard deviation of the average NN intervals in all 5-minute intervals of the entire recording. • rMSSD (Root Mean Squared of Successive Differences), corresponding to the square root of the mean of the sum of the square differences between consecutive NN intervals • NN50, corresponding to the number of differences between 2 NN intervals > 50 ms. • pNN50, corresponding to the number of pairs of adjacent NN intervals differing by more than 50ms in the entire recording, divided by the total number of all NN intervals.

Frequency domain parameters
The following parameters from the frequency domain were assessed: • Low frequency: (LF), between 0.04 and 0.15 Hz • High frequency: (HF), between 0.15 and 0.4 Hz.
In the present study, the LF/HF ratio was used.

Statistical analysis
Statistical analysis was performed using SPSS Statistics 20 (IBM. Chicago, Illinois). Descriptive statistics was used to summarize the characteristics of patients. Normality was assessed by using a Shapiro-Wilk test. Results are expressed as mean ± standard deviation (SD) if normally distributed, or by median and interquartile range otherwise. Categorical variables are presented as counts and proportions (%).
Fisher's exact test was used to compare categorical variables. T-test for independent samples was used to compare normally distributed scalar variables and the Mann-Whitney-Wilcoxon test were used to compare non-normally distributed scalar values. Spearman's correlation coeffi cients were used to assess the relationship between the heart rate variability parameters and varied clinical and para-clinical patients' characteristics.
A p value of < 0.05 was considered statistically signifi cant.

General characteristics of the patients
The main characteristics of the patients included in the present study are presented in Table 1.
There were no statistically signifi cant differences between the diffuse cutaneous and limited cutaneous subgroups of scleroderma patients in what concerns the main clinical characteristics, with the exception of the skin score (15.2 vs. 11.1, p<0.01) and the presence of anti SCL-70 antibodies (15 vs. 5, p<0.01).
On Holter ECG monitoring, scleroderma patients had signifi cantly higher values of the minimum and average heart rate compared to controls ( Table 2). The characteristics of supraventricular and ventricular arrhythmias were similar between the 2 groups of patients. There were no statistically signifi cant differences between the maximum, minimum and average heart rate, total number of PAC, PVC, QT and QTc interval between patients with diffuse cutaneous and limited cutaneous scleroderma.
On transthoracic echocardiography, the main fi nding was the presence of diastolic dysfunction of the left ventricle (25% of patients from the diffuse scleroderma subgroup vs. 36% of patients from the

Cardiac autonomic function parameters
Among scleroderma patients, there was a significant reduction in the values of the following time domain HRV parameters compared to controls: SDNN, SDANN, TI (-) and TINN. The rest of the time domain HRV parameters had similar values in both groups. There were no statistically signifi cant differences in the frequency domain parameters between the 2 groups of patients (Table 3).  Among all scleroderma patients, there was a positive correlation between certain time and frequency domain HRV parameters and the duration from the onset of Raynaud's and non-Rauynaud's phenomenon, the value of sPAP assessed by echocardiography, and a negative correlation between HRV parameters and the total number of PVC on Holter ECG monitoring (Table 5). When separated according to the scleroderma subtype, a negative correlation be- tween time domain HRV parameters and the ANA titer was found among patients with limited cutaneous scleroderma (Table 5).

DISCUSSION
The main fi nding of the present study is that cardiac autonomic dysfunction is present in scleroderma patients, and it can easily be diagnosed by assessing HRV using Holter ECG monitoring.
Heart rate variability is defi ned as subtle beat-tobeat variations / oscillations between 2 consecutive heart beats (19); it is controlled by the autonomic nervous system and refl ects the degree of baroreceptors' sensitivity. Heart rate variability is impaired in scleroderma patients both in an upright and supine position, which was demonstrated with the use of tilt table testing (20). The lack of a signifi cant increase in heart rate when switching from the supine to upright position refl ects an altered baroreceptor modulation by the autonomic nervous system.
A reduction in HRV parameter values indicates an increase in the sympathetic infl uences and a decrease in the parasympathetic infl uences on the sino-atrial node, fostering electrical instability (1,21). The main time domain HRV parameters are SDNN and HRV triangular index. These parameters assess heart rate variability on a global level and have a prognostic role. TINN also characterizes HRV on a global level. SDANN and rMSSD predict long-term alterations of HRV, and short-term, respectively; rMSSD and pNN50, which are parameters obtained by measuring differences of adjacent cardiac cycles can be considered as surrogate markers. They are specifi c markers of the parasympathetic nervous system and refl ect the vagal tone, with rMSSD being more frequently used due to its superior statistical properties; SDNN and SDANNi refl ect the circadian variations of HRV, being infl uenced by daily activities, short-term increase in sympathetic and parasympathetic tone. Regarding frequency domain HRV parameters, LF are considered to characterize the barorefl ex control of the sympatho-vagal balance of arterial blood pressure and represent a marker of the sympathetic nervous system activity. HF characterizes sinus respiratory arrhythmia mediated by the parasympathetic nervous system. The LF/HF ratio is considered to characterize the sympatho-vagal balance.
The existence of a signifi cant correlation between cardiac autonomic nervous system dysfunction and cardiovascular mortality is well known (22,23). An imbalance in the sympatho-vagal nervous system results in an increased risk of potentially fatal ventricular arrhythmias development (24)(25)(26)(27)(28). Previous studies performed on post-myocardial infarction patients have shown that an imbalance of the autonomic nervous system represents an independent risk factor for malignant ventricular arrhythmias and sudden cardiac death (29)(30)(31). Other studies have demonstrated that HRV analysis can be used as a non-invasive diagnostic method for identifying the presence of autonomic neuropathy in diabetic patients (32).
In the present study, on Holter ECG monitoring, patients with scleroderma had signifi cantly higher values of the average and minimal heart rate, both while awake and while asleep compared to controls. This observation, previously described by other authors (9), refl ects an increase sympathetic tone in these patients. Studying HRV, Ferri et al. (9) assessed the cardiac autonomic function in a group of 30 scleroderma patients and found signifi cantly higher heart rate values and lower HRV parameter values compared to controls. In addition, these parameters proved to have a good positive predictive value for mortality. The relative risk for mortality was increased among elderly scleroderma patients with positive anti Scl 70 antibodies. In the studies of Malliani et al. (33) and Appel et al. (34), the authors found an increased sympathetic tone in the studied subjects, manifested by higher heart rate values and higher serum levels of circulating catecholamines in patients with cardiac autonomic dysfunction.
The main fi nding of these studies was a decrease in time and frequency domain HRV parameter values and an increase in the LF/HF ratio. An increased LF/HF ratio refl ects an increased sympathetic tone in this population of patients. In our study, the LF/ HF ratio was increased only in patients with diffuse cutaneous scleroderma, compared to patients with limited cutaneous scleroderma.
Another important fi nding of this study is the correlation between certain HRV parameters and clinical and paraclinical characteristics of scleroderma patients. Among patients with both subtypes of scleroderma, there was a positive correlation between the value of TI and SDANN and the duration from the onset of Raynaud's phenomenon, between the SDANN values and systolic pulmonary arterial pressure (sPAP) values and a negative correlation between the LF/HF values and the patients' age, the duration from the onset of non-Raynaud's phenomenon, echocardiographic value of sPAP and the total number of premature ventricular contractions on Holter ECG monitoring. In the study of Othman et al. (13), the authors found a statistically signifi cant positive correlation between certain HRV parameters and the presence of Raynaud's phenomenon, anti Scl 70 antibodies, as well as the skin score. Another study conducted by di Franco (12) analyzed the relationship between the alteration of HRV determined using Holter ECG monitoring and the presence of Raynaud's phenomenon assessed by capilaroscopy. The authors found a statistically signifi cant correlation between an impaired HRV and the semiquantitative capilaroscopy score. They concluded that the coexistence of cardiac autonomic dysfunction with a more severe microvascular dysfunction could be a useful marker for the identifi cation of patients with a higher mortality risk.
In the present study, in both groups of scleroderma subtypes there was a negative correlation between certain HRV parameters (TI, SDANN and TINN) and the total number of PVC on Holter ECG monitoring. Such a correlation has previously been described by Othman et al (13) and suggests the presence of a higher arrhythmic burden among scleroderma patients with more advanced autonomic dysfunction. These patients might have a worse prognosis compared to the other scleroderma patients, since the presence of cardiac autonomic dysfunction and numerous ventricular arrhythmias are associated with a higher risk of malignant ventricular arrhythmias (24)(25)(26)(27)(28).
Further studies are needed to prospectively evaluate the impact of cardiac autonomic dysfunction on cardiovascular morbidity and mortality among scleroderma patients.

CONCLUSION
Patients with systemic sclerosis have subclinical cardiac autonomic dysfunction, manifested by alterations in parameters characterizing heart rate variability, which can be diagnosed with Holter ECG monitoring. Given the correlation between cardiac autonomic dysfunction and the risk of malignant ventricular arrhythmia development, the Holter EGC monitoring may play an important role in identifying patients at risk for sudden cardiac death, allowing implementation of optimal therapeutic measures in an early stage of the disease.