The changes of HRV in refractory epilepsy: The potential index to predict the onset of epilepsy in children

Abstract

In this study, we examine the potential of heart rate variability (HRV) as an efficient tool for predicting the onset of epilepsy in children. We totally collected 53 seizures EEG and ECG data using Video - EEG - ECG monitoring system. We then separated the ECG data into three segments: ten-minute before onset of each seizure, five-minute before onset of each seizure, and five-minute from the onset of each seizure. After the HRV parameters in all segments were calculated, we compared the differences between pre-ictal period and ictal period. We found that the values of meanHR, LF and LF/HF were greater in onset period. And the values of meanRR and the HF were less in ictal period. And it presented the similar changes when seizures occurred in the daytime and seizures occurred in the nighttime. In brief, we found that the sympathetic nervous system was under a more active status during onset period. We speculated that the HRV parameters such as the LF, HF or LF/HF could have potential to predict the seizures in children with epilepsy.

Keywords

Seizure HRV children

1 Introduction

As one common cerebral dysfunction disease presently, about 50 million people suffered from epilepsy [12]. The occurrence of epilepsy in children without immediate treatment is more likely to cause the sudden dead or damage to brain. According to a large-scale study about the children with epilepsy, the mortality in children with epilepsy is three times higher than that of the general population [26]. Some patients with epilepsy fail to respond to antiepileptic drugs or other treatments. The mortality of the epileptic patient in uncontrolled status is more likely higher than those in controlled status [26]. Furthermore, since the seizure can destroy the balance between sympathetic nervous system and parasympathetic nervous system, it might cause autonomic disorder and cardiac arrhythmia, which might lead to sudden death [6, 18].

Thus, it is of great demand to develop a convenient monitoring tool for the children with epilepsy. Although EEG plays an important role in the diagnosis of epilepsy, the unclear mechanisms causing epilepsy and complicated pattern presentations have restricted the potential of EEG in predicting it. Because the seizure can destroy the balance between sympathetic nervous system and parasympathetic nervous system, it might be possible to predict the epilepsy using the parameters extracted from ECG data. Previous studies have shown that the heart rate changes at the onset of seizure [10, 16]. Because it has been proven that the cardiac activities are affected by autonomic nerve [11], we believe that the seizure may cause the disorder of autonomic nervous system. As Heart Rate Variability (HRV) can reflect the status of autonomic nervous system in real time, the prediction of epilepsy using the function of heart autonomic nerves could be possible. Fatemeh Fadaie et al. reported that the momentary cardiac arrest was observed during seizure, which indicated that the patients with epilepsy were at risk of sudden death. Therefore, it will benefit the patient with epilepsy if cardiac monitoring could be conducted as part of daily management [20]. Although there were many studies regarding the relation between HRV and adult with epilepsy [24], the relation between the HRV and the children with epilepsy has been seldom explored. The aims of our study were to compare the difference of HRV in pre-ictal period and ictal period and to explore whether HRV could be an index to predict seizure.

2 Methods

2.1 Subjects

A total of forty children (22 female and 18 male) with refractory epilepsy were included in this study. The refractory epilepsy refers to that the patients with epilepsy fail to respond to antiepileptic drugs or other treatments. It can be understood as uncontrolled and progressive epilepsy. The demographic data, the localization of epilepsy, the duration of epilepsy and the antiepileptic drug therapy of the subjects were summarized in Table 1. All patients were enrolled into the Department of Pediatric Neurosurgery of Shenzhen Children’s Hospital between June 2014 and March 2015. And the patients were monitored by video-EEG with a sampling frequency of 512 Hz for routine inspection. And the Lead II ECG surveillance was carried on simultaneously. A total of 57 partial seizures were observed initially. In order to avoid the interference causing by the adjacent seizure, four seizures were eliminated because that the interval time between two seizures was less than 20 minutes. Besides, two seizures were excluded because that the RR interval couldn’t be recognized clearly on account of signal interference or electrode shedding. The remaining 53 seizures records were selected finally. This study was approved by the ethics committee of the Shenzhen Children’s Hospital. Though this study was retrospective, we received oral informed from all subjects’ guardians over phone.

2.2 Data

The data were selected retrospectively from children with refractory epilepsy. The experienced clinical experts inspected all the EEG data and labeled the onset of each seizure. The seizures were classified into two groups including seizures occurred in the daytime (8 : 00–20 : 00) and seizures occurred in the nighttime (20 : 00–8 : 00). A total of 20 seizures occurred in the daytime and 33 seizures occurred in the nighttime were obtained. We defined three segments in corresponding ECG data according to the onset of seizure: ten-minute before onset of each seizure, five-minute before onset of each seizure, five-minute from onset of each seizure. The length of time for each segment was 5 minute. We used Kubios HRV Software (version2.2.0.0) to calculate all the parameters of HRV. The cardiac electrophysiological signal was derived from surface electrodes adhered to the chest. The QRS complexes were identified automatically. We checked all the RR intervals manually to avoid the miscount of QRS complex. The parameters of HRV were separated into time-domain and frequency-domain. The time-domain parameters included the mean of the RR interval(meanRR), the standard deviation of RR interval(SDNN), the root mean square of difference between successive normal intervals(RMSSD), number of successive RR interval pairs that differ more than 50 ms(NN50), NN50 divided by the total number of RR intervals(p NN50), the integral of the RR interval histogram divided by the height of the histogram(HRV triangular index), and the baseline width of the RR interval histogram(TINN). The RR interval series was calculated by one power spectrum density (PSD) algorithm. The power values for each frequency band were obtained by simply integrating the spectrum over the band limits. The frequency-domain parameters included very low frequency (VLF,≤0.04 Hz) bands powers, low frequency (LF, 0.04–0.15 Hz) bands powers, high frequency (HF, 0.15–0.4 Hz) bands powers, the ratio between LF and HF band powers (LF/HF).

2.3 Statistical analysis

The values of all the parameters were expressed as mean ± sd. The statistical soft ware SPSS 17.0 (SPSS Inc., Chicago, IL, USA) was used to analyze the statistical differences between every two segments. Paired-Samples t-test was used for the evaluation of continuous variables. And the Two-Related-Samples test was used for evaluating the non-normal distributing variables. P < 0.05 was considered statistically significant.

3 Results

This study involved 40 children with refractory epilepsy at average age of 5.86 ± 2.85 years old including 22 female and 18 male. And the mean duration of epilepsy was 2.66 ± 1.72 years. Besides, all of them were taking antiepileptic drugs for the routine management. The mean number of seizures per week was three (range 1 to 7). 24 patients were left-sided in onset and the other 22 patients were right-sided in onset. None of the patients had congenital heart disease, arterial hypertension, clinical signs of autonomic dysfunction or other cardiovascular disease.

Table 2 describes the mean ± SD values of all the frequency-domain parameters and time-domain parameters. Regarding the time domain parameters, when compared segment 1 with segment 2, there were no statistical difference in any parameters. As compared segment 1 with segment 3, the meanRR values were significantly less in segment 3 no matter seizures occurred in daytime or in nighttime (P = 0.000, P = 0.000). And a trend was noted toward higher of meanRR values in segment 3 compared with segment 1 both in daytime and nighttime (P = 0.00, P = 0.00). Besides, when compared segment 2 with segment 3, the meanRR values were significantly less in segment 3 no matter seizures occurred in daytime or the nighttime (P = 0.000, P = 0.000). And the meanHR values were significantly higher in segment 3 regardless of seizures occurred in daytime or the nighttime (P = 0.00, P = 0.00). In frequency domain parameters, when compared segment 1 with segment 2, no significant differences were found in any parameters including VLF, LF, HF, LF/HF. When compared segment 1 with segment 3, the LF values was significantly higher in segment 3 no matter seizures occurred in daytime or the nighttime (P = 0.002, P = 0.000). And the HF values were significantly less in segment 3 when seizures occurred in the nighttime (P = 0.000). Besides, the ratio of LF/HF was significantly greater in segment 3 regardless of seizures occurred in daytime or nighttime (P = 0.007, P = 0.000). The values of VLF shown no difference in seizures occurred in the day or night. When compared segment 2 with segment 3, the LF values were significantly higher in segment 3 seizures occurred in daytime or the nighttime (P = 0.000, P = 0.000). And the HF values were significantly less in segment 3 when seizures occurred in the nighttime (P = 0.000). Furthermore, the ratio of LF/HF was significantly greater in segment 3 regardless of seizures occurred in daytime or nighttime (P = 0.005, P = 0.000). And the values of VLF also presented no significant changes in seizures occurred in the daytime or nighttime. All of the P-values were presented in Table 3.

Figure 1 shows the trend of meanHR in all segments. Patients tended to display a faster heart rate at ictal period. And the changes of Heart Rate in daytime were similar to those in nighttime.

Figure 2 depicts the changes of the LF, HF and LF/HF in different segments when seizures occurred in the daytime. The HF values showed descending in onset period. And the LF and LF/HF presented increase in ictal time.

Figure 3 depicts the variation of LF, HF and LF/HF in different segments when seizures occurred in the nighttime. The HF values also presented decreasing in ictal time. And the LF and LF/HF presented ascent in onset period.

4 Discussion

It is well known that the autonomic nervous system regulates the activity of the internal organs. Previous studies have proven that cardiac activity is under the control of the autonomic nervous system [25]. The autonomic nervous system is consisted of parasympathetic nervous and sympathetic nervous. The parasympathetic modulates heart rhythm via the following mechanism: the cell membrane K+ conductance increase after the muscarinic acetylcholine receptors respond to the acetycholine [2]. And the sympathetic nerve works on heart rhythm via releasing epinephrine and norepinephrine. The process of slow diastolic depolarization accelerated after the activation of β-adrenergic receptors[7 , 28]. Epileptic discharge has potential to spread to the central autonomic nerve net. And it may disrupt the normal regulation of the heart autonomic nervous system in patients with epilepsy. And the insular cortex and amygdala are thought to be the center of the autonomic nervous system. The severe cardiac arrhythmias and other autonomic manifestations may relate to the discharging of insular cortex and amygdale [5].

There were many studies regarding the relation of heart rate change and seizure. Fan-Gang Menget al. found the heart rate is significantly different among pre-ictal, ictal and post-ictal period. The heart rate increases in ictal period or after seizure [9]. And Greene et al. found no significant difference in the HR during ictal period [4]. Hanna N. Ahmed et al. observed bradycardia at ictal period in a few epileptic patients [1]. In our study, we found that meanHR increased at ictal period in 47 seizures. And the remaining 6 seizures did not accompany with HR change. But we observed tachycardia during seizure in only 2 patients. The meanRR increase significantly only in segment 3 and could not be found significant change in segment 1 and segment 2. It suggested that the cardiac activities were likely to be more active at ictal time. What can be confirmed is that cardiac conduction presents disorders in ictal time in patients with refractory epilepsy. The previous study found that the patient with intractable epilepsy may accompany with heart rate variability impairments. All of the patients in our study had no congenital heart disease or other diseases except epilepsy. So we speculated that the disorganized function of autonomic nervous and the damaged of the heart were due to epilepsy itself probably. But the whole subjects were under the treatment of antiepileptic. So it was hard to distinguish whether the impairment was influenced by drugs. Angelica B. Delogu et al. found that the cardiac autonomic nerve is impaired in patients with epilepsy whether patients take antiepileptic or not[13]. H Ansakorpi et al. also reported that the cardiovascular changes are more likely to have relation to the epilepsy process instead of the antiepileptic [3]. In general, the function of cardiac autonomic nervous was injured indeed in some patients with epilepsy. The change of cardiac activity was more various than those in the normal population. More seriously, patients with epilepsy are at risk of sudden death. Fatemeh Fadaie et al. reported that two patients were observed the cardiac arrest at their seizures [20]. Furthermore, the previous study reported that the sudden death has relationship with the seizure type. The generalized tonic-clonic seizures (GTCS) are more likely to lead to sudden death [14]. So epilepsy is a nervous system disease but can lead to cardiac impairment during its progress. And the prevention of the cardiac arrest should be an important part in epileptic management f. To most of our knowledge, Heart rate variability monitoring was a tool could detect autonomic impairment in early time. And HRV reflects modulation capability of cardiac autonomic. So HRV may have the potential acts as a risk warning indicator for patient with epilepsy because that it has operable, repeatable and noninvasive characteristics [19]. Furthermore, the HRV monitoring will be beneficial to the treatment in some patients with epilepsy theoretically.

In our study, there were no statistical differences between time-domain parameters except the mean RR values, meanHR values and the TINN values. According to the mechanism of HRV monitoring, a long time history record is needed for time domain parameters analysis. But the frequency domain parameters can reflect the short-term heart rate variability. Frequency-domain methods may be better than time-domain methods when investigating short-term recordings [8].

The values of LF and LF/HF were greater while the values of HF were less at ictal period. The physiological significance of VLF is not yet clear. The physiological process which may relate to the change of VLF is still unknown [8]. HF is mainly affected by parasympathetic nervous system. LF is under the influence of sympathetic nervous system and parasympathetic nervous. And the LF/HF ratio reflects the balance of sympathetic nervous system and parasympathetic nervous [27]. Seizures occurred in the nighttime were more frequent than that in the daytime in our study. It may have relation with the circadian timing system. Epilepsy may also affected by consciousness and environmental factors [23]. Although the HRV usually changes differently in day and night under normal physiological conditions, the sympathetic nervous system and parasympathetic nervous were more active during onset period no matter seizures occurred in the daytime or the nighttime. We speculated that the HRV changes in ictal time were not connected to the circadian rhythm. Dr. E. Urrestarazu and Dr. J.A. Palma found that increased sympathetic and decreased parasympathetic cardiac tone in patients with sleep disease was due to the reduced nocturnal oxygen saturation [22]. Seizures are associated with inadequate pulmonary ventilation in some patients. So we speculated that the sympathetic increased when seizure occurred was likely to the reduced nocturnal oxygen saturation. L. Nashef also found that pulmonary ventilation is limited and pressure of oxygen dropped in seizures. And the central apnoea may be the reason for sudden death for patient with epilepsy [21]. Therefore, the change of breathing can affect the cardiac autonomic system. In our study, all the results indicated that the sympathetic nervous system became more active in ictal period.

The LF/HF ratio, which reflects the balance between sympathetic nervous system and parasympathetic nervous system, increased at the onset period regardless of seizure occurred in the daytime or nighttime. The sympathetic nervous system was more active in the onset period when compared with five-minute before onset or ten-minute before onset. Ebru Kolsal et al. reported that LF values and LF/HF ratios significantly increase before the seizure when compare with the baseline values. And The parasympathetic system suppressed during the seizures, but there were predomination in the sympathetic system 1 hour before the seizure [17]. In our study, we found the values of LH/HF in ictal time were significantly different to five-minute before onset and ten-minute before onset. The autonomic nervous system changed at onset period but not changed before onset. Angelica B. Deloguet al. found that the LF and HF values decrease in children with refractory epilepsy which were different to the adult patients [13]. But in our study, the HRV changes were similar to what the adult patients present. To most of our knowledge, it is of great interest for most researchers to discover the relation between HRV and adult with epilepsy. And there are less study regarding the relation between HRV and children with epilepsy. So studies about HRV and children with epilepsy are required. At least, the previous researches have shown that HRV could be used to detect the cardiac modulation capability. And for the epileptic patients with cardiac damage, HRV monitoring actually could be done as a tool of management.

The main imitation of this study is that the sample size of our study was small. Therefore, we should test our conclusion in large samples or multiple centers in future study. Finally, though we found that HRV was different between pre-ictal period and ictal period, further prospective study is needed to determine whether HRV can be use as an index to predict seizure for children with epilepsy.

5 Conclusions

The HRV parameters were different between pre-ictal period and ictal period. The change of frequency domain parameters suggested the predominance of sympathetic system during ictal period. We speculate that HRV has potential to be an index for predicting epileptic seizures in the future. Besides, the abnormal electrical activity of the heart could be detected earlier by HRV monitoring. And we could take interventions to prevent patients from sudden death in some epilepsy patients theoretically.

Conflict of interest

All authors declare no conflict of interest.

Footnotes

Acknowledgments

This work was supported partly by Grant JCYJ20140414170821285, JCYJ20150529164154046 and CXZZ20140909004122087 from Shenzhen Science and Technology Innovation Committee.

References

Allana

S.S.

, Ahmed

H.N.

, Shah

and Kelly

A.F.

, Ictal bradycardia and atrioventricular block: A cardiac manifestation of epilepsy, Oxf Med Case Reports 2014 (2014), 33–35.

Anderson

K.P.

, Vagal control of the heart: Experimental basis and clinical implications, in, LWW, 1995.

Ansakorpi

, Korpelainen

, Huikuri

, Tolonen

, Myllylä

and Isojärvi

, Heart rate dynamics in refractory and well controlled temporal lobe epilepsy, Journal of Neurology, Neurosurgery & Psychiatry 72 (2002), 26–30.

Bateman

L.M.

, Spitz

and Seyal

, Ictal hypoventilation contributes to cardiac arrhythmia and SUDEP: Report on two deaths in video-EEG–monitored patients, Epilepsia 51 (2010), 916–920.

Benarroch

E.E.

, The central autonomic network: Functional organization, dysfunction, and perspective, Mayo Clin Proc 68 (1993), 988–1001.

Bermeo-Ovalle

A.C.

, Kennedy

J.D.

and Schuele

S.U.

, Cardiac and autonomic mechanisms contributing to SUDEP, J Clin Neurophysiol 32 (2015), 21–29.

Brown

, DiFrancesco

and Noble

, How does adrenaline accelerate the heart? (1979).

Camm

A.J.

, Malik

, Bigger

, Breithardt

, Cerutti

, Cohen

, Coumel

, Fallen

, Kennedy

and Kleiger

, Heart rate variability. Standards of measurement, physiological interpretation, and clinical use, Eur Heart J 17 (1996), 354–381.

Chen

, Guo

C.-L.

, Zhang

P.-S.

, Liu

, Qiao

, Zhang

J.-G.

and Meng

F.-G.

, Heart rate changes in partial seizures: Analysis of influencing factors among refractory patients, BMC Neurology 14 (2014), 135.

10.

Chen

, Guo

C.L.

, Zhang

P.S.

, Liu

, Qiao

, Zhang

J.G.

and Meng

F.G.

, Heart rate changes in partial seizures: Analysis of influencing factors among refractory patients, BMC Neurol 14 (2014), 135.

11.

Convertino

V.A.

, Neurohumoral mechanisms associated with orthostasis: Reaffirmation of the significant contribution of the heart rate response, Front Physiol 5 (2014), 236.

12.

H.M.

, Boer, “Out of the shadows”: A global campaign against epilepsy, Epilepsia 43(Suppl 6) (2002), 7–8.

13.

Delogu

A.B.

, Spinelli

, Battaglia

, Dravet

, De Nisco

, Saracino

, Romagnoli

, Lanza

G.A.

and Crea

, Electrical and autonomic cardiac function in patients with Dravet syndrome, Epilepsia 52 (2011), 55–58.

14.

Derby

L.E.

, Tennis

and Jick

, Sudden unexplained death among subjects with refractory epilepsy, Epilepsia 37 (1996), 931–935.

15.

DiFrancesco

, Ferroni

, Mazzanti

and Tromba

, Properties of the hyperpolarizing-activated current (if) in cells isolated from the rabbit sino-atrial node, The Journal of Physiology 377 (1986), 61–88.

16.

Kolsal

, Serdaroglu

, Cilsal

, Kula

, Soysal

A.S.

, Kurt

A.N.

and Arhan

, Can heart rate variability in children with epilepsy be used to predict seizures?, Seizure 23 (2014), 357–362.

17.

Kolsal

, Serdaroğlu

, Çilsal

, Kula

, Soysal

A.Ş.

, Kurt

A.N.Ç.

and Arhan

, Can heart rate variability in children with epilepsy be used to predict seizures?, Seizure 23 (2014), 357–362.

18.

Lathers

C.M.

and Schraeder

P.L.

, Autonomic dysfunction in epilepsy: Characterization of autonomic cardiac neural discharge associated with pentylenetetrazol-induced epileptogenic activity, Epilepsia 23 (1982), 633–647.

19.

Malpas

S.C.

and Purdie

G.L.

, Circadian variation of heart rate variability, Cardiovasc Res 24 (1990), 210–213.

20.

Mehvari

, Fadaie

, Omidi

, Poorsina

, Najafi Ziarani

, Gharekhani

, Beladimoghadam

, Rahimian

, Moein

, Barakatain

, Basiratnia

and Hashemi Fesharaki

S.S.

, Cardiac arrest associated with epileptic seizures: A case report with simultaneous EEG and ECG, Epilepsy Behav Case Rep 2 (2014), 145–151.

21.

Nashef

, Walker

, Allen

, Sander

J.W.

, Shorvon

S.D.

and Fish

D.R.

, Apnoea and bradycardia during epileptic seizures: Relation to sudden death in epilepsy, J Neurol Neurosurg Psychiatry 60 (1996), 297–300.

22.

Palma

J.-A.

, Urrestarazu

, Lopez-Azcarate

, Alegre

, Fernandez

, Artieda

and Iriarte

, Increased sympathetic and decreased parasympathetic cardiac tone in patients with sleep related alveolar hypoventilation, Sleep 36 (2013), 933.

23.

Quigg

, Circadian rhythms: Interactions with seizures and epilepsy, Epilepsy Research 42 (2000), 43–55.

24.

Sarkis

R.A.

, Thome-Souza

, Poh

M.Z.

, Llewellyn

, Klehm

, Madsen

J.R.

, Picard

, Pennell

P.B.

, Dworetzky

B.A.

, Loddenkemper

and Reinsberger

, Autonomic changes following generalized tonic clonic seizures: An analysis of adult and pediatric patients with epilepsy, Epilepsy Res 115 (2015), 113–118.

25.

Shen

M.J.

and Zipes

D.P.

, Role of the autonomic nervous system in modulating cardiac arrhythmias, Circ Res 114 (2014), 1004–1021.

26.

Sillanpaa

and Shinnar

, Long-term mortality in childhood-onset epilepsy, N Engl J Med 363 (2010), 2522–2529.

27.

Stein

P.K.

, Bosner

M.S.

, Kleiger

R.E.

and Conger

B.M.

, Heart rate variability: A measure of cardiac autonomic tone, Am Heart J 127 (1994), 1376–1381.

28.

Trautwein

and Kameyama

, Intracellular control of calcium and potassium currents in cardiac cells, Japanese Heart Journal 27 (1986), 31–50.