If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Pediatric Neurology, Dipartimento di Scienze della Salute della Donna, del Bambino e di Sanità Pubblica, Fondazione, Fondazione Policlinico Universitario A. Gemelli IRCCS, Rome, ItalyDipartimento di Scienze della Vita e Sanità Pubblica, Università Cattolica del Sacro Cuore, Rome, Italy
Pediatrics, Dipartimento di Scienze della Salute della Donna, del Bambino e di Sanità Pubblica, Fondazione, Fondazione Policlinico Universitario A. Gemelli IRCCS, Rome, ItalyDipartimento di Scienze della Vita e Sanità Pubblica, Università Cattolica del Sacro Cuore, Rome, Italy
Pediatric Neurology, Dipartimento di Scienze della Salute della Donna, del Bambino e di Sanità Pubblica, Fondazione, Fondazione Policlinico Universitario A. Gemelli IRCCS, Rome, ItalyDipartimento di Scienze della Vita e Sanità Pubblica, Università Cattolica del Sacro Cuore, Rome, Italy
Pediatric Neurology, Dipartimento di Scienze della Salute della Donna, del Bambino e di Sanità Pubblica, Fondazione, Fondazione Policlinico Universitario A. Gemelli IRCCS, Rome, ItalyDipartimento di Scienze della Vita e Sanità Pubblica, Università Cattolica del Sacro Cuore, Rome, Italy
Pediatric Neurology, Dipartimento di Scienze della Salute della Donna, del Bambino e di Sanità Pubblica, Fondazione, Fondazione Policlinico Universitario A. Gemelli IRCCS, Rome, ItalyDipartimento di Scienze della Vita e Sanità Pubblica, Università Cattolica del Sacro Cuore, Rome, Italy
Pediatric Neurology, Dipartimento di Scienze della Salute della Donna, del Bambino e di Sanità Pubblica, Fondazione, Fondazione Policlinico Universitario A. Gemelli IRCCS, Rome, ItalyDipartimento di Scienze della Vita e Sanità Pubblica, Università Cattolica del Sacro Cuore, Rome, Italy
Pediatric Neurology, Dipartimento di Scienze della Salute della Donna, del Bambino e di Sanità Pubblica, Fondazione, Fondazione Policlinico Universitario A. Gemelli IRCCS, Rome, ItalyDipartimento di Scienze della Vita e Sanità Pubblica, Università Cattolica del Sacro Cuore, Rome, Italy
Pediatric Neurology, Dipartimento di Scienze della Salute della Donna, del Bambino e di Sanità Pubblica, Fondazione, Fondazione Policlinico Universitario A. Gemelli IRCCS, Rome, Italy
Pediatric Neurology, Dipartimento di Scienze della Salute della Donna, del Bambino e di Sanità Pubblica, Fondazione, Fondazione Policlinico Universitario A. Gemelli IRCCS, Rome, ItalyDipartimento di Scienze della Vita e Sanità Pubblica, Università Cattolica del Sacro Cuore, Rome, Italy
Pediatric Neurology, Dipartimento di Scienze della Salute della Donna, del Bambino e di Sanità Pubblica, Fondazione, Fondazione Policlinico Universitario A. Gemelli IRCCS, Rome, ItalyDipartimento di Scienze della Vita e Sanità Pubblica, Università Cattolica del Sacro Cuore, Rome, Italy
Cardiology, Dipartimento di Scienze Cardiovascolari e Toraciche, Fondazione Policlinico Universitario A. Gemelli IRCCS, Rome, ItalyDipartimento di Scienze cardiovascolari e pneumologiche, Università Cattolica del Sacro Cuore, Rome, Italy
Pediatric Neurology, Dipartimento di Scienze della Salute della Donna, del Bambino e di Sanità Pubblica, Fondazione, Fondazione Policlinico Universitario A. Gemelli IRCCS, Rome, Italy
Pediatrics, Dipartimento di Scienze della Salute della Donna, del Bambino e di Sanità Pubblica, Fondazione, Fondazione Policlinico Universitario A. Gemelli IRCCS, Rome, ItalyDipartimento di Scienze della Vita e Sanità Pubblica, Università Cattolica del Sacro Cuore, Rome, Italy
Corresponding author at: Child Neurology Unit, Department of Pediatrics, Fondazione Policlinico Universitario A. Gemelli IRCCS, Largo Francesco Vito 1, 00168 Roma, Rome, Italy.
Pediatric Neurology, Dipartimento di Scienze della Salute della Donna, del Bambino e di Sanità Pubblica, Fondazione, Fondazione Policlinico Universitario A. Gemelli IRCCS, Rome, ItalyDipartimento di Scienze della Vita e Sanità Pubblica, Università Cattolica del Sacro Cuore, Rome, Italy
Patients with Dravet Syndrome display lower heart rate variability compared to healthy controls and patients different epilepsies.
•
Having a lower heart rate variability in DS is predicted by a recent history of SE.
•
Compared to the literature our cohort showed better HRV and lower mortality, reinforcing a possible association between them.
Abstract
Purpose
Preliminary data suggest that patients with Dravet Syndrome (DS) have a reduced heart rate variability (HRV). This seems particularly evident in patients who experienced sudden unexpected death in epilepsy (SUDEP). This study aims at confirming these findings in a larger cohort and at defining clinical, genetic or electroencephalographic predictors of HRV impairment in DS patients.
Methods
DS patients followed at our Institution performed a 24h-ECG Holter to derive HRV parameters. We used as control population patients with epilepsy (PWEs) and healthy controls (HCs). In DS patients, we assessed the impact of different clinical, neurophysiological and genetic features on HRV alterations through multiple linear regression. After a mean follow-up of 7.4 ± 3.2 years since the HRV assessment, all DS patients were contacted to record death or life-threatening events.
Results
56 DS patients had a significantly reduced HRV compared to both HCs and PWEs. A recent history of status epilepticus (SE) was the only significant predictor of lower HRV in the multivariate analysis. At follow-up, only one patient died; her HRV was lower than that of all the controls and was in the low range for DS patients.
Conclusion
We describe for the first time an association between SE and HRV alterations in DS. Further studies on other SCN1A-related phenotypes and other epilepsies with frequent SE will help clarify this finding. Compared to the literature, our cohort showed better HRV and lower mortality. Although limited, this observation reinforces the role of HRV as a biomarker for mortality risk in DS.
Dravet syndrome (DS) is a developmental and epileptic encephalopathy, characterized by epilepsy, cognitive decline, behavior disorders and motor impairment [
], caused by a mutation in SCN1A in almost all the patients. SCN1A encodes the α1 subunit of the voltage-gated sodium channel (Nav1.1) and its loss of function determines a dysfunction of GABAergic interneurons with consequent impairment of inhibitory pathways [
]. The mortality in DS is very high, not only compared to healthy controls but also to other patients with drug-resistant epilepsy (DRE), who are about three times less likely to die [
]. A metanalysis that included 177 fatal cases of patients with DS revealed that sudden unexpected death in epilepsy (SUDEP) was the leading cause of death (49%, n = 87), followed by status epilepticus (SE, 32%, n = 56) [
]. A Japanese study reporting a cohort of 623 DS patients found similar findings, also reporting that the risk of death is higher in children than in adults [
]. High seizure burden, presence of generalized tonic-clonic seizures (GTCSs), use of several antiseizure medications and developmental disability have been recognized as risk factors for SUDEP and are all typical features of DS [
]. A high fatality rate has also been shown in animal models of DS. Different underlying mechanisms such as cardiac dysfunction, respiratory dysfunction and post-ictal brainstem depression have been hypothesized [
]. In mouse SUDEP models, death due to cardiorespiratory arrest was induced by triggering seizures and spreading depolarization involving the dorsal medulla, a key brainstem area for respiratory and cardiac autonomic control [
]. The involvement of the central autonomic system in SUDEP pathophysiology is also supported by postmortem investigations in DS patients, which demonstrated a deficiency in monoaminergic neurons of the brainstem [
Heart rate variability (HRV) is an established proxy measure of cardiac autonomic nervous system activity. A low HRV is a strong, independent predictor of future acute cardiovascular events and of all-cause mortality in elderly people [
Heart rate variability from short electrocardiographic recordings predicts mortality from all causes in middle-aged and elderly men. The Zutphen Study.
] it has become important to identify whether there are clinical and EEG features that may be possibly associated with HRV.
The aim of this study was i) to assess heart rate variability in DS patients in relation to patients with other forms of epilepsy and healthy controls and ii) to identify clinical, neurophysiological and genetic factors that may be associated with altered HRV and, potentially, with higher risk of SUDEP.
2. Methods
2.1 Patients
We consecutively enrolled patients diagnosed with DS attending the department of Child Neurology of Policlinco Universitario A. Gemelli (Rome) between 2008 and 2016. The diagnosis was based on “the occurrence of febrile and afebrile, generalized and unilateral, often prolonged in status epilepticus, clonic or tonic-clonic seizures in the first year of life in an otherwise normal infant, later associated with other types of seizures and developmental decline” [
] there is substantial lack of normative data acquired with our methodology (i.e. 24h-Holter ECG). We therefore decided to include a group of healthy controls (HC) with similar age and gender distribution. They were children referred to our Pediatric Cardiology Unit for complaints such as chest pain or palpitation, who were found to have a normal cardiac examination, electrocardiogram and standard transthoracic Doppler echocardiography. We also included a second control group of patients with other forms of epilepsy (PWE) other than DS but taking similar antiseizure medications to account for their effect. PWE were studied for SCN1A mutations and were all negative. Data from PWE and HC were already published in our previous study [
Parents or guardians of all patients agreed to take part in the study and gave informed consent according to the Declaration of Helsinki.
2.2 Study design
2.2.1 Standard ECG and Holter-ECG
All subjects underwent a standard 12-lead ECG exam and at least 24 h of Holter-ECG monitoring. The QT interval was measured on standard 12-lead ECG, and the corrected QT interval (cQT) was calculated according to the Bazzett formula.
Holter ECG recordings were performed using 3-channel recorders (Medilog FD5 Plus, Schiller, Milan Italy), with a sampling rate of 32,000 Hz. Holter recordings were analyzed to extract heart rate variability (HRV) parameters over the entire 24 h using the software Darwin 2.0 (Schiller, Milan Italy).
HRV was analyzed both in the time and frequency domain (Figure 1). Intervals including premature ventricular beats were interpolated based on previous and following normal-to-normal (NN) intervals; NN intervals shorter by >20% compared to the previous NN interval (premature supraventricular beats) were excluded.
The following parameters were obtained from time-domain HRV analysis: 1) the standard deviation of all NN QRS intervals (SDNN), and 2) the square root of the average of the squares of differences of successive NN intervals (RMSSD). HRV in the frequency-domain was assessed by fast Fourier transform analysis and the amplitude of the variations in NN intervals in the following frequency ranges were obtained: 1) very low frequency (VLF, 0.0033–0.04 Hz); 2) low-frequency (LF, 0.04–0.15 Hz); 3) high frequency (HF, 0.15–0.40 Hz). The average heart rate (HR) over the whole recording was also considered.
2.2.2 Video-EEG
Within 2 days before or after the Holter-ECG recording all patient underwent a 21-electrodes nap video-EEG recording of at least 1 h. The recording was labelled as “markedly abnormal” if interictal epileptic discharges were present for at least 20% of the recording.
2.2.3 Genetic testing
All patients underwent Sanger sequencing and MLPA for SCN1A. If those were negative they were tested with an NGS epilepsy gene panel. SCN1A variants were classified as being either (1)frameshift or truncating, (2)missense or (3)splice site mutations.
2.2.4 Clinical records
Patients demographics, including age and sex, were duly recorded. Clinical details regarding seizures and medications history were also recorded. These included:
1
History of myoclonic seizures, with subsequent label of complete or incomplete phenotype, as defined in existing literature [
History of recent convulsive status epilepticus, as defined by at least one convulsive seizure lasting more than 15 min within the three months preceding the long-term ECG recording.
3
Presence of frequent tonic-clonic seizures (having >3 GTCSs/year vs ≤3 GTCSs/year)
4
Type and number of antiseizure medications (use vs. no use of individual ASMs taken by more than 10 patients; total number of ASMs).
Cut-offs for rates of GTCS and ASMs were drawn from large epidemiological studies on SUDEP that demonstrated they are significant steps in the risk distributions [
In 2020, after 7.2 ± 3.2 years following the electrophysiological study (range 3.2–12.1 years), all the patients were contacted to establish the occurrence of death or near-SUDEP events, recently defined as a resuscitation for more than 1 h after a cardiorespiratory arrest that has no structural cause identified after investigation [
The SPSS 26.0.1 package for IOS was used to analyze the data. Due to relatively small and uneven samples, non-parametric tests were used for comparison between groups. To compare individual HRV parameters among groups (DS, HC, PWE) we used Kruskal-Wallis test and Mann-Whitney test for pairwise post-hoc comparisons. Multiple comparisons were adjusted with Bonferroni correction. A p-value < 0.05 was considered significant.
Exclusively in the DS group, we looked for predictors of an alteration in HRV. We choose as outcome measures SDNN and RMSSD as time-domain parameter and HF as frequency-domain parameter, three reliable and commonly used HRV descriptors. We tested the role of factors/covariates through multiple linear models corrected by age, which is known to affect multiple HRV variables. We then selected the most significant factors to build the multiple linear model with the highest proportion of explained variance. Diagnostics for collinearity and heteroskedasticity have been successfully computed.
To interpret the results of deaths and near SUDEP events, the HRV parameters with the least age-dependency (RMSSD, HF) were used to compare individual values, presented as percentile ranks, with the HC and DS cohorts.
3. Results
3.1 Study population
3.1.1 DS
For the 56 DS patients mean age at enrollment was 7.2 ± 5,5 years (age range 1–23 years); 30/56 were female. 42/56 (75%) patients had a complete form. They were all taking a mean of 2.4 ± 0.9 ASMs (range 1–4); 45 were on sodium valproate, 28 on clobazam, 24 on topiramate 13 on stiripentol, 5 were on ethosuximide, levetiracetam or clonazepam, 2 on phenobarbital or zonisamide, 1 on perampanel, acetazolamide, nitrazepam, potassium bromide and sulthiame. One patient was on ketogenic diet. A pathogenic variant in SCN1A was found in 53/56 patients (95%). In one patient the NGS panel detected a homozygous mutation in SCN1B; in the other 2 patients, NGS studies did not reveal any mutation in the genes known to be asssociated with epilepsy.
3.1.2 PWE
The 37 PWE had a mean age of 7.0 ± 4.3 years. 18/37 were females. They had various types of epilepsy: 27 patients had focal, 4 had generalized and 6 had a combined focal and generalized epilepsy. The etiology was unknown in 19 patients, structural in 15 patients and genetic in 3 patients (16p11.2 duplication,PTE-encephalopathy, Pitt-Hopkins syndrome). A syndromic diagnosis was possible in 12 patients: 8 Childhood epilepsy with centrotemporal spikes, 2 Childhood absence epilepsy, 1 Juvenile myoclonic epilepsy, 1 Myoclonic-atonic epilepsy (Doose Syndrome). 20/37 (54%) were taking ASMs: 20 sodium valproate, 3 clobazam, 5 topiramate. Only 5/37 patients had DRE.
3.1.3 HC
Twenty-four consecutive healthy children, referred to our pediatric cardiology unit for a clinical check-up and who agreed to participate in the study, were enrolled in the HC group. The mean age of HC was 5.9 ± 4.2 years (age range 1–17 years) and 14/24 were females. They were screened through a pediatric visit and echocardiogram for general pediatric and cardiac conditions and found to be healthy. They also were not taking any kind of drugs.
In terms of demographic features, there were no statistically significant differences among groups (see Table 1).
Table 1Demographic features of the different groups are presented as means and standard deviations. The p-values refer to the results of Kruskal-Wallis test for age comparison and of Chi-square test for sex. DS= Dravet Syndrome, HC= Healthy Controls, PWE= Patients With Epilepsy.
Table 2 shows details of HRV variables in the 3 groups. DS patients had a significantly lower HRV both compared to HC and PWE, in all key parameters, including SDNN, RMSSD, LF and HF. There was no difference (p = 1.00) between HC and PWE, even when considering PWE with and without ASMs.
Table 2Comparison between groups for each HRV parameter. Kruskal-Wallis test and Mann-Whitney test are presented for group and pairwise post-hoc comparisons, respectively. HRV=heart rate variability. RR=average of RR intervals, HR=heart rate, SDNN=standard deviation of Normal-to-Normal(NN) QRS intervals, SDNNI= standard deviation of Normal-to-Normal(NN) index, PNN50= Percentage of successive NN QRS intervals that differ by more than 50 ms, RMSSD Root mean square of successive RR interval differences, VLF= very low-frequency band power, LF=low-frequency band power, HF=high-frequency band power, QTc=corrected QT-interval. DS= Dravet Syndrome, HC= Healthy Controls, PWE= Patients With Epilepsy.
The average heart rate (HR) was higher in DS patients compared to HC and PWE (DS 100.82 ± 15.85, PWE 93.54 ± 13.42, HC 94.63 ± 11.74). No significant differences in QTc length were observed (DS 377.20± 18.48, PWE 380.05 ± 20.49, HC 378.58 ± 21.80).
PWE taking ASMs were also analyzed separately and no significant difference with HC and those taking no ASMs was found.
3.3 Predictors of HRV alterations in patients with Dravet Syndrome
Table 3 shows the weight of the different considered factors in predicting different HRV parameters. Age was a significant factor for most of the HRV parameters while was borderline significant for RMSSD and not significant for VLF. Gender was never a significant factor. When taken individually, significant predictors of HRV were age, recent history of SE and frequent GTCS. When considered together with multiple regression only recent history of SE remained significant in predicting a decrease of HRV (HF −13.02, 95% C.I. −22.13;−3.91; SDNN −26.45 ms, 95% C.I. −42.91;−9.99). The three factors together explained more than 55% and 35% of the variance of SDNN and HF respectively, the former being more sensitive to age. Stiripentol treatment was associated with SDNN reduction and was only borderline significant for HF in simple regression analysis but did not survive correction for the presence of recent SE, a significantly correlated occurrence. History of myoclonic seizures, type of variant, markedly abnormal EEG recordings, total number and individual ASMs were not significant factors in determining HRV.
Table 3Predictors of HRV alterations. β = regression coefficient, R2= effect-size measure, HF= high-frequency power, RMSSD= Root mean square of successive RR interval differences, SDNN=standard deviation of Normal-to-Normal(NN) QRS intervals, SE= status epilepticus, GTCS= generalized tonic-clonic seizure. The reference type of variant is frameshift and truncating mutations vs missense and splice-site variations.
Table 4Mortality and HRV in DS.the table summarizes the published data regarding Mortality in DS (top) and HRV (bottom). Age is presented as mean SD. NA= Not Available.
All the patients in the study were successfully contacted after a mean follow-up of 7.2 ± 3.2 years. Only one patient (female, 7-year-old) with DS died of fatal cerebral edema causing mass effect after a fever-associated SE, a well recognized cause of death in DS [
]. Her HRV parameters (Figure 2) were well below those of any HC (<0.01 percentile) and also in the lower range of the DS cohort (RMSSD 29th percentile, HF 27th percentile). It is interesting to note that her sister, who had a much milder epilepsy despite sharing the same mutation, had an HRV in the same range of HC (RMSSD 55th percentile, HF 49th percentile) and in the higher range for DS patients (RMSSD 73th percentile, HF 68th percentile). No other fatality nor near SUDEP events were reported.
Fig. 1The neural communication pathways between the brain and the heart are responsible for the generation of HRV: a pictorial representation of how the brain controls heart rythms through the sympathetic and parasympathetic networks, together with peripheral inputs from chemoceptors and mechanoceptors. On the right is schematized the process of analysis of the electrocardiogram to derive HRV parameters; these can be categorized into time and frequency derivations. HRV=heart rate variability. SDNN=standard deviation of Normal-to-Normal(NN) QRS intervals, SDNNI= standard deviation of Normal-to-Normal(NN) index, PNN50= Percentage of successive NN QRS intervals that differ by more than 50 ms, RMSSD Root mean square of successive RR interval differences, TPF= total power frequency, ULF=ultra low-frequency band power, VLF= very low-frequency band power, LF=low-frequency band power, HF=high-frequency band power. Created with BioRender.com.
Our study reports the largest cohort published so far of DS patients in whom HRV changes were systematically recorded. We confirmed that patients with DS, when considered as a group, have lower values of HRV compared to healthy controls of similar age. By studying patients with different epilepsies but taking similar treatments we also suggest that this difference is not driven by ASMs. This finding, firstly proposed by our group in a study with 20 patients [
] and was recently replicated in a study on 38 patients with DS and two additional patients with other Na-channel mutation associated encephalopathies [
In our cohort, we only had one case of death from acute encephalopathy after SE while we did not have any death related to SUDEP, at variance with the findings by Cooper et al. reporting a much higher rate of mortality, mainly related to SUDEP (Table 4). Our data, reporting low mortality due to SUDEP, despite lower HRV values, appears to be in disagreement with previous papers reporting the association between SUDEP and low HRV. Myers et al. gathered data from 9 DS patients and one SCN2A-related encephalopathy patient who died of SUDEP. They were then able to demonstrate that the patients who died of SUDEP had lower HRV, particularly during the awake state, suggesting that an altered HRV might be a biomarker of a higher risk of SUDEP in DS.
This difference may only be partly due to our relatively short follow-up, as for most of the studies on DS, the majority of our cohort is represented by young patients, and little is inferable on the role which HRV could have been in older DS patients. The risk of death, however, has been reported to be higher in children than in adults [
] and in our cohort the mean age of the patients at the last follow-up was 15.1 ± 6.4 years suggesting that the length of follow-up in our cohort may not be the only reason for the low mortality rate.
A possible alternative explanation may come from the fact that in our cases the median HRV, even if lower than controls and other epilepsy cases, was higher than in the SUDEP cases reported by Myers, even if the comparison should be interpreted with caution because of the different methodology used by Myers et al. (5mins RMSSD vs 24 h). We can therefore hypothesize that these ‘intermediate’ HRV values, higher than in other DS cohorts, may reduce the risk of SUDEP. In fact, the only patient who died in our cohort had an HRV at the lower end of the spectrum found in the DS patients (21st percentile).
These findings highlight the importance of monitoring HRV and prompted the need to better investigate the possible underlying mechanisms that may have contributed to affecting HRV. A recent history of SE was the most robust clinical factor predicting a lower HRV, independently from all the possible confounders and for all HRV parameters. Frequent GTCSs and stiripentol were significantly associated with some of the HRV parameters but were not significant when corrected by other factors. Conversely, the presence of myoclonic seizures, frequent EEG abnormalities, the number and type of medications prescribed and the type of SCN1A mutation were not significantly associated with HRV changes.
These factors have not been previously systematically assessed as possible predictors of HRV in DS. However, their role in determining autonomic dysfunction and, in the most severe cases, SUDEP has already been reported. SE is the second leading cause of death in DS. Evidence that prolonged seizures result in neuronal loss is longstanding [
]. Parvalbumin and somatostatin-positive interneurons, the cell types mostly damaged by SCN1A mutation, are also the most vulnerable to the detrimental effects of SE [
Progression of spontaneous seizures after status epilepticus is associated with mossy fibre sprouting and extensive bilateral loss of hilar parvalbumin and somatostatin-immunoreactive neurons.
]. Cell loss in the ventrolateral medulla, the region responsible for cardiorespiratory autonomic pacemaking, has been demonstrated in patients with DS who died of SUDEP or SE [
]. Therefore it is possible that recurrent and particularly prolonged seizures might irreversibly damage vulnerable structures responsible for central control of heart rhythms, resulting in the alteration of HRV. Moreover, in animal models of SE, it was demonstrated an altered HRV with an imbalance towards an excessive sympathetic tone and myocardial remodeling, with altered calcium homeostasis [
It is also possible a reverse and partially indirect causal link between autonomic dysfunction and SE. It is established that the inhibitory dysfunction seen in DS is widespread and involves the entire central nervous system [
]. This leads to the vast corollary of characteristic signs and symptoms of DS, from epilepsy to sleep dysfunction, from movement disorders to intellectual disability [
]. Seizure termination is driven by a complex brainstem inhibitory network comprising the substantia nigra pars reticulata and the reticular activating system, among others [
]. It is, therefore, reasonable to speculate that the same brainstem dysfunction that causes a predisposition to SE could also cause dysfunction in cardiorespiratory pace-making which manifests itself as an HRV alteration. A unifying hypothesis could be that SE might further aggravate the brainstem autonomic dysfunction already present due to SCN1A mutation. We suggest that future studies investigating HRV in different SCN1A-related epilepsies and other epilepsies with frequent SE will help disentangle the mechanisms behind the relationship between SCN1A, HRV and SE.
To recognize a link between the seizure burden and cardiac autonomic vulnerability would have important practical consequences in mortality prevention. In particular, we could speculate that drugs that significantly reduce the risk of SE could have a significant impact in reducing DS mortality. Individual ASMs used in our patients did not seem to have a direct impact on HRV. However, it is not possible to draw definite conclusions due to the small sample size of the subgroups. Novel ASMs which have higher efficacy in reducing seizure burden and SE in DS, deserve further investigation.
SUDEP is a rare event and recording detailed clinical information in a large cohort is challenging. A reliable biomarker of SUDEP risk would be highly beneficial to design new therapeutical approaches in patients at higher risk. In this context, HRV seems to be a consistent, non-invasive and inexpensive biomarker for autonomic dysfunction. It can be extracted from routine video-EEG recordings or obtained by wearable sensors which monitor heart rate. This would facilitate the acquisition of large datasets and investigate HRV's role in patients with a wide range of disease severity and in a longitudinal manner. Longitudinal assessments would significantly help to reinforce the correlation between HRV and epilepsy severity at the single-subject level and would better account for the effect of ASMs modifications.
Indeed, one limitation of the study is that electrophysiological data is cross-sectional. Our study explored only some superficial features of the genetic and EEG data, which could be further explored. Furthermore, HRV parameters were extracted from ECG Holter recordings, without concomitant EEG monitoring. This does not allow a precise separation of sleep and wake states, which have instead been proposed as important methodological requirements in the recent recommendations for the assessment of HRV in studies on epileptic patients [
]. Unfortunately, our study was conducted before these recommendations were available. However, 24h-monitoring has the advantage that parameters are averaged over longer traces, reducing variability due to random factors such as seizures or circadian variations; moreover, traces closely represent fluctuations due to habitual daily life activities (Fig. 2).
Fig. 2HRV in DS vs HC and PWE: the majority of patients with DS have lower values compared to both HC and PWE. The black arrow indicates the patient who died of fever-associated acute encephalopathy. DS= Dravet Syndrome, HC= Healthy Controls, PWE= Patients With Epilepsy.
Regarding the selection of the most significant parameters to use in the analysis, there is little data to guide the selection of parameters more sensitive to factors of interest (i.e. patients vs controls) and least sensitive to confounders (i.e. age). In our experience, we found that RMSSD was the least affected by age changes and could therefore be suitable to compare heterogeneous age groups and to track longitudinal changes. Other parameters, such as SDNN, HF and LF, were more sensitive in detecting differences between patients and controls. At the same time, we want to underline that the most significant results were replicated across all parameters considered. Therefore, we suggest that sharing complete data regarding the HRV parameters analyzed would help to draw comparisons across studies.
4. Conclusion
In conclusion, patients with DS exhibit a reduced HRV both compared to HC and to PWE. Some factors are associated with HRV alterations, such as a recent history of SE. Frequent GTCSs and the use of stiripentol deserve further investigation. Contributions from the genetic background and brain electrical activity could be further explored. A link between the seizure burden and cardiac autonomic dysfunction could explain these findings. HRV could therefore provide a deeper characterization of DS patients allowing the identification of patients at higher risk of mortality. Longitudinal prospective studies in larger cohorts will help to establish its value and its limitations.
Declaration of Competing Interest
None of the authors has any conflict of interest to disclose.
We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.
Aknowledgment
We thank the patients and their families for their paticipation in the study. We aknowledge the support by the Associazione AREF Onlus.
References
Dravet C.
Bureau M.
Oguni H.
Cokar O.
Guerrini R.
Dravet Syndrome (previously severe myoclonic epilepsy in infancy).
Epileptic Syndromes in Infancy, Childhood and Adolescence - 6th ED. 2019: 139-171
Heart rate variability from short electrocardiographic recordings predicts mortality from all causes in middle-aged and elderly men. The Zutphen Study.
Progression of spontaneous seizures after status epilepticus is associated with mossy fibre sprouting and extensive bilateral loss of hilar parvalbumin and somatostatin-immunoreactive neurons.
00383-6&id=gr1.jpg){kind=link}
00383-6&id=gr2.jpg){kind=link}