Clinical Spectrum and Comorbidities of Dravet Syndrome in Taiwan

Dravet syndrome (DS) is a rare and devastating epilepsy syndrome, and it can affect the patients and their caregivers. However, the lack of a reliable and valid measures of caregiver impact and the characteristic pattern in Taiwan. The purpose of this study was to describe the characteristics of patients with DS and caregivers’ concerns, and to establish a baseline frequency of characteristics of the disease using a cross-sectional survey in Taiwan. We recruit the caregivers of patients with DS and assessed their condition via online anonymous questionnaire. Seizure frequency decreased with age though not statistically signicant. Vaccine may not inuence the condition of DS. We highlighted the greatest impact on the domains that affect caregivers’ daily life, including additional household tasks, symptoms observation, further medical plan and nancial problem. Caregivers may also concern about lack of independence/constant care, seizure control, speech/communication and sibling impacts/long-term care when patents are gone. The current ndings highlight the signicant life effects of caring for a child with DS in Taiwan, and can be used to raise the attention about the need for these families. The possible pathogenic mechanisms of these comorbidities were also discussed.


Introduction
Dravet syndrome (DS), also known as Severe Myoclonic Epilepsy of Infancy, is a rare and devastating epilepsy syndrome. The prevalence rate was estimated to be approximately 1 in 20,000 to 1 in 40,000 children. [1][2][3] Associated mutations of the SCN1A gene were reported in 75% of cases. The characteristics of DS include frequently prolonged hemi-convulsion, developmental delays, speech impairment, and other comorbidities, including ataxia, circadian rhythm disorder, impaired sleep quality, and autistic-like socialinteraction de cits [4]. DS is frequently accompanied with a range of triggering factors, such as fever, infections, hot bath, and photosensitivity. Although DS is usually pharmacoresistent, a trend toward less severe epilepsy and cognitive impairment is usually observed after the age of 5 years. [5] Although current literature had proven that there is no signi cant difference in the clinical and cognitive outcomes, in the past many parents may worry about vaccination-related seizures. [6][7][8] Due to the knowledge gap about frequency of seizures following vaccination, misperception of vaccination side effects and reduced vaccination coverage still exists in many DS families. [8] This study aims to describe the characteristics of patients with DS and caregivers' concerns, and to establish a baseline frequency of characteristics of the disease using a cross-sectional survey in Taiwan. The data may be used to help researchers and clinicians for further studies, and to more understand this refractory epilepsy and most important issues to patients and their families. The possible pathogenic molecular related to comorbidities were also discussed.

Demographics
We identi ed 38 patients with DS from our study, all of whom had a con rmed mutation in SCN1A. However, only 32 patients knew correct mutations data, consisting of 19 missense mutations in 21 patients, 2 nonsense mutations in 2 patients, 1 splice site mutation in 1 patient, 5 frameshift mutations in 6 patients, and 2 chromosome deletion in 2 patients (Fig. 1). Patients ranged in age from 1 to 28 years (mean ± SD, 10.5 ± 6.3 years). Sixteen patients are female (Table 1). About their language and ambulation evaluation excluding patient age less than 2 years, 51% can speak a clear and correct sentence and 78% of patients can ambulate without assistance.

Seizure evaluation
The mean ages of epilepsy onset were about 9.5 ± 16.1 months (median 6 months). The generalized tonic-clonic seizures, absence seizures and focal seizures were the most common at rst time, and occurred in 66%, 37% and 29%, respectively (Table 1). These seizures were often induced by fever (54%) (Fig. 2a), and caregivers reported that fever, infection, sun exposure, warm bath, after exercise and positive mood were the most common triggering factors that induced seizure attack (Fig. 2b). All of each may be > 50%. Besides, vaccination accounted for 32 %, which may trigger seizure attack in DS patients. Photosensitivity and pattern sensitivity were also triggering factors for seizure, and might occur across different age group and might not disappear when age increases.

Vaccine evaluation
When we compared between vaccination-proximate and vaccination-distant groups, there was no statistical difference between the two groups ( Table 2). Only vaccination in seizure triggering factors had been frequently observed in vaccination-proximate group (P = 0.038).
No signi cant difference was present in the language, ambulation, seizure characteristics ( rst seizure onset, seizure triggering factors except vaccine, seizure pattern, and seizure frequency), and number of anti-epileptic drugs (AEDs) in use. * Current seizure control at last follow-up: "controlled" seizure indicates that the patient has less than no seizures in the past 6 months; "uncontrolled" for patients with more than one and less than six times per months seizures in the past 6 months ; "intractable" for one or more seizures each month in the past 6 months.

Comorbidity evaluation
All questionnaires were reported by 100% of caregivers (Table 3). With the exception of nocturnal seizures, there was little variation in sleep issues among patient age groups.
Both bradycardia and tachycardia were reported in 9% and 3%, respectively. Only one patient had arrhythmia history, and 3 of 36 patients reported abnormalities or changes in the heart structure.
Behavioral and psychiatric issues were commonly reported in the survey, and attention de cit disorder or attention de cit hyperactivity disorder were most complained (61%). Other psychiatric symptoms such as di culty with impulse control and autistic-like traits were also noticed in 39% and 31%, respectively. 42% had ever went to pediatric psychiatric clinic for evaluation. Anxiety and psychosis accounted for one-third of DS patients.
As for musculoskeletal issues, hypotonia (50%) was relatively common in the patients' early childhood. Broken bones (24%) and scoliosis (19%) had increased with age in our data, which were more prevalent from the population of middle childhood to adult, compared with that of infant to preschool.
Constipation (47%) was another common issue. About one-third of DS patients had appetite disturbance and frequent/chronic urinary tract infection.
Drowsiness, cognition problem and unsteady gait were most common drug side effect, and occurred in 46, 47 and 49%, respectively. The unsteady gait may also arise from the disease entity of DS.
In most patients, multiple AEDs were needed, and 78% of patients may need more than 3 drugs for seizure control. 100% (12/12) of responders with children age 12 years or older reported having the use of more than 3 antiepileptic drugs. Only 5% (2/38) of patients used 1 drug for seizure control and 16% (6/38) of patients needed 2 drugs for seizure control.
Caregiver issues and family dynamics (supplementary tables) Nearly half of caregivers (47%) reported having suffered from depressed mood via questionnaire, but we did not record whether they had received further help or not. When evaluating the caregiver burden scale in each domain, approximately three quarters of caregivers reported moderate or greater di culty in doing additional household tasks (79%), watching and reporting symptoms (77%), and seeking further medical plans (76%). The rest of the items of the questionnaire were nancial problem (66%), medical or nursing treatments (66%), medication use (63%), patient care (58%) and mobility problem (50%). When being asked to rank their top three concerns in an open response, caregivers highlighted the lack of independence (61%), seizure control (58%), speech and communication challenges (50%), and sibling impacts for long-term caring for patients with DS (50%).

Discussion
In this cross-section cohort, we collected data about patients with DS from caregivers in Taiwan, which add to our understanding about the impact of DS patients' condition and DS on caregivers' concerns. It is important to avoid speci c seizure trigger factors in DS. In our study, we found that the most signi cant trigger factor as shown in other studies is hyperthermia. Therefore, the patients should avoid the predisposing factors of hyperthermia, such as overexcitement, overexertion, sun exposure, and warm bath [9]. The family members should also be educated to seek for medical assistance whenever facing hyperthermia. We also noticed that photosensitivity and pattern sensitivity were trigger factors of seizure, similar to that of Villas et. al. reported in 2017 [10].
In our study, we found that vaccine-related seizures were reported in 12 (34%) of 35 patients at our cohort. This was consistent with previous studies showing that one third of patients with DS developed seizures after vaccination [7,8,11] (Table 4). Because there was no statistically signi cant difference between language, ambulation or seizure characteristics for those with and without vaccine-related seizures, the vaccination should not be withheld in patient with DS, and all clinicians should provide families with proper information before administering vaccination. In previous literatures, seizure frequency may decrease with age, which was independent of their SCN1A mutation type [12][13][14][15]. We also found a tendency of decreasing seizure frequency with age, although this did not reach statistical signi cance. In other study, in adolescence and adult groups with DS, fever sensitivity persisted but had less in uence [14].
According to previous studies in Dravet mouse model, seizure susceptibility in DS is caused by reducing the sodium currents and electrical excitability of GABAergic interneurons, which lower the seizure threshold [16,17]. Therefore, the rst-line treatments for DS include valproic acid and clobazam, and the second-line treatment may include stiripentol, topiramate and the ketogenic diet [18]. As shown in Table 5, valproate was the most commonly used AEDs. Clobazem, topiramate, and stiripentol were also used frequently. In contrast, levetiracetam was the third common AED treatment in Taiwan. About the side effect of AEDs, drowsiness, cognition problem, and unsteady gait were most common in our study. In contrast, hematologic side effects, such as thrombocytopenia, neutropenia, or anemia, did not reach signi cance in our study. Nephrocalcinosis caused by topiramate accounted for 3% of the patients, which was similar to other studies [10]. Appetite disturbance and constipation were also noted in our patients, which might be due to AEDs or DS itself.
The characteristic symptoms of DS in our study include nocturnal seizures, hypotonia, drowsiness, cognition problem, unsteady gait, constipation, and psychiatric issues like ADD or ADHD problem, corresponding to previous literatures [10,19]. In our study, caregivers' reports of nocturnal seizure were 51%, the same as one previous study [20], and was less compared with another study, which reported 77% of nocturnal seizure [10]. It indicated that nocturnal seizures are a major concern for most caregivers. Awareness and attention between DS with SCN1A mutations and heart rate/rhythm and abnormalities are increasing in recent years. Cardiac dysfunction leading to sudden death may be a major concern for caregivers [21]. Although cardiac arrhythmia was noted in some of our patients, no one was found to have sudden death due to cardiac problem.
There was limited discussion about the possible pathogenic mechanisms leading to different comorbidities in the past. Dravet syndrome, which is primarily caused by heterozygous loss of function mutation in the SCN1A gene that encodes voltage-gated sodium channel type-1, termed Nav 1.1. Nav1.1, is a member of a family of voltage-sensitive sodium channels expressed primarily in the brain, including Na v 1.1, Na v 1.2, Na v 1.3, and Na v 1.6. As Na v 1.1 expression is very low in neonates, other alpha subunits such as Na v 1.2 and Na v 1.3 may compensate for reduced Na v 1.1 function during this early stage of development [22]. Therefore, as shown in mouse model, a physiologic decline in Na v 1.3 channel expression in brain development, coupled with the failure of increase in functional Na v 1.1 channels in Dravet syndrome, may lead to wide spread dysfunction of neuronal networks, intractable seizures, and comorbidities, including ataxia, sleep quality and autistic-like behaviors and spatial learning and memory [23].
Electrophysiological studies showed that Na v 1.1 plays a crucial role in the excitability of cerebellar Purkinje neurons, with major contributions to peak, persistent, and resurgent forms of sodium current and to sustained action potential ring. Loss of these channels in Purkinje neurons of mutant mice may cause loss of excitability of Purkinje neurons leading ataxia, as shown in other study [24].
Patients with DS had reported sleep problems, including disturbance of sleep time, impaired sleep quality and increased incidence of nocturnal seizures [25]. In DS mouse model, the mechanism of sleep disorders is related to mutation of Na v 1.1 channel in forebrain GABAergic interneurons without involvement of drug treatment and may correlated with cell-speci c loss of sodium current and excitability of reticular nucleus of the thalamus GABAergic interneurons. This result con rms that impaired action potential ring of these GABAergic interneurons is responsible for the defect in sleep quality [26]. Regarding other sleep problems, DS children also have a circadian rhythm defect, which in uences their sleep-wake cycle [25]. Although our result did not show signi cant nding in this eld, DS mice had an abnormally long circadian cycle length, defects in phase shift after change of their light-dark cycle, and impaired lightinduced shifts of their sleep-wake cycle [25]. In heterozygous Scn1a +/− mice study, it showed that reduced Na v 1.1 channel activity impairs suprachiasmatic nucleus of the hypothalamus, the primary site of the circadian clock [27]. This nding supports the relationship between decreased GABAergic transmission and circadian defect [27]. Therefore, some symptoms of DS sleep problems may be treatable by enhancement of GABAergic neurotransmission [27].
Children with DS also exhibit autistic-like behaviors [10]. DS mice also have substantial de cit in social interactions. This de cit may be explained by the speci c deletion of Na v 1.1 channel in forebrain inhibitory neurons, which may reduce action potential ring in inhibitory neurons. Besides, treating with low doses clonazepam may rescue the autistic-like behaviors in DS mice, caused by channel dysfunction rather than the epileptic activity itself as contributor [28]. Furthermore, another study also showed that GABAergic interneurons may include parvalbumin-(PV+) or somatostatin-expressing (SST+) interneurons [29]. Therefore, deletion of Nav 1.1 in PV + interneurons may cause social interaction de cit rather than hyperactivity; however, deletion of Na v 1.1 in SST + interneuron may cause hyperactivity behavior. In contrast, synergistic effect of PV + and SST + interneurons were found to impair long-term spatial memory [29]. In these studies, they demonstrated that autistic-related phenotypes and spatial learning de cits resulted from decreasing Na v 1.1 activity in GABAergic interneurons in hippocampus and cortical interneurons [17,28,30].
As we known, sudden unexpected death in epilepsy (SUDEP) is the most common cause of death in intractable epilepsies, but the physiological mechanisms leading to SUDEP are unknown [20]. Although we did not document these events in our study, recent works suggested that SUDEP was caused by parasympathetic hyperactivity following hyperthermia-induced tonic-clonic seizures, which resulted in severe bradycardia and death in the Scn1a +/− mouse model of DS [31]. DS mice had been shown to have ventricular myocytes problems via alterations in neuronal excitability and cardiac electrophysiologic abnormalities, which may contribute to the susceptibility for arrhythmogenesis and SUDEP [32]. Furthermore, reductions in Nav1.1 expression may indirectly affect Nav 1.5 activity and cardiac electrical function [32], leading cardiac problems.
Therefore, regaining the impaired GABAergic neurotransmission may not only lead to improved seizure control but also improved function form prefrontal cortex to cerebellar networks, as has been shown in mice [28,29,33] (Fig. 3).
In recent literatures, more and more studies have focused on caregivers of patients with DS owing to different aspects of stress, and, therefore, may need multidisciplinary team to care about the patients. In our study, caregivers viewed additional household tasks, symptoms observation, further medical plan and nancial problems as signi cant factors. A cohort study in Children's Hospital Colorado, they found caregivers suffered from emotional exhaustion and anxiety related to "fear of the next seizure" and "the seizure that kills my child". Besides, caregivers need to quit their jobs or careers to care for their child with DS due to the severity of condition [34]. In another study in Canada, persistent severe seizures, accompanied with developmental, cognitive, behavioral, and sleeping issues increased caregivers' burden [25]. Most caregivers may also concern about sleep deprivation, reduced mental health, deterioration of social relationships, and nancial burden problem [35].
Caregivers in this study ranked their top 3 major concerns in the future, including lack of independence/constant care, seizure control, speech and communication problem and sibling impacts (long-term care when parents are gone). In another study, caregivers ranked their top 4 concerns, as speech and communication challenges, the impact of the patient with DS on siblings, cognitive/developmental delay, and behavioral issues including violence and autistic traits [10]. Therefore, recent studies have begun to raise the attention about caregivers' need and offer further additional support services for the relief the burden of caring for patients with DS in the hope for improving caregivers physical and nancial well-being.
There were several limitations in our study. Not all patients with DS from Taiwan were enrolled in the present study. In addition, the most common seizure pattern in adolescents and adults was generalized tonic-clonic, and mostly nocturnal and in cluster [14,36]. However, in our study, we did not record the serial seizure changes. We could also not prove that broken bones and scoliosis were positively related to DS or other etiologies, such as vitamin D de ciency, due to no related blood tests in patients.
In conclusion, comorbidities are very common in patients with DS. They are related to the involvement of different regions in the brain. Therefore, detailed evaluation of the patients with DS for possible association of different comorbidities may direct the neurologists to correct treatment in addition to seizures.

Survey design
We recruited the patients and caregivers of patients with diagnosis of DS from most medical centers in Taiwan. All cases had been diagnosed and actively followed up by a pediatric neurologist in different hospitals. An online questionnaire regarding demographic data, gene mutation, clinical features, vaccine use and impact to family was designed and the caregivers and their doctors were requested to ll the form. In this online questionnaire study, all participants in the survey were voluntary, and data were collected anonymously. Permission to use deidenti ed data was obtained prior to participation, and each survey included a demographic session as well as content-related sections about characteristics, possible comorbidities, medications and e cacy, and caregiver/family dynamics, including Oberst Caregiving Burden Scale and caregiver concerns. Oberst Caregiving Burden Scale is a two-dimensional 15-item scale that evaluate the time and di culty in performing each of the items measuring tasks to be done. Responses included lists, closed-question multiple choice, and open response. The study had been approved by the Ethical Committee of National Taiwan University Hospital. Informed consent was obtained when the families answered the questionnaire online. All methods were performed in accordance with the relevant guidelines and regulations.
We also assessed the time period between the rst seizure activity and the previous vaccination. We de ned two groups according to whether seizure activity was within 48 hours (vaccination-proximate group) or not (vaccination-distant group) as previous literatures [7,8].

Statistical analysis
Response frequencies and the mean, standard deviation (SD), and median ranges of interest were summarized using descriptive statistics. Statistical comparisons between groups were performed using Chi-square tests, and a p < 0.05 was considered as signi cant.