Address correspondence to Nicola Specchio, MD, Division of Neurology, Bambino Gesù Children’s Hospital – IRCCS, P.za S. Onofrio, 4 00165 Rome, Italy. E-mail: firstname.lastname@example.org
Purpose: To explore the clinical, electroencephalography (EEG), neuropsychological features, and prognosis of Panayiotopoulos syndrome (PS).
Methods: Of 1,794 children aged between 1 and 14 years referred for the first afebrile focal seizure, between January 1992 and December 2004, 93 (5.2%) had PS according to clinical criteria.
Results: Age at onset ranged from 1.1 to 8.6 years, and was earlier in children with more than one seizure. Autonomic seizures followed a stereotypical onset and progression. Emesis, pallor, or flushing was almost always among the first symptoms that usually culminated in vomiting (77.4% of patients). More than half (55%) of seizures were longer than 30 min but these did not appear to affect remission and number of seizures. Interictal EEG showed great variability, with 79.5% of patients showing spikes of variable localizations and evolution over time; 16.1% had background abnormalities only, and 5.4% had consistently normal EEG studies. Onsets in five ictal EEGs were posterior or anterior—left or right. On neuropsychological testing, IQ and subtests of Wechsler Intelligence Scale for Children–Revised (WISC-R) were within normal limits, although some minor statistically significant differences were found in arithmetic, comprehension, and picture arrangement in comparison with controls. Cumulative probability of recurrence was 57.6%, 45.6%, 35.1%, and 11.7% at 6, 12, 24, and 36 months, respectively, after the first seizure. Thirty-four (58.6%) of 59 patients treated with antiepileptic drugs continued having seizures before ultimate remission.
Discussion: PS is a uniform childhood susceptibility to autonomic seizures that is related to early age of development and with excellent prognosis with regard to seizure remission and neuropsychological development.
Panayiotopoulos syndrome (PS) is a significant childhood epilepsy because of high prevalence, unusual clinical features disguised as nonepileptic events, and overall excellent prognosis, despite the frequent occurrence of autonomic status epilepticus (ASE). The clinical, electroencephalography (EEG), and magnetoencephalographic features, and the prognosis of PS have been confirmed through independent studies of more than 500 patients (Ferrie et al., 2006, 2007; Panayiotopoulos et al., 2008; Michael et al., 2010). It appears that PS is a model of autonomic epilepsy specific to childhood, characterized by infrequent autonomic epileptic seizures, peak age at onset in early childhood, and an EEG with marked variability of spike localizations.
We wish to contribute to the understanding of the true dimensions of PS by reporting electroclinical and neuropsychological findings in patients with PS.
We have retrospectively reviewed the medical records of 1,794 children between 1 and 14 years of age referred for the first afebrile focal seizure, between January 1992 and December 2004, to the Neurology Division of the Bambino Gesù Children’s Hospital in Rome. This is a private nonprofit research organization, providing health care and research on behalf of the Italian National Health System. It receives secondary patients for general pediatrics and tertiary patients for specialist consultations. The majority of patients (∼70%) are from the local Region (Lazio); the remaining patients come from other Italian regions and 1.3% come from abroad.
Inclusion and exclusion criteria
In defining inclusion and exclusion criteria we emphasize clinical seizure manifestations. Autonomic seizures are the prerequisite for diagnosis as defined in consensus reports (Ferrie et al., 2006, 2007). These should be of unequivocal epileptic origin, last for 5 min or longer, and be idiopathic. Therefore, patients were included if they had one or more unequivocal autonomic epileptic seizures with or without ictal vomiting.
To ensure unequivocal epileptic origin, we excluded patients with nonepileptic paroxysmal autonomic manifestations only.
To ensure idiopathic cause, we excluded patients with symptomatic epilepsy and preexisting neurodevelopmental abnormalities. However, we have included five patients with abnormal magnetic resonance imaging (MRI) because we consider that this is a coincidental finding.
We did not consider EEG abnormalities and localization as essential inclusion or exclusion criteria in order to comply with current evidence that PS is mainly a multifocal epileptic syndrome with significant EEG variability (Panayiotopoulos et al., 2008; Michael et al., 2010).
Patient population and assessments
Ninety-three patients met inclusion and exclusion criteria. We analyzed pre- and perinatal antecedents, familial history of epileptic seizures, circadian distribution, seizure semiology and duration, termination of seizure (spontaneous, with drugs, or requiring intensive care unit hospitalization), EEG and brain imaging, treatment, and prognosis. The children were evaluated during follow-up every 3–6 months.
EEG was performed in all cases within 24 h from first seizure and at each follow-up visit. EEG abnormalities were classified in terms of localization, frequency, and modification at follow-up.
Neuropsychological assessment was possible for 17 patients using the Wechsler Intelligence Scale for Children–Revised (WISC-R), and the data were compared with an age- and sex-matched control group of 19 healthy children. All verbal and performance subitems (Information, Similarities, Arithmetic, Vocabulary, Comprehension, Picture completion, Picture arrangement, Block design, Object assembly, Coding) were considered for the analysis of results.
Student’s t-test was used to compare continuous variables (or Mann-Whitney if the distribution was significantly different from the normal one with Kolmogorov-Smirnov test), and chi-square for categorical variables (or Fisher’s exact test, as appropriate). Univariate analysis of variance (ANOVA) was applied to compare means of seizure duration, with Bonferroni post hoc test. Product-limit method was applied to describe prognosis (Kaplan-Meier curves), after classifying as failures the patients without recurrence of seizures at least in the 24 months following the first observation. SPSS software was used (version 13; SPSS Inc., Chicago, IL, U.S.A.) and p < 0.05 was considered statistically significant.
Demographic and other characteristics of the 93 patients with PS are shown in Table 1. Prevalence was 5.2% among children between 1 and 14 years of age with a first afebrile focal seizure. Girls (55.9%) slightly predominated (boys 44.1%).
Table 1. Demographic and clinical characteristics
FS, febrile seizure; SD, standard deviation.
Family history (FS/epilepsy)
Age at onset
Age Groups at onset (years)
Total number of seizures
93 (100 )
Age at onset ranged from 1.1–8.6 years (mean 4.4 ± 1.6 years). Three-fourths (71%) started between age 3 and 6 years (Fig. 1). Children with more than one seizure had a younger age at onset than those with a single seizure (3.9 vs. 5.4 years, p < 0.001).
Antecedents and family history of epilepsy
Six patients (6.5%) had febrile seizures, five (5.4%) had delivery distress at birth, and one (1.1%) was premature. One child with two autonomic seizures at age 5 years had a history of a brief spell of unverified epileptic spasms at infancy but with normal development and seizure outcome at follow-up at age 13.
A family history of epilepsy was reported in 54.2% of patients; this was 10.8% among first-degree relatives.
Statistical analysis failed to identify any significant difference regarding duration of first seizure, number of seizures, seizure semiology, and EEG findings between patients with a familial history of seizures and those without. However, age at onset appeared to be lower in patients with a first-degree familial history of seizures than in others, although the difference was not statistically significant (3.79 ± 1.09 years vs. 4.50 ± 1.68 years, p = n.s.).
In addition, there were no statistically significant differences between patients with delivery distress at birth and prematurity and those without.
Only the first seizure of each patient was assessed because the semiology and duration of subsequent seizures could have been modified by rescue benzodiazepines or antiepileptic medications. Table 2 shows data regarding timing of occurrence, termination of seizure, duration, and semiology.
Table 2. First seizure characteristics
aVersus Drugs no significant difference, versus Emergency Department p=0.007.
bVersus Emergency Department p=0.05.
cEach patient can experience more than one symptom.
28 ± 15a
31 ± 19b
44 ± 21
Impairment of consciousness
As determined by inclusion criteria, all patients had autonomic seizures. In seizures witnessed from onset, autonomic signs such as malaise, nausea, retching, pallor, or flushing were almost always the first ictal symptoms, with children complaining as “feeling sick,”“I want to throw up,” or “tummy-ache.” Vomiting was the most common autonomic manifestation that occurred in 72 patients (77.4%).
Consciousness at seizure onset was usually intact, becoming gradually impaired during the seizure course, with some fluctuations of a variable degree of severity, from mild confusion to total unresponsiveness in most patients (89.2%).
Eyes and/or head deviation (73.1%), generalized hypotonia (37.6%), and hemiconvulsions or generalized convulsions (23.6%) were the most common nonautonomic symptoms. Syncopal-like attacks with the child becoming flaccid and unresponsive were reported in 12 patients (12.9%). In others, generalized hypotonia was concomitant with malaise and pallor.
Four patients (4.3%) had visual ictal symptoms “I cannot see” or “I see cloudy,” which always appeared after nausea and vomiting. Two of these patients had a single seizure only; in the other two patients visual symptoms were not a consistent finding in recurrent seizures.
Most seizures (70%) occurred while the child was asleep, 17% during wakefulness, and 13% upon awakening. A statistically significant correlation was found between the circadian distribution of seizures and some clinical features. Vomiting occurred in 84.6% of patients who had seizures while asleep, in 75% of those with seizures after awakening, and in 50% of those with seizures while awake (p = 0.01). Hypersalivation occurred in 14 patients (15.1%) but was more often reported in seizures occurring during wakefulness (41.7%) than during sleep (12.3%) (p = 0.02).
The duration of seizures was usually lengthy, with a median of 30 min [interquartile range (IQR) 20–45 min, range 5–90 min]. More than half of seizures (55%) lasted longer than 30 min.
We divided patients into two groups according to the duration of the first seizure. Group A (with first seizure ≥30 min) consisted of 51 patients, and group B (with first seizure <30 min) consisted of 42 patients. Spontaneous remission was observed in 17 patients (33.3%) in group A versus 22 patients (52.3%) in group B (p = n.s). The number of patients with a second seizure was similar for the two groups: 37 patients (72.5%) in group A versus 26 patients (62%) in group B. The mean number of seizures was 3.7 ± 3.8 in group A versus 2.7 ± 2.7 in group B.
Of 14 patients (15%) with a first seizure longer than 60 min, this was terminated spontaneously in 3 patients, but the other 11 patients needed intrarectal benzodiazepines and five of them required admission to an intensive care unit.
Eighty-eight (94.6%) of the 93 patients had at least one abnormal interictal EEG (Table 3). Interictal spikes were found in 78.5% of patients, background abnormalities without spikes in 16.1%, and 5.4% of patients had consistently normal EEGs.
Table 3. Type, localization and evolution of EEG interictal abnormalities
aEach patient can show more than one localization.
Occipital to centrotemporal
Occipital to frontal
Photo paroxysmal response
Occipital spikes were the most common (76.1%), followed by temporal (23.9%), parietal (15.9%), central (13.6%), and frontal (10.2%) locations alone or in various combinations (Fig. 2 A–E).
Generalized discharges were recorded in one or more EEG studies of 22 patients (23.7%) and consisted of brief 1–3 s, diffuse slow-wave complexes with usually small spikes (Fig. 2F). They were not consistent in series of EEG studies of the same patient and they usually occurred together with focal interictal spikes.
Photo-paroxysmal response (PPR) were observed in EEG studies of eight (8.6%) patients. They consisted of generalized spikes or spike-and-wave discharges of brief duration, and with higher amplitude over posterior areas of both hemispheres, and occurred in one or a few of the series of EEGs of the same patient (Fig. 2G).
In serial follow-up EEGs, 16 patients (17. 2%) showed shift from the occipital to centrotemporal areas and 3 patients (3.2%) to frontal areas.
There were no statistically significant differences between patients with or without interictal abnormalities, or between patients with different types of interictal abnormalities (epileptiform, delta, or theta rhythm) with regard to age of onset, duration of first seizure, and recurrence of seizures.
Ictal EEG, recorded in five patients during sleep, showed left- (three patients) or right-sided (two patients) lateralization. Of four EEGs recorded from seizure onset, two started from the temporooccipital, one from the frontal, and one from the frontotemporal regions. Abnormalities consisted mainly of rhythmic theta activity at onset followed by spikes, polyspikes, and spikes-and-waves with diffusion to both hemispheres. Seizure semiology consisted of pallor, malaise, and sometimes retching, followed by progressive impairment of consciousness and vomiting in two cases and tachycardia in all cases (to as high as 240 bpm Fig. 3), and eye deviation contralateral to ictal EEG onset. Swallowing and chewing automatisms were observed in one case.
The duration of the recorded seizures ranged from 12–60 min (mean 27 min), and all but one were terminated with intrarectal diazepam. Fig. 3 illustrates the most representative sample of ictal EEG of ASE in PS.
Ninety patients had brain MRI, and this was normal in 85 of them. Abnormalities seen in five patients were (1) mild lateral ventricular asymmetry and a left temporal cyst, (2) Arnold-Chiari type 1 malformation, (3) slight periventricular leukomalacia, (4) right temporal lobe cyst, and (5) pre-rolandic gliosis.
Neuropsychological testing was performed after a mean period of 35.6 months (range 3–79 months) from the first seizure. The main clinical characteristics of the tested patients and controls are shown in supplementary Table S1. The controls were age- and sex-matched to the patients. There were no statistically significant differences between patients who underwent neuropsychological evaluation (n = 17) and all others (n = 76) regarding clinical characteristics.
IQ was normal in all patients (median full scale IQ 103, Verbal IQ 102, Performance IQ 104).
Results of subtests of cognitive profile of WISC-R are shown in Fig. 4. All patients had normal values in all subtests, although a statistically significant difference was found in arithmetic (9.8 vs. 11.8, p 0.04), comprehension (9.1 vs. 11.9, p 0.04), and picture arrangement (9.9 vs. 12.6, p 0.05) subtests when comparing patients with controls.
Treatment and prognosis
In total, during the observation period, the 93 patients had 301 seizures, with a median number of two seizures per patient. One-third of patients (31.1%) had a single seizure, and only 6.4% had >10 seizures. There were no patients with atypical evolutions.
Fifty-nine patients (63%) received antiepileptic drugs (AEDs) for various medical or parental reasons; 44 had recurrent seizures. Table S2 shows type and duration of treatment and adverse events reported. Thirty-four of the treated patients continued to have seizures despite treatment. Thirty-seven patients discontinued AEDs after a median period of 37 months, and only four had a seizure-relapse.
Sixty-four patients (68.8%) after 1 year had a complete remission of seizures (Figs. 1 and 5). The cumulative probability [with 95% confidence interval (CI)] of recurrence 6, 12, 24, and 36 months after the first seizure was 57.6 (44.7–68.4), 45.6 (31.1–58.9), 35.1 (19.1–51.7), and 11.7% (0.9–37.4%), respectively.
This is a clinical study over a 13-year observation period that supplements and adds new information on existing evidence in the characterization of PS.
Defining inclusion and exclusion criteria of PS
There is a significant variation on inclusion and exclusion criteria of PS among various published reports. In earlier studies, EEG interictal occipital spikes and ictal vomiting were often considered as mandatory inclusion criteria (see review in Panayiotopoulos, 2002). However, recent expert reports and consensus indicate that approximately one-fourth of patients may not have occipital spikes and one-fourth of seizures do not manifest with ictal vomiting (Ferrie et al., 2006; Michael et al., 2010; Panayiotopoulos, 2010).
Our decision to rely on clinical rather than EEG findings in the inclusion criteria of our study is based on well-documented evidence that although EEG is useful in the diagnosis of PS, there is not a definite EEG pattern that is specific for this syndrome. Autonomic seizure as defined by Ferrie et al. (2006, 2007) is the prerequisite inclusion criterion in this study; seizures without autonomic manifestations may occur rarely in PS (7%) but these affect children who also have additional autonomic seizures (Panayiotopoulos, 2002). The basis for our criterion that seizures should be 5 min or longer is that shorter seizures are common in other epileptic syndromes such as idiopathic childhood occipital epilepsy of Gastaut (ICOE-G) and rolandic epilepsy, whereas these are rare in PS.
That we have included five patients with MRI abnormalities as also others did (Yalcin et al., 2009) may appear contradictory for idiopathic epilepsy. However, these five patients did not differ in any other respect from those with normal MRI results; they had just a few autonomic seizures that were unrelated to the localization of the MRI abnormalities and remitted in long-term follow-up. Furthermore, similar cases have been included in rolandic epilepsy (Gelisse et al., 2003; Boxerman et al., 2007).
There has been a significant debate on the exact prevalence of PS, which may reflect awareness and diagnostic precision between authors (Panayiotopoulos et al., 2008; Taylor et al., 2008). Prevalence was 5.2% among children aged between 1 and 14 years with a first afebrile focal seizure referred to our children’s department.
In the original cohort of Panayiotopoulos, prevalence was ∼13% in children aged 3–6 years with one or more nonfebrile seizures, and 6% in the age group 1–15 years (Panayiotopoulos, 1988). PS is the most common specific cause of afebrile nonconvulsive status epilepticus in childhood (Okanishi et al., 2008).
Sex and age at onset
A uniform finding in all studies and our study is that age at onset is between 3 and 6 years for the majority of patients with PS. Both sexes are probably equally affected, although a female preponderance of around 56–60% was found in some and our cohort (Lada et al., 2003; Durá-Travé et al., 2008). This is in contrast to the male predominance evident in rolandic epilepsy.
Cumulative results indicate a high prevalence of febrile seizures (about 17%) (Panayiotopoulos, 2002) in comparison to only 5.4% in this cohort and other reports such as 4.4% by Lada et al. (2003).
However, the above and our study were not purposely designed for genetic purposes. According to Taylor et al. (2008) PS, like rolandic epilepsy, is probably genetically determined, but conventional genetic influences may be less important than other mechanisms. SCN1A mutations have been recently reported in a child (Grosso et al., 2007) and two siblings (Livingston et al., 2009) with a more severe type of PS and strong association with febrile precipitants, even after the age of 5 years. This may indicate that SCN1A mutations contribute to a more severe phenotype of PS; no SCN1A mutations were found in two sisters with typical PS, infrequent autonomic seizures, and no febrile precipitants (Panayiotopoulos, 2010).
We found no statistically significant differences regarding duration of first seizure, number of seizures, seizure semiology, and EEG findings between patients with or without a family history of epilepsy. Similarly, these features did not differ statistically between patients with and without antecedent medical events.
Autonomic seizures and ASE are the defining features of PS, as recently detailed in an expert consensus statement (Ferrie et al., 2007) and also confirmed in this report.
In PS, autonomic seizures follow a stereotypical pattern, in which emesis predominates (Vigevano & Ricci, 1993; Vigevano et al., 2000; Panayiotopoulos, 2002, 2010; Ferrie et al., 2006, 2007). Seizures start while the child fully conscious and able to speak complains of feeling unwell and nauseated prior to the development of more salient autonomic symptoms, followed by impairment of consciousness and other nonautonomic seizure manifestations, which may end with convulsions. When seizures occur in sleep, the child either wakes up with similar manifestations as described, or is found in the middle of a seizure.
Ictal syncopal-like attacks or ictal syncope is another intriguing clinical manifestation of PS (Oguni et al., 1999; Panayiotopoulos, 2002; Caraballo et al., 2007; Ferrie et al., 2007). The child becomes “completely unresponsive and flaccid like a rag doll,” a state that may precede, be concurrent with other seizure symptoms, or be the sole manifestation of a seizure. These attacks may occur while the patient is standing, sitting, lying down, or sleeping and may last from 1–2 min to half an hour.
In our study, all children by definition of inclusion criteria had at least one autonomic seizure typical of PS. Our findings corroborate previous results documenting the specific pattern of autonomic seizures in PS. Some insignificant discrepancies in the prevalence of reported autonomic manifestations are understandable considering that all these studies are based on clinical descriptions by patients and witnesses. Therefore, it is likely that only conspicuous symptoms are reported. The true magnitude and features of the autonomic seizures of PS can be revealed only by ictal video-EEG monitoring, where for example tachycardia has been documented as a significant sign (Fig. 3), whereas this has not been reported in clinical series.
Relevant to the debate on the differential diagnosis between PS and ICOE-G (Panayiotopoulos et al., 2008; Taylor et al., 2008; Caraballo et al., 2009) is our finding that visual symptoms rarely occurred in PS (4.3%), they always happened after the onset of autonomic symptoms, and they were not consistent in all seizures of the same patient; these are entirely different than in ICOE-G.
Our studies also confirm that seizures occur more often during sleep (70%) than wakefulness (17%) and awakening (13%). A new finding is that vomiting is more likely to occur in seizures during sleep than in wakefulness (p=0.01), whereas hypersalivation is more frequent in wakefulness than in sleep (p=0.02). The reason for this relationship is unclear. We have also found a marginally statistically significant difference between later age at onset and the presence of hypersalivation or oral twitching (p=0.07), which is understandable considering that these are primary features of rolandic epilepsy, which starts at a later age than PS (Panayiotopoulos et al., 2008).
Duration of autonomic seizures and ASE
Seizures in PS are unusual for other epileptic syndromes of long duration as documented in all studies and this study (Panayiotopoulos, 2002; Caraballo et al., 2007; Ferrie et al., 2007; Michael et al., 2010). In 51 patients (54.8%) the first seizure lasted more than 30 min, and in 14 of them it lasted more than 60 min. Adding to the literature is our finding that for patients with a first seizure longer than 60 min the possibility of experiencing a second episode is not higher than for all other patients [11 of 14 (78.6%) vs. 54 of 79 (68.4%), p = n.s., respectively]. Moreover, the mean number of seizures per patient during the illness course is similar for both groups (3.1 ± 2.2 vs. 3.3 ± 3.6, p = n.s.).
This study confirms that the interictal EEG in PS is usually multifocal with occipital spikes predominating (76.1%). Interictal spikes are typically of high voltage and morphologically identical to the centrotemporal spikes, but smaller spikes may occur. Furthermore, EEG foci frequently shift, multiply, and propagate diffusely with age-related changes (Oguni et al., 1999). Brief generalized discharges may occur and a few patients may have normal EEG or EEG without functional spikes.
Adding to the literature is the finding that there are no statistically significant differences between patients with or without interictal abnormalities, or between patients with different types of interictal abnormalities regarding age at onset, duration of first seizure, and recurrence of seizures.
The autonomic seizures and ASE of PS have been confirmed with eight recorded ictal EEGs as recently reviewed (Iannetti et al., 2009). Our five new cases of ictal EEG corroborate the findings that EEG onset may be posterior or anterior despite similar clinical manifestations.
Our findings from the neuropsychological assessment of children with PS are particularly important in view of the increasing number of reports emphasizing cognitive, linguistic, and behavioral abnormalities in rolandic epilepsy that have significant links with PS (see review in Panayiotopoulos et al., 2008).
Compared to matched control children, patients with PS had normal IQ. In addition, all patients had normal values in all subtests of WISC-R that explore attention, short-term memory, and visual perception functions. However, patients showed minor but statistically significant lower scores (always within normal limits) in arithmetic, comprehension, and picture arrangement subtests in comparison with controls. These minor deviations are less than anticipated considering that 76.5% of patients were on chronic AED treatment and 88.2% had abundant interictal epileptiform abnormalities with marked exaggeration during sleep. Furthermore, these patients may represent the worst spectrum of the disorder (i.e., hospital-based populations) or other factors associated to psychosocial impact of “epilepsy” in general (Valeta, 2005). Therefore, our results indicate that patients with PS do not show any significant neuropsychological abnormalities that can be attributed directly to their epileptic seizures and/or EEG abnormalities.
Table S3 summarizes the results of seven studies published between 2000 and 2009 on neuropsychological assessments in patients with idiopathic occipital lobe epilepsy including PS. The overall conclusions of all studies are that IQ is normal in all patients; the most frequent deficit has been found in visual and visuoperceptual functions. Less frequently some minor attention and memory disturbances have been reported.
PS is one of the most benign epileptic syndromes in terms of seizure frequency and remission. Cumulative results indicate that 40% (range 18–83%) have a single seizure, 48% (41–58%) have 2–5 seizures, and only 5% (3–6%) have >10 seizures (Panayiotopoulos, 2002, 2010; Michael et al., 2010). Ninety percent of patients go into complete remission within 1–2 years of onset; the others may have frequent seizures, protracted active seizure period, and may develop rolandic and less often occipital or other seizures, but these are also age-related and remit.
In a recent report of 79 patients with PS who all achieved remission by age 12 years of age, 57% had 1–5, 20% had 6–9, and 23% >9 seizures (Hirano et al., 2009). Of prognostic factors among the three groups only preexisting neurobehavioral disorders were significantly associated with those who had more than nine seizures (p > 0.05).
Our results confirm the excellent prognosis of PS. The overall cumulative probability for seizure recurrence, expressed using survival curve, demonstrates the short duration of the disease. Patients who do not have remission after 1 year have a low probability to remit during the second year, and seizures tend to persist until the third year when all cases remit. The overall disease duration is approximately 3 years. However, we have found that early age at onset is a negative prognostic factor significantly associated with more than one seizure.
Most children with PS might not need medication, as the natural course of PS may include only one seizure. Treatment is suggested for those with seizures that are frequent and/or affect their quality of life. In addition, most parents usually request treatment because of the dramatic psychosocial impact of seizures on themselves (Valeta, 2005; Panayiotopoulos et al., 2008).
Currently, there is no evidence of the optional AED in PS, although carbamazepine appears to be preferred (Ferrie et al., 2006, 2007), whereas valproate was most commonly used by us. In our series, the length of treatment was variable, from 3 months to more than 10 years. However, duration of treatment did not correlate with recurrence of seizures: in fact 34 (77.3%) of 44 patients who received AEDs continued to have seizures. Among patients who discontinued AEDs, only 10.8% relapsed, which is low in comparison with that expected from published data on discontinuation of AEDs in epilepsy (Specchio & Beghi, 2004).
We thank Dr. C P Panayiotopoulos for critically reviewing the manuscript and editorial suggestions. We thank also Lucilla Ravà, for statistical advice.The authors have nothing 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.