JLE

Epileptic Disorders

MENU

Sleep-related, low voltage Rolandic and vertex spikes: an EEG marker of benignity in infancy-onset focal epilepsies Volume 4, numéro 1, March 2002

The existence of benign forms of early-onset epilepsies has been debated. In 1989, it was stated that there were no cases with good outcome in epilepsies with onset between the ages of 2 months (the upper limit of benign neonatal convulsions) and the age of 15 months (the lower limit of benign ''Rolandic'' epilepsies) [1]. However, benign seizures in infancy have been recognized in sporadic cases [2-5], and a syndrome of familial benign infantile convulsions (BIFC) has been described [6-8]. In clinical practice, the occurrence of seizures in an infant or very young child without a clear familial history of BIFC is an ominous event, and the prognosis concerning both epilepsy and neuropsychological development is guarded. No reliable clinical or electroencephalographic (EEG) markers can predict the outcome.

Following the discovery of a very striking EEG trait that we retrospectively found in infants and young children with early onset epilepsies and favourable outcome [9, 10], we saw several, newly referred patients with these EEG changes and could reliably predict such a favourable course. We thus wish to make epileptologists and electroencephalographers aware of this strikingly reliable marker of good prognosis, that might help define a very early onset variant of benign focal epilepsy.

Materials and methods

We looked at 10 patients with a benign, self-limited form of early-onset partial epilepsy and a striking EEG pattern. Cases 1-6 were found during a systematic, retrospective survey of focal epilepsies with onset before age ten, who did not have the typical features of benign Rolandic epilepsy, and who had been evaluated by means of extensive EEGs, including sleep EEG, and had been followed for at least five years. The details of this study, which was performed on newly referred, consecutive patients between 1975 and 1994 and which also included cases with a bad prognosis, have been published elsewhere [9, 10]. These six patients belonged to a group of 18 cases with a full remission of seizures: among these, 8, including the 6 cases with the EEG marker, had experienced their first seizure before the age of 2 years. Cases 7-10 were studied in our EEG laboratory between 1995 and 1998, without a reference population and were selected on the basis of the presence of the typical EEG changes.

All patients had at least one waking EEG and at least 2 daytime sleep EEGs. All were seen within 2 years of the first seizure, and followed for at least 1 year after the last seizure, with the exception of one patient who has not been seen since his second and presumably last seizure at age 2y 2m.

The EEG data were compared with a control group of healthy infants that included four patients with benign infantile spasms [11] and 10 patients with breathholding spells, who all had sleep EEG recorded between the ages of 7 months and 2 years.

Results

The main clinical and EEG data are reported in tables I and II. There were 7 girls and 3 boys. The family history was positive in three, with febrile seizures in two, first-degree parents (father and mother in one, and father in the second), and a case of benign, infancy-onset seizures in a paternal great-aunt in one. There was no significant personal history except coexisting breath-holding spells, a febrile seizure and a low birth weight in patients 4, 7 and 9, respectively. The mean age at onset of seizures was 9.9 months (range 1-20 months). Neurological examination and developmental milestones were normal, and remained normal during follow-up, in all patients. Neuroimaging was performed in six patients: four had a CT scan, which was normal, and only the two most recent patients had MRI, which showed in both cases, non specific posterior white matter hypersignals. The mean age at the last follow-up was 7 years 6 months (range 2 y 2 m-14 y). The mean duration of follow-up after onset of epilepsy was 6 years 6 months (range 6 m-12 y 5 m).

Semeiology of seizures

Seven patients had suffered only one seizure type. Five children had presented with complex partial seizures (CPS), characterized by an arrest of activity, staring or ocular revulsion, unresponsiveness, pallor, cyanosis, chewing, swallowing, hypotonia or stiffening. These signs were not present simultaneously during each seizure in the same infant. One child had simple partial seizures (SPS) with autonomic symptoms and arrest of activity, without apparent loss of consciousness. In one patient, loss of consciousness was impossible to assess.

Three children presented with two types of seizure: 1 with CPS and SPS, 2 with seizures described as tonic and CP. There was no secondary generalization except once in one child. The duration of seizures ranged from a few seconds to 2 minutes. There were no post-ictal signs, a quick recovery being the rule. However, in 1 child, one of the attacks was accompanied by a longlasting (1 hour 30 min) state of obtusion.

The frequency of seizures was variable from several per day to one in 6 months, and did not change during follow-up. Seizures never occurred in clusters or in status. Nearly all seizures occurred during wakefulness.

Electroencephalographic findings

During the waking state, background activity was normal in all children with no paroxysmal abnormalities. EEG during sleep showed physiological patterns in all patients. REM stage could be recorded in 8 children.

The sleep EEG recording was performed at the mean age of 13 months and the mean delay since the first seizures was 5 months (3 weeks-26 months). During drowsiness and all non-REM and REM sleep stages, we observed low-voltage spikes followed or not by a slow wave which were localized over the frontocentral region of one or both hemispheres and the vertex, simultaneously or independently (figure 1). Sometimes it was very difficult to see these spikes because they were intermixed with spindles or K complexes (figure 2). In order to enhance these abnormalities, we found it is often necessary to lower the voltage and record at 30 millimeters per second rather than 15 mm/s (figure 3). In two patients, we also found bursts of diffuse spikes and waves during drowsiness, which then disappeared during slow sleep (figure 4). In one child, a seizure was recorded during a 24-hr ambulatory EEG. However, movement obscured the initial ictal activity and only the end of the episode, represented by diffuse slow waves on the left hemisphere, could be seen.

In a previous work [9], we found this typical interictal EEG changes in six out of eight patients with a typically good outcome: we can thus assume that the sensitivity of this EEG trait is less than 100%, or 75% in this small series (the other 4 patients were not drawn from a systematic study of consecutive cases). However, none of the patients with this EEG pattern had an unfavourable long-term epilepsy or mental outcome, and none of the patients with a more severe form of focal epilepsy had this trait. Moreover, none of the non-epileptic infants that we recorded in the control group had such spikes, and, to our knowledge, such findings have not been reported as incidental EEG traits in normal infants. Thus, in our experience, the specificity of this marker is near 100%.

Outcome

Antiepileptic drugs were given to all children except one: phenobarbital in two, carbamazepine in five, sodium valproate in two. Seizures disappeared in all children. In eight, treatment was withdrawn between the age of 3 and 11 years. One child is still on medication after a 4-year follow-up. No seizures reappeared during a follow-up of 14 months to 8 years 6 months (median: 4 years). The duration of active epilepsy ranged from 2 months to 2 years 9 months (median: 9 months). In one patient there was only one cluster of seizures over 2 days. The EEG abnormalities disappeared a few months after the last seizure, with no change in waveform morphology.

Comments and discussion

The main characteristic of our cases was that they fulfilled only some of the criteria for good outcome in benign focal epilepsies of childhood, as defined by Dalla Bernardina et al., [12] i.e.: normal psychomotor development, no postictal deficit, and good response to medication. But several features are the same as those observed in the classical criteria for bad outcome: onset before 3 years of age, association of several types of seizures (in three cases), high seizure frequency, existence of CPS. However, the favourable outcome allows us to hypothesise that these infants have a benign focal epilepsy as reported in the two series describing benign partial infant epilepsies with complex partial seizures [3, 4] and benign partial epilepsy with secondarily generalized seizures (SGS) [5], and in the series of benign infantile familial convulsions [6-8]. In 2000, Capovilla and Beccaria [13] reported on 12 patients who had both an infancy-onset, self-limited and benign focal epilepsy and the same, striking EEG pattern we had reported in a previous work [9].

The clinical features of patients with benign, sporadic focal epilepsies with onset in infancy are characterized by partial seizures that occur in clusters, with subacute onset in the first or second year of life, without clear etiology, and with normal psychomotor development [3-5]. Age at onset has been reported to range from 3 to 20 months, but in the familial cases, seizures occurred mostly between the 4th and 7th month. History of pregnancy and delivery is unremarkable. All the etiological investigations, especially metabolic screening and neuroimaging, yielded normal results. Familial cases, which were studied in detail by Vigevano et al. [7], have autosomal dominant inheritance. An almost constant characteristic was the occurrence of seizures in clusters: mostly brief, successive seizures, up to a maximum of 8-10 daily, which did not amount to true status epilepticus. Isolated seizures in the 10-15 day period before a cluster were reported in about one third of the children. Seizures were usually longer at the beginning of the cluster, lasting 2-5 minutes, and became shorter after treatment was initiated. The cluster could last 1-3 days.

The differences in seizure symptomatology between the sporadic and familial cases are difficult to assess. Elements common to both include motor arrest, impairment of consciousness, staring and convulsive movements. Limb or oral and facial automatisms were described in patients with CPS [4], and prompt generalization with tonic-clonic manifestations was seen in those with SGS [5]. Ictal slow deviation of head and eyes to one side, diffuse hypertonia, cyanosis and unilateral progressing to bilateral synchronous or asynchronous limb jerks were described in the familial forms [7]. Waking and sleep interictal EEG before and after the cluster is normal. Lateralized interictal occipito-parietal slow waves and spikes in the interictal EEG performed during a cluster of seizures were described in familial cases [8]. The site of seizure origin in the sporadic cases with CPS [4] is the temporal area, whereas the site varies in children with SGS [5]. In familial cases, seizures originated in the parieto-occipital area, with the side varying from one seizure to another [7].

The EEG findings we reported in infants resemble, to some extent, the changes found in older children with the well-known benign partial epilepsy of childhood with centro-temporal spikes (BECTS) [14, 15], including diffuse SW discharges around sleep onset in some patients [16, 17], and the age-related disappearance of spikes. However, BECTS and benign focal epilepsy in infancy show major differences, beyond the age-dependency: the infants never expressed these spikes while awake, the amplitude was always low, and there was no difference in morphology or in activation between sleep stages. The main difference may be that the spikes found in BECTS are very prominent, whereas the spikes found in our patients were easy to overlook.

It is difficult to form a well-founded opinion on the pharmacological sensitivity of these infants. In all cases, seizures remitted quickly on any medication, and our present attitude is to avoid chronic treatment with AEDs, beyond the acute period of seizures, when AEDs are often given in a context of emergency. In their series of 12 patients, Capovilla and Beccaria [13] administered anticonvulsants to only 3 patients.

Received October 12, 2001 ; Accepted February 4, 2002

CONCLUSION

In a small, but increasing number of infants with focal epilepsies that appear as idiopathic, we found a previously unreported EEG marker that appears to be very specific and quite sensitive: low-voltage spikes over the central regions and/or vertex, which can only be seen during sleep. These changes have probably been overlooked in previous reports, that stressed the absence of EEG abnormalities in infants with benign focal epilepsies. We recommend that sleep EEG recordings in such patients be carefully evaluated in the search for this specific marker. Our findings have practical consequences, in that they may help to establish a prognosis of good outcome, early in the course of such benign, self-limited early onset focal epilepsies in which no AED treatment is necessary.