INTRODUCTION: SELF-LIMITED FOCAL EPILEPSIES OF CHILDHOOD

The International League Against Epilepsy’s (ILAE’s) revised classification of seizures and epilepsies (1) argues against the older term benign because it obscures consideration of cognitive, behavioral, and psychiatric comorbidities, as well as mortality (2), aspects that we emphasize in this chapter. Instead, they recommend parsing the “benign” concept into the terms self-limited, to mean having a high likelihood of spontaneous remission at a predictable age, and pharmacoresponsive, to designate the reasonable certainty that seizures will come under rapid control with appropriate medication. The Commission did not go as far as endorsing a definite genetic classification for this group of epilepsies, although accumulated evidence since 2010, discussed in the text later, may lead to a reconsideration in the near future.

In the 2010 revised scheme, the ILAE listed 33 electroclinical syndromes and other epilepsies (1). Only a small number of them satisfy the ILAE criteria for the diagnosis of self-limited focal epilepsy. These are included in the following list in order of age of onset, and ironically many still carry the “benign” epithet:

Together these syndromes account for approximately 30% of all children with epilepsy; along with the genetic generalized epilepsies (GGEs) (see Chapters 17 to 21), the majority of these are managed by nonspecialists. The “core group” of benign focal epilepsies left to consider in this chapter (with their official ILAE names in parentheses) are

The main features of these disorders are compared in Table 23.1. These syndromes are considered to lie at the mild end of a spectrum of rarer epilepsy syndromes and epileptic encephalopathies including atypical benign partial epilepsy (ABPE or pseudo-Lennox syndrome) (3); Landau–Kleffner syndrome (LKS) with acquired aphasia and verbal auditory agnosia (4); continuous spikes in slow-wave sleep (CSWSS) (5); and “epilepsy–aphasia spectrum” with language delay and autistic features (6). Among these, only LKS and CSWSS are currently recognized by the international classification (1).

A NOTE ON TERMINOLOGY

We prefer rolandic epilepsy (RE) to benign childhood epilepsy with centrotemporal spikes because the term is more widely used by pediatricians and in PubMed; RE may occur without centrotemporal spikes; conversely, centrotemporal spikes occur in children without seizures, as well as in children with other self-limited focal epilepsies; centrotemporal spikes are located mainly in the central (rolandic) fissure; they are rarely located in the temporal electrodes. The term temporal may misleadingly suggest the occurrence of temporal lobe symptoms during seizures. We use the term idiopathic childhood occipital epilepsy to include both late-onset childhood occipital epilepsy—Gastaut type—and idiopathic photosensitive occipital lobe epilepsy. This is because both conditions share many common features and it is not clear that they merit separation into two distinct syndromes.

TABLE 23.1

GENETICS OF THE SELF-LIMITED FOCAL EPILEPSIES OF CHILDHOOD

Some of the rarer benign focal epilepsies have an established monogenic basis but with clear evidence of genetic heterogeneity, that is, mutations in different genes resulting in indistinguishable electroclinical syndromes. For example, KCNQ2 and KCNQ3 mutations in BFNE (8,9); and mutations in PRRT2, SCN2A, or less commonly in KCNQ2 and KCNQ3, in benign familial infantile epilepsy (10–13). These are usually autosomal-dominant mutations with incomplete penetrance, meaning that not every person with a mutation will manifest the disease. Additionally, the clinical manifestations are not limited to epilepsy in people with these mutations (“pleiotropy”), for example, PRRT2 mutations are associated with movement disorders (13), while KCNQ2 mutations are also associated with neonatal onset epileptic encephalopathy (14).

Among the more common forms of benign focal epilepsy, RE is the most studied; there are no genetic studies in PS or Gastaut-type focal epilepsy. Unlike the rarer epilepsy syndromes, RE does not have a Mendelian inheritance, except in rare and mostly very severe forms such as autosomal-dominant RE with speech dyspraxia, autosomal-recessive RE with dystonia, and X-linked rolandic seizures with oral and speech dyspraxia and mental retardation (15–18). Screening of genes discovered in other epilepsies has identified rare mutations in KCNQ2 and KCNQ3 in RE cases with and without neonatal seizures (19); mutations in six other genes are now shown to be risk factors for RE including: the NMDA receptor subunit GRIN2A, also seen in ABPE, CSWS and LKS (20,21); neuronal splicing regulators RBFOX1, RBFOX3 (22); GABA-A receptor GABRG2 (23); DEPDC5, a repressor of the mammalian target of rapamycin (mTOR) complex (24); and ZDHHC9 (25).

By contrast, the common form of RE has a complex inheritance, implying the involvement of several gene variants acting in combination, with there being no evidence for environmental factors. Clues to these variants have emerged from studying the components of RE, starting with the characteristic centrotemporal spikes (CTS). Segregation analysis in families shows that CTS is inherited as an autosomal-dominant trait (26). CTS was first localized using linkage analysis to 15q13.3 (originally reported as 15q14) (27), and more recently using linkage and association analysis to ELP4-PAX6 locus at chromosome 11p13 (28). The same 11p13 locus predisposes to the speech sound disorder frequently observed in RE (29). A microRNA-328 binding site in PAX6 is disrupted in 7% of RE patients and strongly associated with CTS (30). Reading disability (RD) is the most prominent comorbidity in RE (see the following text), and genetic analysis suggests strong evidence of linkage at chromosome 1q42, and also at 7q21 in some populations, but no involvement of known dyslexia loci (31).

Structural variation in the genome resulting in copy number variation (CNV) at dosage-sensitive regions of the genome is a strong but rare risk factor in epilepsies (32): for example, deletions at 15q13.3, 15q11.2, and 16p13.11 predispose individuals to GGE (33,34). CNV also plays a role in RE, with 7.5% carrying known pathogenic CNVs, including at recurrent “hotspots” 1p36, 16p12, 16p13.11, and Xp22.31; an additional 10% carry likely pathogenic CNVs (involved in other neurologic disorders or involved in neuronal signaling or development) (Addis et al, unpublished). Interestingly, there is hardly any overlap between CNVs involved in GGEs and RE, and the distribution of CNVs differs according to continental ancestry. CNVs do not correlate with seizure, treatment, or comorbidity variables in RE, and while they increase disease risk, their exact role in etiology remains to be fully elucidated (35).

ROLANDIC EPILEPSY

Introduction and Definition

RE was, in 1958, the first of the benign focal epilepsies to be recognized (36) a few years after the description of CTS (37); later, Lombroso confirmed it as an entity (“sylvian epilepsy”) in the United States (38), and now it is recognized as the single most common epilepsy syndrome. Rolandic epilepsy can best be defined as a developmental focal epilepsy with frequent neurodevelopmental comorbidities. The seizures, which are usually few in number, appear and remit spontaneously within a predictable age range, and for the most part respond well to conventional medication. Although the terms RE and BECTS are regarded as interchangeable, some authors cautiously reserve the term RE for children with typical rolandic seizures (see the following text) (39), and BECTS for children who have the typical EEG feature but less characteristic seizure manifestations, for example, only (secondarily) generalized nocturnal seizures. Approximately 30% of children with RE have a history of speech sound disorder; 30% have RD; 10% to 20% attentional impairments, dyspraxia, or migraine (40); many children have multiple comorbidities, and the cognitive impairments often have a more pervasive impact on quality of life than seizures, especially because they often go undiagnosed (41).

Epidemiology

The prevalence of RE is estimated at approximately 16% of pre-adolescent epilepsies or 1 in 2,500, with an incidence of 10–20/100,000 children per year (42,43). It is diagnosed on every continent and there is no evidence of geographical variation. Boys are affected 50% more often than girls. The peak age range for onset of afebrile seizures is 5 to 8 years, with most commencing between 3 and 10 years (44); onset before 2 or after 12 years of age is unusual (45). Risk factors for RE include Panayiotopoulos syndrome, autism spectrum disorders, and fragile X syndrome (see the following text). Febrile seizures precede RE in approximately 10% of cases, but are not considered a specific risk factor. RE is associated with several specific neurodevelopmental comorbidities including speech sound disorder (odds ratio 3x) and RD (odds ratio 5x) (46). A family history of seizures is observed in 7% to 15% of cases, although RE is only seen in 4% of siblings (26); however, SSD, RD, and migraine strongly cluster in RE families (46,47).

Seizure Manifestations

Between 70% and 80% of seizures are focal, and mostly motor, although sensory phenomena can occur (48). The sequence of symptoms recapitulates the anatomy of the vocal tract, often starting with a gurgling, guttural, or rattling sound from the throat, followed by tonic contraction, clonic jerks, and subsequent paresis of one side of the face, as well as mouth movements and profuse salivation. Seizures may spread to the upper limb with clonic jerks (20%) and less commonly the lower limb. Parents often report that they believe their child is having a stroke. Sensory symptoms most often involve tingling or numbness of the corner of the mouth, the cheek, tongue, or gums. Sometimes, even though parents may note the absence of motor symptoms in follow-up, the child reports periodic sensory symptoms. Speech arrest often follows, although children know what they want to say and may utter inarticulate sounds or gesture with their hands; in over half of cases, children retain awareness and are able to give a full account. The side affected by seizures may remain constant or switch in subsequent seizure episodes; and the seizure manifestations may also vary in features, spread, and duration (usually attenuating) as the child gets older.

Seizures tend to last no longer than 5 to 10 minutes; longer attacks, which may become secondarily generalized, are infrequently accompanied by postictal hemiplegia (49). Sometimes nocturnal generalized seizures may be the only reported seizure type and their focal origin is not appreciated until the EEG reveals CTS—this is a common reason for misdiagnosis. Generalized seizures are more commonly seen in nocturnal seizures and among children with a younger age of onset. Continuous focal motor facial jerks, opercular status, and hemiconvulsive status are very rare complications (50–53). A small proportion of children will either present with mixed rolandic and Panayiotopoulos type autonomic seizures (vomiting, pallor, headaches) or progress to RE from an earlier diagnosis of PS (50) (see Chapter 22).

Electrical status in slow-wave sleep (ESES) or a near-ESES picture is an important complication of RE, and is seen in perhaps 10% of sleep EEGs. The functional impact of ESES can be hard to evaluate in RE, and parental report of academic or behavioral deterioration should be objectively verified. Ideally, the presence of ESES should be documented on two EEGs 2 weeks apart before intervention is contemplated (see Treatment in the following text).

RE, in common with PS and GGEs, shows a definite relation to the sleep cycle (54). Seizures occur predominantly in sleep, and CTS are potentiated by sleep (55). Sleep is organized into cycles of non-rapid eye movement (NREM), REM, and periods of arousal; younger children have shorter and more frequent cycles than older children and adults. Thus, most seizures are reported either around 1 to 2 hours after going to bed, or 1 to 2 hours before arising in the morning, usually during NREM cycles; occasionally, seizures can occur during an unaccustomed daytime nap in the back of a car. Because of their predominantly nocturnal occurrence, seizure observation may be missed, but parents can be asked whether they have noticed a wet pillow in the morning during follow-up.

Comorbidities

Cognitive, behavioral, psychiatric, and somatic comorbidities are well documented in childhood epilepsies. Some of these comorbidities are shared among children with epilepsies traditionally considered to have a favorable prognosis. For example, attention deficit hyperactivity disorder (ADHD), as well as problems in language and executive function, are all common to the syndromes of childhood absence epilepsy, RE, and juvenile myoclonic epilepsy (56–58). However, RE uniquely has a strong and specific association with both reading disability (RD): odds ratio 5.8 (95%CI: 2.8 to 11.7) and speech sound disorder (SSD): OR 2.5 (1.2–5.0) (46). Notably, RD, SSD, and attentional and language impairments also co-aggregate in RE family members without epilepsy (46,59).

Approximately 30% of children with RE will experience RD, while 50% will have language impairment (59). Some affected children will be detected at elementary (primary) school, usually before seizure onset, when they struggle to master basic reading skills, but unfortunately many do not come to notice until high school. Meta-analysis of 23 studies shows the most prominent impairments (with moderate to large effect sizes) are in single word reading, receptive and expressive language, and phonologic processing (58). These literacy and language problems are most likely underpinned by phonologic processing difficulties, which is accepted as a common mechanism in dyslexia (60). However, it may be the case that the route to RD differs in RE and that deficits in attention (61) and auditory processing (59)also play a role in poor literacy skills; this area requires further investigation. RD in RE (and in the general population) is associated with SSD, ADHD, and male sex. There are three times higher odds for siblings to have RD if the RE proband has RD, but no higher odds if the proband does not have RD, highlighting the need for family surveillance in pediatric practice (46). Expressive language impairments are more pronounced in older children, and are even seen in children with full scale IQs greater than 100 (61).

Speech sound disorder refers to the late production of intelligible speech, characterized by substitution, omission, distortion, or addition of phonemes (speech sounds), in the absence of neurologic, mechanical, or sensory impairments (ICD-10 F80.0), and is found in 30% of RE (46). Acoustic analysis of speech in RE shows that a developmental verbal dyspraxic mechanism underlies SSD; in other words, a breakdown in the ability to control the spatial/temporal properties of speech articulation (29). Notably, SSD is diagnosed in the second or third year of life, long preceding the typical onset of seizures in RE, and the presence of SSD increases later risk for RD.

Parents rate the presence of ADHD in up to 20% of RE, mainly of inattentive type, a similar frequency as seen in other childhood epilepsies; and they report certain psychologic symptoms on the child behavior checklist (CBCL): mainly internalizing behaviors and somatic symptoms (30%), as well as low social competency (20%). ADHD symptoms increase the risk for RD. Changes in attentional impairments over time correlate with changes in EEG abnormalities in RE (61). Motor dyspraxia is also occasionally observed, though it is usually mild.

Last, migraine is a well-known comorbidity that is seen in 15% of RE and 15% of siblings, almost three times the risk in healthy controls (47). These are usually migraine attacks without aura, which do not usually occur at the same time as seizures and can be managed in the usual way.

The etiology and mechanism for these comorbidities has been debated over several decades, but there is now clear evidence for a genetic basis for RD, SSD, and migraine in RE. RD is linked to novel genetic loci at 1q42 and, in some populations, at 7q21, which have not been reported in “pure” dyslexia (31). Despite observations of transient and dynamic neurocognitive disturbances in the presence of interictal epileptiform discharges (IEDs) (62), there is no evidence that either IEDs, seizure, or treatment variables are associated with RD when the roles of SSD and ADHD are taken into account (40). SSD appears to share the same inheritance as CTS (29); and migraine in RE is linked to a novel locus at 17q22 as well as the FHM2 locus at 1q23 (63). Only attentional impairments seem to correlate consistently with EEG abnormalities in RE (61).

Triggers and Prevention

Seizure triggers in RE are related to sleep fragmentation or deprivation. This usually occurs during long distance travel (eg, vacations), or following a late night after a party or sleepover. Simple sleep hygiene measures should routinely be recommended at diagnosis.

Rarer Manifestations/Variants

Between 10% and 20% of children with fragile X syndrome (FXS) have epilepsy, and RE is the most common associated epilepsy syndrome, while CTS is a common asymptomatic finding (64). The presence of RE or CTS in the presence of global learning disability, an unexpected finding in RE, should prompt consideration of FXS. CTS is also found in autism spectrum disorders, where seizures are common (65). Rare variants in which speech dyspraxia and language impairment are especially prominent have also been reported (15,16); care should be taken to differentiate perisylvian disorders by 3T MRI, and the Worster–Drought form of cerebral palsy where oromotor and/or upper limb fine motor impairment are significant (66).

EEG Features

The interictal EEG in children with rolandic epilepsy usually shows a normal background but with the hallmark centrotemporal (or rolandic) spikes (Figure 23.1). Despite their name, these are usually high amplitude sharp- and slow-wave complexes localized to the central (C3/C4) electrodes or midway between the central and temporal electrodes (C5/C6). They may be unilateral or bilateral, synchronous or asynchronous. They tend to be highly stereotyped, meaning each discharge bears striking resemblance to the others.

Centrotemporal spikes have been studied with EEG single- or multiple-dipole modeling computerized techniques. The consensus is that their main negative spike component can usually be modeled by a single and stable tangential dipole source along the central (rolandic) region, with the negative pole maximal in the centrotemporal and the positive pole maximal in the frontal regions (25,26). Their source has been localized to the pre- and postcentral gyri in both RE (67) and PS (68).

Centrotemporal spikes are highly activated by sleep; normal sleep architecture is preserved. The frequency, location, and persistence of centrotemporal spikes do not determine the clinical manifestations, severity, and frequency of seizures or the prognosis. The interictal EEG (especially if there is no sleep recording) may occasionally be normal. Mild background slowing may be a postictal effect, reflect antiepileptic drug medication, or may be seen if centrotemporal spikes are particularly abundant. Rarely, small inconspicuous spikes may be seen, rather than those with the more characteristic morphology. Sharp- and slow-wave complexes in areas outside the centrotemporal regions, such as occipital, parietal, frontal, and midline regions, may occur concurrently with centrotemporal spikes. They are of similar morphology to centrotemporal spikes. Rarely, generalized discharges occur.

FIGURE 23.1 Centrotemporal spikes recorded from an 11-year-old girl with rolandic seizures that had been in remission for the last 3 years. Note that the spikes have their maximal amplitude in the central rather than temporal electrodes. (Top) EEG in wakefulness. (Top, left) Infrequent spikes occur independently in the right or left central electrodes (C4 or C3). (Top, right) Stimulation of the tips of the fingers elicited high amplitude central spikes, which were contralateral to the side of stimulation. Self-stimulation of the fingers elicited simultaneous bilateral central spikes. In clinical EEG practice, asking the child to tap together the palmar surface of the tips of his or her fingers of both hands is an easy method of testing for evoked spikes. The child should be instructed to strike them with sufficient strength and at random intervals. This may elicit either bilateral or unilateral central spikes. (Bottom) Frequency and amplitude of central spikes is markedly increased in sleep. In previous EEGs their abundance approached that of electrical status epilepticus in sleep. Despite marked EEG abnormalities over many years, the child was otherwise normal, did well in school, and had no linguistic or neuropsychologic deficits.

There are very few ictal EEG recordings of rolandic seizures. Before the onset of the ictal discharge, centrotemporal spikes become sparse. The ictal discharge then appears and consists of unilateral slow waves intermixed with fast rhythms and spikes located in the central regions. The ictal EEG features of rolandic epilepsy are shown in Figure 23.2 and compared to those of Panayiotopoulos syndrome.

Magnetoencephalography (MEG) studies have shown that the dipoles of the prominent negative sharp waves of rolandic discharges appear as tangential dipoles in the central (rolandic) region, with positive poles being situated anteriorly (69,70). MEG findings in children with RE are shown in Figure 23.3 and compared with those in Panayiotopoulos syndrome.

Diagnostic Evaluation and Differential Diagnosis

A full medical history including sleep pattern (for remediable triggers), developmental, academic, and family history (for treatable neurodevelopmental disorders in patients and siblings) should be taken. There is controversy as to the extent of investigation required in children with suspected RE, and great variation in practice. Investigations, other than EEG, are expected to be normal, but 30% of UK pediatricians admit to not performing it, presumably relying on the clinical history alone, or lacking ready access to EEG (72). European practice is that neuroimaging is not performed if the clinical picture, seizure semiology, and EEG features are typical, and comorbidities are not unusual in nature or severity; in the UK, MRI is nevertheless routinely performed in 25% of cases (72). It is worth noting that routine MRI may detect coincidental minor abnormalities that are seen at the same frequency as in headache controls (73). MRI is certainly indicated if atypical features are present (see Rarer Manifestations/Variants earlier in this chapter) or if the evolution is not as expected (see Course and Prognosis, later in this chapter). Neuropsychologic or educational psychologic evaluation is underutilized (only 5% of UK pediatricians) (72), despite the very high prevalence of literacy, language, and attentional impairments (40), as well as their impact on quality of life (41).

FIGURE 23.2 Samples from video-EEG recorded seizures of rolandic epilepsy and Panayiotopoulos syndrome. (Top) Onset of a rolandic seizure captured during routine sleep video-EEG of a 9-year-old girl. Right-sided centrotemporal spikes (oblique arrows) and brief (2–4 sec) generalized discharges of sharp slow waves intermixed with small spikes occurred in the interictal EEG only during sleep. The exact onset of the electrical ictal event is not clear but started in the right centrotemporal regions (open horizontal arrows) with 2- to 3-Hz slow waves and irregular, random, and monophasic medium-voltage spikes intermixed and superimposed on the slow waves. This activity tended to spread, and the amplitude of the spikes rapidly increased before the first clinical symptoms of left hemifacial convulsions, which occurred approximately 30 seconds from the onset of the EEG ictal changes (black vertical arrow). At age 18 years, she is entirely normal and is not taking medication.

Source: From Ref. (5). With the permission of the publisher, and Ref. (71). With the permission of the authors and the editor of Epilepsy & Behavior.

FIGURE 23.3 Magnetoencephalography (MEG) studies from two patients with rolandic epilepsy (left two columns) and a patient with Panayiotopoulos syndrome (right two columns). In rolandic epilepsy, equivalent current dipoles of spikes are located and concentrated in the rolandic regions and have regular directions. In Panayiotopoulos syndrome, equivalent current dipoles of spikes are located and concentrated bilaterally in the rolandic regions and right occipital area. The directions of each equivalent current dipole in each area are quite regular as if three small round toothbrushes are placed in each of the three areas. Small yellow circles represent locations and yellow arrows represent directions of equivalent current dipoles. Blue circles and arrows represent a bilateral somatosensory evoked magnetic field. Figure courtesy of Dr. Osamu Kanazawa.

Genetic testing is widely available for epilepsy, but currently its clinical utility has not been evaluated for RE. Cases with a family history, ESES, or atypical features, might justify a search for copy number variation or GRIN2A mutation. However, in the future, genetic testing will undoubtedly have increased value for personalized medicine when treatments can be matched to an individual’s genetic makeup.

The diagnosis may not be clear if the seizure has not been witnessed, or only a secondarily generalized seizure is reported, as may occur at night; this is the most common reason for misdiagnosis of RE as GTCS. Other nocturnal phenomena including night terrors or sleep apnea should be considered in the differential. Cases with atypical features (see Rare Manifestations/Variants earlier in this chapter) should be investigated in further detail.

Management

Seizures in RE are self-limited and AEDs do not affect the eventual prognosis. Textbooks and “expert opinion” often advised conservative management (ie, no AEDs) (74); in the UK, about 40% of clinicians follow a no-AED plan (72), but its scientific rationale is unknown (75). Pediatricians cite low seizure frequency, nocturnal seizures, and parental preference as the leading reasons for nontreatment (72). However, quality of life may suffer regardless of seizure number (76); and the no-treatment approach does not take into account the hidden costs of emergency room attendance, hospital admission, and social stigma for unprevented seizures. Discussion with the child and parents about the merits of treatment vs. nontreatment should cover aspects such as the risk of seizure recurrence, risks associated with generalized seizures, and the adverse effects of both seizures and AEDs on learning and behavior.

The evidence base for AED treatment in RE is acknowledged to be poor (77), with only three Class III studies (ie, open-label) suggesting that carbamazepine and sodium valproate are “possibly” effective, and levetiracetam, oxcarbamazepine, gabapentin, and sulthiame are “potentially” effective as initial monotherapy (78–81). International practice and expert opinion vary widely (82,83). Sulthiame is considered first-line in Germany, Austria, Japan, and Israel (84,85), although concerns have been expressed over its effect on cognition (86); sodium valproate is preferred in France, while levetiracetam is popular in the United States. Carbamazepine and lamotrigine are recommended as first-line monotherapy in the UK and by others (87–89), although there are theoretical risks of carbamazepine exacerbating seizures and the EEG abnormality, and worsening speech production (90,91). There is clearly scope for a randomized, controlled trial in this common syndrome, to determine whether treatment is superior to nontreatment, which AED is best, and how AEDs impact on learning and behavior.

Children should ideally be professionally evaluated for language, literacy, attentional, and fine motor impairments, and an educational plan made, ideally with a multidisciplinary team, to reduce the impact of impairments. There are multiple educational interventions used in dyslexia, but no randomized controlled trials to show which are (more) effective in RE-associated RD.

The management of ESES in RE is not standardized. A definite clinical deterioration should be documented, and other factors such as exacerbation by AEDs (carbamazepine) excluded. Ethosuximide or clobazam can be added to the regimen for a short period, and EEG and neurocognitive evaluation repeated after several weeks to assess effectiveness.

Course and Prognosis

There are no large cohort studies of RE, but most children will have fewer than 10 seizures, and about 15% will only have a single seizure; those with more than one seizure have a median period of active seizures of two and a half years; however, perhaps 10% of patients will have frequent seizures, with a lifetime total of >100; two-thirds will experience a generalized seizure (45). The seizure prognosis is excellent—90% are seizure-free for more than 10 years (45). The risk of GTCS in adult life is less than 2% (similar to the normal population). There is no evidence of increased mortality or SUDEP in RE (92,93).

Fewer than 1% of children with RE have so-called atypical evolutions, which include change in semiology toward PS (94); toward CSWSS (50); or toward LKS (95); and intermediate forms (52). Suspicious new symptoms such as the acute or subacute development of severe linguistic, cognitive, or behavioral problems merit a sleep EEG to investigate the presence of continuous spike-and-wave during slow-wave sleep (CSWSS); these evolutions can sometimes be precipitated by carbamazepine (96). CSWSS may also be seen in children with opercular status characterized by continuous positive or negative myoclonias around the mouth or face, as well as pseudobulbar problems. Atypical benign partial epilepsy of childhood (pseudo-Lennox syndrome, see Chapter 24) in which other seizure types, including tonic and atypical absence seizures, occur may also develop in children with otherwise typical rolandic epilepsy (3). There are no simple clinical markers that predict atypical evolutions (97).

The outlook for cognitive functioning is less clear-cut as there are only a few, small-scale studies available: in a study of 28 patients followed up over 2 to 5 years from an original cohort of 42, initial deficits in verbal and phonemic manipulation improved over time, but deficits in phonologic awareness and visual memory showed no change (98); a systematic review of attentional impairments in RE suggests resolution occurs with remission of EEG abnormalities (61). There are no longitudinal studies of literacy or fine motor coordination in RE, but in the general population, children with dyslexia or dyspraxia usually have residual impairments in adult life.

REFERENCES

Stay updated, free articles. Join our Telegram channel

Join