EEG in Adult-onset Idiopathic Generalized Epilepsy

Authors


Address correspondence and reprint requests to Dr. S. Berkovic at Epilepsy Research Institute, Austin and Repatriation Medical Centre, Locked Bag 1, Heidelberg West 3081, Australia. E-mail: s.berkovic@unimelb.edu.au

Abstract

Summary:  Purpose: It remains controversial whether adult-onset idiopathic generalized epilepsy (IGE) is a distinct syndrome or a continuum among IGE syndromes. EEG is the only known biologic marker of IGE and helps differentiate many of its classic subsyndromes. In this study, we looked for the differences in the EEG findings of IGE of classic adolescent onset versus adult onset that may suggest syndromic heterogeneity.

Methods: Seventy-six patients (47 adolescent-onset IGE, 29 adult-onset IGE) with a clinical and EEG diagnosis of IGE were included. We defined IGE with age at onset of 11–20 years as adolescent-onset IGE and age at onset of 20 years or after as adult-onset IGE. Patients with first-decade onset of seizures, delayed EEGs, and no EEG available for review were excluded. The first EEG was performed within 24 h of the seizure, and if negative, a sleep-deprived EEG was done. All EEGs were reviewed in detail with respect to the background activity and the generalized spike–wave (GSW) characteristic.

Results: EEGs (87; 56 adolescent-onset IGE, 31 adult-onset IGE) were systematically reviewed. Background was normal in all patients. The morphology, amplitude, duration, frequency, occurrence, or activation of the GSW pattern did not differ between these two groups.

Conclusions: No differences of EEG features were found between the classic adolescent-onset and the adult-onset IGE. This supports the hypothesis that they share common biologic determinants and exist along a life-long age spectrum of classic IGE.

Epilepsy with seizure onset in adult life is generally thought to have a focal basis and symptomatic etiology (1). However, in some patients, idiopathic generalized epilepsy (IGE) is suspected because of typical generalized tonic–clonic, absence, or myoclonic seizures, the finding of generalized spike–wave complexes on EEG, normal brain imaging, and a family history of seizures (2,3). Two hospital-based studies recently reported that 13.4% and 34.8% of IGE cases had seizure onset in adulthood (4,5), although IGE typically appears within the first two decades of life. Hauser et al. (6) reported that the annual age-specific incidences of IGE patients aged 15–24 years and 25–34 years were 3.6 and 3.5 per 100,000 respectively, whereas the incidences of IGE patients aged 5–9 years and 10–14 years were 10.7 and 15.3 per 100,000 respectively. Whether adult-onset IGE represents a later age-dependent peak or the far right end of the typical age spectrum of typical IGE is uncertain. We previously reported similar family data in the adolescent-onset and adult-onset groups, suggesting these patients have the same syndrome (7). We examined the EEGs in adolescent-onset IGE and adult-onset IGE to see if there were age-related differences that might suggest syndromic heterogeneity, because the EEG is the only known biologic marker of this condition.

METHODS

Patients

We retrospectively reviewed the diagnoses of patients attending the First Seizure Clinic of the Austin and Repatriation Medical Centre (ARMC), Melbourne, from March 1994 to December 2000. The First Seizure Clinic recruits patients with new-onset seizures from emergency departments of hospitals of the northeast sector of Melbourne, representing a community-based cohort (8). The 121 patients with a clinical and EEG diagnosis of IGE, as defined by the Commission on Classification and Terminology of the International League Against Epilepsy (ILAE) (9), were identified. All patients had detailed clinical history, neurologic examination, computed tomography (CT) and/or magnetic resonance imaging (MRI) of the head and EEG. We included only patients with diagnostic EEGs showing generalized spike–wave (GSW) patterns in whom EEGs were available for review.

We defined IGE with age at onset between 11and 20 years as adolescent-onset IGE, and IGE with age at onset of 20 years or after as adult-onset IGE. Based on the seizure type and age at onset of seizure, we categorized the patients in the adolescent-onset IGE group as having juvenile absence epilepsy (JAE), juvenile myoclonic epilepsy (JME), and IGE with only tonic–clonic seizures (TCSs), and in the adult-onset IGE group as adult absence epilepsy (AAE), adult myoclonic epilepsy (AME), and adult-onset IGE with only tonic–clonic seizures (ATCS). We excluded (a) patients with first-decade onset of seizures because these were few and represented a biased proportion at this adult center; (b) patients with delayed EEGs >3 months from seizure onset; and (c) patients with no EEG available for review.

EEG recordings

EEGs were performed within 24 h of seizure onset, or at the earliest opportunity if the seizure occurred >24 h before referral, and invariably before commencement of medication in nearly all cases. If the first EEG was negative, a sleep-deprived EEG was done. We used 16-channel paper recordings or 21-channel digital recordings with bipolar longitudinal and transverse montages, average reference and source-derived montages, and standard 10-20 electrode placement. Recordings lasted for 20–30 min for resting awake and/or sleep states, including hyperventilation for 3 min and intermittent photic stimulation at 1–20 Hz with the eyes open and closed at each frequency. Sleep-deprived recordings lasted for 30 min, or longer if needed for sleep to occur, and included photic stimulation and hyperventilation before going to sleep.

Evaluation

Two investigators (S.Y. and A.S.H.) together reviewed all available EEGs on these patients in a systematic fashion, blinded to their identity and age. The background rhythm was assessed for the posterior dominant rhythm, amplitude, and frequency. We analyzed the occurrence, amplitude, frequency and duration of GSW and generalized polyspike–waves (GPSWs) in the resting state and with activation (during sleep, hyperventilation, and photic stimulation). We also looked for paroxysmal slow activity and fast activity.

Statistical comparison was performed by using the Student t test for the amplitude, frequency, and duration variables and χ2 or Fisher's exact tests for the other categoric variables (significance, p < 0.05).

RESULTS

Patients

Of the initial 121 patients, 45 were excluded by these criteria (five patients with first-decade onset of seizures, four patients with delayed EEG >3 months from seizure onset, and 36 patients with no EEG available for review). Seventy-six patients (47 adolescent-onset IGE, 29 adult-onset IGE) had early diagnostic EEGs available for review. In the adolescent-onset group, 27 male and 20 female patients had a mean age at onset seizure of 16.5 years (range, 11.7–19.9) years. In the adult-onset IGE group, the 15 male and 14 female patients had a mean age at seizure onset of 32.1 years (range, 20–75 years). Figure 1 shows the distribution of age at onset of seizures among the 76 patients.

Figure 1.

The distribution of age at onset of seizures among the 76 patients with idiopathic generalized epilepsy.

Epilepsy syndromes were categorized in each patient group. Of the 47 adolescent-onset IGE patients, 53% had TCS, 26% had JME, and 21% had JAE. Of the 29 adult-onset IGE patients, 69% had ATCS, 24% had AME, and 7% had AAE. Only one patient in each group received antiepileptic drugs (AEDs) before EEG recording. CT and MRI were normal in all cases.

EEG recordings

A total of 87 EEGs (56 adolescent-onset IGE, 31 adult-onset IGE) was reviewed. Eleven (20%) of the adolescent-onset group and four (13%) of the adult-onset group were sleep-deprived EEGs. Ten sleep-deprived EEGs of the adolescent-onset group and three sleep-deprived EEGs of the adult-onset group were the second EEGs. More than 50% of EEGs in both adolescent-onset and adult-onset groups were performed within 24 h after the first seizure attack (57 and 58%, respectively). There was no significant difference between the two groups in the number of sleep-deprived EEGs, number of EEGs performed within 24 h after the seizure onset, number of paper and digital recordings, duration of recording, number of EEGs with recording in different conscious states, and number of EEGs with hyperventilation and photic stimulation.

EEG findings

Background activity was normal in all patients, with no difference in the frequency and amplitude of the posterior dominant rhythm in the two groups. There was no difference in the morphology, amplitude, duration, frequency, occurrence, or activation of the GSW pattern between these two groups (Table 1). Figure 2 shows an example of the similarity in GSW between two patients with recent-onset IGE, one aged 14 years and the other aged 75 years. Only the presence of paroxysmal slow activity and the subjective impression of a frontal predominance of GSWs were increased in the adolescent-onset IGE compared with the adult-onset IGE (32 vs. 13%, and 25 vs. 7% respectively; p < 0.05). When corrected for multiple comparisons with the Bonferroni correction (10), they did not reach statistical significance.

Table 1.  EEG patterns of epileptiform discharges in patients with adolescent-onset and adult-onset idiopathic generalized epilepsy
EEG findingsAdolescent
onset
Adult
onset
  • G, generalized; SW, spike–wave; PSW, polyspike–wave; HV, hyperventilation.

  • a

     Statistically significant difference, p = 0.004 (uncorrected for multiple comparisons).

  • b

     Statistically significant difference, p = 0.003 (uncorrected for multiple comparisons).

Number of patients4729
Number of EEGs5631
GSW occurrence (patients)47 (100%)29 (100%)
GPSW occurrence (patients)30 (64%)14 (48%)
GSW occurrence (EEGs)49 (87%)30 (97%)
GPSW occurrence (EEGs)30 (53%)14 (45%)
GSW with frontally predominancea12 (25%)2 (7%)
GSW in awake record (EEGs)26 (47%)20 (64%)
GSW in drowsy record (EEGs)7 (23%)4 (22%)
GSW in sleep record (EEGs)6 (26%)5 (45%)
GSW in sleep record only, not in awake (EEGs)4 (67%)4 (80%)
Hyperventilation activation of GSW (EEGs)22 (43%)10 (37%)
Maximum frequency SW resting and/or HV (Hz)3–7.5 (5.3)3–10 (5.6)
Maximum amplitude SW resting and/or HV (μV)10–250 (84)10–250 (84)
Maximum duration of SW resting and/or HV (s)0.5–21 (3.1)0.5–11 (2.5)
Photic sensitivity (EEGs)20 (36%)6 (20%)
Photic sensitivity flash frequency (median; mode)15; 1715; 9, 15, 17
GSW on sleep-deprivation EEG (EEGs)10 (90%)3 (75%)
Sleep-deprived activation (EEGs)9 (82%)3 (75%)
Paroxysmal slow activity (EEGs)b18 (32%)4 (13%)
Paroxysmal fast activity (EEGs)2 (3%)1 (6%)
Figure 2.

Example of diagnostic EEGs of a patient with adolescent-onset idiopathic generalized epilepsy (A), age 14 years, and a patient with adult-onset idiopathic generalized epilepsy (B), age 75 years, both showing characteristic generalized spike–wave patterns on a normal awake background.

DISCUSSION

Idiopathic generalized epileptic syndromes are heterogeneous with respect to age at onset and seizure manifestations. Spike–wave frequency, morphology, and activation vary between the different syndromes across the first two decades. The etiologic factors and relations between each syndrome are not clearly understood. Some authors postulate that IGE syndrome in adults is a new syndrome (11,12), whereas the concept of a continuum among IGE subjects is debated (13,14). In this study, we looked for etiologically related syndromic differences between the classic group of IGE and the adult-onset group of IGE.

Our data showed no differences in the EEG findings between these two groups of patients, especially in the duration, amplitude, and frequency of GSW and the number of discharges during resting and activated states. However, the first-seizure clinic ascertainment removed a potential bias of hospital-acquired patients, such as the elderly patients with absence status and metabolically induced spike–wave states that may not truly be representative of IGE.

The lack of difference in EEG features of adult-onset and adolescent-onset IGE strongly supports the hypothesis that they share common biologic determinants. Our previous genetic analysis also supports this view (7). The adult-onset occurrence of seizures may be due to other modifying genetic or acquired factors. When considering the infant, preschool, and peripubertal peaks for presentation and exacerbation of IGE syndromes in general, such a proposition is easily understood. Thus a broad adult-onset peak can be added to the well-recognized onset peak of IGE in infancy, childhood, and adolescence.

Ancillary