1 INTRODUCTION

The most common genetic generalized epilepsy syndromes include childhood and juvenile absence epilepsies (CAE and JAE respectively), juvenile myoclonic epilepsy (JME) and generalized tonic-clonic seizures (TCS) upon awakening. In these four specific syndromes, the previously used concept idiopathic generalized epilepsies (IGE) can still be applied, according to the 2017 International League Against Epilepsy (ILAE) guidelines.1 In addition, other less common but still clinically relevant syndromes have been included in the IGE category, such as IGE with absences of early childhood, perioral myoclonia with absences, IGE phantom absences, Jeavons syndrome (eyelid myoclonia with absences) and monogenic IGE syndromes.2 Furthermore, the precise clinical and EEG phenotype of a single patient is often difficult to ascertain even when all information is suggestive of IGE.3

The main aim of epilepsy treatment is to achieve seizure freedom without adverse effects related to the medication. In general, the treatment of IGE syndromes involves the use of broad-spectrum antiseizure medications (ASM). The results from a few rigorous studies comparing outcomes with ASM for IGE have shown valproate to be the most effective ASM when treating generalized epilepsy.4, 5 However, the use of valproate has become problematic in women with epilepsy (WWE) due to its adverse effects. The European Medicines Agency (EMA) published a press release in 20146 and a further 20187 restriction advising clinicians not to prescribe valproate to women of childbearing age due to its teratogenic effects, unless other treatments fail. A joint report by the Commission on European Affairs of the International League Against Epilepsy (CEA-ILAE) and the European Academy of Neurology (EAN) in 2015 strongly agreed with previous EMA guidelines and provided further guidance on the treatment of women and girls of childbearing age with epilepsy.8 The use of valproate during pregnancy has been connected most significantly to the risk of major congenital malformations (MCMs).9, 10 It has also been associated with lower IQ and increased risk of autism and attention deficit hyperactivity disorder in the child.11, 12 Among other broad-spectrum ASMs, lamotrigine and levetiracetam have a low risk of MCMs, and conversely, topiramate has an elevated risk of MCMs and is associated with significant growth restriction and microcephaly.10

There are fewer treatment options available for generalized epilepsies than for focal epilepsies. The effects of the decreased use of valproate on seizure outcomes in WWE have not been well studied.13 Several other ASMs have been used to treat IGE, but their therapeutic potential has not been firmly established.8 However, use of lamotrigine and levetiracetam was recommended as first-line treatments for IGE, except for CAE, for which ethosuximide is the drug of choice.14

Between 2005 and 2008, a single-centre follow-up study of 128 subjects with IGE was conducted at Tampere University Hospital.15 In the current study, we evaluated whether there has been a change in the prescription patterns in the treatment of IGE due to updated treatment recommendations (with special reference to WWE) and, if so, how the reduced use of valproate16 has affected seizure outcomes compared to those described in our previous study.15

2 MATERIALS AND METHODS 2.1 Subjects

The study was carried out at Department of Neurology, Tampere University Hospital, where all patients with epilepsy were diagnosed according to local treatment guidelines and had their treatment initiated in Pirkanmaa Hospital District (population approximately 500,000). Our department also serves as a referral centre for refractory epilepsy patients from other smaller or regional hospitals. This was an observational, noninterventional retrospective study that did not require ethics committee approval according to Finnish Law on Research. Access to patient records was based on a decision made by the Head of Science Centre, Tampere University Hospital Research and Innovation Services, Science Centre.

2.2 Data collection

Patients with generalized epilepsy (including patients with ICD-10 codes referring to generalized epilepsy; G40.3X) who had their visit at Tampere University Hospital from 1 January 2009 to 31 December 2018 were identified from the hospital's patient registry. Patients’ electronic medical records were retrospectively inspected for demographics and clinical information including age, sex, syndrome, intellectual disability (IQ < 70), time of patient's latest visit, age at diagnosis, EEG and MRI history, and complete ASM history (dosages and reasons for discontinuation). Data concerning ASM initiation and changes as well as seizure outcomes and adverse events were registered systematically. A complete seizure history was recorded for TCS, myoclonic jerks, absence seizures and unclassified seizures. The definite IGE syndromes were divided into CAE, JME, JAE and TCS. Furthermore, the possibility of other putative IGE syndromes as proposed by Panayiotopoulos in 20052 was assessed, and patients with eyelid myoclonia (with or without absences) but no others were found in the study group. Patients diagnosed with Unverricht-Lundborg disease, myoclonic-astatic epilepsy and Dravet syndrome were not included in this study. All the patients’ diagnoses, including the cases of unclassified IGE, were re-evaluated by the investigators (RK, SR and JP) to examine whether a diagnosis of a specific syndrome could be made. If not, then the patient remained in the unclassified IGE category. The patients with unclassified IGE needed to have at least one EEG registration with generalized epileptiform activity to be classified in this category, but they did not fit into any specific generalized syndrome category. We did not include any patients with focal epileptiform activity or normal activity in this category because we only wanted to include patients with unequivocal generalized epilepsy. They also needed to have seizure types concordant with generalized onset seizures and absence of any significant focal findings in MRI. Strict age limitations were not used, and both paediatric and adult patients were included.

Duration of seizure freedom was calculated based on the time between the patients’ latest seizure (whether provoked or unprovoked) and the patients’ latest visit at the neurology clinic. Seizure freedom was defined as 12 months without TCS or absence seizures or days with myoclonic jerks. Patients with a duration of epilepsy of under 12 months were excluded from calculations concerning seizure freedom as one-year seizure freedom could not be reached. The duration of epilepsy was calculated as months between the start of the patients’ first medication and their latest visit at the neurology clinic. The use of different ASMs and their combinations as well as the seizure outcome were examined. The use of different ASMs was examined separately for mono- and combination therapy.

The retention rate, referring to the percentage of patients still using a specific ASM at their latest visit at the neurology clinic, was studied for each drug. The reasons for discontinuation were classified as lack of efficacy, lack of tolerability, long-term seizure freedom and other reasons. The seizure freedom for a given ASM was calculated by dividing the number of seizure-free patients using the ASM at their latest visit by the overall number of patients using the specific ASM at their latest visit. The following classifications were made for the subgroup analyses. Age (at the patients’ latest visit) was categorized into two groups: paediatric (<16 years of age) and adult (≥16 years of age), and the subjects were also categorized by sex.

The data of this study are available on reasonable request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.

2.3 Statistical analysis

Descriptive statistics were reported as frequencies and percentages for all variables. Pearson's chi-squared test was used to test the association between sex and the use of different ASMs. In addition, an adjusted odds ratio (OR) for the association between sex and the seizure freedom was evaluated by binary logistic regression model. We chose the factors in the model which are known to have effects on seizure outcomes. Our main question was regarding seizure freedom rates between males and females, but it is well established that epilepsy syndrome and medication selection have a major effect on the clinical outcome in terms of seizure freedom. Because of the natural evolution of generalized epilepsy, we also included the age of the patient because it does have an impact on seizure outcomes. We considered that other possible factors are probably of less importance. The association was adjusted for epilepsy syndrome, age, use of valproate, use of lamotrigine and levetiracetam at latest clinical visit. Statistical significance was defined as a p-value < .05, and all statistical analyses were performed with SPSS version 26.

3 RESULTS 3.1 Patients

Altogether, 263 subjects (166 females and 97 males) were included in the analysis. The mean age was 24.0 (range 2–81) years at the patients’ latest visit and 13.3 (range 0–80) years at the start of medication. The mean duration of epilepsy for those with available data (n = 227) was 9.0 (range 0.4–54) years. We were unable to obtain data for 36 patients concerning the age at the start of their medication and hence the duration of their epilepsy from the electronic patient records. This was due to the patients transferring from another treatment centre or their treatment being started before the electronic registry was established. Altogether, 252 patients were included in the seizure freedom analysis because 11 patients were diagnosed less than 12 months before the latest visit at the neurology clinic. The distribution of patients based on sex, age and syndromes is presented as a flowchart of the study in Figure 1.

Flowchart of the study with the distribution of patients based on sex, age and syndromes. CAE, childhood absence epilepsy; CLB, clobazam; CZP, clonazepam; DS, dravet syndrome; ESM, ethosuximide; IGE, idiopathic generalized epilepsy; JAE, juvenile absence epilepsy; JME, juvenile myoclonic epilepsy; LEV, levetiracetam; LTG, lamotrigine; MAE, myoclonic-astatic epilepsy; TCS, tonic-clonic seizure; TPM, topiramate; ULD, unverricht-Lundborg disease; VPA, valproate; ZNS, zonisamide

3.2 Epilepsy syndromes and age-specific results

Syndrome- and age-specific results are presented in Table 1. Among all patients, 182 (72.6%) patients achieved one-year seizure freedom. Out of 15 patients with intellectual disability, only 6 (42.9%) were seizure free. There was no significant difference in multivariate analysis between the seizure freedom of male and female patients (female vs. male; OR 1.25, p = .48). Age or the use of a distinct ASM did not influence the outcome. Patients diagnosed with TCS only had the highest rate of seizure freedom with 81.0% (n = 51) seizure free at their latest visit. The proportion of seizure-free patients diagnosed with JME and CAE was similar, 75.0% (n = 51) and 75.8% (n = 25) respectively. Regarding the syndrome distribution in the entire study group, 70 (26.6%) JME, 69 (26.2%) TCS only and 51 (19.4%) unclassified IGE composed the largest syndrome diagnoses. The most common syndromes in the paediatric cohort were CAE (47.3%) followed by unclassified IGE (29.7%) and TCS only (12.2%). To compare, the most common syndromes in the adult cohort were JME (36.0%), TCS only (31.7%) and unclassified IGE (15.3%). Details for the binary logistic regression model with odds ratios for all factors used in the model are presented in the supplementary table (Table S1).

TABLE 1. Syndrome- and age-specific results for entire study group Entire study group Paediatric Adult Total 1-year SFa Total No. of ASMs used at latest visit 1-year SFa Total No. of ASMs used at latest visit 1-year SFa N (%) N (%) N (%) 0 1 2 3 N (%) N (%) 0 1 2 3 N (%) Total 263 (100) 182 (72.6) 74 (100) 35 34 4 1 49 (72.1) 189 (100) 17 121 35 16 134 (72.8) Female 166 (63.1) 119 (73.0) 42 (56.8) 21 21 0 0 31 (75.6) 124 (65.6) 11 76 28 9 87 (72.5)

Male

97 (36.9) 64 (71.9) 34 (43.2) 14 13 4 1 18 (66.7) 65 (34.4) 6 45 7 7 46 (74.2) JME 70 (26.6) 51 (75.0) 2 (2.7) 0 2 0 0 1 (100) 68 (36) 3 46 10 9 50 (74.6) TCS only 69 (26.2) 51 (81.0) 9 (12.2) 7 2 0 0 7 (87.5) 60 (31.7) 4 46 10 0 44 (78.6) CAE 36 (13.7) 25 (75.8) 35 (47.3) 15 20 0 0 25 (78.1) 1 (0.5) 0 0 1 0 0 JAE 26 (9.9) 17 (65.4) 4 (5.4) 3 1 0 0 3 (75) 22 (11.6) 3 12 5 2 14 (63.6) Eyelid myoclonia 11 (4.2) 6 (54.5) 2 (2.7) 1 1 0 0 1 (50) 9 (4.8) 3 5 1 0 5 (55.6) Unclassified IGE 51 (19.4) 33 (66.0) 22 (29.7) 9 8 4 1 12 (57.1) 29 (15.3) 4 12 8 5 21 (72.4) Abbreviations: CAE, childhood absence epilepsy; IGE, idiopathic generalized epilepsy; JAE, juvenile absence epilepsy; JME, juvenile myoclonic epilepsy; SF, seizure freedom; TCS, tonic-clonic seizure. a Patients with a duration of epilepsy of under 12 months were excluded when calculating seizure freedom (n = 11). 3.3 ASMs and their efficacy

Antiseizure medications and their efficacy for the entire study group are presented in Table 2 separately for age subgroups comprising of paediatric and adult patients. In the whole study group, valproate (including both previous and present use) was the most widely used ASM (n = 235) in both age subgroups, followed by lamotrigine (n = 84) and levetiracetam (n = 65) (Table 3). Other drugs were used much less with 30 or less patients having used each of these ASMs (Table 3). Although valproate was the most common drug used in both age subgroups, lamotrigine and levetiracetam were more commonly used in adults than in the paediatric population (Table 2). Phenytoin, tiagabine, phenobarbital and perampanel were also previously used by between one and six patients with no patient retaining them until their latest visit (data not shown in tables).

TABLE 2. All drugs ever used and their efficacy for the entire study group Paediatric Adult Reason for discontinuation n (%) Total exposed Retained until latest visit SFa Total exposed Retained until latest visit SFa ASMs N N (%) N (%) N N (%) N (%) Tolerability Lack of efficacy Wrong syndrome Long-time SF Other VPA 70 30 (42.3) 14 (58.3) 165 115 (69.7) 83 (74.8) 28 (34.1) 6 (7.3) 0 42 (51.2) 6 (7.3) LTG 10 4 (40) 1 (25) 74 48 (64.9) 31 (64.6) 7 (25.9) 12 (44.4) 0 6 (22.2) 2 (7.4) LEV 7 2 (28.6) 0 58 38 (65.5) 22 (59.5) 11 (40.7) 10 (37.0) 0 1 (3.7) 1 (3.7) ESM 18 7 (38.9) 1 (14.3) 12 4 (33.3) 2 (50) 1 (5.9) 6 (35.3) 0 9 (52.9) 1 (5.9) TPM 3 1 (33.3) 0 17 8 (47.1) 6 (75) 4 (44.4) 4 (44.4) 0 1 (11.1) 0 CBZ 0 - - 17 2 (11.8) 2 (100) 2 (18.1) 2 (18.1) 3 (27.3) 3 (27.3) 1 (9.1) CLB 6 1 (16.7) - 14 10 (71.4) 3 (30) 2 (40.0) 3 (60.0) 0 0 0 CZP 2 0 - 9 5 (55.6) 1 (20) 2 (40.0) 2 (40.0) 0 0 1 (20.0) OXC 4 0 - 23 4 (17.4) 1 (25) 1 (5.9) 2 (11.8) 12 (70.6) 2 (11.8) 0 ZNS 0 - - 9 5 (55.6) 2 (40) 1 (50.0) 1 (50.0) 0 0 0 Abbreviations: CBZ, carbamazepine; CLB, clobazam; CZP, clonazepam; ESM, ethosuximide; LEV, levetiracetam; LTG, lamotrigine; OXC oxcarbazepine; SF, seizure freedom; TPM, topiramate; VPA, valproate; ZNS, zonisamide. a Patients with a duration of epilepsy of under 12 months were excluded when calculating seizure freedom (n = 11). Seizure freedom was calculated for patients still using the drug (combination) at their latest visit. TABLE 3. Sex-specific results for drugs use for the entire study group Retained until latest visit ASMs Exposed overall (n) Female:Male ratio Total Monotherapy at latest visit Combination therapy at latest visit Monotherapy SF at latest visith Combination therapy SF at latest visith Overall SF at latest visith (Females 166, males 97) (ratio in entire study group 1.7:1) N (%) N (%) Female: male ratio N (%) Female: male ratio % % N (%) VPA 235 1:1.1 145 (61.7) 106 (73.1) 1:1.4 39 (26.9) 1:1.2 82.5 44.7 97 (71.9) Female 143 81 (56.6) 58 (71.6) 23 (28.4) 80.4 47.8 56 (70.9) Male 92 64 (69.6) 48 (75.0) 16 (25.0) 85.4 40.0 41 (73.2) LTG 84 2.3:1 52 (61.9) 17 (32.7) 2.7:1 35 (67.3) 1.5:1 82.4 51.4 32 (61.5) Female 67 39 (58.2) 14 (35.9) 25 (64.1) 85.7 56.9 26 (66.7) Male 17 13 (76.5) 3 (23.1) 10 (76.9) 66.7 40.0 6 (46.2) LEV 65 1.6:1 40 (61.5) 18 (45.0) 2.9:1 22 (55.0) 1.3:1 70.6 45.5 22 (56.4) Female 48 30 (62.4) 15 (50.0) 15 (50.0) 64.3 46.7 16 (55.2) Male 17 10 (58.8) 3 (30.0) 7 (70.0) 100.0 42.