| |

Received: 20 March 2023    Revised: 21 June 2023    Accepted: 5 July 2023

DOI: 10.1111/epi.17711

RESEARCH ARTICLE

Antiseizure medication selection and retention for adult

onset focal epilepsy in a Swedish health service region:
A population-­based cohort study

David Larsson1,2,3   | Deala Mroué4  | Kerstin Andrén5   | Johan Zelano1,2,3

1
Institute of Neuroscience and
Physiology, Department of Clinical Abstract
Neuroscience, Sahlgrenska Academy, Objective: Historically, approximately half of those with newly diagnosed epi-
Gothenburg University, Gothenburg,
lepsy have responded to and tolerated the first antiseizure medication (ASM),
Sweden
2
Department of Neurology, Sahlgrenska
but there are few contemporary real-­world data. Third-­generation ASMs have
University Hospital, Gothenburg, improved tolerability and are increasingly used according to prescription data.
Sweden We aimed to describe current ASM selection and retention in adult onset focal
3
Wallenberg Center of Molecular and
epilepsy in western Sweden.
Translational Medicine, Gothenburg
University, Gothenburg, Sweden Methods: A multicenter retrospective cohort study was performed at five public
4
Department of Neurology, Södra neurology care providers in western Sweden (nearly complete coverage in the
Älvsborg Hospital, Borås, Sweden area). We reviewed 2607 medical charts and included patients diagnosed with
5
Angered Hospital, Sjukhusen i Väster nongeneralized epilepsy after January 1, 2020 who had a seizure onset after age
Hospital Group, Gothenburg, Sweden
25 years (presumed focal onset) and were started on ASM monotherapy.
Correspondence Results: A total of 542 patients (median age at seizure onset = 68 years, inter-
David Larsson, Department of
quartile range = 52–­ 77) were included. Most patients received levetiracetam
Neurology, Sahlgrenska University
Hospital, Blå Stråket 7, 413 45, (62%) or lamotrigine (35%), with levetiracetam being more common among men
Gothenburg, Sweden. and those with structural causes or short epilepsy duration. During follow-­up
Email: [email protected]
(median = 471.5 days), 463 patients (85%) remained on the first ASM. Fifty-­nine
Funding information (18%) patients discontinued levetiracetam, and 18 (10%) ended treatment with
The Swedish state, Grant/ lamotrigine (p = .010), most commonly because of side effects. In a multivariable
Award Number: ALF-­agreement
(ALFGBG-­965029); Västra
Cox regression model, the discontinuation risk was higher for levetiracetam than
Götalandsregionen, Grant/Award lamotrigine (adjusted hazard ratio = 2.01, 95% confidence interval =  1.16–­3.51).
Number: Innovationsfonden Significance: Levetiracetam and lamotrigine were the dominating first ASMs for
(VGRINN-­992175)
adult onset focal epilepsy in our region, indicating good awareness of problems
with enzyme induction or teratogenicity of older drugs. The most striking finding
is the high retention rates, perhaps reflecting a shift toward an older epilepsy pop-
ulation, higher tolerability of newer ASMs, or suboptimal follow-­up. The finding
that treatment retention differed among patients receiving levetiracetam and la-
motrigine aligns with the recent SANAD II results. It suggests lamotrigine may
be underutilized in our region and that education efforts are needed to ensure it
is considered the first choice more often.

This is an open access article under the terms of the Creative Commons Attribution-NonCommercial License, which permits use, distribution and reproduction in any
medium, provided the original work is properly cited and is not used for commercial purposes.
© 2023 The Authors. Epilepsia published by Wiley Periodicals LLC on behalf of International League Against Epilepsy.

Epilepsia. 2023;00:1–8.  wileyonlinelibrary.com/journal/epi  |  1

 |

15281167, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/epi.17711 by Yenny maharani - Nat Prov Indonesia , Wiley Online Library on [30/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2       LARSSON et al.

KEYWORDS
antiepileptic drug, discontinuation, lamotrigine, levetiracetam, seizure, withdrawal

1  |  I N T RO DU CT ION
Key Points
Selection of initial antiseizure medication (ASM) is a core
element in epilepsy care. Ideally, the type of epilepsy sug- • Levetiracetam and lamotrigine accounted for
gests a range of appropriate ASMs, among which the final >95% of first ASMs prescribed by neurologists
choice is based on patient characteristics such as presumed for newly diagnosed adult onset focal epilepsy
side effect sensitivities, comorbidities, and pharmacologi- • The high retention rates highlight the impor-
cal interactions. Teratogenicity adds to the complexity of tance of selecting appropriate ASMs
ASM selection in women of childbearing age. Treatment • Treatment discontinuation was more common
response is another important aspect, but more challeng- among patients receiving levetiracetam than
ing to integrate into clinical practice, because it remains those receiving lamotrigine
difficult to predict how a particular patient will respond to
a specific drug. Nevertheless, seizure control can still play
a significant role in ASM selection, especially when seek- 2  |  MATERIALS AND METHO D S
ing to attain an effective dose quickly.
In pivotal studies, approximately 50% of adult patients 2.1  |  Study design and setting
treated at the Epilepsy Unit of the Western Infirmary in
Glasgow, Scotland, responded to and tolerated the first We conducted a multicenter retrospective cohort study
ASM.1,2 Despite the introduction of new ASMs, a reanal- based on medical charts review. The study is the initial
ysis incorporating data until 2014 showed no marked in- (baseline) report of a quality improvement project in a
crease in seizure freedom rates or response rates to the geographical region in western Sweden (Västra Götaland
initial regime.3 A Finnish study of older adults with epi- County; population 1.7 million), performed at five of six
lepsy onset in 2000–­2013 found retention of the first ASM public neurology care providers (Figure 1). One regional
in 64%, with carbamazepine and valproic acid being the hospital, with a catchment population of approximately
most common choices.4 Slightly higher rates of retention 260 000 inhabitants, did not participate. Private centers
of initial therapy were reported in the SANAD random- provide only a minor proportion of epilepsy care in the
ized trials; of particular relevance for adult practice has catchment area (no acute or inpatient care), which im-
been the superiority of lamotrigine (LTG), which had plies almost complete coverage.
fewer withdrawals due to side effects than comparators.5,6
For adult epileptologists, the SANAD results and in-
creased awareness of the detrimental effects of enzyme-­ 2.2  |  Study population
inducing drugs have led to a marked shift in ASM selection.
In Sweden, prescription statistics show a decreased use of We searched the electronic systems at each neurology de-
carbamazepine (CBZ) and increased use of levetiracetam partment for epilepsy-­related diagnostic codes (recorded
(LEV) and LTG in adults with epilepsy.7,8 We have performed at each medical appointment or hospitalization). All pa-
several analyses based on prescription data on a nationwide tients with a first diagnostic code of G40 (epilepsy and
scale, and found these two drugs to have the highest reten- recurrent seizures) from January 2020 to June 2022 were
tion rates in agreement with the SANAD results.7,9,10 identified, and their medical charts were examined for in-
In an ongoing project aiming to improve ASM selection formation about epilepsy and ASM therapy. We included
in our health service area (western Sweden), we collected all patients newly diagnosed with epilepsy (after January
preintervention data on ASM selection and retention in 1, 2020) who had a seizure onset after 25 years of age and
adult patients recently diagnosed with epilepsy and who were started on ASM monotherapy. Patients diagnosed
had a seizure onset after age 25 years (presumed focal with generalized epilepsy before treatment initiation
onset). In addition to describing the current clinical situa- (baseline) were excluded; because of the age threshold of
tion regarding the selection and withdrawal of first ASMs, 25 years, patients with unknown epilepsy type were pre-
we wanted to validate previous register-­based analyses of sumed to have focal epilepsy. Figure 2 illustrates the eli-
prescription data. gibility process.
 |

15281167, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/epi.17711 by Yenny maharani - Nat Prov Indonesia , Wiley Online Library on [30/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
LARSSON et al.   
   3

F I G U R E 1   Map illustrating Västra Götaland County in Sweden, its municipality zones, and population density (inhabitants per square
kilometer as of 2016). The orange-­colored zones represent the catchment area of the neurology units that chose to participate in the project.
One regional hospital did not participate; its catchment area is colored in grayscale.

2.3  |  Data collection in GraphPad Prism version 9. Descriptive statistics are pre-
sented as median (interquartile range [IQR]) or number
The review of medical charts was performed by a neu- (%). We used the Fisher exact test, the χ2 test, or the Mann–­
rology resident (D.M.) in August and September 2022. Whitney U-­test to compare baseline group differences. In
Clinical data were extracted from medical charts onto a some circumstances, for example, when >20% of cells had
pseudonymized report form and entered into a study data- expected frequencies of less than five, categories had to be
base for analysis. A sample of the collected data was veri- pooled for statistical significance testing. All tests were two-­
fied by senior researchers (D.L., J.Z.) to ensure accuracy. sided and considered statistically significant at p < .05.
Collected variables included age, sex, comorbidities (focus We used the Kaplan–­Meier method and Cox propor-
on causes of epilepsy), seizure type, date of first seizure, tional hazards regression models to assess ASM retention
date of last known seizure, number of seizures before and discontinuation. The survival time was calculated
ASM start, information on first and second ASMs, reasons from treatment start to either treatment discontinuation
for ASM withdrawal, and date of death. (event), death, or end of observation (the date of the med-
ical chart review), whichever came first. Retention rates
were extracted from survival tables, and the standard
2.4  |  Statistical analysis error was multiplied by 1.96 to estimate the 95% confi-
dence interval (CI). The drug with the highest retention
All analyses were performed in IBM SPSS Statistics version rate became the reference when evaluating the hazard of
28, except the Kaplan–­Meier curves, which were created discontinuation in Cox regression models.
 |

15281167, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/epi.17711 by Yenny maharani - Nat Prov Indonesia , Wiley Online Library on [30/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
4       LARSSON et al.

F I G U R E 2   Flowchart. ASM,
antiseizure medication; ICD, International
Classification of Diseases, 10th Revision.

The multivariable Cox regression model included age initial ASM monotherapy (Figure 2). The median follow-
at the start of ASM therapy (continuous), sex, duration ­up (from treatment start to death or end of observation)
between epilepsy onset and treatment start (continuous), was 471.5 days (IQR =  256–­686). Table 1 presents the study
number of seizures before treatment start, and etiology population's characteristics; most patients had epilepsy
(structural vs. nonstructural; the category "other" was of structural or unknown origin. Twenty-­ two percent
pooled with "unknown" due to low frequency). The covari- (n = 119) were 50 years or younger; 33% (n = 178) were
ates were preselected, but we also performed univariable aged <60 years. Forty-­two percent (n = 228) of the cohort
analyses to avoid missing other variables of importance; received their diagnosis at the tertiary hospital (Table S2).
the only variables with p < .10 were ASM therapy and the
number of seizures before treatment start (Table S1).
3.2  |  First ASM

2.5  |  Ethics statement Three hundred thirty-­four (61.6%) patients were started
on LEV and 190 (35.1%) on LTG (Figure  3). The third
The Swedish Ethical Review Authority approved the study most commonly used ASM was lacosamide (n = 7). There
(approval no. 2022–­214) and waived the need for patient was no significant difference in the ASM prescription pat-
consent. We confirm that we have read the Journal's posi- tern between the tertiary and regional hospitals (Table S2;
tion on issues involved in ethical publication and affirm p = .302). The proportion of LTG prescriptions was some-
that this report is consistent with those guidelines. what higher among younger individuals (25–­ 50 years:
LTG, n = 50 [42%]; LEV, n = 62 [52.1%]).

3  |  R E S U LTS
3.3  |  Retention of first ASM
3.1  |  Study population
Given their overwhelming proportions, only LTG and
The cohort included 542 adult individuals with newly LEV lent themselves to comparative analysis. Patients
diagnosed epilepsy (presumed to have a focal onset) and prescribed LTG were younger at epilepsy onset, had a
 |

15281167, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/epi.17711 by Yenny maharani - Nat Prov Indonesia , Wiley Online Library on [30/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
LARSSON et al.   
   5

T A B L E 1   Characteristics of the cohort, including stratification for the two most common first ASMs.

Characteristic All, n = 542 Levetiracetam, n = 334 Lamotrigine, n = 190 p

Age at epilepsy onset, years, median 68 (52–­77) 70 (54–­79) 65 (41–­77) .017
(IQR)
Sex, n (%) <.001
Male 299 (55) 203 (61) 85 (45)
Female 243 (45) 131 (39) 105 (55)
Focal seizure type, n (%) 416 (77) 260 (78) 139 (73) .242
Duration until ASM start, n (%) <.001
<2 weeks 245 (45) 187 (56) 49 (26)
2 weeks–­3 months 69 (13) 50 (15) 17 (9)
3 months–­2 years 125 (23) 61 (18) 59 (31)
>2 years 103 (19) 36 (11) 65 (34)
Number of seizures before ASM start, <.001
n (%)
1 166 (31) 127 (38) 36 (19)
2 164 (30) 102 (31) 59 (31)
3–­9 91 (17) 48 (14) 36 (19)
≥10 121 (22) 57 (17) 59 (31)
Etiology, n (%) <.001
Structural 295 (54) 211 (63) 72 (38)
Other 4 (1) 2 (1) 2 (1)
Unknown 243 (45) 121 (36) 116 (61)
Comorbidity, n (%)a
Dementia 43 (8) 26 (8) 14 (7) 1.00
Intracranial tumor 35 (7) 28 (8) 7 (4) .045
Stroke 166 (31) 123 (37) 35 (18) <.001
Traumatic brain injury 23 (4) 18 (5) 5 (3) .184
Abbreviations: ASM, antiseizure medication; IQR, interquartile range.
a
Commonly identified causes of adult onset epilepsy.

longer epilepsy duration before starting treatment, and

less frequently had a structural cause (Table 1).
The bulk of the cohort (n = 463 [85.4%]) remained on
the same drug during the observation period. Treatment
discontinuation was more common among patients re-
ceiving LEV than those receiving LTG (Table 2). The dom-
inant reason (78%) for withdrawal of either drug was side
effects. Figure 4 illustrates the survival-­adjusted probabil-
ity of continued treatment for LTG and LEV, assessed with
the Kaplan–­Meier estimator. The 2-­year retention rate was
89.8% (95% CI = 85.3%–­94.3%) for LTG and 79.4% (95%
CI = 74.5%–­84.3%) for LEV. The retention rate for LTG
was similar when stratifying the analysis for age groups
(25–­50 years: 89.6%, 95% CI  =  81.0%–­98.2%; >50 years:
F I G U R E 3   Proportions of first antiseizure medications 89.8%, 95% CI = 84.5%–­95.1%), whereas the retention rate
(ASMs) in the cohort. CBZ, carbamazepine; LCM, lacosamide; for LEV was lower in the younger age strata (25–­50 years:
LEV, levetiracetam; LTG, lamotrigine; OXC, oxcarbazepine; TPM, 69.4%, 95% CI =  56.5.0%–­82.3%; >50 years: 82.0%, 95%
topiramate; VPA, valproic acid. CI =  76.9%–­87.1%).
 |

15281167, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/epi.17711 by Yenny maharani - Nat Prov Indonesia , Wiley Online Library on [30/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
6       LARSSON et al.

T A B L E 2   Information about
Levetiracetam, Lamotrigine,
treatment discontinuation stratified for
n = 334 n = 190 p
the two most common first ASMs.
a
Median follow-­up, days (IQR) 478 (245–­678) 468 (284–­722) .256
Discontinued treatment, n (%) 59 (18) 18 (10) .010
Reason for discontinuation, n (%) .907
Inefficacy 11 (19) 3 (17)
Adverse effects 46 (78) 14 (78)
Other 2 (3) 1 (6)
Most common adverse effects,
n (%)b,c
Mood or behavior changes 27 (59) 1 (7)
Tiredness, somnolence 16 (35) 2 (14)
Rash, pruritus 0 (0) 9 (64)
Dizziness 4 (9) 1 (7)
Psychosis, hallucinosis 4 (9) 0 (0)
b
Most common second ASM (%) Lamotrigine (86) Levetiracetam (83)
Abbreviations: ASM, antiseizure medication; IQR, interquartile range.
a
Follow-­up until death or end of observation.
b
Among those who discontinued their first ASM.
c
Each patient may describe several adverse effects.

F I G U R E 4   Kaplan–­Meier curves illustrating retention rates (treatment continuation) of the two most common antiseizure medications,
including stratification for age groups. LEV, levetiracetam; LTG, lamotrigine; P, p-­value obtained from the log-­rank test.

Patients receiving LEV had a higher risk of treat- significant difference between the drugs in individuals
ment discontinuation (crude hazard ratio [HR] = 1.98, older than 50 years (crude HR = 1.80, 95% CI = .97–­3.34;
95% CI = 1.17–­3.35) than those receiving LTG (reference adjusted HR = 1.64, 95% CI = .86–­3.15).
group). Adjustments for age, sex, epilepsy severity (du-
ration and number of seizures before ASM start), and
etiology (structural vs. nonstructural) had a negligible 4  |  DISC USSION
impact (adjusted HR = 2.01, 95% CI = 1.16–­3.51). In sub-
group analyses stratified by age (25–­50 or >50 years at In this multicenter regional study of current ASM selec-
ASM initiation), patients receiving LEV had a higher risk tion practices, the most striking finding is the domina-
of treatment discontinuation in the younger age group tion of LEV and LTG, which accounted for >95% of ASMs
(crude HR = 2.80, 95% CI = 1.02–­7.64; adjusted HR = 3.53, prescribed by neurologists for newly diagnosed adult
95% CI = 1.20–­10.42). We could not verify a statistically onset focal epilepsy. This pattern resembles the trend of
 |

15281167, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/epi.17711 by Yenny maharani - Nat Prov Indonesia , Wiley Online Library on [30/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
LARSSON et al.   
   7

increasing popularity for both drugs described in Glasgow, sample. Västra Götaland County is very similar to the
Scotland until 2012, at which time they were equally com- whole of Sweden regarding demographic factors (age, sex,
mon.3 In 2020–­2021, LEV was prescribed almost twice as education level) and treatment recommendations (na-
frequently as LTG in our region. tional epilepsy guidelines; LEV, LTG, and CBZ are first-­
The other general observation is the large proportion line alternatives, although the latter is not advised as the
of patients continuing with their first ASM. The high first choice in the elderly), suggesting the findings can
retention might indicate that frequent use of LEV and be generalized nationally. Nonetheless, ASM prescribing
LTG results in fewer withdrawals than previous ASM practices and follow-­up routines vary between countries,
practices or that patients are not followed vigorously which limits the results' generalizability. Another strength
enough regarding side effects. The COVID-­19 pandemic is the medical chart review, which ensures case ascer-
at the time of the present study could have resulted in tainment and provides information on disease severity.
less-­
than-­optimal follow-­ up, with patients not being Weaknesses include the retrospective nature of the inves-
asked sufficiently about side effects or seizures. Another tigation, making it sometimes difficult to extract reasons
important factor might be age; the median age at treat- for withdrawal from the medical charts, and the relatively
ment start was 69 years in our population, compared short follow-­up, which, on the other hand, makes the in-
to 33 years in the Glasgow cohort, where the long-­term formation on ASM selection more up-­to-­date. One must
response rate was 50% in all adults but 83% in the el- also keep in mind that we only included patients with sei-
derly.1 Even so, in our cohort, 81.5% of individuals aged zure onset after age 25 years; children and younger adults
25–­50 years retained their treatment during follow-­up. may have different drug responses.
Interestingly, age was not a significant predictor of ASM LEV and LTG both seem to be relatively well-­tolerated
discontinuation in the regression analyses; it might be ASMs in focal, adult onset epilepsy. LTG is perhaps un-
more important in cohorts with lower median age or derutilized for optimal outcomes in our health service
different ASM prescription patterns, for example, more area. Educational efforts are needed to ensure that LTG is
frequent use of first-­generation ASMs. considered the first choice more often. Information efforts
The relatively short follow-­up (median = 471.5 days) are already underway in our health care organization, and
suggests that our investigation mainly evaluates short-­ a follow-­up study is planned for 2024–­2026.
term tolerability. In both the SANAD studies and the
Glasgow cohort, treatment failure due to seizure re-
lapse did not occur solely in the first year.3,5,6 As ex- 5  |  CONC LUSIONS
pected, we found side effects to be the most common
reason for drug discontinuation in our population. More than 95% of patients with newly diagnosed adult
Interestingly, more patients discontinued LEV than onset focal epilepsy in our health service region were
LTG, and the difference remained statistically signifi- prescribed initial monotherapy with either LEV or LTG.
cant after adjustments, including epilepsy severity. The Most of them remained on the first ASM throughout the
finding provides real-­world evidence in line with the follow-­up period; the high retention may reflect improved
recent SANAD II trial.5 tolerability of newer ASMs, that our cohort was older than
Our inclusion criteria merit some consideration. The many previous ones, or less than ideal follow-­up. Patients
age threshold set at 25 years allowed us to presume that receiving LEV had a higher risk of discontinuation than
the vast majority had focal epilepsy. Patients diagnosed those receiving LTG.
with generalized epilepsy at baseline were excluded; how-
ever, we will have included some subjects who initially AUTHOR CONTRIBUTIONS
have unknown epilepsy types (either because the epilepsy David Larsson: Data validation; analyses; interpreta-
type is truly unknown or due to incomplete/missing infor- tion of results; drafting and revision of manuscript. Deala
mation) but who eventually will be diagnosed with gen- Mroué: Data collection; interpretation of results; revi-
eralized epilepsy. Still, we consider this a study of focal sion of manuscript. Kerstin Andrén: Case identification;
epilepsy and would expect this misclassification to have a interpretation of results; revision of manuscript. Johan
limited and negligible effect on the results. Zelano: Conceptualization; supervision of data collec-
Contemporary regionwide studies of real-­world ASM tion; analysis planning; interpretation of results; drafting
selection and retention are rare, and usually focused on and revision of manuscript.
particular patient groups or based on health care adminis-
trative data.11–­14 The main strength of our investigation is ACKNOWLEDGMENTS
the unbiased inclusion of all patients at five neurology de- The study was funded by the VGR Innovations fund and
partments in our region, resulting in a population-­based the Swedish state through the ALF agreement.
 |

15281167, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/epi.17711 by Yenny maharani - Nat Prov Indonesia , Wiley Online Library on [30/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
8       LARSSON et al.

8. Bolin K, Berggren F, Berling P, Morberg S, Gauffin H, Landtblom

CONFLICT OF INTEREST STATEMENT AM. Patterns of antiepileptic drug prescription in Sweden:
J.Z. reports outside the submitted work: speaker honoraria a register-­
based approach. Acta Neurol Scand. 2017;136:​
for unbranded educational presentations from UCB and 521–­7.
Eisai; consultancy fee from the Swedish Medical Products 9. Hakansson S, Zelano J. Big data analysis of ASM retention rates
agency; honoraria from Neurologi i Sverige, Liber, and and expert ASM algorithm: A comparative study epilepsia.
Studentlitteratur; and as an employee of Sahlgrenska Epilepsia. 2022;63:1553–­62.
University hospital, being investigator or subinvestigator 10. Hakansson S, Karlander M, Larsson D, Mahamud Z, Garcia-­
Ptacek S, Zelezniak A, et al. Potential for improved reten-
in clinical trials sponsored by Bial, SK Life Science, GW
tion rate by personalized antiseizure medication selection: a
Pharma, and UCB. The remaining authors have no con- register-­based analysis. Epilepsia. 2021;62:2123–­32.
flicts of interest. 11. Shouman W, Delaney JA, Kowalec K, Ng M, Ruth C, Falk J,
et al. Trends of utilization of antiseizure medications among
ORCID pregnant women in Manitoba, Canada: a 20-­year population-­
David Larsson  https://orcid.org/0000-0002-8334-951X based study. Pharmacol. 2022;13:871136.
Kerstin Andrén  https://orcid.org/0000-0001-6797-6348 12. Terman SW, Niznik JD, Slinger G, Otte WM, Braun KPJ, Aubert
CE, et al. Incidence of and predictors for antiseizure medica-
Johan Zelano  https://orcid.org/0000-0001-9445-4545
tion gaps in Medicare beneficiaries with epilepsy: a retrospec-
tive cohort study. BMC Neurol. 2022;1(22):328.
REFERENCES 13. Wojcik K, Franciszek Kolek M, Dec-­Cwiek M, Slowik A, Bosak
1. Mohanraj R, Brodie MJ. Diagnosing refractory epilepsy: re- M. Trends in antiseizure medications utilization among women
sponse to sequential treatment schedules. Eur J Neurol. 2006 of childbearing age with epilepsy in Poland between 2015 and
Mar;13:277–­82. 2019. Epilepsy Behav. 2023;20(139):109091.
2. Kwan P, Brodie MJ. Early identification of refractory epilepsy. N 1 4. Terman SW, Youngerman BE, Choi H, Burke JF. Antiseizure
Engl J Med. 2000 Feb;3(342):314–­9. medication treatment pathways for US Medicare bene-
3. Chen Z, Brodie MJ, Liew D, Kwan P. Treatment Outcomes ficiaries with newly treated epilepsy. Epilepsia. 2​ 022;​6 3:​
in Patients With Newly Diagnosed Epilepsy Treated With 1571–­9 .
Established and New Antiepileptic Drugs: A 30-­year longitudi-
nal cohort study. JAMA Neurol. 2018 Mar;1(75):279–­86.
4. Bruun E, Kalviainen R, Keranen T. Outcome of initial antiepi- SUPPORTING INFORMATION
leptic drug treatment in elderly patients with newly diagnosed Additional supporting information can be found online
epilepsy. Epilepsy Res. 2016 Nov;127:60–­5. in the Supporting Information section at the end of this
5. Marson A, Burnside G, Appleton R, Smith D, Leach JP, Sills G, et al.
article.
The SANAD II study of the effectiveness and cost-­effectiveness of
levetiracetam, zonisamide, or lamotrigine for newly diagnosed
focal epilepsy: an open-­label, non-­inferiority, multicentre, Phase
4, randomised controlled trial. Lancet. 2021;397:1363–­74.
6. Marson AG, Al-­Kharusi AM, Alwaidh M, Appleton R, Baker How to cite this article: Larsson D, Mroué D,
GA, Chadwick DW, et al. The SANAD study of effectiveness Andrén K, Zelano J. Antiseizure medication
of carbamazepine, gabapentin, lamotrigine, oxcarbazepine, or selection and retention for adult onset focal
topiramate for treatment of partial epilepsy: an unblinded ran-
epilepsy in a Swedish health service region: A
domised controlled trial. Lancet. 2007;24(369):1000–­15.
7. Larsson D, Asberg S, Kumlien E, Zelano J. Retention rate of
population-­based cohort study. Epilepsia.
first antiepileptic drug in poststroke epilepsy: a nationwide 2023;00:1–8. https://doi.org/10.1111/epi.17711
study. Seizure. 2019 Jan;64:29–­33.