Shift from older‐ to newer‐generation antiseizure medications in people with acute ischemic stroke in Australia: A population‐based study

Abstract

Objective

To investigate the trends in antiseizure medications (ASMs) use following ischemic stroke and to examine factors associated with use of newer‐ and older‐generation ASMs.

Methods

A retrospective cohort study was conducted using state‐wide linked health datasets. Patients who were hospitalized with a first‐ever ischemic stroke between 2013 and 2017 and were dispensed ASM within 12 months from discharge were included. Logistic regression was used to examine the predictors of receiving newer‐generation ASMs. Generalized linear modeling was used to identify factors associated with ASM use after ischemic stroke.

Results

Of 19 601 people hospitalized with a first‐ever ischemic stroke, 989 were dispensed an ASM within 12 months from discharge. The most prevalent first ASMs were levetiracetam (38.0%), valproate (25.8%), and carbamazepine (10.3%). Most people were dispensed ASM monotherapy (86.9%). There was a shift toward the use of newer‐generation ASMs between 2013 and 2017 (odds ratio [OR] 2.82, 95% confidence interval [CI] 1.92–4.16). Metropolitan residents were more likely to be dispensed newer‐generation ASMs as a first‐line treatment (OR 1.79, 95% CI 1.31–2.45). People over 85 years (OR 0.38, 95% CI 0.23–0.64), with dementia (OR 0.35, 95% CI 0.19–0.63) and psychotic comorbidities (OR 0.29, 95% CI 0.09–0.96) were less likely to be dispensed newer‐generation ASMs. Older age (coefficient [β] 0.23, P = 0.030), history of beta blocker use (β 0.17, P = 0.029), multiple ASMs (β 0.78, P < 0.001), and newer‐generation ASM (β 0.23, P = 0.001) were associated with higher defined daily dose (DDD) of ASM whereas female sex and being married were associated with lower DDD.

Significance

There has been a shift toward newer‐generation ASMs for poststroke seizures and epilepsy. Concerningly, vulnerable patient groups were more likely to be dispensed older‐generation ASMs. This may lead to unnecessary exposure to adverse events and drug–drug interactions. Further research is needed to evaluate comparative effectiveness and safety of newer‐ and older‐generation ASMs in poststroke populations.

Key Points.

• Five percent of people who were hospitalized with an ischemic stroke between 2013 and 2017 received an ASM within 12 months from discharge.

• The most prevalent first ASMs were levetiracetam, valproate, carbamazepine and clonazepam.

• There has been a considerable shift toward newer‐generation ASMs for poststroke seizures and epilepsy between 2013 and 2017.

• People over 85 years, with dementia and psychotic comorbidities were less likely to receive newer‐generation ASMs as a first‐line treatment.

1. INTRODUCTION

Stroke is a major cause of disability and death worldwide. ¹ Compared to the general population, people who experience a stroke are at higher risk of developing epilepsy, ² which is a common and serious neurological disorder characterized by recurrent unprovoked seizures. ³ Cerebrovascular disease is the cause of 50% of newly diagnosed epilepsy in people aged over 60 years. ⁴ Meta‐analyses of observational studies suggest that the pooled incidence of poststroke seizures and epilepsy is 7% and 5%, respectively. ⁵ Although hemorrhagic stroke poses a greater risk than ischemic stroke for the development of poststroke seizures, ⁶ ischemic stroke accounts for 70% of all strokes in Australia. ⁷ Therefore, the burden of seizures and epilepsy among people with ischemic stroke should not be underestimated.

Antiseizure medications (ASMs) are the main treatment option for controlling seizures after stroke. The European Stroke Organization's Guidelines for the Management of Post‐Stroke Seizures and Epilepsy recommend commencing ASM treatment after one unprovoked seizure. ⁸ However, current guidelines do not make specific recommendations regarding the optimal seizure control strategy. ASM treatment needs to be individualized according to factors such as seizure type, demographic factors, comorbidities, and concurrent medications. ⁹ To date, there are no robust data on the optimal management of poststroke seizures and epilepsy. Prescribing patterns of ASMs in people with poststroke epilepsy differ across countries and over time. Early studies in Taiwan ¹⁰ and Sweden ¹¹ reported that carbamazepine and phenytoin were the most prevalent ASM dispensed, while levetiracetam was most prevalent in more recent studies from Sweden, ¹² Italy, ¹³ and Japan. ¹⁴

The objective of this study was to investigate the trends in the utilization of ASMs following ischemic stroke and to examine factors associated with ASM use in people after ischemic stroke in Australia.

2. MATERIALS AND METHODS

2.1. Data sources

We analyzed administrative health data from Victoria, Australia. Victoria has a population of 6.5 million people and is the second most populous Australian state. Records of all public and private hospital admissions for ischemic stroke throughout the study period were identified from the Victorian Admitted Episodes Dataset (VAED). These records were linked to data from the Pharmaceutical Benefits Scheme dataset (PBS), Victorian Emergency Minimum Dataset (VEMD), and National Death Index (NDI). The linkage was performed by the Australian Institute of Health and Welfare (AIHW). Medical diagnoses during admissions in VAED or emergency visits in VEMD were coded using the International Classification of Diseases and Related Health Problems, Tenth Revision, Australian Modification (ICD‐10‐AM) diagnosis codes. ¹⁵ PBS data contained records of prescribed medications dispensed at community pharmacies or at hospital discharge. The majority of Australian medication prescriptions are captured in PBS data under “concessional,” “general,” “under co‐payment,” and “repatriation”. ¹⁶ Dispensed medications were coded using the World Health Organization Anatomical Therapeutic Chemical (ATC) codes and PBS item codes. ¹⁷ Death status was identified using date of death from the NDI.

2.2. Standard protocol approvals, registrations and patient consents

This study was approved by the AIHW Ethics Committee (EO2018/4/468) and the Monash University Human Research Ethics Committee (14339). Patient consent was waived as the data were acquired retrospectively and were made nonidentifiable for the investigators. All authors have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.

2.3. Cohort selection and study design

We conducted a cohort study including people aged ≥30 years with incident ischemic stroke (ICD‐10‐AM I63, I64) and discharged from hospital between July 1, 2013, and June 30, 2017. Children and young adults less than 30 years old were not included in this study due to a small number of stroke cases and different risk profiles. The earliest acute hospital admission for ischemic stroke during this period was considered as the index admission. Those who had history of stroke (ICD‐10‐AM I60, I61, I62, I63, I64, I69), were hospitalized for nonacute stroke care, or who had a history of epilepsy (ICD‐10‐AM G40, G41) in the 5‐year “look‐back” period from the time of their index admission were excluded from the study. Likewise, people with ASM use (ATC N03) within a 1‐year look‐back period were excluded from the study. All individuals were followed up from the date of discharge for 1 year or until death, whichever occurred first, to identify those who received ASM following ischemic stroke. Those who initiated treatment with two or more types of ASM on the same day were excluded from the analyses, as a prescription of multiple ASMs on the same day was considered a surrogate marker of pre‐existing ASM users. A flow chart showing patient selection for this study is shown in Figure S1.

2.4. Antiseizure medications

All ASMs dispensed after stroke were identified using the ATC code N03A. Specific drug codes for ASMs can be found in Table S1. Patients were classified according to the initial ASM they were dispensed. Oxcarbazepine, gabapentin, lamotrigine, levetiracetam, and topiramate were considered “newer‐generation” ASMs whereas carbamazepine, phenytoin, primidone, valproate, and clonazepam were considered “older‐generation” ASMs. ¹⁸ , ¹⁹

2.5. Outcome measures

The outcome measures were the prevalence of different types of ASMs dispensed following acute ischemic stroke each year and the average defined daily dose (DDD) of ASMs dispensed. ²⁰ Prevalence rates of ASMs dispensed after stroke were calculated as the proportion of people receiving at least one dispensing record of ASM within 1 year from the total stroke cohort during each financial year. DDD is a unit of measurement defined by the World Health Organization (WHO), which assumes the average maintenance dose per day for a specific drug when used for its main indication in adults. ²⁰ It allows direct comparison of drug utilization at national and international levels by standardizing the quantity of drugs dispensed to a common parameter. DDDs for each ASM are shown in Table S1. The DDD of ASM dispensed per day per person for any ASM was calculated as below:

$$\mathrm{DDD}\mathrm{per}\mathrm{day}\mathrm{per}\text{person}=\frac{\text{Number of orescriptions}\left(n\right)\times \text{quantity of drugs}\mathrm{per}\text{prescription}\left(q\right)\times \text{drug strength}\left(g\right)}{\mathrm{DDD}\times \text{follow}-\mathrm{up}\text{time}}$$

The average DDD per person was calculated by dividing the sum of DDD per person by the stroke population who received ASM following ischemic stroke.

2.6. Predictor variables

Age group, sex, marital status, insurance status, and geographical area identified on the index admission date were included as potential sociodemographic predictors of ASM use. Marital status was binarized into “married or de facto,” and “single.” Single included people who were widowed, never married, divorced, separated or not stated. Patient type was grouped into public, private and others based on their insurance status. Others included compensable, department of Veterans' Affairs (DVA) and ineligible patients. The geographical region of residence was grouped into metropolitan, regional, and other areas. Financial year was determined by the year of discharge from the index admission for acute ischemic stroke. Comorbidities were identified 5 years prior to the index admission including the admission date of the index admission. ICD‐10‐AM codes for comorbidities are listed in Table S2. The history of cardiovascular disease medication use was identified 1 year prior to the index admission date. Specific ATC codes for cardiovascular medications are listed in Table S3.

2.7. Statistical methods

Descriptive statistics were used to summarize baseline characteristics of the study cohort and patterns of newer‐generation ASMs dispensed after ischemic stroke. Categorical variables were presented as frequencies and percentages while continuous variables were presented as means, standard deviations (SD), medians, and interquartile ranges (IQR). Odds ratios (OR) and 95% confidence intervals (CI) were obtained from the logistic regression models to examine the predictors of receiving newer‐generation ASM versus older‐generation ASM. Beta coefficients (β), standard errors (SE), and P‐values were derived from the generalized linear model to identify factors associated with DDD of ASMs dispensed after ischemic stroke. Logistic regression and generalized linear models were conducted with all predictor variables included in the model. All statistical analyses were performed using SAS® 9.4.

3. RESULTS

3.1. Baseline characteristics

Of the 19 601 people with incident acute ischemic stroke between July 2013 and June 2017, 989 people (5.0%) were dispensed an ASM for the first time within 1 year from the discharge date (ie ASM initiators). Of 989 ASM initiators, the majority were aged ≥60 years (81.1%), males (52.5%), and married or in de facto relationship (58.8%). Approximately 70% resided in metropolitan areas and two‐thirds were public patients with universal healthcare insurance (65.0%). The median length of stay of the index stroke admission was 20.0 days (IQR 5.0–55.0). The most prevalent comorbid conditions among ASM initiators were hypertension (67.4%) followed by atrial fibrillation (30.2%) and diabetes (26.7%). The most common cardiovascular medications used were renin‐angiotensin system inhibitors (70.4%) followed by lipid modifying agents (60.3%) and diuretics (47.8%). Baseline characteristics of the study cohort are shown in Table 1.

TABLE 1.

Baseline characteristics of the study cohort and predictors of initiating newer‐generation ASMs following ischemic stroke.

Variables	Total ASM initiators	Older‐generation ASM	Newer‐generation ASM
N (%)	989 (5.0)	511 (51.7)	478 (48.3)	OR (95% CI)
Baseline characteristics
Stroke type at index admission
Ischemic stroke (ICD‐10‐AM I63)	808 (81.7)	408 (79.8)	400 (83.7)
Unknown (ICD‐10‐AM I64)	181 (18.3)	103 (20.2)	78 (16.3)
Age group
30–59	187 (18.9)	83 (16.2)	104 (21.8)
60–74	301 (30.4)	137 (26.8)	164 (34.3)	0.94 (0.62–1.42)
75–84	270 (27.3)	140 (27.4)	130 (27.2)	0.74 (0.47–1.15)
85+	231 (23.4)	151 (29.5)	80 (16.7)	0.38 (0.23–0.64)
Sex
Male	519 (52.5)	267 (52.3)	252 (52.7)	Ref
Female	470 (47.5)	244 (47.7)	226 (47.3)	1.05 (0.79–1.40)
Financial year
2013/14	245 (24.8)	154 (30.1)	91 (19.0)
2014/15	222 (22.4)	133 (26.0)	89 (18.6)	1.24 (0.84–1.83)
2015/16	257 (26.0)	113 (22.1)	144 (30.1)	2.47 (1.69–3.61)
2016/17	265 (26.8)	111 (21.7)	154 (32.2)	2.82 (1.92–4.16)
Marital status
Married/De facto	582 (58.8)	297 (58.1)	285 (59.6)	0.79 (0.58–1.07)
Single	407 (41.2)	214 (41.9)	193 (40.4)
Region
Metro	700 (70.8)	342 (66.9)	358 (74.9)	1.79 (1.31–2.45)
Rural	268 (27.1)	158 (30.9)	110 (23.0)
Others	21 (2.1)	11 (2.2)	10 (2.1)	1.12 (0.44–2.86)
Patient type
Public	643 (65.0)	336 (65.8)	307 (64.2)	0.97 (0.71–1.32)
Private	289 (29.2)	140 (27.4)	149 (31.2)
Others	57 (5.8)	35 (6.8)	22 (4.6)	1.03 (0.53–2.02)
Length of stay (days)
Mean (SD)	37.4 (51.1)	36.3 (51.6)	38.6 (50.7)
Median (IQR)	20.0 (5.0–55.0)	20.0 (5.0–50.0)	20.0 (5.0–59.0)
Comorbid conditions
Hypertension	667 (67.4)	357 (69.9)	310 (64.9)	0.98 (0.71–1.36)
Atrial fibrillation	299 (30.2)	157 (30.7)	142 (29.7)	1.14 (0.82–1.58)
Diabetes	264 (26.7)	150 (29.4)	114 (23.8)	0.77 (0.55–1.08)
Congestive heart failure	138 (14.0)	86 (16.8)	52 (10.9)	0.71 (0.45–1.13)
Transient ischemic attack	122 (12.3)	68 (13.3)	54 (11.3)	0.91 (0.60–1.39)
Mood disorder	111 (11.2)	59 (11.5)	52 (10.9)	0.87 (0.55–1.37)
Substance misuse	104 (10.5)	60 (11.7)	44 (9.2)	0.68 (0.43–1.08)
Myocardial infarction	102 (10.3)	57 (11.2)	45 (9.4)	0.90 (0.55–1.47)
Osteoarthritis	77 (7.8)	41 (8.0)	36 (7.5)	1.09 (0.65–1.82)
Dementia	74 (7.5)	56 (11.0)	18 (3.8)	0.35 (0.19–0.63)
Anxiety disorder	73 (7.4)	39 (7.6)	34 (7.1)	1.06 (0.62–1.83)
Angina	66 (6.7)	37 (7.2)	29 (6.1)	1.06 (0.59–1.90)
Parkinson's disease	19 (1.9)	>6	<6	0.55 (0.18–1.71)
Traumatic brain injury	18 (1.8)	12 (2.3)	6 (1.3)	0.46 (0.16–1.33)
Psychotic disorder	15 (1.5)	>6	<6	0.29 (0.09–0.97)
Prior medications
CVD medications	783 (79.2)	419 (82.0)	364 (76.2)
Types of CVD medication
Renin‐angiotensin system inhibitors	551 (70.4)	299 (71.4)	252 (69.2)	0.91 (0.65–1.25)
Lipid modifying agent	472 (60.3)	248 (59.2)	224 (61.5)	1.12 (0.81–1.53)
Diuretics	374 (47.8)	207 (49.4)	167 (45.9)	1.02 (0.72–1.43)
Beta blocker	338 (43.2)	176 (42.0)	162 (44.5)	1.17 (0.85–1.62)
Calcium channel blocker	330 (42.1)	174 (41.5)	156 (42.9)	1.11 (0.80–1.52)
Antiplatelet	243 (31.0)	146 (34.8)	97 (26.6)	0.90 (0.63–1.27)
Antiarrhythmics and nitrates	196 (25.0)	112 (26.7)	84 (23.1)	1.02 (0.69–1.51)
Anticoagulant	163 (20.8)	88 (21.0)	75 (20.6)	0.86 (0.58–1.28)
Other antihypertensives	84 (10.7)	48 (11.5)	36 (9.9)	0.85 (0.51–1.41)
Patient follow‐up period
Mean (SD)	323.6 (96.5)	309.8 (109.6)	338.3 (77.6)
Median (IQR)	365.0 (365.0–365.0)	365.0 (327.0–365.0)	365.0 (365.0–365.0)

Abbreviations: aOR, adjusted odds ratio; ASM, antiseizure medication; CI, confidence interval; CVD, cardiovascular disease; IQR, interquartile range; SD, standard deviation.

Note: Marital status: “single” includes widowed, never married, divorced, separated, not stated; Region: “others” includes interstate, overseas, unknown or missing data; Insurance type: “others” includes compensable, DVA and ineligible patients; Other antihypertensive: includes antiadrenergic agents, agents acting on arteriolar smooth muscle and antihypertensives for pulmonary arterial hypertension; Patient follow‐up period: measured from the discharge date of index stroke admission until the completion of 1 year or death, whichever occurred first. aORs and 95% CIs were derived from the logistic regression model adjusting for age, sex, region, marital status, insurance type, comorbidities (hypertension, atrial fibrillation, diabetes, congestive heart failure, transient ischemic attack, mood disorder, substance misuse, myocardial infarction, osteoarthritis, dementia, anxiety disorder, angina, Parkinson's disease, traumatic brain injury, psychotic disorder) and prior CVD medications (antiplatelet, anticoagulant, antiarrhythmics and nitrates, other antihypertensives, diuretics, beta blocker, calcium channel blocker, renin‐angiotensin system agent, lipid modifying agent); references for comorbidities and CVD medication history were set as “No.”

Among the 989 ASM initiators, 51.7% received older‐generation ASM and 48.3% received newer‐generation ASM for their initial treatment. Over half of the older‐generation ASM initiators were aged ≥75 years (56.9%), while less than half of the newer‐generation ASM initiators (43.9%) were aged ≥75 years. A lower proportion of older‐generation ASM initiators (66.9%) lived in metropolitan areas compared to newer‐generation ASM initiators (74.9%). The prevalence of dementia was about three times higher in older‐generation ASM initiators (11.0%) than in newer‐generation ASM initiators (3.8%). There were no differences in the proportion of other baseline characteristics, comorbidities, and cardiovascular medication use between older and newer‐generation ASM initiators (Table 1).

3.2. ASM dispensing patterns and trends

Of people who initiated ASM following ischemic stroke, 75.9% were dispensed ASM within the first 6 months from the discharge date. The median time taken from discharge to first ASM dispensing was 61 days (IQR 7.0–176.0). Levetiracetam accounted for 38.0% of first ASMs dispensed, followed by valproate (25.8%), carbamazepine (10.3%), and clonazepam (10.1%). Most people receiving ASMs used a single ASM (86.9%) during the 1‐year follow‐up period. Dispensing patterns for ASMs among people with ischemic stroke are shown in Table 2.

TABLE 2.

Dispensing patterns for ASMs among people with ischemic stroke.

ASM dispensing	ASM initiators
	(N = 989)
Time of first ASM
Within 1 month	387 (39.1)
Within 2 months	498 (50.4)
Within 3 months	579 (58.5)
Within 6 months	751 (75.9)
Within 12 months	989 (100.0)
Time taken to receive first ASM
Mean (SD)	101.6 (106.5)
Median (IQR)	61.0 (7.0–176.0)
Types of first ASM received following stroke
Levetiracetam	376 (38.0)
Valproate	255 (25.8)
Carbamazepine	102 (10.3)
Clonazepam	100 (10.1)
Phenytoin	49 (5.0)
Gabapentin	44 (4.4)
Topiramate	40 (4.0)
Lamotrigine	17 (1.7)
Primidone	<6
Oxcarbazepine	<6
ASM dispensing patterns
Single type of ASM	859 (86.9)
Multiple types of ASM	130 (13.1)

Abbreviations: ASM, antiseizure medication; IQR, interquartile range; SD, standard deviation.

The proportion of people initiating ASMs after stroke remained relatively stable over time (5.3% in 2013/14 to 5.2% in 2016/17). The proportion of people receiving newer‐generation ASMs increased from 37.1% in 2013/14 to 58.1% in 2016/17 whereas the proportion of people receiving older‐generation ASMs decreased from 62.9% in 2013/14 to 41.9% in 2016/17 (Figure 1). By ASM types, there was an increase in the proportion of people who received levetiracetam (26.9% to 47.2%) or gabapentin (3.7% to 5.3%), while there was a decrease in the proportion of people who received carbamazepine (13.1% to 6.4%), phenytoin (7.8% to 2.6%) or valproate (31.8% to 21.5%) over time (Figure S2). Similar trends were observed across all age groups. The proportion of people who received newer‐generation ASMs was higher than those who received older‐generation ASMs in all age groups except those aged 85 years and over (Figure S3).

FIGURE 1.

Trends in the utilization of ASM after ischemic stroke. ASM, antiseizure medication; DDD, defined daily dose.

There was a small decrease in the average DDD of ASMs received after ischemic stroke during the study period (0.36 DDD per day per person in 2013/14 to 0.34 DDD per day per person in 2016/17). The average DDD among those who initiated on older‐generation ASM decreased from 0.40 per day, per person in 2013/14 to 0.28 per day, per person in 2016/17. The average DDD among those who initiated on newer‐generation ASM increased from 0.31 per day, per person in 2013/14, peaked to 0.42 in 2014/15, and slightly decreased to 0.38 per day, per person in 2016/17. Trends in the utilization of ASM following ischemic stroke between 2013/14 and 2016/17 are shown in Figure 1. Among people aged between 30 and 59 years, the average DDD of older‐generation ASM increased from 0.39 per day, per person in 2013/14 to 0.48 per day, per person in 2016/17. Trends in the use of older‐ vs. newer‐generation ASMs by age group is shown in Figure S3.

3.3. Correlates of daily ASM consumption

In 2016/17, people had lower DDD of ASM per day (β −0.19; P‐value 0.049) compared to 2013/14. The DDD of ASM per day was significantly higher among people aged 75–84 years (β 0.23; P‐value 0.030) compared to people aged 30–59 years after adjusting for other factors. Females (β −0.18; P‐value 0.008) had lower DDD of ASM per day compared to males and people who were married or in a de facto relationship (β −0.21; P‐value 0.004) had lower DDD of ASM than those who were single. People who received multiple types of ASMs (β 0.78; P‐value <0.001) during the study period had higher DDD of ASM per day compared to those who received single type of ASM. People who received newer‐generation ASM (β 0.23; P‐value 0.001) as their first treatment had higher DDD per day compared to those who received older‐generation ASM. Moreover, people who received beta blockers had higher DDD of ASM per day (β 0.17; P‐value 0.030). Tables 3 shows correlates of ASM consumption following ischemic stroke.

TABLE 3.

Correlates of quantity of ASM use following ischemic stroke, measured using defined daily dose.

Predictors	Beta coefficient	SE	P‐value
Age group
30–59	Ref
60–74	0.18	0.10	0.071
75–84	0.23	0.11	0.030
85+	0.06	0.12	0.622
Sex
Male	Ref
Female	−0.18	0.07	0.008
Financial year
2013/14	Ref
2014/15	−0.04	0.09	0.704
2015/16	0.01	0.09	0.880
2016/17	−0.19	0.09	0.049
Marital status
Married/De facto	−0.21	0.07	0.004
Unmarried	Ref
Region
Metro	−0.07	0.07	0.365
Rural	Ref
Others	−0.13	0.23	0.568
Patient type
Public	−0.07	0.07	0.355
Private	Ref
Others	−0.14	0.16	0.389
Hypertension
Yes	−0.09	0.08	0.217
No	Ref
Atrial fibrillation
Yes	0.03	0.08	0.665
No	Ref
Diabetes
Yes	0.07	0.08	0.355
No	Ref
Congestive heart failure
Yes	−0.08	0.11	0.447
No	Ref
Transient ischemic attack
Yes	0.00	0.10	0.993
No	Ref
Mood disorder
Yes	0.09	0.11	0.423
No	Ref
Substance misuse
Yes	0.15	0.11	0.181
No	Ref
Myocardial infarction
Yes	−0.05	0.12	0.692
No	Ref
Osteoarthritis
Yes	−0.08	0.13	0.513
No	Ref
Dementia
Yes	−0.06	0.13	0.649
No	Ref
Anxiety disorder
Yes	−0.10	0.13	0.436
No	Ref
Angina
Yes	0.03	0.15	0.828
No	Ref
Parkinson's disease
Yes	0.11	0.24	0.634
No	Ref
Traumatic brain injury
Yes	0.19	0.25	0.441
No	Ref
Psychotic disorder
Yes	0.23	0.27	0.384
No	Ref
Renin‐angiotensin system agent
Yes	0.09	0.08	0.249
No	Ref
Lipid modifying agent
Yes	−0.11	0.08	0.147
No	Ref
Diuretics
Yes	0.02	0.08	0.809
No	Ref
Beta blocker
Yes	0.17	0.08	0.030
No	Ref
Calcium channel blocker
Yes	0.11	0.08	0.158
No	Ref
Antiplatelet
Yes	0.02	0.08	0.859
No	Ref
Antiarrhythmics and nitrates
Yes	−0.19	0.10	0.053
No	Ref
Anticoagulant
Yes	−0.10	0.10	0.307
No	Ref
Other antihypertensives
Yes	−0.18	0.12	0.140
No	Ref
ASM dispensing patterns
Single ASM	Ref
Multiple ASMs	0.78	0.10	<0.001
Type of first ASM
Older‐generation ASM	Ref
Newer‐generation ASM	0.23	0.07	0.001

Abbreviations: ASM, antiseizure medication; SE, standard error.

Note: Beta coefficients, standard errors and P‐values were derived from the generalized linear model adjusting for age, sex, region, marital status, insurance type, comorbidities (hypertension, atrial fibrillation, diabetes, congestive heart failure, transient ischemic attack, mood disorder, substance misuse, myocardial infarction, osteoarthritis, dementia, anxiety disorder, angina, Parkinson's disease, traumatic brain injury, psychotic disorder), prior CVD medications (antiplatelet, anticoagulant, antiarrhythmics and nitrates, other antihypertensives, diuretics, beta blocker, calcium channel blocker, renin‐angiotensin system agent, lipid modifying agent), ASM dispensing patterns and type of first ASM; references for comorbidities and CVD medication history were set as “No.”

3.4. Predictors of newer versus older‐generation ASMs

After adjusting for other factors, people were more likely to receive newer‐generation ASM for their first treatment in 2015/16 (OR 2.47; 95% CI 1.69–3.61) or 2016/17 (OR 2.82; 95% CI 1.92–4.16) compared to 2013/14. People aged ≥85 years (OR 0.38; 95% CI 0.23–0.64) were less likely to receive newer‐generation ASM for their first treatment compared to people aged 30–59 years. People residing in metropolitan areas (OR 1.79; 95% CI 1.31–2.45) were more likely to receive newer‐generation ASM for their first treatment than those living in rural areas. Moreover, those with comorbidities such as dementia (OR 0.35; 95% CI 0.19–0.63) and psychotic disorders (OR 0.29; 95% CI 0.09–0.97) were less likely to receive newer‐generation ASM for their first treatment compared to people without those conditions (Table 1).

4. DISCUSSION

Our retrospective data linkage study of ASM use following ischemic stroke noted several important findings. Firstly, there was a significant shift over time toward the use of newer‐generation ASMs, and people living in metropolitan reasons were significantly more likely to be dispensed these. Conversely, people who were older and those with dementia or psychotic comorbidities were less likely to be dispensed newer‐generation ASM. Factors associated with higher ASM consumption were older age, history of beta blocker use, multiple ASMs, and newer‐generation ASM. Meanwhile, female sex and being in a marital or nonmarital relationship were associated with lower ASM consumption. These observations establish an important evidence‐base going forward for the optimal treatment of poststroke epilepsy.

Throughout the study period, levetiracetam was the most frequently dispensed ASM following ischemic stroke. After adjusting for other factors, patients discharged in 2016/17 were significantly more likely to receive newer‐generation ASMs compared to those discharged in 2013/14. This change may represent a preference toward ASMs with better safety profiles and fewer potential drug–drug interactions. A retrospective population study conducted in Australia between 2002 and 2007 reported that valproate was the most frequently dispensed ASM regardless of the indication followed by carbamazepine and phenytoin. ²¹ However, the proportion of these older‐generation ASMs decreased while newer‐generation ASMs such as levetiracetam (0.4% to 4.6%) and lamotrigine (7.2% to 10.7%) increased progressively during the study period. ²¹ In recent studies from Sweden ¹² (2012–2016), Italy ¹³ (2013–2018) and Japan ¹⁴ (2014–2019), levetiracetam has replaced older‐generation ASMs, and it became the most frequently used ASM for poststroke epilepsy. This transition was also evident in observational cohort studies conducted in England ²² (2003–2016) and Brazil ²³ (2009–2019), demonstrating an increased use of newer‐generation ASMs over time among people with epilepsy. This trend could be attributed to the improved safety and tolerability of newer‐generation ASMs. ²⁴

We demonstrated that people residing in metropolitan areas were significantly more likely to be initiated on newer‐generation ASMs compared to people living in regional areas. Treatment disparities across different geographical regions have been reported in Germany ²⁵ (2010–2012) and Taiwan ²⁶ (2013–2016) among people with epilepsy. Both studies observed that newer‐generation ASMs were more likely to be prescribed in metropolitan areas compared to regional or rural areas. ²⁵ , ²⁶ This might be partially explained by higher socioeconomic advantage in metropolitan areas. Previous studies reported that people in rural areas or of lower socioeconomic status were more likely to be prescribed older‐generation ASMs. ²⁷ , ²⁸ Another factor that might have impacted the choice of ASM was the density of medical specialists and physician preference. Studies conducted in Sweden ²⁷ and the United States ²⁹ found that patients were more likely to receive newer‐generation ASMs from neurologists compared to other specialists including general practitioners. The uneven distribution of neurologists between metropolitan and regional Victoria could have accounted for different treatment preferences. Since there are no Australian clinical practice guidelines on poststroke seizures, physicians may prefer to choose ASMs with long history of established safety profiles.

We also observed older patients and those with dementia and psychotic comorbidities were less likely to receive newer‐generation ASMs, despite this group perhaps being more susceptible to adverse drug events associated with older‐generation ASMs. Similar results were reported in Sweden ²⁷ (2006) and Taiwan. ²⁶ Older people with epilepsy in Sweden were more likely to be prescribed carbamazepine than levetiracetam or lamotrigine. ²⁷ Furthermore, comorbidities such as stroke, dementia, chronic obstructive pulmonary disease, and neuroses were associated with receiving older‐generation ASMs for initial treatment. ²⁶ Despite declining use, 42.9% of older people in the United States received older‐generation ASMs between 2004 and 2015. ³⁰ This may be due to the potential effect of ASMs on the existing comorbidities that are more common in older people. For example, carbamazepine and valproate are also prescribed as mood stabilizers in people with dementia. ³¹ , ³² Another possible explanation for older patients receiving older‐generation ASMs is that these agents may have been effective or tolerated for a patient in the past, and therefore, physicians may have decided to prescribe a ASM known to that person.

Older age was associated with increasing daily ASM consumption. Similar results were observed in a previous Australian study on ASM dispensing patterns by age and sex. ²¹ This result may partially reflect failure in seizure control among older patients, leading to a dose increase, switching, or combination therapy. ⁹ , ³³ Older people tend to be on multiple medications for various health conditions and some medications may induce the metabolism of ASMs, resulting in a dosage increase. ³⁴ Though older patients are recommended to be treated with ASMs at a lower dosage and increase gradually, DDD calculation used in our study did not reflect this recommendation. ⁹ , ³⁵ Another factor that increased ASM consumption was the history of beta blocker use. A possible explanation is that beta blocker prescription is a surrogate marker for cardiovascular disease. People with pre‐existing cardiovascular disease may experience more severe or more frequent strokes, which contributes to increased seizure risk. ³⁶ Further exploration is warranted to investigate the association between cardiovascular medication use and ASM consumption.

We showed that females had lower ASM consumption regardless of age. This finding is in line with studies conducted in Australian general population. ²¹ A cross‐sectional study in Taiwan (2016) reported lower mean prescribed daily dose among females with epilepsy. ³⁷ This could be explained by the lower body weights of females and potentially lower prevalence of poststroke seizures in females. ³⁸ Also, being in a marital or de facto relationship was associated with lower ASM consumption. A cross‐sectional Ethiopian study showed having a married caregiver contributed to better adherence to ASMs. ³⁹ This suggests family support may improve ASM adherence and allow patients to remain seizure‐free on lower doses.

4.1. Strengths and limitations

One of the main strengths of this study is that it represents the entire stroke population in Victoria and captures all subsidized ASMs dispensed in community pharmacies and at hospital discharge from 2013 to 2017. This is the first Australian population‐based study on patterns and predictors of ASM consumption after stroke. Nevertheless, our study population focused on people over 30 years of age as our data sources did not provide information on children and young adults. Overall, this likely represents very few missed cases as stroke incidence increase with age. The most recent data from the Australian Institute of Health and Welfare revealed that 71% of people who had a stroke were aged 65 and over, ⁴⁰ and a Victorian‐based epidemiology study identified only 3 incident ischemic stroke cases in people aged 24–35 years over a 12‐month period, and no cases in younger people in that same period. ⁴¹ Also, our data did not include records of ASMs that are not subsidized via the PBS, and we did not have access to information on specific indications. It was possible that ASMs were prescribed for other neurological conditions such as migraine (eg topiramate) and neuropathic pain (eg gabapentin). The WHO DDDs are assigned based on the adult dose for the main indication and may not represent the recommended daily dose for older people or for other indications. Not all dispensed medications are consumed as directed by patients. We did not have data on stroke severity, which is a risk factor for developing poststroke seizures.

5. CONCLUSION

Our study found a shift toward the use of newer‐generation ASMs between 2013 and 2017. However, older age, dementia, psychotic comorbidities, and regional residence were negatively associated with receiving newer‐generation ASMs. Older age, history of beta blocker use, multiple ASMs, and newer‐generation ASM was associated with higher daily consumption of ASM. On the contrary, female sex and being married was associated with lower daily consumption. More study is needed to examine the safety and effectiveness of widely used newer‐generation ASMs among patients with poststroke seizures and epilepsy.

FUNDING INFORMATION

No funding was received for this study.

CONFLICT OF INTEREST STATEMENT

The authors report no disclosures relevant to this study. For other disclosures, EF is supported by the Monash Partners STAR Clinician Fellowship, and The Royal Australian College of Physicians Fellows Research Establishment Fellowship. EF reports grants from Brain Foundation Australia, LivaNova, Lundbeck Australia, and Sylvia and Charles Viertel Charitable Foundation. This research funding is unrelated to this study. JSB is supported by a National Health and Medical Research Council (NHMRC) Boosting Dementia Research Leadership Fellowship and has received funding or consulting funds from the NHMRC, Medical Research Future Fund (MRFF), Victorian Government Department of Health and Human Services, Dementia Australia Research Foundation (DARF), Yulgilbar Foundation, Aged Care Quality and Safety Commission, Dementia Centre for Research Collaboration, Pharmaceutical Society of Australia, Society of Hospital Pharmacists of Australia, GlaxoSmithKline Supported Studies Programme, Amgen, and several aged care provider organizations unrelated to this work. All grants and consulting funds were paid to the employing institution. JI has received funding from NHMRC, MRFF, National Breast Cancer Foundation, DARF, Yulgilbar Foundation, Channel 7 children's research foundation, Amgen and AstraZeneca, unrelated to this work.

ETHICS STATEMENT

This study was approved by the AIHW Ethics Committee (EO2018/4/468) and the Monash University Human Research Ethics Committee (14339). Patient consent was waived as the data were acquired retrospectively and were made nonidentifiable for the investigators. All authors have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.

Supporting information

Appendix S1:

Kim SJ, Wood S, Marquina C, Foster E, Bell JS, Ilomäki J. Shift from older‐ to newer‐generation antiseizure medications in people with acute ischemic stroke in Australia: A population‐based study. Epilepsia Open. 2023;8:1413–1424. 10.1002/epi4.12809

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are not publicly available due to privacy agreements between the researchers and linkage authorities. Data can be requested via an application process from Centre for Victorian Data Linkage and AIHW.