Dementia,Subtype of Seizures,and the Risk of New Onset Seizures: A Cohort Study

Abstract

Background:

Seizure disorders have been identified in patients suffering from different types of dementia. However, the risks associated with the seizure subtypes have not been characterized.

Objective:

To compare the occurrence and risk of various seizure subtypes (focal and generalized) between patients with and without a dementia diagnosis.

Methods:

Data from 40.7 million private insured patient individual electronic health records from the U.S., were utilized. Patients 60 years of age or more from the Optum Insight Clinformatics-data Mart database were included in this study. Using ICD-9 diagnoses, the occurrence of generalized or focal seizure disorders was identified. The risk of new-onset seizures and the types of seizures associated with a dementia diagnosis were estimated in a cohort of 2,885,336 patients followed from 2005 to 2014. Group differences were analyzed using continuity-adjusted chi-square and hazard ratios with 95%confidence intervals calculated after a logistic regression analysis

Results:

A total of 79,561 patient records had a dementia diagnosis, and 56.38%of them were females. Patients with dementia when compared to those without dementia had higher risk for seizure disorders [Hazard ratio (HR) = 6.5 95%CI = 4.4–9.5]; grand mal status (HR = 6.5, 95%CI = 5.7–7.3); focal seizures (HR = 6.0, 95%CI = 5.5–6.6); motor simple focal status (HR = 5.6, 95%CI = 3.5–9.0); epilepsy (HR = 5.0, 95%CI = 4.8–5.2); generalized convulsive epilepsy (HR = 4.8, 95%CI = 4.5–5.0); localization-related epilepsy (HR = 4.5, 95%CI = 4.1–4.9); focal status (HR = 4.2, 95%CI = 2.9–6.1); and fits convulsions (HR = 3.5, 95%CI = 3.4–3.6).

Conclusion:

The study confirms that patients with dementia have higher risks of generalized or focal seizure than patients without dementia.

Keywords

Databases dementia epilepsy new-onset seizures

INTRODUCTION

Dementia is typically defined as a clinical syndrome of cognitive decline that is sufficiently severe to interfere with social or occupational functioning [1]. The Diagnostic and Statistical Manual of Mental Disorders, 5th edition (DSM-5) replaces the term “dementia” with major neurocognitive disorder and mild neurocognitive disorder [2], and considers them as acquired, permanent, neurocognitive syndromes characterized by short and long-term memory deficits, associated with abstract thinking problems. In contrast, the Alzheimer’s Association outlines different diagnostic criteria for Alzheimer’s disease and retains the use of the word dementia. Dementia is a universal public health problem. Moreover, its incidence is related to the widespread aging of the world’s population. Estimates show that around to 6%of people over 65 years of age have some form of dementia, with a prevalence of 8–10%in the general population of developed countries [3 –5].

Studies in elderly with dementia older than 55 years of age show that close to 9%of them experience at least one seizure during their lifetime [6, 7]. The risk of unprovoked seizure in elderly patient increases six times with a diagnosis of Alzheimer’s disease dementia and eight times in case of non-Alzheimer dementia [8]. Among patients older than 65 years with dementia, 3.6%experienced seizures recurrences, i.e., develop epilepsy [9].

While the increased risk of seizures and epilepsy is well established in patients with dementia, a recent view found that the prevalence of various seizure types is very scarce [10]. Such knowledge is critical for managing any patient population with epilepsy. However, its particularly important in elderly with dementia as some seizure types like focal onset with impaired awareness can be mistaken for confusional episodes in this population [10]. Increased risk of generalized convulsive seizures, which account for more than 80%of seizure-related injuries [11], could have more grievous consequences in this patient population given the increased risk of osteoporosis in the elderly, especially the ones on anti-epileptic drugs [12]. Therefore, the objective of this study was not only to assess the frequency risk, but also to try to elucidate the seizure types that affect patients with dementia after controlling for traditional risk factors for new onset seizure (NOS) in a population sample representative of the United States (U.S.) general population.

MATERIALS AND METHODS

This study used data from the Optum Insight Clinformatics-data Mart (OICM) database. All patients were part of the OICM database that represents privately insured populations and capture administrative claims with patient-level de-identified data from in and outpatient visits and pharmacy claims of multiple insurance plans, primarily The United Health Group. The use of data from databases has been previously documented as valid information source [13]. Valuable results of dementia studies have been published by different authors, including the present ones, using this same approach [14]. The OICM database is a fully de-identified health information, and protection accountability act (HIPAA) compliant database. OICM contains claims from individuals covered by 69 self-insured companies with locations in all census areas of the U.S. Data are available for all company beneficiaries, including employees, spouses, dependents, and retirees. Data has been included in this database for more than 40 million managed-care lives. The average duration of observation for this database is 2.7 years. Since the information is completely de-identified at the source, Institutional review board (IRB) approval and informed consent were not required for this study.

The study population was the group of patients with at most one International Classification of Diseases, 9th version clinical modification (ICD-9-CM) codes [15] for dementia, through the years 2005–2014 from the OICM database. To identify disease diagnoses, Thomson Medstat disease staging coding criteria was used. This method is based on electronic screening and identification of a comprehensive map of ICD-9-CM codes and has been widely used as a classification system for diagnostic categories [16]. Using these coding criteria, all potential dementia types from the population selected were identified for the inpatient and outpatient claims.

The study population consisted of patients of 60 years or older diagnosed with any type of dementia, and a comparison group of patients of similar age group but no dementia diagnosis. Non-dementia patients were identified in the database during the same years as the patients with dementia. They were matched by age (at the time of entrance in the database) and gender. Selecting both groups from the same population source and during the same period of time avoids the problem of no comparability of patients negative for the expected outcome, at the beginning of cohort studies [17]. To be included, patients had to be present in the OICM database for at least 12 continuous months before their first dementia diagnosis. To avoid misdiagnosis of dementia, patients needed to have the same dementia diagnosis during the study’s duration. Otherwise, they were excluded from the study. For example, if patient received first a diagnosis of undifferentiated dementia and subsequently during the study period, the diagnosis was changed; for instance, to vascular dementia, this patient was excluded from the study. In contrast, patients could have different seizure types. Patients who were taking any type of anti-seizure medications (ASMs) or had diagnosis of any seizure, seizure disorder, or epilepsy during the year prior to the cohort entry date were excluded. The follow up of the patient’s cohort started the date of their dementia/non-dementia determination. Irrespective of the absence or presence of new-onset seizures, patients were followed up until December 31, 2014, or to the date of their last record in the database. A NOS was defined as the new presence of any type of ICD-9-CM code for seizure, seizure disorder, or epilepsy diagnosis. All of the ICD-9-CM codes used during this study to either classify a dementia, or a seizure disorder are shown in detail in Table 1. As noted in Table 1, although the ICD codes for seizure used terms ‘Seizure Partial’ and ‘Seizure Undifferentiated’, for the purpose of this manuscript, we will be using the ILAE 2017 seizure classification terms of ‘focal onset seizure’ and ‘unclassified’ seizures, respectively [18]. Similarly, ‘Epilepsy Partial’ is replaced by ‘Focal Epilepsy’ in the text, according to the ILAE classification system [19]. Patients with both focal and generalized seizures or epilepsy were considered ‘unclassified’ NOS in our analysis. The validity of ICD-9-CM and ICD-10 diagnoses in administrative data recording clinical conditions has been assessed and documented using unique dually coded databases [20].

Table 1

Dementia and seizure disorder classifications according to ICD-9-CM^a code and description

Classification	Code	Description
Alzheimer’s dementia, late onset, uncomplicated	290.0	Senile dementia, uncomplicated
	290.1	Presenile dementia
Alzheimer’s dementia, early onset, uncomplicated	290.10	Presenile dementia, uncomplicated
Alzheimer’s dementia, early onset, with delirium	290.11	Presenile dementia with delirium
Alzheimer’s dementia, early onset, with delusions	290.12	Presenile dementia with delusional features
Alzheimer’s dementia, early onset, with depressed mood	290.13	Presenile dementia with depressive features
Alzheimer’s dementia, late onset, with delusions	290.2	Senile dementia with delusional or depressive features
	290.20	Dementia of the Alzheimer’s, late onset, with delusions
Alzheimer’s type dementia, late onset, with depressed mood	290.21	Senile dementia with depressive features
Alzheimer’s type dementia, late onset, with delirium	290.3	Senile dementia with delirium
Vascular dementia, uncomplicated	290.4	Arteriosclerotic dementia.
	290.40	Arteriosclerotic dementia, uncomplicated
Vascular dementia, with delirium	290.41	Arteriosclerotic dementia with delirium
Vascular dementia, with delusions	290.42	Arteriosclerotic dementia with delusional features
Vascular dementia, with depressed mood	290.43	Arteriosclerotic dementia with depressive features
Dementia due to general medical condition	294.1	Dementia in conditions classified elsewhere
Amnestic disorder NoS^b (dementia NoS)	294.8	Other specified organic brain syndromes (chronic)
Dementia due to HIV^c disease (cognitive disorder NoS)	294.9	Unspecified organic brain syndrome (chronic)
Alzheimer’s disease	331.0	Alzheimer’s disease
Pick’s disease	331.1	Pick’s disease
Parkinson’s disease	332.0	Parkinson’s disease
Seizures Partial	780.34	With consciousness impairment
	780.35	With no consciousness impairment
Seizures Generalized	780.30	Generalized non-convulsive convulsion
	780.31	Generalized tonic-clonic convulsion
Seizures Undifferentiated	780.3	Undifferentiated convulsion
Epilepsy Partial	345.5	With no consciousness impairment
	345.6	With consciousness impairment
Epilepsy Generalized	345.0	Generalized non-convulsive epilepsy
	345.1	Generalized tonic-clonic epilepsy
Epilepsy Undifferentiated	345	Undifferentiated epilepsy

^aInternational classification of diseases 9th version, clinical modification; ^bNo otherwise specified; ^cHuman immunodeficiency virus.

Statistical analysis

Differences between proportions were analyzed using the continuity-adjusted chi-square (χ²) statistics. We compared proportions of dementia and non-dementia patients with NOS disorder subtypes. The data analysis software, International Business Machines Statistical Package for Social Sciences, version 21 (IBM-SPSS-21®) was used to determine Cox proportional hazard mixed regression models. This method was used to assess the risk of NOS adjusting for variables that have been identified as being able to induce seizure disorders in this age group [21]. These included age at entry into the cohort as a continuous variable, and comorbid medical diagnostic categories (e.g., neurological or psychiatric conditions, etc.) and the use of medications including acetylcholinesterase inhibitors, antipsychotics, and antidepressants, as binary (yes/no) variables. The statistical significance accepted for all the differences calculations was p≤0.05.

RESULTS

Patients’ demographics

During the years 2005 to 2014, a total of 2,885,336 subjects 60 years or older were identified. The study population age ranged from 60 to 93 years. Among the total 79,561 (2.76%) had a diagnosis of dementia, of which 44,857 (56.38%) were women. As seen in Table 2, the distribution of gender, as well as the age groups (60 to 69, 70 to 79 and ≥80) between the two study cohorts (dementia versus without dementia) were significantly different (p≤0.05).Table 2 also shows the inter-cohort comparison of medical comorbidities, specified either by organ system or disease type. The only exceptions to this rule were infections, hepatic and immunological problems that are equally distributed between the two study groups.

Table 2

Demographic characteristics and medical comorbidities^a in patients with/without Dementia

Demographics		Dementia (%) (N = 79,561)	Without Dementia (%) (N = 2,805,775)	Statistics^c	p ^d

Sex (Female)		56.38	49.97	1271.6	< 0.05^*
		20.27	72.98	105544.4	< 0.05^*
Age	60–69 y	24.51	17.42	2679.8	< 0.05^*
	70–79 y	55.82	29.58	25222.0	< 0.05^*
	≥80 y
Medical comorbidities
Neurological		46.98	2.61	51.6	< 0.05^*
Gastrointestinal		56.42	37.60	6.5	< 0.05^*
Musculoskeletal		72.21	39.46	22.0	< 0.05^*
Hepatic		6.27	2.60	1.0	0.30
Endocrine		32.54	13.68	10.0	< 0.05^*
Cancer		12.89	3.66	5.2	< 0.05^*
Infection		52.25	38.52	3.4	0.067
Respiratory		61.15	45.42	5.1	< 0.05^*
Hematological		27.66	11.91	8.0	< 0.05^*
Cardiovascular		61.03	23.14	29.6	< 0.05^*
Metabolic		52.73	24.72	16.5	< 0.05^*
Dermatologic		54.75	35.10	8.1	< 0.05^*
Genital and urinary		42.71	23.61	8.1	< 0.05^*
Allergic		41.66	25.19	6.5	< 0.05^*
Immune		29.08	18.88	2.7	0.09
Renal		16.66	4.36	8.9	< 0.05^*
Psychiatric		35.71	0.22	40.6	< 0.05^*

^aComorbidity was classified by system-organ class. At least one diagnosis for each category; ^bPercentage; ^cChi-square; ^dSignificance: ^*p≤0.05; ^**p≤0.0001.

As expected, regarding the use of medications, patients with dementia had significant higher percentages of use of acetylcholinesterase inhibitors (16.33 versus 0.19; p ≤0.05), antidepressants (48.97 versus 17.22; p ≤0.05) and antipsychotics (38.64 versus 19.81; p≤0.05).

Seizure subtypes

The overall incidence of NOS in patients with dementia was 12.34%per year compared to 2.21%in the non-dementia cohort. Compared to patients without dementia, focal, generalized, or undifferentiated seizures were more frequently in patients with dementia (p≤0.0001) (Table 3).

Table 3

Percentage (%) difference of new onset seizure (NOS) disorders subtypes between patients with and without dementia

ICD-9-CM^a description			With dementia versus Without dementia
			%Difference	Chi-Square 95%CI^b	p ^c
SEIZURES	Focal	(780.34) (780.35)	1.1	3813.2 (1.0 to 1.2)	0.0001^**
	Generalized	(780.30) (780.31)	2.2	5995.2 (2.1 to 2.4)	0.0001^**
	Unclassified	(780.3)	4.5	17369.3 (4.4 to 4.7)	0.0001^**
EPILEPSY	Focal	(345.5) (345.6)	0.4	1135.1 (0.4 to 0.5)	0.0001^**
	Generalized	(345.0) (345.1)	1.5	4242.3 (1.4 to 1.6)	0.0001^**
	Unknown	(345)	0.5	1458.8 (0.5 to 0.6)	0.0001^**

^aInternational classification of diseases 9th version, clinical modification; ^bConfidence interval; ^cStatistical significance: ^*p≤0.05; ^**p≤0.0001.

The frequency of focal, generalized, and unclassified seizures was progressively incremental in both groups of patients with and without dementia. However, the frequencies of generalized and undifferentiated epilepsies are inverted to the mentioned progression, being greater the generalized in patients with and without dementia, compared to the undifferentiated. Figure 1 illustrates the distribution of subgroups of seizures/epilepsy in patients with and without dementia. The Hazard ratios (HR) of each seizure and epilepsy type for the dementia cohort are shown in Table 4.

Fig. 1

Frequency (%) of subtypes of new onset seizures (NOS) in patients with and without dementia.

Table 4

Risk of new onset seizures (NOS) of a Dementia diagnosis

New onset seizure disorder type		Risk HR^a [CI 95%]^b
Seizures	Focal	6.0 [5.5 –6.6]
	Generalized	4.8 [4.5 –5.0]
	Unclassified	6.5 [4.4 –9.5]
Epilepsy	Focal	4.7 [4.4 –5.1]
	Generalized	5.2 [4.9 –5.5]
	Unknown	5.0 [4.8 –5.2]

^aHazard Ratio; ^b95%confidence interval.

DISCUSSION

To our knowledge, this is the first report comparing subtypes of NOS in patients of 60 years and older with and without dementia. The large sample size with ten years of follow up period, allowed us to confirm the finding of increased prevalence of NOS as reported in literature [7, 8]. We found the prevalence of NOS in more than 12%of individuals 60 years or older with dementia, which was significantly higher compared to the non-dementia cohort (2.21%). The unique finding of our study is the seizure type for NOS in the dementia cohort. As noted in Fig. 1, the most common NOS type was unclassified, with the dementia cohort being 6.5 times [HR = 6.5, 95%confidence interval (CI) = 5.5–6.6] more likely to have such seizure type and 5.0 times [HR = 5.0, 95%CI = 4.8–5.2] times more likely to have such epilepsy type compared to non-dementia cohort. Interestingly, studies from various case series of older adults and elderly show that seizures are commonly focal in origin [22] in this population. Similarly, in a small case series, close to three-fourth of dementia, patients with epilepsy suffered focal seizures [9]. In contrast, our data found that NOS in the dementia cohort were classified as generalized seizure or epilepsy in 2.83%and 1.81%, respectively, which was remarkably higher than focal seizures and epilepsy. The most likely explanation for this discrepant finding is that the ICD codes used by the care provider meant to reflect generalized tonic-clonic seizures, which is the most likelihood in this patient cohort were of focal to bilateral tonic-clonic type [18]. Regardless of the epilepsy type leading to NOS, the fact that patients with dementia seem to be suffering generalized seizures is worrisome as these seizures are more likely to predispose to physical injuries [11].

Our study did not investigate various dementia types. Other studies have found the incidence of having at least one new-onset seizure in patients with mild to moderate Alzheimer’s disease of 0.48 per 1,000 person-year [23]. This observation repeated for seizures/epilepsy in another group of Alzheimer’s disease patients from the nationwide inpatient sample, especially after correcting for age, but not for gender or race [24]. Lastly, another study found that patients with vascular dementia (n = 4,438) had an incidence rate (IR) of seizures of 7.5 per 1,000 persons-year (py), higher than (n = 7,086) Alzheimer’s disease patients with an IR = 5.6/1,000 py, and also higher than (n = 11,524) not demented subjects with an IR = 0.8/1,000 py [25].

In our study, we found a significantly high percentage of dementia population was taking psychotropic medications. Some medications frequently used in dementia patients, such as olanzapine, quetiapine, donepezil, ginkgo biloba [26 –29], or the simultaneous combination of anti-muscarinic and acetylcholinesterase inhibitors [30] can also produce seizure disorders in elder. The increased frequency of NOS in the dementia cohort remained significant after adjusting for these medications. However, it may be prudent to be extra cautious in the use of these medications in older dementia populations. If alternatives available, medications that do not affect seizure risk should be preferred.

Cohort studies are supposed to be subjected to bias and attrition [31]. However, in this case, by dividing the patients in with and without dementia groups at the beginning, and by selecting them from the same pool of patients, we avoided selection bias. Also, because all the information from the patients was consigned at the starting time directly into the database, recollection bias was controlled. Thus, probably the only factor we need to be careful during the analysis was attrition.

There are several limitations to our study. Medical claims data contain information on the interaction of the patient with the medical system (office, clinic, outpatient, and inpatient hospital, etc.) that generate a claim to the insurer. In order to verify the diagnosis of NOS, clinical review of administrative claims profiles was conducted to ensure that the claim’s series was consistent with the diagnosis determined by the claims algorithm which does present limitations, especially with regards to the accuracy of the study diagnoses. On the other hand, inaccuracies in coding, processing errors, or coding of “rule-out” diagnoses may lead to misclassification of the outcome, albeit likely equally in both cohorts. Additional data sources are needed to improve the sensitivity and accuracy of identifying comorbidity in patients with dementia. While we looked at several commonly used medication groups in the elderly, we did not have data on the use of benzodiazepines, which is likely used more often in patients with seizure or epilepsy and alternatively, can potentially mask the expression of seizures.

This type of observational data cannot give information about healthcare services if the patient sought care outside of the plan’s coverage or if the patient chose to pay for the services out-of-pocket. Another limitation of insurance claims data is that the recorded diagnoses may represent either suspected conditions being ruled out or confirmed conditions being treated. The former is more likely to lead to misdiagnosis of the patient to dementia cohort. In addition, important risk factors for seizure events cannot be captured through health claims data, such as a family history of seizure disorders. Nonetheless, the validity and reliability [32] of the information in administrative claims data with respect to seizures are largely unknown, and further research is warranted in this area. Another major limitation of our study is that an overwhelming majority of people who are 65 years and older are on Medicare. While 20 million Medicare beneficiaries had private healthcare insurance as well in 2016 [33], our study population may not be a true representative of the general population. Future studies are required to analyze seizure and epilepsy risk in Medicare and uninsured elderly patients and compare the difference in incidence of these conditions between differently insured patient population. This is especially important because chronically uninsured patients lack regular medical care, which increases the risk of untreated cardiovascular risk factor predisposing them to increased vascular dementia risk. Due to de-identification of patient identifiers, we lacked data on other demographical factors. The lack of data on social determinants of health, which can critically contribute to modifiable risks of dementia development were not available in the database. Additionally, future studies should utilize the advanced machine learning analytical tools like neural networks, and random forest regression to analyze the big administrative databases to help uncover hitherto unknown facets about seizure and epilepsy in dementia population.

In summary, our study uses a large population-based database of older adults and elderly with and without dementia to confirm the increased risk of seizure in this population. More importantly, we report the various seizures types encountered in the dementia population. Our findings raise concern that a small, but substantial, the population of dementia patients could be having a generalized, convulsive seizure of focal origin.

DISCLOSURE STATEMENT

Authors’ disclosures available online (https://www.j-alz.com/manuscript-disclosures/21-0028r1).

Previous presentations: Data from this study were previously presented at the Alzheimer’s Association International conference 2019, held at Los Angeles, CA from July 14-18, 2019.

References

Chertkow

, Feldman

, Jacova

, Massoud

(2013) Definitions of dementia and predementia states in Alzheimer’s disease and vascular cognitive impairment: Consensus from the Canadian conference on diagnosis of dementia. Alzheimers Res Ther 5(Suppl 1), S2.

American Psychiatric Association (2013) Diagnostic and statistical manual of mental disorders 5th edition (DSM-V). APA Publications, Washington, D.C.

(1994) Canadian study of health and aging: Study methods and prevalence of dementia. CMAJ 150, 899–912.

Helmer

, Pérès

, Letenneur

, Guttiérez-Robledo

, Ramaroson

, Barberger-Gateau

, Fabrigoule

, Orgogozo

, Dartigues

(2006) Dementia in subjects aged 75 years or over with the PAQUID cohort: Prevalence and burden by severity. Dement Geriatr Cogn Disord 22, 87–94.

Lobo

, Launer

, Fratiglioni

, Andersen

, Di Carlo

, Breteler

, Copeland

, Dartigues

, Jagger

, Martinez-Lage

, Soininen

, Hofman

(2000) Prevalence of dementia and different subtypes in Europe: A collaborative study of population based cohorts. Neurologic diseases in the elderly research group. Neurology 54(Suppl 5), S4–S9.

Hesdorffer

, Hauser

, Annegers

, Kokmen

, Rocca

(1996) Dementia and adult-onset unprovoked seizures. Neurology 46, 727–730.

McAreavey

, Ballinger

, Fenton

(1992) Epileptic seizures in elderly patients with dementia. Epilepsia 33, 657–660.

Schachter

, Cramer

, Thompson

, Chaponis

, Mendelson

, Lawhorne

(1998) An evaluation of antiepileptic drug therapy in nursing facilities. J Am Geriatr Soc 46, 1137–1141.

Rao

, Dove

, Cascino

, Petersen

(2009) Recurrent seizures in patients with dementia: Frequency, seizure types, and treatment outcome. Epilepsy Behav 14, 118–120.

10.

Nicastro

, Assal

, Seeck

(2016) From here to epilepsy: The risk of seizure in patients with Alzheimer’s disease. Epileptic Disord 18, 1–12.

11.

Lawn

, Bamlet

, Radhakrishnan

, O’Brien

, So

(2004) Injuries due to seizures in persons with epilepsy: A population-based study. Neurology 63, 1565–1570.

12.

Miziak

, Błaszczyk

, Chrościńska-Krawczyk

, Danilkiewicz

, Jagiełło-Wójtowicz

, Czuczwar

(2014) The problem of osteoporosis in epileptic patients taking antiepileptic drugs. Expert Opin Drug Saf 13, 935–946.

13.

Motheral

, Fairman

(1997) The use of claims databases for outcomes research: Rationale, challenges and strategies. Clin Ther 19, 346–366.

14.

Castilla-Puentes

, Habeych

(2010) Subtypes of depression among patients with Alzheimer’s disease and other dementias. Alzheimers Dement 6, 63–69.

15.

INGENIX (2010) International classification of diseases, 9th revision, clinical modification for hospitals, (ICD-09-CM), 6th edition, vols. 1, 2 and 3, USA.

16.

Issa

, Fenter

, Black

, Grogg

, Kruep

(2006) An assessment of the diagnosed prevalence of diseases in men 50 years of age or older. Am J Manag Care 12(Suppl), S83–S89.

17.

Anderson

, Rocca

, Rosário

(1998) Pitfalls in neuroepidemiological research. Neuroepidemiology 17, 55–62.

18.

Fisher

, Cross

, French

, Higurashi

, Hirsch

, Jansen

, Lagae

, Moshé

, Peltola

, Roulet Perez

, Scheffer

, Zuberi

(2017) Operational classification of seizure types by the international league against epilepsy: Position paper of the ILAE Commission for Classification and Terminology. Epilepsia 58, 522–530.

19.

Scheffer

, Berkovic

, Capovilla

, Connolly

, French

, Guilhoto

, Hirsch

, Jain

, Mathern

, Moshé

, Nordli

, Perucca

, Tomson

, Wiebe

, Zhang

, Zuberi

(2017) ILAE classification of the epilepsies: Position paper of the ILAE Commission for Classification and Terminology. Epilepsia 58, 512–521.

20.

Quan

, Li

, Saunders

, Parsons

, Nilsson

, Alibhai

, IMECCHI Investigators (2008) Assessing validity of ICD-9-CM and ICD-10 administrative data in recording clinical conditions in a unique dually coded database. Health Serv Res 43, 1424–1441.

21.

Pourhoseingholi

, Baghestani

, Vahedi

(2012) How to control confounding effects by statistical analysis. Gastroenterol Hepatol Bed Bench 5, 79–83.

22.

Sen

, Jette

, Husain

, Sander

(2020) Epilepsy in older people. Lancet 395, 735–748.

23.

Irizarry

, Jin

, He

, Emond

, Raman

, Thomas

, Sano

, Quinn

, Tariot

, Galasko

, Ishihara

, Weil

, Aisen

(2012) Incidence of new-onset seizures in mild to moderate Alzheimer’s disease. Arch Neurol 69, 368–372.

24.

Sherzai

, Losey

, Vega

, Sherzai

(2014) Seizures and dementia in the elderly: Nationwide inpatient sample 1999-2008. Epilepsy Behav 36, 53–56.

25.

Imfeld

, Bodmer

, Schuerch

, Jick

, Meier

(2013) Seizures in patients with Alzheimer’s disease or vascular dementia: A population based nested case-control analysis. Epilepsia 54, 700–707.

26.

Bonelli

(2003) Olanzapine associated seizure. Ann Pharmacother 37, 149–150.

27.

Dogu

, Sevim

, Kaleagasi

(2003) Seizures associated with quetiapine treatment. Ann Pharmacother 37, 1224–1227.

28.

Babic

, Zurak

(1999) Convulsion induced by donepezil. J Neurol Neurosurg Psychiatry 66, 410.

29.

Granger

(2001) Ginkgo biloba precipitating epileptic seizures. Age Aging 30, 523–525.

30.

Piecoro

, Wermeling

, Schmitt

, Ashford

(1998) Seizures in patients receiving concomitant anti-muscarinic and acetylcholinesterase inhibitors. Pharmacotherapy 18, 1129–1132.

31.

Logroscino

(2009) Design and methods of cohort studies. Neuroepidemiology 33, 317–319.

32.

Bartko

, Carpenter

(1976) On the methods and theory of reliability. J Nerv Ment Dis 163, 307–317.

33.

https://www.kff.org/health-reform/issue-brief/whats-the-role-of-private-health-insurance-today-and-under-medicare-for-all-and-other-public-option-proposals/, Accessed April 15, 2020.