1 Master Thesis Clinical Neuropsychology
Faculty of Behavioural and Social Sciences – Leiden University (February, 2017)
Student number: 1228323
Daily Supervisors: Edo Grevers & Bert Aldenkamp
Gedragswetenschappelijke Dienst (GWD) Kempenhaeghe CNP-Supervisor: Ilse Schuitema, Department of Health, Medical and Neuropsychology; Leiden University
Cognitive deterioration in chronic refractory
epilepsy
2 Index
Abstract 3
Introduction 4-9
Cognitive impairment vs cognitive deterioration 4
Age at onset and duration 4-5
Epileptic seizures and status epilepticus 5-6
Comorbid diseases 6-7
Antiepileptic drugs (AEDs) 7
Cognitive deterioration: two models explained 7-9
Methods 9-12
Design and study population 9-10
Procedure 10-11
Measures 11
Statistical analysis 11-12
Results 12-17
First research question 12-13
Second research question 13-15
Third research question 15-17
Discussion 18-21
References 22-24
3 Abstract
Objective
To investigate the cognitive deterioration in chronic refractory epilepsy patients whom are assumed on the basis of clinical observations to have substantially deteriorated in comparison with premorbid levels.
Methods
With a retrospective case study design we investigated the clinical characteristics of a group (30 patients) of chronic refractory epilepsy patients. Later the individual disease course was charted. Linear regression analyses were preformed to investigate which characteristics statistically influence the deterioration on a group level.
Major results
On a group level we found the following characteristics: an early age at onset (10.0±6.0), long duration of active epilepsy (36.1±16.0), high seizure frequency (60% of the patients has weekly/daily seizures), high percentage of generalized tonic-clonic seizures (60%), high percentage of status epilepticus (36.7%), dependency to AED polytherapy for many years, and high comorbidity. By charting the important life-events and disease course for each patient we were able to find evidence for both disease course models. In the statistical analysis no significant results were found. We did find a statistically significant difference for all IQ scales between estimated premorbid levels and current IQ (p < .001)
Conclusions and implications of the work
Our findings give insight in the clinical characteristics of a patient population whom have substantially deteriorated from premorbid levels, currently living in a Dutch residential care facility (Providentia, Sterksel) for patients with epilepsy and a mental handicap. Through our multiple case study we were able to show that in a relatively small sample size (from the same population) two completely different disease course models can be present. Because of several limitations no significant results were found in regard to the effect of clinical characteristics on the deterioration.
4 Introduction
Epilepsy is a brain disease that approximately 50 million people around the world are diagnosed with (Brodie et al., 1997). Like many different neurological diseases epilepsy can have consequences on neurobiological, cognitive, social and psychological functioning. In some cases epilepsy is even associated with severe deterioration. The following patient will serve as an example. This patient had a normal development till the age of thirteen, when the debut of epileptic seizures occurred. In the following years he deteriorated from an average intelligence level to living in a specialized Dutch residential care facility for patients with epilepsy and a mental handicap (admitted at the age of nineteen). Several factors may have played a role in this deterioration, like high seizure frequency (daily/weekly), many head injuries because of his seizures, polytherapy (multiple antiepileptic drugs), etc. This thesis will hopefully create a clearer image of the cognitive deterioration seen in patients with chronic refractory (drug-resistant) epilepsy, also in the context of earlier identification of (risks factors for) deterioration so that we can adjust and implement proper guidance.
Cognitive impairment vs cognitive deterioration
Cognitive impairment is often seen in people with chronic epilepsy (Elger, Helmstaedter, & Kurthen, 2004). Many studies have examined this topic, focusing on specific kinds of epilepsy syndromes, like Dravet or the Lennox-Gastaut Syndrome (Vezyroglou & Cross, 2016). Others have assessed cognitive impairment found in specific localized types of epilepsy, like for example memory impairments found in patients with temporal lobe epilepsy (Hendriks et al., 2004). Global cognitive deterioration is the decline in general cognitive functioning/intelligence and is therefore different than impairment in specific cognitive functions. There are several factors that might play a role in causing cognitive deterioration in patients with chronic refractory epilepsy.
Age at onset and duration
Age at onset of epilepsy (either early or adult onset) is associated with global cognitive decline. Vignoli et al. (2016) investigated the long-term outcome of epilepsy with an onset in the first three years of life. They found that an early age at onset is significantly linked to general intellectual deterioration, but found no significant relations between age at onset and etiology, antiepileptic drug resistance and neuro-radiological findings (Vignoli et al., 2016). There are several possible explanations for this global cognitive deterioration suggested. Kaaden and Helmstaedter (2009) for example claimed that the origin of the intellectual deterioration (in early onset epilepsy) is the altered cognitive development that occurs because of the epilepsy and confounding factors (Kaaden & Helmstaedter, 2009). Seizures in an
5 immature brain can have irreversible and long-term consequences on the developing neuronal networks and their connectivity, in turn causing cognitive deterioration (Holmes & Ben-Ari, 2001). Hermann et al. (2006) offered a different explanation. They investigated children with new-onset epilepsy (age: 8-18 years), i.e. the relationship between cognitive/academic achievement and brain abnormalities, and the patterns of development of the immature brain in children with epilepsy versus controls. They found that cognitive impairment and underachievement at school was already present before the onset of epilepsy. There were indications of an altered structure-function relationship in the brain for the children with new-onset epilepsy at the time of the study (Hermann et al., 2006).
Adult onset epilepsy (after the age of 20) is far less investigated, but more research is being devoted to this topic (Breuer et al., 2016). Here, other aspects play a role than in early onset epilepsies. It is known that older age is more often accompanied by a comorbid condition (Stefan et al., 2014). For example risk of developing certain diseases increases with age (e.g. cerebral tumors, cerebrovascular disease, etc.), which in turn increase the chance of developing seizures as a consequence. Furthermore the brain's neural plasticity decreases with age. Some studies have found global cognitive decline in patients with adult onset epilepsy (Taylor & Baker, 2010b), while others did not find any cognitive deterioration, although cognitive function was already impaired at baseline compared to healthy controls (Aikia et al., 2001). The difference between the two study groups was that the last one was seizure-free on anti-epileptic medication monotherapy at the time of the follow-up, indicating a relationship between these factors and cognitive deterioration.
Another important aspect that may play a role in cognitive deterioration is the duration of the disease, which is linked to the age at onset described above. The most straightforward and common answer is that cognitive impairment will increase the longer someone has epileptic seizures. However it is hard to pinpoint this impairment to just the duration as there are many other factors that come to play, for example the long-lasting poor seizure control, accumulating effects of anti-epileptic drug treatment, specific comorbid syndromes or diseases, etc. (Helmstaedter & Kockelmann, 2006).
Epileptic seizures and status epilepticus
The effect of frequent epileptic seizures on cognition has been researched. Vlooswijk et al. (2008) found that a high life-time number of secondary generalized tonic-clonic seizures (which causes a disruption of consciousness and the epileptic activity is spread throughout the whole brain) is associated with general intellectual deterioration (Vlooswijk et al., 2008). The effect of frequent complex partial seizures (epileptic activity in only one part of the brain and
6 consciousness is lost or impaired) is usually less global, effecting specific cognitive subdomains such as memory or executive functioning depending on where in the brain epileptic activity is present (Thompson & Duncan, 2005). The influence of simple partial seizures (where consciousness remains intact) on cognition is less well known (Black et al., 2010; Thompson & Duncan, 2005).
A status epilepticus (SE) is a neurological emergency associated with poor functional outcome (Leitinger, 2015). In 2015 the ILAE (International League Against Epilepsy) revised its definition of SE: ‘’ Status epilepticus is a condition resulting either from the failure of the mechanisms responsible for seizure termination or from the initiation of mechanisms, which lead to abnormally, prolonged seizures. It is a condition, which can have long-term consequences, including neuronal death, neuronal injury, and alteration of neuronal networks, depending on the type and duration of seizures.’’ (Trinka et al., 2015, p.1515). A SE can be either convulsive (tonic-clonic seizure > 5 minutes) or non-convulsive (for example complex partial seizure >10 minutes). Different hypotheses about SE have been brought forward including: SE is a risk factor for cognitive deterioration, the underlying etiology of the SE causes the cognitive decline (Helmstaedter, 2007), and if the SE hits on an already vulnerable brain cognitive deterioration can be a consequence ('second hit model', Breuer et al, 2016).
Comorbid diseases
Neurodegenerative diseases, head trauma, abnormalities in brain structures (e.g. heterotopy), meningitis, brain tumors, etc. can all be causes of developing a seizure syndrome. Vascular disease, especially cerebral stroke, is the most common cause of symptomatic epilepsy (epileptic seizures as a result of one or more identifiable structural brain lesions) in adults (Rodriguez-Sainz et al., 2013). Cerebral stroke is associated with cognitive deterioration and the long-term cognitive outcome is worse for patients with repeated post- stroke seizures compared to those patients that do not develop seizures (De Reuck et al., 2006a). Recurrent post-stroke seizures have been associated with ischemic changes and long-lasting worsening of the original damage caused by the stroke itself (Kumral et al., 2013). Head trauma (traumatic brain injury, TBI) can also be either a cause or a consequence of epileptic seizures. Head trauma can, depending on the severity and location of the injury, have long-lasting effects on cognition. Multiple head trauma as a consequence of seizures is also an important factor in cognitive deterioration, as it can have an accumulative effect on the already existing trauma.
7 After the occurrence of epilepsy (either symptomatic, cryptogenic or idiopathic), patients can also be more susceptible to co-existing diseases/disorders, for example migraine or dementia. Kanner (2016) found that having a co-existing disease often has a severe impact on the quality of life and cognition in patients with treatment resistant epilepsy (Kanner, 2016).
Antiepileptic drugs (AED)
AEDs are employed as a means of suppressing seizures. While the working mechanisms of some AEDs have yet to be unraveled, these drugs essentially address the balance between neuronal excitation and inhibition which can affect cognition. There are several cognitive side effects known to be caused by AEDs, the most common are psychomotor slowness, decreased alertness, and slowing of information processing speed (Aldenkamp, 2011). Patients with refractory epilepsy often are on polytherapy.Complaints of tiredness, memory problems and difficulty concentrating were higher in patients with polytherapy compared to patients on monotherapy (Andrew, Milinis, Baker & Wieshmann, 2012). These patients often have a higher total drug load (If this number exceeds 2, there is a higher chance for adverse cognitive side-effects (St.Louis, 2009)).
There have only been a few studies that found evidence for cognitive deterioration possibly caused by AEDs (Breuer et al., 2016; Vermeulen & Aldenkamp, 1995). For example phenytoin has been found to have a negative effect on memory, mental speed and motor speed. This negative effect however was found to be reversible when the intake of phenytoin was discontinued (Trimble, 1987). The AED topiramate has been reported to cause a significant persistent decline in verbal IQ, fluency and verbal memory (Fritz et al., 2005). Another AED which has reportedly been associated with decrease in intelligence scores (both verbal and performance IQ) is phenobarbital, and this effect has been found to be persistent on the long-term. Most studies have been focused on the effect of phenobarbital on children, and specifically on the negative effect (mainly on the language/verbal developmental skills) this AED has on the developing brain (Sulzbacher, Farwell, Temkin, Lu & Hirtz, 1999). Cognitive deterioration: two models explained
This thesis will be an extension of a topic brought forward in a recent review of Breuer et al. (2016). In this review two different courses/trajectories of global cognitive deterioration in epilepsy are proposed. The first is the ''accumulation chronic model''. This model states that cognitive decline is slow and gradual and a consequence of chronicity of the epilepsy and the accumulation of negative effects of different epilepsy-related factors (e.g. epileptic seizures and polytherapy) on cognition (Breuer et al., 2016). In 1889, Gowers labeled this kind of
8 deterioration ‘epileptic dementia’ (Rose, 2010). The 'persona' (characteristics of a person) that has been described in the literature for this kind of deterioration is one with chronic refractory epilepsy with an early age at onset (which results in a longer epilepsy duration). This deterioration model has remained standing exclusively ever since, but in daily practice another possible form of cognitive deterioration is encountered. This is the second model brought forward in the described review, namely the ''second hit model''. This model states that cascadic cognitive deterioration can occur when epilepsy hits on a vulnerable brain. A first hit for example can be traumatic brain injury (TBI) after an accident, in consequence reducing the cognitive reserve of the brain. Cognitive reserve is a potential compensatory mechanism of the brain, and declines with age. A person with a higher cognitive reserve can sustain greater brain damage before functional impairment occurs (Stern, 2002). After e.g. TBI a second hit can occur in the form of epileptic seizures further diminishing the brains cognitive reserve. The ''second hit model'' explains a cascadic cognitive deterioration, that accelerates the effects of ageing by diminishing the cognitive reserve. For this phenomenon Breuer et al. (2016) coined the term ‘Accelerated Cognitive Ageing’ (ACA, Breuer et al., 2016). Figure 1 graphically represents the ''second hit model''
Figure 1. Graphical representation of Accelerated Cognitive Ageing (ACA). Published by: Breuer et al. (2016). Cognitive deterioration in adult epilepsy: does accelerated cognitive ageing exist? Neuroscience and biobehavioral reviews, 64, p. 8.
9 1)What are the clinical characteristics (on group level) of a group of chronic refractory epilepsy patients whom are assumed on the basis of clinical observations to be substantially deteriorated in comparison with premorbid levels?
2)What is the disease course and the course of deterioration for individual patients? 3)Which characteristics statistically influence on (sub)group level the deterioration?
In regard of the first research question we expect patients with active severe epilepsy, with a high seizure frequency (on polytherapy), and possible high frequency of tonic-clonic seizures (as a link has been found between these seizures and deterioration).
For the disease course we hypothesize that there are different subgroups of patients, which have different clinical characteristics. It is feasible that we will find patients where the disease course is gradual, as described in the long standing ''accumulation chronic model''. As mentioned above this model has been found in patients with chronic refractory epilepsy with an early age at onset. In another subgroup of adult patients with chronic refractory epilepsy, cognitive deterioration may be cascadic as described in the ''second hit model'' rather than gradual. Based on the literature this subgroup is hypothesized to be characterized by an adult age at onset, resulting in a shorter duration of epilepsy, is on polytherapy (multiple AEDs), with a history of SE, with a lower premorbid education level (resulting in a lower cognitive reserve), and high comorbidity (Breuer et al., 2016).
We hypothesize the following characteristics to influence the course of deterioration: age at onset, duration, frequency and type of seizures, history of SE, comorbidity, and AED treatment. We believe that these factors have a significant influence if there is: a long duration of active epilepsy (which is linked with an early age at onset and older age), high seizure frequency (weekly/daily), the presence of generalized tonic-clonic seizures, a history of one or multiple status epilepticus (especially convulsive SE), comorbid disease and a long history and current intake of AED polytherapy (especially on phenytoin, topiramate and phenobarbital).
Methods Design and study population
A retrospective design was chosen, 30 patients (20 men; mean age 48 years old) who were diagnosed with refractory epilepsy (either symptomatic, cryptogenic or idiopathic) who show evidence of worrisome cognitive deterioration (assumed on the basis of clinical observations and existing neuropsychological data) were included in this study. These patients were selected by the clinicians working in a Dutch residential care facility (Providentia, Sterksel) for patients with epilepsy and a mental handicap. All patients are currently living in
10 this residential care facility. Most of the patients underwent a neuropsychological examination for diagnostic purposes in a Dutch tertiary care center for epilepsy (Kempenhaeghe, Heeze). Exclusion criteria were other than a Dutch nationality and/or suspicions of neurodegenerative diseases (e.g. diagnosis of dementia) which can underlie cognitive deterioration. This thesis is part of the recently established project ‘deterioration de novo’ in Kempenhaeghe.
Procedure
We have reviewed existing psychometric/cognitive data. All patient related data was collected from an electronic patient database (EPD). Before this electronic database was introduced patient related information was gathered on paper files which are partly still available on site. The following variables were collected: age, gender, premorbid educational level, age at onset, duration of active epilepsy, type of epilepsy, frequency and type of seizures, history of status epilepticus, number of taken AEDs (and which kind of AEDs), total anti-epileptic drug load, and comorbidity (cardiovascular disease, cerebrovascular disease, traumatic brain injury). The seizure frequency was calculated using patient records and seizure diaries. The total anti-epileptic drug load was calculated by dividing the Prescribed Daily Dose (PDD, the dose a patient takes of that drug a day) with the Defined Daily Dose (DDD, average maintenance dose of the drug, according to criteria of the WHO Collaborating Centre for Drug Statistics Methodology).
A timeline was created for each patient featuring all the important life-events. This timeline was later used to create a deterioration + life-events figure showing the line of cognitive deterioration for each patient (see appendix 1 for two examples). Based on qualitative analyses the individual cognitive deterioration was charted and after this a selection was made of patients in which deterioration seemed to have occurred in a cascadic or in a gradual course. The neuropsychological assessments (NPA) performed were spread between 1981 and 2016. A great variety of intelligence measures was used, and therefore it was not possible to create a uniform outcome variable to reliably calculate the degree of deterioration (current IQ- premorbid IQ) for the whole group. Therefore, for the third research question (Which characteristics statistically influence on (sub)group level the deterioration?) a subgroup of 15 patients (out of 30) was selected who had in approximately the last ten years (2005-2016) a Dutch version of the Wechsler Adult Intelligence Scale-III or IV (Wechsler, 1997; Wechsler, 2008) administered. For these patients the level of deterioration could be determined more precisely, which made analyses on group level possible.
11 All patients signed a generic informed consent and gave permission to use clinically obtained data for scientific purposes. Those patients who refused this consent are not included in this study. This research has been approved by the Committee Medical Ethics and the research and development committee of Kempenhaeghe.
Measures
Deterioration was operationalized by subtracting the current IQ of the estimated premorbid IQ. Current intelligence was measured with the Dutch version of the Wechsler Adult Intelligence Scale-III or IV (Wechsler, 1997; Wechsler, 2008). Measures of current full scale IQ (FSIQ), verbal IQ (VIQ) and performance IQ (PIQ) were used. The WAIS-IV has no indexes for VIQ and PIQ, so in order to make reliable comparisons the verbal comprehension index (VCI) and the perceptual organization index (POI) of the WAIS-III were used, as these are comparable to the verbal comprehension index and perceptual reasoning index (PRI) of the WAIS-IV.
For an estimation of the premorbid IQ the Oklahoma Premorbid Intelligence Estimate (OPIE-3) was used (Schoenberg et al., 2002; Schoenberg et al., 2003). The OPIE-3 predicts premorbid intelligence with the use of several variables: age, education, gender, ethnicity, region of country (only for the United States) and the raw scores of two WAIS-subtests (Vocabulary and Matrix Reasoning). The original formula uses the following categorization for education (in years): ≤8, 9-11, 12, 13-15, ≥16. This categorization was not usable for our sample because of a different education system in the Netherlands compared to the United States. We made our own 5-point variable for education by using the coding system of Verhage (Verhage, 1964). This resulted in the following scale: 1= Verhage score 1-3, 2= Verhage score 4, 3= Verhage score 5, 4=Verhage score 6, and 5=Verhage score 7. The variables ethnicity and region of country did not account for variability in premorbid IQ scores in our sample, as all patients were Caucasian and had a Dutch nationality. Three measures of premorbid IQ were calculated by the use of different OPIE-algorithms. The OPIE-3 (2ST) was used for the estimation of premorbid FSIQ, the OPIE-3V for premorbid verbal IQ and the OPIE-3MR for premorbid performance IQ.
Statistical analyses
The statistical analyses were performed with the Statistical Package for Social Sciences (SPSS, version 21.0, IBM Crop., Armonk, NY, USA) for Windows. A significance level of P ≤ 0.05 was chosen. Demographic and clinical characteristics for the whole group were assessed using descriptive statistics. Through the qualitative case study we were able to create two different groups (either cascadic or gradual). Descriptive statistics were computed
12 to assess the demographic and clinical characteristics for each group. We performed an independent-samples t-test and chi-square test (for categorical variables) to evaluate the similarities and differences between these groups. For our third research question a paired-samples t-test was performed to calculate if the difference between the current IQs and premorbid IQs were significant. Multiple linear regression analyses were performed to evaluate the explained variance of the deterioration score by main epilepsy characteristics and other clinical and demographic factors. The independent variables were of interval/ratio level, the categorical variables were converted into 'dummy' variables. The Backward method of multiple regressions was chosen.
Results: first research question (clinical characteristics of the whole group)
Main epilepsy characteristics and other clinical and demographic data are shown in table 1. Remarkable characteristics are an early age at onset (10.0±6.0), on average long duration of epilepsy (36.1±16.0), mostly of cryptogenic origin (60%), with a high mean seizure frequency (60% of the patients has weekly/daily seizures), mostly with complex partial seizures (36.7%). A very high number of patients (60%) had tonic-clonic seizures in the two years preceding neuropsychological assessment, which are known to be associated to general intellectual deterioration (Vlooswijk et al., 2008). Status epilepticus is a common occurrence in this patient group, 36.7% has had at least one status epilepticus in the past (convulsive or non-convulsive). Almost all patients are on polytherapy (86,7%), which is in line with the high total drug load (3.3±1.5). Premorbid educational level is below average to average. The presence of comorbid diseases was high (see table 2). The most common comorbid disease is traumatic brain injury (46.7%).
Table 1
Demographic and clinical variables of all patients
Characteristics Mean ± SD or n(%)
Demographic
Age (years) 48.0 ±15.7 (range: 17-87)
Gender Male/female 20 (66.7)/10 (33.3)
Educational levelᵇ Only elementary school
Elementary school + two years further education
Lower vocational education Secondary vocational education Higher education 5 (16.7) 6 (20) 10 (33.3) 7 (23.3) 2 (6.7) Epilepsy variables
Age at onsetª (years) 10.0 ±6.0 (range: 1-22)
Durationª (years) 36.1±16.0 (range: 3-85)
13 epilepsyª (years)
Type of epilepsy Cryptogenic Symptomatic Idiopathic 18 (60.0) 11 (36.7) 1 (3.3) Dominant seizure typeᶜ Tonic-clonic seizures Complex partial seizures Tonic seizures Seizure free 8 (26.7) 11 (36.7) 6 (20.0) 5 (16.7) Total seizure frequency
Seizure free without AEDs Seizure free with AEDs 1 seizure per 2 months Monthly seizures Weekly seizures Daily seizures 1 (3.3) 4 (13.4) 3 (10.0) 4 (13.3) 11 (36.7) 7 (23.3)
Tonic-clonic seizuresᶜ Yes/no 18 (60.0)/12 (40.0)
Status epilepticus Yes/no 11 (36.7)/19 (63.3)
Medication variables Number of taken AEDs (anti-epileptic drugs) 0 AEDs Monotherapy Polytherapy 1 (3.3) 3 (10.0) 26 (86.7)
Total drug load 3.3 ± 1.5 (range: 1.1-6.7)
Phenobarbital Yes/no 2 (6.7)/28 (93.3)
Phenytoin Yes/no 10 (33.3)/20 (66.7)
Topiramate Yes/no 8 (26.7)/22 (73.3)
ª Two missing values (N=28)
ᵇ Education level is based on Verhage coding system (Verhage, 1964) ᶜ Determined for the two years preceding neuropsychological assessment.
Table 2
Comorbid disease and other brain abnormalities n (%) yes/no Cardiovascular 8 (26.7)/22 (73.3) Cerebrovascular 4 (13.3)/26 (86.7) Traumatic brain injury (TBI) 14 (46.7)/16 (53.3) Other (i.e., inflammatory,
immunologic)
4 (13.3)/26 (86.7)
Results: second research question (disease course)
In the qualitative analyses, the course of the deterioration in 15 of the 30 patients seemed to have developed cascadicly and are labeled in the group 'cascadic deterioration', and for 6 patients the deterioration course seemed to have developed more gradually and are labeled in the group 'gradual deterioration'. A total of 8 patients could not be classified in one of the groups, because on the basis of the available data the course of decline was uncertain (6 patients), or only impairments in specific cognitive functions were found (2 patients). One patient was excluded because we concluded that the cognitive deterioration that occurred was a consequence of a diffuse congenital brain abnormality. Demographic and clinical
14 characteristics were assessed using descriptive statistics for the 'cascadic deterioration' group and the 'gradual deterioration' group. These results are shown in table 3.
Table 3
Demographic and clinical variables for the 'gradual deterioration' (n=6.) and the 'cascadic deterioration' group (n=15) Gradual deterioration Cascadic deterioration Demographic Age 49.2±15.9 (range: 29-77) 42.3±13.0 (range: 17-60) Gender Male/female 9 (60)/6 (40) 5 (83.3)/1 (16.7)
Educational level Only elementary school Elementary school + two years further education Lower vocational education Secondary vocational education Higher education 1 (16.7) 2 (33.3) 1 (16.7) 2 (33.3) 0 (0.0) 0 (0.0) 3 (20) 8 (53.3) 2 (13.3) 2 (13.3) Epilepsy variables Age at onset 8.2±4.7 (range: 4-16) 10.9±6.3 (range: 1-22) Duration (years) 35.4±8.4 (range: 23-44) 31.3±15.2 (range: 3-50) Duration active epilepsy (years) 34.4±9.2 (range: 23-44) 30.2±14.8 (range: 3-48) Type of epilepsy Cryptogenic
Symptomatic 2 (33.3) 4 (66.7) 11 (73.3) 4 (26.7) Dominant seizure type Tonic-clonic Complex partial Tonic
Seizure free with AEDs
0 (0.0) 3 (50) 2 (33.3) 1 (16.7) 4 (26.7) 6 (40.0) 4 (26.7) 1 (6.7) Seizure frequency Seizure free with AEDs
1 seizure per 2 months Monthly Weekly Daily 1 (16.7) 0 (0.0) 0 (0.0) 3 (50) 2 (33.3) 1 (6.7) 2 (13.3) 1 (6.7) 6 (40.0) 5 (33.3) Tonic-clonic seizures
(last two years)
Yes/no 4 (66.7)/2 (33.3) 9 (60)/6 (40)
Status epilepticus Yes/no 1 (16.7)/ 5 (83.3) 7 (46.7)/8 (53.3)
Medication variables
Number of AEDs Monotherapy Polytherapy
1 (16.7) 5 (83.3)
2 (13.3) 13 (86.7)
Total drug load 3.0±1.4
(range: 1.5-4.7)
3.5±1.6
(range: 1.3-6.7) Phenobarbital Yes/no 0 (0.0)/6 (100.0) 1 (6.7)/14 (93.3)
15 Topiramate Yes/no 1 (16.7)/5 (83.3) 4 (26.7)/11 (73.3) Comorbid disease Cardiovascular disease Yes/no 2 (33.3)/4 (66.7) 3 (20)/12 (80) Cerebrovascular disease Yes/no 2 (33.3)/ 4 (66.7) 1 (6.7)/14 (93.3) TBI Yes/no 5 (83.3)/1 (16.7) 4 (26.7)/ 11 (73.3)
Values are mean ± SD or n(%)
Between groups similarities and differences
A statistical analysis was used to check for potential significant differences between the two groups on the different variables. An independent-samples t-test (for interval/ratio variables) and chi-square test (for categorical variables) was conducted to compare the described variables in the 'gradual deterioration' and the 'cascadic deterioration' conditions. We only found a significant effect for the variable TBI: χ²(1) = 5.62, p = .018. TBI seems to have occurred more often in the 'gradual deterioration' group. Another variable which shows a difference between the two groups is 'type of epilepsy', with a higher percentage of patients in the 'cascadic deterioration group' having a cryptogenic origin of epilepsy (73.3% vs 26.7% in the 'gradual deterioration' group). The result found was a trend: χ²(1) = 2.91, p = .088. For the other variables no significant result was found, indicating that the between group similarities are large. The biggest similarities between the two groups are found in: seizure frequency (χ²(4) = 1.75, p = .782), a high percentage of the presence of tonic-clonic seizures in the two years preceding neuropsychological assessment (both around 60%, χ²(1) = .08, p = .776), and more than 80% of the patients are on polytherapy in both groups: χ²(1) = .04, p = .844 (which is in line with the high total drug load).
Results: third research question (clinical factors that influence deterioration)
As described in the method section we could only reliably calculate the degree of deterioration for 15 of our 30 patients. Because of this missing data we decided not to use the two created subgroups from the case study (second research question) for our third research question (we only had the premorbid and current FSIQ, VIQ and PIQ data of one patient in the 'gradual deterioration' group in comparison with 14 patients in the 'cascadic deterioration' group). Which clinical factors influence the deterioration is evaluated on a group level below. Intelligence measures
Estimated premorbid IQ, current IQ and deterioration scores are presented in table 4. When looking at these results we see that for premorbid IQ the FSIQ, VIQ and PIQ are close
16 in mean scores, respectively of 87.2 (SD 14.2), 89.2 (SD 12.3) and 94.5 (SD 11.1). For current IQ we find a different result. The VIQ and PIQ are close in means (76.2±18.0 ; 74.2±16.5), while the FSIQ is lower with a mean of 67.9 (SD 12.2).
Table 4
Estimated premorbid IQ, Current IQ and Deterioration score
N Premorbid IQ N Current IQ N Deterioration score (Premorbid IQ - Current IQ) Full Scale IQ 15 87.2 (14.2) 15 67.9 (12.2) 13 19.3 (6.7) Verbal IQ 15 89.2 (12.3) 13 76.2 (18.0) 13 13.0 (8.0) Performance IQ 15 94.5 (11.1) 13 74.2 (16.5) 13 20.3 (10.7) Values are mean±SD
Figure 2 shows the comparisons between the premorbid and current IQs for all three IQ scales. A paired-samples t-test was conducted to see if there are significant differences between the premorbid and current IQ. There was a significant difference in the scores for premorbid FSIQ (M=87.2, SD=14.2) and current FSIQ (M=67.9, SD=12.2); t(14) = 11.2, p < .001, in the scores for premorbid VIQ (M=89.2, SD=12.3) and current VIQ (M=76.2, SD=18.0); t(12) = 5.9, p < .001 and in the scores for premorbid PIQ (M=94.5, SD=11.1) and current PIQ (M=74.2, SD=16.5); t(14) = 11.2, p < .001.
17 Cognitive deterioration: epilepsy variables
A total of three linear regression analyses were performed with the deterioration score of the FSIQ, VIQ and PIQ as the dependent variables. The following predictors were entered in our analyses: age at onset, duration, type of epilepsy, seizure frequency, presence of tonic-clonic seizures, status epilepticus, and total drug load. First we checked the assumptions of linear regression:
1) Normality: a non-significant result for the shapiro-wilk test was found for FSIQ (p = .741), VIQ (p = .848), and PIQ (p = .716). This shows that the dependent variables are normally distributed.
2) No multicollinearity: this assumption was met. The correlation between predictors was lower than .70.
3) Linear relationship between the three dependent variables and the predictors: this assumption was met based on the normal P-P plot (units roughly follow the linear line) and scatterplot (units fall between -3 and 3 on both the X- and Y-axis) for all variables.
The Backward method was used. The linear regressions showed no significant results, the predictors only accounted for 4% of the variance of the deterioration score FSIQ, 11% of the deterioration score VIQ, and 8% of the deterioration score PIQ.
Cognitive deterioration: linear regression AEDs
Another variable discussed in the introduction is AEDs. Three AEDs in particular are mentioned for their proven negative effect on cognitive functions and/or causing cognitive deterioration. These are phenobarbital, phenytoin, and topiramate. Phenobarbital was excluded for this analysis as for only one patient (out of 15) this AED was prescribed. Three different linear regression analyses were computed, with respectively the deterioration scores FSIQ, VIQ and PIQ as dependent variables. The results are shown in table 5. No significant results were found.
Table 5
Results linear regression AEDS
Dependent variable Model AED Unstandardized B Sig. Variance accounted for by predictors FSIQ 1 Phenytoin Topiramate -4.3 1.3 .27 .74 5.8% (p = .74) 2 Phenytoin -4.6 .20 VIQ 1 Phenytoin .62 .90 Topiramate 5.5 .32 2 Topiramate 5.3 .29 1.9% (p = .90) PIQ 1 Phenytoin -5.2 .41 Topiramate 6.4 .36 2 Topiramate 8.0 .23 4,9% (p = .41)
18 Discussion
Deterioration in adult patients with epilepsy has been the subject of a recent established project 'deterioration de novo' in Kempenhaeghe, a Dutch tertiary care center for epilepsy. In this thesis we put our focus on two models that have been described in the literature in the context of cognitive deterioration in epilepsy. The first is the 'accumulation chronic model' which assumes that deterioration occurs slow and gradual (progressive) in the light of chronicity and accumulation of epilepsy-related factors. The second model is relatively new, this ''second hit model'' explains a cascadic cognitive deterioration, that accelerates the effects of ageing by diminishing the cognitive reserve. For this phenomenon Breuer et al. (2016) coined the term ‘Accelerated Cognitive Ageing’ (Breuer et al., 2016).
In this study, we examined the disease course and characteristics for an unique patient group. These patients all live in a Dutch residential care facility for patients with epilepsy and a mental handicap (‘Providentia’), but were not born with a mental handicap, so cognitive deterioration is suspected. The cognitive deterioration and the characteristics that may influence this deterioration were evaluated.
Characteristics of the patient group
When looking at the group as a whole a few characteristics mainly stand out, one of these is the age at onset. Almost all patients had an early age at onset, with only one patient having an epilepsy onset above 20 years of age. Next to an early age at onset, we found a high current seizure frequency in our patient sample. A relationship between high seizure frequency and cognitive decline has been proven in previous research. For example Thompson and Duncan (2005) found that cognitive decline was severe across a wide range of cognitive functions in patients with severe refractory epilepsy. The strongest predictor of deterioration was the frequency of generalized tonic-clonic seizures, as complex partial seizures only caused specific cognitive impairments but no decline in IQ scores (Thompson & Duncan, 2005). In accordance to these findings, in our sample with patients that deteriorated from premorbid level, we found a very high percentage (60%) of generalized tonic-clonic seizures which based on previous research can be seen as an important risk factor for deterioration. In the total epilepsy population only about 25% of the people have generalized tonic-clonic seizures (according to the American Epilepsy Society).
In short several risk factors for cognitive deterioration as described in previous studies have been found in our research group. This is an early age at onset (which may cause an altered cognitive development, Kaaden & Helmstaedter, 2009), long duration of active epilepsy, high seizure frequency, high percentage of generalized tonic-clonic seizures, high
19 percentage of status epilepticus, dependency to AED polytherapy for many years, and high comorbidity. These factors combined may all together have influenced the worrisome disease course witnessed in this patient sample. Before we evaluated this aspect we first charted the disease course for each patient.
Disease course
By doing a multiple case study we were able to chart all the important life-events and the deterioration course for each patient. We found evidence for both disease course models for cognitive deterioration, with the group in which deterioration seemed to have developed in a cascadic way being notable larger. We found similarities and differences between the groups, but most differences were not great enough to be significant. The only significant difference between the groups was the existence of TBI, which percentage was larger in the 'gradual deterioration' group. This finding is contrary to what we would expect, as the existence of TBI in previous studies has been linked to cascadic deterioration. Unfortunately this finding could not be reliably explained with the information available. It is possible that this finding will not be replicated in a larger patient sample.
An interesting finding was the early age at onset in both groups. In a previous empirical study of Breuer et al. (2016) the cognitive deterioration in adult epilepsy was also investigated (the clinical characteristics of 'accelerated cognitive ageing'). In this group, most patients had an adult age of onset and a relatively short duration of epilepsy and they found an adult age of epilepsy onset to be a possible determining factor for cascadic deterioration. Our results show that cascadic cognitive deterioration can also occur in a patient group with an early age at onset (chronic refractory epilepsy). This implies that in different patient samples cascadic deterioration can occur.
Influence of the evaluated characteristics on the cognitive deterioration
In further analyses we examined which characteristics statistically influence the deterioration. All factors described in the introduction (which can possibly cause deterioration) are added in this analyses, except comorbidity. We eventually concluded that there was too much concordance between the different comorbid variables, as many patients have more than one comorbid disease. For this reason a clear distinction could not be made.
In the analyses no significant results were found. We did find a statistical difference for all IQ scales between estimated premorbid level and current IQ (FSIQ, VIQ, and PIQ). However we speculate that the largest deterioration occurred for FSIQ and PIQ. A possible explanation for this lower FSIQ is that besides VIQ and PIQ, the FSIQ consist of two more indexes (Processing speed and Working memory). These indexes represent fluid functions of the
20 brain, which are usually affected first when there is some form of brain disease (Malojcic, Brinar, Susnic, Coric, & Mubrin, 2000). Because of our non-significant findings no definite conclusions can be drawn for the question which characteristics influence cognitive deterioration.
Reflection and suggestions for future research
After the first evaluation of the characteristics of the patient sample we soon realized the large heterogeneity of the group (e.g. in age and clinical characteristics, but also in different cognitive measures, etc.). Therefore we chose to perform a retrospective multiple case study (30N). Trough this method we were able to collect a lot more detail about the patients disease course and their characteristics over a long period of time. A limitation of this method was that we had to rely on the information available. Some electronic patient databases missed important information or contradicting information was found. The paper files on site were helpful in partly solving this problem. The case study design has the disadvantage that neither generalization to the wider population nor drawing definite conclusions about cause and effect is possible. Therefore we selected a smaller group (15 patients) from our sample, which consisted of patients that have a recent intelligence test administered to study the data on a group level. This method was chosen to create a uniform outcome variable (deterioration score) for this subgroup. We believe an important reason for the many non-significant results between deterioration and the characteristics lies in the sample size. For future research we suggest at least 20 patients per added predictor in the regression analysis (Palmer & O'Connell, 2009). Only then definite conclusions can be made on the notion of which characteristics significantly influence the deterioration course of patients with chronic refractory epilepsy. Another recommendation is adding a matched control group with chronic epilepsy patients where no deterioration has occurred.
While our small sample size makes it difficult to draw definite conclusion about the relationships between the characteristics and the disease course, our case study gave us important new insights. We were able to describe the characteristics of this specific patient group and conclude that the deterioration course can be substantially different for each patient. We were able to point out possible risk factors for deterioration (in this patient population) based on the group characteristics. This thesis is the first step toward better understanding of the decline of chronic refractory epilepsy patients who have deteriorated from premorbid level.
21 Conclusions
In conclusion we can state the following. Our sample can be described as patients with chronic refractory epilepsy, with an early age at onset, mainly high seizure frequency (daily/weekly), with a high percentage that is not seizure-free for generalized tonic-clonic seizures. Also, a history of status epilepticus, high total drug load and comorbidity are characteristics that occurred frequently. Trough our case study we found indications for both deterioration models represented in our group, with a larger percentage in which deterioration seemed to have developed in a cascadic course. This result shows us that in a relatively small sample size (from the same population) two completely different disease course models can be present. Unfortunately no conclusions can be made on the ground of what causes these differences in the disease course in the same patient population, as no significant results were found between the groups except for TBI. Also, on a group level we found no clinical characteristics statistically influence the deterioration. We believe both results are mainly caused by the small sample size. Future research including larger numbers and possibly a matched control group may hopefully shed some light on these important questions.
22 References
Aikia, M., Salmenpera, T., Partanen, K., & Kalviainen, R. (2001). Verbal Memory in Newly Diagnosed Patients and Patients with Chronic Left Temporal Lobe Epilepsy. Epilepsy & Behavior, 2(1), 20- 27.
Aldenkamp, A.P. (2011). Antiepileptic drugs and cognitive disorders. In: M.R. Trimble & B. Schmitz. (Ed.), The Neuropsychiatry of Epilepsy (p. 153-163). Cambridge: Cambridge University Press.
Andrew, T., Milinis, K., Baker, G., & Wieshmann, U. (2012). Self reported adverse effects of mono and polytherapy for epilepsy. Seizure-European journal of epilepsy, 21(8), 610-613.
Black, L.C., Schefft, B.K., Howe, S.R., Szaflarski, J.P., Yeh, H.-S., & Privitera, M.D. (2010). The effect of seizures on working memory and executive functioning performance. Epilepsy & Behavior, 17(3), 412-419.
Breuer, L.E.M., Boon, P., Bergmans, J.W.M., Mess, W.H., Besseling, R.M.H., de Louw, A., Tijhuis, A.G., Zinger, S., Bernas, A., Kloosterm, D.C.W., & Aldenkamp, A.P. (2016). Cognitive deterioration in adult epilepsy: does accelerated cognitive ageing exist? Neuroscience and biobehavioral reviews, 64, 1-11.
Breuer, L.E.M, Grevers, E., Boon, P, Bernas, A., Bergmans, J.W.M, Besseling, R.M.H., Klooster, D.C.W., de Louw, A., Mestrom, R.M.C., Vonck, K., Zinger, S., Aldenkamp, A.P. (2016). Cognitive deterioration in adult epilepsy: clinical characteristics of ''Accelerated Cognitive Ageing''. Acta Neurologica Scandinavica, 00, 1-7.
Brodie, M.J., Shorvon, S.D., Canger, R., Halász, P., Johannessen, S., Thompson, P., Wieser, H.G., & Wolf, P. (1997). Commission on European Affairs: appropriate standards of epilepsy care across Europe: ILEA. Epilepsia, 38(11), 1245-1250.
De Reuck, J., De Clerck, M., & Van Maele, G. (2006a). Vascular cognitive impairment in patients with late-onset seizures after an ischemic stroke. Clinical Neurology and Neurosurgery, 108(7), 643-637.
Elger, C.E., Helmstaedter, C., & Kurthen, M. (2004). Chronic epilepsy and cognition. The Lancet. Neurology, 3(11), 663-672.
Fritz, N., Glogau, S., Hoffmann, J., Rademacher, M., Elger, C.E., & Helmstaedter, C. (2005). Efficacy and cognitive side effects of tiagabine and topiramate in patients with epilepsy. Epilepsy & Behavior, 6(3), 373-381.
23 Helmstaedter, C., & Kockelmann, E. (2006). Cognitive outcomes in patients with chronic
temporal lobe epilepsy. Epilepsia, 47, 96-98.
Helmstaedter, C. (2007). Cognitive outcome of status epilepticus in adults. Epilepsia, 48, 85-90.
Hendriks, M.P., Aldenkamp, A.P., Alpherts, W.C., Ellis, J., Vermeulen, J. , & Vlugt, H. van der (2004). Relationships between epilepsy-related factors and memory impairment. Acta Neurologica Scandinavica, 110(5), 291-300.
Hermann, B., Jones, J., Sheth, R., Dow, C., Koehn, M., & Seidenberg, M. (2006a). Children with new onset epilepsy: neuropsychological status and brain structure. Brain, 129, 2609-2619.
Holmes, G.L., & Ben-Ari, Y. (2001). The neurobiology and consequence of epilepsy in de developing brain. Pediatric research, 49(3), 320-325.
Kaaden, S., & Helmstaedter, C. (2009). Age at onset of epilepsy as a determinant of intellectual impairment in temporal lobe epilepsy. Epilepsy & Behavior, 15(2), 213-217.
Kanner, A.M. (2016). Management of psychiatric and neurological comorbidities in epilepsy. Nature reviews neurology, 12 (2), 106-116.
Kumral, E., Uncu, G., Dönmez, I., Cerrahoglu Şirin, T., Alpaydın, S., Çallı, C., & Kitiş, Ö. (2013). Impact of Poststroke Seizures on Neurological Deficits: Magnetic Resonance Diffusion-Weighted Imaging Study. European Neurology, 69(4), 200-206.
Leitinger, M., Kalss, G., Rohracher, A., Pilz, G., Novak, H., Hofler, J., Deak, I., Kuchukhidze, G., Dobesberger, J., Wakonig, A., & Trinka, E. (2015). Predicting outcome of status epilepticus. Epilepsy and behavior, 49, 126-130.
Malojcic, B., Brinar, V., Susnic, M., Coric, B., & Mubrin, Z. (2000). Cognitive functions in mild to moderate traumatic brain injury. Neurologica croatica, 49(3), 159-173.
Palmer, P.B., & O'Connell, D.G. (2009). Regression analysis for prediction: understanding the process. Cardiopulmonary physical therapy journal, 20(3), 23-26.
Rodriguez-Sainz, A., Pinedo-Brochado, A., Sanchez-Menoyo, J.L., Ruiz-Ojeda, J., Escalza- Cortina, I., & Garcia-Monco, J.C. (2013). Migraine, stroke and epilepsy: underlying and interrelated causes, diagnosis and treatment. Current treatment options cardiovascular medicine, 15(3), 322-334.
Rose F.C. Chapter 39: an historical overview of British neurology. In: Finger S, Boller F, Tyler KL, eds. Handbook of Clinical Neurology: History of Neurology. Edinburgh: Elsevier; 2010:95:613-628.
24 St.Louis, E.K. (2009). Truly ''rational'' polytherapy: maximizing efficacy and minimizing drug interactions, drug load, and adverse effects. Current Neuropharmacology, 7(2), 96-105.
Stefan, H., May, T.W., Pfäfflin, M., Brandt, C., Füratsch, N, Schmitz, B., Wandschneider, B., Kretz, R., Runge, U., Geithner, J., Karakizlis, C., Rosenow, F., & Kerling, F. (2014). Epilepsy in the elderly: comparing clinical characteristics with younger patients. Acta Neurologica Scandinavica, 129(5), 283-293.
Stern, Y. (2002). What is cognitive reserve? Theory and research application of the reserve concept. Journal of the international neuropsychological society, 8(3), 448-460.
Sulzbacher, S., Farwell, J.R., Temkin, N., Lu, A.S., & Hirtz, D.G. (1999). Late cognitive effects of early treatment with phenobarbital. Clinical pediatrics, 38(7), 387-394. Taylor, J., & Baker, G.A. (2010b). Newly diagnosed epilepsy: Cognitive outcome at 5 years.
Epilepsy & Behavior, 18(4), 397-403.
Thompson, P.J., & Duncan, J.S. (2005). Cognitive Decline in Severe Intractable Epilepsy. Epilepsia, 46(11), 1780-1787.
Trimble, M.R. (1987). Anticonvulsant drugs and cognitive function: a review of the literature. Epilepsia, 28, 37-45.
Trinka, E., Cock, H., Hesdorffer, D., Rosetti, A.O., Scheffer, I.E., Shinnar, S., & Lowenstein, D.H. (2015). A definition and classification of status epilepticus – report of the ILAE task force on classification of status epilepticus. Epilepsia, 56(10), 1515-1523.
Verhage, F. (1964). Intelligentie en leeftijd. Assen: van Gorcum
Vermeulen, J., & Aldenkamp, A.P. (1995). Cognitive side-effects of chronic antiepileptic drug treatment: a review of 25 years of research. Epilepsy research, 22, 65-95.
Vezyroglou, K., & Cross, J.H. (2016). Targeted treatment in childhood epilepsy syndromes. Current treatment options in neurology, 18(6), 1-12.
Vignoli, A., Peron, A., Turner, K., Scornavacca, G.F., La Briola, F., Chiesa, V, Zambrelli, E., & Canevini, M.P. (2016). Long-term outcome of epilepsy with onset in the first three years of life: findings from a large cohort of patients. European journal of pediatric neurology, 20(4), 566-572.
Vlooswijk, M.C.G., Jansen, J.F.A., Reijs, R.P., Krom de, M.C.T.F., Kooi, M.E., Majoie, H.J.M., Hofman, P.A.M, Backes, W.H., & Aldenkamp, A.P. (2008). Cognitive fMRI and neuropsychological assessment in patients with secondarily generalized seizures. Clinical neurology and neurosurgery, 110(5), 441-450.
25 Appendix 1: 'gradual deterioration' (1) and 'cascadic deterioration' (2)
