Incidence, predictors and cognitive dysfunction in post-traumatic epilepsy following a traumatic brain injury: A study protocol

Introduction:
51 At least 10 million people sustain head trauma every year [2,3], making 
52 traumatic brain injury (TBI) one of the leading global causes of death and disease 
53 burden [1]. One of TBI's major complications but poorly understood sequelae is post54 traumatic epilepsy (PTE) [4–6]. PTE is a significant form of symptomatic epilepsies 
55 resulting from various brain insults [6]. In the general population, it causes 10–20% of 
56 symptomatic epilepsy and 5% of all epilepsy [7]. 
57 PTE poses significant health and socio-economic problem worldwide [8]. 
58 People with epilepsy are three times more likely to die prematurely than the general 
59 population, with the absolute number of deaths increasing by 39% [1]. Post-traumatic 
60 epileptic seizures are distinguished from non-epileptic post-traumatic seizures (PTS) 
61 by their etiology and timing after the trauma. Late seizures, occurring more than a 
62 week after the initial event, characterise PTE. [9]. The time between a TBI and the 
63 onset of the first seizure varies greatly [10]. In about 80% of cases, the first seizure 
64 following PTE occurs within the first year, and more than 90% of cases occur by the 
65 end of the second year [11]. Eighty-two percent of patients who experienced a late66 onset seizure (more than a week after injury) also experienced a second seizure within 
67 the following two years [8]. Although a seizure during the first week after injury (early 
68 PTS) is linked to a higher incidence of late PTS (after the first week of injury), having 
It is made available under a CC-BY 4.0 International license . 
perpetuity. 
preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in
medRxiv preprint doi: https://doi.org/10.1101/2022.11.10.22282171; this version posted November 11, 2022. The copyright holder for this
4
69 a late PTS is correlated with an even higher risk of seizure recurrence [12–14]. The 
70 likelihood of seizures following a TBI has been predicted mathematically by several 
71 authors, but their clinical applicability for the majority of brain-injured people has not 
72 been tested [13]. Researchers have tried to pin down the true prevalence of PTE for 
73 decades, and their findings have yielded a wide range of reported numbers, from 1.3% 
74 to 53.3% [15–17]. This variation may be attributable to differences in the severity of 
75 the initial TBI. The cumulative incidence of PTE was reported to be 4.4 per 100 people 
76 with mild TBI, 7.6% per 100 people with moderate TBI, and 13.6 per 100 people with 
77 severe TBI in a three-year population-based study conducted in the United States [4]. 
78 In most TBI cases, cognitive difficulties are the most prominent impairments 
79 [18]. However, little is known about how PTE can exaggerate neuropsychological 
80 deficits. Preclinical TBI studies provided scientific proof that PTE may further 
81 deteriorate cognitive performance. This comorbidity is also reported in other 
82 secondary epilepsies, such as post-stroke epilepsy [19]. Cognitive performance 
83 includes motor function, memory, intelligence, attention or concentration, processing 
84 speed, language, and visual-spatial abilities [20]. Researchers found that mice 
85 induced with TBI followed by electroconvulsive shock seizures performed worse in the 
86 Barnes maze than mice who underwent TBI alone [21]. These findings on the "double 
87 hit" mice were related to increased glial activation.
88 Meanwhile, in a study by Shultz and colleagues [22] comparing rats that did or 
89 did not develop PTE at six months post-injury, no significant differences between the 
90 groups were observed in Morris water maze (MWM) performance. The few preclinical 
91 studies that evaluated the impacts of PTE on cognitive functions have yielded 
92 inconclusive results. In some studies, temporal complexities may contribute to the 
93 failure to distinguish between PTE versus non-PTE groups of post-TBI rats. One 
It is made available under a CC-BY 4.0 International license . 
perpetuity. 
preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in
medRxiv preprint doi: https://doi.org/10.1101/2022.11.10.22282171; this version posted November 11, 2022. The copyright holder for this
5
94 previous study found an alarming statistic: half of the people with epilepsy exhibit 
95 cognitive dysfunctions in multiple domains, including learning, memory, attention, and 
96 executive functioning [23]. Memory loss, however, is far more common [24]. As a 
97 result, it is prudent to postulate that epilepsy resulting from TBI adds deficits to an 
98 individual with TBI, thereby worsening neuropsychological performance.
99 This study involves two parts. Part I of the study is to determine the etiology of 
100 PTE in TBI individuals to quantify the incidence rate and stratify risks for developing 
101 PTE. Part II of the study examines the cognitive comorbidity of PTE in a diverse multi102 ethnic population. Previously, several authors [5,6,13,25,26] had analysed predictors 
103 for PTE in a large sample of TBI people. Unfortunately, the geographical limitation 
104 prevents the broad generalisation of the results of these investigations. To the best of 
105 our knowledge, this study represents the first investigation in Malaysia with multi106 ethnicity and various demographical backgrounds who were retrospectively reviewed 
107 for up to 2 years after brain injury. The results of this study are intended to add to the 
108 emerging body of knowledge for both professionals and the general public. Individuals 
109 with TBI and PTE face unique challenges that healthcare providers will consider. Due 
110 to this, better services and strategies to enhance these people's lives could be created. 
111 Additionally, people with PTE and their families may be better prepared to deal with 
112 upcoming issues such as memory deficit, driving limits, intelligence, and the effect on 
113 caregivers, family members, and significant others. Furthermore, reducing the burden 
114 of comorbidities linked with decreased cognitive abilities and PTE would substantially 
115 impact the individual, society, and the economy.
116
117 Significance of the study
It is made available under a CC-BY 4.0 International license . 
perpetuity. 
preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in
medRxiv preprint doi: https://doi.org/10.1101/2022.11.10.22282171; this version posted November 11, 2022. The copyright holder for this
6
118 Despite a wealth of epidemiological data, there are still many questions regarding the 
119 epidemiology and determinants of outcomes of TBI patients with no definitive answers 
120 and no robust evidence base to guide their management. There is a relative lack of 
121 data on PTE in Malaysia. As a consequence, it has become an underappreciated 
122 condition. Patients with similar injuries may receive very different opinions about their 
123 likely prognosis and the best treatment, depending on their unit and the surgeon who 
124 treats them. Once a risk has been unequivocally associated with PTE, a logical final 
125 step is targeting those individuals with specific predictors in advocating a policy that 
126 aims to strategise the pharmacological or non-pharmacological treatment toward 
127 preventing epilepsy and its cognitive comorbidity. The findings can be used to plan 
128 and evaluate strategies to prevent further deterioration and guide the management of 
129 patients whom PTE has already developed, thereby helping to lighten the burden of ill 
130 health upon the individual and the community.
131
132 Aim and objectives
133 Study aim:
134 This project aims to describe the incidence and predictors for PTE among TBI in a 
135 tertiary care center in Malaysia and generate evidence on the cognitive burden of PTE.
136 Study objectives:
137 a. To estimate the incidence of PTE among TBI patients attending Universiti 
138 Kebangsaan Malaysia Medical Centre (UKMMC), Kuala Lumpur, between 
139 2019 to 2020.
140 b. To identify potential predictors of PTE among TBI patients in UKMMC, Kuala 
141 Lumpur.
It is made available under a CC-BY 4.0 International license . 
perpetuity. 
preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in
medRxiv preprint doi: https://doi.org/10.1101/2022.11.10.22282171; this version posted November 11, 2022. The copyright holder for this
7
142 c. To determine the association between PTE and cognitive impairment in the 
143 study population
144
145 Research Question
146 i. What is the incidence of PTE among the TBI patients attending UKMMC, Kuala 
147 Lumpur, between 2019 to 2020?
148 ii. What are the potential predictors of PTE among the TBI population in UKMMC, 
149 Kuala Lumpur?
150 iii. Does PTE correlate with cognitive impairment in TBI patients?
151
152 Materials and method
153 Study design and setting: 
154 This retrospective cohort study will take place in UKMMC, Kuala Lumpur, from March 
155 2022 until March 2024. All TBI patients admitted from 2019 to 2020 to UKMMC that 
156 meet the inclusion criteria will be retrieved from the hospital ward census. 
157 Inclusion criteria:
158 Inclusion criteria are (a) above 18 years old, (b) patient or guardian able to understand 
159 and speak English/Malay, and (c) with a history of admission to the Neurosurgery ward 
160 in UKMMC for traumatic brain injury caused between 2019 to 2020.
161 Exclusion criteria:
162 Patients are excluded if (a) the patient or guardian does not understand Malay/English, 
163 (b) declines to participate, (c) individuals with pre-existing epilepsy before TBI, (d) 
164 brain disorder that may cause epilepsy, e.g., brain tumor, stroke, meningitis, dementia. 
165
It is made available under a CC-BY 4.0 International license . 
perpetuity. 
preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in
medRxiv preprint doi: https://doi.org/10.1101/2022.11.10.22282171; this version posted November 11, 2022. The copyright holder for this
8
166 Sample and sampling:
167 We will approach All TBI patients meeting the inclusion criteria retrieved from the 
168 ward census.
169
170 a) Study Part I: Determine the incidence and predictors of PTE
171 Sample size calculation:
172 The reported incidence of PTE range from about 1% to 18.5% in year 1 [8,27–30]. 
173 Hence for a cumulative incidence of 18.5%, the sample size for 95% CI is 232 with an 
174 alpha error of 5% (https://www.openepi.com/SampleSize/SSPropor.htm). Assuming 
175 a drop out of 10%, thus 232+ 23= 255, rounded off to 260. Therefore, the TBI subjects 
176 to be recruited are 260. 
177
178 Data collection process:
179 The study flowchart is depicted in Fig 1. A TBI database will be formed based on the 
180 list. The patient's medical records in UKMMC will be retrieved from the medical record 
181 unit to find related demographic and clinical information, which are:
182  Demographic and basal variables: 
183 Age, gender, ethnicity, comorbid illness, and medical history focus on 
184 neurological conditions.
185
186  TBI characteristics during admission: 
187 Glasgow Coma Scale (GCS), loss of consciousness, vomiting, post-traumatic 
188 seizure (PTS), cranial surgery, and neuroimaging (looking for different 
189 aspects, such as subdural hematoma (SDH), subarachnoid hemorrhage 
It is made available under a CC-BY 4.0 International license . 
perpetuity. 
preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in
medRxiv preprint doi: https://doi.org/10.1101/2022.11.10.22282171; this version posted November 11, 2022. The copyright holder for this
9
190 (SAH), extradural hemorrhage (EDH), Intraparenchymal hemorrhage (IPH), 
191 midline shift, hydrocephalus, cerebral edema, and skull fracture).
192
193  Seizure and epilepsy evaluation: 
194 Seizures occurrence in-ward (if any), type, duration of seizure episode, EEG 
195 monitoring for evidence of epileptogenesis, epilepsy diagnosis, and anti196 seizure medication. 
197
198 An invitation letter that mentions the researcher's background, the study's aims, and 
199 the design will be sent to the patients' home addresses. Next, patients will be 
200 contacted by telephone to ascertain their epilepsy using a validated instrument 
201 adapted from a study by Ottman et al. [31] and the Malay version by Fong et al. [32]. 
202 Permission to use the original screening instruments has been obtained from both 
203 authors.
204 We made some modifications to the scoring system as below. We classified the 
205 response as positive having epilepsy following TBI if the subject answered as:
206 i. Q1 and Q3d: "No" or "don't know" and
207 ii. Any Q2, Q3a, b, c, e, f, g: "Yes" or "possible."
208 The following questions were also asked of patients with positive symptoms. 
209 Responses were recorded in a subject data sheet (APPENDIX III).
210 i. Before the TBI, did the patient experience seizures or epilepsy?
211 ii. Have the most recent seizure(s) occurred more than seven days after the 
212 TBI?
It is made available under a CC-BY 4.0 International license . 
perpetuity. 
preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in
medRxiv preprint doi: https://doi.org/10.1101/2022.11.10.22282171; this version posted November 11, 2022. The copyright holder for this
10
213 iii. Has the patient been diagnosed with epilepsy or a seizure disorder since 
214 the TBI diagnosis? If so, when was the date of diagnosis, and which 
215 clinician/hospital made the diagnosis?
216 iv. Any anti-seizure /anti-epileptic medication received?
217
218 All screened cases will be discussed with a consultant neurologist. Patients who 
219 declared having symptoms of PTE but have not sought any treatment determined by 
220 a record of follow-up in UKMMC will be directed to a neurologist for clinical observation 
221 and further treatment. Electroencephalography (EEG) will be used as an adjunct 
222 diagnostic tool. Then, patients who consented to cognitive assessment will be 
223 grouped into two, with PTE and without PTE. The cognitive assessment will be 
224 performed by our clinical psychologist and trained researcher at Neurology Clinic 
225 UKMMC or by home visit. 
226 Consent taking:
227 The consent-taking process will be done verbally at the beginning of the phone call 
228 (APPENDIX IV). First, the researcher will introduce herself and explain the purpose 
229 of the phone call. Then, the researcher will play a pre-recorded Patient Information 
230 Sheet (PIS) script to ensure transparency (APPENDIX V). The patient/ guardian will 
231 be given sufficient time to think and ask questions afterward. Next, verbal consent 
232 from the patient or guardian will be taken before the interview, and the process will be 
233 recorded. The phone call will be discontinued if no permission is given. Suppose the 
234 patient is not fit to be interviewed (e.g., hearing impairment or cognitively not fit to 
235 answer); in that case, the patient's guardian will be contacted. 
236
It is made available under a CC-BY 4.0 International license . 
perpetuity. 
preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in
medRxiv preprint doi: https://doi.org/10.1101/2022.11.10.22282171; this version posted November 11, 2022. The copyright holder for this
11
237 b) Study Part II: Cognitive assessments
238 Sample size calculation:
239 The next part of the study is the sub-analysis of cognitive performance. The ratio of 
240 individuals from Part I in the PTE group to individuals in the non-PTE group (as the 
241 control group) to be recruited is 1:3. The number of PTE patients depends upon the 
242 incidence within the sample of 260. For example, the incidence of PTE is 18.5% which 
243 means 48, so 48 subjects in the PTE group vs. 144 subjects in the non-PTE group. 
244 The control group will be age-matched and gender-matched with the study group.
245
246 Data collection process:
247 The following is the cognitive assessments and anxiety/depression test involved:
248 i. Addenbrooke's Cognitive Examination-III (ACE-III) test [33] with both 
249 English and Malay versions [34].
250 ii. Wechsler Adult Intelligence Scale (WAIS-IV) subtest: symbol search and 
251 coding [35].
252 iii. Comprehensive Trail Making Test (CTMT) [36,37]. 
253 iv. Depression, Anxiety and Stress Scale (DASS 21) [38,39].
254
255
256 Data analysis:
257 Statistical analysis will be performed using SPSS 28.0. Chi-square test or Fisher exact 
258 test will be conducted to compare categorical data, independent sample t-test for 
259 continuous variables, and ANOVA with repeated measures to study between subjects' 
260 factor group, within subjects' factor time, and post-doc tests with Bonferroni correction 
It is made available under a CC-BY 4.0 International license . 
perpetuity. 
preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in
medRxiv preprint doi: https://doi.org/10.1101/2022.11.10.22282171; this version posted November 11, 2022. The copyright holder for this
12
261 method. Pearson correlations and logistic regression analysis will determine if the 
262 variables are significantly related. The cognitive score difference between the two 
263 groups will be analysed using SPSS.
264
265 Ethics approval:
266 This study has been approved by the Research Ethics Committee, The National 
267 University of Malaysia (UKM PPI/111/8/JEP-2022-072).
268
269 Status and timeline of study:
270 The TBI patients' database has been formed, and participant recruitment is in 
271 progress. The study is expected to be completed by June 2023. The data analysis 
272 will be performed by July 2023, and the research article will be completed by 
273 September 2023.
274 Discussion:
275 This study will benefit a multidisciplinary team involved in managing TBI patients, like 
276 critical care physicians, neurosurgeons, neurologist rehabilitation specialists, and 
277 clinical psychologists. The results will help clinicians to understand the best way to 
278 treat TBI and PTE cases and serve as a source of information to support care 
279 decisions in designing therapies that would improve the quality of life of TBI patients.
280 There are several limitations of this study. It is a single-center study; thus, the 
281 sample may not represent all geographic regions of Malaysia. Nonetheless, to our 
282 knowledge, our study constitutes the first analysis of PTE incidence in Malaysia. As 
283 with the other retrospective study designs, the results will rely on whatever information 
284 is available in the medical record during data collection. Thus, there might be 
285 unrecorded or unavailable information. To deal with it, we will try to capture important 
It is made available under a CC-BY 4.0 International license . 
perpetuity. 
preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in
medRxiv preprint doi: https://doi.org/10.1101/2022.11.10.22282171; this version posted November 11, 2022. The copyright holder for this
13
286 details such as the patients' TBI history, education level, and any signs and symptoms 
287 of seizure/epilepsy during the phone interview session.
288 We anticipate that patients willing to come for cognitive assessment might be 
289 cognitively well, thus leading to the skewness of results. We offer a home visit for 
290 individuals with ambulatory problems to minimise bias.
291 Another challenge is that the contacts provided may no longer be in service. 
292 The previous study shows that some may not want to answer calls due to business or 
293 worry about answering unknown phone numbers [40]. To reduce this problem, we will 
294 send a letter to the individual's home address to mention our study purpose and action 
295 plan to contact them. In the letter, we will provide the team telephone numbers and a 
296 Q.R. code linked to Whatsapp so they can propose their preferred date and time for a 
297 phone interview. A token of RM20 will also be provided to cover the petrol cost of 
298 patients participating in the cognitive assessments.