AIM: To determine the prevalence and risks of suicidality in a group of patients with epilepsy.

METHODS: Included were 200 adult patients and 100 matched healthy subjects. The clinical interview using The Diagnostic and Statistical Manual of Mental Disorders (4th edition), Beck Depression Inventory (2nd edition) (BDI-II), Hamilton Anxiety Rating Scale (HAM-A), Yale-Brown Obsessive Compulsive Scale (Y-BOCS) and Eysenck Personality Questionnaire-Revised Rating Scale testings were used for diagnosis and assessment of severity of psychiatric symptoms. Blood concentrations of serotonin, catecholamines and dopamine were also measured.

RESULTS: Suicidality was reported in 35% (compared to 9% for controls), of them 80%, 72.86%, 55.71% and 52.9% had depression, anxiety, obsession and aggression respectively. Patients with suicidality had higher scores of BDI-II (P = 0.0001), HAM-A (P = 0.0001), and Y-BOCS (P = 0.037) and lower scores of psychotic (P = 0.0001) and extroversion (P = 0.025) personality traits. Regardless the presence or absence of suicidality, patients with epilepsy had low serotonin (P = 0.006), noradrenaline (P = 0.019) and adrenaline (P = 0.0001) levels. With suicidality, significant correlations were identified between: (1) age and scores of BDI-II (r = 0.235, P = 0.0001) and HAM-A (r = 0.241, P = 0.046); (2) age at onset and concentrations of noradrenaline (r = -0.502, P = 0.024); (3) duration of illness and scores of BDI-II (r = 0.247, P = 0.041), Y-BOCS (r = 0.270, P = 0.025) and neurotic personality trait (r = -0.284, P = 0.018); and (4) doses of antiepileptic drugs and scores of psychotic personality traits (r = -0.495, P = 0.006 for carbamazepine; r = -0.508, P = 0.0001 for valproate).

CONCLUSION: This is the first study which systematically estimated the prevalence and risks of suicidality in a homogenous group of patients with epilepsy. This study emphasizes the importance of epilepsy itself as a risk for suicidality and not its treatment.

Epilepsy is one of the most important and common chronic medical problems with prevalence rate of 8.2-12.9 per 1000 in general population[1]. Fortunately 70% of patients with epilepsy become seizure free on antiepileptic drugs (AEDs) while the remaining had medically refractory seizures (interactable epilepsy). In such patients, resection of the region of of the brain causing seizures is one approach of treatment[2]. The research literature indicates that people with epilepsy are at higher risk for suicide and suicidality (suicidal thoughts and attempts) with an estimated lifetime prevalence rate that varies between 3.3% and 14.3% or even up to 32.5%. This rate has been reported to be 6 to 25 times higher with temporal lobe epilepsy (TLE)[3,4] compared to 1.4%-6.9% in general population[5] and even higher for those who had temporal lobectomy[2]. Suicide accounts for > 10% of deaths with epilepsy compared with 1.4% in general population[6].

Suicidality in patients with epilepsy has complex and multifactorial aetiologies. Several factors have been identified as risks which include: male sex[6], early age at onset[7,8], TLE, high frequency of primary generalized seizures, polymorphic seizures combination, severe epilepsy, lateralization of the epileptic focus, recent control of seizures, absence of seizures for a long time especially after being very frequent, cognitive deterioration and hypofrontality and psychiatric comorbidity[6-9]. Retrospective studies indicate that the majority (81%-100%) of suicides occurs in subjects with psychiatric illnesses of which depression is the most common with a risk of 15%-18.9%, and even as high as 50 times than of general population. Others include anxiety, mood disorders and epileptoid personality and past or family histories of psychiatric disorders[6-9].

Recently, AEDs have been suggested as a risk of suicide with epilepsy[10,11]. The mechanisms of the negative psychotropic effects of AEDs are complex and vary dramatically between patients. It has been suggested to be related to the direct (i.e., anticonvulsant action) and indirect mechanisms of the drug action, particulary with rapid dose titration, slow drug metabolism, polypharmacy, drug-drug interactions, drug toxicity, drug withdrawal and metabolic derangements (as folate deficiency)[12]. Also non-pharmacological treatments as surgery[2] and vagus nerve stimulation[13] have been suggested to increase the risk of suicide up to 5-folds higher than pharmacological therapy. The concept of forced normalization (or alternative psychosis) in which a good control of seizures regardless control of EEG changes by pharmacological and non-pharmacological therapy results in appearance of behavioral abnormalities or psychosis, has been suggested as a cause of behavioral adverse effects or even suicid, although the exact mechanisms are not fully understood[14].

Despite the above information, the prevalnece of suicidality and the spectrum of comorbid psychiatric disorders with epilepsy tended to range greatly from one study to another. This reflects the heterogeneity of epilepsy and the differences in study designs and studied populations.

Early recognition and possible modification of clinical and psychosocial variables will have significant impact on the medical management and quality of life of patients with epilepsy.

To our knowledge, this is the first study in our population which aimed to estimate the prevelance and risks of suicidality in a homogenous group of patients with epilepsy.

The out-patient epilepsy clinic of Assiut University Hospital manages patients with epilepsy who were not covered by insurance service and all families were of low-income. The number patients of epilepsy of unknown etiology (idiopathic epilepsy) attending the clinic monthly for follow up is 601 (adults = 474 and children = 127). Clinical psychiatric interviewing was the primary method for examining the patients and in addition by applying the criteria for diagnosis of Diagnostic and Statistical Manual of Mental Health Disorders (4th edition) (DSM-IV)[15]. Accordingly, patients were divided into two groups: (1) those with no psychiatric symptoms (n = 107); and (2) those without psychiatric symptoms (n = 367). Also 100 healthy subjects matched for age, sex, educational level and socioeconomic states were included as controls for statistical comparisons. Healthy subjects were recruited from the general population. The patient’s seizure type was diagnosed according to Berg et al[16] Revised terminology and concepts for organization of seizures and epilepsies: report of The International League Against Epilepsy Commission on Classification and Terminology, 2005-2009”. Excluded were subjects with: (1) intelligence quotient < 70 as assessed by the Arabic version of Wechsler Adult Intelligence Scale Revised[17,18]; (2) other neurologic, systemic or medical diseases that result in psychiatric abnormalities; (3) psychiatric disorder prior to the onset of epilepsy; (4) regular medication(s) in addition to AEDs; and (5) alcoholism or substance abuse. The study protocol was approved by the ethical committee of The Faculty of Medicine of Assiut University and all participants gave their informed consent to participate in the study.

Neurological evaluation included information regarding seizure variables (as age at onset, precipitating factors, duration of illness, type and frequency of seizures, type of utilized AED(s), duration of treatment, degree of control on AEDs and family history of psychiatric illness. The frequencies of seizures were defined as described before[19] as follow: (1) very frequent: occurring several times a day or at intervals shorter than 7 d; (2) frequent: occurring at intervals longer than 7 d but shorter than 30 d; (3) occasional: occurring at intervals longer than 30 d but shorter than 1 year; and (4) rare: occurring at intervals longer than 1 year. Regarding the degree of control on AEDs, patients were considered controlled when seizure free for ≥ 1 year, partially controlled when seizure were occasional or rare and uncontrolled when seizures were frequent or very frequent.

Patients were divided into two groups according to the presence or absence of sucidality. The following Arabic translated questionnaires and scales were used for screening and assessment of severity of psychiatric symtoms and disorders: (1) Beck Depression Inventory (2nd edition) (BDI-II)[20,21]; (2) Hamilton Anxiety Rating Scale (HAM-A)[22,23]; (3) Yale-Brown Obsessive Compulsive Scale (Y-BOCS)[24,25]; and (4) Eysenck Personality Questionnaire-Revised Rating Scale (EPQ-R)[26,27].

BDI-II consists of 21 items and each corresponds to a symptom of depression is summed to give a single score. Accordingly, subjects were divided into: (1) those with no or minimal symptoms (scoreing: 0-13); (2) those with mild symptoms (scoring: 14-19); (3) those with moderate symptoms (scoring: 20-28); and (4) those with severe symptoms (scoring: 29-63). HAM-A consists of 14 items, each is scored on a 5-point scale, ranging from 0 (not present) to 4 (severe). Accordingly, subjects were divided into: (1) those with no anxiety symptoms (scoring: 0-13); (2) those with mild symptoms (scoring: 14-17); (3) those with moderate symptoms (scoring: 18-20); and (4) those with severe symptoms (scoring: 25-30). Y-BOCS is formed of two subscales: obsessions and compulsions. Each subscale consists of 5-items: (1) time spent in the symptoms; (2) subjective distress; (3) interference from symptoms; (4) resistance from symptoms; and (5) control over symptoms. Each item is scored from 0 (no symptoms) to 4 (extreme symptoms). The total Y-BOCS score ranges from a minimum of 0 to a maximum of 40. Accordingly, subjects are divided into: (1) those with no OCSs (scoring: 0-7); (2) those with mild symptoms (scoring: 8-15); (3) those with moderate symptoms (scoring: 16-23); (4) those with severe symptoms (scoring: 24-31); and (5) those with extreme symptoms (scoring: 32-40). EPQ-R consists of 4 scales: (1) Extroversion (E) which is formed of 23 items; (2) Neuroticism (N) which is formed of 24 items; (3) Psychoticism (P) which is formed of 32 items; and (4) social desirability or Lie (L) which is formed 21 items. Each scale is measured and has a dichotomous response format (yes or no).

Laboratory investigations: Subjects included in this study were controlled for timing of blood samples which were drawn at (8.00-10.00 am) after an overnight fast and patients were seizure free for at least 72 h (as any postictal central neurochemical dysfunction, is recognized to reverse within hours). The following laboratory tests were done: (1) Standardized which included: measurements of complete blood count and levels of creatinine, liver enzymes and fasting blood glucose; and (2) Specific which included: measurements of serum levels of serotonin and plasma levels of catecholamines (adrenaline and noradrenaline) and dopamine. The levels of serotonin, catecholamines and dopamine were measured by enzyme immunoassay method kits (IBL international GMBH, Hamburg, Germany). For confirmation, levels were assessed twice at two different days and combined with the cross sectional assessment while clinical evaluation and interviewing of included subjects. The serum levels of AEDs were measured as part of the investigation in batched assays in The Therapeutic Drug Monitoring laboratory of Assiut University hospital, Assiut, Egypt using Fluorescence Polarization Immunoassay System of Abbott, TDxFLX apparatus (Abbott Lab, Wiesbaden, Germany). The approximated reference therapeutic serum level of carbamazepine (CBZ) was 4-10 μg/mL and 50-100 μg/mL for valproate (VPA)[19].

This study included 200 adults (males = 100; females = 100) with epilepsy psychiatric symptoms. Their ages ranged between 20 to 48 years (mean: 30.47 ± 7.56 vs 29.14 ± 6.90 for control subjects; P = 0.229) and duration of illness ranged from 3 to 35 years (mean: 13.94 ± 7.24). The majority (62%) had complex partial or partial epilepsy with secondary generalization, while 38% had generalized tonic-clonic convulsions (GTC). Sixty patients (32.2%), 58 (46.77%) and 5 (4.03%) had frontal, temporal (TLE) and parietal lobe epilepsies. Seventy four (59.68%) had left sided foci of epileptic activity while 50 (40.32%) had right sided foci. The majority of patients were treated with CBZ (57%, n = 114 and/or VPA 19%, n = 38 and 24%, n = 48, respectively) for 2-30 years (mean: 9.28 ± 4.02). The majority of patients (54.0%, n = 108) were uncontrolled on AEDs, 23.5% (n = 47) were partially controlled and 22.5% (n = 45) were controlled.

Suicidatlity was reported in 35% (suicidal attempts = 23, 32.9%, suicidal thoughts = 47, 67.1%; male = 34, 17%; female = 36, 18%) compared to 9% of healthy subjects (suicidal thoughts = 9, 100%; male = 4, 4%; female = 5, 5%, Table 1). The majority of patients with suicidality (regardless of gender) had depression (80%). While anxiety, obsession, extroversion personality traits and aggression were reported in 72.86%, 55.71%, 21.5% and 52.9%, respectively (Tables 2 and 3). In general and regardless to the presence or absence of suicidality, patients with epilepsy had: (1) high scores of BDI-II, HAM-A and psychotic and neurotic personality traits regardless the type and focus and side of epilepsy and type of AEDs; (2) high scores of Y-BOCS particularly with frontal lobe epilepsy (P = 0.001) and left foci of epileptic activity (P = 0.015) and regardless to the type of AEDs; (3) the lack of control of seizures on AEDs was a risk for high scores of BDI-II (P = 0.0001), HAM-A (partially controlled: P = 0.02; uncontrolled: P = 0.0001) and Y-BOCS (partially controlled: P = 0.01; uncontrolled: P = 0.0001) (Table 4); (4) low serotonin concentrations particularly with GTC (P = 0.002) and TLE (P = 0.003) and with VPA (P = 0.009) and polytherapy (P = 0.007); (5) low catecholamines (noradrenaline and adrenaline) concentrations regardless the type, focus and side of epilepsy and type of AEDs; and (6) low dopamine concentrations particularly with GTC (P = 0.051) and lack of seizures control on AEDs (P = 0.004) (Table 5).

When patients with and without suicidality were compared, the former group had higher scores of BDI-II (P = 0.0001), HAM-A (P = 0.0001) and Y-BOCS (P = 0.037) and lower concentrations of serotonin (P = 0.006), noradrenaline (P = 0.019) and adrenaline (P = 0.0001), while the latter group had lower scores of psychotic (P = 0.0001), neurotic (P = 0.003) and extroversion (P = 0.025) personality traits (Tables 6 and 7).

In presence of suicidality, significant correlations were identified between: (1) age and scores of BDI-II (r = 0.235, P = 0.0001) and HAM-A (r = 0.241, P = 0.046); (2) age at onset and concentrations of noradrenaline (r = -0.502, P = 0.024); (3) duration of illness and scores of BDI-II (r = 0.247, P = 0.041), Y-BOCS (r = 0.270, P = 0.025) and neurotic personality trait (r = -0.284, P = 0.018); and (4) doses of AEDs and scores of psychotic personality traits (r = -0.495, P = 0.006 for CBZ; r = -0.508, P = 0.0001 for VPA). In absence of suicidality, significant correlations were identified between: (1) age at onset and scores of BDI-II (r = 0.350, P = 0.043) and serotonin concentrations (r = -0.432, P = 0.014); and (2) doses of AEDs and scores of psychotic (r = -0.271, P = 0.020) and neurotic (r = -0.348, P = 0.003) personality traits.

Several longitudinal and large sample-sized population studies with longer durations of follow-up, have shown that suicide is more frequent with epilepsy than in general population and represents the chief cause of death with epilepsy[3,4,6]. However, big differences in rates of suicidality have been shown across epilepsy studies. In this study, we reported high prevalence of suicidality (35% vs 9% for control subjects). In the literature, the estimated lifetime prevalence of suicide in epileptic population is 3.3%-14.3% or even up to 25%-32.5%[2,4] compared to 1.4%-6.9% for general population[5]. In contrast, some investigators did not find excess mortality with epilepsy due to suicide[28]. These great differences reflect the heterogeneity of epilepsy and epileptic disorders and differences in culture and risk factors across populations.

We and others reported a link between epilepsy, suicidality and comorbid interictal psychiatric abnormalities[4,9,29]. Depression was the most prevalent (80%) psychiatric comorbidity in patients with suicidality, followed by anxiety (72.86%), obsession (55.71%) and aggression (52.9%). Several studies reported that depression is the most prevalent interictal with epilepsy and frequently associated with emotional instability, poor impulse control, anxiety, obsession and aggression, reaching about 15%-18.9% or even as high as 50 times that of the general population[9,30,31]. Anxiety is the second most frequent psychiatric comorbiditiy with epilepsy and commonly associated with depression, inattention, aggression and obsession[32]. Obsession is the third most frequent psychiatric comorbiditiy with epilepsy and commonly associated with depression and anxiety, hostility, aggression, paranoid ideation, low intelligence, seizure frequency, duration of education, brain magnetic resonance imaging abnormalities and number of AEDs[33]. Male sex, increasing age, long duration of illness and lack of control on AEDs were the identified risks for higher scores of depression, anxiety and obsession[6,30,31]. It was observed that the earlier onset of seizures (< 18 years old), the better quality of life in adulthood. This reflects the more effective copying mechanisms and the adjustment to social and psychological consequences of the disease, while the late onset seizures, the more disruption of patient’s life (e.g., due to loss of the ability to drive and unemployment)[34].

It seems that the increase in the prevalence of suicidality with comorbid depression, anxiety and obsession in patients with epilepsy is due to the followings: (1) poor adjustment to seizures occur as a result of low socio-economic status, financial stress and poor cultural approach to epilepsy, the lower the expectations of parents and teachers, poor academic achievement, unemployment, inability to drive and diminished sexual desire and marital stresses. All result in low self-esteem, social isolation, stigmatization, depression, low quality of life and increase the risk of suicidality with epilepsy[34]; (2) misinformation about the disorder, results in low self-esteem, stigmatization, restrictions of normal daily activites, depression, obsessions, anxiety and increase the risk of suicidality with epilepsy[32]; and (3) giving in to obsessions and compulsions, knowing that they are unreal and incapacity to stop them, result in lack of self-confidence, depression, anxiety and even suicide[33].

We and others[35] observed no difference in the prevalence of suicidality and scores of psychiatric questionnaires and scales in relation to the dose, drug level and duration of treatment with AEDs. This indicates that the impact of epilepsy itself on suicidality is more significant than its medications. In contrast, suicidality was found in 4.3/1000 patients treated with AEDs in the active arm, compared with 2.2 per 1000 patients in the comparison arm as a result of a meta-analysis done by the Food and Drug Administration (FDA) on 199 randomized clinical trials and encompassed on 43 892 patients treated with different 11 AEDs for epilepsy, psychiatric disorders, pain and others. Which were: CBZ, VPA, felbamate, gabapentin, lamotrigine, levetracitam, oxcarbazepine, pregabalin, tiagabine, topiramate, and zonisamide. Thus, in January 2008 and based on this meta-analysis, FDA reported that exposure to AEDs increased the risk of suicidality by 1.80-fold. The FDA issued an alert and decided to insert suicide warnings in the package inserts of all AEDs[10]. However, the results of FDA meta-analysis must be considered with caution because of the followings[36]. (1) not all patients with epilepsy but some of them may be more susceptible to negative psychotropic effect of AEDs which may be resulted from the direct (i.e., anticonvulsant mechanisms) and/or the indirect effect of the AEDs as: (a) metaboolic adverse effects (folate deficiency); (b) interactions between the AED and the underlying epileptic process; (c) presence of hippocampal sclerosis; (d) forced normalization; (e) polytherapy; and (f) past or family history of psychiatric disorder; and (2) some AEDs (e.g., CBZ, VPA, topiramate and gabapentin) are well known mood stabilizers and are used in treatment of bipolar, impulse-control, borderline personality and episodic dyscontrol disorders. Thus patients with epilepsy should be closely followed whenever a new AED is introduced.

To summarize, it seems that the comorbidity between suicidality and epilepsy is correlated to epilepsy itself-related clinical, psychosocial and biological variables and not to its medications (AEDs). This is supported by the findings of this and other studies as follow: (1) the frequency of suicidality was associated with increase in seizures’ frequency, severity of seizures and intractability to medications. On the other hand, the frequency of depression and suicidality was associated with the type of epilepsy (e.g., GTC and frontal and temporal lobe epilepsies[37] and the lateralized foci of epileptic activity[38]; (2) lower blood concentrations of serotonin and catecholamines were identified in patients regardless to the presence or absence of suicidality[39,40]; (3) some investigators observed the followings: (a) the absence of a correlation between the severity of seizures and depression; (b) seizure-free state couldn’t protect patients from developing depression and consequently suicide; (c) improvement of psychiatric manifestations with disappearance of seizures after temporal lobectomy[41]; and (d) In fact: (i) some brain areas and their connections are incriminated in generating both epileptic discharges and psychiatric symptoms which include: frontal-limbic-subcortical circuits, frontal-striatal systems and limbic-brainstem, amygdale-hypothalamic and amygdale-locus coeruleus connections[37,38,42-44]; and (ii) disturbances in serotonin metabolism has been found by many investigators to be involved in the pathogenesis of suicide irrespective of primary diagnosis. Thus Serotonin could be a link between suicidality, epilepsy, depression, anxiety and obsession[45].

In conclusion, this is the first study in our population which systematically estimated the prevalence and risks of suicidality in a homogenous group of patients with epilepsy. This study emphasizes the importance of epilepsy itself as a risk for suicidality and not its treatment. Thus, early recognition of seizures, optimizing seizure control and managing its comorbid psychiatric symptomatology and disorders are of paramount importance while managing patients with epilepsy (i.e., biopsychosocial model of treatment which focuses on the whole person rather than simply treating convulsions). However and despite the importance of this study, its main limitation is that the majorities of patients had chronic and severe illness and were intractable to medications. This can be explained by the fact that patients were recruited from a University Hopital (a tertiary referral hospital). Thus performing multicenter studies will represent the accurate prevalence of suicidality with epilepsy of different severities.