JLE

Epileptic Disorders

MENU

Prevalence of psychiatric comorbidities in temporal lobe epilepsy: the value of structured psychiatric interviews Volume 12, numéro 4, December 2010

Auteur(s) : José Augusto Bragatti1, Carolina Machado Torres1, Renata Gomes Londero1, Juliana Bohn Assmann1, Vivian Fontana1, Kelin Cristine Martin1, Maria Paz Loayza Hidalgo2, Márcia Lorena Fagundes Chaves1, Marino Muxfeldt Bianchin1

1Division of Neurology
2Hospital de Clinicas de Porto Alegre, Division of Psychiatry, Porto Alegre, Brazil

Article reçu le 28 Mai 2009, accepté le 27 Octobre 2010

Epilepsy is a common neurological disorder. The world prevalence of epilepsy is estimated to range from 0.5 to 1.5% (Sander, 2003). The term “epilepsy” encompasses different neurological disorders characterized by a tendency towards recurrent epileptic seizures. Epileptic seizures are the clinical correlates of paroxysmal events generated by an enduring condition of hyperexcitability and hypersynchrony of brain electrical activity. The clinical spectrum of epilepsy encompasses many different neurobehavioural comorbidities (Elger and Schmidt, 2008). Epilepsy and neurobehavioural conditions may share some physiopathological, genetic, and environmental mechanisms (Gaitatzis et al., 2004; Hermann et al., 2008).

The association between epilepsy and psychiatric disorders has been known since ancient times, but has been marked by an explosion of studies during the last two decades (Devinsky, 2003). Prevalence of psychiatric comorbidities ranges from 20 to 40% in patients with epilepsy, and in selected populations may be two-fold higher (Pond and Bidwell, 1960; Silberman et al., 1994; Perini et al., 1996; Blumer et al., 1998; Davies et al., 2003; Swinkels et al., 2005; Tellez-Zenteno et al., 2005). Different definitions of psychiatric comorbidities, different study populations, and most importantly, different forms of psychiatric evaluation are factors that may explain the observed variability. Studies with structured psychiatric interviews are still lacking.

The objective of the present study was to determine the prevalence of major psychiatric disorders in a cohort of patients with TLE living in South Brazil using a structured psychiatric evaluation, and compare the findings with studies conducted around the world. Our study may therefore contribute to a better understanding of the worldwide prevalence of psychiatric comorbidities in epilepsy.

Methods

We studied a cohort of 98 consecutive Caucasian patients (59 women and 39 men) with TLE, from March 2007 to December 2008. Patients were selected from the epilepsy outpatient clinic at the Hospital de Clinicas de Porto Alegre, a tertiary hospital located in the Southern region of Brazil. Porto Alegre is the capital of Rio Grande do Sul state. The city has a population of 1,416,735, mostly composed of European immigrants (Portuguese, German, and Italian), distributed in an area of 496.8 km2. The annual per capita income is U$ 4840.91 (IBGE Cidades@ 2009). Economy is based on industry, commerce and services. In Brazil, health is the responsibility of the state and access is universal. For the rest of the country, health, education and safety are provided by both public and private services. It is estimated that about two thirds of the population uses governmental services.

The inclusion criterion for the study was the presence of electroclinical and neuroimaging features of TLE, according to the ILAE classification for epileptic seizures and syndromes (Commission on Classification and Terminology of the International League Against Epilepsy, 1989; Maillard et al., 2004; Pascual, 2007). Patients less than 18 years old or with generalised epilepsies, extratemporal epilepsies, mental retardation (IQ scores below 70), brain tumour, systemic disease (e.g. systemic erythematous lupus, AIDS), or penetrating head trauma were excluded.

After giving written informed consent, all patients participated in the Structured Clinical Interview for DSM-IV (SCID) (First et al., 2001) which was divided into six modules for the detection of one or more lifetime diagnoses using the Axis I Diagnostic and Statistical Manual, fourth edition (DSM-IV) (American Psychiatric Association, 2000). Inter-ictal spikes were independently reviewed by two board-certified electroencephalographers (JAB and CMT) who were blind to the psychiatric evaluation. Whenever the results were discordant, EEGs were reviewed by the two examiners together to reach a consensus. When available, all MRI examinations were reviewed to determine aetiology. Seizure control was assessed by an events calendar completed by the patient. Seizures occurring more than once a month were considered uncontrolled. Data regarding prior and current antiepileptic treatments, as well as the use of any psychotropic or sedative drug (e.g. antidepressants, antipsychotics), were registered in a database for statistical analyses.

Results were displayed in a percentage form. We analyzed the association of psychiatric diagnosis with the main aspects of TLE (control of seizures, inter-ictal EEG, MRI abnormalities, presence of aura). We compared results of patients with and without a positive SCID for these aspects using Pearson's chi-square test. All results were expressed using OR (95% CI). A significant level was considered when p < 0.05. The study was approved by the Ethics Committee of the Hospital de Clínicas de Porto Alegre.

Results

Mean age of the study population was 43.3 (± 12.3) years (range: 20-75 years), with a mean age at first seizure of 18.1 (± 14.3) years (range: 3-67 years) and a mean duration of epilepsy of 25.3 (± 12.9) years (range: 2-51 years). The main clinical characteristics of the study population are shown in table 1. Fifty-three patients (54.1%) were diagnosed with at least one lifetime psychiatric disorder. Forty-two of the SCID-positive patients (42.9%) had a mood disorder, 18 (18.4%) had an anxiety disorder, six (6.1%) had a psychotic disorder and six patients had alcohol or drug abuse (table 2). An association between mood and anxiety disorders was the most common psychiatric comorbidity observed, present in 22 patients (41.5%).

Major depression was the most frequent mood disorder observed in our series, present in 57% of the patients with mood disorders and in 25% of all patients. Dysthymic disorder was observed in 19% of patients with mood disorders (8% of all patients). A past depressive episode was reported in 14% of patients with mood disorders and in 6% of the total patient series. Generalised anxiety disorder was present in five patients (28% of patients with anxiety disorders and 5% of all patients), and panic disorder in six (two with agoraphobia). Post-traumatic stress disorder was observed in three patients (table 2).

Our findings were compared with other reports with different methodological characteristics and are presented in table 3. The overall prevalence of psychiatric disorders in our patients was 54%, with mood disorders twice as common as anxiety disorders (table 2). Relative to previous reports, these values are high, but similar to those observed in other European or South American studies which used similar structured interviews in populations of epileptic patients attending tertiary centres (Edeh and Toone, 1987; Araújo Filho et al., 2008). Moreover, we found a greater prevalence of psychiatric disorders in our patients when compared with the general population of Porto Alegre; data published in a previous study (Almeida-Filho et al., 1997) (table 4).

Although there was a tendency of patients with uncontrolled epilepsy to present some lifetime psychiatric disorder, we did not find any statistical difference between patients with and without lifetime psychiatric disorders, based on the presence of MRI abnormalities, inter-ictal EEG features, control of seizures, or presence of aura (table 5).

EEG and MRI features were analyzed separately in order to search for risk factors of psychiatric disorders in TLE patients. We categorized EEG data in only right, only left, and bilateral inter-ictal temporal discharges. Involvement of the left side (dominant hemisphere) was significant for lifetime mood disorders in TLE patients, with a risk of 7.1 (p = 0.007). Less than half of our patients had MRI scans (n = 43). There was no association between presence (uni- or bilateral) or absence of hippocampal sclerosis with lifetime psychiatric disorders, or specifically depression, in our patients.
Table 1 Clinical features of the patients studied.

Number of patients (%)

Gender

Men Women

39 (39.8%) 59 (60.2%)

Controlled seizures

Yes No

40 (40.8%) 58 (59.2%)

Aura

Yes No

64 (65.3%) 34 (34.7%)

Family history of epilepsy

Yes No

45 (45.9%) 53 (54.1%)

Family history of psychiatric disorders

Yes No

38 (38.8%) 60 (61.2%)

Initial precipitant insult

Yes No

24 (24.5%) 74 (75.5%)

EEG temporal focus lateralization

Right Left Not lateralized

32 (32.7%) 58 (59.2%) 8 (8.2%)

MRI

Normal Abnormal Not available

20 (20.4%) 23 (23.5%) 45 (45.9%)

Antiepileptic drugs

Monotherapy Polytherapy

47 (48.0%) 51 (52.0%)

Psychotropic drugs

No drugs One drug Combined therapy

78 (79.6%) 16 (16.3%) 4 (4.1%)


Table 2 DSM-IV Axis I psychiatric diagnoses.

Diagnosis

Number of patients (%)

Mood disorders

42 (42.9%)

Major depression

24 (24.5%)

Dysthymic disorder

8 (8.1%)

Past depressive episode

6 (6.1%)

Past manic episode

2 (2.0%)

Bipolar disorder

1 (1.0%)

Anxiety disorders

18 (18.4%)

Generalised anxiety disorder

5 (5.1%)

Panic disorder

4 (4.1%)

Post-traumatic stress disorder

3 (3.1%)

Panic with agoraphobia

2 (2.1%)

Specific phobia

2 (2.1%)

Obsessive compulsive disorder

2 (2.1%)

Psychotic disorders

6 (61.%)

Substance abuse

6 (61.%)


Table 3 Geographical distribution of psychiatric comorbidities in epilepsy.

Continent

Country

Authors

N

Instrument

Population

Psychiatric disorders

Mood disorders

Anxiety disorders

Psychosis

Substance abuse

North America

USA

Victoroff et al., 1994

60

SCID– DSM-III-R

TLE; candidates for surgery

70%

58.3%

31.7%

13.3%

-

Ettinger et al., 2004

775

CES-D

Epilepsy; community-based

-

36.5%

-

-

-

Strine et al., 2005

427

Kessler 6 scale

Epilepsy; community-based

-

32.6%

14.4%

-

-

Kobau et al., 2006

131

Health Style Survey (self-reported depression and anxiety)

Epilepsy; community-based

-

39%

39%

-

-

Canada

Tellez-Zenteno et al., 2007

253

CIDI

Epilepsy; community-based

23.5%

17.4%

12.8%

-

-

Europe

UK

Pond and Bidwell, 1960

245

Unstructured psychiatric interview

Children with epilepsy; community-based

29%

-

-

-

-

Graham and Rutter, 1970

63

Unstructured psychiatric interview

Children with epilepsy; community-based

28.6%

-

-

-

-

Edeh and Toone, 1987

88

CIS

Epilepsy; selected by general practitioners (GP)

48%

22%

15%

3.4%

-

Davies et al., 2003

67

SCID

Epilepsy; community-based

37%

-

-

-

-

Gaitatzis et al., 2004

5834

ICD-9

Epilepsy; selected from a database generated by GP

41%

18.2%

11.1%

9%

2.4%

Mensah et al., 2006

499

HADS

Epilepsy; from GP

-

11.2%

-

-

-

Italia

Perini et al., 1996

38

SADS, BDI, STAIX1, STAIX2

JME and TLE (selected) patients

80% (TLE), 22% (JME)

55% (TLE), 17% (JME)

15% (TLE), 11% (JME)

-

-

Netherlands

Swinkels et al., 2001

209

CIDI

Epilepsy; tertiary epilepsy centre

-

24.9%

29.7%

0.5%

20.1%

Czech Republic

Havlová, 1990

225

Chart review (unstructured)

Cohort of epileptic children

6.7%

-

-

-

-

Iceland

Gudmundsson, 1966

654

Clinical interview (unstructured)

Epilepsy (community-based)

54.5%

-

-

9%

-

Stefansson et al., 1998

241

ICD-9

Epileptic patients receiving benefits

35.3%

-

-

6.2%

5%

Sweden

Forsgren, 1992

713

Chart review (unstructured)

Epilepsy; community-based

5.9%

-

-

0.7%

-

Finland

Jalava and Sillanpaa, 1996

94

Chart review and ICD-9

Epilepsy; selected from different sources

24%

-

-

3.1%

-

Denmark

Bredkjaer et al., 1998

67

ICD-8

Epilepsy; community-based

16.8%

-

-

-

-

Asia

India

Hackett et al., 1998

26

ICD-10

Epilepsy; community-based

23.1%

-

-

-

-

Africa

Nigeria

Gureje, 1991

204

CIS

Epilepsy; tertiary centre

37%

-

-

30%

-

South America

Brazil

Araújo Filho et al., 2008

270

SCID

Refractory TLE and JME from a tertiary epilepsy centre

50% (TLE), 49% (JME)

25.8% (TLE), 19% (JME)

14.1% (TLE), 23% (JME)

15.8% (TLE), 3% (JME)

2% (JME)

Our study

98

SCID

TLE; selected from a tertiary epilepsy centre

54.1%

42.9%

18.4%

6.1%

6.1%


Table 4 Prevalence of psychiatric comorbidities in our patients and in Porto Alegre (Almeida-Filho et al., 1997).

Psychiatric diagnosis

TLE patients (n = 98)

General population (n = 6,471)

Overall

54.1%

42.5%

Mood disorders

42.9%

11.3%

Anxiety disorders

18.4%

9.6%

Psychotic disorders

6.1%

2.4%

Substance abuse

6.1%

9.2%


Table 5 Analysis of associations between psychiatric disorders and main clinical aspects of TLE.

SCID +

SCID -

Risk (95%CI)

p

Controlled epilepsy

20

18

Uncontrolled epilepsy

42

18

2.1 (0.9-4.9)

0.09

Unilateral temporal EEG spikes

32

19

Bilateral temporal EEG spikes

30

17

1.1 (0.5-2.4)

1.0

Abnormal MRI

12

11

Normal MRI

12

8

0.6 (0.2-1.6)

0.4

Presence of aura

43

22

Absence of aura

19

14

1.4 (0.6-3.4)

0.5

Discussion

We observed a high prevalence of lifetime psychiatric disorders in our TLE patients, of whom 54.1% presented with psychiatric comorbidities. The main psychiatric diagnoses found in our series were mood disorders (42 patients, 42.9% of the total), followed by anxiety disorders (18 patients, 18.4% of the total). Psychotic disorders and substance abuse were each observed in six patients (6.1%) (table 2).

Our results are consistent with the literature. Most reports show that mood disorders are the most frequent psychiatric comorbidity in TLE patients (Kanner, 2005; Schmitz, 2005). According to previous reports, a higher prevalence of psychiatric comorbidities was observed in epileptic patients studied at tertiary centres (40-60%) (Victoroff et al., 1994; Ring et al., 1998; Grabowska-Grzyb et al., 2006; Briellmann et al., 2007), while population-based studies showed an intermediate prevalence of about 20% (Edeh and Toone, 1987; Jacoby et al., 1996; Ettinger et al., 2004; Mensah et al., 2006). Nevertheless, in all studies the frequencies of psychiatric disorders among epileptic patients were higher than in the general population (12.2-16.2%) (Kessler et al., 2003; Hasin et al., 2005; Patten et al., 2006).

Several authors have reported a wide variability of psychiatric comorbidities in epileptic patients. The prevalence of these comorbidities varies according to the type of patient and psychiatric disorder studied, the duration of the study (last 12 months or lifetime), and the type of diagnostic procedure used (structured interview or self-applicable questionnaire) (Silberman et al., 1994; Perini et al., 1996; Blumer et al., 1998). For example, community-based studies of epileptic patients with structured interviews have identified prevalence of psychiatric comorbidities ranging from 23.5% to 37.5%, always higher than in the general population (10-20%). In contrast, studies using ICD diagnoses and data from administrative registries have shown more varied results (ranging from 16.8 to 60%) (Pond and Bidwell, 1960; Shukla et al., 1979; Jalava and Sillanpaa, 1996; Bredkjaer et al., 1998; Hackett et al., 1998; Stefansson et al., 1998). The highest prevalence was found in populations extracted from lists of individuals with some other associated disease, and therefore probably with a selection bias (Pond and Bidwell, 1960; Shukla et al., 1979).

The increase in prevalence of psychiatric disorders seems to be directly proportional to the increase in the severity of neurological disorders; from patients with chronic non-neurological diseases, non-epileptic neurological diseases, generalised epilepsies, extratemporal focal epilepsies, non-surgically treatable TLE through to patients eligible for surgery (Manchanda et al., 1996; Glosser et al., 2000; Wrench et al., 2004). The prevalence of psychiatric disorders in TLE patients is, in general, two-fold greater than in the general population (Tellez-Zenteno and Wiebe, 2008). Our data closely resemble data observed from other selected populations of epileptic patients (Davies et al., 2003; Araújo Filho et al., 2008). The high prevalence of psychiatric disorders in our study is therefore consistent with the finding that most of our patients (60%) did not have proper seizure control (table 1), as previously described (Almeida-Filho et al., 1997).

Another interesting aspect is the observation that studies conduced with structured interviews tend to point to higher frequencies of neuropsychiatric disorders in epilepsy (table 3). Because the use of structured psychiatric interviews is relatively more recent and limited to smaller populations, it is possible that larger epidemiological studies might underestimate the true prevalence of psychiatric disorders in epilepsy. Thus, further observations are necessary to clarify these matters.

For about 40% of our patients, we observed more than one type of lifelong psychiatric disorder; most frequently mood and anxiety disorders. Although this association has been recognized since ancient times, the pathophysiological mechanisms are still poorly understood (Temkin, 1971). Studies with adults and children suffering from epilepsy have shown a high prevalence of this comorbidity in association with epilepsy, sometimes up to 70% (Jones et al., 2005; Kobau et al., 2006). Depression, anxiety and epilepsy seem to share some biological and structural mechanisms related to limbic system dysfunction; an area of research intensely investigated over the last few years.

Fear is a frequent type of aura, observed in about 15% of TLE patients (Devinsky et al., 1995) which sometimes mimics panic attacks (Kanner et al., 2004). A previous study (Strine et al., 2005) found a high prevalence of post-ictal anxiety symptoms in epileptic patients. However, we could not observe this association because there were too few patients with anxiety symptoms in our sample. Goldstein et al. (1999) observed an inverse correlation between seizure frequency and post-ictal anxiety symptoms. The authors suggested that this inverse association might be caused by “habituation” of the anxiety generator circuits (mostly amygdala) due to high seizure frequency, causing seizures to be processed as ordinary events. Another possibility is that this inverse correlation could be due to the “learned helplessness” phenomenon (Hermann et al., 1996). Further research is needed to clarify these aspects.

There is much evidence suggesting that TLE and depression may share common pathogenic mechanisms (Kondziella et al., 2007). For example, in both TLE and depression small volumes of frontal lobes have been found (Lavretsky et al., 2007; Mueller et al., 2007). High-resolution MRI studies have shown that hippocampal volumes in depression are decreased bilaterally (Sheline, 2003) or in the left hippocampus only (Bremner et al., 2000). In TLE, volumes may be reduced at the site of seizure origin (Baxendale et al., 2005; Mueller et al., 2007) or, when combined with depression, reduced bilaterally (Baxendale et al., 2005). Nevertheless, 1H magnetic resonance spectroscopy (MRS) studies revealed reduced glutamate concentrations in the anterior cingulate cortex in depressed adults (Auer et al., 2000) and children (Mirza et al., 2004). In a study by Hasler et al. (2007), levels of glutamate/glutamine and GABA were also decreased in prefrontal dorsomedial and ventromedial regions. In TLE, most studies using inter-ictal fluorodeoxyglucose-positron emission tomography (FDG-PET) have confirmed hypometabolism of epileptogenic temporal regions (Manno et al., 1994) such as the hippocampus (Semah et al., 1995), often bilaterally (Joo et al., 2004; Kim et al., 2006). Indeed, orbitofrontal hypometabolism of glucose has been suggested to be a predisposing risk factor for the development of depression in patients with TLE (Salzberg et al., 2006). The relevant mechanisms may include extension of sclerosis and cell loss from the temporal lobe to extratemporal structures (Semah, 2002) or compensatory neuronal inhibition (Salzberg et al., 2006). Alternatively, orbitofrontal hypometabolism may occur secondary to depression or be merely a marker for general cerebral dysfunction associated with TLE (Salzberg et al., 2006).

A strong hypothesis derived from these data, which requires further research, is that neuronal hyperexcitability may be expressed either as impairment of emotions or seizure activity.

One limitation of our study was the inability to identify mood disorders not yet classified by DSM-IV (Krishnamoorthy et al., 2007). Another limitation was the cross-sectional design which did not permit us to identify psychiatric disorders temporally related to seizures (peri-ictal and inter-ictal symptoms). In these situations, mood disorders are different in epileptic patients when compared to subjects from the general population. There is increasing recognition of an association between epilepsy and an affective-somatoform disorder called “inter-ictal dysphoric disorder”. The main symptoms of this provisional psychiatric diagnosis are short temper and euphoria. Other less specific symptoms such as depression, pain, insomnia, fear and anxiety, also compose this newly recognised entity (Blumer et al., 2004). As is the case for mood disorders, there is also a common diagnosis of psychotic symptoms in epileptic patients not listed in DSM-IV. This disorder has been named “alternate psychosis”, a concept proposed by Tellenbach (1965) based on observations by Landolt (1953), and typifies a genuine psychosis of epilepsy (POE) due to its close relationship with epileptic activity. At times, ictal psychotic symptoms may be due to a focal non-convulsive status epilepticus, with continuous subclinical epileptiform activity involving one frontal or temporal lobe (Schmitz and Trimble, 2008). Post-ictal psychosis has a prevalence of about 7% in refractory epileptic patients (Tellez-Zenteno et al., 2007), especially when a double independent epileptic focus is present (Graham and Rutter, 1970). Landolt (1953) observed a paradoxical EEG normalisation in epileptic patients during the manifestation of psychotic symptoms, and called this phenomenon “Forced Normalization”. Finally, mood changes preceding (Blanchet and Frommer, 1986) or following the epileptic event (Kanner and Balabanov, 2002) are relatively frequent. As ictal phenomena, however, depression (Taylor and Lochery, 1987; Robertson, 1992) and mania (Barczak et al., 1988, Humphries and Dickinson, 1988) are much less frequent.

Although structured interviews are necessary for accurate determination of psychiatric diagnoses in epilepsy, their application in a busy clinical setting is not always feasible. Most rating scales and self-report questionnaires have been developed to screen for psychopathology in non-epileptic patients. Nevertheless, validated screening instruments (such as the Mood Disorder Questionnaire and NDDI-E) were specifically developed to screen for the presence of psychiatric disorders (especially mood disorders) in patients with epilepsy. These instruments are self-rating, can be completed in a few minutes, and can be used with confidence since the risk of overlap with adverse AED effects or pre-existing cognitive problems is minimised (Hirschfeld et al., 2000; Jones et al., 2005; Gilliam et al., 2006).

Conclusion

In our series of TLE patients, we have found a high prevalence of psychiatric disorders. The most frequent diagnoses were mood and anxiety disorders, which occurred simultaneously in 40% of patients. Our data are consistent with the literature and, in particular, similar to data from European studies (table 3). This is of particular interest since our studied population was composed of European descendents. Thus, the prevalence of psychiatric comorbidities in epileptics in a given population may remain similar to an ancestral population, independent of location in the world. This observation suggests that genetic predisposing factors may be more relevant than eventual environmental factors, an interesting aspect which merits further research. Moreover, with regards to the high prevalence of psychiatric comorbidities in TLE patients, our study is consistent with growing evidence in the literature indicating that TLE and psychiatric disorders share similar physiopathological mechanisms.Disclosure.

The present study was supported by Brazilian governmental funds (MS/CNPq/FAPERGS-06/2006/0615286 and CNPq 305501/2007-0, 504430/2008-4, and 481222/2008-1).

None of the authors has any conflict of interest to disclose.