- Email: [email protected]
Depression and schizophrenia in epilepsy: social and biological risk factors
Epilepsy Research 35 (1999) 59 – 68
Depression and schizophrenia in epilepsy: social and biological risk factors E. Bettina Schmitz a,*, Mary M. Robertson b, Michael R. Trimble b a
Neurologische Klinik und Poliklinik, Charite´, Campus Virchow Klinikum, Humboldt-Uni6ersita¨t, Augustenburger Platz 113353, Berlin, Germany b Raymond Way Neuropsychiatry Research Group, Institute of Neurology, Queen Square, London, UK Received 5 January 1998; received in revised form 30 November 1998; accepted 3 December 1998
Abstract Background: In a retrospective study we investigated the role of social and biological risk factors for the development of major depression and schizophreniform psychoses in epilepsy. We tested the hypotheses that social risk factors are associated with depression and biological risk factors are associated with schizophreniform psychoses. Method: We studied 25 patients with epilepsy and paranoid-hallucinatory psychosis, 25 patients with epilepsy and major depression, and 50 non-psychiatric epilepsy patients (controls) with respect to biological and psychosocial variables. Results: Schizophrenic patients had an earlier age of onset of epilepsy and a more severe epilepsy as characterised by history of status epilepticus, multiple seizure types and severity of seizures compared to non-psychiatric controls. Simple seizure symptoms were often vegetative and EEGs showed various abnormalities including temporal lobe discharges but no lateralisation to either side. With respect to antiepileptic drugs (AED) there were only few significant differences between groups: Polytherapy as well as treatment with phenytoin (DPH) was more frequent in psychotic patients as compared to non-psychiatric patients. Patients with psychoses were also characterized by a disturbed familial background, lack of interpersonal relationships, social dependency and professional failure. Depressive patients were significantly older than non-psychiatric controls and they suffered more frequently from focal epilepsies arising from the temporal lobe. They did not differ from controls with respect to severity of epilepsy. Treatment with valproate (VPA) was inversely linked with depression, suggesting that VPA may have prophylactic antidepressive properties in epilepsy patients. There were no psychosocial variables significantly linked with depression. Conclusions: In this study, patients with different forms of psychiatric complications in epilepsy could clearly be distinguished from controls. However, we could not confirm the simple hypothesis that there are biological predictors for schizophreniform psychoses and psychosocial predictors for major depression. Neurological and sociological variables seem linked with both, suggesting a multifactorial etiology. © 1999 Elsevier Science B.V. All rights reserved. Keywords: Epilepsy; Depression; Psychosis; Schizophrenia; Risk factors; Psychosocial; Biological
* Corresponding author. 0920-1211/99/$ - see front matter © 1999 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 0 - 1 2 1 1 ( 9 8 ) 0 0 1 2 9 - 6
60
E.B. Schmitz et al. / Epilepsy Research 35 (1999) 59–68
1. Introduction There is an increased risk for both depression and schizophreniform psychosis in epilepsy. The literature on risk factors for such complications has been highly controversial. With respect to psychoses many authors found a link with temporal lobe epilepsy (TLE) or more specifically limbic epilepsy (Trimble, 1991). Some studies have however failed to identify a significant bias towards TLE and have stressed the role of epilepsy severity (Schmitz and Wolf, 1995). With respect to depression the impact of biological variables is even less clear (Robertson et al., 1987). Hermann et al. (1990) have suggested that psychosocial complications are better predictors for depression than biological variables. It is possible that some of the controversies in the literature relate to selection bias and the insufficient consideration of both biological and social risk factors. We therefore conducted a study on both psychiatric complications and examined the hypothesis of Hermann et al. in relation to biological and psychosocial risk factors. Our hypothesis was that biological variables relate to schizophreniform psychosis and psychosocial risk factors relate to depression.
2. Method Retrospectively we compared biological and social variables in three patient groups: A first group comprising of 25 patients with epilepsy and a paranoid-hallucinatory psychosis according to the following DSMIV diagnostic categories (American Psychiatric Association, 1994): schizophrenia, schizophreniform disorder, delusional disorder, schizoaffective disorder, organic mental disorder (hallucinosis); a second group of 25 patients with epilepsy and major depression according to the diagnostic criteria of the DSMIV and a third group of non-psychiatric epilepsy control patients, who were selected by the next hospital number (in other words, the next epilepsy patient without comorbid psychopathology registered at the hospital). All patients were referred to the National Hospital for Neurology and Neuro-
surgery (NHNN), Queen Square, London, for management of epilepsy. The depressive as well as the psychotic epilepsy patients were personally seen by one of the authors during the psychiatric episode. All psychiatric patients were on a data base from the Department of Neuropsychiatry. They were examined either as inpatients or outpatients. We recorded biological variables which might be interlinked with the occurrence of psychiatric complications in epilepsy. EEG findings were extracted from all available reports. Seizure severity was estimated according to a modified version of the Chalfont Seizure Severity Scale (CSSS) for the dominating seizure type (Duncan and Sander, 1991). This scale considers features which patients and carers associate with seizure severity, such as duration and frequency of seizures, presence of a warning, falls, injuries, incontinence, etc. Seizure frequency was looked at with respect to the number of different seizure types in a 1-year period prior to two time points. The first being the onset of psychiatric disorder in the psychiatric group and onset of treatment at the NHNN in the non-psychiatric group. The second time point was when last seen in the clinic. For the estimation of the individual number of different seizure classes we distinguished simple partial, complex partial, generalised tonic clonic, and non-convulsive generalised seizures. We also compared the antiepileptic drugs prescribed prior to the onset of the psychiatric disorder. In the non-psychiatric control patients, we registered the drug medication when last seen in the clinic. We also examined the following biological variables: neurological examination, psychometry, anamnesis for brain damage (perinatal complications, complicated febrile convulsions, encephalitis, brain trauma, vascular accidents) and neuroradiology (existence and laterality of pathology on CCT or MRI, not type). With respect to psychosocial variables, because of the retrospective design, we focused on basic information relating to immigrant status, disturbed familial background (death of one parent or parental separation prior to the patient’s 18th year of life), education (school and professional qualification), work (employment status), housing
E.B. Schmitz et al. / Epilepsy Research 35 (1999) 59–68
61
Table 1 Demographic data Patients with psychosis (N =25)
Patients with depression (N= 25)
Non-psychiatric controls (N =50)
36.0 9 15.8**
24.9 912.4
42.5 910.5
47.8 9 14.6
39.4 915
13.2 910.9 36/64
11.4 9 9.0 72/28
12.3 99.9 56/44
Mean age of first visit (years) 23.9 910.6 Mean age at end of follow up Duration of follow up Females/males (%)
** PB0.01 (depressive patients versus controls).
(independent living) and interpersonal relationships (marital status, sexual relationships). Data were analysed using the SPSS PC version 7.0 for Windows. All statistics refer to the comparison between psychotic patients versus nonpsychiatric controls and between depressive patients versus non-psychiatric controls. For comparisons of means we performed Student t-tests. For comparisons of binominal data we performed x 2 tests or Fisher’s exact test when appropriate (expected frequency in at least one cell less than five).
3. Results
3.1. Psychiatric data 3.1.1. Patients with psychosis There was a male preponderance (not significant compared to controls). Mean follow up at the NHNN was 13.2 years (Table 1). The mean age at onset of psychosis was 29.3 years after a mean duration of epilepsy of 18.0 years. There was no history of schizophrenia-like illnesses in first degree relatives. Seven patients had a history of suicide attempts (Table 2). The psychiatric diagnosis was schizophrenia in six cases. Seven cases had delusional disorder, seven schizophreniform disorder, four had a schizoaffective disorder, and three had a hallucinosis (two patients had two diagnoses); 21 patients had delusions; 22 had hallucinations which were of auditory character in 19 patients. Nine patients had Schneiderian first rank symptoms.
Impairment of consciousness was noted in six cases and disorientation in seven. Catatonic symptoms occurred in only one patient. Six patients had a primarily chronic course (duration of at least two years). Only one patient had a single psychotic episode, 18 patients had recurrent episodes with complete remission in 14 and incomplete remission in four patients. The number of episodes ranged between one and 50 (mean 13.29 10.8). The mean duration was days in seven cases, weeks in eight cases and months in three cases (two unclear). There were only four patients without any etiological factors at any psychotic episode (Table 3). Another seven patients had both episodes with identifiable etiologies and episodes without identifiable etiologies. Twenty-one patients had EEGs during the psychotic episode which were unchanged to previous recordings in 15 patients, improved in two, deteriorated in two and showed a non-convulsive status epilepticus in the one case of an ictal psychosis. Acute stressful life events as defined by the DSM axis IV were relevant for the occurrence of psychosis in three cases (moderate, severe and extreme each), enduring problems in seven cases (moderate in two and severe in five).
3.1.2. Patients with depression Some of these patients have been analysed in a study on type and severity of depression in epilepsy (Robertson et al., 1987). There were more females in the depressive group than in the psychotic group. However, there were no significant sex differences when compared to controls. De-
E.B. Schmitz et al. / Epilepsy Research 35 (1999) 59–68
62 Table 2 Psychiatric data
Onset of psychiatric disorder (years) Duration of epilepsy Family history of depression (%)1 Family history of schizophrenia (%)1 History of pseudoseizures (%) History of suicide attempts (%)
Patients with psychosis (N = 25)
Patients with depression (N= 25)
Non-psychiatric controls (N =50)
29.39 12.0
36.49 14.9
–
18.09 11.4 8 –
18.2 9 14.1 4 –
– 4 –
0 28**
24* 20*
2 4
1
First degree relatives. * PB0.05 (depressive patients versus controls). ** PB0.01 (psychotic patients versus controls).
pressive patients were older as compared to controls when referred to the NHNN (Table 1). Follow up at the NHNN was 11.4 years. Age at first depressive episode was 36.4 years after a mean interval of epilepsy of 18.2 years. A family history of affective disorder was recorded in two cases (Table 2). Two patients had bipolar disorder; all others had monopolar depression. Severity of depression according to the DSMIV classification was mild in three cases, moderate in 10 cases, severe plus delusions or hallucinations in six and severe plus Table 3 Precipitating factors of psychiatric episodes related to seizures and AED Patients with psyPatients with depreschosis (%, N= 261) sion (%, N= 25) Preictal Ictal Postictal Parictal2 Alternative AED withdrawal AED intoxication No etiology
4 4 42 19 8 4
– – 12 12 24 –
4
4
15
48
1 One patient had psychotic episodes with different etiologies. 2 Parictal, gradual development of psychosis parallel to an increase in seizure frequency (Schmitz and Wolf, 1991).
mood-incongruent psychotic features in four. A history of pseudoseizures was recorded in six patients and suicide attempts in five patients. Five patients had a single depressive episode, 14 had recurrent episodes (with full remission in seven and incomplete remission in seven cases). Two patients had a primarily chronic course (unclear course in four cases). Duration of episodes was variable, weeks in eight patients and months in the remaining 17 cases. The number of episodes varied among patients between 1 and 100 (mean 11.49 9.0). Only 13 patients had an identifiable etiology related to seizure activity or AED (Table 3). Six patients had a significant decrease in seizure frequency during the depressive episode and were classified as alternative syndromes. Twenty-one patients had an EEG during the depressive episode which was unchanged in all but one case, in which the EEG had deteriorated. Acute life events were noted in five cases (moderate in four, extreme in one case) and enduring life events were relevant in another five cases (two moderate, two severe and one extreme).
3.2. Neurological 6ariables 3.2.1. Type of epilepsy A significantly high proportion of depressive patients had been diagnosed as TLE (Table 4). None of the depressive patients had frontal lobe epilepsy. In the psychotic group a relatively large proportion had an unclassifiable epilepsy
E.B. Schmitz et al. / Epilepsy Research 35 (1999) 59–68
63
Table 4 Electroclinical type of epilepsy
Temporal lobe epilepsy Frontal lobe epilepsy Other focal epilepsy Generalized epilepsy Generalized and focal epilepsy Unclassifiable epilepsy
Patients with psychosis (%, N =25)
Patients with depression (%, N= 25)
Non-psychiatric controls (%, N =50)
40 12 8 20 –
60* 0 12 28 –
34 12 2 36 4
20
–
12
* PB0.05 (depressive patients versus controls).
which was generalised in two and focal in three cases. When looking at simple seizure symptoms in more detail, vegetative auras such as epigastric rising were significantly more common in both psychiatric groups compared to controls (Table 5).
3.2.2. EEG EEG findings were extracted from all reported EEGs (Table 6). Psychotic patients had more EEGs than patients with depression (P B 0.05, t-test) which might account for the increased frequency of pathological findings in this group. There was more generalised slowing, multiple focal, temporal abnormalities and bilateral discharges in the psychosis group. There was no preponderance of either laterality in the psychosis group. 3.2.3. Se6erity of epilepsy Compared to controls, patients with schizophrenia-like psychosis were characterized by an earlier age of onset of epilepsy (Table 7). A history of status epilepticus (non-convulsive or convulsive) was more common in this group such as multiple seizure classes (three or more). The CSSS was highly significantly increased as compared to controls (P B 0.001). There were no differences between groups with respect to the frequency and type of symptomatic etiology of epilepsy, structural lesions as identified by CCT or MRI and verbal or performance IQ. When looking at seizure frequency at onset of the psy-
chiatric disorder and at the end of the follow up at the NHNN there were also no group differences.
3.3. Medication A significantly higher number of patients with schizophrenia-like psychosis received two or more AED. In accordance with the frequency of polytherapy, all AED were prescribed more frequently in psychotic patients compared to the control group. However, this was significant only for DPH. In depressive patients, a significantly low number were on VPA (Table 8).
3.4. Psychosocial 6ariables Although the level of school qualifications was similar as compared to controls, 28% of psychotic patients did not have a professional qualification and 88% were unemployed (Table 9). With respect to psychological variables, in the psychotic group there was a high frequency of severely disturbed paternal relationships. A minority had a relationship and about half of them had never had any sexual experiences with the opposite sex (Table 10). There were no negative psychosocial variables related to depression. In agreement with a relatively high proportion of subjects living independently, 80% of patients with depression had long-lasting relationships. This was however not significant when compared to controls.
E.B. Schmitz et al. / Epilepsy Research 35 (1999) 59–68
64 Table 5 Simple partial seizure symptoms
Vegetative Dysmnesic Olfactory/gustatory Dysphasic Auditory Visual Somato-sensory Motor Cognitive Affective
Patients with psychosis (%, N = 25)
Patients with depression (%, N= 25)
Non-psychiatric controls (%, N = 50)
36** 8 4 – 4 8 – 12 4 8
32* 16 8 12 4 12 – 4 4 12
8 8 2 8 4 2 2 20 2 –
* PB0.05 (depressive patients versus controls). ** PB0.01 (psychotic patients versus controls).
4. Discussion
4.1. Methodological considerations A retrospective study design has disadvantages and a prospective study would clearly be more appropriate in order to investigate risk factors for psychiatric complications of epilepsy. Especially psychosocial variables are insufficiently recorded in medical case notes. The advantage of the retrospective design is pragmatic and relates to the fact that schizophreniform psychoses as well as major depressive episodes are severe complications of epilepsy with a low incidence rate which makes prospective studies in a single center difficult. Patients were obviously not matched for age, sex, IQ, seizure type and severity because we do not know whether these are independent or dependent variables.
4.2. Psychiatric data It should be noted that there was a range of diagnoses and severity within our group of schizophrenia-related psychoses and only a minority of patients had pure interictal psychoses of the type identified in some earlier studies (Slater and Beard, 1963; Perez and Trimble, 1980). In most patients there were identifiable links between the psychiatric complications and seizure fre-
quency or treatment in at least one psychiatric episode. In accordance with other series of psychotic patients (Schmitz and Wolf, 1995) postictal psychoses were most common. The high rate of suicide attempts in the psychotic group underlines the potential danger of these complications. Only two patients had a diagnosis of bipolar depression which is in accordance with the suggested rarity of manic-depressive illness in epilepsy (Toone, 1981; Trimble, 1991). The later age of onset and the female preponderance as compared to psychotic patients are in agreement with the epidemiological data from endogenous monopolar depression (Gelder et al., 1989). Our data suggest an increased risk for the development of pseudoseizures in patients with depression which confirms other findings (Roy, 1979).
4.3. Neurological 6ariables The most significant link was between psychosis and seizure severity as quantified by the CSSS. Other variables which define severity of epilepsy such as early onset of epilepsy, status epilepticus and multiple seizure classes were also related to paranoid-hallucinatory psychosis. Because the number of recorded EEGs was relatively high in the psychotic group the excess of pathological findings may not be reliable. A patient-to-patient analysis was not carried out. However, the in-
E.B. Schmitz et al. / Epilepsy Research 35 (1999) 59–68
65
Table 6 EEG dataa Patients with psychosis (N= 25) Number of EEG recordings Diffuse slowing (%) Temporal focus (%) Multiple foci (%) Right laterality (%) Left laterality (%) Bilateral foci (%)
Patients with depression (N= 25) 2.0 9 0.8
3.49 2.2 52** 68** 40** 36 24 48**
20 24 28 28 28 32
Non-psychiatric controls (N = 50) 2.7 9 2.7 18 28 24 20 24 14
a
Focal abnormalities relate to epileptiform or unspecific discharges. ** PB0.01 (psychotic patients versus controls).
creased frequency of EEG abnormalities over the temporal lobe is supported by the finding of a higher proportion of vegetative auras in this group which suggests a limbic or mesial temporal seizure origin. We found a high proportion of bilateral abnormalities in our psychotic group, which has been described in previous studies (Kristensen and Sindrup, 1978). The mixture of psychotic patients, which included ictal and postictal cases, and contained only a minority with classical interictal schizophreniform illness, precludes us from examining the laterality hypothesis in any more detail in this group. There were no neurological variables linked with depression, except for the syndromatic diagnosis of TLE. This suggests that there may be consistent biological determinants and is in keeping with several other reports of associations between either TLE (Robertson et al., 1987) or complex partial seizures (Mendez et al., 1963) and affective disorders in epilepsy.
4.4. Medication It is well known that AED influence the mental state. In epilepsy patients, the main adverse effect is the link between AED and depression. It has been suggested that polytherapy, and treatment with barbiturates increase the risk for depression in epilepsy patients (Robertson et al., 1987; Brent et al., 1987). The effects of barbiturates are in
contrast to those seen with carbamazepine (CBZ). As has been demonstrated by several studies, treatment with CBZ is inversely correlated with depressive symptomatology in epilepsy patients (Andrewes et al., 1986; Robertson et al., 1987). We found an inverse link between VPA and depression, but not between CBZ and depression. The mood-stabilising properties of VPA have been clearly established in non-epileptic patients, but to date there is no evidence for positive psychotropic effects in epilepsy. In a comparison of 101 depressive patients to 202 non-depressive epilepsy patients, Mendez et al. (1963) also found a lower frequency of VPA prescriptions in the psychiatric group, in keeping with the data presented here. We could not confirm a link between treatment with primidone (PRM) or phenobarbitone (PHB) and depression. This might imply that psychiatric adverse effects of barbiturates (with respect to depression) have been overestimated in the recent literature. For schizophreniform psychosis in epilepsy, the role of AED is less clear. All AED have been associated with psychosis, either as an idiosyncratic dose-related effect or as a result of alternative psychosis with ‘forced normalisation’ of the EEG, a rare phenomenon paralleling seizure control (Landolt, 1958; Tellenbach, 1965). Only two of our patients had an alternative psychosis and this was related to CBZ plus PRM in one case and CBZ plus DPH in the other case. In the six
E.B. Schmitz et al. / Epilepsy Research 35 (1999) 59–68
66
patients with alternative depressive episodes the associated AED were: CBZ four cases, PHB two cases, PRM four cases, DPH three cases. The role of the drugs in these cases is unclear. Alternative psychoses are not usually described in association with standard anticonvulsant drugs, but more with either ethosuximide or some of the more powerful newer agents that readily switch off seizures in some patients with intractable epilepsy (Trimble, 1996). However, there are cases in the literature with all of the standard agents. The frequency noted in our study suggests this may be more common than usually reported.
Polytherapy was not linked with depression, but with schizophrenia-like psychosis. Patients with psychosis were characterized by complicated and severe epilepsies. Therefore, it is likely that polytherapy does not act as a primary risk factor. Of all AED, only DPH was significantly linked with psychosis. It is well known that some individuals develop a paranoid psychosis with high doses of DPH (McDanal and Bolman, 1975). Unfortunately, we did not record exact serum levels. We only distinguished between toxic, normal and subtherapeutic levels. Using these simple categories we did not find intergroup differences.
Table 7 Severity of epilepsy
Age at onset of epilepsy (years) Symptomatic etiology (%) Structural lesions CT/MRI (%) Status epilepticus (%) Multiple seizure classes (3 or more, %) Verbal IQ Performance IQ Severity of seizures1
Patients with psychosis (N= 25)
Patients with depression (N= 25)
Non-psychiatric controls (N =50)
11.3 9 10.4* 48 44 36* 44*
18.2 912.9 40 40 4 16
18.3 914.6 34 34 10 18
91.7 9 17.0 90.69 16.9 33.59 18.9***
96.0 911.3 95.8 912.8 22.0 913.2
87.4 917.8 87.4 921.3 16.9 9 13.8
1
Mean score, based on the Chalfont Seizure Severity Scale (Duncan and Sander, 1991). * PB0.05 (psychotic patients versus controls). *** PB0.001 (psychotic patients versus controls). Table 8 Polytherapy and antiepileptic drugs
Polytherapy Barbiturates Primidone Phenytoin Ethosuximide Carbamazepine Valproate Benzodiazepines Vigabatrin
Patients with psychosis (%, N = 25)
Patients with depression (%, N =25)
Non-psychiatric controls (%, N =50)
80** 16 36 72** – 48 28 4 –
56 28 20 68 – 44 4* 8 –
46 28 12 48 6 46 28 10 2
* PB0.05 (depressive patients versus controls). ** PB0.01 (psychotic patients versus controls).
E.B. Schmitz et al. / Epilepsy Research 35 (1999) 59–68
67
Table 9 Social variables
Immigrant status School qualification Professional qualification Unemployment Independent living
Patients with psychosis (%, N =25)
Patients with depression (%, N =25)
Non-psychiatric controls (%, N= 50)
16 36 28* 88* 56
4 36 40 72 84
20 18 44 54 62
* PB0.05 (psychotic patients versus controls).
Table 10 Psychological variables
Divorce or death of parents No sexual experiences Long-lasting relationships
Patients with psychosis (%, N =25)
Patients with depression (%, N =25)
Non-psychiatric controls (%, N= 50)
28** 56** 36***
16 12 80
8 10 60
** PB0.01 (psychotic patients versus controls). *** PB0.001 (psychotic patients versus controls).
4.5. Psychosocial 6ariables
5. Conclusions
Acute or enduring stressful life events prior to the psychiatric complication were noted in 40% in each group. In psychotic patients there was a high rate of severely disturbed paternal relationships. It is interesting that achievement at school was not worse in the psychotic group, but there was a significant proportion of subjects who did not complete any professional training and who became unemployed which perhaps stresses the social sequelae of recurrent psychosis rather than risk factors. Although only six patients had a chronic psychosis, the striking lack of sexual experiences and lasting relationships suggests severe psychosocial impairment exceeding the recognised psychotic episodes. With respect to depressive episodes we did not identify any specific psychological or social risk factors. It is however possible that we failed to detect any predictive variables due to the poor sensitivity of a retrospective method.
Our retrospective data on the relationship between biological and social variables and psychiatric complications in epilepsy must be carefully interpreted. The correlations between some of the variables and depression or schizophrenia-related psychoses are unlikely to reflect simple etiological relationships. However, it seems that patients who develop psychoses are characterised by a combination of biological and psychosocial risk factors which relate to an early onset severe epilepsy involving the limbic system in addition to an unfavourable familial background. Depressive patients are not as easy to distinguish on neurological factors or severe psychosocial problems, with the exception of an association to TLE. Our working hypothesis for this study was that affective and psychotic disorders in epilepsy had defined etiological and pathological relationships that would be different. In fact, combinations of
68
E.B. Schmitz et al. / Epilepsy Research 35 (1999) 59–68
psychosocial and neurological variables seem linked with both, and simple ideas regarding the psychological as opposed to biological associations with depression, or the development of psychoses in epilepsy are not supported by this study. This is similar to the situation without epilepsy, where simpler divisions between depression and psychotic illnesses in relation to underlying neurological abnormalities and social precipitants have succumbed to the advances of investigations of the nervous system.
References American Psychiatric Association, 1994. Diagnostic and Statistical Manual of Mental Disorders. American Psychiatric Press, Washington, DC. Andrewes, D.G., Bullen, J.G., Tomlinson, L., et al., 1986. A comparative study of cognitive effects of phenytoin and carbamazepine in new referrals with epilepsy. Epilepsia 26, 128 – 134. Brent, D.A., Crumrine, P.K., Varma, R.R., et al., 1987. Phenobarbital treatment and major depressive disorder in children with epilepsy. Pediatrics 80, 909–917. Duncan, J.S., Sander, W.A.S., 1991. The Chalfont Seizure Severity Scale. J. Neurol. Neurosurg. Psychiatry 54, 873– 876. Gelder, M., Gath, D., Mayou, R., 1989. Oxford Textbook of Psychiatry. Oxford University Press, Oxford. Hermann, B.P., Whitman, S., Wyler, A.R., et al., 1990. Psychosocial predictors of psychopathology in epilepsy. Br. J. Psychiatry 156, 98 – 105. Kristensen, O., Sindrup, H.H., 1978. Psychomotor epilepsy and
.
psychosis. II Electroencephalographic findings. Acta Neurol. Scand. 57, 361 – 369. Landolt, H., 1958. Serial electroencephalographic investigations during psychotic episodes in epileptic patients and during schizophrenic attacks. In: Lorentz de Haas, A.M. (Ed.), Lectures on Epilepsy. Elsevier, Amsterdam, pp. 91 – 131. McDanal, C.E., Bolman, W.M., 1975. Delayed idiosyncratic psychosis with diphenylhydantoin. J. Am. Med. Assoc. 231, 1063. Mendez, M.F., Doss, R.C., Taylor, J.L., et al., 1963. Depression in epilepsy. Relationship to seizures and anticonvulsant therapy. J. Nerv. Ment. Dis. 181, 444 – 447. Perez, M.M., Trimble, M.R., 1980. Epileptic psychoses — comparison with process schizophrenia. Br. J. Psychiatry 137, 245 – 249. Robertson, M.M., Trimble, M.R., Townsend, H.R.A., 1987. Phenomenology of depression in epilepsy. Epilepsia 28, 364 – 372. Roy, A., 1979. Some determinants of affective symptoms in epileptics. Can. J. Psychiatry 24, 554 – 556. Schmitz, B., Wolf, P., 1991. Epilepsy and psychosis. In: Theodore, W., Devinsky, O. (Eds.), Epilepsy and Behaviour. Frontiers in Clinical Neurosciences, vol. 12. Wiley, New York. Schmitz, B., Wolf, P., 1995. Psychosis with epilepsy: frequency and risk factors. J. Epilepsy 8, 295 – 305. Slater, E., Beard, A.W., 1963. The schizophrenia-like psychoses of epilepsy. V. Discussion and conclusions. Br. J. Psychiatry 109, 143 – 150. Tellenbach, H., 1965. Epilepsie als Anfallsleiden und als Psychose. Nervenarzt 36, 190 – 202. Toone, B.K., 1981. Psychoses of Epilepsy. In: Reynolds, E.H., Trimble, M.R. (Eds.), Epilepsy and Psychiatry. Churchill Livingstone, Edinburgh, pp. 113 – 137. Trimble, M.R., 1991. The Psychoses of Epilepsy. Raven Press, New York. Trimble, M.R., 1996. Biological Psychiatry. Wiley, Chichester.
Depression and schizophrenia in epilepsy: social and biological risk factors E. Bettina Schmitz a,*, Mary M. Robertson b, Michael R. Trimble b a
Neurologische Klinik und Poliklinik, Charite´, Campus Virchow Klinikum, Humboldt-Uni6ersita¨t, Augustenburger Platz 113353, Berlin, Germany b Raymond Way Neuropsychiatry Research Group, Institute of Neurology, Queen Square, London, UK Received 5 January 1998; received in revised form 30 November 1998; accepted 3 December 1998
Abstract Background: In a retrospective study we investigated the role of social and biological risk factors for the development of major depression and schizophreniform psychoses in epilepsy. We tested the hypotheses that social risk factors are associated with depression and biological risk factors are associated with schizophreniform psychoses. Method: We studied 25 patients with epilepsy and paranoid-hallucinatory psychosis, 25 patients with epilepsy and major depression, and 50 non-psychiatric epilepsy patients (controls) with respect to biological and psychosocial variables. Results: Schizophrenic patients had an earlier age of onset of epilepsy and a more severe epilepsy as characterised by history of status epilepticus, multiple seizure types and severity of seizures compared to non-psychiatric controls. Simple seizure symptoms were often vegetative and EEGs showed various abnormalities including temporal lobe discharges but no lateralisation to either side. With respect to antiepileptic drugs (AED) there were only few significant differences between groups: Polytherapy as well as treatment with phenytoin (DPH) was more frequent in psychotic patients as compared to non-psychiatric patients. Patients with psychoses were also characterized by a disturbed familial background, lack of interpersonal relationships, social dependency and professional failure. Depressive patients were significantly older than non-psychiatric controls and they suffered more frequently from focal epilepsies arising from the temporal lobe. They did not differ from controls with respect to severity of epilepsy. Treatment with valproate (VPA) was inversely linked with depression, suggesting that VPA may have prophylactic antidepressive properties in epilepsy patients. There were no psychosocial variables significantly linked with depression. Conclusions: In this study, patients with different forms of psychiatric complications in epilepsy could clearly be distinguished from controls. However, we could not confirm the simple hypothesis that there are biological predictors for schizophreniform psychoses and psychosocial predictors for major depression. Neurological and sociological variables seem linked with both, suggesting a multifactorial etiology. © 1999 Elsevier Science B.V. All rights reserved. Keywords: Epilepsy; Depression; Psychosis; Schizophrenia; Risk factors; Psychosocial; Biological
* Corresponding author. 0920-1211/99/$ - see front matter © 1999 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 0 - 1 2 1 1 ( 9 8 ) 0 0 1 2 9 - 6
60
E.B. Schmitz et al. / Epilepsy Research 35 (1999) 59–68
1. Introduction There is an increased risk for both depression and schizophreniform psychosis in epilepsy. The literature on risk factors for such complications has been highly controversial. With respect to psychoses many authors found a link with temporal lobe epilepsy (TLE) or more specifically limbic epilepsy (Trimble, 1991). Some studies have however failed to identify a significant bias towards TLE and have stressed the role of epilepsy severity (Schmitz and Wolf, 1995). With respect to depression the impact of biological variables is even less clear (Robertson et al., 1987). Hermann et al. (1990) have suggested that psychosocial complications are better predictors for depression than biological variables. It is possible that some of the controversies in the literature relate to selection bias and the insufficient consideration of both biological and social risk factors. We therefore conducted a study on both psychiatric complications and examined the hypothesis of Hermann et al. in relation to biological and psychosocial risk factors. Our hypothesis was that biological variables relate to schizophreniform psychosis and psychosocial risk factors relate to depression.
2. Method Retrospectively we compared biological and social variables in three patient groups: A first group comprising of 25 patients with epilepsy and a paranoid-hallucinatory psychosis according to the following DSMIV diagnostic categories (American Psychiatric Association, 1994): schizophrenia, schizophreniform disorder, delusional disorder, schizoaffective disorder, organic mental disorder (hallucinosis); a second group of 25 patients with epilepsy and major depression according to the diagnostic criteria of the DSMIV and a third group of non-psychiatric epilepsy control patients, who were selected by the next hospital number (in other words, the next epilepsy patient without comorbid psychopathology registered at the hospital). All patients were referred to the National Hospital for Neurology and Neuro-
surgery (NHNN), Queen Square, London, for management of epilepsy. The depressive as well as the psychotic epilepsy patients were personally seen by one of the authors during the psychiatric episode. All psychiatric patients were on a data base from the Department of Neuropsychiatry. They were examined either as inpatients or outpatients. We recorded biological variables which might be interlinked with the occurrence of psychiatric complications in epilepsy. EEG findings were extracted from all available reports. Seizure severity was estimated according to a modified version of the Chalfont Seizure Severity Scale (CSSS) for the dominating seizure type (Duncan and Sander, 1991). This scale considers features which patients and carers associate with seizure severity, such as duration and frequency of seizures, presence of a warning, falls, injuries, incontinence, etc. Seizure frequency was looked at with respect to the number of different seizure types in a 1-year period prior to two time points. The first being the onset of psychiatric disorder in the psychiatric group and onset of treatment at the NHNN in the non-psychiatric group. The second time point was when last seen in the clinic. For the estimation of the individual number of different seizure classes we distinguished simple partial, complex partial, generalised tonic clonic, and non-convulsive generalised seizures. We also compared the antiepileptic drugs prescribed prior to the onset of the psychiatric disorder. In the non-psychiatric control patients, we registered the drug medication when last seen in the clinic. We also examined the following biological variables: neurological examination, psychometry, anamnesis for brain damage (perinatal complications, complicated febrile convulsions, encephalitis, brain trauma, vascular accidents) and neuroradiology (existence and laterality of pathology on CCT or MRI, not type). With respect to psychosocial variables, because of the retrospective design, we focused on basic information relating to immigrant status, disturbed familial background (death of one parent or parental separation prior to the patient’s 18th year of life), education (school and professional qualification), work (employment status), housing
E.B. Schmitz et al. / Epilepsy Research 35 (1999) 59–68
61
Table 1 Demographic data Patients with psychosis (N =25)
Patients with depression (N= 25)
Non-psychiatric controls (N =50)
36.0 9 15.8**
24.9 912.4
42.5 910.5
47.8 9 14.6
39.4 915
13.2 910.9 36/64
11.4 9 9.0 72/28
12.3 99.9 56/44
Mean age of first visit (years) 23.9 910.6 Mean age at end of follow up Duration of follow up Females/males (%)
** PB0.01 (depressive patients versus controls).
(independent living) and interpersonal relationships (marital status, sexual relationships). Data were analysed using the SPSS PC version 7.0 for Windows. All statistics refer to the comparison between psychotic patients versus nonpsychiatric controls and between depressive patients versus non-psychiatric controls. For comparisons of means we performed Student t-tests. For comparisons of binominal data we performed x 2 tests or Fisher’s exact test when appropriate (expected frequency in at least one cell less than five).
3. Results
3.1. Psychiatric data 3.1.1. Patients with psychosis There was a male preponderance (not significant compared to controls). Mean follow up at the NHNN was 13.2 years (Table 1). The mean age at onset of psychosis was 29.3 years after a mean duration of epilepsy of 18.0 years. There was no history of schizophrenia-like illnesses in first degree relatives. Seven patients had a history of suicide attempts (Table 2). The psychiatric diagnosis was schizophrenia in six cases. Seven cases had delusional disorder, seven schizophreniform disorder, four had a schizoaffective disorder, and three had a hallucinosis (two patients had two diagnoses); 21 patients had delusions; 22 had hallucinations which were of auditory character in 19 patients. Nine patients had Schneiderian first rank symptoms.
Impairment of consciousness was noted in six cases and disorientation in seven. Catatonic symptoms occurred in only one patient. Six patients had a primarily chronic course (duration of at least two years). Only one patient had a single psychotic episode, 18 patients had recurrent episodes with complete remission in 14 and incomplete remission in four patients. The number of episodes ranged between one and 50 (mean 13.29 10.8). The mean duration was days in seven cases, weeks in eight cases and months in three cases (two unclear). There were only four patients without any etiological factors at any psychotic episode (Table 3). Another seven patients had both episodes with identifiable etiologies and episodes without identifiable etiologies. Twenty-one patients had EEGs during the psychotic episode which were unchanged to previous recordings in 15 patients, improved in two, deteriorated in two and showed a non-convulsive status epilepticus in the one case of an ictal psychosis. Acute stressful life events as defined by the DSM axis IV were relevant for the occurrence of psychosis in three cases (moderate, severe and extreme each), enduring problems in seven cases (moderate in two and severe in five).
3.1.2. Patients with depression Some of these patients have been analysed in a study on type and severity of depression in epilepsy (Robertson et al., 1987). There were more females in the depressive group than in the psychotic group. However, there were no significant sex differences when compared to controls. De-
E.B. Schmitz et al. / Epilepsy Research 35 (1999) 59–68
62 Table 2 Psychiatric data
Onset of psychiatric disorder (years) Duration of epilepsy Family history of depression (%)1 Family history of schizophrenia (%)1 History of pseudoseizures (%) History of suicide attempts (%)
Patients with psychosis (N = 25)
Patients with depression (N= 25)
Non-psychiatric controls (N =50)
29.39 12.0
36.49 14.9
–
18.09 11.4 8 –
18.2 9 14.1 4 –
– 4 –
0 28**
24* 20*
2 4
1
First degree relatives. * PB0.05 (depressive patients versus controls). ** PB0.01 (psychotic patients versus controls).
pressive patients were older as compared to controls when referred to the NHNN (Table 1). Follow up at the NHNN was 11.4 years. Age at first depressive episode was 36.4 years after a mean interval of epilepsy of 18.2 years. A family history of affective disorder was recorded in two cases (Table 2). Two patients had bipolar disorder; all others had monopolar depression. Severity of depression according to the DSMIV classification was mild in three cases, moderate in 10 cases, severe plus delusions or hallucinations in six and severe plus Table 3 Precipitating factors of psychiatric episodes related to seizures and AED Patients with psyPatients with depreschosis (%, N= 261) sion (%, N= 25) Preictal Ictal Postictal Parictal2 Alternative AED withdrawal AED intoxication No etiology
4 4 42 19 8 4
– – 12 12 24 –
4
4
15
48
1 One patient had psychotic episodes with different etiologies. 2 Parictal, gradual development of psychosis parallel to an increase in seizure frequency (Schmitz and Wolf, 1991).
mood-incongruent psychotic features in four. A history of pseudoseizures was recorded in six patients and suicide attempts in five patients. Five patients had a single depressive episode, 14 had recurrent episodes (with full remission in seven and incomplete remission in seven cases). Two patients had a primarily chronic course (unclear course in four cases). Duration of episodes was variable, weeks in eight patients and months in the remaining 17 cases. The number of episodes varied among patients between 1 and 100 (mean 11.49 9.0). Only 13 patients had an identifiable etiology related to seizure activity or AED (Table 3). Six patients had a significant decrease in seizure frequency during the depressive episode and were classified as alternative syndromes. Twenty-one patients had an EEG during the depressive episode which was unchanged in all but one case, in which the EEG had deteriorated. Acute life events were noted in five cases (moderate in four, extreme in one case) and enduring life events were relevant in another five cases (two moderate, two severe and one extreme).
3.2. Neurological 6ariables 3.2.1. Type of epilepsy A significantly high proportion of depressive patients had been diagnosed as TLE (Table 4). None of the depressive patients had frontal lobe epilepsy. In the psychotic group a relatively large proportion had an unclassifiable epilepsy
E.B. Schmitz et al. / Epilepsy Research 35 (1999) 59–68
63
Table 4 Electroclinical type of epilepsy
Temporal lobe epilepsy Frontal lobe epilepsy Other focal epilepsy Generalized epilepsy Generalized and focal epilepsy Unclassifiable epilepsy
Patients with psychosis (%, N =25)
Patients with depression (%, N= 25)
Non-psychiatric controls (%, N =50)
40 12 8 20 –
60* 0 12 28 –
34 12 2 36 4
20
–
12
* PB0.05 (depressive patients versus controls).
which was generalised in two and focal in three cases. When looking at simple seizure symptoms in more detail, vegetative auras such as epigastric rising were significantly more common in both psychiatric groups compared to controls (Table 5).
3.2.2. EEG EEG findings were extracted from all reported EEGs (Table 6). Psychotic patients had more EEGs than patients with depression (P B 0.05, t-test) which might account for the increased frequency of pathological findings in this group. There was more generalised slowing, multiple focal, temporal abnormalities and bilateral discharges in the psychosis group. There was no preponderance of either laterality in the psychosis group. 3.2.3. Se6erity of epilepsy Compared to controls, patients with schizophrenia-like psychosis were characterized by an earlier age of onset of epilepsy (Table 7). A history of status epilepticus (non-convulsive or convulsive) was more common in this group such as multiple seizure classes (three or more). The CSSS was highly significantly increased as compared to controls (P B 0.001). There were no differences between groups with respect to the frequency and type of symptomatic etiology of epilepsy, structural lesions as identified by CCT or MRI and verbal or performance IQ. When looking at seizure frequency at onset of the psy-
chiatric disorder and at the end of the follow up at the NHNN there were also no group differences.
3.3. Medication A significantly higher number of patients with schizophrenia-like psychosis received two or more AED. In accordance with the frequency of polytherapy, all AED were prescribed more frequently in psychotic patients compared to the control group. However, this was significant only for DPH. In depressive patients, a significantly low number were on VPA (Table 8).
3.4. Psychosocial 6ariables Although the level of school qualifications was similar as compared to controls, 28% of psychotic patients did not have a professional qualification and 88% were unemployed (Table 9). With respect to psychological variables, in the psychotic group there was a high frequency of severely disturbed paternal relationships. A minority had a relationship and about half of them had never had any sexual experiences with the opposite sex (Table 10). There were no negative psychosocial variables related to depression. In agreement with a relatively high proportion of subjects living independently, 80% of patients with depression had long-lasting relationships. This was however not significant when compared to controls.
E.B. Schmitz et al. / Epilepsy Research 35 (1999) 59–68
64 Table 5 Simple partial seizure symptoms
Vegetative Dysmnesic Olfactory/gustatory Dysphasic Auditory Visual Somato-sensory Motor Cognitive Affective
Patients with psychosis (%, N = 25)
Patients with depression (%, N= 25)
Non-psychiatric controls (%, N = 50)
36** 8 4 – 4 8 – 12 4 8
32* 16 8 12 4 12 – 4 4 12
8 8 2 8 4 2 2 20 2 –
* PB0.05 (depressive patients versus controls). ** PB0.01 (psychotic patients versus controls).
4. Discussion
4.1. Methodological considerations A retrospective study design has disadvantages and a prospective study would clearly be more appropriate in order to investigate risk factors for psychiatric complications of epilepsy. Especially psychosocial variables are insufficiently recorded in medical case notes. The advantage of the retrospective design is pragmatic and relates to the fact that schizophreniform psychoses as well as major depressive episodes are severe complications of epilepsy with a low incidence rate which makes prospective studies in a single center difficult. Patients were obviously not matched for age, sex, IQ, seizure type and severity because we do not know whether these are independent or dependent variables.
4.2. Psychiatric data It should be noted that there was a range of diagnoses and severity within our group of schizophrenia-related psychoses and only a minority of patients had pure interictal psychoses of the type identified in some earlier studies (Slater and Beard, 1963; Perez and Trimble, 1980). In most patients there were identifiable links between the psychiatric complications and seizure fre-
quency or treatment in at least one psychiatric episode. In accordance with other series of psychotic patients (Schmitz and Wolf, 1995) postictal psychoses were most common. The high rate of suicide attempts in the psychotic group underlines the potential danger of these complications. Only two patients had a diagnosis of bipolar depression which is in accordance with the suggested rarity of manic-depressive illness in epilepsy (Toone, 1981; Trimble, 1991). The later age of onset and the female preponderance as compared to psychotic patients are in agreement with the epidemiological data from endogenous monopolar depression (Gelder et al., 1989). Our data suggest an increased risk for the development of pseudoseizures in patients with depression which confirms other findings (Roy, 1979).
4.3. Neurological 6ariables The most significant link was between psychosis and seizure severity as quantified by the CSSS. Other variables which define severity of epilepsy such as early onset of epilepsy, status epilepticus and multiple seizure classes were also related to paranoid-hallucinatory psychosis. Because the number of recorded EEGs was relatively high in the psychotic group the excess of pathological findings may not be reliable. A patient-to-patient analysis was not carried out. However, the in-
E.B. Schmitz et al. / Epilepsy Research 35 (1999) 59–68
65
Table 6 EEG dataa Patients with psychosis (N= 25) Number of EEG recordings Diffuse slowing (%) Temporal focus (%) Multiple foci (%) Right laterality (%) Left laterality (%) Bilateral foci (%)
Patients with depression (N= 25) 2.0 9 0.8
3.49 2.2 52** 68** 40** 36 24 48**
20 24 28 28 28 32
Non-psychiatric controls (N = 50) 2.7 9 2.7 18 28 24 20 24 14
a
Focal abnormalities relate to epileptiform or unspecific discharges. ** PB0.01 (psychotic patients versus controls).
creased frequency of EEG abnormalities over the temporal lobe is supported by the finding of a higher proportion of vegetative auras in this group which suggests a limbic or mesial temporal seizure origin. We found a high proportion of bilateral abnormalities in our psychotic group, which has been described in previous studies (Kristensen and Sindrup, 1978). The mixture of psychotic patients, which included ictal and postictal cases, and contained only a minority with classical interictal schizophreniform illness, precludes us from examining the laterality hypothesis in any more detail in this group. There were no neurological variables linked with depression, except for the syndromatic diagnosis of TLE. This suggests that there may be consistent biological determinants and is in keeping with several other reports of associations between either TLE (Robertson et al., 1987) or complex partial seizures (Mendez et al., 1963) and affective disorders in epilepsy.
4.4. Medication It is well known that AED influence the mental state. In epilepsy patients, the main adverse effect is the link between AED and depression. It has been suggested that polytherapy, and treatment with barbiturates increase the risk for depression in epilepsy patients (Robertson et al., 1987; Brent et al., 1987). The effects of barbiturates are in
contrast to those seen with carbamazepine (CBZ). As has been demonstrated by several studies, treatment with CBZ is inversely correlated with depressive symptomatology in epilepsy patients (Andrewes et al., 1986; Robertson et al., 1987). We found an inverse link between VPA and depression, but not between CBZ and depression. The mood-stabilising properties of VPA have been clearly established in non-epileptic patients, but to date there is no evidence for positive psychotropic effects in epilepsy. In a comparison of 101 depressive patients to 202 non-depressive epilepsy patients, Mendez et al. (1963) also found a lower frequency of VPA prescriptions in the psychiatric group, in keeping with the data presented here. We could not confirm a link between treatment with primidone (PRM) or phenobarbitone (PHB) and depression. This might imply that psychiatric adverse effects of barbiturates (with respect to depression) have been overestimated in the recent literature. For schizophreniform psychosis in epilepsy, the role of AED is less clear. All AED have been associated with psychosis, either as an idiosyncratic dose-related effect or as a result of alternative psychosis with ‘forced normalisation’ of the EEG, a rare phenomenon paralleling seizure control (Landolt, 1958; Tellenbach, 1965). Only two of our patients had an alternative psychosis and this was related to CBZ plus PRM in one case and CBZ plus DPH in the other case. In the six
E.B. Schmitz et al. / Epilepsy Research 35 (1999) 59–68
66
patients with alternative depressive episodes the associated AED were: CBZ four cases, PHB two cases, PRM four cases, DPH three cases. The role of the drugs in these cases is unclear. Alternative psychoses are not usually described in association with standard anticonvulsant drugs, but more with either ethosuximide or some of the more powerful newer agents that readily switch off seizures in some patients with intractable epilepsy (Trimble, 1996). However, there are cases in the literature with all of the standard agents. The frequency noted in our study suggests this may be more common than usually reported.
Polytherapy was not linked with depression, but with schizophrenia-like psychosis. Patients with psychosis were characterized by complicated and severe epilepsies. Therefore, it is likely that polytherapy does not act as a primary risk factor. Of all AED, only DPH was significantly linked with psychosis. It is well known that some individuals develop a paranoid psychosis with high doses of DPH (McDanal and Bolman, 1975). Unfortunately, we did not record exact serum levels. We only distinguished between toxic, normal and subtherapeutic levels. Using these simple categories we did not find intergroup differences.
Table 7 Severity of epilepsy
Age at onset of epilepsy (years) Symptomatic etiology (%) Structural lesions CT/MRI (%) Status epilepticus (%) Multiple seizure classes (3 or more, %) Verbal IQ Performance IQ Severity of seizures1
Patients with psychosis (N= 25)
Patients with depression (N= 25)
Non-psychiatric controls (N =50)
11.3 9 10.4* 48 44 36* 44*
18.2 912.9 40 40 4 16
18.3 914.6 34 34 10 18
91.7 9 17.0 90.69 16.9 33.59 18.9***
96.0 911.3 95.8 912.8 22.0 913.2
87.4 917.8 87.4 921.3 16.9 9 13.8
1
Mean score, based on the Chalfont Seizure Severity Scale (Duncan and Sander, 1991). * PB0.05 (psychotic patients versus controls). *** PB0.001 (psychotic patients versus controls). Table 8 Polytherapy and antiepileptic drugs
Polytherapy Barbiturates Primidone Phenytoin Ethosuximide Carbamazepine Valproate Benzodiazepines Vigabatrin
Patients with psychosis (%, N = 25)
Patients with depression (%, N =25)
Non-psychiatric controls (%, N =50)
80** 16 36 72** – 48 28 4 –
56 28 20 68 – 44 4* 8 –
46 28 12 48 6 46 28 10 2
* PB0.05 (depressive patients versus controls). ** PB0.01 (psychotic patients versus controls).
E.B. Schmitz et al. / Epilepsy Research 35 (1999) 59–68
67
Table 9 Social variables
Immigrant status School qualification Professional qualification Unemployment Independent living
Patients with psychosis (%, N =25)
Patients with depression (%, N =25)
Non-psychiatric controls (%, N= 50)
16 36 28* 88* 56
4 36 40 72 84
20 18 44 54 62
* PB0.05 (psychotic patients versus controls).
Table 10 Psychological variables
Divorce or death of parents No sexual experiences Long-lasting relationships
Patients with psychosis (%, N =25)
Patients with depression (%, N =25)
Non-psychiatric controls (%, N= 50)
28** 56** 36***
16 12 80
8 10 60
** PB0.01 (psychotic patients versus controls). *** PB0.001 (psychotic patients versus controls).
4.5. Psychosocial 6ariables
5. Conclusions
Acute or enduring stressful life events prior to the psychiatric complication were noted in 40% in each group. In psychotic patients there was a high rate of severely disturbed paternal relationships. It is interesting that achievement at school was not worse in the psychotic group, but there was a significant proportion of subjects who did not complete any professional training and who became unemployed which perhaps stresses the social sequelae of recurrent psychosis rather than risk factors. Although only six patients had a chronic psychosis, the striking lack of sexual experiences and lasting relationships suggests severe psychosocial impairment exceeding the recognised psychotic episodes. With respect to depressive episodes we did not identify any specific psychological or social risk factors. It is however possible that we failed to detect any predictive variables due to the poor sensitivity of a retrospective method.
Our retrospective data on the relationship between biological and social variables and psychiatric complications in epilepsy must be carefully interpreted. The correlations between some of the variables and depression or schizophrenia-related psychoses are unlikely to reflect simple etiological relationships. However, it seems that patients who develop psychoses are characterised by a combination of biological and psychosocial risk factors which relate to an early onset severe epilepsy involving the limbic system in addition to an unfavourable familial background. Depressive patients are not as easy to distinguish on neurological factors or severe psychosocial problems, with the exception of an association to TLE. Our working hypothesis for this study was that affective and psychotic disorders in epilepsy had defined etiological and pathological relationships that would be different. In fact, combinations of
68
E.B. Schmitz et al. / Epilepsy Research 35 (1999) 59–68
psychosocial and neurological variables seem linked with both, and simple ideas regarding the psychological as opposed to biological associations with depression, or the development of psychoses in epilepsy are not supported by this study. This is similar to the situation without epilepsy, where simpler divisions between depression and psychotic illnesses in relation to underlying neurological abnormalities and social precipitants have succumbed to the advances of investigations of the nervous system.
References American Psychiatric Association, 1994. Diagnostic and Statistical Manual of Mental Disorders. American Psychiatric Press, Washington, DC. Andrewes, D.G., Bullen, J.G., Tomlinson, L., et al., 1986. A comparative study of cognitive effects of phenytoin and carbamazepine in new referrals with epilepsy. Epilepsia 26, 128 – 134. Brent, D.A., Crumrine, P.K., Varma, R.R., et al., 1987. Phenobarbital treatment and major depressive disorder in children with epilepsy. Pediatrics 80, 909–917. Duncan, J.S., Sander, W.A.S., 1991. The Chalfont Seizure Severity Scale. J. Neurol. Neurosurg. Psychiatry 54, 873– 876. Gelder, M., Gath, D., Mayou, R., 1989. Oxford Textbook of Psychiatry. Oxford University Press, Oxford. Hermann, B.P., Whitman, S., Wyler, A.R., et al., 1990. Psychosocial predictors of psychopathology in epilepsy. Br. J. Psychiatry 156, 98 – 105. Kristensen, O., Sindrup, H.H., 1978. Psychomotor epilepsy and
.
psychosis. II Electroencephalographic findings. Acta Neurol. Scand. 57, 361 – 369. Landolt, H., 1958. Serial electroencephalographic investigations during psychotic episodes in epileptic patients and during schizophrenic attacks. In: Lorentz de Haas, A.M. (Ed.), Lectures on Epilepsy. Elsevier, Amsterdam, pp. 91 – 131. McDanal, C.E., Bolman, W.M., 1975. Delayed idiosyncratic psychosis with diphenylhydantoin. J. Am. Med. Assoc. 231, 1063. Mendez, M.F., Doss, R.C., Taylor, J.L., et al., 1963. Depression in epilepsy. Relationship to seizures and anticonvulsant therapy. J. Nerv. Ment. Dis. 181, 444 – 447. Perez, M.M., Trimble, M.R., 1980. Epileptic psychoses — comparison with process schizophrenia. Br. J. Psychiatry 137, 245 – 249. Robertson, M.M., Trimble, M.R., Townsend, H.R.A., 1987. Phenomenology of depression in epilepsy. Epilepsia 28, 364 – 372. Roy, A., 1979. Some determinants of affective symptoms in epileptics. Can. J. Psychiatry 24, 554 – 556. Schmitz, B., Wolf, P., 1991. Epilepsy and psychosis. In: Theodore, W., Devinsky, O. (Eds.), Epilepsy and Behaviour. Frontiers in Clinical Neurosciences, vol. 12. Wiley, New York. Schmitz, B., Wolf, P., 1995. Psychosis with epilepsy: frequency and risk factors. J. Epilepsy 8, 295 – 305. Slater, E., Beard, A.W., 1963. The schizophrenia-like psychoses of epilepsy. V. Discussion and conclusions. Br. J. Psychiatry 109, 143 – 150. Tellenbach, H., 1965. Epilepsie als Anfallsleiden und als Psychose. Nervenarzt 36, 190 – 202. Toone, B.K., 1981. Psychoses of Epilepsy. In: Reynolds, E.H., Trimble, M.R. (Eds.), Epilepsy and Psychiatry. Churchill Livingstone, Edinburgh, pp. 113 – 137. Trimble, M.R., 1991. The Psychoses of Epilepsy. Raven Press, New York. Trimble, M.R., 1996. Biological Psychiatry. Wiley, Chichester.