Depression and anxiety before and after temporal lobe epilepsy surgery

Seizure 2004; 13: 129 135 doi:10.1016/s1059 1311(03)00073-6 Depression and anxiety before and after temporal lobe epilepsy surgery MARKUS REUBER, BJÖRN ANDERSEN, CHRISTIAN E. ELGER & CHRISTOPH HELMSTAEDTER University Hospital of Epileptology, Sigmund-Freud Str. 25, 53105 Bonn, Germany Correspondence to: Dr C. Helmstaedter, University Hospital of Epileptology, Sigmund-Freud Str. 25, 53105 Bonn, Germany. E-mail: c.helmstaedter@uni-bonn.de Purpose: To examine the course of depression and anxiety in patients with temporal lobe epilepsy (TLE) treated with epilepsy surgery (and anticonvulsant drugs) or medical means alone. Methods: Of 94 TLE patients evaluated for epilepsy surgery, 76 underwent a resective procedure, 18 continued on medical treatment alone. Depression (Beck Depression Inventory (BDI)) and anxiety scores (Self-Rating Anxiety Scale (SRAS)) were examined during presurgical evaluation (T1) and after a mean of 16 months (T2), or 12 months after surgery. Depression and anxiety scores were related to type of intervention, underlying epileptogenic lesion, change of seizure control and anticonvulsant therapy. Results: At T1, depression and anxiety scores were higher in patients with TLE than scores in published normal populations. At T2, depression but not anxiety scores were significantly lower than at T1. Change of depression scores interacted with improvements of seizure control. Conclusions: Evidence of depression and anxiety is commonly found in patients with TLE. Depression improves not because of epilepsy surgery per se, but because of improved seizure control. This is more commonly achieved by surgery than medical treatment. The results are consistent with the hypothesis that depression in TLE is caused by pathological epileptic activity rather than a fixed structural defect. 2003 BEA Trading Ltd. Published by Elsevier Science Ltd. All rights reserved. Key words: epilepsy surgery; depression; anxiety. INTRODUCTION Depression and anxiety are the most common psychiatric disorders in patients with epilepsy, and their prevalence is increased compared to the general population 1, 2. Psychiatric symptoms contribute to the reduction of quality of life and cognitive function found in patients with epilepsy 3, 4, and they are one cause of the increased risk of self-harm, parasuicide and completed suicide in this patient group 5 9. The risk of depression and anxiety may be higher in patients with epilepsy than in patients with other chronic or neurological disorders 10 12. Psychiatric symptoms may be more commonly associated with focal, especially temporal lobe, than idiopathic generalised epilepsies 13 16. Many patients with temporal lobe epilepsy (TLE) prove refractory to medical treatment and are considered for epilepsy surgery. Given the high prevalence and clinical significance of co-existing psychiatric disorders in these patients the effects of surgery on psychiatric symptoms is highly relevant. Previous studies have reported improvement 17 19, deterioration 20, 21, or no change in postoperative depression and anxiety prevalence or scores 22. However, these studies were small 18, 19, 21, were based on follow-up data gathered shortly after surgery 19, 21, or did not differentiate between patients with good or poor postoperative seizure outcome 21, 22. With the exception of one retrospective study 20, no previous studies have included a control group of medically treated TLE patients. Our study was designed to examine the effects of epilepsy surgery or medical anticonvulsant treatment on depression and anxiety in patients with TLE. METHODS All patients with refractory TLE undergoing in-patient presurgical evaluation between 1998 and 2000 were 1059 1311/$30.00 2003 BEA Trading Ltd. Published by Elsevier Science Ltd. All rights reserved.

130 M. Reuber et al. screened for participation in this study. Patients below the age of 15, and patients with a premorbid IQ of less than 70 were excluded. In total, 94 patients fulfilled all inclusion criteria and were willing to participate. Patients were divided into a group of 76 who underwent TLE surgery, and a group of 18 control subjects who continued on anticonvulsant treatment alone. They were not randomised to surgical or medical epilepsy treatment. Of the 76 patients who underwent surgery, 47% were operated on the left and 53% on the right temporal lobe, 60% had a selective amygdalohippocampectomy (SAH), 15% a standard anterior 2/3 temporal lobe resection and 25% an extended temporal lesionectomy. The surgical techniques used in the operated patients have been described elsewhere 23, 24. Histological diagnoses included ammonshorn sclerosis (AHS, 63%), tumor (13%), migration disorder (7%) and other changes including gliosis and cavernoma (16%). In one patient, no histopathological results were available. The medically treated control group consisted of nine patients with TLE who decided against an operation although they were considered good candidates, and nine patients in whom no reasonable hypothesis for an operation could be formulated. In these 18 patients, MRI revealed hippocampal sclerosis in 56%, and other temporal lobe abnormalities in 22%. MRI was unremarkable in 22%. Patients were first examined during in-patient presurgical investigation (T1). Repeat examination (T2) was undertaken on an out-patient basis after a mean of 16 months (SD 3.5). In the surgically treated patients, T2 was performed 12 months postoperatively (SD 3.9). Depression and anxiety levels at T1 and T2 were determined using the Beck Depression Inventory 25 (BDI) and the Self-Rating Anxiety Scale 26 (SRAS). Patients were categorised as probably depressed if they had a sum BDI score of greater than 12. This threshold was chosen because a large previous study (N = 1250) showed that the sensitivity and specificity of this score for clinical depression are 100 and 99%, respectively 27. Following the suggestion of the authors of the test, an SRAS score of greater than 35 was defined as evidence of clinically relevant abnormal anxiety 26. To describe seizure outcome at T2, patients were categorised as follows: (1) completely seizure free, (2) greater than 50% reduction of the frequency of seizures involving loss of consciousness, and (3) less than 50% reduction of seizures with loss of consciousness. Medical anticonvulsant treatment was described using the categories monotherapy and polytherapy. Psychiatric medications were divided into antidepressant and anxiolytic drugs. No patient received antipsychotic medication. Multivariate covariance analysis for repeat measurements (MANCOVA) was used to look for interaction effects between both affective scores (BDI and SRAS) and treatment (surgical/medical) or change of seizure frequency between T1 and T2 in percent. Chi-squared tests (χ 2 -tests) were performed to test the significance of effects of binary variables (e.g. sex, monotherapy or polytherapy). Ordinal or not normally distributed data were compared with the Mann Whitney U-test for independent samples. Continual, normally distributed data were analysed using t-tests. Multiple stepwise regression analysis was performed to identify possible predictors of change of affective scores over time. Findings were reported as significant with P < 0.05 and as a trend with P<0.10. RESULTS At T1 there were no significant differences between the surgically or medically treated groups of TLE patients with regard to age, sex, IQ, anticonvulsant monotherapy or polytherapy, medical treatment for anxiety or depression, seizure frequency, age at onset or duration of epilepsy (Table 1). There was a difference between the groups in terms of lateralisation of the epileptogenic cerebral lesion as determined by MRI: there were significantly more patients with evidence of bilateral or right-sided lesions in the medically treated group. There was no significant difference between the T1 and T2 latency in the two groups. At T2, significantly more patients in the surgically treated group had experienced an improvement in seizure frequency (χ 2 = 55.857; P<0.001), and a reduction of their anticonvulsant therapy (χ 2 = 9.209; P = 0.010). The majority of medically treated patients continued to receive pharmacological polytherapy and seizure frequency remained unchanged (Table 2). Depression and anxiety scores at T1 At T1, patients in the surgical group had significantly higher depression scores than a group of healthy German subjects (T = 3.838; P < 0.001) 28 ; patients continuing on medical treatment only showed a trend towards higher depression scores (T = 2.062; P = 0.056). Twenty-nine percent of surgically and 18% of medically treated patients had a BDI sum score greater than 12. Neither the difference between the frequency of a score elevated to this level (χ 2 = 0.903; P = 0.342) nor the difference in mean sum BDI scores (T = 0.104; P = 0.918) between the two patient groups was statistically significant.

Depression and anxiety 131 Table 1: Demographic and clinical patient characteristics at T1. Epilepsy surgery (N = 76) Medical treatment (N = 18) Significance (two-tailed) Age, mean (SD) 35.0 (9.0) 35.8 (10.1) n.s. b Sex (male/female) 33/43 9/9 n.s. a IQ, mean (SD) 101.0 (13.1) 98.5 (12.2) n.s. b Medical treatment Anticonvulsants None (N) 1% (1) 0% (0) n.s. a Monotherapy (N) 20% (15) 28% (5) Polytherapy (N) 79% (60) 72% (13) Antidepressants Yes (N) 3% (2) 11% (2) n.s. a No (N) 97% (74) 89% (16) Anxiolytics Yes (N) 24% (16) 11% (2) n.s. a No (N) 76% (60) 89% (16) Seizure frequency per year, mean (SD) 110.9 (214.5) 132.2 (188.7) n.s. b Age at onset, mean (SD) 13.1 (9.0) 11.2 (6.6) n.s. b Duration of epilepsy in years, mean (SD) 22.0 (11.7) 24.8 (10.1) n.s. b Lateralisation of cerebral lesion (MRI) Bilateral (N) 0% (0) 11% (2) Right (N) 47% (36) 28% (5) P = 0.007 a Left (N) 53% (40) 61% (11) n.s., not significant. a χ 2 -test; b Mann Whitney U-test. Table 2: Change of seizure frequency and medical therapy. Epilepsy surgery (N = 76) Medical treatment (N = 18) Significance (two-tailed) a Change of seizure frequency Seizure-free (N) 75% (57) 5% (1) P < 0.001 >49% improvement (N) 20% (15) 11% (2) <50% improvement or worse (N) 5% (4) 84% (15) Change of anticonvulsant therapy Fewer anticonvulsants (N) 43% (33) 5% (1) P = 0.01 No change (N) 49% (37) 78% (14) More anticonvulsants (N) 8% (6) 17% (3) Change of antidepressant therapy Started (N) 3% (2) 5% (1) n.s. No change (N) 91% (69) 95% (17) Discontinued (N) 6% (5) 0% (0) Change of anxiolytic therapy Started (N) 17% (13) 5% (1) n.s. No change (N) 79% (60) 78% (14) Discontinued (N) 4% (3) 17% (3) a χ 2 -test. At T1 both, operated (T = 7.932; P < 0.001) and conservatively treated patients (T = 4.081; P = 0.001) had significantly higher anxiety scores than a population of 489 healthy German controls who had a mean SRAS score of 27.5 29. Based on a previously determined SRAS threshold score of 35 26, 48% of pa- tients in the surgical and 39% in the conservative treatment groups were considered abnormally anxious. Mean SRAS scores (T = 0.036; P = 0.971) and the frequency patients with abnormal anxiety in the two TLE groups (χ 2 = 0.485; P = 0.486) did not differ significantly (Table 3). Table 3: Mean depression and anxiety scores and standard deviations at T1. Surgical treatment (N = 76) Medical treatment (N = 18) t-value Significance (two-tailed) a Depression (BDI) 10.07 (8.21) 8.41 (3.92) 0.104 n.s. Anxiety (SRAS) 34.68 (8.11) 34.78 (7.57) 0.036 n.s. n.s., not significant. a t-test for independent samples.

132 M. Reuber et al. gorised as probably depressed dropped significantly in the surgical group (Sign-test, P = 0.021), there was a non-significant increase of depressed patients in the medical group (P = 0.289) (Fig. 1). Seventeen percent of patients who had undergone epilepsy surgery and 12% of patients treated medically who had BDI scores greater than 12 at T1 dropped below this threshold at T2. On the other hand, 4% of surgical and 35% of medical patients who were not categorised as probably depressed at T1 became so at T2. Regardless of treatment mode, a significant relationship between seizure control and change in depression scores could be demonstrated (Table 4). Patients whose seizure frequency had improved by less than 50% or whose seizures were worse had significantly higher depression scores at T2 than at T1 (T = 2.261; P = 0.037). However, depression scores at T2 were significantly lower than at T1 in patients who had at least 50% improvement in seizure frequency (T = 2.284; P = 0.036) or in patients who became seizure-free (T = 4.327; P < 0.001). Anxiety scores at T2 Fig. 1: Proportion of 76 surgically and 18 medically treated patients with TLE, probable depression and abnormal anxiety at the time of the initial assessment (T1) and 16 months (SD 3.5) later (T2). Depression scores at T2 At T2, the mean depression scores of patients with surgically treated TLE were not significantly different from those of a healthy German control population 28 (T = 0.300; P = 0.765), whereas scores in medically treated patients were significantly elevated (T = 3.916; P = 0.001). The proportion of patients with a BDI > 12 at T2 was significantly smaller in the surgically than in the medically treated group (16% vs. 41%, χ 2 = 5.508; P = 0.019). Whereas the number of patients cate- Anxiety scores in both treatment groups remained significantly higher at T2 than those of healthy controls (surgical group: T = 7.872; P < 0.001, medical group: T = 3.869; P = 0.001). At T2, 39% of surgically and 56% of medically treated patients had a SRAS score above the predefined threshold for abnormal anxiety. The prevalence of abnormal anxiety in the two patient groups at T2 did not differ significantly (χ 2 = 1.700; P = 0.192). In neither group was the prevalence of abnormal anxiety significantly different from T1 (Sign-test, surgical group: P = 0.248, medical group: P = 0.375) (Fig. 1). Twenty-three percent of surgically and 6% of medically treated patients above the threshold at T1 had SRAS scores of 35 or less at T2. On the other hand, 13% of surgically and 22% of medically treated patients who were not abnormally anxious at T1 had become so at T2. T2 Table 4: Mean depression and anxiety scores at T1 and T2 and relation between seizure control. T1 T2 t-value Significance (two-tailed) a Depression <50% improved, mean (SD) 8.3 (3.7) 12.9 (6.4) 2.26 P = 0.037 >49% improved, mean (SD) 10.4 (5.9) 7.9 (6.7) 2.28 P = 0.036 Seizure-free, mean (SD) 10.0 (8.9) 6.0 (6.9) 4.33 P < 0.001 Anxiety <50% improved, mean (SD) 33.6 (5.7) 36.8 (6.7) 3.07 P = 0.007 >49% improved, mean (SD) 33.8 (8.2) 32.2 (6.5) 0.74 n.s. Seizure-free, mean (SD) 35.6 (8.5) 33.9 (7.6) 1.68 n.s. n.s., not significant. a t-test for dependent samples.

Depression and anxiety 133 scores only differed significantly from T1 scores in patients who had experienced a deterioration of seizures or a less than 50% improvement of seizure frequency (Table 4). In this group, scores increased significantly from T1 to T2 (T = 3.074; P = 0.007). Multivariate analysis of affective scores (BDI and SRAS) Multivariate analysis revealed significant interactions between change in affective scores between T1 and T2 and change in seizure frequency (Hotelling s F = 4.365; P = 0.016). However, neither repeat measurement (affective scores at T1 and T2, Hotelling s F = 0.148; P = 0.862) nor the in-between group factor (surgical/non-surgical treatment, Hotelling s F = 2.106; P = 0.128) or the covariate (change of seizure frequency, Hotelling s F = 0.824; P = 0.442) showed significant main effects. There was a trend towards a significant interaction between repeat measurement in-between group factor (Hotelling s F = 2.897; P = 0.060). Univariate analysis revealed that the significant interaction effect repeat measurement covariate depended solely on depression scores (F = 8.429; P = 0.005) and not anxiety scores (F = 2.630; P = 0.108). Multiple linear regression analysis of depression scores Multiple linear regression analysis revealed that the change of depression scores between T1 and T2 was affected by the degree of seizure improvement (T = 2.439; P = 0.017), the seizure frequency at T2 (T = 2.190; P = 0.031), and the BDI score at T1 (T = 2.721; P = 0.008). This means that depression scores improved most in the patients with the highest pre-operative BDI scores, the greatest seizure reduction after surgery and the best seizure control at the time of T2. Taken together these three variables explained 36% of the variance of change in BDI. Age at the time of the test, sex, IQ, age of onset of epilepsy, duration of epilepsy, interval between T1 and T2, change of anticonvulsant and antidepressive treatment category, side of cerebral pathology, diagnosis of AHS, abnormal MRI or mode of treatment (not taking into account seizure control) did not affect BDI scores at T2. Use of medical treatments for depression and anxiety At T1 four patients (4%) used antidepressants, 20 patients (21%) anxiolytic drugs. There was no significant difference in terms of use of psychotropic medications between the two groups (Table 1). At T2 six patients (6%) took antidepressant drugs, 12 patients (13%) anxiolytic agents. There was no significant difference in the change of psychiatric medications for depression (χ 2 = 1.59; P = 0.450) or anxiety (χ 2 = 4.995; P = 0.082) from T1 to T2 between the groups (Table 2). According to the results of our multiple linear regression analysis, changes in antidepressant treatment did not contribute to the improvements seen in depression scores at T2. DISCUSSION Prior to therapeutic intervention (T1), mean depression and anxiety scores were higher in patients with TLE than in healthy control populations. Based on threshold scores determined in other studies 26, 27, 27% of patients had probable depression and 46% abnormal anxiety. At T2, there were no significant changes of both affective scores in the medically treated TLE patients. In the surgical group depression but not anxiety scores had improved. It is possible that the test retest interval of 16 months (or 12 months after surgery) was too short to detect improvements in anxiety scores or that the SRAS is not sufficiently sensitive in the setting of TLE and epilepsy surgery. However, the lack of change of anxiety scores observed by us is in line with the findings of a previous study 30 which found that 54% of patients continued to suffer from significant anxiety symptoms although 78% were seizure-free after TLE surgery. Postoperative improvements in depression scores after epilepsy surgery have also been observed in the past 18, 21. In contrast to previous prospective studies 17 19, 21, 22, our study included a control group of patients with TLE treated with medical means alone. This allowed us to differentiate between the effects of temporal lobe surgery and the effects of improved seizure control. Our results reveal that surgery as such did not improve depression scores. Scores only dropped after surgical intervention, if seizure control was improved postoperatively. As clinically relevant improvements of seizure control between T1 and T2 were achieved in 95% of surgically but only 17% of medically treated patients, improvements in BDI scores were significantly greater in the surgical patient group. The correlation between the degree of improvement of BDI score and seizure control found here is in line with one previous study 18 although it is in contrast with several others which suggested that only complete seizure-freedom leads to changes of depression scores 19, 21, 31. The fact that absolute BDI scores at T1 were correlated with the degree of change at T2 has been

134 M. Reuber et al. reported before 18, 20, and is hardly surprising. Very low and very high scores offer greater scope for change than scores in the middle of the BDI scale. Taken together, our results support the notion that depression in patients with TLE is predominantly caused by endogenous factors 15. The findings suggest that the improvement of depression scores was caused by a reduction of abnormal ictal (and presumably interictal) epileptic activity rather than the removal of tissue from, or the deafferentation of neurons within the temporal lobe. That is to say in TLE patients, depression seems to be related to pathological dys-functional processes rather than fixed structural defects. A purely reactive effect, with patients being less depressed because of fewer seizures, seems less likely because anxiety scores did not fall in the same way. Psychiatric medications did not explain the improvement of depression scores at T2. However, it was notable at both, T1 and T2, that only 4 or 6% of patients were treated with antidepressants although 27 or 20% had BDI scores greater than 12 patients. Likewise, the number of patients with recognised and treated anxiety disorders (T1: 21%; T2: 13%) was much smaller than the number with SRAS scores above the threshold of 35 (T1: 46%; T2: 42%). In accord with other studies in patients with epilepsy 32, this suggests that depression and anxiety are under-recognised in this patient group. There are several limitations to our results. Patients undergoing evaluation for epilepsy surgery represent a highly selected group. Our results should, therefore, not be extrapolated to other patients with epilepsy. In view of the proven effectiveness of surgery for refractory TLE 33, it would be unethical to randomise patients to surgical or medical treatment groups. It is, therefore, possible that our results were affected by clinical differences between the two patient groups (for instance, more patients in the medical group had bilateral temporal lobe lesions). Despite the high sensitivity and specificity of BDI scores greater than 12 for a clinical diagnosis of depression 27, the validity of self-assessment by questionnaire is necessarily limited. With this caveat we conclude that (1) the prevalence of probable depression and abnormal anxiety are increased in patients with refractory TLE; (2) depression but not anxiety levels can be reduced by therapeutic (antiepileptic) intervention; (3) improvements of depression scores are greater in surgically than in medically treated patients with previously refractory TLE, because surgery controls seizures more effectively; (4) depression is not improved by surgery per se; and (5) depression in patients with TLE seems to be related to pathological epileptic activity rather than fixed structural defects. ACKNOWLEDGEMENT B. Andersen was supported by the German Department of Health (212-43346-14/19), M. Reuber by the St. James s Trust for Nervous System Diseases and the Special Trustees of the General Infirmary, Leeds, UK. REFERENCES 1. Dodrill, C. B. and Batzel, L. W. Interictal behavioral features of patients with epilepsy. Epilepsia 1986; 27 (Suppl. 2): S64 S76. 2. Pulliainen, V., Kuikka, P. and Kalska, H. Are negative mood states associated with cognitive function in newly diagnosed patients with epilepsy? Epilepsia 2000; 41: 421 425. 3. Lehrner, J., Kalchmayr, R., Serles, W. et al. Health-related quality of life (HRQOL), activity of daily living (ADL) and depressive mood disorder in temporal lobe epilepsy patients. Seizure 1999; 8: 88 92. 4. Sawrie, S. M., Martin, R. C., Kuzniecky, R. et al. Subjective versus objective memory change after temporal lobe epilepsy surgery. Neurology 1999; 53: 1511 1517. 5. Barraclough, B. M. Suicide and epilepsy. In: Epilepsy and Psychiatry (Ed. M. R. Trimble). London, Churchill Livingstone, 1981: pp. 339 350. 6. Barraclough, B. M. The suicide rate of epilepsy. Acta Psychiatry Scandinavian 1987; 76: 339 345. 7. Hawton, K., Fagg, J. and Marsack, P. Association between epilepsy and attempted suicide. Journal of Neurology, Neurosurgery, and Psychiatry 1980; 43: 168 170. 8. Mackay, A. Self-poisoning. A complication of epilepsy. British Journal of Psychiatry 1979; 134: 277 279. 9. Mendez, M. F., Cummings, J. L. and Benson, D. F. Depression in epilepsy. Significance and phenomenology. Archive Neurology 1986; 43: 766 770. 10. Kogeorgos, J., Fonagy, P. and Scott, D. F. Psychiatric symptom patterns of chronic epileptics attending a neurological clinic: a controlled investigation. British Journal of Psychiatry 1982; 140: 236 243. 11. Standage, K. F. and Fenton, G. W. Psychiatric symptom profiles of patients with epilepsy: a controlled investigation. Psychological Medicine 1975; 5: 152 160. 12. Kanner, A. M. and Barry, J. J. Controversies in epilepsy and behavior. Is the psychopathology of epilepsy different from that of nonepileptic patients? Epilepsy & Behavior 2001; 2: 170 175. 13. Fenwick, P. Epilepsy and psychiatric disorders. In: Epilepsy (Ed. A. Hopkins). New York, Chapman & Hall, 1987: pp. 511 521. 14. Indaco, A., Carrieri, P. B., Nappi, C. et al. Interictal depression in epilepsy. Epilepsy Research 1992; 12: 45 50. 15. Lambert, M. V. and Robertson, M. M. Depression in epilepsy: etiology, phenomenology, and treatment. Epilepsia 1999; 40 (Suppl. 10): S21 S47. 16. Septien, L., Giroud, M., Didi-Roy, R. et al. Depression and partial epilepsy: relevance of laterality of the epileptic focus. Neurological Research 1993; 15: 136 138. 17. Blumer, D., Wakhlu, S., Davies, K. et al. Psychiatric outcome of temporal lobectomy for epilepsy: incidence and treatment of psychiatric complications. Epilepsia 1998; 39: 478 486. 18. Derry, P. A., Rose, K. J. and McLachlan, R. S. Moderators of the effect of preoperative emotional adjustment on postoperative depression after surgery for temporal lobe epilepsy. Epilepsia 2000; 41: 177 185.

Depression and anxiety 135 19. Hermann, B. P., Wyler, A. R., Ackerman, B. et al. Short-term psychological outcome of anterior temporal lobectomy. Journal of Neurosurgery 1989; 71: 327 334. 20. Altshuler, L. L., Rausch, R., Delrahim, S. et al. Temporal lobe epilepsy temporal lobectomy and major depression. Journal of Neuropsychiatry and Clinical Neuroscience 1999; 11: 436 443. 21. Glosser, G., Zwil, A. S., Glosser, D. S. et al. Psychiatric aspects of temporal lobe epilepsy before and after anterior temporal lobectomy. Journal of Neurology, Neurosurgery, and Psychiatry 2000; 68: 53 58. 22. Anhoury, S., Brown, R. J., Krishnamoorthy, E. S. and Trimble, M. R. Psychiatric outcome after temporal lobectomy: a predictive study. Epilepsia 2000; 41: 1608 1615. 23. Behrens, E., Zentner, J., Van Roost, D., Hufnagel, A., Elger, C. E. and Schramm, J. Subdural and depth electrodes in the presurgical evaluation of epilepsy. Acta Neurochirurgica (Wien) 1994; 128: 84 87. 24. Zentner, J., Hufnagel, A., Wolf, H. K. et al. Surgical treatment of temporal lobe epilepsy: clinical radiological and histopathological findings in 178 patients. Journal of Neurology, Neurosurgery, and Psychiatry 1995; 58: 666 673. 25. Hautzinger, M., Bailer, M., Worall, H. and Keller, F. Das Beck Depressions Inventar (BDI). Bern, Huber, 1993. 26. Zung, W. W. K. SAS. Self-Rating Anxiety Scale. In: ECDEU Assessment Manual for Psychopharmacology (Ed. W. Guy). Maryland, Rev. Ed. Rockville, 1976. 27. Lasa, L., Ayuso-Mateos, J. L., Vazquez-Barquero, J. L., Diez-Manrique, F. J. and Dowrick, C. F. The use of the Beck Depression Inventory to screen for depression in the general population: a preliminary analysis. Journal of Affective Disorders 2000; 57: 261 265. 28. Hautzinger, M. Das Beck Depressions Inventar (BDI) in der Klinik. Der Nervenarzt 1991; 62: 689 696. 29. Merz, W. A. and Ballmer, U. Demographic factors influencing psychiatric rating scales (Zung SDS and SAS). Pharmacopsychiatry 1984; 17: 50 56. 30. Bladin, P. F. Psychosocial difficulties and outcome after temporal lobectomy. Epilepsia 1992; 33: 898 907. 31. Kellett, M. W., Smith, D. F., Baker, G. A. and Chadwick, D. W. Quality of life after epilepsy surgery. Journal of Neurology, Neurosurgery, and Psychiatry 1997; 63: 52 58. 32. Paradiso, S., Hermann, B. P., Blumer, D., Davies, K. and Robinson, R. G. Impact of depressed mood on neuropsychological status in temporal lobe epilepsy. Journal of Neurology, Neurosurgery, and Psychiatry 2001; 70: 180 185. 33. Wiebe, S., Blume, W. T., Girvin, J. P. and Eliasziw, M. A randomized controlled trial of surgery for temporal-lobe epilepsy. New England Journal of Medicine 2001; 345: 311 318.