Postictal psychosis in temporal lobe epilepsy: Risk factors and postsurgical outcome?

Epilepsy Research (2013) 106, 264—272

journal homepage: www.elsevier.com/locate/epilepsyres

Postictal psychosis in temporal lobe epilepsy: Risk factors and postsurgical outcome? Rebecca A. Cleary a,b,∗, Pamela J. Thompson a, Maria Thom a, Jacqueline Foong a,b a b

Department of Clinical and Experimental Epilepsy, Institute of Neurology, University College London, London, UK Department of Neuropsychiatry, The National Hospital for Neurology and Neurosurgery, Queen Square, London, UK

Received 22 November 2012 ; received in revised form 19 March 2013; accepted 25 March 2013 Available online 1 May 2013

KEYWORDS Temporal lobe epilepsy; Epilepsy surgery; Neuropsychiatry; Neuropsychology; Seizures

Summary Purpose: Postictal psychosis (PIP) occurs in ∼7% of patients with temporal lobe epilepsy (TLE). The aims of this study were to identify risk factors that may predispose patients to developing PIP and determine whether the severity of PIP predicts postsurgical outcome. Methods: We compared 20 patients with a history of TLE + PIP to 60 age-matched TLE patients without any psychiatric history (TLE-only), with respect to pre-surgical clinical and neuropsychological variables. Group differences in postsurgical psychiatric, histopathological, cognitive and seizure outcomes were also examined. Key findings: TLE + PIP patients were significantly less likely to have localised ictal epileptiform activity than the TLE controls (p = 0.05) and were significantly more likely to have a positive family psychiatric history than TLE controls (p = 0.04). Other pre-surgical clinical and neuropsychological variables did not distinguish between the groups. Patients with two or more PIP episodes had significantly increased odds of developing de novo psychopathology within 4 years of surgery, after controlling for comorbid pre-surgical psychiatric status and a history of SGTCS (OR: 9.11, 95% CI: 1.53—54.10, p = 0.02). A history of PIP did not significantly predict other postsurgical outcomes (seizure freedom (ILAE = 1) or cognitive outcome). Significance: Our results suggest that more widespread or diffuse brain abnormalities as reflected by pre-surgical EEG findings and positive genetic determinants may contribute to the development of PIP. Furthermore, patients with recurrent PIP episodes who undergo TLE surgery are at increased risk of developing de novo psychiatric disorders, particularly mood disorders. This has implications for pre-operative counselling and highlights a need for postsurgical psychiatric monitoring for these patients. © 2013 Elsevier B.V. All rights reserved.

∗ Corresponding author at: Department of Neuropsychiatry, The National Hospital for Neurology and Neurosurgery, Queen Square, London WC1N 3BG, UK. Tel.: +44 01494 601353; fax: +44 01494 874136. E-mail address: [email protected] (R.A. Cleary).

0920-1211/$ — see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.eplepsyres.2013.03.015

Postictal psychosis in temporal lobe epilepsy

Introduction Post-ictal psychosis (PIP) is a recognised psychiatric complication of temporal lobe epilepsy (TLE); usually manifesting after a cluster of complex partial or generalised seizures (Logsdail and Toone, 1988). Psychotic symptoms tend to emerge following a symptom-free (lucid) period ranging from 12 to 72 h (Kanner et al., 1996) and are transient; typically lasting several days (Leutmezer et al., 2003), but may be up to weeks (Devinsky et al., 1995). Incidence rates of PIP in studies of patients undergoing video-electroencephalography recordings (VEEG) are ∼7% (Kanner et al., 1996; Alper et al., 2001). Patients who have a recurrent PIP may be at risk of developing chronic interictal psychosis (Tarulli et al., 2001). The clinical risk factors reported are diverse (Devinsky, 2008), with consistent reports of a history of encephalitis (Devinsky et al., 1995; Alper et al., 2008); head trauma (Devinsky et al., 1995); secondary generalised tonic-clonic seizures (SGTCS) (Devinsky et al., 1995; Alper et al., 2008; Kanner and Ostrovskaya, 2008); a range of EEG abnormalities such as bilateral interictal (Devinsky et al., 1995; So et al., 1990; Christodoulou et al., 2002; Schulze-Bonhage and van Elst, 2010; Oshima et al., 2011); ictal (So et al., 1990); extra-temporal (Alper et al., 2008; Christodoulou et al., 2002; Kanemoto et al., 1996; Ferguson et al., 1969); nonlocalised/lateralised epileptiform activity (So et al., 1990; Ferguson et al., 1969; Falip et al., 2009); and EEG slowing (Kanner et al., 1996; Alper et al., 2008; So et al., 1990; Mathern et al., 1995; Seeck et al., 1999; Takeda et al., 2001; Kuba et al., 2012). The pathophysiological mechanisms of PIP are not fully understood. Results from functional imaging studies have implicated that widespread or diffuse brain dysfunction. Case reports using SPECT during an acute episode have found increased perfusion in contralateral basal ganglia (Fong et al., 2000); bi-temporal and bi-frontal areas (Leutmezer et al., 2003); and widespread ipsilateral fronto-temporal networks (Nishida et al., 2006). Interictal bitemporal hypometabolism has also been documented in single case studies (Seeck et al., 1999; Kuba et al., 2012). Structural imaging studies have reported bilateral pathology, generalised cerebral volume loss (Morrow et al., 2006) and cortical abnormalities in PIP patients (DuBois et al., 2011). Neuropathological studies are limited, focussing mainly on patients with chronic interictal psychosis and not PIP. One small uncontrolled retrospective study reported that preoperative PIP was associated with temporal dysplasia (Briellman et al., 2000), in contrast to reports of an increased prevalence of ‘alien tissue’ (small tumours, hamartomas, focal cortical dysplasia) (Taylor, 1975) and gangliogliomas (Roberts et al., 1990) in interictal psychosis. Further research into the histopathological correlates of PIP is warranted. Few studies have investigated the relationship between cognitive status and PIP. No study has systematically explored the association of PIP with neuropsychological test performance apart from studies reporting unspecified bi-temporal memory dysfunction (So et al., 1990; Christodoulou et al., 2002; Falip et al., 2009; Mathern et al., 1995; Savard et al., 1991).

265 Patients with a history of PIP are not excluded from being considered for epilepsy surgery and it has been argued that there may be an indication to consider resective surgery in order to prevent the evolution into chronic (interictal) psychosis (Kanner, 2009). However, there is limited data of postsurgical outcome of patients with PIP who undergo epilepsy surgery in relation to psychiatric status (Kanemoto et al., 1998, 2001) or seizure outcome (Kanner and Ostrovskaya, 2008; Falip et al., 2009). The aims of this study therefore were to examine the clinical risk factors associated with PIP in TLE patients, including EEG, cognitive and neuropathological measures. Furthermore, we wanted to investigate whether a history or frequency of PIP predicted psychiatric, cognitive or seizure outcome following TLE surgery.

Method Study sample Patients who had undergone epilepsy surgery at the National Hospital for Neurology and Neurosurgery between 1997 and 2007 were identified from a surgical database. Of the 280 TLE case described elsewhere (Cleary et al., 2012), a single coder (RAC) systematically identified 20 TLE patients with a history of PIP (TLE + PIP) and 60 age-matched TLE controls with no history of psychopathology (TLE-only; n = 60). The sample was selected blind for neurological history, family history of seizures and psychopathology, VEEG and MRI results. Demographic and clinical characteristics are summarised in Table 1. This study was approved by UCL Hospital NHS Foundation Trust.

Neuropsychiatric classification All TLE patients were routinely assessed by a neuropsychiatrist as part of their pre-surgical evaluation and psychiatric diagnoses pre- and postoperative were recorded (DSM-IV-TR Axis I). For the TLE + PIP group, we recorded the number of PIP episodes documented. None of these PIP episodes were as a result of drug reduction/withdrawal during pre-surgical VEEG. Any family psychiatric history was also documented. For co-morbid psychopathology (e.g. a history of PIP and depression), a positive entry was made into each diagnostic category. For the TLE + PIP group, de novo psychopathology was only coded if a patient had no preoperative psychiatric history of that psychiatric disorder. For example, if a TLE + PIP patient had preoperative and postoperative depression, but developed and anxiety disorder post-surgically, only the latter condition would be considered de novo. For the TLE control group, all postoperative psychiatric conditions were coded as de novo as these patients had no previous psychiatric history. All de novo cases could only be termed as such at one mutually exclusive time frame. All patients had neuropsychiatric evaluation only at 3 and 6 months post-surgically and subsequently diagnosed psychiatric disorder(s) were then coded as appropriate into the following postoperative time-frames: 12, 24, 36 and 48 months.

266 Table 1

R.A. Cleary et al. Demographic and clinical characteristics of TLE + PIP and TLE-only patients who underwent TLE surgery.

Variable

TLE + PIP (n = 20)

TLE-only (n = 60)

All (n = 80)

Age of seizure onset, yb Epilepsy duration, yb Seizure laterality (R/L) Predisposing factors History of febrile convulsions History of status epilepticus History of head trauma History of CNS infection Family history of epilepsy Pre-surgical seizure typea Simple partial Complex partial History of SGTCS Medication Total number of AEDs trialled preoperativelyb Monotherapy at the time of surgery Polytherapy at the time of surgery Presurgical investigations SEEG Ictal SPECT PET Localised ictal EEG Localised interictal EEG Pre-surgical psychiatric history Mood disorder Anxiety disorder Interictal psychosis Psychiatric treatment Suicidal ideation Suicidal attempts Psychiatric hospitalisation Family history of psychopathology Histopathological findingsc Hippocampal sclerosis HS + FCDIIIa Glioneuronal tumour Cavernomas/vasucular malformations Non-specific gliosis only Dual pathology

6 (2.1—14) 29 (21—40.3) 40/60%

9 (4—15) 24 (16—30) 55/45%

8.5 (3—15) 25 (17—32) 51/48%

55% 30% 20% 10% 25%

50% 20% 12% 7% 15%

51% 23% 14% 8% 18%

75% 100% 65%

78% 80% 80%

78% 99% 76%

5 (3—7) 5% 95%

4 (3—6) 10% 90%

5 (3—6) 9% 91%

5% 5% 0% 40% 25%

7% 0% 0% 65% 41%

6% 1% 0% 58% 38%

25% 15% 0% 60% 15% 0% 55% 25%

0% 0% 0% 0% 0% 0% 0% 7%

6% 4% 0% 15% 4% 0% 14% 11%

71% 12% 12% 0% 0% 6%

67% 9% 15% 4% 0% 4%

68% 10% 14% 3% 0% 5%

AED = anti-epileptic drugs. a May have >1 seizure type. b Median (IQR). c For available data.

Seizure data

Neuropathology

The following seizure related variables were recorded; a history of febrile convulsion(s), status epilepticus, head trauma, CNS infection (e.g. meningitis/encephalitis) and familial history of epilepsy. Postoperative seizure outcome was taken from a surgical follow-up database and classified according to the International League Against Epilepsy (ILAE) surgical classification scheme (Wieser et al., 2001).

All tissue from each surgical resection was routinely processed as previously described (Martinian et al., 2012). The pathology findings in the cases were groups as follows (1) hippocampal sclerosis (HS) alone; (2) HS with additional focal cortical dysplasia (FDC type IIIa); (3) glioneuronal tumour (dysembryoplastic neuroepithelial tumour (DNT) or ganglioglioma); (4) cavernomas or other vascular formations; (5) Non-specific gliosis only, with no specific

Postictal psychosis in temporal lobe epilepsy pathological lesion identified and (6) dual pathology comprising of cases with HS and another lesions (e.g. cavernoma, DNT, but HS+FCDIIIa was not regarding as dual pathology) (Blümcke et al., 2011).

267 networks involving more lateral temporal region. Two were taken from the WAIS-III; namely the Vocabulary and Similarities subtests. In addition, we used data from the Graded Naming Test (GNT) (McKenna and Warrington, 1980) and a category fluency test.

MRI and VEEG data The laterality of the epileptogenic focus was determined using preoperative MRI scans and VEEG reports. Pre-surgical VEEG reports were used to classify cases demonstrating ‘localised’ or ‘non-localised’ ictal or interictal epileptiform abnormalities. Non-localised ictal or interictal EEG was based on any epileptiform discharges that were discordant with neuro-radiological findings as defined by a neuro-physiologist. For example, independent bi-temporal spikes, multifocal spikes, fronto-temporal discharges and documented ‘non-localising/lateralising’ EEG features in the presence of unilateral HS.

Neuropsychological measures All patients had undergone a neuropsychological assessment preoperatively, 3—6 months and 12—18 months postoperatively. Measures were selected to provide a measure of general ability and to tap temporal and extra-temporal brain regions. General ability Intellectual level was measured using the WAIS-III and verbal IQ (VIQ) and performance IQ (PIQ) were derived. Patients also underwent the Nelson Adult Reading Test (NART) (Nelson, 1982) preoperatively. This is a measure of reading competence that has been validated as a measure of intellectual potential (McGurn et al., 2004). A discrepancy between NART predicted IQ and assessed IQ indicates more widespread cerebral disturbance. Memory Scores were available from the List Learning, Story Recall, Design Learning and Figure Recall subtests from the Adult Memory and Information Processing Battery (Coughlan and Hollows, 1985). The learning paradigms in particular have been demonstrated as sensitive to mesial temporal lobe pathology and surgical treatment (Baxendale et al., 2008). Contralateral memory weakness We anticipated that contralateral memory dysfunction would be more marked in the TLE + PIP patients compared to controls (TLE-only). Performance in List- and Design Learning were translated into Z-scores and used as measures of hippocampal integrity. The contralateral memory score for each case (i.e. Z-score on verbal memory (List Learning) on right TLE patients or visual memory (Design Learning) Z-score in left TLE patients) was recorded. This composite measure of contralateral memory function (i.e. overall group mean contralateral Z-score) was statistically compared between the groups.

Executive functions For two measures working memory was considered a major component, namely the Digit Span Test from the WAIS-III and a phonemic fluency test. Working memory is an early and key stage in most cognitive processing, but is fundamental to executive functioning and evidence supports a fronto-parietal cerebral network (Schlösser et al., 2006).

Memory decline The difference in pre- and postoperative List- and Design Leaning performance was calculated by a change in Z-scores at 3 and 12 months following TLE surgery. Patients’ ipsilateral and contralateral Z-scores, according to resection laterality, were then entered into linear regression analyses.

Statistical analyses Due to sample attrition, analyses were based on a 4-year follow-up. Chi-square (2 ; with Fisher Exact Test (FET) when warranted) was used to evaluate comparisons between TLE+PIP and control patients (TLE-only) on categorical variables. Differences between continuous variables were evaluated using the independent samples t-test, with the exception of variables that were nonnormally distributed, for which Mann—Whitney U tests were employed. Uni- and multivariable logistic regression was used to examine whether the frequency of PIP predicted postsurgical (de novo) psychiatric outcome. Specifically, we computed a dummy variable with four levels (Field, 2000) to control for the possible confounding influence of a history of SGTCS and co-morbid psychopathology other than PIP (e.g. PIP and depression), as both have been demonstrated as independent predictors of postoperative psychopathology (Cleary et al., 2012). This dummy (predictor) variable was entered into multivariable logistic regression analysis. Univariable logistic regression was also used to explore whether a history (yes vs no) or frequency (0 vs 1 vs ≥2) of PIP predicted seizure freedom (ILAE = 1) at 12 or 48 months. Whereas, linear regression explored whether a history (yes vs no) or frequency (0 vs 1 vs ≥2 episodes) of PIP predicted ipsilateral or contralateral memory decline at 3—6 and 12—18 months postoperatively.

Results Clinical characteristics

Semantic processing Data from four tests were available that measured semantic processing and that have been shown to involve neural

The two groups did not differ in terms of age of seizure onset duration, epilepsy duration, history of febrile

268

R.A. Cleary et al.

convulsions, status epilepticus, head trauma, CNS infection or family history of epilepsy. We did not find differences in the frequency of simple partial, complex partial or history of SGTCS or lateralisation of seizure onset. TLE + PIP patients had significantly less localised ictal EEG recordings than controls (40% vs 65%; 2 (1) = 3.8, p = 0.05). The interictal EEG indicated a similar but non-significant trend (TLE + PIP: 25% vs controls: 41%, p = 0.20). Regarding the frequency of pre-surgical investigations, no significant differences were evident. A family psychiatric history was significantly more common in the TLE + PIP patients (25%) compared to the controls (7%) (p = 0.04, FET). Univariable logistic regression analysis indicated that TLE + PIP patients had over four and a half times the odds of a positive family psychiatric history compared to controls (n = 5/20 vs 4/60; OR: 4.67, 95% CI: 1.11—19.55, p = 0.04).

Neuropsychology Table 2 indicates that both groups demonstrated similar neuropsychological profiles with weak naming and memory. There were no significant differences in cognitive findings between the groups (p values > .05; see Table 2).

Table 2

VIQ PIQ NART NART and VIQ discrepancy Memory Verbal learning: trials Verbal learning: delay Verbal recall: immediate Verbal recall: % retained Visual learning: trials Visual learning: delay Visual recall: immediate Visual recall: % retained Contralateral memory weakness Semantic knowledge Vocabulary Similarities Graded Naming Test Category fluency Executive functions Working memory (Digit Span) Phonemic fluency (‘‘S’’ only) a

Table 3 presents the point prevalence of de novo psychopathologies for each group at specific time-intervals following TLE surgery.

TLE + PIP group Nine out of 20 (45%) patients with a history of PIP developed de novo psychopathology during the four year follow-up. De novo psychopathologies were most apparent within the first post-surgical year (IQR: 3—33 months) and the majority (78%; n = 7/9) persisted for at least 6 months (IQR: 5—33 months). The most common de novo psychiatric condition were mood disorders (depression, n = 3; mixed affective state, n = 1; bipolar disorder, n = 1) followed by interictal psychosis (n = 3) and anxiety disorders (not otherwise specified, n = 1). Six patients (67%; n/9) required psychiatric treatment (pharmacological and/or psychological) within four year follow-up.

Predictors of de novo psychopathology: group analysis In univariable analyses, only a history of PIP was a significant predictor of de novo psychopathology within 4 years

Preoperative cognitive performance of TLE + PIP and TLE-only patients.

Cognitive measurea

b

De novo psychopathology

For available data. Based on group mean age, y.

TLE + PIP (n = 20)

TLE-only (n = 60)

Mean (SD)

Centileb

Mean (SD)

Centileb

93 (10) 95 (14) 97 (12) −2.8 (8.8)

— — — —

94 (15) 96 (19) 100 (14) −5.2 (9.8)

— — — —

43 (10) 8 (3) 23 (12) 72 (23) 30 (8) 6 (3) 50 (20) 93 (15) −0.5(1.4)

5—10th 10th 10—25th 5—10th 10th 5—10th <2nd 25—50th —

46 (10) 8 (3) 24 (12) 79 (25) 31 (9) 6 (3) 65 (22) 94 (19) −0.5(1.1)

10—25th 10th 25th 10—25th 10—25th 5—10th 10—25th 25—50th —

8 (2) 9 (2) 15 (6) 17 (6)

— — 10th 50—55th

8 (3) 9 (3) 16 (5) 18 (5)

— — 10—25th 55—63rd

10 (2) 15 (3)

— —

10 (3) 16 (6)

— —

Postictal psychosis in temporal lobe epilepsy Table 3

269

New cases of de novo psychopathologies for each group at specific time-intervals following TLE surgery.

De novo psychopathologies Mood disorder TLE + PIP TLE-only Anxiety disorder TLE + PIP TLE-only Adjustment disorder TLE + PIP TLE-only Interictal psychosis TLE + PIP TLE-only Postictal psychosis TLE + PIP TLE-only NES TLE + PIP TLE-only

3 months

6 months

12 months

24 months

36 months

48 months

Overall total

3 4

0 0

1 1

1 2

0 1

0 1

5 9

0 0

0 1

0 0

0 0

0 0

1 0

1 1

0 1

0 0

0 0

0 0

0 0

0 0

0 1

1 0

1 0

1 0

0 0

0 1

0 0

3 1

0 0

0 0

0 0

0 1

0 0

0 0

0 1

0 0

0 0

0 0

0 0

0 0

0 0

0 0

(OR: 3.0, 95% CI: 1.01—8.65, p = 0.05). In the multivariable analysis, a history of PIP remained a significant predictor of de novo psychopathology after controlling for preoperative psychiatric status and a history of SGTCS (OR: 5.06, 95% CI: 1.23—20.82, p = 0.03).

the increased chance (rather than odds) of de novo psychopathology as a function of PIP episodes (0 vs 1 vs ≥2), after adjusting for SGTCS and co-morbid psychiatric conditions (Kanner et al., 2000) (see Fig. 1).

Histopathology Recurrent episodes of PIP and de novo psychopathology Patients with only one episode of PIP did not have significantly increased odds of de novo psychiatric morbidity within 4 years after surgery (OR: 1.36, 95% CI: 0.31—5.85, p = 0.68). However, patients with two or more PIP episodes had significantly increased risk of de novo psychiatric morbidity within 4 postoperative years, even after controlling for SGTCS and co-morbid psychopathology (OR: 9.11, 95% CI: 1.53—54.10, p = 0.02). For clinical translation, we assessed

There were no significant differences in histopathological findings between TLE + PIP and control patients (see Table 1).

Cognitive outcome Linear regression analyses indicated that neither a history of PIP or the number of episodes (0 vs 1 vs ≥2) predicted ipsilateral or contralateral memory decline at 3- or 12 month follow-up (p values > .05).

Figure 1 The modifying effect of preoperative PIP episodes (0 vs 1 vs ≥2) on a patient’s chance (%) of developing de novo psychopathology within 4 years following TLE surgery.

270

R.A. Cleary et al.

Seizure outcome

Postsurgical outcome

There were no significant differences in the number of patients who were seizure free (ILAE = 1) at 12 months (TLE-only: 62% vs TLE + PIP: 50%, p = 0.44) or remained seizure free over 4 years (TLE-only: 45% vs TLE + PIP: 35%, p = 0.6). In addition, there were no significant associations in seizure freedom and the number of PIP episodes (p values >.05).

Consistent with previous reports (Kanemoto et al., 1998, 2001), we found that a history of PIP was associated with postsurgical de novo psychopathology. However, our study extends previous findings in several ways. We had a much larger sample size, identified patients with co-morbid presurgical psychiatric diagnoses and/or SGTCS and controlled for these possible confounds in our analyses. Furthermore, we determined that patients with a history of two or more PIP episodes had a significantly increased risk of developing de novo psychopathology within four postoperative years (see Fig. 1). Notably, the 3 patients in the TLE + PIP group who developed a postoperative chronic interictal psychosis had a history of more than 2 PIP episodes (range: 2—6). This is concordant with Tarulli et al.’s (2001) findings that 5 out of 6 of their patients with interictal psychosis had a previous history of at least 2 or more PIP episodes. Our findings may suggest that resective TLE surgery should not be delayed in patients with a history of PIP (Kanner, 2009). Few studies have examined the postsurgical seizure outcome of PIP patients (Kanner and Ostrovskaya, 2008). To our knowledge, this is the first study to investigate surgical outcome (ILAE = 1) in TLE + PIP patients compared to controls over a four year period (at yearly intervals). We found that postsurgical seizure outcome (ILAE 1 vs 2—6) did not significantly differ between the groups at 12 or during 48 months, nor did the number of PIP episodes predict seizure freedom. Alper et al. (2008) also found no difference between seizure freedom rates in PIP vs non-PIP epilepsy patients. In Kanemoto et al.’s (1998) study, whilst the majority (81%) of the 52 TLE patients in their study achieved an Engel class I, the surgical outcomes of 5 PIP patients were not specifically reported. In contrast, a retrospective controlled study found patients without a history of PIP (n = 17/20; 85%) were significantly more likely to become seizure free compared to a small sample of PIP patients (n = 2/3; 67%) (Kanner and Ostrovskaya, 2008). The lack of consensus with regards to seizure freedom in patients with PIP may reflect methodological differences (e.g. sample size and/or seizure classification systems). We found that neither a history nor frequency of PIP episodes significantly predicted ipsi- or contralateral memory decline at 3-and 12 months. Future research to investigate whether extra-temporal (executive function) decline is evident in PIP following TLE surgery is warranted. There are several limitations of our study due to its retrospective design. We were not able to investigate follow-up longer than 4 years due to sample attrition arising from patients being referred back to local services. Despite all patients undergoing post-surgical postoperative neuropsychiatric evaluations at 3 and 6 months, it is possible that psychiatric symptoms may have emerged later or were missed, as research indicates that psychiatric morbidity in TLE is often unrecognised or overlooked in routine neurological consultations (Kanner et al., 2000). While future research should address this limitation, these scenarios would have been the same for both patient groups. We were also unable to determine whether the development of de novo psychopathology was modified by the temporal relationship between the onset of PIP and surgical intervention. A lack of statistical power precluded an investigation into

Discussion In keeping with others, we found a greater prevalence of non-localised ictal epileptiform activity in TLE + PIP patients (Kanner and Ostrovskaya, 2008; So et al., 1990; Falip et al., 2009) compared to those without PIP. This suggests that more widely distributed epileptogenic networks may contribute to the development of PIP. Our findings also support the emerging evidence from functional neuroimaging studies that more diffuse brain abnormalities can be detected in PIP patients interictally (Seeck et al., 1999; Takeda et al., 2001). There have also been some reports of diffuse structural brain abnormalities in patients with PIP (Morrow et al., 2006; DuBois et al., 2011). However, these changes may be subtle and warrant more advanced or sensitive imaging techniques. There were no histopathological differences between the TLE + PIP and TLE-only groups. In contrast, a small study (n = 6) found PIP to be associated with temporal dysplasia (Briellman et al., 2000). Our sample was larger and applied a refined clinico-pathological classification system was applied (Blümcke et al., 2011). There were no significant differences in the other epilepsy variables between the groups. This is in consistent with previous retrospective (Devinsky et al., 1995; Falip et al., 2009) and prospective (Alper et al., 2008) studies. However, investigators have indicated that SGTCS (Devinsky et al., 1995; Alper et al., 2008), status epilepticus (Falip et al., 2009) and encephalitis (Devinsky et al., 1995; Alper et al., 2008) are risk factors for PIP. However, we lacked reliable information to subtype CNS infections (e.g. meningitis/encephalitis) or quantify the frequency of SGTCS and status epilepticus, which may explain the lack of association in our study. Our findings that a positive family psychiatric history was significantly associated with PIP, and the most common diagnosis was mood disorders, is consistent with previous research (Alper et al., 2008). Notably, no first- or second degree relative included in the analysis had a diagnosis of psychosis/schizophrenia. The genetic mechanism(s) underlying this association is unclear. Preoperative cognitive measures did not differentiate between our TLE + PIP and the TLE-only controls. While most previous studies do not report the neuropsychological status of PIP patients, a few have documented bi-temporal neuropsychological dysfunction (So et al., 1990; Christodoulou et al., 2002; Falip et al., 2009; Mathern et al., 1995; Savard et al., 1991). However, it may be difficult to compare our results to these studies given their smaller sample sizes and unspecified cognitive measures.

Postictal psychosis in temporal lobe epilepsy the possible linear relationship between PIP episodes and de novo psychopathology or seizure outcome. Further, we were unable to explore subtypes of de novo psychiatric morbidity with our sample size, but this could be addressed by larger cohort studies. In conclusion, our study has a number of clinical implications. Firstly it confirms that preoperative PIP is associated with poorly localised ictal EEG indicating that surgical candidates with may require the recording of several seizures during VEEG to determine the extent of the epileptogenic zone (Kanner and Ostrovskaya, 2008). Our data also suggest that genetic predisposing factors may increase the risk of PIP. Lastly, patients with recurrent episodes of PIP are at greater risk of developing de novo psychopathology following surgery. TLE patients with recurrent episodes of PIP should be counselled about the increased psychiatric risks following TLE surgery, and closely monitored for postsurgical psychiatric morbidity.

Acknowledgements Medical Records Department, The National Hospital for Neurology and Neurosurgery, Queen Square, London, UK, WC1N 3BG. Miss Jane De Tisi for access to the surgical database. The work was undertaken at University College London Hospital and University College London who receive a proportion of funding from the Department of Health’s NIHR Biomedical Research Centre’s Funding Scheme.

References Alper, K., Devinsky, O., Westbrook, L., et al., 2001. Premorbid psychiatric risk factors for postictal psychosis. J. Neuropsychiatry Clin. Neurosci. 13, 492—499. Alper, K., Kuzniecky, R., Carlson, C., et al., 2008. Postictal psychosis in partial epilepsy: a case—control study. Ann. Neurol. 63, 602—610. Baxendale, S., Thompson, P.J., Duncan, J.S., 2008. Improvements in memory function following anterior temporal lobe resection for epilepsy. Neurology 71, 1319—1325. Blümcke, I., Thom, M., Aronica, E., et al., 2011. The clinicopathologic spectrum of focal cortical dysplasias: a consensus classification proposed by an ad hoc Task Force of the ILAE Diagnostic Methods Commission. Epilepsia 52, 158—174. Briellman, R.S., KalninsRM, Hopwood, M.J., Ward, C., Berkovic, S.F., Jackson, G.D., 2000. TLE patients with postictal psychosis: mesial dysplasia and anterior hippocampal preservation. Neurology 10, 1027—1030. Christodoulou, C., Koutroumanidis, M., Hennessy, M.J., Elwes, R.D., Polkey, C.E., Toone, B.K., 2002. Postictal psychosis after temporal lobectomy. Neurology 59, 1432—1435. Cleary, R.A., Thompson, P.J., Fox, Z., Foong, J., 2012. Predictors of psychiatric and seizure outcome following temporal lobe epilepsy surgery. Epilepsia 53, 1705—1712. Coughlan, A.K., Hollows, S.E., 1985. The Adult Memory and Information Processing Battery (AMIPB) Test Manual. St James University Hospital, Leeds. Devinsky, O., 2008. Postictal psychosis: common, dangerous, and treatable. Epilepsy Curr. 2, 31—34.

271 Devinsky, O., Abramson, H., Alper, K., et al., 1995. Postictal psychosis: a case control series of 20 patients and 150 controls. Epilepsy Res. 20, 247—253. DuBois, J.M., Devinsky, O., Carlson, C., et al., 2011. Abnormalities of cortical thickness in postictal psychosis. Epilepsy Behav. 21, 132—136. Falip, M., Carre˜ no, M., Donaire, A., et al., 2009. Postictal psychosis: a retrospective study in patients with refractory temporal lobe epilepsy. Seizure 18, 145—149. Ferguson, S.M., Rayport, M., Gardner, R., Kass, W., Weiner, H., Reiser, M.F., 1969. Similarities in mental content of psychotic states, spontaneous seizures, dreams, and responses to electrical brain stimulation in patients with temporal lobe epilepsy. Psychosom. Med. 31, 479—498. Field, A., 2000. Discovering Statistics Using SPSS. SAGE Publications, London. Fong, G.C., Fong, K.Y., Mak, W., et al., 2000. Postictal psychosis related regional cerebral hyperfusion. J. Neurol. Neurosurg. Psychiatry 68, 100—101. Kanemoto, K., Kawasaki, J., Kawai, I., 1996. Postictal psychosis: a comparison with acute interictal and chronic psychoses. Epilepsia 37, 551—556. Kanemoto, K., Kawasaki, J., Mori, E., 1998. Postictal psychosis as a risk factor for mood disorders after temporal lobe surgery. J. Neurol. Neurosurg. Psychiatry 65, 587—589. Kanemoto, K., Kim, Y., Miyamoto, T., Kawasaki, J., 2001. Presurgical postictal and acute interictal psychoses are differentially associated with postoperative mood and psychotic disorders. J. Neuropsychiatry Clin. Neurosci. 13, 243—247. Kanner, A.M., 2009. Does a history of postictal psychosis predict a poor postsurgical seizure outcome? Epilepsy Curr. 9, 96—97. Kanner, A.M., Ostrovskaya, A., 2008. Long-term significance of postictal psychotic episodes I. Are they predictive of bilateral ictal foci? Epilepsy Behav. 12, 150—153. Kanner, A., Stagno, S., Kotagal, P., Morris, H., 1996. Postictal psychiatric events during prolonged video-electroencephalographic monitoring studies. Arch. Neurol. 53, 258—263. Kanner, A.M., Kozak, A.M., Frey, M., 2000. The use of sertraline in patients with epilepsy: is it safe? Epilepsy Behav. 1, 100—105. Kuba, R., Brázdil, M., Rektor, I., 2012. Postictal psychosis and its electrophysiological correlates in invasive EEG: a case report study and literature review. Epilepsy Behav. 23, 426—430. Leutmezer, F., Podreka, I., Asenbaum, S., et al., 2003. Postictal psychosis in temporal lobe epilepsy. Epilepsia 44, 582—590. Logsdail, S., Toone, B., 1988. Post-ictal psychoses. A clinical and phenomenological description. Br. J. Psychiatry 152, 246—252. Martinian, L., Catarino, C.B., Thompson, P., Sisodiya, S.M., Thom, M., 2012. Calbindin D28K expression in relation to granule cell dispersion, mossy fibre sprouting and memory impairment in hippocampal sclerosis: a surgical and post mortem series. Epilepsy Res. 98, 14—24. Mathern, G.W., Pretorius, J.K., Babb, T.L., Quinn, B., 1995. Unilateral hippocampal mossy fiber sprouting and bilateral asymmetric neuron loss with episodic postictal psychosis. J. Neurosurg. 82, 228—233. McGurn, B., Starr, J.M., Topfer, J.A., et al., 2004. Pronunciation of irregular words is preserved in dementia, validating premorbid IQ estimation. Neurology 62, 1184—1188. McKenna, P., Warrington, E.K., 1980. Testing for nominal dysphasia. J. Neurol. Neurosurg. Psychiatry 43, 781—788. Morrow, E.M., Lafayette, J.M., Bromfield, E.B., Fricchione, G., 2006. Postictal psychosis: presymptomatic risk factors and the need for further investigation of genetics and pharmacotherapy. Ann. Gen. Psychiatry 21 (5), 9. Nelson, H.E., 1982. The National Adult Reading Test (NART): Test Manual. NFER-Nelson, Windsor. Nishida, T., Kudo, T., Inoue, Y., et al., 2006. Postictal mania versus postictal psychosis: differences in clinical features,

272 epileptogenic zone, and brain functional changes during postictal period. Epilepsia 47, 2104—2114. Oshima, T., Motooka, H., Kanemoto, K., 2011. SPECT findings during postictal psychoses: predominance of relative increase of perfusion in right temporal lobe. Epilepsia 52, 1192—1194. Roberts, G.W., Done, D.J., Bruton, C., Crow, T.J., 1990. A ‘‘mock up’’ of schizophrenia: temporal lobe epilepsy and schizophrenia-like psychosis. Biol. Psychiatry 15, 127—143. Savard, G., Andermann, F., Olivier, A., Rémillard, G.M., 1991. Postictal psychosis after partial complex seizures: a multiple case study. Epilepsia 32, 225—231. Schlösser, R.G., Wagner, G., Sauer, H., 2006. Assessing the working memory network: studies with functional magnetic resonance imaging and structural equation modelling. Neuroscience 139, 91—103. Schulze-Bonhage, A., van Elst, L.T., 2010. Postictal psychosis: evidence for extrafocal functional precursors. Epilepsy Behav. 18, 308—312.

R.A. Cleary et al. Seeck, M., Alberque, C., Spinelli, L., et al., 1999. Left temporal rhythmic electrical activity: a correlate for psychosis? A case report. J. Neural Transm. 106, 787—794. So, N.K., Savard, G., Andermann, F., Olivier, A., Quesney, L.F., 1990. Acute postictal psychosis: a stereo EEG study. Epilepsia 31, 188—193. Takeda, Y., Inoue, Y., Tottori, T., Mihara, T., 2001. Acute psychosis during intracranial EEG monitoring: close relationship between psychotic symptoms and discharges in amygdala. Epilepsia 42, 719—724. Tarulli, A., Devinsky, O., Alper, K., 2001. Progression of postictal to interictal psychosis. Epilepsia 42, 1468—1471. Taylor, D.C., 1975. Factors influencing the occurrence of schizophrenia-like psychosis in patients with temporal lobe epilepsy. Psychol. Med. 5, 249—254. Wieser, H.G., Blume, W.T., Fish, D., et al., 2001. Commission on Neurosurgery of the International League Against Epilepsy (ILAE). ILAE Commission Report. Proposal for a new classification of outcome with respect to epileptic seizures following epilepsy surgery. Epilepsia 42, 282—286.