Characteristics of postictal psychosis in a psychiatric center

Authors


address: Hsing-cheng Liu, Department of Psychiatry, Taipei City Psychiatric Center, 309 Sung-teh Road, Taipei, Taiwan. Email: apple@mail.tcpc.gov.tw

Abstract

Abstract The clinical characteristics of 12 cases of postictal psychosis treated at Taipei City Psychiatric Center, Taipei, Taiwan, were retrospectively reviewed. Increased seizure frequency, especially with generalized tonic–clonic seizures, was the major risk factor predisposing to postictal psychosis. The psychotic symptoms were variable with delusions and/or hallucinations. These patients showed a much longer history of epilepsy (21.9 ± 10.7 years) prior to the development of postictal psychosis than has been previously reported. The possible mechanisms in the pathophysiology of psychosis in epileptics were discussed.

INTRODUCTION

Epileptic patients have a higher incidence of various psychopathologies. The relationship between psychosis and epilepsy remains unclear.1 Antagonism between epilepsy and psychosis and the problems associated with the phenomenon of forced normalization are well known.2–4 The link between epilepsy and psychosis was established by Slater et al.5 in 1963. The incidence of psychosis in epileptic patients has been reported to vary from 2 to 60% in different study populations.5–7 Most studies also showed a higher rate of psychosis in temporal lobe epilepsy.

The classification of psychosis in epileptic patients is controversial, and may be based on psychopathology, or on an etiological or longitudinal relationship between epileptic seizures and psychosis. According to the temporal relationship to seizures, the psychosis may be classified as ictal, postictal, peri-ictal, or interictal.7 Most previous studies have focused on chronic interictal psychosis. Postictal psychosis was first thoroughly investigated by Logsdail and Toone in 1988.8 According to their criteria, postictal psychosis occurs in a distinct group of patients and has a different clinical presentation and course from interictal psychosis.

Therefore, we tried to examine the characteristic of postictal psychosis patients in our setting, a psychiatric center. We employed the criteria of Logsdail and Toone to survey all epileptic patients treated during the period 1988 to 1999 in Taipei City Psychiatric Center, Taipei, Taiwan.

METHODS

We retrospectively reviewed the medical records of patients with the diagnosis of epilepsy and psychosis between 1988 and 1999. Each record was reviewed by two neurologists and two psychiatrists.

The criteria for a diagnosis of postictal psychosis were published by Logsdail and Toone8 and required that four of the following criteria be met. First, the episode of confusion or psychosis manifested immediately upon a seizure or emerged within 1 week of the return of apparently normal mental function. Second, the psychosis had a minimum length of 24 h and a maximum length of 3 months. Third, the mental state was characterized by one of the following: (i) clouding of consciousness, disorientation, or delirium; (ii) delusions, hallucinations, in clear consciousness; (iii) a mixture of (i) and (ii). Fourth, the exclusion criteria was (i) anticonvulsant toxicity; (ii) interictal psychosis; (iii) electroencephalogram (EEG) evidence of minor status; (iv) recent history of head injury, or alcohol or drug intoxication.

The diagnosis of postictal psychosis required a consensus agreement between psychiatrists. The epileptic seizure classifications were based on the 1989 International League Against Epilepsy criteria9 and all were also made based on consensus agreement between neurologists.

RESULTS

A total of 12 cases (eight males, four females) fulfilled the criteria of postictal psychosis. The description of each case is summarized in Table 1. The average age was 48.7 ± 12.3 years. The age of onset of epilepsy in these patients varied from 3 to 49 years of age (mean ± SD, 20.7 ± 5.1 years). The baseline epileptic seizure type was generalized tonic–clonic convulsion (GTC) in eight patients and complex partial seizure (CPS) with or without secondary generalization in four patients.

Table 1.  Clinical characteristics of 12 patients with postictal psychosis
PatientNo. 1No. 2No. 3No. 4No. 5No. 6No. 7No. 8No. 9No. 10No. 11No. 12
  1. CPS, complex partial seizure; GTC, generalized tonic-clonic seizure; AH, auditory hallucination; VH, visual hallucination; HAL, haloperidol; BZD, benzodiazepines; CBZ, carbamazepine; DPH, diphenylhydantoin; VPA, valproic acid; AVM, arteriovenous malformation; NA, not available or not assessable; P, present but could not be assessed accurately; d, days; yr, years; m, month.

Epilepsy
onset
301614143638102149740
Psychosis
onset
553947382946504029533055
Baseline
EEG
NANADiffuse slow
wave
Left
tempoparietal
slow wave
Diffuse slow
wave
NADiffuse slow
wave
NANANADiffuse slow
wave
Right rolandic
spike wave
Seizure type
(before)
GTCGTCGTCCPS, GTCCPS, GTCGTCCPS, GTCGTCGTCCPS, GTCGTCGTC
Seizure type
(prepsychotic)
NANAGTCCPSCPS, GTCGTCGTCGTCGTCGTCGTCGTC
Increased
seizure
frequency
++++++++++
Lucid
interval
4 dP7 d6 d2 dPP18 hrPPP1 d
Psychosis
duration
7 d37 d30 d4 d6 d2 d3 dNA5 d1 d5 d5 d
Psychotic
symptoms
Persecutory
delusion,
jealous
delusion, AH
Persecutory
delusion, AH
Somatic
delusion, AH,
VH,
depression
Persecutory
delusion, AH
Mania,
grandiose
delusion,
confusion
Mania,
confusion
Persecutory
delusion,
AH, VH
AH,
confusion
Persecutory
delusion,
AH, VH,
depression
Persecutory
delusion, AH
Reference
delusion, AH
Mania,
persecutory
delusion, AH
EEG during
psychosis
NANADiffuse slow
wave
Diffuse slow
wave
Diffuse slow
wave
NormalDiffuse slow
wave
NADiffuse slow
wave
Diffuse slow
wave
Diffuse slow
wave
Normal
TreatmentDPHDPH, HAL,
BZD
HAL, CBZCBZHAL, CBZ,
BZD
CBZ, DPHHAL, BZDCBZ, DPHDPH, BZDVPA, DPHHAL, CBZ,
DPH
HAL, DPH
OthersLeft frontal and
temporal
encephalomalacia
Right
temporal
AVM
Meningitis at
6 year-old
History of
alcoholism
Follow up
duration
2 yr7 yr2 yr 10 m5 yr 10 m1 yr 9 m3 yr6 yr1 yr3 yr 8 m3 yr 3 m7 yr1 yr
Recurrence322433
Familiar
psychosis

The age onset of first postictal psychosis variedfrom 29 to 55 years (mean ± SD, 42.6 ± 9.9 years). The interval between the onset of epilepsy and the first episode of postictal psychosis ranged from 4 to 40 years (mean ± SD, 21.9 ± 10.7 years). The prepsychotic seizure type could be assessed in 10 patients with GTC in eight, CPS in one and CPS with secondary GTC in one. Ten patients had an obviously increased seizure frequency before the onset of postictal psychosis. A lucid interval between the cessation of seizure episodes to the onset of psychosis was noted in all cases and the records of six patients had detailed information on these interval times which lasted from 18 h to 7 days (mean ± SD, 3.46 ± 2.4 days). The follow-up duration ranged from 1 to 7 years. During follow-up, six patients had 2–4 attacks of postictal psychosis and six patients had only one attack. Electroencephalograms were obtained during the period of postictal psychosis in nine cases and seven showed diffuse slow waves and two were normal.

The psychotic symptoms included auditory hallucinations in 10 patients, visual hallucinations in three, persecutory delusions in seven, reference delusion in one, grandiose delusion in one, jealous delusion in one, and somatic delusion in one. Manic-like mood change was noted in three patients and depressive features in two patients. The psychotic symptoms lasted from 1 to 37 days (mean ± SD, 21.92 ± 10.76 days). Antipsychotics were prescribed in six cases, and benzodiazepines were used in four cases. No patient needed long-term antipsychotics treatment.

Central nervous system pathology included left frontotemporal encephalomalacia due to sequela of traumatic brain injury in one, history of meningitis in one, and right temporal arteriovenous malformation in one. Although case 12 had a history of alcoholism, the attack was not related to recent alcohol use. None of these 12 cases had family history of psychosis.

DISCUSSION

Psychotic disorders can occur in chronic epileptic patients and have a higher incidence in long-term seizure disorders.5 In the review of Szabo et al., 93% of cases of postictal psychosis occurred in patients older than 18 years of age and the average at onset was 38.3 ± 9.8 years.10 This finding is similar to the results of the present study where the earliest onset of psychosis was 29 years and the average onset was 42.6 ± 9.9 years. However, the interval between the onset of epilepsy and the first episode of ranged from 4 to 40 years (mean ± SD, 21.9 ± 10.7 years), which was much longer in the present series than in previous reports (range 13–15 years),10 and the interval was longer than 20 years in 67% of our patients. The prevalence of postictal psychosis is uncertain. The previous reported frequency under specific investigational settings ranged from 1.3 to 6.4%.11,12 Taipei City Psychiatric Center is the largest psychiatric referral center in the Taipei area; these 12 patients were all referred due to the psychotic problems. Therefore, we might see the most severe cases and it is difficult to estimate the prevalence of postictal psychosis in Taiwan.

In addition to its association with a long epileptic history, some studies have found that male patients had a higher risk of postictal psychosis with up to 65% of cases occurring in men.8,12–16 In the present study, 67% of cases were male and this is compatible with most previous reports.8,12–16 However, clear determination of the gender effect needs further investigation with a larger patient sample. As in previous studies,8,12–16 no patient in the present study had a family history of psychosis.

Increased seizure frequency prior to the onset of postictal psychosis was noted in 10 (83%) of our cases. Increased seizure frequency seemed to predispose patients to an attack of postictal psychosis, especially serial attacks of moderate degree.8,12,13,15,16 Although there were two patients’ prepsychotic seizure type could not be assessed from retrospective chart review, most of the prepsychotic epileptic seizures in this series were of the GTC (8/10, 80%) type. This finding is compatible with the review of Szabo et al., which found that GTC was the major risk for postictal psychosis, with 81% of cases having episodes of GTC seizures before the onset of psychosis.10 Kanner et al. proposed that increased CPS attack was a risk factor for postictal psychosis in their controlled study.12 More than half of the EEG showed increased interictal epileptiform activities during the period of postictal psychosis.10 Some EEG data were not available in this retrospective study. However, no patient showed epileptiform discharges and two patients even had normal EEG as the pattern of forced normalization. Surface EEG could not reflect the actual deep neuronal activities. Therefore, it is reasonable to suspect whether postictal psychosis is an ictal event. So et al. reported a single case in which continuous sterostatic depth and epidural EEG recordings confirmed that the phenomenon was postictal rather than an ictal event.13

The lucid interval between the end of seizure and psychosis is invariably present. All of the present patients had lucid intervals but only six of them could be accurately assessed from chart review. The intervals varied from 18 h to 7 days in the six patients. This characteristic feature can be differentiated from postictal confusion state. Recurrent postictal psychosis was noted in six (50%) of our patients and no patient developed into chronic interictal psychosis during the 1–7 year follow up. Logsdail and Toone found that only two of 14 cases develop into chronic interictal psychosis during a mean follow-up of 8 years.8 Most postictal psychosis patients remained in possibly recurrent postictal psychosis and this was also strong evidence of a different disease entity from a clinical point of view.16

In the present series, the psychotic symptoms, which occurred in the course of postictal psychosis, were variable and included delusions and hallucinations. However, thought disorders were not observed in our patients. Thought disorders have rarely been reported in postictal psychosis11 and this finding suggests a differentiation from schizophrenia. Mood symptoms have been previously reported in 75% of patients.8,12,15 However, only five of our patients reported experiencing mood symptoms. The psychotic symptoms in postictal psychosis are not persistent. In the present study, psychotic symptoms lasted 1–37 days and this finding is similar to previous results.12,15,16 The management for each episode of postictal psychosis varied, mostly depending on the clinical symptoms severity and duration. Antipsychotics were prescribed in six patients and benzodiazepines in four. In this series, antipsychotics were prescribed in patients with longer psychotic duration or increased severity. Kanner et al. recommended the use of low dose antipsychotics treatment.12 However, most patients’ psychotic symptoms resolve spontaneously.8,12 Lancman et al. observed that the clinical pictures were similar in each recurrence in each patient and they suggested the use of chloral hydrate or benzodiazepines to the recurrent cases.16

The causes of postictal psychosis are not fully understood. In the present study, there were three CNS pathological conditions: one left frontotemporal encephalomalacia, one history of meningitis, and one right temporal arteriovenous malformation. It was difficult to make specific conclusions from these three cases. However, Manchanda et al. suggested that decreased activity or hypofunction of the right temporal lobe might play a role in the pathogenesis of postictal psychotic disorders.11 Dopamine might play a role in anticonvulsant actions.17–20 It has therefore been postulated that a hyperdopaminergic state after repetitive seizures plays an important role in postictal psychosis. Ring et al. reported that the binding of 123I-IBZM to striatal D2 receptor is significantly reduced in patients with epileptic psychosis.21 There were seven patients with epileptic psychoses in their series. Reviewing the clinical descriptions in their report, two cases had residual symptoms since their last psychotic episodes years ago, thus these two cases had similar features to chronic interictal psychosis. Five other cases in their series had recurrent psychosis or were in the first psychotic episode. These five cases had a disease history and presentation similar to postictal psychosis. The lower striatal D2 binding may reflect an increased activity of the nigrostriatal dopamine pathway.22,23 Baumgartner et al. also demonstrated frontal hyperperfusion, especially in the mesial frontal areas in postictal psychosis.24 This finding implied that the frontal–subcortical structures are involved in the development of postictal psychosis. Other neurotransmitters, such as γ-aminiobutyric acid or N-methyl-D-aspartate, may also play a role in the pathogenesis of postictal psychosis; however, the evidence for this is still circumstantial.10

Our postictal psychosis patients shared the majority of characteristics of previous series except for a much longer interval between the onset of epilepsy and the first episode of postictal psychosis. Kindling is an important mechanism for the development of psychosis in epileptic patients in altering the neuromodulatory function of limbic nuclei and with regard to dopamine.25,26 This longer interval in our patients provides additional evidence of the kindling mechanism in the pathogenesis of psychosis in chronic epileptics through long-term epileptic discharges. Although the detailed pathophysiology of psychosis in epileptics remains obscure, frontal–subcortical circuits and dopaminergic activity are involved. These resemble the present known model of psychosis.27,28 Exploring the causes of postictal psychosis may provide a better understanding of the nature of functional psychosis as in schizophrenia.

Acknowledgements

This study was supported by Taipei City Psychiatric Center Grant 8803.

Ancillary