Psychotic illness in patients with epilepsy

Abstract

Apart from the rather rare ictal psychotic events, such as non-convulsive status epilepticus, modern epileptic psychoses have been categorized into three main types; chronic and acute interictal psychoses (IIPs) and postictal psychosis (PIP). Together, they comprise 95% of psychoses in patients with epilepsy (PWE). Four major questions, that is, “Is psychosis in PWE a direct consequence of epilepsy or schizophrenia induced by epilepsy?”, “Is psychosis in PWE homogeneous or heterogeneous?”, “Does psychosis in PWE have symptomatological differences from schizophrenia and related disorders?”, “Is psychosis in PWE uniquely associated with temporal lobe epilepsy (TLE)?” are tried to be answered in this review with relevant case presentations. In the final section, we propose a tentative classification of psychotic illness in PWE, with special attention to those who have undergone epilepsy surgery. Psychotic disorders in PWE are often overlooked, mistreated, and consequently lingering on needlessly. While early diagnosis is unanimously supported as a first step to avoid this delay, necessity of switching from antiepileptic drugs with supposedly adverse psychotopic effects. to others is more controversial. To elucidate the riddle of alternative psychosis, we need badly further reliable data.

Introduction

If asked to name the three most important studies in the modern history of epileptic psychosis, those of Landolt, Slater and Beard, and Logsdail and Toone would be the most likely candidates [Landolt, 1953, 1963; Slater and Beard, 1963; Logsdail and Toone, 1988]. This is because, apart from the rather rare ictal psychotic events, such as nonconvulsive status epilepticus, modern epileptic psychoses have been categorized into three main types; chronic and acute interictal psychoses (IIPs) and postictal psychosis (PIP). In 1953, Landolt stressed the seesaw relationship between epileptic discharges on EEG and psychosis, and proposed the concept of forced normalization, which led to the concept of alternative psychosis proposed by Tellenbach [Tellenbach, 1963], while Slater and Beard produced a rather comprehensive report of chronic psychoses in patients with epilepsy (PWEs) [Slater and Beard, 1963]. However, it was as late as 1988 before the concept of PIP was clearly demarcated by Logsdail and Toone, which soon turned out to be a very important concept in modern understanding of epileptic psychosis [Logsdail and Toone, 1988]. Together, the psychotic states described in those three papers, that is chronic and acute IIPs, along with PIP, comprise 95% of psychoses in PWEs [Kanemoto et al. 2001]. It should be noted that psychotic disorder is a seriously challenging problem among patients with a long-standing, intractable epilepsy. In this specific subgroup, patients are uniquely susceptible to psychiatric problems including psychotic disorder [Kanner et al. 2004]

Furthermore, if forced to choose not three but only one study, Slater and Beard’s [Slater and Beard, 1963] would probably be top rated, not because it correctly answers all of the pertinent questions, but because it raised nearly all of the important questions. In Table 1, major questions and answers presented in Slater and Beard’s paper, as well as the most likely answers to those questions based on present understanding are listed. These can be summarized into the following four major questions. Is psychosis in PWEs a direct consequence of epilepsy or schizophrenia induced by epilepsy? Is psychosis in PWEs homogeneous or heterogeneous? Does psychosis in PWEs have symptomatological differences from schizophrenia and related disorders? Is psychosis in PWEs uniquely associated with temporal lobe epilepsy (TLE)?

Table 1.

Slater and Beard’s questions, answers, and the most likely answers based on the status quo.

• Does epilepsy and schizophrenia coexist as a chance combination?	→	No	No
• Does pre-existing epilepsy predispose the development of schizophrenia?	→	No	Yes
• Could occurrence of epileptic seizures be the cause of the psychosis?	→	Yes / No	Yes / No
• Could the basic disorder of function, which manifests itself in epileptic seizures, also cause the psychosis?	→	Yes	Yes / No
• Possibilities of heterogeneity	→	No	Controversial
• Psychoses observed symptomatically deviate from schizophrenia norms	→	Yes	Controversial
• Does the risk of psychosis differ by type of epilepsy? (Is psychosis in a special relationship with temporal lobe pathology? )	→	Yes	Controversial
		S&B’s answers	Status quo

Is psychosis in PWEs explicable as a type of schizophrenia incidentally induced by epilepsy?

If ‘epileptic psychosis’ is understood as a psychotic state in a causal relationship with epilepsy, there are two major historical arguments in opposition to this concept: Argument A is that concurrence of psychotic states and epilepsy is only coincidental; and Argument B is that epilepsy is nothing but a nonspecific precipitating factor leading to psychosis in patients genetically predisposed to psychosis.

Argument A is becoming increasingly improbable. Bredkjaer and colleagues concluded that incidences of the spectrum of nonorganic nonaffective psychoses (16.8%) and schizophrenia (6.0%) were significantly increased in both men and women PWEs, even after exclusion of individuals diagnosed as suffering from a learning disability or substance misuse [Bredkjaer et al. 1998]. Qin and colleagues found an increased risk of schizophrenia (relative risk 2.48) and schizophrenia-like psychosis (relative risk 2.93) in individuals with a history of epilepsy [Qin et al. 2005]. According to Vestergaard and colleagues, even a history of febrile seizures was associated with a 44% increased risk of schizophrenia [Vestergaard et al. 2005]. Furthermore, several studies have noted that the annual incidence of schizophreniform psychosis is twice as high as that of schizophrenia in the general population (Table 2) [Linsay et al. 1979; Onuma et al. 1995; Bredkjaer et al. 1998; Tadokoro et al. 2007; World Health Organization, 1992].

Table 2.

Annual incidence of newly occurring cases of psychotic illness.

Schizophreniform psychosis in patients with epilepsy
Lindsay et al. [1979]	0.75%	Temporal lobe epilepsy; child)
Onuma et al. [1995]	0.30%	General epilepsy outpatient; adult
Bredkjaer et al. [1998]	0.38%	General epilepsy patient
Tadokoro et al. [2007]	0.42%	General epilepsy outpatient; adult (ours)
Schizophrenia in general population
World Health Organization [1992]	0.22%	[median value] (15-54 years old)

In contrast, in opposition to Slater and Beard’s original contention, Argument B has accumulated strong supportive evidence, especially from recent large-scale family history studies. Except for the rather ambiguous report of Flor-Henry [Flor-Henry, 1969], an increased incidence of psychotic episodes has been shown in family members of PWEs and psychosis (Table 3) [Adachi et al. 2000; Jensen and Larsen, 1979; Qin et al. 2005]. A simple deduction from these findings of patients with psychosis and epilepsy suggests that epilepsy is nothing but a nonspecific precipitating factor to psychosis in an already genetically predisposed population.

Table 3.

Psychotic episodes in family members in patients with epilepsy and psychosis.

Study	Incresed psychotic episodes in families	n	incidence
Flor-Henry [1969]	No	100	No incidence reported
Jensen and Larsen [1979]	Yes	74	Odds ratio = 1.66
Ad hi lAdachi et al. [2000]	Yes	904	n = 904; odds ratio = 18.5 [interictal psychosis]; 16.5 [postictal psychosis]
Qin et al. [2005]	Yes	2.3 million	Odds ratio 7.6

However, when attention is focused on individual cases of ictal and postictal psychotic states, this conclusion appears to be implausible [Trimble, 1991; Kanemoto, 2011]. Psychotic states closely parallel epileptic activity on EEG in aura continua or spike-wave stupor. Also, nuclear cases of PIP are closely associated with seizure activities, and can be stopped in most cases after eliminating seizures with either medication or surgical intervention. This leads us immediately to the second question regarding the heterogeneity of psychosis in PWEs.

Is psychosis in PWEs unitary or multiple?

One of the key pieces clinical data for answering this question is derived from studies on age at onset of psychosis. It is well known that there is a long interval (average 15 years) between epilepsy and psychosis onset [Trimble, 1991; Kanemoto et al. 2001; Tadokoro et al. 2007]. Independent Japanese studies that directly compared age at psychosis onset between schizophrenia, IIP, and PIP confirmed that the youngest age was for onset of schizophrenia and the oldest for that of PIP (Figure 1) [Kanemoto et al. 1996a, 2001; Adachi et al. 2000, 2002; Oshima et al. 2006; Tadokoro et al. 2007].

Figure 1.

Age at psychosis onset in patients with epilepsy and patients with schizophrenia. (Data from Kanemoto et al. [1996a, 2001], Adachi et al. [2000, 2002], Oshima et al. [2006], and Tadokoro et al. [2007].)

Adachi and colleagues investigated the interval between onset of epilepsy with various clinical backgrounds and that of psychosis, and found the shortest interval for idiopathic generalized epilepsy (IGE), an intermediate interval for epileptic encephalopathy, and the longest interval for symptomatic partial epilepsy [Adachi et al. 2010]. In addition, they compared the intervals with regards to intelligence quotient (IQ). Patients with apparent mental retardation showed the longest interval, while patients with normal IQ showed the shortest. Similarly, a comparison between patients with and without family history revealed that absence of family history was associated with a shorter interval.

Viewing these findings together, age at psychosis onset in patients with IGE, normal intelligence, and family history of psychotic episodes becomes closer to that in those with schizophrenia. That is to say, the effect of epilepsy may play a more essential role in patients with no family history of psychosis and a damaged brain. In contrast, the impact of epilepsy may only be subsidiary in the pathogenesis of psychosis in patients with IGE.

If genetic predisposition is allowed to be regarded as a main biological determinant for the development of schizophrenia, the different impacts of genetic and epileptic determinants can be schematized, as shown in Figure 2. In this diagram, the impact of genetic factors becomes greater in the lower left quadrant, while the relationship between epileptic activity and psychosis becomes more direct in the upper right quadrant.

Figure 2.

Different impacts of genetic and epileptic determinants among different subsets of psychosis in patients with epilepsy.

Last but not least, it should be mentioned that IIP and PIP are now acknowledged to be co-existent in 3–8% of PWEs and psychosis [Tarulli et al. 2001; Kanemoto, 2002, 2011; Adachi et al. 2002]. Based on the study of such bimodal psychoses, Adachi and colleagues pointed out that certain clinical backgrounds, such as family history of psychosis, predominance of TLE, the presence of complex partial seizures, appear to have equal influence on postictal and IIPs and stressed common determinaters in both types of psychosis [Adachi et al. 2002]. However, when it comes deeper into details, besides later age at psychosis onset, family history of manic depressive illness and IGE are more frequently encountered in patients with PIP. In the last section of the present review, this topic and discussion will be returned to shortly. Another important clinical determinant possibly contributory to the heterogeneity of psychosis in PWEs is corresponding neuroanatomical correlates, which will be discussed in the next section.

Do psychotic symptoms in epilepsy differ from those in schizophrenia and related disorders?

Originally, Slater and Beard affirmatively answered this question, pointing out that affective responsiveness tends to be preserved, while personality, even in later stages, is left substantially undamaged. However, only five control studies are available regarding this topic (Table 4) [Perez and Trimble, 1980; Toone et al. 1980; Oyebode and Davison, 1989; Matsuura et al. 2004; Tadokoro et al. 2007]. The first four of those studies examined patients at an undetermined point after psychosis onset, while the last one (ours) used rating scales at the first examination after psychosis started and before treatment was initiated.

Table 4.

Do psychotic symptoms differ between psychosis in PWEs and schizophrenia?

Study	Method	Index group	Control group	Answer to the question
Perez and Trimble [1980]	PSE	PWE	Schizophrenia	Yes
Toone et al. [1980]	PSE	PWE	Schizophrenia and related disorder	Yes
Oyebode and Davison [1989]	PSE	PWE	Schizophrenia	Yes
Matsuura et al. [2004]	OPCRIT	IIP, PIP	Schizophrenia and related disorder	No
Tadokoro et al. [2007]	PANSS	IIP	Schizophrenia	Yes

IIP, interictal psychosis; PANSS, Positive and Negative Syndrome Scale; PSE, Present State Examination; PWE, patient with epilepsy; PIP, postictal psychosis.

Perez and Trimble stressed that affective components supervene in patients with epileptic psychosis [Perez and Trimble, 1980], while Toone and colleagues pointed out better premorbid personality trait, more paranoid delusions, and fewer catatonic features [Toone et al. 1980]. Oyebode and Davison also reported that delusions of passivity were significantly more often experienced by patients with schizophrenia than by those with epilepsy and psychosis [Oyebode and Davison, 1989]. Our recent study [Tadokoro et al. 2007], which utilized the Positive and Negative Syndrome Scale (PANSS) rating scales, reconfirmed that negative symptoms are less apparent in patients with epileptic psychosis.

On the other hand, a strong opposing view was expressed by Matsuura and colleagues, who stated that these differences are only quantitative and not qualitative in nature [Matsuura et al. 2004]. Exploratory factor analysis revealed that both groups shared a similar factor profile, although the negative symptoms, e.g. blunted affect, as well as positive symptoms, e.g. third-person auditory hallucinations, were less severe in PWEs. This view is in agreement with the critical review of Mace, which regarded the asserted difference in psychotic symptoms between PWEs and schizophrenia as a nonvalidated speculation [Mace, 1993].

Figure 3 demonstrates the correlation between epileptic psychoses and Diagnostic and Statistical Manual of Mental Disorders (DSM)-based subcategories of schizophrenia and related disorders [American Psychiatric Association, 1994]. It is well known that the duration of psychotic episodes has a decisive impact on how psychosis is manifested. Thus, the duration of psychosis should be carefully matched before any conclusion is drawn in the future. Otherwise, comparisons among different reports, as well as between index and control groups may be obscured.

Figure 3.

Duration of illness of DSM-IV psychoses and corresponding epileptic psychoses psychoses. (Reproduced with permission from Kanemoto et al. [2008].)

Is TLE specially associated with psychosis in PWEs?

Now, we turn to the last question. Does the risk for psychosis differ by type of epilepsy? Qin and colleagues concluded, based on data derived from the Denmark National Study, that the increased risks for schizophrenia or schizophrenia-like psychosis did not differ by type of epilepsy [Qin et al. 2005]. However, their conclusion seems rather conflicting after surveying previous reports on this topic. The incidence of TLE in PWEs amounts to 67.5% when including results of previous reports (Table 5) [Gastaut et al. 1956; Bartlet, 1957; Guerrant, 1962; Slater and Beard, 1963; Boudin et al. 1963; Bruens, 1971; Shukla, 1979; Perez and Trimble, 1980; Parnas et al. 1982; Garyfallos et al. 1988; Adachi et al. 2000; Kanemoto et al. 2001; Matsuura et al. 2004]. Although the study of Qin and colleagues seemed to be more reliable than these previous studies at first glance owing to the overwhelming number of participants and the strict study design, the credibility of their data is diminished to some extent because epilepsy types were judged as unclassified in two thirds of their patients.

Table 5.

Incidences of temporal lobe epilepsy in patients with epilepsy and interictal psychosis.

Gastaut et al. [1956]	63% (n = 83)
Bartlet [1957]	88% (n = 8)
Guerrant [1962]	88% (n = 8)
Slater and Beard [1963]	80% (n = 69)
Boudin et al. [1963]	70% (n = 27)
Bruens [1971]	94% (n 16)
Shukla et al. [1979]	79% (n = 14)
Perez and Trimble [1980]	71% (n = 24)
Parnas et al. [1982]	86% (n = 29)
Garryfallos et al. [1988]	100% (n = 9)
Adachi et al. [2000]	65% (n = 246)*
Kanemoto et al. [2001]	56% (n = 132)
Matsuura et al. [2004]	53% (n = 58)
TOTAL	67 5% ( n = 477)

Supplement to Trimble’s [1991] table

^*Patients with complex partial seizures

Furthermore, IIP occurred more frequently in patients with TLE than in those with extratemporal focal epilepsies (Table 6) [Gibbs, 1951; Sherwin et al. 1982; Kanemoto et al. 2001]. Except for Gudmundsson’s study, which reported a rather implausible proportion of more than 50% of IIP among patients with generalized epilepsy [Gudmundsson, 1966], the prevalence of IIP is higher in patients with TLE than in those with generalized epilepsy. In our series [Kanemoto et al. 2001] as well as that of Shukla and colleagues [Shukla et al. 1979], this difference was statistically significant.

Table 6.

Comparisons of incidences of psychosis as a function of epilepsy subgroups.

	TLE	Other focal epilepsies
Gibbs [1951]	17%	2% 1
Sherwin et al. [1982]	9% (7/80)	0% (0/42)
Kanemoto et al. [2001]	14% (74/529)	4% (35/902)*
	TLE	GE
Gudmundsson [1966]	7% (5/71)	54% (38/71)2
Shukla et al. [1979]	18% (11/62)	4% (3/70)*
Sengoku et al. [1983]	6% (21/350)	4% (14/326)
Kanemoto et al. [2001]	14% (74/529)	2% (12/689)**

Supplement to Trimble’s [1991] table¹, incalculable;², GE apparently includes cases with secondarily generalized seizures; *p < 0.05; **p < 0.01.

GE, generalized epilepsy; TLE, temporal lobe epilepsy.

Previous reports unanimously suggest even higher proportions of TLE in patients with PIP. The ratio of TLE in previous cases is as high as 81% (Table 7) [Logsdail and Toone, 1988; Savard et al. 1991; Lancman, 1994; Umbricht et al. 1995; Devinsky et al. 1995; Kanemoto et al. 1996a, 2008; Kanner et al. 1996; Adachi et al. 2002]. However, as Alper and colleagues pointed out, frontal involvement might have remained unnoticed in those reports [Alper et al. 2008]. In addition, direct comparisons between PIP and IIP confirmed that the prevalence of TLE was significantly higher in patients with PIP than in those with IIP (Table 8) [Adachi et al. 2002; Kanemoto et al. 2008]. These data suggest that a long duration of epileptic activity serves as a fertile background for development of psychosis, especially in patients with TLE (Figure 4). However, some notable exceptions to this rule were reported even in patients with PIP [Sakakibara et al. 2012; Chakrabarti et al. 1999].

Table 7.

Incidences of TLE in patients with epilepsy and postictal psychosis.

Logsdail and Toone [1988]1	79% (11/14)
Savard et al. [1991]	100% (9/9)
Lancman et al. [1994]2	100% (7/7)
Umbricht et al. [1995]	– (?/8)
Devinsky et al. [1995]3	90% (18/20)
Kanemoto et al. [1996a]4	– [30/30]
Kanner et al. [1996]3, 4	– [11/11]
Adachi et al. [2002]5	75% (27/36)
Kanemoto et al. [2008]	77% (41/53)
TOTAL	81.2% (113/139)

1, partial epilepsy; 2, recurrent cases only; 3, seizure monitoring unit; 4, analysis limited to TLE; 5: complex partial seizures instead of TLE.

TLE, temporal lobe epilepsy.

Table 8.

Incidences of temporal lobe epilepsy in interictal psychosis (IIP) and postictal psychosis (PIP).

	PIP	IIP
Adachi et al [2002]	75% (27/36)	63% (141/224)
Kanemoto et al. [2008]	77% (41/53)	53% (74/140)
TOTAL	76% (68/89)	59% (215/364)*

^*p = 0.02

Figure 4.

Induced schizophrenia or epileptic psychosis determined by association with temporal lobe epilepsy (TLE).

A series of MRI studies conducted in patients with TLE and psychosis support this idea. In this group of patients, not only atrophic changes that encompassed the medial temporal lobe structures, but also those which extend to the lateral temporal and extratemporal regions, have often been found [Kanemoto et al. 1996b; Sundram et al. 2010]. Recently, Gutierrez-Galve and colleagues provided further evidence to specify anatomic substrate [Gutierrez-Galve et al. 2012]. In patients with TLE and IIP, cortical thickness proved to be reduced in the inferior frontal gyrus, implicated also in schizophrenia. The authors regarded this as further supporting evidence that IIP reflects the interplay of psychosis-related genetic factors and the cumulative effects of seizure activity on the brain.

Illustrative cases

Case 1: Chronic interictal psychosis

The first case was a 17-year-old male high-school student at the first contact, who was qualified as a middle-class karate wrestler. His uncle suffered from schizophrenia and committed suicide. At the age of 15, this high-school student experienced his first seizures, which initially occurred exclusively during sleep. Typical complex partial seizures with oral automatism and postictal confusion started at the age of 16. No auras preceded the complex partial seizures. With carbamazepine, seizure frequency decreased dramatically from weekly to monthly. However, seizures continued to occur monthly, even though all available antiepileptic medications were tried thereafter.

Surface EEG revealed predominantly RaT sharp waves with rare LaT sharp waves. No lesions were found in MRI or neurological findings. Full scale IQ was 99. The patient became increasingly irritable and somewhat paranoiac. Finally, he seriously injured a school comrade following the slightest provocation at the age of 17. At the age of 20, both parents and the patient asked us to consider surgical intervention. We explained the increased risk for postoperative psychiatric complications and predicted pending difficulties during seizure monitoring, because of his explosive and vulnerable character traits.

The parents and patient consulted another surgical center, which resulted in immediate presurgical evaluations. During seizure monitoring, a left medial temporal origin of complex partial seizures was confirmed, however, the patient escaped from the hospital on the 7th day after initiation of seizure monitoring. Upon arrival at our institution, the patient insisted that several staff from the surgical center had shadowed him and he was in a state of extreme agitation. Although the patient became swiftly composed with a small amount of dopamine-blocker, vague feelings of being watched and followed by someone could not be eliminated completely, even under antipsychotic medication.

At the age of 21, after all of our warnings, the patient and family decided to undergo an operation. After surgery, the patient became completely seizure free, although mental status was deteriorated drastically. One day, he suddenly appeared in the first author’s consultation room, saying, ‘Dr. A [the neurosurgeon of the surgical center] asked me to come and get an operation. My mother opposes it, but I cannot go anywhere without surgery. I came here to thank you for all you have done for me before going to the surgical center’. A half year after that visit, even though the patient regularly received adequate amounts of a dopamine blocker, he attacked his neighbors in response to verbal hallucinations. Over the most recent 10 years, the patient has remained continuously psychotic. The total duration of admission has amounted to more than 6 years.

The salient features of this case are as follows: a short interval between epilepsy and psychosis onset (5 years), family history of schizophrenia, and absence of structural brain damage (normal IQ and MRI findings). In this case, we presume that a genetic factor is predominating. However, without a past medical history, it is difficult to distinguish him from other patients with functional psychosis (Figure 5)

Figure 5.

Case 1: IIP.

IIP, interictal psychosis; IQ, intelligence quotient; P, psychosis; PIP, postictal psychosis; TLE, temporal lobe epilepsy.

Case 2: Repetitive acute interictal psychoses inclusive of alternative psychosis

Case 2 was a 29-year-old female pharmacist at the first contact who was an heiress of a famous pharmacy founded in the medieval era. At the age of 3 years old, prolonged febrile convulsions with subsequent transient paralysis of the right upper extremities occurred. No family history was confirmed.

This patient had been experiencing complex partial seizures as well as epigastric aura since she was 12. Complex partial seizures occurred every week despite administrations of all available antiepileptic medications in Japan. Complex partial seizures became secondarily generalized once every 3 or 4 months.

At the age of 22, a trial of high-dose phenytoin monotherapy (25 μg/ml) succeeded to completely control her seizures. However, suspicions that every comment of those in the vicinity were directed at her and that an eye was kept on her constantly insidiously haunted the patient. Meanwhile, these ideas of reference became more and more intensified, culminating in a delusional belief that her parents had implanted a communications apparatus into her heart and spied on her every thought. By replacing phenytoin with carbamazepine and beginning haloperidol (3 mg/day), these bizarre ideas completely disappeared within 2 weeks. However, complex partial seizures recurred.

At the age of 25, she participated in a trial of vigabatrin, which caused a nearly identical psychotic episode in the form of alternative psychosis. This time, frank psychotic manifestation disappeared swiftly in replacement of seizures without beginning dopamine-blockers. However, different from the first episode, a hypochondriac idea that she might be infected with HIV could not be erased easily, in spite of repeated negative blood test results.

Surface EEG revealed predominant LaT sharp spikes. MRI findings suggested left hippocampal sclerosis. Repeated ictal EEG recordings indicated focal maximum discharges at the left sphenoidal electrode. There were no neurological findings. IQ was normal.

While surgical intervention was considered, we repeatedly warned the patient beforehand that there could be a recurrence of psychotic episodes after surgery and that she should be prepared if she was determined to undergo surgery. One month after the operation, she began to be afflicted with an obsessive idea with initially psychotic tint, ‘Kill yourself immediately now’. This idea came to her again and again in the midst of daily activities, such as face washing and cooking. Then, 12 months after surgery, she noticed that the haunting idea was her own and it was she herself who wanted to die as soon as possible. She avoided her grandparents who were living together in the same house, as she was afraid that the spirits of longevity residing in the old people contaminated her, resulting in her own longevity.

After a 5-year follow-up with intensive psychotherapy as well as antipsychotic medications, the patient began to work as a clerk and enjoyed playing the violin as a member of an amateur orchestra. Except for several sleep grand mal seizures within 1 year after surgery, no other seizures have recurred for more than 10 years after surgery. Interestingly, no remarkable amelioration of mental symptoms was found after these grand mal seizures.

In summary, Case 2 can be characterized as follows: a moderate interval between epilepsy and psychosis onset (10 years), no family history of psychotic episode, and the presence of mesial temporal sclerosis. In this patient, long-standing epileptic seizures seemed to contribute to development of psychosis to a greater extent than in Case 1 (Figure 6).

Figure 6.

Case 2: IIP.

IIP, interictal psychosis; P, psychosis; PIP, postictal psychosis; TLE, temporal lobe epilepsy.

Case 3: Postictal psychosis

Case 3 was a 28-year-old housewife at the first contact. At the age of 3 years, a prolonged febrile convulsion occurred along with subsequent transient paralysis of the left upper extremities. At age 12, the first episode of complex partial seizures occurred, in which she unknowingly handed over an examination paper to a classmate who happened to be sitting next to her. After that, each time during the seizure, she would unconsciously reiterate the same phrase; ‘He’s coming to collect my examination paper. What should I do?’. At age 15, paroxysmal feelings of a peculiar familiarity began to precede the complex partial seizures, during which she felt that the atmosphere of her environment suddenly changed and it seemed as if she had dissolved into the immediate surroundings. Her seizures remained uncontrollable despite intensive medication.

The first manifest PIP occurred after bouts of complex partial seizures at 29 years old. After an intervening 36-hour lucid interval, the patient would rapidly become more and more elevated in mood, with loud rapid speech that was difficult to interrupt. She would change subjects kaleidoscopically from one to another. She screamed to her husband repeatedly, ‘I love you, darling!’, and hugged and kissed him in public in a sensual manner. Three days after the cluster of seizures, the euphoric state culminated in agitated exaltation. She said, ‘I am directly feeling all that is happening in every corner of the world through the palpitating movements of my teeth. The circular movements of my teeth are synchronized with the circular movements of the world. Through nerves in my teeth, I can sense the future two thousand years from now’. Owing to extreme psychomotor agitation, a short stay in the psychiatric ward as well as potent sedatives were required. This state disappeared completely within 10 days. Throughout the episode, her orientation and memory remained intact.

As the patient’s age advanced, she joined a local religious sect as a devoted member and became increasingly eccentric. After several episodes of PIP states that necessitated admissions to the psychiatric ward, she agreed to surgery.

MRI findings revealed marked asymmetry of the hippocampi (left side smaller than the right) along with a lower signal intensity from the left hippocampus on reversed T2 images. Although she was right-handed, her dominant language side proved to be right in an Amytal test. Ictal EEG recordings unanimously suggested that the left hippocampus was the origin of both simple and complex partial seizures. During the course of intensive seizure monitoring, PIP recurred once after a cluster of complex partial seizures.

Surgical intervention was considered. According to results of a standard presurgical evaluation, a left inferior lobectomy with a hippocampoamygdalotomy was performed. One year after the operation, the patient began to work as manager of a Japanese restaurant and completely lost interest in the religious activity as if exorcised. She has been completely seizure free for 15 years postoperatively, received no related medication for 8 years, and neither psychosis nor seizures have recurred.

There was a long interval between epilepsy and psychosis onset (17 years). No family history of psychotic episode was confirmed. There was evidence of mesial temporal sclerosis. In this patient, a causal relationship between seizures and psychosis can be more directly presumed than in Cases 1 and 2 (Figure 7).

Figure 7.

Case 3: PIP.

IIP, interictal psychosis; P, psychosis; PIP, postictal psychosis; TLE, temporal lobe epilepsy.

Case 4: Ictal psychosis

In Case 4 [Kanemoto, 1997], recurrent secondarily generalized seizures began in a girl at 2 years old, which were replaced by complex partial seizures at the age of 4. Thereafter, her first episodes of ictal fear occurred at 20 years of age. Seized by this particular fear, she suddenly felt a conviction that someone was behind her, although she knew that it was not true. She had several complex partial seizures each week, even though nearly all major antiepileptic drugs available in Japan at that time were given up to the maximum dose.

At the age of 22, bouts of complex partial seizures as well as ictal fear led to episodes of brief psychosis, in which the patient felt that close friends or relatives had become totally different persons overnight, replaced by some unknown impostor. During this state, she tried hard to seclude herself and avoid any personal contact. The duration of this state ranged from several hours to a few weeks. These episodes of psychotic states occurred every 2–3 months, which prevented her from seeking stable employment.

The patient had never experienced a psychotic state independent of a cluster of seizures nor had she suffered from other psychiatric symptoms, including dissociative disorder. A strong feeling of depersonalization to her environment accompanied this periictal Capgras syndrome.

Neurological and routine laboratory examinations revealed no particular findings. Both ictal and interictal recordings of scalp EEG including sphenoidal electrodes were inconclusive concerning the origin of the seizures. Even in the midst of ictal fear, no epileptiform discharge could be picked up by the scalp EEG. However, MRI findings revealed marked asymmetry of the hippocampi (volume of the left side only 41% of that of the right). An 80-mg injection of Amytal into the left (but not right) internal carotid artery caused the patient to be totally aphasic. Since scalp EEG was limited for detecting and localizing epileptiform activity in this patient, subdural and depth electrodes were indicated.

An episode of Capgras syndrome-like symptoms began 9 days after implantation of depth-electrodes, though did not occur until the intervals between consecutive ictal fears became less than 1 hour. It is noteworthy that the psychotic episodes appeared regularly when an interval between any two consecutive ictal discharges in the left amygdalohippocampal region was recorded within 1 hour. During this state of clustered ictal fear, she insisted that even her best friends, who were visiting her incidentally during this abnormal mental state, assumed a totally different personality. She suspected even my identity as well, asking me repeatedly, ‘Are you really Dr. Kanemoto? You smile in an awkward manner that I’ve never seen before’. Throughout this psychotic episode, there was an intense feeling of derealization to the surroundings, whereas there was no trace of clouding of consciousness and no disorientation of time, place, and person. We showed her seven pictures during this episode and asked her to remember them, and she recalled these pictures perfectly, even 2 hours later. All 17 seizures including isolated ictal fears recorded during monitoring showed clear-cut epileptiform discharges in the left amygdalohippocampal region.

A left inferior temporal lobectomy with a hippocampoamygdalotomy was performed, and a histological examination revealing mesial temporal sclerosis. Since the lobectomy, she has been free from seizures for 15 years, with no further episodes of either ictal fear or Capgras syndrome. Epileptic activity was apparently a direct and main cause of the psychotic state in this patient (Figure 8).

Figure 8.

Case 4: Ictal P.

IIP, interictal psychosis; P, psychosis; PIP, postictal psychosis; TLE, temporal lobe epilepsy.

Concluding remarks

In general, psychotic illness in PWEs has been subcategorized into IIP and PIP. However, while PIP constitutes a rather homogenous clinical entity, IIP is apparently quite heterogeneous. Most often, IIP is subdivided into chronic and acute interictal types, with the latter comprising also patients with alternative psychosis [Tadokoro and Kanemoto, 2012; Trimble, 1991].

In a review of epileptic psychosis, Sachdev provided a new perspective of postictal and alternative psychoses as a large unified group [Sachdev, 1998]. According to that proposal, these groups of transient epileptic psychoses can be regarded as true epileptic psychosis, because both are closely associated with epileptic activity, even if the link is inversely related in cases of alternative psychosis. Along this line, chronic psychosis in PWEs is consequently left as a remnant, which may be heavily dependent on genetic predisposition.

In Table 9, data derived from our series of 200 PWEs and psychosis is rearranged as a function of the closeness of the association with TLE. Interestingly, just as PIP is quite closely associated with TLE, most PWEs whose initial psychotic episodes began with repetitive episodic IIP and who ultimately became continuously psychotic showed extremely high affinity with TLE. To be noted, this group of patients showed a long interval between psychosis and epilepsy onset, just as those with PIP.

Table 9.

Subcategories of epileptic psychosis with special attention to the associations with TLE. (Reproduced with permission from Kanemoto et al. [2008].)

	TLE (%)
AIP evolving into CP (n = 23)	19 (82.6%)*,**
PIP (n = 53)	41 (77.4%)*,**
AIP with complete remission ( n = 81)	40 ( 49.4%)*,**
CP without preceding AIP episodes (n = 36)	15 (41.7%)**

^*p = 0.002; **p = 0.005

AIP, acute interictal psychosis; PIP, postictal psychosis; CP, chronic psychosis; TLE, temporal lobe epilepsy.

In Table 10, we propose a tentative classification of psychotic illness in PWEs, with special attention to those who have undergone epilepsy surgery. Type A psychosis consists of ictal psychosis and PIP, and is characterized by a long epilepsy–psychosis interval and high affinity with TLE. Patients with Type A psychosis are well known to be excellent candidates for surgical intervention to control intractable seizures, as in the present Cases 3 and 4. In Table 10, psychotic episodes which began initially with repetitive episodic IIP and which ultimately became continuous are classified into Type B psychosis. As we mentioned, patients with Type B psychosis share common clinical properties with Type A. Notably, as in Case 2, patients with Type B psychosis, although chronic in nature, can also be considered surgical candidates with inexorable preparedness on the part of the patient and extreme caution on the part of medical staff, because up to 30% of these patients experience postoperative exacerbation of psychosis, although long-term amelioration of the mental state is expected. As we mentioned earlier in the preceding section, Type A psychosis can progress to Type B in some cases [Tarulli et al. 2001; Kanemoto, 2002, 2011; Adachi et al. 2002]. This also supports the notion of kinship between Type A and Type B psychoses. In contrast, Type C psychosis seems to be strongly influenced by genetic predisposition. Therefore, in contrast to Type A and Type B psychoses, elimination of epileptic activity may not be expected to improve psychosis in this group, as in the present Case 1. With the present state of knowledge, surgical intervention is not recommended for this group.

Table 10.

A tentative trial of classification of psychotic illness in PWEs.

	Evolution of psychosis	Acute or chronic	Long or short interval*	Affinity with TLE	Main determinant	Case
A	PIP, ictal psychosis (or aura continua)	Acute	Long	Very high	Epileptic activity	3,4
B	Initial repetitive transient psychotic episodes ultimately resulting in CP	Chronic	Long	Very high	Epileptic activity	2
C	Alternative psychosis CP without preceding transient psycotic episodes	Acute / Chronic	Short	Relatively low	Genetic predisposition	1

^*Interval between psychosis and epilepsy onset.

CP, chronic psychosis; PIP, postictal epilepsy; PWE, patient with epilepsy; TLE, temporal lobe epilepsy.

Last, but not least, it should be stressed that psychotic disorders in PWEs are often overlooked, mistreated, and consequently lingering on needlessly [Cornaggia et al. 2002]. These failures stem mainly from two sources. The first is based on misdiagnosis. Mild, beginning psychoses are often mistaken for simple depression and given antidepressants, which are not very effective against psychosis. This delay of administration of dopamine-blockers may well complicate the situation seriously. The other problem is more controversial. Some antiepileptic drugs, such as vigabatrin [Weber et al. 2012], phenytoin, zonisamide [Noguchi et al. 2012], and topiramate [Khan et al. 1999], have been reported to show adverse psychotropic effects. If these suggestions are reliable enough, replacement of these antiepileptics in the case of the occurrence of psychosis seems to be mandatory. However, robust arguments against this noxious drug theory have been constantly provided [Mula and Trimble, 2003]. According to the investigators in support of this line of arguments, it is freedom from seizures in patients with particular clinical backgrounds that matters, not adverse effect profiles of individual antiepileptic drugs. To elucidate the riddle of alternative psychosis, we desperately require further reliable data.