Basic treatment principles for psychotic disorders in patients with epilepsy


Address correspondence to Naoto Adachi, Adachi Mental Clinic, Kitano 7-5-12, Kiyota, Sapporo 004-0867, Japan. E-mail:


In patients with epilepsy, coexisting psychoses, either interictal (IIP) or postictal (PIP), are associated with serious disturbance in psychosocial function and well-being, and often require the care of a specialist. Unfortunately, evidence-based treatment systems for psychosis in patients with epilepsy have not yet been established. This article aims to propose concise and practical treatment procedures for IIP and PIP based on currently available data and international consensus statements, and primarily targeting nonpsychiatrist epileptologists who are often the first to be involved in the management of these complex patients. Accurate and early diagnosis of IIP and PIP and their staging in terms of acuity and severity form the essential first step in management. It is important to suspect the presence of psychosis whenever patients manifest unusual behavior. Knowledge of psychopathology and both individual and epilepsy-related vulnerabilities relevant to IIP and PIP facilitate early diagnosis. Treatment for IIP involves (1) obtaining consent to psychiatric treatment from the patient, whenever possible, (2) optimization of antiepileptic drugs, and (3) initiation of antipsychotic pharmacotherapy in line with symptom severity and severity of behavioral and functional disturbance. Basic psychosocial interventions will help reinforce adherence to treatment and should be made available. Due consideration must be given to patients’ ability to provide informed consent to treatment in the short term, with the issue being revisited regularly over time. Given the often prolonged and recurrent nature of IIP, treatment frequently needs to be long-term. Treatment of PIP consists of two aspects, that is, acute protective measures and preventive procedures in repetitive episodes. Protective measures prioritize the management of risk in the early stages, and may involve sedation with or without the use of antipsychotic drugs, and the judicious application of local mental health legislation if appropriate. As for preventative procedures, optimizing seizure control by adjusting antiepileptic drugs or by surgical treatment is necessary.

Although psychotic illnesses in patients with epilepsy have been long studied, the literature thus far has leaned toward descriptive psychopathology and establishment of etiologic and neuropathologic links. In contrast, there are only a few systematic studies that have evaluated treatment strategies (Adachi, 2005; Sachdev, 2007). The responsibility for initial management of these patients can fall on a range of professionals in addition to epileptologists, and experience in the management of psychotic disorders can vary, depending for instance on whether the treatment setting is in a tertiary referral center with access to neuropsychiatric expertise. We aim to propose treatment procedures, based on available data, for two main types of epilepsy-related psychoses (i.e., interictal psychosis and postictal psychosis), that are practical and clinically meaningful.

Interictal Psychosis

People with epilepsy have a twofold to threefold increased risk of developing psychosis (Qin et al., 2005), which may indicate that epilepsy per se plays some role, although establishing a causal association remains controversial (Adachi, 2006). With regard to their temporal association with seizures, the psychosis may be regarded as interictal psychosis (IIP) or postictal psychosis (PIP; Sachdev, 1998). In contrast to recent advances in clinical research in PIP, studies of IIP remain limited. In particular, there are few systematic studies of effective treatment strategies for IIP. Existing recommendations (Kerr et al., 2011) are extrapolated from those established for the treatment of schizophrenia and related psychotic illnesses with some additional guidance from anecdotal evidence or expert opinions.

A range of factors influence the choice of psychiatric treatments for IIP patients. In this article, we summarize treatment procedures in three stages depending on the severity of psychotic symptoms and the level of psychosocial disturbance.

  1. Mild: Although patients clearly show some psychotic symptoms, they are not distressed by the symptoms, and their overall psychosocial function is maintained.
  2. Moderate: Patients have distinct psychotic symptoms; day to day functioning is disrupted, and overall psychosocial functioning is affected to some degree.
  3. Severe: Patients have overwhelming psychotic symptoms, behavioral consequences of these symptoms put them or other people at risk, and their psychosocial function is significantly impaired.

Definition of IIP

The term psychosis encompasses a broad spectrum of mental disturbance. Clinically, in accordance with International Classification of Diseases, Tenth Revision (ICD 10) (World Health Organization, 1992), the syndrome of psychosis is defined by the presence of hallucinations, delusions, and/or a limited number of severe behavioral abnormalities, such as gross excitement and overactivity, marked psychomotor retardation, and catatonic behavior. Thought disorder is a salient feature of psychosis, which may be characterized by disordered content, as represented by delusions, or disordered thought form. Disordered thought form broadly comprises disorders of the mechanisms of thinking, that is, concrete thinking, loosening of associations, incoherence, overinclusion, and illogicality. There may be associated disorder of language and speech, that is, derailment, neologisms, retarded/poverty of speech, pressured speech, and flight of ideas (Cutting & Murphy, 1988). Despite reports of frequent occurrence in patients with epilepsy, in particular those with psychiatric or neurobehavioral disorders (Slater et al., 1963; Caplan et al., 1997), formal thought disorder is not commonly considered in diagnostic workup for IIP (Adachi, 2006).

Contributions from British neuropsychiatrists (Hill, 1953; Pond, 1957; Slater et al., 1963; Slater & Roth, 1969) defined IIP as any psychosis in clear consciousness that occurs in someone who has previously been diagnosed with epilepsy, and the psychosis is not exclusively during or immediately following a seizure. This criterion is concise and practical, and has often been used for approximately three decades. There is an implication that the psychosis is, in some manner, “secondary” to the epilepsy (Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition; DSM-IV, American Psychiatric Association, 1994) or is an “organic” mental disorder (ICD-10) in line with a conceptually orthodox dichotomy between organic and functional psychoses (Adachi et al., 2010b). Such orthodoxy is often not possible to sustain. Several recent studies have demonstrated some vulnerability factors, such as the presence of family history of psychosis (Adachi et al., 2000a; Qin et al., 2005) and impaired intellectual functioning (Adachi et al., 2000a; Mellers et al., 2000) that are common to both “non-organic” psychosis and psychosis in epilepsy (IIP). These findings have shed light on pathogenesis of psychosis in patients with epilepsy in a wider context. Although evidence is still limited, recent findings suggest common vulnerabilities among learning disability, developmental disorders, and psychosis (Craddock & Owen, 2010).

In this article, we have adopted Slater's definition for practical use. Previous authors have employed various forms of subclassifications in IIP according to its duration (e.g., acute/brief/transient and chronic) or presumptive etiologies (e.g., drug-induced and forced normalization; Sachdev, 1998, 2007; Fig. 1). These subclassifications have not been validated. It needs to be kept in mind that the same patient can exhibit different forms of IIP at different times, for example, drug-induced and non–drug-induced, or acute and chronic.

Figure 1.

Possible pathophysiologic mechanisms for the association between epilepsy and schizophrenia-like psychosis (with permission from Sachdev, 2007).

Diagnosing IIP

People with epilepsy are at an increased risk of psychosis. Every year, 3–4 of 1,000 adult patients with epilepsy with no past history of psychosis exhibit some psychotic symptoms for the first time (Onuma et al., 1995; Tadokoro et al., 2007). The risk is approximately 2–3 times higher than that of first-episode schizophrenia in the general population aged 15–65 years. The first step for making a diagnosis is to be aware that any patient with epilepsy can develop IIP, and to suspect IIP when they show unusual behavior.

Identifying IIP may be straightforward when patients exhibit frank psychotic symptoms. It is, however, not uncommon that patients conceal their psychotic experiences (Koutroumanidis et al., 2010). Viron et al. (2012) proposed a set of screening questions for psychosis (Table 1). Nonverbal cues (e.g., sullen, preoccupied, monotonous, or unnatural behaviors) are vital in understanding abnormal inner experiences, as verbally provided information cannot be relied upon exclusively. Obtaining collateral history from their family, friends, and carers is also important (Koutroumanidis et al., 2010).

Table 1. Screening questions for psychoses (Viron et al., 2012)

A lead-in statement helps to normalize the experiences for the patient and reduces the potential shame and embarrassment of sensitive topic, for example:

“Sometimes when people are [under stress/feeling anxious/feeling depressed], they can have strange experiences such as trouble with their

thinking or seeing or hearing things that others don't.” Affirmative responses to the question below should be followed by: “Tell me about that.”

Questions to elicit delusional thinking
Have you had any strange or odd experiences lately that are difficult to explain or that others would find hard to believe?
Have you felt like people are watching or following you or they want to harass or hurt you?
Have you felt like others can hear your thoughts or that you can hear another person's thoughts?
Questions to elicit hallucinations
Do your eyes or ears ever play tricks on you?
Have there been times when you heard or saw things that other people could not?

To confirm the diagnosis of IIP, it is necessary to exclude ictal, preictal, and postictal psychotic phenomena. Criteria for these phenomena are detailed in the PIP section of this article. Because patients may not clearly remember precedent seizures, clinical clues indicating the presence of seizure activity around the time of onset of the psychotic episode should be sought meticulously.

Associated characteristics

Constant characteristics

Characteristics that remain constant during the course of illness in each patient appear to be associated with vulnerabilities relevant to the development of IIP (Fig. 1). A number of recent studies have shown common vulnerabilities between epilepsy and psychosis (Qin et al., 2005; Adachi et al., 2011). Epilepsy patients with a family history of psychosis are predisposed to developing IIP, even in the absence of overt cerebral insult related to epilepsy (Adachi et al., 2010a). The rate of family history of psychosis is estimated at 3–10% in IIP (Adachi et al., 2000a, 2002a; Qin et al., 2005). Because the actual risk of IIP based on genetic vulnerabilities is difficult to calculate at present, the presence of family history of psychosis could be explored not only within the first-degree relatives as in most genetic studies with strict criteria but also in the more extended family. Epilepsy patients with a learning disability are more vulnerable to IIP than those without (Adachi et al., 2000a; Mellers et al., 2000). In particular, those with borderline intellectual functioning have 2–5 times higher risk for various psychotic states (Adachi et al., 2002a).

Several epilepsy-related variables have been considered in the pathogenesis of IIP (Trimble, 1991; Sachdev, 1998). They include epilepsy types, seizure variables, and the lateralization of epileptogenesis. Indeed, patients with focal epilepsy have a higher risk of IIP than those with generalized epilepsy (Adachi et al., 2000a,b). It is noteworthy that 15–20% of IIP patients have idiopathic generalized epilepsy (Adachi et al., 2000a, 2010a). As for seizure variables, most reports have pointed to the occurrence of complex partial seizures and in particular those occurring in temporal lobe epilepsy as suggested by Gibbs's (1951) report. Patients with other forms of focal epilepsy can also develop IIP (Adachi et al., 2000b, 2002b). Because most of the recent large studies failed to find associations between left-sided epileptogenicity and IIP (Mendez et al., 1993; Kanemoto et al., 2001ab; Adachi et al., 2002a), the importance of lateralization of epileptogenesis may have been overestimated.

Inconsistent characteristics

Features that are subject to change, such as regimens and doses of antiepileptic drugs (AEDs) and seizure frequency, are also likely to affect IIP. Modifying such conditions by appropriate medical approaches can consequently prevent or improve IIP.

AEDs may sometimes induce psychotic symptoms. A high serum level of AED, particularly in the context of polypharmacy, can often cloud consciousness, precipitating development of psychotic symptoms (Onuma, 1987). Whereas any of the AEDs has the potential to cause psychotic symptoms (Schmitz, 1999), some AEDs, for example, ethosuximide (Fischer & Pedersen, 1965), phenytoin (Noguchi et al., 2012), zonisamide (Matsuura & Trimble, 1997; Noguchi et al., 2012), topiramate (Mula & Monaco, 2009), and vigabatrin (Mula & Monaco, 2009), seem to be more potent in this regard than others. Although most findings on AED-related psychosis have been based on single case reports or small series of cases without systematic examinations, it is arguably safer to avoid these agents in patients at high risk. Approximately 30–40% of patients who had AED-induced psychosis also developed chronic IIP or AED-unrelated IIP at different times (Matsuura, 1999; Kanemoto et al., 2001a,b). These results may suggest that patients with individual vulnerabilities to psychosis tend to exhibit IIP irrespective of types of AEDs.

The association of seizure frequency with occurrence of IIP has been a matter of debate. Several early studies (Slater et al., 1963; Flor-Henry, 1969; Kristensen & Sindrup, 1978) reported a lower frequency of seizures in patients with IIP. This is in line with the concept of antagonism between epilepsy (seizures) and psychosis, which may be represented by descriptions of forced normalization (Landolt, 1958) and alternative psychosis (Tellenbach, 1965). Applying these theories to clinical practice needs careful evaluation for each IIP episode. The same individual may develop an episode of IIP consistent with the phenomenon of forced normalization on one occasion while having an episode of IIP without any changes in electroencephalography (EEG) and/or seizure frequency at a different time. Furthermore, Mendez et al. (1993) reported a higher frequency of complex partial seizures (CPS) in IIP. Systematic study has not as yet been conducted to verify these hypothetical theories.

Treatment principles for IIP

In the absence of specific treatment guidelines, psychotic symptoms in IIP have hitherto been treated in line with well-established treatment protocols for primary schizophrenia and related psychotic illnesses (Kerr et al., 2011). Few data are available on the treatment of IIP specifically.

It is not uncommon in day to day clinical practice for nonpsychiatrist epileptologists to be the first clinicians to face the management of a patient developing IIP. Awareness of the potential for the syndrome to develop in any epilepsy patient and knowledge of common psychopathologic phenomena as described above is important.

The next step is to set a treatment target. IIP symptoms can be difficult to manage for a variety of reasons that commonly include intractability of psychotic symptoms, the individual variation in tolerability of antipsychotic treatment, the treatment's potential for adverse consequences on seizure control, and the individual's adherence to psychiatric treatment (Onuma, 1987; Adachi, 2005). In cases of primary psychotic episodes, at least one half of patients are partially or completely nonadherent (Lacro et al., 2002; Abdel-Baki et al., 2012). If this is the case, treatment modalities should be offered aiming at overall functional recovery and improved quality of life, rather than aiming at complete symptomatic remission.

There are some advantages for patients with IIP to remain in the care of their treating epileptologists at the time of initial psychiatric treatment. The epileptologists are positioned to observe and identify early signs of IIP even before the individual reports it. They are also best suited to simultaneously manage factors common in epilepsy and IIP (Onuma, 1987). The existing therapeutic relationship helps ensure that epilepsy patients with psychosis are more likely to be cooperative with their treating doctors than are patients with first-episode psychosis who are treatment-naive (Adachi, 2005). The good doctor–patient relationship enhances therapeutic outcome (Day et al., 2005).

A significant proportion of patients with psychosis have limited insight into their psychotic symptoms and related behavior, and the need for treatment (McGorry & McConville, 1999; Mintz et al., 2004). It is not uncommon for patients to consent to epilepsy treatment but refuse psychiatric treatment. Assessing their capacity to provide informed consent to treatment is important (Brabbins et al., 1996; Pollack & Billick, 1999; Capdevielle et al., 2009). The management of epilepsy should continue alongside regardless. At the very least, this will allow for monitoring of the evolving psychiatric presentation and arranging appropriate intervention when needed.

How to use medical treatment

Timing of initiating antipsychotic drug (APD)

In general, psychosis is better managed earlier rather than at a later, more advanced stage. Early initiation of APD therapy relative to time of onset of IIP shortens the duration of the IIP episode (Adachi et al., 2012). This finding is in line with studies on first-episode psychosis and schizophrenia (White et al., 2009). Therefore, an early introduction of APD is recommended where clinically indicated, in particular if the patients present with their first episode. On the other hand, approximately 15% of IIP episodes remitted without any APD treatment (Adachi et al., 2012), indicating that a short delay in commencement of APDs may not be harmful for first-episode patients if they are well supported with psychosocial interventions (Francey et al., 2010). For patients with chronic IIP, it is prudent to seek patient consent to continued psychiatric treatment.

Choice of APD

The recent consensus report of the International League Against Epilepsy (ILAE) commission (Kerr et al., 2011) states that IIP can be treated with APDs similarly to functional psychoses, for example, schizophrenia, brief/transient psychosis, and delusional disorders. APDs are often divided into two categories: first-generation (mostly typical/conventional) APDs (FAPDs); that is, butyrophenones, phenothiazines, benzamides, and thiepins), and second-generation (atypical) APDs (SAPDs); that is, serotonin-dopamine antagonists, dibenzothiazepines, and multiacting receptor-targeted antipsychotics). In principle, choosing an APD depends on its efficacy and tolerability. Although some differences in efficacy have been reported, there is no conclusive evidence that SAPDs are more effective than FAPDs (Crossley et al., 2010; Girgis et al., 2011; Hara et al., 2012). The dose of APDs is often converted into chlorpromazine equivalent (CE mg/day) to help compare different APDs (American Psychiatric Association, 1998; Toru, 1998; Table 2). This is not always possible or reliable, especially for some SAPDs.

Table 2. Types of antipsychotic drugs and their chlorpromazine equivalent dosages
  1. Modified from American Psychiatric Association, 1998; Toru, 1998. Chlorpromazine equivalent dose (CE): 1 mg of each medicine can be converted into certain dose (mg) of CP, for example, 1 mg haloperidol is 50 mg CP.

First-generation APDPhenothiazinesChlorpromazine1
Second-generation APDSerotonin-dopamine antagonists (SDA)Risperidone67 (50–100)
DibenzodiazepinesQuetiapine1.4 (1–2)
Multiacting receptor-targeting antipsychotics (MALTA)Olanzapine40 (30–50)
Dopamine system stabilizerAripiprazole40 (30–50)

Given the paucity of evidence for the relative efficacy of one APD over another in treating IIP, the choice is largely based on adverse effects. The main adverse effects of APDs are as follows:

  1. Sedation, such as somnolence and hypersomnia, is the most common side effect of APDs, either FAPDs or SAPDs. Low potency APDs except for sulpiride, for example, chlorpromazine, are more likely to cause these effects.
  2. Weight gain and metabolic syndrome occur with any of the APD treatments, although SAPDs are generally less favorable in this regard. In particular, most SAPDs (e.g., olanzapine, quetiapine, risperidone; Correll et al., 2009), with the possible exception of aripiprazole, and some FAPDs (e.g., sulpiride) are likely to be associated with these problems.
  3. Cardiovascular effects can occur with any APD treatment (Mackin, 2008). Orthostatic hypotension is the most common adverse effect of APDs, in particular with low potency FAPDs. QTc prolongation and subsequent arrhythmia can be caused with any APDs; however, this risk is increased with pimozide, thioridazine, sertindole, and zotepine.
  4. Hyperprolactinemia and sexual dysfunction are sometimes in the context of treatment with most FAPDs (e.g., haloperidol and sulpiride) and some SAPDs (e.g., risperidone and amisulpride; Rosenbloom, 2010; Holt & Peveler, 2011).
  5. Extrapyramidal symptoms (EPS), such as akathisia and parkinsonism, occur mostly with FAPDs (in particular high potency agents, that is, haloperidol), although some degree of EPS can be seen with SAPDs. EPS can be a primary cause of noncompliance with medication. Low doses of anti-parkinsonism/antispasmodic drugs (e.g., trihexyphenidyl 4–10 mg/day or biperiden 2–6 mg/day) can be prescribed together with FAPDs to ameliorate EPS (Toru, 1998).

Cost–benefit analyses are also important to take into account, since treatment for IIP is often required over long periods. In addition to already long-term financial commitment to epilepsy treatment, treatment cost for IIP could be a burden, becoming a barrier to adherence to treatment (Perkins, 2002; Abdel-Baki et al., 2012). In the WHO multinational cost-effectiveness study (Chisholm et al., 2008), the most cost-effective interventions in the developing world were those using typical APDs combined with psychosocial treatment, delivered via a community-based service.

Some AEDs, such as carbamazepine, valproate, and lamotrigine, are known to have certain psychotropic effects; however, these are mainly for affective and not for psychotic symptoms, and these drugs are not useful for treating IIP per se. Minor tranquilizers (particularly benzodiazepines) have some antiepileptic and psychotropic effects. Although minor tranquilizers normally suppress anxiety, irritability, and psychomotor-excitements, they often induce somnolence, loss of concentration, and disinhibition and have dependence potential (Toru, 1998). Judicious use in the short term when required is necessary to avoid such undesirable effects. Antidepressants may have some beneficial effects for IIP with depressive/affective symptoms (Blumer et al., 2000), and judicious use of these drugs is recommended.

Special considerations

Concerns over the propensity of APDs to lower seizure threshold can lead to the undertreatment of IIP. Some APDs are recognized to increase seizure risk more than others, including chlorpromazine and zotepine. Among SAPDs, clozapine, olanzapine, and quetiapine appear to have a higher seizure risk. It is important to bear in mind that in addition to drug-related factors, predisposing factors in the individuals contribute to drug-induced aggravation of seizures. Most reports are based on animal studies, APD overdose cases, or patients with mental disorders who are not treated with AEDs. In epilepsy patients treated with AEDs, most APDs, either FAPD or SAPD, can be safely administered in patients with epilepsy (Okazaki et al., 2012).

Pharmacokinetic interactions between APDs and AEDs need to be considered (Trimble & Mula, 2008). Via cytochrome P450 (CYP) systems, serum concentrations will be altered depending on the combination of the drugs of interest, potentially leading to adverse events or reduced efficacy. Enzyme inducers, such as carbamazepine, may decrease the serum level of some APDs. Patients with multiple disorders often take a combination of drugs, either enzyme inducers, enzyme inhibitors, or both. The drug interactions are often difficult to calculate accurately when the patients are on polypharmacy, since interactions between two drugs (e.g., one AED and one APD) observed experimentally or clinically can often be altered in unexpected directions when further drugs are added. Pragmatic approaches are (1) to simplify medication regimens as far as possible, (2) to monitor efficacy and adverse effects clinically, and (3) to check the serum level if appropriate and adjust the dose of the medication accordingly (Onuma, 1987; Adachi, 2005).

APD dose titration

General rules of APDs titration are to commence at low doses, titrate slowly to a minimum therapeutic dose, and continue at a steady dose for a sufficient period of time. Of 320 IIP episodes in Japanese neuropsychiatric institutions, the maximum dose of APDs required ranged from 12.5 to 3,425 mg/day CE (mean 454.7 mg/day CE; Hara et al., 2012). Empirically, for patients with mild IIP episodes, the target dose of APD ranges from 50 to 300 mg/day CE. Some mild psychotic episodes disappear without any APD treatment; therefore, IIPs can be observed without medication under some conditions. Examples include situations in which (1) the patients and people around them are not distressed or disturbed by the symptoms, (2) the patients do not give a consent to take APDs and the risk profile is favorable, and (3) risks of administering APDs is greater than their benefits. Slow and careful titration from low dose (around 50–100 mg/day CE) can significantly reduce the occurrence of adverse effects (Mula & Monaco, 2009). When IIP patients show moderate episodes, which are determined not only by symptom severity but also by functional disturbance, the patients with moderate episodes should take 300–1,000 mg/day CE of APDs. For serious episodes, in particular those accompanying characteristics such as excessive psychomotor excitement or attempted suicide, they may need to take sufficient dosages (1,000–3,000 mg/day CE) of APDs. Because dose required does not necessarily correlate with present severity, other conditions, such as individual responsiveness, course of illness, and history of violent behaviors, should also be considered. Depending on their risks to self and/or others, a transfer of care to psychiatric services, including psychiatric emergency, may be required. Keeping up to date with mental health legislation and local procedures would minimize delay for involvement of specialist services.

Careful titration is recommended for patients with particular conditions by an international consensus study (Gardner et al., 2010). They include patients with distinct brain damage, that is, postencephalitis and cerebrovascular disease, with physical conditions, in the younger or older age groups, or in certain ethnic groups. The starting dose should be smaller, and titration regimens should be slower than those for patients without such conditions. For patients with brain damage, the target dose should be set at 30–50% lower. Patients with physical diseases, such as impaired hepatic function (−45%) and impaired renal function (−30%) should also be prescribed a lower dosage. Age-related dosing changes of APD treatment should be as follows; that is, children (age 6 to puberty; −60%), adolescence (−30%), and elderly (>65 years; −50%). Likewise dosages of APDs reflect ethnic metabolic characteristics; East Asian patients may take a 20–40% lower dose than Western patients (Lawson, 1986; Matsunaga et al., 2009).

Duration of APD treatment

There is insufficient evidence on the duration of treatment with APDs for IIP episodes. Effectiveness of APD use for maintenance and prevention of IIP is also little known. In pharmacologic management in functional psychoses, it is advised that the patients remain on the optimized dose of APD for at least 6 months after reaching remission for the first couple of episodes, and then come off medication gradually. In patients who already have had multiple episodes are advised to stay on medication long term. Due to its chronic nature and relatively high rate of exacerbation, the international commission recommended a long-term administration of APDs after a remission of IIP (Kerr 2011). This is in line with findings in first-episode psychosis that the maintenance of treatment is more effective for preventing relapse and deterioration than targeted intermittent treatment (Gaebel et al., 2011). Sudden discontinuation of APDs may disrupt serum levels of AEDs; thus, seizure control and adverse effects of AEDs should be carefully monitored. It is possible that IIP episodes recur after discontinuing APDs (Onuma et al., 1991; Adachi, 2005); careful follow-up is needed.

Psychotherapeutic approaches

Psychosocial functions are adversely affected by psychotic symptoms as well as nonpsychotic symptoms, that is, apathy and neurocognitive dysfunction, which are sometimes difficult to treat with medication. Although there is no systematic study on effectiveness of psychotherapeutic approaches in epilepsy patients with psychoses, ample evidence in functional psychoses (Francey et al., 2010) gives us insight into the value of such interventions in the management of these patients, leading to better adherence to treatment and improved quality of life. Limited resources and experience in this area may be a deterrent, although basic interventions listed below could be provided by nonspecialists.

  1. Psychoeducation: It is essential to ensure that patients understand the diagnosis of psychosis and need for treatment over a long period (Perkins, 2002). Improved awareness and understanding of the condition can help them remain motivated in continuing treatment. Tackling stigma surrounding psychiatric conditions and their treatment may be one of the keys for successful treatment (Iyer et al., 2011; McCann et al., 2011).
  2. Self-help and reframing: Stress may play a role in development of IIP. Poor coping skills could have detrimental effects on the patients, leading to reduced quality of life and low self-esteem. Assisting the patients in developing more helpful coping strategies can improve their self-monitoring of the conditions, and long-term management, particularly in patients with mildly impaired to borderline intellectual functioning (Harris, 1998; Adachi, 2005).
  3. Support for family and carers, and networking: Good support by family members, carers, and social networks is a significant determinant for adherence to medication (Rabinovich et al., 2009). In patients with schizophrenia, the relapse rate can be reduced by 20% if relatives are involved in treatment (Addington & Addington, 2008). Providing psychoeducation and practical and emotional support for family and carers often leads to better management of the conditions as a whole (Onuma, 1987; Adachi, 2005).

Prognosis of IIP

Episodes of IIP generally take a longer time than PIP, lasting for months in many cases. In a large cohort study, approximately 75% of all IIP episodes lasted for 1 month or more, regardless of APD treatment; the mean duration of 320 IIP episodes in 155 patients was 82.7 weeks (median 17 weeks; Adachi et al., 2012). This study considered only the periods during which the patients manifested hallucinations and/or delusions. Episodes are likely to be more prolonged if formal thought disorders and/or nonpsychotic symptoms (e.g., emotional withdrawal, depression, anxiety, and irritability) were included.

Episodes often recur in patients with IIP. Onuma et al. (1991) reported that almost two thirds of IIP patients had episodes of relapse after remission or worsening of chronic episodes during a 10-year follow-up period. A longitudinal observation disclosed that approximately 90% of the IIP patients had psychotic episodes that lasted for 1 month or more and 65% had episodes that lasted for 6 months or more (Adachi et al., 2012). In the course of illness, patients sometimes exhibit different types of psychosis, for example, acute-transient and chronic (Adachi et al., 2012), AED-induced and AED-unrelated (Matsuura, 1999; Kanemoto et al., 2001a,b), and even IIP and PIP (Tarulli et al., 2001; Adachi et al., 2003). Some IIPs recur after discontinuation of APD treatment (Onuma et al., 1991; Adachi, 2005).

Patients with mental disorders have an increased risk of premature death, from either natural or unnatural causes (Harris & Barraclough, 1998; Tiihonen et al., 2009). Patients with epilepsy also have particular risks to early death, for example, sudden unexpected death, status epileptics, suicide, and accidents, in addition to common etiologies (Hitiris et al., 2007). It is therefore possible that synergism of the two conditions can lead to an increase in mortality.

Future issues

We have proposed treatment guidelines for IIP based on limited data and the consensus statements currently available (Table 3). Whereas treatment systems for IIP are predominantly “borrowed” from standard treatment systems for functional psychoses, they have been found to be safe and effective. Specific aspects of treatment that would improve outcome in IIP have not been researched. Particularly, few data are available for maintenance and prophylaxis in IIP (Adachi, 2006). Large-scale studies in patients with IIP are required in a prospective manner, which can be challenging due to the relatively low prevalence of IIP. In addition, despite growing interest in quality of life in people with epilepsy, usefulness of psychosocial interventions has been rarely mentioned or investigated in patients with IIP. The field remains open for more research.

Table 3. Treatment principles for IIP
  1. CE, chlorpromazine equivalent.

Establish a diagnosis of IIP episodes by excluding ictal/periictal/postictal psychotic phenomena; evaluate the severity of their symptoms and level of their disturbances
Assess the patient's capacity to give a consent to treatment and/or to participate in the decision-making process; seek views from family members and/or carer when necessary
Make a treatment strategy, for example, psychosocial interventions or watchful wait in outpatient clinics, pharmacologic treatment (with or without other treatment options) as an outpatient or inpatient
Optimize AED regimens where possible by reducing polypharmacy and adjusting the dose to aim for therapeutic serum levels
Administer APDs: following issues should be taken into account
Timing: early intervention is preferable
Choice of APD: any APD with fewer or lesser adverse effects, availability, and affordability
Dose titration: 50–300 mg (CE)/day for mild episodes, 300–1,000 mg (CE)/day for moderate episodes, 1,000–3,000 mg (CE)/day for severe episodes
Duration: long-term treatment (usually months or years) with APD is required
Provide basic psychosocial approaches, for example, psychoeducation, self-help, and reframing, and support for family and carers

Postictal Psychosis

Postictal psychosis (PIP) had not drawn wide attention until the iatrogenic provocation of seizures in seizure monitoring units for preoperative evaluation became prevalent (Savard et al., 1991; Devinsky et al., 1995; Kanner et al., 1996), and neurologists and neurosurgeons were frequently confronted with PIP requiring urgent intervention. Pertinent knowledge of this phenomenon has therefore become indispensable for clinicians (Trimble, 1991; Kanemoto et al., 1996a; Kanemoto, 2011). Although PIP is, in most cases, a self-remitting condition, individuals in the midst of PIP can be astonishingly violent, extremely agitated, or confused, such that immediate protection under strict custody often becomes mandatory (Devinsky, 2008). PIP is a psychiatric emergency that every neurologist or neurosurgeon involved in epilepsy surgery or treatment of long-standing limbic epilepsy could be confronted with.

Definition of PIP

As is the case for other types of epileptic psychoses, PIP is defined by two features: its chronologic relationship to seizures and the nature of the mental state. For the chronologic relationship, most authors have adapted the criterion proposed by Logsdail and Toone (1988), that is, episodes of psychosis that develop within 1 week after a seizure, or usually a cluster of seizures, qualify as PIP. This operational delineation is not as arbitrary as it seems in light of accumulated data that now indicate that PIP generally occurs within 3 days after the last seizure (Logsdail & Toone, 1988; Devinsky et al., 1995; Kanemoto et al., 1996a). Indeed, the presence of a lucid interval between the last seizure and start of changes in mental state guarantees a qualitative difference between PIP and simple postictal confusion, and constitutes an integral part of the clinical picture of “nuclear PIP” (Fig. 2). However, inclusion of a lucid interval in the essential diagnostic criteria would exclude nearly half of the cases from PIP (Levin, 1952; Kanemoto et al., 1996a) and most authors have not automatically excluded cases without a lucid interval from their study subjects. Indeed, the lucid interval could be very short and may therefore go unnoticed.

Figure 2.

Definition of PIP and closely related conditions (with permission from Kanemoto, 2011).

The second part of the definition, the nature of the mental state, relates to the psychotic presentation. Indeed, the original criteria proposed by Logsdail and Toone (1988) included the presence of delusions or hallucinations in clear consciousness. However, because a manic presentation is a frequent precursor of nuclear PIP (Kanemoto et al., 1996a; Kanner et al., 1996), it seems reasonable that postictal manic state should be included in this definition. The fact that appropriate and timely therapeutic intervention can prevent subsequent transition into full-brown psychosis also shows the practical value of this idea. In Figure 2, PIP and various related conditions are schematically summarized.

Diagnosing PIP

In seizure-monitoring units, PIP lends itself to being easily recognized such that no specific advice seems necessary, except for knowing that such a state exists. However, without careful attention, an initial sign of elevated mood or emotional lability may be overlooked, which could result in failure to apply early intervention, and an otherwise avoidable psychiatric emergency may ensue. It should be also noted that a sensitive inquiry can disclose subtle changes of subjective feeling such as derealization (“outer world looks as if it was not real”) or an otherwise barely noticeable lowering of cognitive function even in a seemingly normal lucid interval (Ito, 2010; Schulze-Bonhage & van Elst, 2010). In principle, it is highly advisable that a careful examination of psychiatric as well as neuropsychological status should be repeated at least within 3 days following a seizure cluster in seizure monitoring units.

As for outpatients, especially in cases of the first episode of PIP, a preceding seizure or seizure cluster can be overlooked, which may obscure the true nature of the presentation. If an otherwise taciturn person with long-standing temporal lobe epilepsy (TLE) becomes suddenly eloquent or if a reticent one becomes unexpectedly irritable or sticky, medical personnel should be on alert, as it may herald the beginning of PIP even if preceding seizures cannot be confirmed. If it is truly an initial sign of PIP, frank psychosis or seriously annoying behavioral changes including menacing aggression (Gerard et al., 1998; Kanemoto et al., 2010) will develop, in the absence of appropriate intervening measures, within 48 h.

Associated characteristics

Long-standing (average duration >15 years) TLE is a typical comorbid condition of PIP (Savard et al., 1991; Devinsky et al., 1995; Kanemoto et al., 1996a; Kanner et al., 1996; Adachi et al., 2002a; Kanemoto, 2011). Other types of focal epilepsies, especially frontal lobe epilepsy, are also associated at times (Adachi et al., 2000b; Nishida et al., 2006; Alper et al., 2008). On the other hand, coexistence of idiopathic generalized epilepsies and PIP is quite exceptional (Chakrabarti et al., 1999).

In a substantial proportion of cases, involvement of the medial temporal structures is indicated on MRI, which may also extend to the lateral temporal and extratemporal regions (Kanemoto et al., 1996b; Sundram et al., 2010). Furthermore, epileptiform discharges on surface EEG in temporal regions tend to be bilateral (Kanner & Ostrovskaya, 2008a; De Toffol, 2009). In patients with repetitive episodes of PIP, a specific personality trait such as hyperreligiosity or viscosity may supervene (Trimble, 2007; Devinsky & Lai, 2008).

Treatment principles for PIP

The therapeutic approach to PIP comprises acute protective measures and the prevention of repeat episodes (Lancman et al., 1994; Kanemoto, 2002; Devinsky, 2008).

Acute management of PIP or imminent PIP

Initial stage preceding PIP

In the nuclear type, elevated mood or peculiar irritability often precedes a full-blown PIP episode (Schulze-Bonhage & van Elst, 2010). If patients are successfully treated at this stage, development of frank psychosis might be thwarted. As noted above, PIP is closely linked with various abnormal behaviors that can do great harm to the subsequent social life of the patient. In cases of long-standing TLE, patients and families should be warned beforehand to seek medical help as soon as possible and not leave the patient alone if any suspicious change in behavioral patterns appear after a seizure or seizure cluster (Krauss & Theodore, 2010). In the seizure monitoring unit, every seizure cluster or secondarily generalization should be taken as a possible precipitant of PIP, and indicates the need for intensified observation.

Full-blown PIP

Once PIP develops fully, immediate protective custody is unavoidable in many cases and any delay in protection may lead to serious problems, including destructive or self-harming behavior (Devinsky et al., 1995; Kanner et al., 1996; Kanemoto et al., 1999; Fukuchi et al., 2002). In patients with a history of violent PIP episodes in the past, the risk of similar behaviors being repeated is particularly high. A potent sedative agent is worth trying, which may shorten the duration of the episode. Paradoxically, ECT may be helpful in terminating violent PIP attacks in exceptional cases (Pascual-Aranda et al., 2001). Repetitive transcranial magnetic stimulation might be more safely applied than ECT in the future, although it remains an experimental treatment for PIP at present (Krauss & Theodore, 2010).

Preventive strategy in the long term

In contrast to interictal psychosis inclusive of alternative psychosis, seizure control generally prevents recurrence of PIP episodes (Kanemoto et al., 2001a,b). In some patients, a simple reappraisal of pharmaceutical therapy can succeed in stopping seizures or seizure clusters, consequently preventing PIP episodes as well. However, a considerable number of patients have already tried nearly all available antiepileptic agents by the time PIP appears, which shows a rather discouraging prospect for drug therapy to achieve complete seizure freedom. In addition, in a clinical setting like this, intensive antiepileptic pharmacotherapy may be prone to induce various psychiatric symptoms, including interictal psychosis and depression. Operative intervention, if applicable, is often the last hope for a breakthrough in affected patients. Indeed, successful operative intervention leads to cessation of PIP episodes as well, except in some cases (Christodoulou et al., 2002). Nevertheless, it should be noted that epilepsy-related variables associated with PIP such as structural changes beyond medial temporal lesion and bilateral EEG foci may predict poor surgical outcome (Alper et al., 2001; De Toffol, 2009), although that has not yet been confirmed by actual investigations (Winesett & Benbadis, 2010). An additional cautionary note is also in order. Patients with preoperative episodes of PIP or even prolonged postictal confusion may have a higher risk of developing serious depression (or even a manic state) within 3 months after surgery (Kanemoto et al., 2001a,b). Furthermore, in presurgical evaluations of patients with a history of PIP, seizure cluster or secondary generalization is highly likely to replicate similar episodes, so that every step should be taken in advance to avoid inducing PIP episodes and, once PIP manifests, to treat it as soon as possible (Kanner, 2009). Although these anticipated troubles may discourage surgical intervention, successful surgery can dramatically change the life of affected patients, even more dramatically than in those with epilepsy but without PIP. Therefore, patients with PIP should not be deprived of a chance for surgical intervention beforehand, as some are in fact excellent candidates.

How to use medical treatment

For psychoses associated with epilepsy, a systemic review of literature up to 2008 revealed only one underpowered, randomized controlled trial, from which obviously few conclusions can be drawn (Farooq & Sherin, 2008). When it comes to pharmacotherapy for PIP, limitations in the evidence should be always kept in mind, since nearly all advice remains at the level of expert opinion. Furthermore, the initial target of pharmacotherapy often needs to be directed toward violent or agitated behavior, rather than the psychosis itself. This means that measures established for rapid tranquilization of patients with schizophrenia may not automatically be applicable to those with PIP.

Acute phase

In both the initial stage preceding PIP and during the episode, sedative drugs are generally needed, as they not only stop socially inappropriate behaviors but can also abort or alleviate symptoms of PIP. Some experts have recommended benzodiazepines (Lancman et al., 1994), whereas others prefer a combination of both benzodiazepines and antipsychotics (Kanner et al., 1996). Although antipsychotics are known to lower seizure threshold in vitro, except for some notable exceptions, such as zotepine and clozapine (Devinsky et al., 1991), their proconvulsive effects seem to be at a tolerable level, at least as long as they are given within a therapeutic range. In most cases, imminent need of sedation has priority over a possible proconvulsive risk.

In some exceptional cases, early stage PIP can be managed and thwarted with a relatively low dose of oral benzodiazepines. However, in other extreme cases, sedatives should be given swiftly as well as massively enough to force immediate tranquilization. Depending on how serious the detrimental impact is on the patient and their surroundings, and whether an inpatient or outpatient situation is in question, different combinations of pharmaceutical interventions can be chosen, as follows.

  1. Oral administration of benzodiazepine (e.g., lorazepam).
  2. Combined oral administration of benzodiazepine and dopamine-blocker (e.g., risperidone, olanzapine, quetiapine, chlorpromazine).
  3. Intramuscular administration of dopamine-blocker (e.g., haloperidol plus promethazine).

In cases without seriously alarming behaviors either during precedent or current PIP episodes, regimen A may suffice. However, even when observed signs are confined to slightly elevated mood or irritability, the patient should be kept under strict scrutiny at least for the next 24 h. Meanwhile, if any sign of increasing excitability is noticed, a switch to regimen B may be necessary, because benzodiazepines alone may precipitate paradoxical excitement and are not as potent as dopamine-blockers for rapid tranquilization of violent or agitated patients in a psychiatric emergency (Alexander et al., 2004; Allen et al., 2005). When confronted with seriously alarming current behaviors or in the presence of past violent episodes during PIP, if the patient is unable to cooperate, regimen C may be left as the only choice. Again, it should be noted that there are no reliable data to favor one antipsychotic over another.

Interepisodic phase

Robust manifestations of PIP usually fade away within a week and patients seem to superficially return to their former selves. However, unexpected outbursts caused by minimal stimuli can occur sporadically within the next few weeks afterward (Oshima et al., 2006). In this unstable stage, a reduced amount of regimen A or B often needs to be maintained. Sedatives should be reduced gradually and can be tapered completely on average within 4 weeks and mostly within 3 months. Except for cases with combined chronic psychosis and severe personality change, continuous use of psychotropic agents is not generally recommended during the interepisodic phase. No reliable data are available to judge whether a switch of antiepileptic drug to one with more psychotropic efficacy, such as valproate, carbamazepine, and lamotrigine, is beneficial for the prophylaxis of PIP (Krauss & Theodore, 2010).

Prognosis of PIP

In considering prognosis, each episode of PIP is essentially a benign, self-remitting condition. Only temporary measures such as physical restriction or medical sedation are needed to cope with risk behaviors aimed at protecting the safety of patients and others. In comparison with IIP, the duration of PIP is clearly shorter. In a recent collaborative study, approximately 95% of PIP episodes resolved within 1 month (Adachi et al., 2007). In another study, PIP episodes tended to resolve within 1 week in an overwhelming majority of patients in a seizure monitoring setting (Devinsky et al., 1995; Kanner et al., 1996). Nevertheless, it should be noted that suicidal attempts are frequent complications of PIP episodes and without intervention there may be a heightened risk of mortality (Devinsky et al., 1995; Kanemoto et al., 1996a; Kanner et al., 1996; Fukuchi et al., 2002).

In half of affected patients, PIP remains as a single episode. In the other half, PIP episodes tend to repeat, which may occur even quasiregularly after a cluster of complex focal seizures or secondary generalization in extreme cases (Logsdail & Toone, 1988; Kanemoto et al., 1996a,b; Kanner et al., 1996; Liu et al., 2001). In some patients, repetitive PIP episodes finally develop into chronic interictal psychosis (Tarulli et al., 2001; Adachi et al., 2003; Kanner & Ostrovskaya, 2008b).

Future issues

Although data gathered from cases with invasive EEG recordings during PIP are highly limited (Wieser et al., 1985; So et al., 1990; Mathern et al., 1995; Kanemoto, 1997; Seeck et al., 1999; Takeda et al., 2001; Schulze-Bonhage & van Elst, 2010), it has become increasingly clear that only a limited proportion of PIP or related conditions correspond to ictal EEG correlates. Indeed, sporadic single photon emission computed tomography (SPECT) studies have suggested that early frontal and late medial temporal hyperactivities may play some role in the genesis of PIP (Fong et al., 2000; Leutmezer et al., 2003; Nishida et al., 2006; Oshima et al., 2011). However, in a substantial number of cases, these hyperactivities do not seem to be directly linked with epileptiform discharge. Some authors have speculated that postictal alterations of neurochemical settings may be found somewhere in the limbic circuit. Stevens (1983) regarded these alterations as excessive inhibitory surround and, more specifically, Bortolato and Solbrig (2007) assumed that increased hippocampal dynorphin release may serve as a psychotomimetic agent via overstimulation of kappa opioid receptors in PIP. From an electrophysiologic point of view, Kuba et al. (2012) recently found a unique slow rhythmicity in depth-EEG recordings during PIP episodes, which clearly differed from ictal epileptiform discharge appearing in limbic status epilepticus. Also, as was recently suggested (Oshima et al., 2006), subcategorizing of PIP into nuclear type with a lucid interval and initial hypomanic stage and atypical periictal type may be helpful for understanding the underlying pathophysiology of PIP (Fig. 3A,B). Further accumulation of relevant cases is highly anticipated.

Figure 3.

(A) Nuclear PIP; (B) Periictal psychosis (with permission from Kanemoto, 2011).

There are genetic as well as phenomenologic data suggesting a close affinity of PIP with bipolar disorder (Alper et al., 2001; Kanemoto et al., 2001a,b). PIP episodes can be well observed and investigated because of their limited duration and clear-cut on-and-off psychotic states. Increased knowledge of how PIP develops may also help to clarify the mechanisms related to how relevant psychiatric disorders without epilepsy develop, which may help bridge neurology and psychiatry.


Psychotic disorders are not infrequent in epileptic patients. IIP reportedly occurs in 4.5–7.0% (Trimble, 1991) and PIP in <2% (Kanemoto et al., 2001b). In those with PEs, the incidence of PIP is doubled (Kanner et al., 1996; Kanemoto et al., 2002; Oshima et al., 2006; Falip et al., 2009), whereas the incidence is threefold in seizure monitoring units (Kanner et al., 1996; Alper et al., 2001, 2008; Falip et al., 2009). IIP and PIP are serious disorders with significant effects on patients and their families. A number of cases may become psychiatric emergencies or pose significant risk to others. Such risks can be reduced or avoided if prompt and adequate treatment is put in place. There is now international consensus among experts that all medical personnel involved in providing health care to patients with epilepsy should be well-informed about IIP and PIP (Kerr et al., 2011).

Based on the limited data and consensus statements currently available, we have proposed treatment procedures for IIP and PIP that are concise and practical. We hope that this article will raise awareness of the importance of these matters and the need for further studies. With the accumulation of further evidence, the proposed treatment systems should be refined and placed on even greater empirical footing.


None of the authors have any conflicts of interests. We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines. The contents of this supplement reflect the opinions of the individual authors and do not necessarily represent official policy or position of the ILAE.