Historical Aspects and Nomenclature
An association between psychotic phenomena and epilepsy has been described since ancient times. However, it was not until the 19th century when the close relationship between these two conditions was studied in a systematic manner. Epilepsy continued to be associated with “madness” in one form or another to the point where various authors considered psychotic episodes to be an “epileptic equivalent.”Reference Trimble 1
In the 1950s, several investigators recognized that patients with epilepsy, particularly those with temporal lobe epilepsy (TLE), suffered from psychotic phenomena that differed in many ways from the schizophrenias,Reference Hill 2 ie, these patients did not exhibit the lack of affect and the “asocial or withdrawn attitude” that were typical in patients with schizophrenia. In their classic paper, Slater et al Reference Slater, Beard and Glithero 3 also emphasized these differences.
Today, psychotic disorders identified in patients with epilepsy are commonly referred to in the medical literature as psychosis of epilepsy (POE). This is a term applied to a group of psychotic disorders with distinct phenomenology and suspected etiopathogenic mechanisms that are likely to be closely related to the seizure disorder.Reference Ferguson and Rayport 4 For example, in a review of the literature, Rayport and FergusonReference Rayport and Ferguson 5 described how episodic POE is related to seizure recurrence and remits when seizures are controlled.
The purpose of this manuscript is to review the different clinical manifestations of the various forms of POE, their potential pathogenic mechanisms, and current management.
Epidemiology
The prevalence rates of primary schizophreniform disorders in the general population range between 0.4% and 1%. In contrast, the prevalence of POE has been estimated to be between 7% and 10%.Reference Trimble 1 A 3-year epidemiological study held in The Netherlands revealed a prevalence of psychosis in 5.6% of epilepsy patients.Reference Van der Feltz-Cornelis, Aldenkamp, Adèr, Boenink, Linszen and Van Dyck 6 The overall evidence suggests that schizophrenia-like psychosis is believed to be 6 to 12 times more likely to occur in people with epilepsy than in the general population.Reference Mendez, Grau, Doss and Taylor 7
Most of the above-cited data are based on cross-sectional studies. Yet, a bidirectional relationship between psychotic disorders and epilepsy has been suggested in population-based studies that investigated the temporal relation between the onset of the psychiatric and epileptic disorders. For example, in a study conducted in Sweden, the age-adjusted odds ratio (OR) for unprovoked seizures was 2.3 higher for patients with a primary psychotic disorder compared to a control group of subjects matched for gender and year of diagnosis who were selected randomly from the register of the Stockholm County population.Reference Adelöw, Andersson, Ahlbom and Tomson 8 Likewise, in a population based-study conducted in Taiwan, patients with schizophrenia had a 6-fold higher risk to develop epilepsy than controls,Reference Chang, Chen and Tsai 9 while in a population-based study using the United Kingdom’s General Practice Research Data Base, the incidence rate ratios of psychosis, anxiety, and depression were significantly higher during the 3 years preceding and during the 3 years that followed the diagnosis of epilepsy.Reference Hesdorffer, Ishihara, Mynepalli, Webb, Weil and Haurser 10
Clinical Manifestations
Traditionally POE has been classified according to the temporal relationship of seizure occurrence, with psychotic episodes divided into ictal (eg, the psychotic symptoms are the clinical expression of the seizure), postictal (eg, episodes occur up to 120 hours after the seizure); and interictal (eg, episodes that are independent of seizures).Reference Kanner 11 Alternative psychotic episode (APE), also known as the phenomenon of forced normalization, is another type of POE that occurs in patients with a chronic history of treatment-resistant epilepsy who suddenly become seizure-free.Reference Landolt 12 , Reference Tellenbach 13
Different classifications of POE have been proposed. For instance, Rayport and Ferguson grouped POE into 2 categories: episodic psychosis of epilepsy and chronic (or non-episodic) psychosis of epilepsy.Reference Ferguson and Rayport 4 , Reference Rayport and Ferguson 5 The former has been identified primarily in patients with TLE, and, according to these authors, is closely linked to the effectiveness of seizure control. Episodic psychotic episodes last from a few days to several weeks. Rayport and Ferguson also pointed out that episodic exacerbations could also occur in patients with non-episodic disorders if the seizures worsen. Thus, Rayport and Ferguson’s episodic psychosis of epilepsy corresponds to postictal psychotic episodes (PIPE) and APE, while the non-episodic psychoses are equivalent to the interictal psychosis of epilepsy (IPE). Finally, in addition to IPE, PIPE, and APE, any classification should also include psychotic processes resulting from iatrogenic effects of antiepileptic drugs (AEDs) or surgical treatment.
Interictal psychosis of epilepsy (IPE)
Interictal psychotic episodes may be indistinguishable from primary schizophreniform disorders,Reference Perez, Trimble, Murray and Reider 14 but in a significant percentage of patients, the semiology and the course of the psychotic disorder may differ. A distinguishing feature of IPE is the late onset of symptoms in comparison to patients with schizophrenia.Reference Slater, Beard and Glithero 3
In contrast to schizophrenia, IPE is characterized by the absence of negative symptoms, a better premorbid condition, and an unlikely deterioration of personality. This is stressed on Slater’s observation that “the delusions and hallucinations of patients with POE were empathizable (the patient remains in our world).”Reference Slater, Beard and Glithero 3 There is general agreement about the lesser severity, and also that better response to therapy can be anticipated in POE.Reference Toone, Garralda and Ron 15
The annual prevalence of IPE has been estimated to range between 0.3% to 0.4% of adults with epilepsyReference Onuma, Adachi, Ishida, Katou and Uesugi 16 , Reference Tadokoro, Oshima and Kanemoto 17 ; it can present as an acute episode, lasting from days or weeks, or a chronic disorder, lasting for months or years. Reference Adachi, Kato and Sekimoto 18 , Reference Sachdev 19 For example, in a group of 155 patients with epilepsy followed for a period of 16.9 years, the mean duration of the psychotic episode was 82.7 weeks, with the longest episode lasting up to 150.1 weeks; of note, no clinical characteristics were significantly associated with the duration of the IPE.Reference Adachi, Akanuma, Ito, Okazaki, Kato and Onuma 20
Postictal psychosis of epilepsy (PIPE)
PIPE has been estimated to represent 25% of all types of POE.Reference Dongier 21 Yet, many of these cases remain under-recognized and/or often are misdiagnosed as an IPE.Reference Trimble, Kanner and Schmitz 22 While PIPEs are of limited duration and remit following treatment with low doses of antipsychotic drugs or, sometimes, even benzodiazepines, they must nonetheless be taken seriously, as they are associated with increased mortality caused by suicide. Postictal psychotic phenomena can present as isolated symptoms or as fully developed psychotic episodes.
Logsdail and Toone’s diagnostic criteria for PIPE continue to be widely used.Reference Trimble, Kanner and Schmitz 22 , Reference Logsdail and Toone 23 These include
-
1. Onset of confusion or psychotic symptoms within 1 week of the return of apparently normal mental function after a seizure (and /or cluster of seizures)
-
2. Duration of 1 day to 3 months
-
3. Mental state characterized by (a) louding of consciousness, disorientation, or delirium; (b) delusions or hallucinations, in clear consciousness; (c) a mixture of (a) and (b)
-
4. No evidence of factors that may have contributed to the abnormal mental state, such as anticonvulsant toxicity, a previous history of interictal psychosis, electroencephalogram (EEG) evidence of status epilepticus, and a recent history of head injury or alcohol/drug intoxication
While treatment-resistant focal epilepsy has been identified in a majority of the studies of PIPE,Reference Devinsky, Kelley, Porter and Theodore 24 – Reference Umbricht, Degreef, Barr, Lieberman, Pollack and Schaul 26 primary generalized epilepsy has also been reported.Reference Logsdail and Toone 23 , Reference Devinsky, Kelley, Porter and Theodore 24 In most studies, a “lucid interval” between the restoration of an apparent “normal mental state” after the seizure and the beginning of the psychotic episode is described. This could last from 12 hours to up to 7days.Reference Mendez, Grau, Doss and Taylor 7 , Reference Devinsky, Abramson and Alper 27 – Reference Lancman, Craven, Asconape and Penry 29 This symptom-free time period has accounted for the failure to recognize PIPE, as clinicians do not recognize the temporal relation between the seizure and the psychotic episode.Reference Trimble, Kanner and Schmitz 22
Postictal psychotic symptoms
Kanner et al Reference Kanner, Soto and Gross-Kanner 30 investigated the prevalence of postictal psychiatric symptoms in 100 consecutive patients with pharmaco-resistant focal epilepsy during a 3-month period.Reference Kanner, Soto and Gross-Kanner 30 The postictal period was defined as the 72 hours that followed recovery from the last seizure. Questions were asked about the frequency of occurrence of each symptom. Only symptoms that were identified after more than 50% of seizures were included in this study, so as to reflect a “habitual” occurrence. To ensure that patients were reporting postictal psychiatric symptoms, each question of the survey also asked about occurrence of the same symptoms during the interictal period. When similar symptoms were identified during interictal and postictal periods, the only symptoms considered to be postictal were those that were significantly more severe during the postictal period. These were then classified as a postictal exacerbation of interictal psychiatric symptoms. Among the 100 patients, 79 patients had seizures of temporal origin, and 21 had seizures of extratemporal origin. Half of the patients had only complex partial seizures (CPS); the other half had both CPS and generalized tonic clonic (GTC). Fifty-two patients had a history of psychiatric conditions: depression, anxiety disorders, and attention deficit disorders. Seven patients experienced postictal psychotic symptoms after more than 50% of their seizures; the duration of the “postictal psychotic symptoms only” ranged between 1 and 108 hours (median: 15 hours). Such symptoms always occurred together with postictal symptoms of depression and anxiety.
Postictal psychotic episodes
As mentioned previously, postictal psychiatric symptoms may cluster into PIPE. Since the advent of video-EEG monitoring (V-EEG) more than 4 decades ago, recognition of PIPE has increased substantially. This is not surprising, because the circumstances of V-EEG are such that the likelihood of psychiatric symptoms occurring is increased (due, for example, to the facilitation of frequent seizures over a short time period following discontinuation or dose reduction of AEDs). In a 1996 study, the yearly incidence of postictal psychiatric disorders among patients with focal epilepsy undergoing V-EEG was 7.9%.Reference Kanner, Stagno, Kotagal and Morris 31 The majority (6.4%) presented as PIPE. Among the 10 patients with PIPE, 4 patients had a delusional psychosis, 1 patient had a mixed manic-depressive–like psychosis, 2 had psychotic depression, 1 had a hypomanic-like psychosis, and 1 had a manic-like psychosis. The 10th patient presented with bizarre behavior associated with a thought disorder. In every case, the onset of symptoms lagged the last seizure by a mean period of 24 hours (range 12–72 hours). The mean duration of the PIPE was 69.6 hours (range 24–144). In 5 patients, psychotic episodes remitted with low doses of a neuroleptic drug (2–5 mg/day of haloperidol), although 1 patient required high doses of haloperidol (40 mg/day), and remission occurred in 4 patients without pharmacotherapy. Six of the 10 patients had an average of 2.4 PIPE prior to V-EEG, while in the remaining 4 patients, it was the first one.
Other authors have reported similar findings with respect to clinical characteristics, course, and response to pharmacotherapy.Reference Logsdail and Toone 23 , Reference Kanner and Ostrovskaya 25 – Reference Kanemoto, Kawasaki and Kawai 28 , Reference Savard, Andermann, Olivier and Rémillard 32 , Reference So, Savard, Andermann, Olivier and Quesney 33 Common findings among the different case series include (1) a symptom-free interval between the last seizure and the onset of psychiatric symptoms; (2) a relatively short duration (even shorter durations were noted among our patients, as only 2 had episodes lasting more than 5 days, while this was true in 5 of 9 patients in Savard et al’sReference Savard, Andermann, Olivier and Rémillard 32 and in 7 of 14 patients in Logsdail and Toone’s seriesReference Logsdail and Toone 23 ; (3) affect-laden symptoms; (4) clustering of symptoms into delusional and affective-like psychosis; (5) increased numbers of secondarily generalized tonic-clonic seizures before the onset of PIPE; (6) onset of PIPE after having had seizures for more than 10 years; and (7) a prompt response to low dose neuroleptic medication or benzodiazepines.
Kanemoto et al Reference Kanemoto, Kawasaki and Kawai 34 studied the clinical differences between PIPE and acute and chronic interictal psychosis of epilepsy. They noted that patients with PIPE were more likely to experience grandiose and religious delusions in the presence of elevated moods, as well as a sense of mystic fusion of the body with the universe. On the other hand, perceptual delusions or commenting voices were less frequent in PIPE, though feelings of impending death were common among patients with PIPE.
Various investigators have tried to identify the pathogenic mechanisms of PIPE. For example, Kanner and OstrovskayaReference Kanner and Ostrovskaya 25 compared 18 consecutive adults with focal seizure disorders and PIPE, and 36 patients with focal epilepsy without PIPE. Data analysis included the number and location of ictal foci recorded by V-EEG, seizure type, etiology, age at seizure onset, duration of seizure disorder, magnetic resonance imaging (MRI) abnormalities, and psychiatric history before the index V-EEG (other than PIPE).Reference Kanner and Ostrovskaya 25 Significant differences included the presence of bilateral independent ictal foci on V-EEG (these were identified as an independent predictor of the development of PIPE in univariate and multivariate analyses). The logistic regression model correctly classified 89% of patients. A second independent predictor of PIPE was having only secondarily generalized tonic-clonic seizures. Conversely, the occurrence of PIPE and cryptogenic focal epilepsy were predictive of bilateral independent ictal foci in univariate analyses. In a recent study that included 59 consecutive patients with focal epilepsy and a history of PIPE, and 94 controls with focal epilepsy and no history of PIPE, Alper et al Reference Alper, Kuzniecky and Carlson 35 found that predictors of PIPE included ambiguous or extratemporal epileptogenic localization, a family history of psychiatric disorders, multifocal interictal EEGs, and encephalitis. Other investigators have also identified bilateral independent interictalReference Devinsky, Abramson and Alper 27 ) and ictalReference Umbricht, Degreef, Barr, Lieberman, Pollack and Schaul 26 foci, as well as the presence of secondarily GTC.Reference Logsdail and Toone 23 , Reference Umbricht, Degreef, Barr, Lieberman, Pollack and Schaul 26 , Reference Savard, Andermann, Olivier and Rémillard 32
Prompt recognition of PIPE has important clinical implications. First, its recurrence can be minimized by starting an atypical neuroleptic drug at a low dose at the first sign of PIPE (eg, risperidone 1–2 mg). In most cases, insomnia is the initial presenting symptom. Family members need to be informed about early symptoms so that they can assist in the timely administration of the antipsychotic drug. Patients should be maintained on a dose of the drug for 2–5 days, after which it can be discontinued, unless the symptoms fail to remit or progress.
Patients with a family psychiatric history, who have a cluster of secondarily generalized tonic-clonic seizures in the course of V-EEG monitoring, and who are found to have bilateral independent ictal foci should be watched carefully for the development of PIPE during the first week after the last seizure.
Occurrence of PIPE has implications with respect to the localization of ictal foci. As suggested by the data cited above, development of PIPE should raise the possibility of bilateral independent ictal foci. This is especially important in patients who are being considered for epilepsy surgery. The consequence of this is that such patients may require longer V-EEG monitoring studies and, possibly, use of intracranial electrodes. If recordings with depth or subdural electrodes are undertaken, prophylactic treatment with low-dose atypical antipsychotic drugs can avert the occurrence of PIPE during the period of invasive V-EEG monitoring.Reference Kanner, Stagno, Kotagal and Morris 31
Relation between PIPE and IPE
There appears to be a bidirectional relationship between PIPE and IPE. For example, in a retrospective study of 18 consecutive adults who have focal seizure disorders with PIPE and 36 patients with focal seizure disorders without PIPE, Kanner and OstrovskayaReference Kanner and Ostrovskaya 25 found that 7 patients with PIPE and 1 control patient went on to develop IPE.
Whether preventing further PIPE protects patients from developing IPE has not been established. Such risk, however, supports the recommendation to consider epilepsy surgery whenever possible, even if it is only palliative.Reference Kanner and Ostrovskaya 25 Remission of IPE and PIPE has been reported after epilepsy surgery in a woman who had suffered from TLE since the age of 15 years.Reference Marchetti, Tavares, Gronich, Fiore and Ferraz 36 She developed recurrent PIPE at the age of 35, and this evolved to chronic refractory IPE. The presurgical evaluation demonstrated right hippocampal atrophy and a right mesial temporal epileptogenic focus. Following a right temporal lobe resection, she has been seizure-free and the IPE has remitted.
While an association between a history of PIPE and the development of IPE has been identified in non-surgical case series,Reference Marchetti, Tavares, Gronich, Fiore and Ferraz 36 , Reference Tarulli, Devinsky and Alper 37 this has not been reported in studies of patients who have undergone epilepsy surgery. However, Kanemoto et al Reference Kanemoto, Kim, Miyamoto and Kawasaki 38 found a significant risk of postsurgical mood disorders in patients with a history of presurgical PIPE. Indeed, preoperative occurrence of PIPE was 5 times more frequent among patients with post-operative mood disorders (38%) than among those who did not develop them (7%). The relation between mood disorders and PIPE was further established in a study by Alper et al,Reference Alper, Devinsky and Westbrook 39 which found a higher prevalence of mood disorder among first- and second-degree relatives as the only psychiatric variable that predicted the development of PIPE on logistic regression analyses (odds ratio=3.49, P=0.001).Reference Alper, Devinsky and Westbrook 39
A bidirectional relationship between PIPE into IPE was also reported by Tarulli et al,Reference Tarulli, Devinsky and Alper 37 who conducted a retrospective study of 43 patients with PIPE. Five (13.9%) patients developed IPE after multiple documented PIPE, while in 1 patient, IPE preceded PIPE. The length of time between PIPE and IPE ranged from 7 to 96 months. Adachi et al Reference Adachi, Kato and Sekimoto 40 also reported a bidirectional relationship between IPE and PIPE (which they called “bimodal psychosis”) in a study of 14 patients.Reference Adachi, Kato and Sekimoto 40 Ten patients who had PIPE went on to develop IPE, while 4 patients had IPE that remitted but later developed PIPE. Mean age at onset of epilepsy was 10.8±4.3 years, at the time of the first IPE 24.4±6.1 years, and at the time of the initial PIPE 33.8±4.5 years. These patients did not differ with respect to the epilepsy-related characteristics found in patients with only PIPE: they had bilateral EEG abnormalities and borderline (or decreased) intellectual function.
In the study by Tarulli et al,Reference Tarulli, Devinsky and Alper 37 the symptomatology of PIPE and IPE was similar or identical in 5 of 6 cases, and this was also true in our study. However, differences were found in other studies. For example, Kanemoto et al Reference Kanemoto, Kawasaki and Kawai 34 compared the clinical semiology among 30 patients with PIPE, 33 patients with acute IPE, and 25 patients with chronic IPE. The most striking feature that distinguished PIPE from both acute interictal and chronic psychoses was occurrence of grandiose, religious, and mystical delusions with markedly elevated moods. In addition, patients with PIPE exhibited few schizophreniform psychotic traits such as perceptual delusions or auditory hallucinations.
Alternative psychosis or forced normalization
In 1953, LandoltReference Landolt 12 coined the concept of forced normalization after observing the development of psychotic symptoms in patients who had suffered from persistent seizures for a long time and who became seizure free. Tellenbach proposed the term “alternative psychosis.”Reference Tellenbach 13 This antagonism between psychosis and epilepsy has been considered by some as the explanation for the therapeutic effect of electroconvulsive therapy (ECT) of psychotic disorders. This phenomenon is relatively rare. Landolt reported 47 cases between 1951 and 1958. Single case studies were reported by others, and in 1988, Schmitz and WolfReference Schmitz and Wolf 41 estimated the prevalence of alternative psychosis to be 1% among 697 patients followed at a university epilepsy center. Forced normalization has been reported in patients with TLE and generalized epilepsies. As with other forms of POE, WolfReference Wolf 42 identified the psychotic manifestations of an APE after a mean 15.- year history of epilepsy in 23 patients. Both Landolt and Wolf reported a pleomorphic clinical presentation with a paranoid psychosis without clouding of consciousness being the most frequent manifestation. As with other POE, a richness of affective symptoms has been identified.
The phenomenon of forced normalization has been observed following the use of various AEDs, including phenytoin and primidone,Reference Wolf 42 , Reference Gibbs 43 carbamazepine and valproic acid,Reference Pakalnis, Drake, John and Kellum 44 and vigabatrin,Reference Sander, Hart, Trimble and Shorvon 45 and this author (AMK) has had 3 cases associated with benzodiazepines, 1 with clonazepam, and 2 with clobazam (unpublished data). In these cases, the psychotic disorder was thought to result from the suppression of seizures and not from an adverse event of the AED.Reference Ried and Mothersill 46
The pathogenic mechanisms mediating APE are yet to be identified. Some hypotheses have been proposed. TrimbleReference Trimble 1 postulated that an excess of dopaminergic action could explain the seizure cessation and the psychotic symptomatology. Rayport and Ferguson suggested that forced normalization is not the expression of seizure cessation, but rather, of a “voltage depression or suppression” in neocortical derivations concurrent with ictal activity in amygdala or hippocampal structures.Reference Ferguson and Rayport 4 , Reference Rayport and Ferguson 5 Depth electrode studies by Wieser et al Reference Wieser, Hailemariam, Regard and Landis 47 appear to lend support to this hypothesis.
The treatment of APE remains the source of debate. Ried and Mothersill,Reference Ried and Mothersill 46 for example, concluded that the treatment of alternative psychotic episodes should include the reduction and/or discontinuation of AEDs until overt seizure recurrence cause remission of psychotic symptoms. The rapidity with which AED should be tapered is not clear. TellenbachReference Tellenbach 13 suggested a rapid tapering under EEG monitoring. LandoltReference Landolt 12 advocated the use of ECT or metrazol if necessary. Following the recurrence of seizures and the remission of psychotic symptoms, AEDs should be re-introduced slowly.Reference Ried and Mothersill 46
Iatrogenic psychotic episodes
AED-related psychotic episodes
Psychotic symptoms and episodes as an expression of an iatrogenic phenomenon have been reported with most AEDs, including the first generation AEDs ethosuximide, phenytoin, phenobarbital, and primidone,Reference Perlo and Schwab 48 , Reference Rivinus 49 as well as the newer AEDs, such as vigabatrin, topiramate, levetiratecetam, zonisamide, lacosamide, pregabalin, and parampanel.Reference Chatzistefanidis, Karvouni, Kyritsis and Markoula 50 – Reference Pedroso, Nakama, Carneiro Filho and Barsottini 54 The clinical differentiation between APE and a toxic reaction can be difficult if a seizure-free state followed the introduction of the AED.
Psychotic disorders can occasionally follow the discontinuation of AEDs, particularly those with mood stabilizing properties. Ketter et al Reference Ketter, Malow, Flamini, White, Post and Theodore 55 reported the development of anxiety and depression, primarily, but also of some cases among 32 inpatients who experienced psychosis and who were withdrawn from carbamazepine and valproic acid. Acute withdrawal from benzodiazepines is well known to result in an acute psychotic episode.Reference Sironi, Franzini, Ravagnati and Marossero 56
Postsurgical psychotic episodes
Postsurgical psychotic disorders are a relatively rare complication of epilepsy surgery with prevalence rates estimated to range between 3% and 10% among patients undergoing anterior-temporal lobe resections.Reference Jensen and Vaernet 57 – Reference Trimble 59 Several case series have included a mixture of patients with presurgical and de novo postsurgical psychotic disorders. TrimbleReference Trimble 59 calculated that the prevalence rate of postoperative de novo psychoses ranged between 3.8% and 35.7% (mean, 7.6%). He suggested that at least some of these cases may be an expression of forced normalization.Reference Trimble 59 The observation that epileptic activity in the hippocampus leads to plastic changes and the aberrant reinervation phenomenon and reorganization of synapses supports this suggestion.Reference Foong and Flugel 60 Many epilepsy centers currently do not consider patients with a preoperative history of psychosis as potential candidates for epilepsy surgery. Thus, the lower prevalence rates of more recent reports of postsurgical psychotic episodes are primarily de novo psychoses.
De novo postsurgical psychotic episodes can mimic schizophreniform disorders, manic episodes, or present as PIPE. In a study carried out at the Rush Epilepsy Center, Kanner et al (unpublished data) identified de novo psychotic episodes in 4 patients within the first 6 months after an anterior temporal lobe resection. Manifestations consisted of manic episodes in 2 and paranoid episodes in 2 (unpublished data). Symptoms remitted in 3 patients using pharmacotherapy without need for hospitalization. One patient had to be hospitalized in a psychiatric unit. Two of the patients had lesional epilepsy: in one it was caused by dysembryoplastic neuroepithelioma (DNET), and in the other by a ganglioglioma. Shaw et al Reference Shaw, Mellers, Henderson, Polkey, David and Toone 61 reported 11 patients with postoperative new-onset schizophreniform psychosis among 320 consecutive patients (3.2%) who underwent anterior temporal lobe resection. Psychotic symptoms became apparent within the first postoperative year in all patients. They were significantly more likely to have bilateral epileptiform activity, a smaller amygdala in the non-operated side, and pathologies other than mesial temporal sclerosis than a comparison non-psychotic group.
In a study of 57 consecutive patients who underwent anterior temporal lobe resection, Leinonen et al Reference Leinonen, Tuunainen and Lepola 62 identified 5 (8.8%) who developed postoperative psychotic episodes.
Some investigators have associated the risk of postsurgical psychotic episodes with a right temporal seizure focus. For example, Mace and TrimbleReference Mace and Trimble 63 reported 7 consecutive patients who developed a de novo psychotic disorder following a right anterior temporal lobe resection: 1 developed a delusional depression and 4 a schizophrenic-like psychosis, and 1 patient was diagnosed with Capgras’ syndrome.Reference Mace and Trimble 63 Calvet et al Reference Calvet, Caravotta and Scévola 64 described 3 patients with refractory temporal lobe epilepsy who underwent temporal anterior lobectomy and, within the first year after surgery, developed de novo psychotic symptoms associated with complete remission of seizures. Nonetheless, the relationship between side of seizure focus and seizure freedom and the risk of developing postsurgical psychosis cannot be decided upon on the bases of these small case series.
Other authors have associated the presence of gangliogliomas or DNET with the development of de novo postsurgical psychotic disorders. Andermann et al Reference Andermann, Savard, Meencke, McLachlan, Moshe and Andermann 65 reported 6 such patients. Such association remains to be established in larger studies, however.
Postsurgical psychiatric complications of anterior temporal lobe resection can present as manic episodes. For example, Carran et al Reference Carran, Kohler, O’Connor, Bilker and Sperling 66 identified 16 patients who developed a de novo manic episode from a case series of 415 consecutive patients (corresponding to a prevalence of 3.8%) who had undergone anterior temporal lobe resection. These patients were compared to a control group of asymptomatic patients matched for age and gender and a second group of 30 patients who experienced a postsurgical depression. Manic episodes occurred within the first postoperative year and were short-lived in all but 1 patient. Compared to the 2 control groups, patients with postsurgical mania were more likely to display bilateral EEG abnormalities and to have a right temporal seizure focus, although this difference did not reach significance when compared to the depressed group. Symptomatic postsurgical patients were more likely to have experienced generalized tonic-clonic seizures before surgery and to have had continuing seizures postoperatively.
Postictal psychotic episodes can also occur de novo after anterior temporal lobe resections. For example, Christodoulou et al Reference Christodoulou, Koutroumanidis, Hennessy, Elwes, Polkey and Toone 67 reported 3 cases (1%) among 282 consecutive patients who had undergone an anterior temporal lobe resection. All 3 patients had seizures predominantly from the contralateral (nonsurgical) site or had bilateral independent seizures, while none of the patients who failed surgery and continued to have seizures from the site of the surgery developed de novo postictal psychosis. This supports the conclusion that patients with PIPE (chronic or de novo) have bilateral independent temporal lobe dysfunction. Manchanda et al Reference Manchanda, Miller and McLachlan 68 identified 4 patients (1.3%) who developed a de novo PIPE among a group of 298 consecutive patients who had undergone an anterior temporal lobe resection. All 4 patients had a right-sided resection and had no pre-operative psychiatric history.
A history of psychosis should not be considered a contraindication to epilepsy surgery, provided that the patient can cooperate during the presurgical evaluation and has a clear understanding of the nature of the surgical procedure, the potential risks, and probability of seizure freedom. In support of this view is the article by Reutens et al Reference Reutens, Savard, Andermann, Dubeau and Olivier 69 on 5 patients with medically intractable epilepsy and chronic psychosis who had temporal lobe resection for epilepsy. Postoperatively, all patients were seizure-free. Surgery had no apparent effect on the course or severity of the patients’ psychoses. However, seizure freedom allowed 2 of the 5 patients to return to work or a supervised workshop. Da Conceição et al Reference Da Conceição, De Araujo Filho, Mazetto, Alonso and Yacubian 70 followed 18 patients with a history of refractory TLE secondary to mesial temporal lobe sclerosis and POE. They underwent cortico-amygdalohippocampectomy and displayed a complete remission of the psychotic symptoms 12 months after surgery. Patients with epilepsy have a lower quality of life and higher rates of comorbidities in comparison to the general population; therefore, epilepsy surgery should be considered even in the presence of persisting psychosis.Reference Mohammed, Kaufman and Limbrick 71
Although it is often difficult to predict which patients will develop psychiatric complications after epilepsy surgery, several studies have suggested potential risk factors for the development of psychosis. Among these are surgery over the age of 30 years, family history of psychosis,Reference Glosser, Zwil, Glosser, O’Connor and Sperling 72 and, as stated above, the presence of a ganglioglioma or DNET.Reference Andermann, Savard, Meencke, McLachlan, Moshe and Andermann 65 In addition, several studies have examined the issue of postoperative psychosis and the laterality of surgery. Most studies report psychosis to be more common in patients following right temporal lobectomy.Reference Leinonen, Tuunainen and Lepola 62 , Reference Andermann, Savard, Meencke, McLachlan, Moshe and Andermann 65 , Reference Manchanda, Miller and McLachlan 68 The explanation for this lateralizing correlation is yet to be established.
Etiology, Pathogenesis, and Pathophysiology
The etiology of POE has not been elucidated yet. Psychotic symptoms have been associated with focal epilepsy of frontal and mesial temporal origin.Reference Wieser, Hailemariam, Regard and Landis 47 , Reference Bertram 73 , Reference Kandratavicius, Lopes-Aguiar, Bueno-Júnior, Romcy-Pereira, Hallak and Leite 74 However, Kanemoto et al Reference Kanemoto, Tadokoro and Oshima 75 have opposed a causal relationship between psychosis and epilepsy. First, they indicate that epileptic psychosis is not an independent identity, and therefore epilepsy is no more than a nonspecific precipitating factor that leads to psychosis in patients who have a genetic predisposition for this disorder. Their second argument is that the presence of the 2 disorders is only coincidental.Reference Kanemoto, Tadokoro and Oshima 75 This last argument is becoming less improbable, however, due to the increasing evidence of the high risk of developing schizophrenia or schizophrenia-like psychosis in individuals with a history of epilepsy.Reference Bredkjaer, Mortensen and Parnas 76 , Reference Qin, Xu, Laursen, Vestergaard and Mortensen 77
Very few studies in the literature have focused on the pathogenesis of POE.Reference Kandratavicius, Lopes-Aguiar, Bueno-Júnior, Romcy-Pereira, Hallak and Leite 74 , Reference Butler, Weisholtz and Isenberg 78 Whether these same abnormalities have a pathogenic role in the development of POE remains to be established.
Focal brain lesions have been associated with the development of both epilepsy and psychosis, particularly if the lesion involves limbic structures.Reference Serafetinides and Falconer 79 For example, gangliogliomas, which have a predilection for the temporal lobes,Reference Roberts, Done, Bruton and Crow 80 have been associated with the development of postsurgical psychotic episodes. Another structural abnormality associated with psychotic-like symptoms in patients with epilepsy is sclerosis of the limbic structures.Reference Roberts, Done, Bruton and Crow 80
According to Bear,Reference Bear 81 psychopathological states in TLE would derive from intrinsic mechanisms leading to anatomical sensory limbic hyper-connectivity. The psychiatric manifestation would be the result of progressive changes in limbic structures secondary to seizure activity which would be the underlying physiopathology.Reference Bear 81 The extra-temporal commitment would be due to the spread of seizures by critical subcortical synchronizers, possibly represented by thalamo-hippocampal circuits.Reference Bertram 73 Thalamic nuclei project to the nucleus accumbens and are interconnected with the prefrontal cortex, forming a behaviorally flexible circuit whose dysfunction contributes to the pathophysiology of epilepsy-related psychosis and schizophrenia.Reference Floresco, Zhang and Enomoto 82
Neuroimaging studies of epilepsy-associated psychosis focus on interictal and postictal states, which have shown bilateral or left-sided loss of cerebral volume in the temporal lobes of patients with IPE as compared to patients with TLE without psychosis.Reference Butler, Weisholtz and Isenberg 78 , Reference Maier, Mellers, Toone, Trimble and Ron 83 Functional neuroimaging studies of IPE using single photon emission computed tomography (SPECT) or positron emission tomography (PET) also point toward left greater than right temporalReference Marshall, Syed, Fenwick and Lishman 84 , Reference Mellers, Adachi, Takei, Cluckie, Toone and Lishman 85 or more diffuse (frontal, temporal, basal ganglia) dysfunction,Reference Gallhofer, Trimble, Frackowiak, Gibbs and Jones 86 though interpretation is limited by the very small numbers of patients. Careful manual tracing of MRI images, with differentiation of amygdala and hippocampus, showed that patients with IPE, as compared to patients with epilepsy without psychosis, have larger bilateral amygdalae.Reference Tebartz van Elst, Baeumer and Lemieux 87 This finding is perhaps counterintuitive (further discussed below), since most structural neuroimaging studies of IPE patients, as well as of TLE patients in general, show volume loss in this region.Reference Bernasconi, Bernasconi, Caramanos, Antel, Andermann and Arnold 88 , Reference Cendes, Andermann and Gloor 89
Neuroimaging studies of PIP are also rare. One functional neuroimaging study using PET has shown that patients with PIP have regional hyperperfusion in the temporal lobe ipsilateral to the seizure focus,Reference Fong, Fong and Mak 90 while others have shown hyperperfusion of bilateral frontal and temporal regions during PIP.Reference Fong, Ho, Tsoi, Fong and Ho 91 – Reference Nishida, Kudo and Inoue 93
During periods of active psychosis, patients without epilepsy showed dramatic spike and slow wave activity in the septal region, as well as less prominent, perhaps propagated, epileptic activity in medial temporal structures (hippocampi and amygdalae). In contrast, actively psychotic patients who also had epilepsy (but who were not experiencing overt seizures at the time of the recording) showed a reverse pattern, with prominent spike and slow wave activity in medial temporal regions, and less prominent septal epileptic activity. These findings suggest that psychosis may be associated with epileptic activity in deep, interconnected limbic brain regions in patients both with and without overt epilepsy.
Management
Pharmacologic treatment of POE
The treatment of all forms of POE, with the exception of ictal psychotic episodes, requires the use of antipsychotic drugs (APD). The treatment of PIPE and alternative psychosis has already been discussed above and will not be reviewed in this section. In patients with IPE, optimal seizure control is desirable, as episodic psychotic exacerbations can accompany clusters of seizures.
The proconvulsant properties of conventional antipsychotic drugs (CAPD) have been recognized for a long time and have often led to some clinicians’ reluctance to use these drugs in patients with epilepsy for fear of worsening seizure frequency and severity. Among the CAPD, the incidence rates of seizures in non-epileptic patients has ranged between 0.5% and 1.2%.Reference Logothetis 94 , Reference Whitworth and Fleischhacker 95 The risk is higher with certain drugs and in the presence of the following factors: (1) a history of epilepsy, (2) abnormal EEG recordings, (3) history of CNS disorder, (4) rapid titration of the CAPD dose, (5) high doses of AP, and (6) the presence of other drugs that lower the seizure threshold. Chloprenazine, fluphenazine, perphenazine, and trifluoperazine are CAPD that can cause seizures in a dose-related manner. Haloperidol and molindone are among the CAPD with a lower seizure risk.Reference Whitworth and Fleischhacker 95
In a recent study, Alper et al Reference Alper, Schwartz, Kolts and Khan 96 compared the incidence of seizures between (non-epileptic) patients randomized to an atypical antipsychotic drug (AAPD) or placebo in the course of regulatory studies submitted to the FDA between 1985 and 2004. The incidence of seizures was higher among those randomized to clozapine and olanzapine, and to a significant lesser degree with quetiapine, but there was no difference in seizure occurrence between patients randomized to placebo and 4 other AAPDs. In non-epilepsy patients, clozapine has been reported to cause seizures in 4.4% when used at doses above 600 mg/day, while at a doses lower than 300 mg, the incidence of seizures has been found to be less than 1%.Reference Stahl 97 On the other hand, in patients with epilepsy, clozapine can increase the seizure frequency at any dose.Reference Zaccara, Messori and Cincotta 98
Among the CAPD, chlorpromazine and loxepine have the highest risk of seizure occurrence. Those with a lower seizure risk include haloperidol, molindone, fluphenazine, perfenazine, and trifluoperazine. Whether the presence of AEDs at adequate levels protects patients with epilepsy from breakthrough seizures upon the introduction of APD with pro-convulsant properties is yet to be established.
Most APD can cause EEG changes consisting of slowing of background activity, above all when used at high doses. In addition some of these drugs, and particularly clozapine, can cause interictal sharp waves and spikes; this epileptiform activity, however, is not predictive of seizure occurrence.Reference Zaccara, Messori and Cincotta 98 There are data suggesting that a severe disorganization of the EEG recordings is a more likely predictor of seizure occurrence. As a rule, any APD should be started at low doses and should undergo slow dose increments to minimize the risk of seizures in patients with epilepsy.
Pharmacokinetic interaction between AEDs and APD
Induction of hepatic enzymes upon the introduction of AEDs with enzyme-inducing properties (such as phenobarbital, primidone, carbamazepine, and phenytoin) may result in an increase of the clearance of most APD. This is, in fact, the most frequent and clinically relevant pharmacokinetic interaction encountered in clinical practice. It may potentially result in recurrence of psychotic symptoms previously controlled at higher serum concentrations of APD. By the same token, discontinuation of an AED with enzyme-inducing properties may result in a decrease in the clearance of APD, which in-turn can lead to toxicity, such as extrapyramidal adverse events caused by an increase of the serum concentrations of CAPD. Finally, certain AEDs, like valproic acid, can inhibit the glucuronidation process of AAPDs like clozapine.
Pharmacodynamic interactions
Worsening of CNS-related adverse events including sedation, ataxic gait, and cognitive disturbances are the most frequent consequences of the pharmacodynamic interactions between APD and AEDs with sedative properties.Reference Zaccara, Messori and Cincotta 98
Clozapine and carbamazepine can cause leucopenia on their own, and their combination can worsen the severity of the leucopenia.Reference Zaccara, Messori and Cincotta 98 In addition, the combination of these two drugs has been also reported to cause an increase risk of neuroleptic malignant syndrome. Thus, this combination should be avoided. In addition, all 6 AAPDs can cause weight gain and type-II diabetes mellitus, with olanzapine having the greatest impact on weight gain and aripiprazole the least. Thus, glycemias and lipid profiles need to be monitored in all patients on a regular basis, but particularly in patients taking AEDs that are associated with an increased risk of weight gain, such as valproic acid, gabapentin, pregabalin, and carbamazepine.
Electroshock therapy (ECT)
Electroconvulsive therapy is not contraindicated in patients with severe POE and should be considered in patients with epilepsy with psychotic depressive or manic episodes that fail to respond to APD.Reference Abrams 99 – Reference Viparelli and Viparelli 105 Blackwood et al Reference Blackwood, Cull, Freeman, Evans and Mawdsley 100 found that the incidence of seizures in patients following treatment with ECT was no higher than in the general population. In fact, several studies have shown that ECT increases the seizure threshold by 50–100%.Reference Fink, Kellner and Sackeim 102 – Reference Sackeim 104 Krystal and Coffey have suggested the use of unilateral ECT in the non-dominant frontotemporal region (101). AEDs should be held in the mornings before a treatment, but otherwise should be kept at baseline doses.
Concluding Remarks
Psychosis of epilepsy is a multifaceted neuropsychiatric disorder that reflects the complex relations between epilepsy and psychotic disorders. Their pathogenic and pathophysiologic mechanisms remain unknown, but the existence of common pathogenic mechanisms operant in the two conditions may explain their bidirectional relation and higher comorbid occurrence. Several contradictory observations remain without an explanation, the most obvious being the existence of postictal psychotic symptoms following clusters of secondarily GTC seizures, while ECT increases the seizure threshold and is used to treat postictal catatonia and other types of psychotic disorders.
Pharmacologic treatment of POE remains a source of frequent misconception among psychiatrists and neurologists who are reluctant to use adequate doses to yield a remission of psychotic symptoms. The availability of several second-generation APD that do not appear to lower the seizure threshold at therapeutic doses should alleviate those concerns.
Since there is such a bidirectional relation between epilepsy and psychotic disorders, it remains, perhaps even from the time of Hippocrates, an essential neuropsychiatric disorder.
Disclosures
Andres Kanner and Ana Maria Rivas do not have anything to disclose.
