Risk factors for behavioral and psychotic dysregulation at the epilepsy monitoring unit in patients with intracranial electrodes

Objective: Aberrant behavior in patients with epilepsy (PWE) admitted to an epilepsy monitoring unit (EMU) can endanger their safety. We sought to identify predictive factors for post-ictal behavioral dysregulation and psychosis in patients with refractory epilepsy being monitored at an EMU. Methods: Retrospective data were gathered from electronic patient ﬁles of all patients with refractory epilepsy who underwent intracranial registration at our EMU. We assessed behavioral and psychotic dys-regulations by reviewing clinical notes, administered emergency medication, and reports of injuries or casualties in patients and nurses. In addition, we compared patient demographic characteristics, clinical characteristics, and antiepileptic drug (AED) proﬁles between patients with and without behavioral and/ or psychotic dysregulation. Results: Out of 73 admissions, 23 patients (32%) experienced behavioral dysregulation, and ﬁve patients experienced psychosis (7%). Behavioral dysregulation was only signiﬁcantly associated with a previous history of interictal or postictal psychosis. Psychotic dysregulation is signiﬁcantly associated with a psychiatric history, including a history of agitation or psychosis, whether or not epilepsy-related. For both types of dysregulations, there was no relation with a pre-admission frequency of seizures, clustering of seizures during monitoring, or a temporal focus of seizures. We could not report a relationship between AED use, tapering, and the occurrence of dysregulation. Conclusion: We conclude that a psychiatric history, including a history of agitation and psychosis, is related to an increased risk of behavioral and psychotic dysregulation in patients undergoing invasive seizure monitoring at the EMU. (cid:1) 2023 The Authors. Published by Elsevier Inc. ThisisanopenaccessarticleundertheCCBYlicense(http:// creativecommons.org/licenses/by/4.0/).


Introduction
An epilepsy monitoring unit (EMU) puts special demands on patient safety, especially because the patients themselves can display unsafe behavior.Patients are admitted to an EMU for, e.g., invasive diagnostic procedures in refractory epilepsy to identify the epileptogenic zone using intracranial electrodes.To provoke seizures (within the limited time frame), antiepileptic drugs (AED) are often tapered.This puts patients at risk for recurrent seizures, which can result in behavioral or psychotic dysregulation.Psychosis in this context is defined as the presence of visual, auditory, or other hallucinations or delusions, which may be accompanied by mood disturbances such as fear, agitation, or aggression.Psychosis in relation to epilepsy is commonly categorized according to the temporal relation of the psychotic episode to the seizure.Given the fact that epileptic seizures are usually brief, ictal psychosis is rarely seen, except in the case of nonconvulsive status epilepticus.Postictal psychosis commonly occurs after a lucid interval of hours after the seizure.Interictal psychosis presents during seizure-free intervals in patients with long-term, uncontrollable seizures.It differs from primary psychiatric psychosis, which is unrelated to seizures, by its usually less severe episodes and more benign course.
To prevent psychosis and behavioral dysregulation in PWE, seizure control is important [5].However, as the registration of several epileptic seizures is the main clinical goal in the EMU, this is not feasible.Therefore, it is necessary to correctly anticipate risk factors for psychotic and behavioral dysregulation [10], although specific risk factors for dysregulation in the EMU are unknown.
In our current study, we assessed possible risk factors for psychotic and behavioral dysregulation, with special attention to the patient's history, clinical aspects of epilepsy, and AED tapering during their EMU stay.

Design and patients
In our retrospective observational cohort study, we included all patients with an EMU admission between December 2007 and January 2021 at the Maastricht University Medical Center (MUMC+).The METC (Medical Ethics Committee of the MUMC) gave consent for the study (2019-1007-A-10), which had no obligations to the Dutch Law for Medical Research (WMO).
All patients had refractory epilepsy and a registration with intracranial electrodes.We excluded two patients younger than 16 years and nine patients with incomplete clinical data.Some patients were admitted more than once; we included every admission separately in our study, resulting in 67 patients with a total of 69 EMU admissions.

Primary outcome
We defined the primary outcome measures 'behavioral dysregulation' and 'psychotic dysregulation' as follows: any behavioral dysregulation during the EMU admission, including behavior for which intervention with medication (benzodiazepines or antipsychotics, except midazolam nasal spray) was necessary during the episode, and the behavior was dangerous (such as aggression) for the patient or the EMU staff.We only included behavior that clearly differed from a person's normal behavior as behavioral dysregulation.We excluded behavior during an epileptic seizure.For psychotic dysregulation, this diagnosis was made by the attending psychiatrist using the diagnostic criteria of the Diagnostic and Statistical Manual (DSM) IV (prior to 2014) or DSM 5 (since 2014) for 'psychotic disorder due to another medical condition' (293.8x) for seizure-related psychosis, or any other psychotic disorder, for psychotic symptoms assumed unrelated to seizures.For the diagnosis of aggressive dysregulation, no specific diagnostic criteria were followed.

Clinical variables
We gathered the abovementioned outcome data and the following clinical data from the electronic patient files: age at EMU admission, sex, age at onset of epilepsy, seizures with their respective focus (with special attention to temporal seizure foci), previous clustering of seizures (3 seizures within 24 h), average frequency of seizures (categorized in >1 per month, >1 per week), previous psychiatric history, including both primary psychiatric diagnoses as well as a history of seizure-related psychoses, as reported in the electronic patient file.Furthermore, previous post-ictal behavior, especially agitation (restlessness or agitation persistent >2.5 h after a seizure, not defined as psychosis), and psychosis (previous ictal, interictal, or postictal psychosis), as well as the use of antipsychotics at the start of EMU admission, were recorded.To quantify the AED load, we used the defined daily dose (DDD) [1].We gathered the following information regarding AED use during EMU admission for each participant: Number of AEDs tapered (0, >1, >2, >3, >4), the average pace of tapering (total amount of DDD tapered divided by the number of days taken to taper; DDDD per days), total amount of DDD tapered (sum of DDD), and the type of AED which was tapered enabled us to separately analyze anticonvulsants with mood stabilizing effects and other effects on mood (carbamazepine, lamotrigine, valproate, levetiracetam, oxcarbazepine, topiramate), and benzodiazepines (clobazam, clonazepam) or the combination of these AED.

Statistical analysis
We compared frequencies between patients with and without dysregulation using parametric and nonparametric tests (whichever was appropriate).Chi-square tests and Fisher's exact test were used for categorical data, and an independent sample t-test was used, which was appropriate for continuous data.For variables with a significant difference between groups, we performed a logistic regression to test for an independent relationship (with correction for possible confounders).Lastly, we performed a post hoc power analysis with G*Power software.We considered pvalues < 0.05 significant.Analysis was conducted using of IBM SPSS Statistics version 27.

Patient demographic characteristics and clinical features
We assessed a total of 69 EMU admissions with patients who had intracranial electrodes (grids and strips, sometimes combined with single depth electrodes (in the hippocampus), no S-EEG implantations).In 23 EMU admissions (33%), there was behavioral dysregulation, appearing e.g., more or less severe agitation and/or aggression.In all patients, this occurred directly after a seizure and remitted within one day.In five EMU admissions (7%), a frank psychotic dysregulation occurred (see below for more clinical details).During all episodes, there were no epileptiform EEG abnormalities and no indication of a non-convulsive status epilepticus.Demographic and clinical variables of patients with behavioral dysregulation (postictal) and psychotic dysregulation and those without dysregulations are shown in Table 1.We found a significantly higher frequency of a previous history of psychosis in patients with behavioral dysregulation than in those without.For patients with psychotic dysregulation, frequencies of female sex, psychiatric history, history of agitation, history of non-epilepsy-related psychosis, and of epilepsy-related psychosis were higher than in those without psychotic dysregulation.In a tentative logistic regression (because of the low event rate), we found that psychiatric history (OR 10.2; 95% CI [1.527-171.851];p = 0.021), history of agitation (OR 16.2; 95% CI [1.713-196.190];p = 0.016), history of psychosis (OR 15.0; 95% CI [1.652-136.172];p = 0.016), and history of psychosis related to epilepsy (OR 63.0; 95%-CI [1.652-136.172];p = 0.004) were significantly related to a (later) psychotic dysregulation during EMU admission.

Patients with psychotic dysregulation
Of the patients with a psychotic dysregulation, two had a formal preoperative psychiatric assessment, which reported no psychiatric abnormalities at that time.All patients who had seizures had their seizures during their stay at the EMU, however, the exact time relation between the seizures and the first symptoms of the psychosis was not well recorded.Our clinical impression is that this usually occurs within a few hours after the seizure.We used the EEG to exclude (ongoing) electroencephalographic seizure activity.Psychotic dysregulation presented with disorganized thought (in five patients), hallucinations (auditory, in three patients), aggressive behavior (in two patients), delusions (in one patient), and social withdrawal (in one patient).The duration of the psychotic dysregulation ranged from 2 to 13 days.

(Tapering of) antiepileptic drugs
Our findings regarding AED use and tapering are shown in Table 2.We could not find any relation between AED use or AED tapering and the occurrence of behavioral dysregulation or psychotic dysregulation.

Discussion
We found that a psychiatric history, especially a previous history of agitation and epilepsy-related as well as non-epilepsyrelated psychosis, is related to psychotic dysregulation during an EMU admission.Other clinical epileptological factors, such as epilepsy type or clustering of seizures during the EMU registration, did not relate to the dysregulation.Moreover, there was no relation between the speed of AED tapering, or the tapering of anxiolytic or mood-stabilizing AEDs, and behavioral or psychotic dysregulation.
In our EMU cohort, we found similar incidences of psychotic dysregulation (7%) and behavioral dysregulation (33%), as previously described [2][3][4][5][6], though not all previous studies were performed in the same setting as our study.Furthermore, the incidence of behavioral dysregulation was higher (up to 52%) in another study, which probably relates to the prospective character with inclusion of self-reported behavioral changes by the patients in that study [7].We did not use self-reported signs and symptoms in our study, which might have led to a lower incidence of behavioral disturbance.
We set out to find clinical markers for psychotic or behavioral dysregulation.We found that a psychiatric history (also independent from seizures) and behavior surrounding previous seizures (like agitation) were related to psychotic dysregulation in our EMU.Others also found that prior psychotic illness was related to dysregulation [4].As such, our findings confirm but also extend these findings, as we also found that other psychiatric illnesses are related to psychotic dysregulation.This finding can help EMU teams identify patients at risk.PWE at the EMU mostly have refractory epilepsy and are therefore already susceptible to suffering from psychiatric illnesses [8,9].A thorough screening of the history of psychiatric disease and a behavioral history of periictal events will identify patients at risk for psychotic dysregulation during their EMU admission.This will help to increase awareness, hopefully leading to an early identification of psychotic dysregulation and possible early treatment.This is also true for the behavioral dysregulation with postictal aggression, which mainly relates to previous postictal psychosis [10].
AED tapering in itself may also cause psychotic and other psychiatric symptoms [11], also because some AED have moodstabilizing and sedating effects.However, studies in different settings (AED tapering in general or AED tapering at the EMU) do not report these problems as major [12][13][14].In our study, we confirmed that AED tapering (with all its aspects, like pace, type of AEDs tapered, or number of AEDs tapered) did not relate to the occurrence of behavioral or psychotic dysregulation.Despite our small patient group experiencing dysregulation, we only find a minimal difference in the pace of tapering between those with behavioral dysregulation and those without, and the group of patients with psychotic dysregulation has a slower tapering rate than the group without psychotic dysregulation.Given the inversed relationship with only a small, non-significant difference, we did not consider this effect clinically relevant.Therefore, there is no need for caution in AED tapering in the EMU, and no extra precautionary measures have to be considered when tapering specific AED.Other possibly relevant factors, which we did not analyze in our study, are sleep deprivation, stress throughout the admission, and the lack of orientation due to prolonged admission without any considerable events.These factors could also influence aberrant behavior [15,16].However assessment of these factors in an EMU setting requires further, ideally prospective, research.Our study has a number of limitations.First, our study has a retrospective design.We had to rely on data that was recorded in the patient's notes, which might have led to an underreport of behavioral disturbances, especially if these were less severe and not disruptive for patients or staff.Also, patient histories regarding the specific preceding psychiatric diagnoses (i.e., postictal or periictal psychosis) were mostly very succinct, precluding more detailed analyses regarding this point.Second, we could not consider sleep deprivation as a possible contributing factor, as the reports in medical files were not always consistent enough.Third, the number of patients with dysregulation is rather low, leading to fairly large confidence intervals and perhaps underestimation of some of the associations.Because the number of events was too small, the logistic regression is only tentative, and its results have to be interpreted cautiously.
Nevertheless, the strength of our study is the consecutive recruitment of patients admitted to the EMU in a fixed timeframe without selection, which results in a representative sample and, therefore, a high external validity.Also, by defining behavioral dysregulation, in addition to psychotic dysregulation, as a clinically significant event, we included all relevant episodes for patient safety in the EMU.This maximizes the possible usefulness of this study for clinical applications.

Conclusions
We conclude that a psychiatric history, including a history of agitation and psychosis, whether or not related to epilepsy, relates to behavioral and/or psychotic dysregulation in patients undergoing invasive seizure monitoring at the EMU, whereas AED tapering and withdrawal do not significantly contribute to the occurrence of dysregulation.This may help identify patients at risk and can be helpful to increase safety and prevent further complications due to dysregulation.

Table 1
Demographic and clinical features -control versus behavioral dysregulation and Control versus psychotic dysregulation.
* Missing values for one patient.