- Email: [email protected]
Safety in the epilepsy monitoring unit: A retrospective study of 524 consecutive admissions
Epilepsy & Behavior 61 (2016) 162–167
Contents lists available at ScienceDirect
Epilepsy & Behavior journal homepage: www.elsevier.com/locate/yebeh
Safety in the epilepsy monitoring unit: A retrospective study of 524 consecutive admissions Firas Fahoum a,⁎, Nurit Omer a, Svetlana Kipervasser a,b, Tal Bar-Adon a, Miri Neufeld a,b a b
Epilepsy and EEG Unit, Department of Neurology, Tel Aviv Sourasky Medical Center, Tel Aviv 6423906, Israel Sackler School of Medicine, Tel Aviv University, Tel Aviv 69978, Israel
a r t i c l e
i n f o
Article history: Received 18 February 2016 Revised 3 May 2016 Accepted 2 June 2016 Available online xxxx Keywords: Epilepsy monitoring unit Safety adverse events Antiepileptic drugs Seizure clusters Psychogenic nonepileptic seizure Status epilepticus
a b s t r a c t The yield of monitoring patients at an epilepsy monitoring unit (EMU) depends on the recording of paroxysmal events in a timely fashion, however, increasing the risk of safety adverse events (AEs). We aimed to retrospectively study the frequency and risk factors for AE occurrences in all consecutive admissions to an adult EMU in a tertiary medical center. We also compared our findings with published data from other centers. Between January 2011 and June 2014, there were 524 consecutive admissions to the adult EMU at the Tel Aviv Sourasky Medical Center. Adverse events were recorded in 47 (9.0%) admissions. The most common AE was 4-hour seizure cluster (58.7% of AEs) and, in decreasing frequency, AEs related to antiepileptic drugs (AEDs, 11.1%), falls and traumatic injuries (9.5%), intravenous line complications (9.5%), electrode-related (4.8%), status epilepticus (SE, 3.2%), and cardiac (1.6%) and psychiatric (1.6%) complications. There were significantly more AEs among patients with a younger age at disease onset (p = 0.005), a history of temporal lobe epilepsy (p = 0.046), a history of focal seizures with altered consciousness (p = 0.008), a history of SE (p = 0.022), use of a vagal nerve stimulator (p = 0.039), and intellectual disability (p = 0.016) and when the indication for EMU monitoring was noninvasive or invasive presurgical evaluation (p = 0.001). Adverse events occurred more frequently when patients had more events in the EMU (p = 0.001) and among those administered carbamazepine (p = 0.037), levetiracetam (p = 0.004), clobazam (p = 0.008), and sulthiame (p = 0.016). Patients with a history of psychogenic nonepileptic seizures (PNESs) had significantly fewer AEs (p = 0.013). Adverse events were not associated with the age, gender, duration of hospitalization or monitoring, AED withdrawal and renewal, seizure frequency by history, presence of major psychiatric comorbidities, abnormal neurological exam, or the presence of a lesion as on brain magnetic resonance imaging. In conclusion, this study reveals that AEs are not unusual in the EMU and that seizure clustering is the most common among them. Adverse events occur more frequently in patients with more severe epilepsy and intellectual disability and in patients undergoing presurgical evaluations and less frequently in patients with PNESs. © 2016 Elsevier Inc. All rights reserved.
1. Introduction Epilepsy monitoring units (EMUs) use video-electroencephalography (video-EEG) recordings for several indications, such as distinguishing epileptic seizures from psychogenic nonepileptic seizures (PNESs) and other nonepileptic paroxysmal events, evaluating patients for epilepsy surgery, characterizing seizure types, and adjusting antiepileptic drug (AED) treatment. The yield of the monitoring is highly dependent on the recording of epileptic and/or nonepileptic events. To increase the likelihood of capturing events in a timely fashion, it is a standard practice to use activating procedures, such as AED withdrawal and sleep deprivation.
⁎ Corresponding author at: Epilepsy and EEG Unit, Department of Neurology, Tel Aviv Sourasky Medical Center, 6 Weizmann St., Tel Aviv 6423906, Israel. E-mail address: fi[email protected] (F. Fahoum).
http://dx.doi.org/10.1016/j.yebeh.2016.06.002 1525-5050/© 2016 Elsevier Inc. All rights reserved.
These measures may provoke patients to experience an increase in the frequency and severity of seizures, therefore increasing the risks of injuries and medical complications. Surveys of EMUs in the United States and Europe demonstrated that the most frequent safety adverse events (AEs) were falls, status epilepticus (SE), and postictal psychosis [1,2]. According to one study, major AEs appeared only during generalized tonic–clonic seizures (GTCSs), and the risk factors for AEs were epilepsy duration, existence of psychiatric comorbidity, and history of SE [1]. Seizure clusters (SCs) are other common AEs [2,3]. Studies on interventions to reduce the frequency of AEs highlight the importance of adherence to stricter safety measures [4], decreasing the number of missed seizures [5], using AED withdrawal protocols [6–8], and providing medical interventions in cases of SE or SC [2,6]. Whereas PNESs are frequently encountered in the EMU, in up to 32% patients in tertiary epilepsy centers [7], many studies on AEs in EMUs excluded those patients and included only patients with epilepsy
F. Fahoum et al. / Epilepsy & Behavior 61 (2016) 162–167
163
[2,3,8,9]. In this study, we analyzed all consecutive admissions to an EMU in a tertiary center and examined the frequency and risk factors for the occurrences of AEs. We also compared our findings with published data from other centers.
convulsive seizures lasting longer than 5 min or when consciousness was not regained between two consecutive seizures [11]. The study was approved by the local research ethics committee at TASMC.
2. Methods
2.2. Technical details and work routines at EMU
2.1. Patients
The EMU is located in physical proximity to the Department of Neurology. During the study period, it had 3 adult monitoring beds with a digital 32-/64-/128-channel video-EEG system (NicOne LTM system, Viasys, Madison, WI, USA). Surface EEG electrodes (Ag/AgCl) were placed according to the international 10–20 system, with additional inferior temporal or closely spaced electrodes when needed. Subdural strips and grids and depth electrodes (Ad-Tech Medical, Racine, WI, USA) were used for invasive EEG recordings. Two electrocardiography (ECG) electrodes were applied in all patients. The EMU staff included 2 EEG technicians from 08:00 to 16:00 on weekdays (Sunday–Thursday) and one nurse who provided 24/7 coverage. A specialized epileptologist was in the EMU during the daytime and on call 24/7 (FF, SK, MN). There was full continuous access to the medical staff of the Department of Neurology. The patients were hospitalized in single rooms with private bathrooms and were monitored by the nurse via a monitoring screen located at the nurse's station. One companion received instructions from the EMU nurse and was encouraged to stay overnight in the patient's room. Thick padded side rails in the up position were used to minimize the risk of seizure-related injuries. In order to ensure intravenous (IV) access in case of emergency, IV catheters were placed in the antecubital fossa of patients who underwent AED withdrawal, as well as those with a high seizure frequency or who had been implanted with intracranial electrodes. Intravenous catheters were flushed twice daily with IV heparin 100 i.u. Antiepileptic drug withdrawal followed the following outline: it was avoided during the first day of admission and, if required, started on the second day or later, taking into account the reported frequency of events prior to admission, the number and dosages of AEDs, and any history of SE. Starting from the second day of admission, the dose of a
This retrospective study included all consecutive patients older than 18 years who had been admitted to the EMU at the Tel Aviv Sourasky Medical Center (TASMC) between January 1, 2011 and June 30, 2014. This period was chosen because fully computerized and standardized medical files had become operable in the EMU. Prior to monitoring, all patients had signed informed consent forms that explained the purpose of monitoring, the possibility of AED withdrawal when required, and its potential risks, as well as the safety measures to ensure patient safety. Medical files were reviewed and the following data documented: patient's age, gender, age of onset of epileptic or paroxysmal events, classification of paroxysmal syndrome (focal temporal epilepsy, focal extratemporal epilepsy, generalized epilepsy, PNES), seizure types according to ILAE classification [10], reported frequency of events prior to EMU admission, history of SE (Yes/No), presence of prediagnosed intellectual disability (Yes/No), presence of major psychiatric comorbidities (major mood disorder, schizophrenia, personality disorder, severe behavioral disturbances, history of suicide attempt, and substance abuse) (Yes/No), use of vagal nerve stimulation (VNS) (Yes/No), AED type, presence of abnormal neurological examination (Yes/No), and presence of a lesion on brain magnetic resonance imaging (MRI) (Yes/No). We also documented the indication for EMU admission, duration of hospitalization, duration of EEG monitoring, number of paroxysmal events during monitoring, first day of AED withdrawal and first day of AED renewal since admission, first day of paroxysmal events, number of AEs, and dates of occurrence and types of AEs as listed in Table 1. An SC was defined as 3 focal seizures with impaired consciousness or 2 GTCSs occurring within 4 h [3]. Status epilepticus was defined as
Table 1 Characterization of safety adverse events (AEs) in the AED withdrawal and in the no-withdrawal groups. AEs (in decreasing order of frequency)
Total number of AEs
AEs in withdrawal group
AEs in no-withdrawal group
AE characterization
Seizure clustersa
37
18
19
AED-related
7
2
5
Falls/traumatic injuries
6
4
2
IV line-related
6
4
2
Electrode-related
3
1
2
2 1 1 0 0 0 63
1 0 1 0 0 0 31
1 1 0 0 0 0 32
33 focal seizures with impaired consciousness 3 GTCSs 1 atypical absence 3 VPA-induced hyperammonemia started in EMU 2 VPA-induced hyperammonemia taken prior to EMU admission 1 CBZ–LTG interaction 1 PB-induced increase in liver enzymes 2 atonic seizures 1 GTCS and nasal fracture outside the EMU 1 GTCS with eyebrow laceration 1 fall after PNES 1 fall and lip laceration due to extrapyramidal syndrome 3 phlebitis 2 Staphylococcus aureus bacteremia 1 extravasation of CT contrast agent 1 extra-axial hematoma following subdural grid insertion 1 extra-axial swelling following subdural grid insertion 1 scalp discomfort due to scalp EEG electrodes leading to electrodes removal 2 focal SE with impaired consciousness 1 ictal asystole 1 nonlethal suicide attempt
Status epilepticusb Cardiac Psychiatric Respiratory Thrombotic Resuscitation or death Total
AED — antiepileptic drug, GTCS — generalized tonic–clonic seizure, PNES — psychogenic nonepileptic seizure, EMU — epilepsy monitoring unit, SE — status epilepticus, VPA — valproate, CBZ — carbamazepine, LTG — lamotrigine, PB — phenobarbital, CT — computerized tomography. a Defined as 3 focal seizures with impaired consciousness or 2 GTCSs occurring within 4 h [3]. b Defined as convulsive seizures lasting longer than 5 min or when consciousness was not regained between two consecutive seizures [11].
164
F. Fahoum et al. / Epilepsy & Behavior 61 (2016) 162–167
single AED was reduced each day by one-half until fully discontinued. If no paroxysmal events occurred, the dose of the second AED was reduced by one-half each day and so on, until enough paroxysmal events were acquired to answer the question of referral. Barbiturates and benzodiazepines were not withdrawn. Antiepileptic drug withdrawal was avoided in patients with a history of SE unless it was crucial for presurgical evaluation. In addition, sleep deprivation was asked from selected patients. In case of SC after AED withdrawal, all AEDs were renewed simultaneously to the prewithdrawal dosages. Intravenous benzodiazepines were administered if SC persisted despite AED renewal or in cases where AEDs were not withdrawn. Patients with SC or high seizure frequency during monitoring were kept in the EMU for medical observation without EEG monitoring until full medical treatment had been reinstated. 2.3. Statistical analysis For univariate analysis, we used a parametric test (Student's t-test) when the variables had a normal distribution and a nonparametric test (Mann–Whitney test) for a non-normal distribution. For categorical variables, we applied the chi-square or Fisher's exact test according to the expected frequency in the cell. For multivariate analysis, we used stepwise logistic regression analysis by entering all potential risk factors that showed a significant association at a p value b0.05 in the univariate analyses. A nonparametric test (Spearman's test) was used to calculate the correlation coefficient. We repeated the analyses including only the first admission to account for potential bias towards patients that were admitted more than once to the EMU. The significance level was set at p ≤ 0.05. All statistical analyses were performed using IBM SPSS Statistics 23. 3. Results 3.1. Patients and monitoring data There were 524 admissions to the EMU during the study period. The patients' ages ranged between 18 and 82 years (mean ± standard deviation: 35.5 ± 13.9 years; 296 females and 238 males). Three patients were admitted 4 times, 4 patients were admitted 3 times, and 34 patients were admitted twice to the EMU. Indications for admission were characterization of paroxysmal events (309 admissions, 59.0%), noninvasive presurgical workup (8 admissions, 15.3%), AED adjustment (122 admissions, 23.3%), and intracranial EEG studies (13 admissions, 2.5%). The types of paroxysmal syndrome for all admissions were as follows: 155 (29.6%) patients with focal temporal epilepsy, 231 (44.1%) patients with focal extratemporal epilepsy, 32 (6.1%) patients with generalized epilepsy, and 211 (40.3%) patients with PNESs. There were 105 admissions (20%) of patients with combined epilepsy and PNESs, 87 (16.6%) patients with a history of SE, and 21 (4%) patients with VNS. In addition, 51 (9.7%) admissions involved patients with major psychiatric comorbidities, 39 (7.4%) patients with intellectual disability, and 63 (12%) of patients with abnormal findings on the neurological exam. A total of 442 patients had a brain MRI, and a lesion was demonstrated in 287 of them (64.9%). The mean duration of hospitalization at the EMU was 7.5 ± 2.5 days, and the mean duration of EEG monitoring was 7.2 ± 2.2 days. Paroxysmal events were recorded in 414 admissions (79%), and none were recorded in 110 admissions (21%).
53 (20.4%) for noninvasive presurgical workup, 54 (20.7%) for AED adjustment, and 10 (3.8%) for intracranial EEG studies. The AEDs were renewed in 214 (82%) patients in the withdrawal group and not renewed in 46 (18%). Antiepileptic drug withdrawal was started after a mean of 2.4 ± 1.5 days since admission and AED renewal after a mean of 6.2 ± 3.3 days since admission. Antiepileptic drugs were not withdrawn (no-withdrawal group) in 264 admissions (50.4%). The indications for admission in that group were as follows: 166 (62.9%) for characterization of paroxysmal events, 27 (10.2%) for noninvasive presurgical workup, 68 (25.8%) for AED adjustment, and 3 (1.1%) for intracranial EEG studies. The mean durations of hospitalization and monitoring in the withdrawal group (8.1 ± 2.2 and 7.9 ± 2.0 days, respectively) were longer than in the no-withdrawal group (6.9 ± 2.6 and 6.5 ± 2.1 days; p b 0.001 and p b 0.001, respectively). Paroxysmal events occurred in 210 of 260 (80.8%) patients in the withdrawal group and in 204 of 264 (73.7%) patients in the no-withdrawal group (p = 0.33). 3.3. Safety adverse events (AEs) Sixty-three AEs were documented in 47 EMU admissions (AE group, 9.0% of all admissions). The mean day of AE occurrence was 3.7 ± 2.1 days from admission. The distribution of the day of AE occurrence is shown in Fig. 1. Antiepileptic drug withdrawal was not associated with an increase in AEs (p = 0.37) (Table 2), and there was no significant correlation between the first day of AED withdrawal and the day of AE occurrence (rs = 0.047, p = 0.28). The most common AE was SC, which was reported 37 times (58.7% of all AEs) in 28 admissions. Seizure clusters were documented once in 22 admissions, twice in 3 admissions, and 3 times in 3 admissions. Clusters of focal seizures with altered consciousness occurred 29 times in 21 admissions, clusters of focal seizures evolving to bilateral convulsive seizures occurred in 5 admissions, and clusters of atypical absence occurred 3 times in 2 admissions. Seizure clusters occurred 18 times in the withdrawal group and 19 times in the no-withdrawal group (p = 0.9). Antiepileptic drug-related AEs occurred 7 times (11.1% of all AEs). Valproate-induced hyperammonemia occurred 3 times when it was started in the EMU and twice when patients had been treated with valproate prior to EMU admission. The interaction of carbamazepine and lamotrigine started in the EMU caused reversible nausea, vertigo, and ataxia in one case, and an increasing phenobarbital dose caused an increase in hepatocellular enzymes in another case. Falls and traumatic injuries occurred 6 times (9.5% of all AEs). All cases were outside the bed: 4 were in the monitoring room, 1 was in the bathroom, and 1 was outside the EMU. The fall accompanied an epileptic seizure in 4 cases (2 GTCSs and 2 atonic seizures), followed a PNES in one case, and was secondary to an extrapyramidal syndrome
3.2. AED withdrawal and renewal Antiepileptic drugs were withdrawn (withdrawal group) in 260 admissions (49.6%). The indications for EMU admission for that group were as follows: 143 (55%) for characterization of paroxysmal events,
Fig. 1. Histogram showing the distribution of the day of safety adverse event (AE) occurrence (n = 63) since admission to the epilepsy monitoring unit (EMU).
F. Fahoum et al. / Epilepsy & Behavior 61 (2016) 162–167
165
Table 2 Comparison of the admissions with and without safety adverse events (univariate analysis). Variable
Age Onset age Hospitalization duration Monitoring duration Number of events during monitoring Gender Indication for admission
AED withdrawal Focal temporal epilepsy Focal extratemporal epilepsy Generalized epilepsy Major psychiatric comorbidities Intellectual disability History of focal seizures without altered consciousness History of focal seizures with altered consciousness History of focal seizures evolving to bilateral convulsive seizures History of GTCS History of absence History of myoclonic seizures History of PNES History of SE Abnormal neurological exam MRI lesion VNS
Unit/category
Years Years Days Days Events Males = 227 Females = 297 Event characterization = 309 Noninvasive presurgical workup = 80 AED adjustment = 122 Invasive presurgical workup = 13 No = 264 Yes = 260 No = 369 Yes = 155 No = 293 Yes = 231 No = 492 Yes = 32 No = 473 Yes = 51 No = 485 Yes = 39 No = 239 Yes = 285 No = 171 Yes = 353 No = 228 Yes = 296 No = 489 Yes = 35 No = 498 Yes = 26 No = 496 Yes = 28 No = 313 Yes = 211 No = 437 Yes = 87 No = 461 Yes = 63 No = 155 Yes = 287 No = 502 Yes = 22
Non-AE group (N = 477)
AE group (N = 47)
Mean ± SD
Mean ± SD
35.6 ± 14.0 21.3 ± 16.8 7.5 ± 2.3 7.2 ± 2.1 9.3 ± 12.5 204 (89.9%) 273 (91.9%) 293 (94.8%) 65 (81.3%) 111 (91.0%) 8 (61.5%) 243 (92.0%) 234 (90.0%) 342 (92.7%) 135 (87.1%) 266 (90.8%) 211 (91.3%) 450 (91.5%) 27 (84.4%) 432 (91.3%) 45 (88.2%) 446 (92.0%) 31 (79.5%) 222 (92.9%) 256 (89.8%) 164 (95.9%) 313 (88.7%) 213 (93.4%) 264 (89.2%) 447 (91.4%) 30 (85.7%) 456 (91.6%)) 21 (80.8%) 452 (91.1%) 25 (89.3%) 277 (88.5%) 200 (94.8%) 404 (92.4%) 73 (83.9%) 422 (91.5%) 55 (87.3%) 143 (92.3%) 256 (89.2%) 458 (91.6%) 17 (77.3%)
34.8 ± 12.5 14.2 ± 13.2 7.9 ± 3.5 7.4 ± 3.0 17.2 ± 19.0 23 (10.1%) 24 (8.8%) 16 (5.2%) 15 (18.8%) 11 (9.0%) 5 (38.5%) 21 (8.0%) 26 (10.0%) 27 (7.3%) 20 (12.9%) 27 (9.2%) 20 (8.7%) 42 (8.5%) 5 (15.6.5%) 41 (8.7%) 6 (11.8%) 39 (8.0%) 9 (20.5%) 18 (7.5%) 29 (10.2%) 7 (4.1%) 40 (11.3%) 15 (6.6%) 32 (10.8%) 42 (8.6%) 5 (14.3%) 42 (8.4%) 5 (19.2%) 44 (8.9%) 3 (10.7%) 36 (11.5%) 11 (2.2%) 33 (7.6%) 14 (16.1%) 39 (8.5%) 8 (12.7%) 12 (7.7%) 31 (10.8%) 42 (8.4%) 5 (22.7%)
p value
.728 .005a .246 .470 .001a .549
.001a .001a .373 .046a 0.878 .193 .441 .016a .358 .008a .122 .228 .073 .731 .013a .022a .247 .400 .039a
AED — antiepileptic drugs, GTCS — generalized tonic–clonic seizures, PNES — psychogenic nonepileptic seizure, VNS — vagal nerve stimulation therapy, SE — status epilepticus, MRI — magnetic resonance imaging. a Significance set at ≤0.05.
in another case. Three of the seizure-related falls were associated with other traumatic injuries (nasal fracture, eyebrow laceration, lip laceration). The number of falls did not differ significantly between the withdrawal group (4 cases) and the no-withdrawal group (2 cases) (p = 0.4). Intravenous line-related AEs occurred 6 times (9.5% of all AEs). Superficial phlebitis not necessitating antibiotic treatment occurred 3 times, and Staphylococcus aureus bacteremia necessitating IV antibiotics occurred twice. Contrast extravasation to the arm's soft tissues during computerized tomography occurred once. Electrode-related AEs occurred in 3 admissions (4.8% of all AEs). Two cases involved intracranial subdural EEG recordings (one extra-axial hematoma and one extra-axial swelling following subdural grid insertion). In the third case, scalp EEG electrodes resulted in skin discomfort necessitating electrode removal and aborting of the EEG monitoring. Status epilepticus occurred twice (3.2% of all AEs). The two cases were of focal SE with impaired consciousness, and both occurred without AED withdrawal. One case was a focal SE with impaired consciousness of a frontal lobe origin: treatment with IV benzodiazepines and valproate
stopped the SE but caused valproate-induced hyperammonemia. The second case was a focal SE with impaired consciousness of temporal lobe origin, which was controlled with IV benzodiazepines, valproate, and levetiracetam. Neither of these patients had a prior history of SE. Cardiac AE occurred once (1.6% of all AEs) in the form of suspected ictal asystole lasting 18 s. The asystole occurred during a focal seizure with altered consciousness, with ictal rhythmic delta activity over the right anterobasal temporal region. The asystole was self-limited, and the patient fully recovered and was offered cardiac pacemaker implantation which he declined. Psychiatric AE in the form of a suicide attempt occurred once (1.6% of all AEs) in a patient with a past medical history of genetic generalized epilepsy, borderline personality disorder, posttraumatic stress disorder, depression, alcohol and illicit drug abuse, and several suicide attempts in the past. The suicide attempt followed a number of PNES events, and no epileptic seizures were documented during monitoring. Suicide was attempted by wrist cutting without hemodynamic compromise: the cuts were sutured, and the patient was transferred to a psychiatric hospital for further treatment.
166
F. Fahoum et al. / Epilepsy & Behavior 61 (2016) 162–167
Table 3 Significant risk factors for safety adverse events (multivariate logistic regression analysis). Factor
Significance
Noninvasive presurgical workup Invasive presurgical workup Number of events during monitoring Levetiracetam Sulthiame
.004 .002 .004 .012 .014
Odds ratio
3.149 7.722 1.027 2.319 5.133
95% C.I. for odds ratio Lower
Upper
1.431 2.114 1.008 1.207 1.384
6.929 28.205 1.046 4.456 19.038
C.I.—confidence interval.
There were no respiratory or thrombotic AEs and no deaths during the study period, and there was no need for resuscitation. The results of the univariate analysis that compared the AE and non-AE groups are shown in Table 2. There were significantly more AEs when patients were younger at disease onset (p = 0.005) and when there was a history of temporal lobe epilepsy (p = 0.46), focal seizures with altered consciousness (p = 0.008), SE (p = 0.022), VNS (p = 0.039), and intellectual disability (p = 0.016) and when there was an indication for EMU monitoring for noninvasive presurgical evaluation or intracranial monitoring (p = 0.001). Adverse events occurred more often when patients had more overall events in the EMU (p = 0.001). The patients in the AE group more frequently used the following AEDs compared with those in the non-AE group: carbamazepine (p = 0.037), levetiracetam (p = 0.004), clobazam (p = 0.008), and sulthiame (p = 0.016). The median number of AEDs in the AE group was 3 AEDs and in the non-AE group 2 AEDs. Patients with PNESs had fewer AEs compared with patients without PNESs (p = 0.013). Age, gender, duration of hospitalization and duration of monitoring, number of days to AED withdrawal and renewal, seizure frequency by history, presence of a major psychiatric comorbidity, abnormal findings on the neurological exam, and presence of a lesion on the MRI were not associated with AEs. An additional analysis to test for the frequency of specific AEs in the different paroxysmal syndromes showed no significant increase of
specific AEs in the different epilepsy syndromes (temporal, extratemporal, and generalized); however, there were fewer SCs in the group with PNES compared with the group with epilepsy (p = 0.013). In the multivariate analysis, AEs occurred more when the indication for EMU admission was a presurgical evaluation (noninvasive p = 0.004 and intracranial p = 0.002), when there were more events during monitoring (p = 0.004), and when patients were prescribed levetiracetam and sulthiame (p = 0.012 and p = 0.014, respectively) (Table 3). Repeating the univariate and multivariate analyses while including only the first admission led to similar results when including all admissions (compare Supplementary Tables 1 and 2 with Tables 2 and 3). The results of the comparison of AE rates in our study with those published in the literature are shown in Table 4. 4. Discussion Our analysis of the frequency and risk factors for AE occurrences in an adult EMU revealed that AEs occurred in nearly 9% of admissions. This rate is comparable with the AE rate reported in a recent comprehensive review of the current literature by Rheims and Ryvlin [12]. The most common AE was SC, comprising more than one-half of all AEs. The rest of the AEs in decreasing order of frequency were related to AEDs, falls and injuries, IV lines, electrodes, SE, and cardiac and psychiatric causes. Adverse events were linked to younger age at disease onset, history of SE and VNS use, and admission for presurgical evaluation, all pointing to more severe epilepsy. The association between younger age at disease onset and AEs confirms the findings of Dobesberger et al. with duration of epilepsy as an important predictive factor for AEs [1]. Furthermore, patients with mental retardation, history of temporal lobe epilepsy, and history of focal seizures with altered consciousness and those with many events during monitoring also had more AEs. These findings emphasize the importance of adhering to strict safety routines for these patients. Unlike many earlier studies, the current report included patients with a history of PNES, and the results revealed that PNESs were associated with a lower AE risk.
Table 4 The frequency of safety adverse events (AEs) reported in the literature and the current study. Study [ref.] year
Study cohort (age group)
AED withdrawal policy
[8] 2001
Presurgical workup in TLE (Adult) All EMU admissions (N/A) All EMU admissions (Adult)
AED reduction start day 2 Stop all AED on days 4–6 Individualized
[3] 2003 [9] 2009
[2] 2009 [1] 2011 [19] 2012 [6] 2012
[18] 2013 [16] 2014 Current 2016
Patients with epilepsy (adult) All EMU admissions (adult/ped) All EMU admissions (adult/ped) Presurgical workup (adult) All EMU admissions (adult) All EMU admissions (adult) All EMU admissions (adult)
Percentage of AEs in admissions (%) 24-h SC
1/2 of all AEDs reduced at admission and discontinued at 24 h Patients with history of SE — 1/4 of all AEDs daily PB was not withdrawn Individualized
4-h SC
AED-related
Electrode-related
48.5
17.8
Cardiac
Psychiatric
2.75
1.8
3 2
23
SE
0
9
2
3.7 0.2
N/A
3.2
2.2
3.4
2 39
11
5.6
0
0
2.3
Individualized AED reduction started on day 2 1/2 of single AED each day
IV-related
48.3
Individualized
1/3 of all AEDs reduced daily at admission except in history of SE or daily seizures N/A
Falls and injuries
0.6 N/A
5.4
1.3
1.1
0.8 1.1
0.6
0.4
0.2
0.2
AED — antiepileptic drug, SC — seizure cluster, SE — status epilepticus, IV — intravenous, TLE — temporal lobe epilepsy, EMU — epilepsy monitoring unit, Ped — pediatric, N/A — not available, PB — phenobarbital, BDZ — benzodiazepines.
F. Fahoum et al. / Epilepsy & Behavior 61 (2016) 162–167
Patients with AEs were treated more frequently with carbamazepine, levetiracetam, clobazam, and sulthiame. These AEDs were used in different polytherapy combinations, and their increased frequency in the AE group is probably due to more severe epilepsy and to local preferences for AED treatment. We could not find specific relations between the use or withdrawal of certain AEDs and increased risk of AEs. Adverse events in the EMU are usually linked to AED withdrawal, which is routinely undertaken to activate seizures [6,8,9]. However, whereas AED withdrawal might be crucial in some cases (e.g., for seizure characterization, presurgical evaluation, and others), it is not implemented when the indication for admission is AED adjustment or when there is a history of high seizure frequency. In our cohort, while AEDs were withdrawn in nearly one-half of all admissions, there was no significant increase in AEs in those patients compared with that in patients without AED withdrawal. There are a few possible explanations for the lack of increase of AEs in the withdrawal group: 1. our implementation of slow AED withdrawal starting from the second day or later since the real seizure frequency is sometimes underreported by patients [13,14]; 2. withdrawing of AEDs sequentially and never more than one AED at a time; 3. refraining from AED withdrawal in patients with a history of SE and/or high seizure frequency, with the exception of those admitted for presurgical evaluation and those who experienced no seizures during the first days in the EMU; 4. keeping patients with high seizure frequency during monitoring for medical observation in the EMU for an extra one or more days (without EEG monitoring); and 5. encouraging the patient's companion to stay alert overnight in the patient's room, as this might reduce seizures missed by the staff [15]. Another possible explanation for the lack of difference in the AE rate between the withdrawal and no-withdrawal groups is the higher seizure frequency by history in the latter group, pointing to more severe disease. This finding stresses that AEs occurring in the EMU are not necessarily related to AED withdrawal, a point further supported by the lack of correlation between the first day of AED withdrawal and the day of AE occurrence. The distribution of AEs throughout the length of stay at the EMU showed that 93% of AEs occurred during 6 days of monitoring, whereas 91% of AEs occurred during the first 4 days of monitoring in an earlier study [1]. It is difficult to compare published studies, as often different patient populations were studied (e.g., only patients with epilepsy, only presurgical evaluation), and different categories of AEs were defined; nevertheless, comparing our AE rates with those published in the literature (Table 4), we found 4-h SCs in 5.4% of all EMU admissions, in contrast to 9–18% of patients in the literature [2,3,6]. The events of SE recorded in 0.4% of admissions were comprised of events of focal SE, while the reported rates of convulsive SE in comparable cohorts were 0.8–3% [1,3,16]. In agreement with other studies, all falls occurred when our patients were outside the bed [17,18], although we had fewer fall events (n = 6, 1.1% of admissions) than the 0.6–5.6% reported in comparable cohorts [1,9,16,18,19]. We had no events of postictal psychosis and fewer psychiatric AEs than reported by others in the literature [1,2], but on the other hand, we had relatively high AED-related AEs compared with Dobesberger et al. [1]. In the latter study, patients with minor or major psychiatric comorbidities had a more than 16-fold increased risk of sustaining psychiatric AEs during or after monitoring; however, in 13 of 20 patients, the psychiatric AE occurred after discharge from EMU [1]. Two main reasons could explain the difference between our study and the study of Dobesberger et al. regarding the frequency of psychiatric complications: we only included major psychiatric comorbidities, and we did not have the data of possible complications occurring after the discharge from the EMU. Recently published data showed that patients with epilepsy have an increased risk of thromboembolism [20]. Interestingly, there were no thrombotic events in our cohort, despite the fact that the patients did not receive prophylactic anticoagulation and were relatively immobile during monitoring. The main limitations of our study are its retrospective design and the lack of follow-up data on occurrences of AEs in the days following
167
discharge from the EMU, as AEs have been reported in that period [1]. The strength of this study is the inclusion of all consecutive patients admitted to an adult EMU and not limiting the study to a specific patient subgroup, therefore providing a comprehensive review of safety issues in adult EMUs. In summary, we found that AEs are not uncommon in the EMU and that they tend to occur more frequently in patients with more severe epilepsy and intellectual disability and those undergoing presurgical evaluations. They are less common in patients with PNESs. Seizure cluster was the most common AE, comprising nearly one-half of all AEs. While AEs are known to occur in patients undergoing and not undergoing AED withdrawal, we believe that our careful and selective AED withdrawal policy in addition to other safety measures was related to no increase in AE risk in patients undergoing AED withdrawal. This work further emphasizes the need for internationally standardized protocols to improve the patient safety in the EMU. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.yebeh.2016.06.002.
Acknowledgments The authors thank Ms. Ilana Gelernter from the Department of Statistics and Operations Research at Tel Aviv University for her help with statistical analysis and Ms. Esther Eshkol for editorial assistance.
References [1] Dobesberger J, Walser G, Unterberger I, Seppi K, Kuchukhidze G, Larch J, et al. VideoEEG monitoring: safety and adverse events in 507 consecutive patients. Epilepsia 2011;52:443–52. [2] Noe KH, Drazkowski JF. Safety of long-term video-electroencephalographic monitoring for evaluation of epilepsy. Mayo Clin Proc 2009;84:495–500. [3] Rose AB, McCabe PH, Gilliam FG, Smith BJ, Boggs JG, Ficker DM, et al. Occurrence of seizure clusters and status epilepticus during inpatient video-EEG monitoring. Neurology 2003;60:975–8. [4] Spritzer SD, Riordan KC, Berry J, Corbett BM, Gerke JK, Hoerth MT, et al. Fall prevention and bathroom safety in the epilepsy monitoring unit. Epilepsy Behav 2015;48:75–8. [5] Kandler R, Lai M, Ponnusamy A, Bland J, Pang C. The safety of UK video telemetry units: results of a national service evaluation. Seizure 2013;22:872–6. [6] Di Gennaro G, Picardi A, Sparano A, Mascia A, Meldolesi GN, Grammaldo LG, et al. Seizure clusters and adverse events during pre-surgical video-EEG monitoring with a slow anti-epileptic drug (AED) taper. Clin Neurophysiol 2012;123:486–8. [7] Mohan KK, Markand ON, Salanova V. Diagnostic utility of video EEG monitoring in paroxysmal events. Acta Neurol Scand 1996;94:320–5. [8] Yen DJ, Chen C, Shih YH, Guo YC, Liu LT, Yu HY, et al. Antiepileptic drug withdrawal in patients with temporal lobe epilepsy undergoing presurgical video-EEG monitoring. Epilepsia 2001;42:251–5. [9] Moien-Afshari F, Griebel R, Sadanand V, Vrbancic M, Hernandez-Ronquillo L, Lowry N, et al. Safety and yield of early cessation of AEDs in video-EEG telemetry and outcomes. Can J Neurol Sci 2009;36:587–92. [10] Proposal for revised classification of epilepsies and epileptic syndromes. Commission on Classification and Terminology of the International League Against Epilepsy. Epilepsia 1989;30:389–99. [11] Lowenstein DH, Bleck T, Macdonald RL. It's time to revise the definition of status epilepticus. Epilepsia 1999;40:120–2. [12] Rheims S, Ryvlin P. Patients' safety in the epilepsy monitoring unit: time for revising practices. Curr Opin Neurol 2014;27:213–8. [13] Blum DE, Eskola J, Bortz JJ, Fisher RS. Patient awareness of seizures. Neurology 1996; 47:260–4. [14] Hoppe C, Poepel A, Elger CE. Epilepsy: accuracy of patient seizure counts. Arch Neurol 2007;64:1595–9. [15] Dericioglu N, Albakir M, Saygi S. The role of patient companions in long-term videoEEG monitoring. Seizure 2000;9:124–7. [16] Sauro KM, Macrodimitris S, Krassman C, Wiebe S, Pillay N, Federico P, et al. Quality indicators in an epilepsy monitoring unit. Epilepsy Behav 2014;33:7–11. [17] Atkinson M, Hari K, Schaefer K, Shah A. Improving safety outcomes in the epilepsy monitoring unit. Seizure 2012;21:124–7. [18] Pati S, Kumaraswamy VM, Deep A, Chung SS, Plueger M, Kiyota G, et al. Characteristics of falls in the epilepsy monitoring unit: a retrospective study. Epilepsy Behav 2013;29:1–3. [19] Spanaki MV, McCloskey C, Remedio V, Budzyn D, Guanio J, Monroe T, et al. Developing a culture of safety in the epilepsy monitoring unit: a retrospective study of safety outcomes. Epilepsy Behav 2012;25:185–8. [20] Martz GU, Wilson DA, Malek AM, Selassie AW. Risk of venous thromboembolism in people with epilepsy. Epilepsia 2014;55:1800–7.
Contents lists available at ScienceDirect
Epilepsy & Behavior journal homepage: www.elsevier.com/locate/yebeh
Safety in the epilepsy monitoring unit: A retrospective study of 524 consecutive admissions Firas Fahoum a,⁎, Nurit Omer a, Svetlana Kipervasser a,b, Tal Bar-Adon a, Miri Neufeld a,b a b
Epilepsy and EEG Unit, Department of Neurology, Tel Aviv Sourasky Medical Center, Tel Aviv 6423906, Israel Sackler School of Medicine, Tel Aviv University, Tel Aviv 69978, Israel
a r t i c l e
i n f o
Article history: Received 18 February 2016 Revised 3 May 2016 Accepted 2 June 2016 Available online xxxx Keywords: Epilepsy monitoring unit Safety adverse events Antiepileptic drugs Seizure clusters Psychogenic nonepileptic seizure Status epilepticus
a b s t r a c t The yield of monitoring patients at an epilepsy monitoring unit (EMU) depends on the recording of paroxysmal events in a timely fashion, however, increasing the risk of safety adverse events (AEs). We aimed to retrospectively study the frequency and risk factors for AE occurrences in all consecutive admissions to an adult EMU in a tertiary medical center. We also compared our findings with published data from other centers. Between January 2011 and June 2014, there were 524 consecutive admissions to the adult EMU at the Tel Aviv Sourasky Medical Center. Adverse events were recorded in 47 (9.0%) admissions. The most common AE was 4-hour seizure cluster (58.7% of AEs) and, in decreasing frequency, AEs related to antiepileptic drugs (AEDs, 11.1%), falls and traumatic injuries (9.5%), intravenous line complications (9.5%), electrode-related (4.8%), status epilepticus (SE, 3.2%), and cardiac (1.6%) and psychiatric (1.6%) complications. There were significantly more AEs among patients with a younger age at disease onset (p = 0.005), a history of temporal lobe epilepsy (p = 0.046), a history of focal seizures with altered consciousness (p = 0.008), a history of SE (p = 0.022), use of a vagal nerve stimulator (p = 0.039), and intellectual disability (p = 0.016) and when the indication for EMU monitoring was noninvasive or invasive presurgical evaluation (p = 0.001). Adverse events occurred more frequently when patients had more events in the EMU (p = 0.001) and among those administered carbamazepine (p = 0.037), levetiracetam (p = 0.004), clobazam (p = 0.008), and sulthiame (p = 0.016). Patients with a history of psychogenic nonepileptic seizures (PNESs) had significantly fewer AEs (p = 0.013). Adverse events were not associated with the age, gender, duration of hospitalization or monitoring, AED withdrawal and renewal, seizure frequency by history, presence of major psychiatric comorbidities, abnormal neurological exam, or the presence of a lesion as on brain magnetic resonance imaging. In conclusion, this study reveals that AEs are not unusual in the EMU and that seizure clustering is the most common among them. Adverse events occur more frequently in patients with more severe epilepsy and intellectual disability and in patients undergoing presurgical evaluations and less frequently in patients with PNESs. © 2016 Elsevier Inc. All rights reserved.
1. Introduction Epilepsy monitoring units (EMUs) use video-electroencephalography (video-EEG) recordings for several indications, such as distinguishing epileptic seizures from psychogenic nonepileptic seizures (PNESs) and other nonepileptic paroxysmal events, evaluating patients for epilepsy surgery, characterizing seizure types, and adjusting antiepileptic drug (AED) treatment. The yield of the monitoring is highly dependent on the recording of epileptic and/or nonepileptic events. To increase the likelihood of capturing events in a timely fashion, it is a standard practice to use activating procedures, such as AED withdrawal and sleep deprivation.
⁎ Corresponding author at: Epilepsy and EEG Unit, Department of Neurology, Tel Aviv Sourasky Medical Center, 6 Weizmann St., Tel Aviv 6423906, Israel. E-mail address: fi[email protected] (F. Fahoum).
http://dx.doi.org/10.1016/j.yebeh.2016.06.002 1525-5050/© 2016 Elsevier Inc. All rights reserved.
These measures may provoke patients to experience an increase in the frequency and severity of seizures, therefore increasing the risks of injuries and medical complications. Surveys of EMUs in the United States and Europe demonstrated that the most frequent safety adverse events (AEs) were falls, status epilepticus (SE), and postictal psychosis [1,2]. According to one study, major AEs appeared only during generalized tonic–clonic seizures (GTCSs), and the risk factors for AEs were epilepsy duration, existence of psychiatric comorbidity, and history of SE [1]. Seizure clusters (SCs) are other common AEs [2,3]. Studies on interventions to reduce the frequency of AEs highlight the importance of adherence to stricter safety measures [4], decreasing the number of missed seizures [5], using AED withdrawal protocols [6–8], and providing medical interventions in cases of SE or SC [2,6]. Whereas PNESs are frequently encountered in the EMU, in up to 32% patients in tertiary epilepsy centers [7], many studies on AEs in EMUs excluded those patients and included only patients with epilepsy
F. Fahoum et al. / Epilepsy & Behavior 61 (2016) 162–167
163
[2,3,8,9]. In this study, we analyzed all consecutive admissions to an EMU in a tertiary center and examined the frequency and risk factors for the occurrences of AEs. We also compared our findings with published data from other centers.
convulsive seizures lasting longer than 5 min or when consciousness was not regained between two consecutive seizures [11]. The study was approved by the local research ethics committee at TASMC.
2. Methods
2.2. Technical details and work routines at EMU
2.1. Patients
The EMU is located in physical proximity to the Department of Neurology. During the study period, it had 3 adult monitoring beds with a digital 32-/64-/128-channel video-EEG system (NicOne LTM system, Viasys, Madison, WI, USA). Surface EEG electrodes (Ag/AgCl) were placed according to the international 10–20 system, with additional inferior temporal or closely spaced electrodes when needed. Subdural strips and grids and depth electrodes (Ad-Tech Medical, Racine, WI, USA) were used for invasive EEG recordings. Two electrocardiography (ECG) electrodes were applied in all patients. The EMU staff included 2 EEG technicians from 08:00 to 16:00 on weekdays (Sunday–Thursday) and one nurse who provided 24/7 coverage. A specialized epileptologist was in the EMU during the daytime and on call 24/7 (FF, SK, MN). There was full continuous access to the medical staff of the Department of Neurology. The patients were hospitalized in single rooms with private bathrooms and were monitored by the nurse via a monitoring screen located at the nurse's station. One companion received instructions from the EMU nurse and was encouraged to stay overnight in the patient's room. Thick padded side rails in the up position were used to minimize the risk of seizure-related injuries. In order to ensure intravenous (IV) access in case of emergency, IV catheters were placed in the antecubital fossa of patients who underwent AED withdrawal, as well as those with a high seizure frequency or who had been implanted with intracranial electrodes. Intravenous catheters were flushed twice daily with IV heparin 100 i.u. Antiepileptic drug withdrawal followed the following outline: it was avoided during the first day of admission and, if required, started on the second day or later, taking into account the reported frequency of events prior to admission, the number and dosages of AEDs, and any history of SE. Starting from the second day of admission, the dose of a
This retrospective study included all consecutive patients older than 18 years who had been admitted to the EMU at the Tel Aviv Sourasky Medical Center (TASMC) between January 1, 2011 and June 30, 2014. This period was chosen because fully computerized and standardized medical files had become operable in the EMU. Prior to monitoring, all patients had signed informed consent forms that explained the purpose of monitoring, the possibility of AED withdrawal when required, and its potential risks, as well as the safety measures to ensure patient safety. Medical files were reviewed and the following data documented: patient's age, gender, age of onset of epileptic or paroxysmal events, classification of paroxysmal syndrome (focal temporal epilepsy, focal extratemporal epilepsy, generalized epilepsy, PNES), seizure types according to ILAE classification [10], reported frequency of events prior to EMU admission, history of SE (Yes/No), presence of prediagnosed intellectual disability (Yes/No), presence of major psychiatric comorbidities (major mood disorder, schizophrenia, personality disorder, severe behavioral disturbances, history of suicide attempt, and substance abuse) (Yes/No), use of vagal nerve stimulation (VNS) (Yes/No), AED type, presence of abnormal neurological examination (Yes/No), and presence of a lesion on brain magnetic resonance imaging (MRI) (Yes/No). We also documented the indication for EMU admission, duration of hospitalization, duration of EEG monitoring, number of paroxysmal events during monitoring, first day of AED withdrawal and first day of AED renewal since admission, first day of paroxysmal events, number of AEs, and dates of occurrence and types of AEs as listed in Table 1. An SC was defined as 3 focal seizures with impaired consciousness or 2 GTCSs occurring within 4 h [3]. Status epilepticus was defined as
Table 1 Characterization of safety adverse events (AEs) in the AED withdrawal and in the no-withdrawal groups. AEs (in decreasing order of frequency)
Total number of AEs
AEs in withdrawal group
AEs in no-withdrawal group
AE characterization
Seizure clustersa
37
18
19
AED-related
7
2
5
Falls/traumatic injuries
6
4
2
IV line-related
6
4
2
Electrode-related
3
1
2
2 1 1 0 0 0 63
1 0 1 0 0 0 31
1 1 0 0 0 0 32
33 focal seizures with impaired consciousness 3 GTCSs 1 atypical absence 3 VPA-induced hyperammonemia started in EMU 2 VPA-induced hyperammonemia taken prior to EMU admission 1 CBZ–LTG interaction 1 PB-induced increase in liver enzymes 2 atonic seizures 1 GTCS and nasal fracture outside the EMU 1 GTCS with eyebrow laceration 1 fall after PNES 1 fall and lip laceration due to extrapyramidal syndrome 3 phlebitis 2 Staphylococcus aureus bacteremia 1 extravasation of CT contrast agent 1 extra-axial hematoma following subdural grid insertion 1 extra-axial swelling following subdural grid insertion 1 scalp discomfort due to scalp EEG electrodes leading to electrodes removal 2 focal SE with impaired consciousness 1 ictal asystole 1 nonlethal suicide attempt
Status epilepticusb Cardiac Psychiatric Respiratory Thrombotic Resuscitation or death Total
AED — antiepileptic drug, GTCS — generalized tonic–clonic seizure, PNES — psychogenic nonepileptic seizure, EMU — epilepsy monitoring unit, SE — status epilepticus, VPA — valproate, CBZ — carbamazepine, LTG — lamotrigine, PB — phenobarbital, CT — computerized tomography. a Defined as 3 focal seizures with impaired consciousness or 2 GTCSs occurring within 4 h [3]. b Defined as convulsive seizures lasting longer than 5 min or when consciousness was not regained between two consecutive seizures [11].
164
F. Fahoum et al. / Epilepsy & Behavior 61 (2016) 162–167
single AED was reduced each day by one-half until fully discontinued. If no paroxysmal events occurred, the dose of the second AED was reduced by one-half each day and so on, until enough paroxysmal events were acquired to answer the question of referral. Barbiturates and benzodiazepines were not withdrawn. Antiepileptic drug withdrawal was avoided in patients with a history of SE unless it was crucial for presurgical evaluation. In addition, sleep deprivation was asked from selected patients. In case of SC after AED withdrawal, all AEDs were renewed simultaneously to the prewithdrawal dosages. Intravenous benzodiazepines were administered if SC persisted despite AED renewal or in cases where AEDs were not withdrawn. Patients with SC or high seizure frequency during monitoring were kept in the EMU for medical observation without EEG monitoring until full medical treatment had been reinstated. 2.3. Statistical analysis For univariate analysis, we used a parametric test (Student's t-test) when the variables had a normal distribution and a nonparametric test (Mann–Whitney test) for a non-normal distribution. For categorical variables, we applied the chi-square or Fisher's exact test according to the expected frequency in the cell. For multivariate analysis, we used stepwise logistic regression analysis by entering all potential risk factors that showed a significant association at a p value b0.05 in the univariate analyses. A nonparametric test (Spearman's test) was used to calculate the correlation coefficient. We repeated the analyses including only the first admission to account for potential bias towards patients that were admitted more than once to the EMU. The significance level was set at p ≤ 0.05. All statistical analyses were performed using IBM SPSS Statistics 23. 3. Results 3.1. Patients and monitoring data There were 524 admissions to the EMU during the study period. The patients' ages ranged between 18 and 82 years (mean ± standard deviation: 35.5 ± 13.9 years; 296 females and 238 males). Three patients were admitted 4 times, 4 patients were admitted 3 times, and 34 patients were admitted twice to the EMU. Indications for admission were characterization of paroxysmal events (309 admissions, 59.0%), noninvasive presurgical workup (8 admissions, 15.3%), AED adjustment (122 admissions, 23.3%), and intracranial EEG studies (13 admissions, 2.5%). The types of paroxysmal syndrome for all admissions were as follows: 155 (29.6%) patients with focal temporal epilepsy, 231 (44.1%) patients with focal extratemporal epilepsy, 32 (6.1%) patients with generalized epilepsy, and 211 (40.3%) patients with PNESs. There were 105 admissions (20%) of patients with combined epilepsy and PNESs, 87 (16.6%) patients with a history of SE, and 21 (4%) patients with VNS. In addition, 51 (9.7%) admissions involved patients with major psychiatric comorbidities, 39 (7.4%) patients with intellectual disability, and 63 (12%) of patients with abnormal findings on the neurological exam. A total of 442 patients had a brain MRI, and a lesion was demonstrated in 287 of them (64.9%). The mean duration of hospitalization at the EMU was 7.5 ± 2.5 days, and the mean duration of EEG monitoring was 7.2 ± 2.2 days. Paroxysmal events were recorded in 414 admissions (79%), and none were recorded in 110 admissions (21%).
53 (20.4%) for noninvasive presurgical workup, 54 (20.7%) for AED adjustment, and 10 (3.8%) for intracranial EEG studies. The AEDs were renewed in 214 (82%) patients in the withdrawal group and not renewed in 46 (18%). Antiepileptic drug withdrawal was started after a mean of 2.4 ± 1.5 days since admission and AED renewal after a mean of 6.2 ± 3.3 days since admission. Antiepileptic drugs were not withdrawn (no-withdrawal group) in 264 admissions (50.4%). The indications for admission in that group were as follows: 166 (62.9%) for characterization of paroxysmal events, 27 (10.2%) for noninvasive presurgical workup, 68 (25.8%) for AED adjustment, and 3 (1.1%) for intracranial EEG studies. The mean durations of hospitalization and monitoring in the withdrawal group (8.1 ± 2.2 and 7.9 ± 2.0 days, respectively) were longer than in the no-withdrawal group (6.9 ± 2.6 and 6.5 ± 2.1 days; p b 0.001 and p b 0.001, respectively). Paroxysmal events occurred in 210 of 260 (80.8%) patients in the withdrawal group and in 204 of 264 (73.7%) patients in the no-withdrawal group (p = 0.33). 3.3. Safety adverse events (AEs) Sixty-three AEs were documented in 47 EMU admissions (AE group, 9.0% of all admissions). The mean day of AE occurrence was 3.7 ± 2.1 days from admission. The distribution of the day of AE occurrence is shown in Fig. 1. Antiepileptic drug withdrawal was not associated with an increase in AEs (p = 0.37) (Table 2), and there was no significant correlation between the first day of AED withdrawal and the day of AE occurrence (rs = 0.047, p = 0.28). The most common AE was SC, which was reported 37 times (58.7% of all AEs) in 28 admissions. Seizure clusters were documented once in 22 admissions, twice in 3 admissions, and 3 times in 3 admissions. Clusters of focal seizures with altered consciousness occurred 29 times in 21 admissions, clusters of focal seizures evolving to bilateral convulsive seizures occurred in 5 admissions, and clusters of atypical absence occurred 3 times in 2 admissions. Seizure clusters occurred 18 times in the withdrawal group and 19 times in the no-withdrawal group (p = 0.9). Antiepileptic drug-related AEs occurred 7 times (11.1% of all AEs). Valproate-induced hyperammonemia occurred 3 times when it was started in the EMU and twice when patients had been treated with valproate prior to EMU admission. The interaction of carbamazepine and lamotrigine started in the EMU caused reversible nausea, vertigo, and ataxia in one case, and an increasing phenobarbital dose caused an increase in hepatocellular enzymes in another case. Falls and traumatic injuries occurred 6 times (9.5% of all AEs). All cases were outside the bed: 4 were in the monitoring room, 1 was in the bathroom, and 1 was outside the EMU. The fall accompanied an epileptic seizure in 4 cases (2 GTCSs and 2 atonic seizures), followed a PNES in one case, and was secondary to an extrapyramidal syndrome
3.2. AED withdrawal and renewal Antiepileptic drugs were withdrawn (withdrawal group) in 260 admissions (49.6%). The indications for EMU admission for that group were as follows: 143 (55%) for characterization of paroxysmal events,
Fig. 1. Histogram showing the distribution of the day of safety adverse event (AE) occurrence (n = 63) since admission to the epilepsy monitoring unit (EMU).
F. Fahoum et al. / Epilepsy & Behavior 61 (2016) 162–167
165
Table 2 Comparison of the admissions with and without safety adverse events (univariate analysis). Variable
Age Onset age Hospitalization duration Monitoring duration Number of events during monitoring Gender Indication for admission
AED withdrawal Focal temporal epilepsy Focal extratemporal epilepsy Generalized epilepsy Major psychiatric comorbidities Intellectual disability History of focal seizures without altered consciousness History of focal seizures with altered consciousness History of focal seizures evolving to bilateral convulsive seizures History of GTCS History of absence History of myoclonic seizures History of PNES History of SE Abnormal neurological exam MRI lesion VNS
Unit/category
Years Years Days Days Events Males = 227 Females = 297 Event characterization = 309 Noninvasive presurgical workup = 80 AED adjustment = 122 Invasive presurgical workup = 13 No = 264 Yes = 260 No = 369 Yes = 155 No = 293 Yes = 231 No = 492 Yes = 32 No = 473 Yes = 51 No = 485 Yes = 39 No = 239 Yes = 285 No = 171 Yes = 353 No = 228 Yes = 296 No = 489 Yes = 35 No = 498 Yes = 26 No = 496 Yes = 28 No = 313 Yes = 211 No = 437 Yes = 87 No = 461 Yes = 63 No = 155 Yes = 287 No = 502 Yes = 22
Non-AE group (N = 477)
AE group (N = 47)
Mean ± SD
Mean ± SD
35.6 ± 14.0 21.3 ± 16.8 7.5 ± 2.3 7.2 ± 2.1 9.3 ± 12.5 204 (89.9%) 273 (91.9%) 293 (94.8%) 65 (81.3%) 111 (91.0%) 8 (61.5%) 243 (92.0%) 234 (90.0%) 342 (92.7%) 135 (87.1%) 266 (90.8%) 211 (91.3%) 450 (91.5%) 27 (84.4%) 432 (91.3%) 45 (88.2%) 446 (92.0%) 31 (79.5%) 222 (92.9%) 256 (89.8%) 164 (95.9%) 313 (88.7%) 213 (93.4%) 264 (89.2%) 447 (91.4%) 30 (85.7%) 456 (91.6%)) 21 (80.8%) 452 (91.1%) 25 (89.3%) 277 (88.5%) 200 (94.8%) 404 (92.4%) 73 (83.9%) 422 (91.5%) 55 (87.3%) 143 (92.3%) 256 (89.2%) 458 (91.6%) 17 (77.3%)
34.8 ± 12.5 14.2 ± 13.2 7.9 ± 3.5 7.4 ± 3.0 17.2 ± 19.0 23 (10.1%) 24 (8.8%) 16 (5.2%) 15 (18.8%) 11 (9.0%) 5 (38.5%) 21 (8.0%) 26 (10.0%) 27 (7.3%) 20 (12.9%) 27 (9.2%) 20 (8.7%) 42 (8.5%) 5 (15.6.5%) 41 (8.7%) 6 (11.8%) 39 (8.0%) 9 (20.5%) 18 (7.5%) 29 (10.2%) 7 (4.1%) 40 (11.3%) 15 (6.6%) 32 (10.8%) 42 (8.6%) 5 (14.3%) 42 (8.4%) 5 (19.2%) 44 (8.9%) 3 (10.7%) 36 (11.5%) 11 (2.2%) 33 (7.6%) 14 (16.1%) 39 (8.5%) 8 (12.7%) 12 (7.7%) 31 (10.8%) 42 (8.4%) 5 (22.7%)
p value
.728 .005a .246 .470 .001a .549
.001a .001a .373 .046a 0.878 .193 .441 .016a .358 .008a .122 .228 .073 .731 .013a .022a .247 .400 .039a
AED — antiepileptic drugs, GTCS — generalized tonic–clonic seizures, PNES — psychogenic nonepileptic seizure, VNS — vagal nerve stimulation therapy, SE — status epilepticus, MRI — magnetic resonance imaging. a Significance set at ≤0.05.
in another case. Three of the seizure-related falls were associated with other traumatic injuries (nasal fracture, eyebrow laceration, lip laceration). The number of falls did not differ significantly between the withdrawal group (4 cases) and the no-withdrawal group (2 cases) (p = 0.4). Intravenous line-related AEs occurred 6 times (9.5% of all AEs). Superficial phlebitis not necessitating antibiotic treatment occurred 3 times, and Staphylococcus aureus bacteremia necessitating IV antibiotics occurred twice. Contrast extravasation to the arm's soft tissues during computerized tomography occurred once. Electrode-related AEs occurred in 3 admissions (4.8% of all AEs). Two cases involved intracranial subdural EEG recordings (one extra-axial hematoma and one extra-axial swelling following subdural grid insertion). In the third case, scalp EEG electrodes resulted in skin discomfort necessitating electrode removal and aborting of the EEG monitoring. Status epilepticus occurred twice (3.2% of all AEs). The two cases were of focal SE with impaired consciousness, and both occurred without AED withdrawal. One case was a focal SE with impaired consciousness of a frontal lobe origin: treatment with IV benzodiazepines and valproate
stopped the SE but caused valproate-induced hyperammonemia. The second case was a focal SE with impaired consciousness of temporal lobe origin, which was controlled with IV benzodiazepines, valproate, and levetiracetam. Neither of these patients had a prior history of SE. Cardiac AE occurred once (1.6% of all AEs) in the form of suspected ictal asystole lasting 18 s. The asystole occurred during a focal seizure with altered consciousness, with ictal rhythmic delta activity over the right anterobasal temporal region. The asystole was self-limited, and the patient fully recovered and was offered cardiac pacemaker implantation which he declined. Psychiatric AE in the form of a suicide attempt occurred once (1.6% of all AEs) in a patient with a past medical history of genetic generalized epilepsy, borderline personality disorder, posttraumatic stress disorder, depression, alcohol and illicit drug abuse, and several suicide attempts in the past. The suicide attempt followed a number of PNES events, and no epileptic seizures were documented during monitoring. Suicide was attempted by wrist cutting without hemodynamic compromise: the cuts were sutured, and the patient was transferred to a psychiatric hospital for further treatment.
166
F. Fahoum et al. / Epilepsy & Behavior 61 (2016) 162–167
Table 3 Significant risk factors for safety adverse events (multivariate logistic regression analysis). Factor
Significance
Noninvasive presurgical workup Invasive presurgical workup Number of events during monitoring Levetiracetam Sulthiame
.004 .002 .004 .012 .014
Odds ratio
3.149 7.722 1.027 2.319 5.133
95% C.I. for odds ratio Lower
Upper
1.431 2.114 1.008 1.207 1.384
6.929 28.205 1.046 4.456 19.038
C.I.—confidence interval.
There were no respiratory or thrombotic AEs and no deaths during the study period, and there was no need for resuscitation. The results of the univariate analysis that compared the AE and non-AE groups are shown in Table 2. There were significantly more AEs when patients were younger at disease onset (p = 0.005) and when there was a history of temporal lobe epilepsy (p = 0.46), focal seizures with altered consciousness (p = 0.008), SE (p = 0.022), VNS (p = 0.039), and intellectual disability (p = 0.016) and when there was an indication for EMU monitoring for noninvasive presurgical evaluation or intracranial monitoring (p = 0.001). Adverse events occurred more often when patients had more overall events in the EMU (p = 0.001). The patients in the AE group more frequently used the following AEDs compared with those in the non-AE group: carbamazepine (p = 0.037), levetiracetam (p = 0.004), clobazam (p = 0.008), and sulthiame (p = 0.016). The median number of AEDs in the AE group was 3 AEDs and in the non-AE group 2 AEDs. Patients with PNESs had fewer AEs compared with patients without PNESs (p = 0.013). Age, gender, duration of hospitalization and duration of monitoring, number of days to AED withdrawal and renewal, seizure frequency by history, presence of a major psychiatric comorbidity, abnormal findings on the neurological exam, and presence of a lesion on the MRI were not associated with AEs. An additional analysis to test for the frequency of specific AEs in the different paroxysmal syndromes showed no significant increase of
specific AEs in the different epilepsy syndromes (temporal, extratemporal, and generalized); however, there were fewer SCs in the group with PNES compared with the group with epilepsy (p = 0.013). In the multivariate analysis, AEs occurred more when the indication for EMU admission was a presurgical evaluation (noninvasive p = 0.004 and intracranial p = 0.002), when there were more events during monitoring (p = 0.004), and when patients were prescribed levetiracetam and sulthiame (p = 0.012 and p = 0.014, respectively) (Table 3). Repeating the univariate and multivariate analyses while including only the first admission led to similar results when including all admissions (compare Supplementary Tables 1 and 2 with Tables 2 and 3). The results of the comparison of AE rates in our study with those published in the literature are shown in Table 4. 4. Discussion Our analysis of the frequency and risk factors for AE occurrences in an adult EMU revealed that AEs occurred in nearly 9% of admissions. This rate is comparable with the AE rate reported in a recent comprehensive review of the current literature by Rheims and Ryvlin [12]. The most common AE was SC, comprising more than one-half of all AEs. The rest of the AEs in decreasing order of frequency were related to AEDs, falls and injuries, IV lines, electrodes, SE, and cardiac and psychiatric causes. Adverse events were linked to younger age at disease onset, history of SE and VNS use, and admission for presurgical evaluation, all pointing to more severe epilepsy. The association between younger age at disease onset and AEs confirms the findings of Dobesberger et al. with duration of epilepsy as an important predictive factor for AEs [1]. Furthermore, patients with mental retardation, history of temporal lobe epilepsy, and history of focal seizures with altered consciousness and those with many events during monitoring also had more AEs. These findings emphasize the importance of adhering to strict safety routines for these patients. Unlike many earlier studies, the current report included patients with a history of PNES, and the results revealed that PNESs were associated with a lower AE risk.
Table 4 The frequency of safety adverse events (AEs) reported in the literature and the current study. Study [ref.] year
Study cohort (age group)
AED withdrawal policy
[8] 2001
Presurgical workup in TLE (Adult) All EMU admissions (N/A) All EMU admissions (Adult)
AED reduction start day 2 Stop all AED on days 4–6 Individualized
[3] 2003 [9] 2009
[2] 2009 [1] 2011 [19] 2012 [6] 2012
[18] 2013 [16] 2014 Current 2016
Patients with epilepsy (adult) All EMU admissions (adult/ped) All EMU admissions (adult/ped) Presurgical workup (adult) All EMU admissions (adult) All EMU admissions (adult) All EMU admissions (adult)
Percentage of AEs in admissions (%) 24-h SC
1/2 of all AEDs reduced at admission and discontinued at 24 h Patients with history of SE — 1/4 of all AEDs daily PB was not withdrawn Individualized
4-h SC
AED-related
Electrode-related
48.5
17.8
Cardiac
Psychiatric
2.75
1.8
3 2
23
SE
0
9
2
3.7 0.2
N/A
3.2
2.2
3.4
2 39
11
5.6
0
0
2.3
Individualized AED reduction started on day 2 1/2 of single AED each day
IV-related
48.3
Individualized
1/3 of all AEDs reduced daily at admission except in history of SE or daily seizures N/A
Falls and injuries
0.6 N/A
5.4
1.3
1.1
0.8 1.1
0.6
0.4
0.2
0.2
AED — antiepileptic drug, SC — seizure cluster, SE — status epilepticus, IV — intravenous, TLE — temporal lobe epilepsy, EMU — epilepsy monitoring unit, Ped — pediatric, N/A — not available, PB — phenobarbital, BDZ — benzodiazepines.
F. Fahoum et al. / Epilepsy & Behavior 61 (2016) 162–167
Patients with AEs were treated more frequently with carbamazepine, levetiracetam, clobazam, and sulthiame. These AEDs were used in different polytherapy combinations, and their increased frequency in the AE group is probably due to more severe epilepsy and to local preferences for AED treatment. We could not find specific relations between the use or withdrawal of certain AEDs and increased risk of AEs. Adverse events in the EMU are usually linked to AED withdrawal, which is routinely undertaken to activate seizures [6,8,9]. However, whereas AED withdrawal might be crucial in some cases (e.g., for seizure characterization, presurgical evaluation, and others), it is not implemented when the indication for admission is AED adjustment or when there is a history of high seizure frequency. In our cohort, while AEDs were withdrawn in nearly one-half of all admissions, there was no significant increase in AEs in those patients compared with that in patients without AED withdrawal. There are a few possible explanations for the lack of increase of AEs in the withdrawal group: 1. our implementation of slow AED withdrawal starting from the second day or later since the real seizure frequency is sometimes underreported by patients [13,14]; 2. withdrawing of AEDs sequentially and never more than one AED at a time; 3. refraining from AED withdrawal in patients with a history of SE and/or high seizure frequency, with the exception of those admitted for presurgical evaluation and those who experienced no seizures during the first days in the EMU; 4. keeping patients with high seizure frequency during monitoring for medical observation in the EMU for an extra one or more days (without EEG monitoring); and 5. encouraging the patient's companion to stay alert overnight in the patient's room, as this might reduce seizures missed by the staff [15]. Another possible explanation for the lack of difference in the AE rate between the withdrawal and no-withdrawal groups is the higher seizure frequency by history in the latter group, pointing to more severe disease. This finding stresses that AEs occurring in the EMU are not necessarily related to AED withdrawal, a point further supported by the lack of correlation between the first day of AED withdrawal and the day of AE occurrence. The distribution of AEs throughout the length of stay at the EMU showed that 93% of AEs occurred during 6 days of monitoring, whereas 91% of AEs occurred during the first 4 days of monitoring in an earlier study [1]. It is difficult to compare published studies, as often different patient populations were studied (e.g., only patients with epilepsy, only presurgical evaluation), and different categories of AEs were defined; nevertheless, comparing our AE rates with those published in the literature (Table 4), we found 4-h SCs in 5.4% of all EMU admissions, in contrast to 9–18% of patients in the literature [2,3,6]. The events of SE recorded in 0.4% of admissions were comprised of events of focal SE, while the reported rates of convulsive SE in comparable cohorts were 0.8–3% [1,3,16]. In agreement with other studies, all falls occurred when our patients were outside the bed [17,18], although we had fewer fall events (n = 6, 1.1% of admissions) than the 0.6–5.6% reported in comparable cohorts [1,9,16,18,19]. We had no events of postictal psychosis and fewer psychiatric AEs than reported by others in the literature [1,2], but on the other hand, we had relatively high AED-related AEs compared with Dobesberger et al. [1]. In the latter study, patients with minor or major psychiatric comorbidities had a more than 16-fold increased risk of sustaining psychiatric AEs during or after monitoring; however, in 13 of 20 patients, the psychiatric AE occurred after discharge from EMU [1]. Two main reasons could explain the difference between our study and the study of Dobesberger et al. regarding the frequency of psychiatric complications: we only included major psychiatric comorbidities, and we did not have the data of possible complications occurring after the discharge from the EMU. Recently published data showed that patients with epilepsy have an increased risk of thromboembolism [20]. Interestingly, there were no thrombotic events in our cohort, despite the fact that the patients did not receive prophylactic anticoagulation and were relatively immobile during monitoring. The main limitations of our study are its retrospective design and the lack of follow-up data on occurrences of AEs in the days following
167
discharge from the EMU, as AEs have been reported in that period [1]. The strength of this study is the inclusion of all consecutive patients admitted to an adult EMU and not limiting the study to a specific patient subgroup, therefore providing a comprehensive review of safety issues in adult EMUs. In summary, we found that AEs are not uncommon in the EMU and that they tend to occur more frequently in patients with more severe epilepsy and intellectual disability and those undergoing presurgical evaluations. They are less common in patients with PNESs. Seizure cluster was the most common AE, comprising nearly one-half of all AEs. While AEs are known to occur in patients undergoing and not undergoing AED withdrawal, we believe that our careful and selective AED withdrawal policy in addition to other safety measures was related to no increase in AE risk in patients undergoing AED withdrawal. This work further emphasizes the need for internationally standardized protocols to improve the patient safety in the EMU. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.yebeh.2016.06.002.
Acknowledgments The authors thank Ms. Ilana Gelernter from the Department of Statistics and Operations Research at Tel Aviv University for her help with statistical analysis and Ms. Esther Eshkol for editorial assistance.
References [1] Dobesberger J, Walser G, Unterberger I, Seppi K, Kuchukhidze G, Larch J, et al. VideoEEG monitoring: safety and adverse events in 507 consecutive patients. Epilepsia 2011;52:443–52. [2] Noe KH, Drazkowski JF. Safety of long-term video-electroencephalographic monitoring for evaluation of epilepsy. Mayo Clin Proc 2009;84:495–500. [3] Rose AB, McCabe PH, Gilliam FG, Smith BJ, Boggs JG, Ficker DM, et al. Occurrence of seizure clusters and status epilepticus during inpatient video-EEG monitoring. Neurology 2003;60:975–8. [4] Spritzer SD, Riordan KC, Berry J, Corbett BM, Gerke JK, Hoerth MT, et al. Fall prevention and bathroom safety in the epilepsy monitoring unit. Epilepsy Behav 2015;48:75–8. [5] Kandler R, Lai M, Ponnusamy A, Bland J, Pang C. The safety of UK video telemetry units: results of a national service evaluation. Seizure 2013;22:872–6. [6] Di Gennaro G, Picardi A, Sparano A, Mascia A, Meldolesi GN, Grammaldo LG, et al. Seizure clusters and adverse events during pre-surgical video-EEG monitoring with a slow anti-epileptic drug (AED) taper. Clin Neurophysiol 2012;123:486–8. [7] Mohan KK, Markand ON, Salanova V. Diagnostic utility of video EEG monitoring in paroxysmal events. Acta Neurol Scand 1996;94:320–5. [8] Yen DJ, Chen C, Shih YH, Guo YC, Liu LT, Yu HY, et al. Antiepileptic drug withdrawal in patients with temporal lobe epilepsy undergoing presurgical video-EEG monitoring. Epilepsia 2001;42:251–5. [9] Moien-Afshari F, Griebel R, Sadanand V, Vrbancic M, Hernandez-Ronquillo L, Lowry N, et al. Safety and yield of early cessation of AEDs in video-EEG telemetry and outcomes. Can J Neurol Sci 2009;36:587–92. [10] Proposal for revised classification of epilepsies and epileptic syndromes. Commission on Classification and Terminology of the International League Against Epilepsy. Epilepsia 1989;30:389–99. [11] Lowenstein DH, Bleck T, Macdonald RL. It's time to revise the definition of status epilepticus. Epilepsia 1999;40:120–2. [12] Rheims S, Ryvlin P. Patients' safety in the epilepsy monitoring unit: time for revising practices. Curr Opin Neurol 2014;27:213–8. [13] Blum DE, Eskola J, Bortz JJ, Fisher RS. Patient awareness of seizures. Neurology 1996; 47:260–4. [14] Hoppe C, Poepel A, Elger CE. Epilepsy: accuracy of patient seizure counts. Arch Neurol 2007;64:1595–9. [15] Dericioglu N, Albakir M, Saygi S. The role of patient companions in long-term videoEEG monitoring. Seizure 2000;9:124–7. [16] Sauro KM, Macrodimitris S, Krassman C, Wiebe S, Pillay N, Federico P, et al. Quality indicators in an epilepsy monitoring unit. Epilepsy Behav 2014;33:7–11. [17] Atkinson M, Hari K, Schaefer K, Shah A. Improving safety outcomes in the epilepsy monitoring unit. Seizure 2012;21:124–7. [18] Pati S, Kumaraswamy VM, Deep A, Chung SS, Plueger M, Kiyota G, et al. Characteristics of falls in the epilepsy monitoring unit: a retrospective study. Epilepsy Behav 2013;29:1–3. [19] Spanaki MV, McCloskey C, Remedio V, Budzyn D, Guanio J, Monroe T, et al. Developing a culture of safety in the epilepsy monitoring unit: a retrospective study of safety outcomes. Epilepsy Behav 2012;25:185–8. [20] Martz GU, Wilson DA, Malek AM, Selassie AW. Risk of venous thromboembolism in people with epilepsy. Epilepsia 2014;55:1800–7.