New-onset refractory status epilepticus (NORSE) — The potential role for immunotherapy

Epilepsy & Behavior 47 (2015) 17–23

Contents lists available at ScienceDirect

Epilepsy & Behavior journal homepage: www.elsevier.com/locate/yebeh

New-onset refractory status epilepticus (NORSE) — The potential role for immunotherapy Ayaz M. Khawaja a,⁎, Jennifer L. DeWolfe a, David W. Miller b, Jerzy P. Szaflarski a,c a b c

Department of Neurology, University of Alabama at Birmingham Hospital, Birmingham, AL, USA Department of Anesthesiology, University of Alabama at Birmingham Hospital, Birmingham, AL, USA University of Cincinnati Department of Neurology, Cincinnati, OH, USA

a r t i c l e

i n f o

Article history: Received 3 March 2015 Revised 20 April 2015 Accepted 21 April 2015 Available online xxxx Keywords: NORSE Status epilepticus Immunotherapy Outcome Chemotherapy

a b s t r a c t New-onset refractory status epilepticus (NORSE) is defined as a state of persistent seizures with no identifiable etiology in patients without preexisting epilepsy that lasts longer than 24 h despite optimal therapy. Management of NORSE is challenging, and the role of immunotherapy (IT) is unclear. We identified patients fulfilling the criteria for NORSE at a single institution. These patients were described, analyzed, and compared with NORSE cases available from the literature. Finally, a pooled analysis of available case series was conducted to compare the outcomes in patients who received IT with those not treated with IT during the course of NORSE in order to generate hypotheses for further research. In our case series, NORSE was diagnosed in 11 patients (9 females) with a mean age of 48 years and a mean duration of 54.4 days. Autoantibodies were identified in 7 patients, of which anti-GAD (glutamic acid decarboxylase) and anti-NMDAR (N-methyl-D-aspartate receptor) were most frequent. Of the 11 patients, 8 were treated with IT (intravenous steroids, immunoglobulins, plasmapheresis, or a combination), and 4 received chemotherapy. Of the 8 patients treated with IT, 6 had favorable outcomes (defined as any outcome other than death, vegetative state, or inability to take care of oneself) compared with 0 out of 3 patients who did not receive IT. Difference in outcomes was significant (p = 0.026). Pooled analysis of all identified case series, including ours, showed a statistically significant effect (p = 0.022), with favorable outcomes in 42% of the patients who received any IT compared with 20% in those who did not. In all patients with refractory SE and negative comprehensive investigations, a diagnosis of NORSE should be considered. This would aid planning for early immunotherapy. Currently, only Class IV evidence for the use of immunotherapy in NORSE is available. Prospective multicenter studies are necessary to assess the true efficacy of IT in NORSE. © 2015 Elsevier Inc. All rights reserved.

1. Introduction Status epilepticus (SE) is defined as a “seizure lasting more than 30 min in duration or the occurrence of two or more seizures without return to baseline mental status” [1]. When either clinical or

Abbreviations: HSV, herpes simplex virus; VZV, varicella zoster virus; CMV, cytomegalovirus; EBV, Epstein–Barr virus; WNV, West Nile virus; HIV, human immunodeficiency virus; JCV, JC virus; PCR, polymerase chain reaction; GAD, glutamic acid decarboxylase; NMDAR, N-methyl-D-aspartate receptor; VGKC, voltage-gated potassium channel; VGCC, voltage-gated calcium channel; ANNA, antineuronal nuclear antibody; AGNA, antiglialnuclear antibody;PCA, Purkinje cell antibody; CRMP, collapsin response mediator protein; TPO, thyroid peroxidase; ACE, angiotensin-converting enzyme; VDRL, Venereal Disease Research Laboratory; IT, Immunotherapy; IVSM, intravenous solumedrol; IVIg, intravenous immunoglobulin; PLEX, plasma exchange; NORSE, new-onset refractory status epilepticus; SE, status epilepticus; AEDs, antiepileptic drugs; cEEG, continuous electroencephalography; LZP, lorazepam; fPHT, fosphenytoin; LCM, lacosamide; VPA, valproic acid; LEV, levetiracetam; FLAIR, fluid-attenuated inversion recovery. ⁎ Corresponding author at: University of Alabama at Birmingham Hospital, 1720 7th Avenue South, Birmingham, AL 35233, USA. Tel.: +1 205 975 3509. E-mail address: [email protected] (A.M. Khawaja).

http://dx.doi.org/10.1016/j.yebeh.2015.04.054 1525-5050/© 2015 Elsevier Inc. All rights reserved.

electrographic seizures continue after initial treatment with first- and second-line antiepileptic drugs, the episode is often termed refractory status epilepticus (RSE) [2], and such prolonged SE contributes to increased mortality [3]. Common identifiable causes of RSE include cerebrovascular disease, traumatic brain injury, central nervous system infections, autoimmune conditions, and low serum levels of antiepileptic drugs (AEDs) [1]. With control of RSE, good outcomes (as identified by modified Rankin scale score of 3 or less) may be expected, especially with normal neuroimaging and a reactive EEG at onset [4]. When all investigations fail to find an etiology of RSE in patients without prior history of epilepsy, the RSE is often called new-onset refractory status epilepticus (NORSE). Limited case series have documented several similarities between patients with NORSE including female gender, young age, extensive negative workup, and no history of epilepsy [5]. Limited evidence from case series provides data regarding the use of early immunotherapy with or without chemotherapy for achieving control of NORSE when these therapies are combined with or added to AEDs (Table 1). In the early phase of SE, the etiology typically remains unknown as the results of the majority of investigations are not

18

A.M. Khawaja et al. / Epilepsy & Behavior 47 (2015) 17–23

Table 1 Outcomes in patients receiving or not receiving immunotherapy as reported in various published case series. Author

No. of patients

Age range

Baxter et al. [6]

6

0–6

Van Lierde et al. [7]

6

18–30

Wilder-Smith et al. [5] Costello et al. [8]

7 6

20–52 24–36

Gall et al. [9]

5

22–34

Li et al. [10]

3

39–51

Kaneko et al. [11]

6

15–38

Mikaeloff et al. [12] Kramer et al. [13]

14 8

4–11 2.5–15

Sahin et al. [14] Our series

22 11

b1–18 21–90

IT tried

Favorable outcome

IVSM 5/6 IVIg 2/6 IVSM 2/6 PLEX 2/6 IVIg 3/7 IVSM 2/6 IVIg 1/6 PLEX 1/6 IVSM 3/5 IVIG 2/5 AZT 1/5 IVSM 2/3 IVIG 1/3 PLEX 3/3 “First-line IT” 4/6 CPH 2/6 IVSM 6/14 IVIg 7/8 IVSM 5/8 IVSM 1/22 IVSM 7/11 IVIG 7/11 PLEX 4/11 CH 3/11

0/5

IT not tried 1/6

Favorable outcome 0/1

0/2

4/6

1/4

0/3 2/2

4/7 4/6

0/4 2/4

3/3

2/5

0/2

2/3

0/3

0/0

2/4

2/6

0/2

0/6 4/8b

8/14 0/8

0/8a 0/0

0/1 6/8

21/22 3/11

7/21c 0/3

IT — immunotherapy (intravenous immunoglobulin [IVIg]; intravenous methylprednisolone [IVSM]; plasmapheresis [PLEX]; immunosuppressants [IMSs]; azathioprine [AZT]; cyclophosphamide [CPH]; chemotherapy [CH]). Poor outcome is defined as death, vegetative state, or inability to take care of oneself; favorable outcome is defined as any outcome other than death, vegetative state, or inability to take care of oneself. a All patients had poor performance on neuropsychological testing, so this was interpreted as an unfavorable outcome. b There was no detailed neuropsychological assessment reported. Outcome was interpreted after reviewing comments regarding overall cognition and frequency of seizures. c These 7 patients were interpreted as “returned to baseline” in the paper; thus, they were designated a favorable outcome.

available to guide treatment decisions. In this case series, we describe our experiences in managing 11 patients with suspected NORSE. We then review the available literature in order to provide guidance regarding the treatment and discuss the possible role of immunotherapy in the management of NORSE.

2. Methods 2.1. Patient selection We conducted a retrospective chart review of all patients admitted to the intensive care unit between 1/1/2009 and 09/30/2014 who received continuous video-EEG (cEEG) monitoring (CPT 95951); were diagnosed with seizures (any ICD-9 345.XX or 780.39); and received propofol, midazolam, phenobarbital, pentobarbital, or ketamine infusions during the course of their ICU stay (appropriate National Drug Code (NCD)). From the pool of approximately 100 patients who fulfilled the above criteria, we further selected patients based on a predefined definition of NORSE: 1) transient or sustained focal or generalized ictal epileptiform activity comprising spikes, sharp waves, spike-and-slow-wave complexes, and polyspike-andslow-wave complexes for at least seven days when detected either during sedation or following discontinuation of sedation; 2) lack of response to 3 or more AEDs (including initial treatment with a benzodiazepine); 3) no history of epilepsy or seizures; and 4) unknown etiology of seizures for at least seven days. A seven-day period was considered an inclusion criterion in order to allow sufficient time for initial investigations to be reported, as standard testing including initial reports of bacterial, fungal, and viral testing typically takes up to 7 days, and to permit exclusion of patients with readily identifiable etiologies such as bacterial or viral infections, traumatic brain injury, and stroke. The Institutional Review Board at the University of Alabama at Birmingham approved this study.

2.2. Data collection and analysis Demographic data collection included age, gender, symptoms present upon initial presentation (fever, encephalopathy, seizures, constitutional symptoms, and comatose state), and medical comorbidities. Continuous video-EEG reports were generated by a board-certified epileptologist on a daily basis. Patients with any transient or sustained epileptiform activity (as defined in Section 2.1) persisting more than seven days were included. As indicated by clinical history and presentation, all patients received metabolic and infectious investigations including serum; urine and/or cerebrospinal fluid for cell count; electrolytes; glucose; protein; cultures (bacterial, fungal, and, if indicated, viral); viral polymerase chain reaction (HSV and, if indicated, VZV, CMV, EBV, WNV, arbovirus, enterovirus, HIV, and JCV); paraneoplastic autoantibodies (Ma2, NMDAR, Hu, Yo, GAD65, amphiphysin, VGKC, VGCC, ANNA-1 /2/3, AGNA-1, PCA-1/2, CRMP-5, and striated muscle); other autoantibodies (anti-TPO and antithyroglobulin); VDRL; serum immunoglobulins (EBV, VZV, WNV, HSV, CMV, and, if indicated, toxoplasmosis); and, if suggested by a patient's history, Cerebrospinal fluid (CSF) flow cytometry, CSF cytology, ACE, and heavy metal screen. All patients underwent 3-Tesla brain MRI (Fig. 1). Specific sequences acquired on MRI for all patients included T1- and T2-weighted, diffusion-weighted (DWI), apparent-diffusion coefficient, susceptibility-weighted, T2-FLAIR, and intracranial and extracranial angiography images. The names, number, and duration of all antiepileptic medications, sedative infusions (propofol, midazolam, pentobarbital, phenobarbital, ketamine, and lorazepam), and antimicrobial agents (vancomycin, ceftriaxone, cefepime, piperacillin–tazobactam, azithromycin, micafungin, penicillin G, tobramycin, and aztreonam) were recorded. Immunotherapies (ITs) included high-dose intravenous (IV) methylprednisolone (IVSM — 1 g daily for five days), IV immunoglobulins (IVIg — 0.4 g/kg daily for five days), plasmapheresis (five sessions), and chemotherapy (rituximab, cyclophosphamide, cyclosporine, tacrolimus, and everolimus). Outcomes for all patients were recorded as discharged to home, inpatient

A.M. Khawaja et al. / Epilepsy & Behavior 47 (2015) 17–23

Patient 2

Patient 7

Patient3

Patient 9

Patient 4

Patient 10

19

Patient 6

Patient 11

Fig. 1. Brain MRI scans for 8/11 patients. Note that the images for patients P1, P5, and P8 are not included as their scans were normal. All images show a T2-weighted FLAIR sequence except for P3 which shows a diffusion-weighted imaging sequence.

rehabilitation, nursing home, long-term acute care hospital and another hospital, or death. Identified cases of NORSE were then divided into two groups based on whether IT was administered (IT vs. no IT) and on outcome profile (favorable vs. unfavorable — see below for definition) and compared with respect to age, duration of NORSE, number of AEDs and antimicrobials used, and antibodies discovered. As indicated in the legend of Table 1, a pooled analysis of the reports from the literature was conducted to assist with addressing the question whether immunotherapy may be beneficial for the management of seizures/NORSE. Studies were included after carrying out a search of keywords “NORSE”, “refractory seizures”, “immunotherapy”, “immune-mediated therapies”, “autoimmune epilepsy” and/or “refractory status epilepticus” within PubMed and Google Scholar. Both adult and pediatric studies that were either published or presented as an abstract were screened. Pediatric studies were considered on the premise that the syndrome described in the screened studies matches the description of NORSE in adults and that IT was used for treatment, thereby allowing comparison between groups of patients who did or did not receive IT [3]. Only studies published in the English language were considered. Once a study abstract was identified from the electronic search, a full manuscript was requested from the respective journal. Single case studies were excluded as were series with none of the patients receiving IT. Patients were divided into two groups — patients who received IT and those that did not for the treatment of NORSE; their clinical outcomes were compared. For the purpose of this analysis, poor outcome is defined as death, vegetative state, or inability to take care of oneself, while favorable outcome was defined as any outcome other than death, vegetative state, or inability to take care of oneself. As a general rule, comments stating moderate-to-severe cognitive deficits, unemployment, uncontrolled epilepsy, and bedbound status were used to assign a poor outcome. Outcomes in pediatric case series were assessed based on the results of neuropsychological assessments and comments regarding overall cognition and seizure refractoriness. A favorable outcome was usually only given if explicitly mentioned by the authors. If follow-up data were available, they were used for outcome classification. In all other cases, discharge status was used for classification of outcome.

3. Results 3.1. Demographics and etiology A total of 11 patients (9 females) were identified from our institution during the study period; in all patients, etiology of status epilepticus was undetermined prior to day 7 (Table 2). The mean age was 48 (range = 21–90). The mean duration of NORSE was 54.4 days (range = 12–110). Nonspecific encephalopathy was the most common presentation (N = 9), followed by generalized convulsions (GTCs; N = 7) and fever (N = 4). One patient (P8) had only nonconvulsive SE. Three patients had known underlying conditions present at the time of initial presentation possibly contributing to the presentation such as postoperative state (N = 2) (one following liver transplant surgery and the other following general surgery) and shingles (N = 1). Autoantibodies were detected in seven patients with anti-GAD65 present in N = 3, anti-NMDAR alone in N = 2, anti-NMDAR and anti-VGCC in N = 1, and anti-VGKC in N = 1. Central nervous system (CNS) infection was suspected in one patient based on the presence of shingles, but CSF VZV PCR test was negative. One patient (P6) eventually developed posterior reversible leukoencephalopathy syndrome (PRES). Three cases remained cryptogenic, with one patient (P5) lacking paraneoplastic antibody analysis; another patient (P8) having no paraneoplastic antibody and CSF analysis, and the third patient (P11) lacking CSF studies.

3.2. Diagnostic investigations Cerebrospinal fluid was obtained in 9 patients (Table 3). In one patient (P11), lumbar puncture was technically challenging and not successful, whereas the reason for not performing lumbar puncture in the other patient (P8) remains unknown. Glucose was within normal range for all of the 9 patients. Cerebrospinal fluid pleocytosis was present in 6 patients (range = 7–35 cells/mm3), and 5 patients had abnormally elevated protein level (range = 55–91 mg/dl). Testing for microorganisms in CSF was negative in all patients whose CSF was collected.

20

A.M. Khawaja et al. / Epilepsy & Behavior 47 (2015) 17–23

Table 2 Demographics, etiology, and presentation.

P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 P11

Age

Gender

NORSE (days)

Known underlying conditions

Etiology

Seizure type

Presentation

28 22 61 53 21 50 48 84 50 21 90

F F F F F F F F F M M

110 110 70 40 90 80 16 16 39 12 15

None (ovarian teratoma discovered later) None None Shingles None Liver transplant None Postoperative state Schizophrenia None None

Anti-NMDAR Anti-GAD Anti-GAD Anti-VGKC, VZV Cryptogenic PRES Anti-NMDAR Cryptogenic Anti-GAD Anti-VGCC, anti-NMDAR, echovirus Cryptogenic

GTC GTC GTC GTC GTC GTC GTC NCSE GTC GTC GTC

AMS, Szs AMS, fever, Szs AMS AMS, fever, Szs AMS, fever AMS Szs AMS, trauma (not TBI) AMS, fever, Szs AMS, Szs, agitation Szs

AMS = altered mental status; Szs = seizures; TBI = traumatic brain injury; NMDAR = N-methyl-D-aspartate receptor; GAD = glutamic acid decarboxylase; VGKC = voltage-gated potassium channels; PRES = posterior reversible leukoencephalopathy; VZV = varicella zoster virus; HSV = herpes simplex virus.

All patients underwent at least one brain MRI (Fig. 1; Table 3) using gadolinium contrast when not contraindicated. Three patients had normal brain MRI, and 8 had at least one abnormality, with T2-FLAIR hyperintensities within the limbic system structures being most common (4/8), followed by hyperintense signals within the insular cortex (2/8), basal ganglia (1/11), occipital/parietal cortex (1/11), subarachnoid space (1/11), and diffuse cortical diffusion restriction on DWI (1/11). One patient had T2-FLAIR hyperintensities within the left occipital and parietal lobes, leading to a diagnosis of unilateral PRES (P6). One patient (P11) had chronic bilateral subdural hygromas along with bilateral hippocampal T2-FLAIR hyperintensities and no evidence of an acute hemorrhage. Another patient (P9) had T2-FLAIR hyperintensities distributed diffusely within the subarachnoid space attributed to hyperoxygenation secondary to oxygen supplementation via mechanical ventilation. 3.3. Therapeutic interventions All patients were treated with multiple antimicrobials throughout the course of their hospitalization (Table 4). The use of antimicrobials was triggered by the discovery of fever on initial presentation (defined as oral temperature N 100.8 °F), preceding febrile illness, or CSF analysis showing a white cell count N 5/mm3. Vancomycin was used in all patients, with acyclovir and piperacillin/tazobactam being the other commonly used antimicrobials. All patients were treated with at least 2 or more antiepileptic drugs (AEDs; range = 2–8). Ten patients were prescribed AEDs at the time of discharge. Antiepileptic drugs were withdrawn in one patient (P11) after transitioning care to the palliative medicine service. The most commonly used AEDs were PHT, LEV, LCM, and VPA. One patient (P2) was treated with the ketogenic diet for forty days. At the time of diagnosis of seizures/SE on clinical or cEEG basis, all patients received intravenous LZP along with an infusion of

fPHT at a standard dose of 20 mg PE/kg. Based on the initial response measured via cEEG, additional AEDs were added, most commonly LCM and LEV. All patients received mechanical ventilation and continuous sedative infusion with either propofol or midazolam (9 and 8 patients, respectively), followed by ketamine (7 patients) and pentobarbital (6 patients). Two patients were sedated with fentanyl infusion (P8 and P9). Information regarding sedatives was incomplete in one patient. Seven patients were treated with high-dose IVSM for at least 5 days followed by an oral prednisone taper. Seven patients received IVIg infusion for 5 days. Six patients received both IVSM and IVIg, with three also receiving PLEX for a total of five treatments. One patient received both IVIg and PLEX but no IVSM (P10). Two patients were treated with rituximab infusion when concurrent immunotherapy failed (P2 and P10), with one patient (P2) also receiving cyclophosphamide. Patient 6 was already on treatment with rituximab, cyclosporine, prednisone, and plaquenil prior to the development of NORSE for immune suppression after liver transplant surgery. Time to initiation of IT ranged from 5 to 21 days after the development of NORSE. One patient (P1) received oophorectomy for anti-NMDAR encephalitis with subsequent resolution of NORSE. Another patient with antiNMDAR encephalitis had a total abdominal hysterectomy with bilateral salpingo-oophorectomy for uterine fibroids and birth control five years prior to the onset of NORSE; this patient (P7) succumbed to NORSE. 3.4. Outcomes Four patients were discharged home, three patients expired, two patients were discharged to nursing home, one patient was discharged to another hospital, and one patient was discharged to an acute care facility for ventilator weaning (Table 4). Of the eight patients who received any immunotherapy including IVSM, IVIg, PLEX, and/or immunosuppressants, four patients were discharged home. Of our patients,

Table 3 Results of various investigations.

P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 P11

Positive serum labs

CSF protein

CSF WBC

Other CSF tests

Results of brain MRI

Anti-NMDAR; anti-TPO EBV IgG Negative Anti-VKGC; VZV IgM and IgG “PNPP not tested” Negative Anti-NMDAR “PNPP not tested” Negative Anti-VGCC; echovirus Negative

36 72 55 82 37 91 55 N/A 27 19 N/A

7 21 3 13 12 2 35 N/A 3 18 N/A

Neg Neg Neg Neg Neg Neg Neg N/A Neg Neg N/A

Normal Limbic and insular T2F Cortical DWI Limbic, insular T2F Normal Occipital T2F Subcortical T2F Normal Subarachnoid T2F Limbic T2F Limbic T2F

RPR — rapid plasma reagin; anti-TPO — antithyroid peroxidase; anti-VGKC — antivoltage-gated potassium channel; anti-VGCC — antivolate-gated calcium channel; anti-NMDAR — anti-Nmethyl-D-aspartate receptor; VZV — varicella zoster virus; DWI — diffusion-weighted image; T2F — T2 flair; Abs — antibodies; PNPP — paraneoplastic antibody panel; Neg — indicates negative results for testing of all tested antibodies and microorganisms.

A.M. Khawaja et al. / Epilepsy & Behavior 47 (2015) 17–23

21

Table 4 Treatment and outcomes.

P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 P11

nAED

nAED at DC

nIVSM

nIVIg

nPLEX

Chemotherapy

Time to initiation of any IT (days)

n-antimicrobials

Sedatives

Disposition at DC

5 8 6 6 6 6 7 5 3 2 4

4 3 4 5 6 3 7 4 3 1 0

1 1 0 1 1 1 1 0 1 0 0

2 2 0 1 1 1 0 0 1 1 0

1 2 0 0 1 0 0 0 0 1 0

None RIT, CPH None None None TAC, EVE, CYC None None None RIT None

21 12 N/A 15 5 PTAa 10 N/A 17 7 N/A

9 7 4 4 5 12 5 2 11 5 6

PRO, MID, PEN, KET PRO, MID, PEN, KET PRO, MID, PEN, KET PRO, MID PRO, MID, PEN, KET PRO, LOR, PEN, KET PRO, MID, PEN, KET PRO, MID, FEN MID, KET, FEN Unknown PRO, MID

Home Home Nursing home Death Home Home Death LTAC Nursing home Other hospitals Death

nAED — number of AEDs; nIVSM — number of IVSM treatments; nIVIg — number of IVIg treatments; nPLEX — number of plasmapheresis treatments; PRO — propofol; MID — midazolam; PEN — pentobarbital; KET — ketamine; LOR — lorazepam; IVSM — steroids; IVIG — intravenous immunoglobulins; PLEX — plasma exchange; RIT — rituximab; CYC — cyclosporine; CPH — cyclophosphamide; TAC — tacrolimus; EVE — everolimus; DC — discharge; AEDs — antiepileptics; LTAC — long-term acute care; IT — immunotherapy; N/A — not applicable; PTA — prior to admission. a These patients were already on chemotherapy before the development of NORSE.

those receiving any IT had better outcomes (6/8 patients were designated a favorable outcome) compared with those not receiving IT (0/3 had a favorable outcome), which was statistically significant (chi-square; p = 0.026). The underlying causes of death were sepsis due to Pseudomonas aeruginosa (P7) and withdrawal of medical care (P4 and P11). Detailed comparison of our patients is provided in Table 5. Overall, patients who received any IT were younger (average = 36.6 years vs. 78.3; p = 0.003) and had better outcomes (p = 0.026). Duration of NORSE, number of AEDs, and antimicrobials used were not significantly different. Patients who had favorable outcome were younger compared with those with unfavorable outcome (average = 32 years vs. 67; p = 0.006) and were treated with a higher number of antimicrobials (p = 0.025). Duration of NORSE and treatments with AEDs were not significantly different between the two groups. 3.5. Results of pooled analysis Over 150 abstracts were screened, with the majority excluded as they either described an entity different from NORSE, described patients that did not receive IT, or were case reports (less than 2 patients). Eleven case series (see Table 1) were identified, four of which are pediatric case series and one was an unpublished abstract presentation [11]. Although the pediatric case series did not contain the keyword “NORSE”, the syndrome described matched that of NORSE seen in adults. As ascertaining

a favorable outcome in pediatric case series was challenging in comparison with adult series, we have purposefully biased towards assigning a poor outcome regardless of use of IT, unless explicitly mentioned by authors whether the child was seizure-free and/or exhibited favorable cognitive outcomes. All included papers described two categories of patients: one treated with IT and the other not treated with IT. The total number of patients identified from the case series, including ours, was 94, with 45 receiving any IT. Of patients receiving any IT, 19/45 (42%) had favorable outcomes. Of patients not receiving IT (49), only 10 (20%) had favorable outcomes. The difference in outcomes between groups was statistically significant (chi-square; p = 0.022). 4. Discussion In this study, we analyzed a series of NORSE cases from our institution and combined it with several case series available from the literature in order to derive guidance regarding the treatment of patients with NORSE with particular attention to the role of immunosuppression and immunomodulation. In our case series, favorable outcomes were observed in patients receiving any IT compared with those untreated. Patients treated with IT were, on average, younger; there was a trend towards a longer duration of NORSE and use of a higher number of AEDs and antimicrobials, but this was not statistically significant (Table 5). Similarly, patients with favorable outcomes were younger and treated with a higher number of antimicrobials. A similar trend (not statistically

Table 5 Patient treatment data from our institution divided by treatment (IT vs. no IT) and outcome (favorable vs. unfavorable).

Number Average age +/− SD Average duration of NORSE +/− SD (days) Average total number of AEDs administered +/− SD Average total number of antimicrobials +/− SD Favorable outcome Unfavorable outcome

Number Average age +/− SD Average duration of NORSE +/− SD (days) Average total number of AEDs administered +/− SD Average total number of antimicrobials +/− SD Antibodies discovered (n = patients) a b

Patients receiving IT

Patients not receiving IT

p-Value

8 36.6 +/− 14.8 62.1 +/− 40.3 5.4 +/− 2.0 7.3 +/− 3.1 6 2

3 78.3 +/− 15.3 33.7 +/− 31.5 5.0 +/− 1.0 4.0 +/− 2.0 0 3

0.003a 0.304 0.754 0.126 0.022b

Patients with favorable outcome

Patients with unfavorable outcome

6 32.0 +/− 14.2 73.5 +/− 39.9 5.0 +/− 2.2 8.2 +/− 3.0 4 (one anti-NMDAR, two anti-GAD, one anti-VGCC)b

5 67.2 +/− 18.8 31.4 +/− 24.0 5.6 +/− 1.1 4.2 +/− 1.5 3 (one anti-GAD, one anti-NMDAR, one anti-VGKC)

Denotes a statistically significant p-value below 0.05. Favorable and unfavorable outcomes for patients receiving IT and for patients not receiving IT were compared using a chi-square test.

0.006a 0.067 0.595 0.025a

22

A.M. Khawaja et al. / Epilepsy & Behavior 47 (2015) 17–23

significant) was observed with prolonged duration of NORSE. Patients with outcomes categorized as favorable or unfavorable had similar rates of an autoimmune etiology. However, patients not treated with IT tended to be older, demonstrating that age influenced treatment planning more than suspected autoimmune etiology. These patients also had poorer outcomes as older age correlates with poor prognosis in status epilepticus [15]. Better outcomes were also observed in pooled analysis (p = 0.022). The studies included are highlighted in Table 1. A retrospective analysis of five patients with NORSE by Gall et al. showed that three out of five patients who received immunotherapy (steroids, IVIg, and/or azathioprine) had seizures subsequently controlled on AEDs, with only one patient developing mild cognitive decline [9]. The fourth patient died (no IT administered), and the fifth patient's outcome was unknown. One patient had anti-TPO antibodies discovered in serum. These authors argued for an autoimmune basis for NORSE and discussed early IT in such patients [11]. Establishment of the underlying etiology of NORSE in the early stages may be challenging because of limited diagnostic information that is immediately available. Ultimately, exhaustive investigations may identify an etiology, as outlined in our study where these were initially negative in two patients (P2 and P3); however, repeated selective investigations showed positive serum titers of anti-GAD in both patients, prompting initiation of immunotherapy in P2. Unfortunately, the window of early intervention was missed in both patients. True cryptogenic cases are expected to be rare, as in our study where only one patient was truly cryptogenic whereas the other two had incomplete investigations. The role of plasmapheresis for the management of NORSE was emphasized by Li et al. [10] in which three patients were treated with PLEX, including one who received IVSM and the other two who received both IVSM and IVIg. Two patients had good outcomes. Seizures were controlled in the third patient treated with IVSM, IVIg, and PLEX; however, the patient's condition deteriorated secondary to intestinal ischemia and necrosis eventually leading to death. The authors suggested that PLEX and not IVSM or IVIg controlled the seizures; however, the onset of action from other ITs may have been delayed. The role of cyclophosphamide was elaborated by Kaneko et al. [11]. Four patients received IT, of which two patients treated with cyclophosphamide improved overall outcome despite persistent seizures. The role of cyclophosphamide in the treatment of autoimmune epilepsy was discussed in a case report that described a patient with refractory status epilepticus secondary to anti-GAD antibodies who improved after infusion of cyclophosphamide and relapsed into refractory seizures due to poor treatment compliance [16]. Cyclophosphamide, in conjunction with rituximab, was also administered in our patient (P2), who as of this writing is experiencing occasional breakthrough seizures, has mild cognitive deficits, and is on maintenance rituximab infusions. Other examples of favorable outcomes from IT include the case series published by Costello et al. [8] in which two patients were treated with IVSM and IVIg. One patient became seizure-free, and the other patient was able to return to employment despite intractable multifocal epilepsy. Kramer et al. [13] demonstrated the benefit of IVIg and IVSM in a pediatric population of patients presenting with a syndrome similar to NORSE. All patients (8) received some form of IT (7 received IVIg; 5 received IVSM). Four patients had a favorable outcome characterized by no seizures in two, normal cognition with few seizures in one, and mild cognitive delay with no seizures in the fourth patient. Two patients died, whereas the other two had refractory seizures with neurological deficits. In a retrospective review of 29 patients seen at Mayo Clinic in Rochester, MN with epilepsy of a proven or suspected autoimmune etiology, it was determined that 18 patients responded to IVSM, IVIg, or both [17]. Ten patients reportedly became seizure-free; however, patients with NORSE were not included in this case series. These and other studies have led to the recognition of the role of an autoimmune etiology in refractory epilepsy that could be mediated by various autoantibodies, some of which are still awaiting discovery.

A beneficial response to IT was not universally reported. Van Lierde et al. published a report of 6 patients with cryptogenic multifocal febrile status epilepticus (described features similar to those of NORSE) [7]. Intravenous solumedrol was attempted in one patient and PLEX in two patients. However, the authors described poor outcomes in all patients, with one patient dying, another developing severe physical and mental disabilities, and four others developing refractory multifocal epilepsy. Brain pathology was available in four patients (two biopsies and two autopsies). The histology was normal in two and showed changes due to seizures in the remaining patients [7]. Another case series demonstrated only nonspecific changes on brain biopsy in 3 patients with NORSE [8]. In a pediatric case series reported by Mikaeloff et al. [12], 6/14 patients received IVSM, with one likely experiencing a transient benefit in seizure control. However, all 14 patients had major cognitive deficits on subsequent neuropsychological assessments, with only one patient being seizure-free. In another pediatric series published by Baxter et al. [6], 5/6 patients received IVSM, with two also receiving IVIg. All patients experienced poor outcomes including death in three who received IT. Although etiology of seizures is more heterogeneous in the pediatric population, these two case series likely represent a syndrome similar to NORSE observed in adults. Wilder-Smith et al. report outcomes in seven patients identified with NORSE from Singapore, five of whom died [5]. While three patients received IVIg infusion at least once, it is not clear from the report how many of the expired patients received IVIg. The two remaining patients survived but were in a persistent vegetative state. In our series, one patient (P6) was already on multiple chemotherapeutic agents and yet developed NORSE. The IT for the patients described by Wilder-Smith et al. may not have been optimized for the three patients who received IVIg, as the infusions were administered once daily and for only three days [5]. As expected, investigations were unrevealing as to the etiology, although, notably, serum or CSF paraneoplastic autoantibodies were not tested in any case. Based on this, we suggest sending all appropriate metabolic, infectious, and paraneoplastic labs from serum, urine and/or, CSF as elaborated in Section 2.2 as soon as NORSE is suspected. Metabolic results would aid in identifying reversible causes, whereas testing for infectious causes would help tailor a directed antimicrobial therapy, especially in the setting of a recorded fever on initial presentation, preceding febrile illness or elevated CSF white cell count. Although paraneoplastic lab results may take a few weeks to report, it would be beneficial to aid planning of early IT, especially if any of the autoantibody tests are positive. However, the association between testing positive for autoantibodies and developing autoimmune epilepsy is not straightforward as recently discussed by Ruegg and Panzer. [18] in their critical appraisal of the aforementioned study from Mayo Clinic [17]. Immunotherapy may be associated with significant adverse effects, and description of its use is unfortunately limited to small retrospective reports. While the approach in treating NORSE hinges on maximal use of available resources, the decision to employ immunotherapy is more challenging in patients with suspected autoimmune epilepsy.

5. Limitations This retrospective study is based on patents identified from a single center, using appropriate ICD/NCD codes which could potentially miss patients that would otherwise meet the inclusion criteria. Comparison of favorable outcomes may be skewed in favor of patients receiving immunotherapy in our study because of different numbers of patients in each group (8 vs. 3 patients not receiving any IT). Because of a limited number of patients, the association between IT and favorable outcome was not adjusted for potentially confounding variables (e.g., age) with multivariate analysis. Patients from this study are included in a pooled analysis with those from previously published studies, resulting in significant heterogeneity in patient demographics, etiologies, criteria for NORSE diagnosis, investigations completed, and treatments offered.

A.M. Khawaja et al. / Epilepsy & Behavior 47 (2015) 17–23

6. Conclusion While NORSE has been defined in patients without any identifiable etiology, we propose a change to this approach based on the report of eleven cases of suspected NORSE managed at our institution in combination with the cases from literature. All patients with status epilepticus refractory to initial treatment and in whom an etiology is not immediately apparent should be subjected to comprehensive metabolic, infectious, and paraneoplastic investigations. New-onset refractory status epilepticus should be suspected, especially due to an autoimmune etiology. Consideration should be given to early administration of immunotherapy using IV steroids and/or IVIg then perhaps followed by chemotherapy based on the response and risk benefit assessment. Although the data from the retrospective studies provide support to design prospective studies to determine the impact of immunotherapy in patients with NORSE, the current level of evidence supporting this practice is Class IV. Because of the relatively rare occurrence of NORSE, single-center studies are unlikely to be feasible, thereby supporting the need for multicenter randomized trials assessing the efficacy of immunotherapy, choice of first- vs. second-line IT, and the ideal timing of such interventions. Disclosures and funding sources Data included in this manuscript were presented, in part, at the Annual Meeting of the Neurocritical Care Society in 2014. References [1] Trinka E, Hofler J, Zerbs A. Causes of status epilepticus. Epilepsia 2012;53(Suppl. 4): 127–38. [2] Brophy GM, Bell R, Claassen J, Alldredge B, Bleck TP, Glauser T, et al. Guidelines for the evaluation and management of status epilepticus. Neurocrit Care 2012;17:3–23.

23

[3] Shorvon S, Ferlisi M. The treatment of super-refractory status epilepticus: a critical review of available therapies and a clinical treatment protocol. Brain 2011;134: 2802–18. [4] Kilbride RD, Reynolds AS, Szaflarski JP, Hirsch LJ. Clinical outcomes following prolonged refractory status epilepticus (PRSE). Neurocrit Care 2013;18:374–85. [5] Wilder-Smith EP, Lim EC, Teoh HL, Sharma VK, Tan JJ, Chan BP, et al. The NORSE (new-onset refractory status epilepticus) syndrome: defining a disease entity. Ann Acad Med Singapore 2005;34:417–20. [6] Baxter P, Clarke A, Cross H, Harding B, Hicks E, Livingston J, et al. Idiopathic catastrophic epileptic encephalopathy presenting with acute onset intractable status. Seizure 2003;12:379–87. [7] Van Lierde I, Van Paesschen W, Dupont P, Maes A, Sciot R. De novo cryptogenic refractory multifocal febrile status epilepticus in the young adult: a review of six cases. Acta Neurol Belg 2003;103:88–94. [8] Costello DJ, Kilbride RD, Cole AJ. Cryptogenic new onset refractory status epilepticus (NORSE) in adults — infectious or not? J Neurol Sci 2009;277:26–31. [9] Gall CR, Jumma O, Mohanraj R. Five cases of new onset refractory status epilepticus (NORSE) syndrome: outcomes with early immunotherapy. Seizure 2013;22:217–20. [10] Li J, Saldivar C, Maganti RK. Plasma exchange in cryptogenic new onset refractory status epilepticus. Seizure 2013;22:70–3. [11] Kaneko J, Iizuka T, Asari H, Ishima D, Kasakura S, Hara A, et al. Retrospective review of 6 Patients with new-onset refractory status epilepticus (NORSE) syndrome: early intervention with intravenous cyclophosphamide may improve outcome. Neurology 2013;80 [Meeting Abstracts 1]. [12] Mikaeloff Y, Jambaque I, Hertz-Pannier L, Zamfirescu A, Adamsbaum C, Plouin P, et al. Devastating epileptic encephalopathy in school-aged children (DESC): a pseudo encephalitis. Epilepsy Res 2006;69:67–79. [13] Kramer U, Shorer Z, Ben-Zeev B, Lerman-Sagie T, Goldberg-Stern H, Lahat E. Severe refractory status epilepticus owing to presumed encephalitis. J Child Neurol 2005; 20:184–7. [14] Sahin M, Menache CC, Holmes GL, Riviello JJ. Outcome of severe refractory status epilepticus in children. Epilepsia 2001;42:1461–7. [15] Gaspard N, Foreman B, Judd LM, Brenton JN, Nathan BR, McCoy BM, et al. Intravenous ketamine for the treatment of refractory status epilepticus: a retrospective multicenter study. Epilepsia 2013;54:1498–503. [16] Kanter IC, Huttner HB, Staykov D, Biermann T, Struffert T, Kerling F, et al. Cyclophosphamide for anti-GAD antibody-positive refractory status epilepticus. Epilepsia 2008;49:914–20. [17] Toledano M, Britton JW, McKeon A, Shin C, Lennon VA, Quek AM, et al. Utility of an immunotherapy trial in evaluating patients with presumed autoimmune epilepsy. Neurology 2014;82:1578–86. [18] Ruegg S, Panzer JA. Immune therapy for pharmacoresistant epilepsy: ready to go? Neurology 2014;82:1572–3.