Published Ahead of Print on September 4, 2019 as 10.1212/WNL.0000000000008159 ARTICLE The epileptology of alternating hemiplegia of childhood

Julie Uchitel, Ashley Helseth, MD, PhD, Lyndsey Prange, CPNP, Melissa McLean, BSBA, Ryan Ghusayni, MD, Correspondence Monisha Sachdev, MD, Arsen Hunanyan, PhD, and Mohamad A. Mikati, MD Dr. Mikati mohamad.mikati@ Neurology 2019;93:e1-e12. doi:10.1212/WNL.0000000000008159 ® dm.duke.edu Abstract Objective To report our experience and investigate 5 original hypotheses: (1) multiple types of epileptic seizures occur in alternating hemiplegia of childhood (AHC), and these can be the initial presentation; (2) epileptiform abnormalities often appear well after clinical seizures; (3) nonepileptic reduced awareness spells (RAS) occur frequently; (4) epilepsy is commonly drug resistant but may respond to vagal nerve stimulation (VNS); and (5) status epilepticus (SE) is common and is usually refractory and recurrent.

Methods We analyzed a cohort of 51 consecutive patients with AHC.

Results Thirty-two of 51 patients had epilepsy: 18 focal seizures, frontal more frequently than temporal, and then posterior. Eleven had primary generalized seizures (tonic-clonic, myoclonic, and/or absence). Epileptic seizures preceded other AHC paroxysmal events in 8 (lag 5.63 ± 6.55 months; p = 0.0365). In 7 of 32, initial EEGs were normal, with the first epileptiform EEG lagging behind by 3.53 ± 4.65 years (p = 0.0484). RAS occurred equally in patients with epilepsy (16 of 32) and patients without epilepsy (10 of 19, p = 1.0). Twenty-eight patients had video-EEG; captured RAS showed no concomitant EEG changes. Nineteen patients (59%) were drug resistant. VNS resulted in >50% reduction in seizures in 5 of 6 (p < 0.04). Twelve patients (38%) had SE (9 of 12 multiple episodes), refractory/superrefractory in all (p < 0.001), and 4 of 12 had regression after SE.

Conclusions Epilepsy in AHC can be focal or generalized. Epileptic seizures may be the first paroxysmal symptom. EEG may become epileptiform only on follow-up. Epilepsy, although frequently drug resistant, can respond to VNS. RAS are frequent and nonepileptic. SE often recurs and is usually refractory/superrefractory. Our observations are consistent with current data on AHC- ATP1A3 pathophysiology.

From the Division of Pediatric Neurology, Duke University Health System, Duke University School of Medicine, Durham NC.

Go to Neurology.org/N for full disclosures. Funding information and disclosures deemed relevant by the authors, if any, are provided at the end of the article.

Copyright © 2019 American Academy of Neurology e1 Copyright © 2019 American Academy of Neurology. Unauthorized reproduction of this article is prohibited. Glossary AHC = alternating hemiplegia of childhood; ILAE = International League Against Epilepsy; MRS = magnetic resonance spectroscopy; RAS = reduced awareness spells; SE = status epilepticus; VNS = vagal nerve stimulation.

– Alternating hemiplegia of childhood (AHC) is a rare neuro- criteria.12 14 In addition, for the purposes of this study and logic disorder characterized by recurrent episodes of hemi- because the AHC spells can often be difficult to distinguish plegia, double hemiplegia, , and frequently epilepsy.1 from epileptic seizures, we used the following additional cri- De novo mutations in the ATP1A3 are the cause of AHC teria. An epileptic seizure was considered to have occurred if at – in ≈75% of patients.2 5 AHC is diagnosed according to the least one of the following conditions was satisfied: (1) repre- Aicardi clinical criteria.1,3,6,7 sentative event recorded on video-EEG was proven to be an epileptic seizure by concurrent electrographic seizure activity; Epilepsy occurs in about half of all patients with AHC and is (2) according to the treating epileptologist, the semiology of – often drug resistant.8 10 However, a full characterization of the the event is definitively indicative of an epileptic seizure (e.g., epilepsy of these patients, including seizure types and local- focal clonic seizure with motor march); or (3) according to the izations, EEG characteristics, and other seizure-related mani- treating epileptologist, semiology of the event is consistent with festations, is still needed. Furthermore, status epilepticus (SE) an epileptic seizure and the patient also has definitive interictal is known to occur in AHC and is, at times, followed by de- EEG changes consistent with such a semiology (e.g., episode of velopmental regression and psychomotor deterioration.11 staring with eyelid flutter and an interictal EEG of 3-Hz spike However, the characteristics of SE in these patients remain to slow waves). Seizure types were defined according to ILAE be fully described. definitions as focal, generalized, or unknown onset.14 Classifi- cation of seizure type and localization were performed On the basis of our experience in managing epilepsy in patients according to the following procedure. First, we carefully ana- with AHC, we generated the following original hypotheses that lyzed the description of the seizure and correlated it with the we aimed to investigate in our study: (1) epilepsy in AHC can EEG interictal findings. Semiology was, thus, classified as focal be localization related (focal) or generalized, and epileptic (with localization to a specific lobe) or generalized, as long as seizures can be the initial presenting symptom; (2) the ap- the EEG and semiology corresponded to each other and were pearance of epileptiform abnormalities often lags behind epi- clearly indicative of a specific classification and localization. lepsy onset; (3) nonepileptic reduced awareness spells (RAS) Second, in patients on whom video-EEG monitoring was occur frequently; (4) epilepsy is commonly drug resistant but performed, monitoring information was used. Third, in those may respond to vagal nerve stimulation (VNS); and (5) SE is cases in whom it was not possible to classify the seizure type common, usually refractory, and recurrent. with the above 2 procedures, the seizure type was classified as unknown onset. The above procedures are fully consistent with the ILAE recommendations for the classification of seizures Methods and epilepsy and for application of this classification.12,14 Focal seizures in which lobe localization was not possible were clas- We analyzed the data of a cohort of 51 consecutive patients sified as undetermined localization. Severity of developmental seen in our multidisciplinary AHC clinic who fulfilled the disability was categorized as mild, moderate, severe, or pro- Aicardi criteria (see appendix e-1 available from Dryad, doi. found on the basis of the neurologic and developmental org/10.5061/dryad.qn3287b).1,3,6,7 Patients underwent clin- examinations performed at the time of last follow-up as de- ical evaluations and testing according to clinical need and our 1,15,16 scribed and used in previous studies. AHC clinical pathway1 (including genetic, EEG, and MRI investigations). Genetic testing was performed through EEG studies whole-exome sequencing or through a targeted panel that Our pathway includes performance of a minimum of a 4-hour included ATP1A3, ATP1A2, SLC2A1, SCN1A, PRRT2, and outpatient video-EEG on all patients with inpatient admission CACN1A, as well as other , performed with next- when clinically indicated. generation sequencing and then, when positive, confirmed by Sanger sequencing (table e-1 available from Dryad, doi.org/ Neuroradiologic studies 10.5061/dryad.qn3287b). Patients provided consent, and MRI was performed on all patients. Magnetic resonance their prospectively collected data from our center and retro- spectroscopy (MRS) was acquired on 7 patients according to spectively available data from previous centers were entered clinical need to assess specific areas of interest. Interictal PET into our Institutional Review Board–approved database. was performed on 3 patients.

Epilepsy and seizure types Therapies Presence of epilepsy and epileptic seizures was defined We classified epilepsy response to therapy by the ILAE clas- according to the International League Against Epilepsy (ILAE) sifications of drug resistant, seizure-free, epilepsy resolved, or e2 Neurology | Volume 93, Number 13 | September 24, 2019 Neurology.org/N Copyright © 2019 American Academy of Neurology. Unauthorized reproduction of this article is prohibited. undetermined response.13 To collect exploratory data on the response of seizures to specific medications, including anti- Figure 1 Summary of epilepsy characteristics of 32 patients epileptics, off-label use of repurposed drugs, steroids, or keto- with alternating hemiplegia of childhood with genic diet, we used the following empirical definitions for this epilepsy study. A medication was considered effective in a patient if that patient achieved seizure freedom (as per the ILAE definition) on it, possibly effective if both the physician and family agreed that the medication helped but did not result in seizure freedom, or ineffective otherwise. Off-label repurposed drugs were used in 11 patients due to their mechanisms of action. The intent was that they might reduce the severity of epileptic seizures and of other AHC manifestations. These included dextromethorphan, which reduces spreading depression, a presumed mechanism of increased excitability in AHC; amiloride, which has antiseizure effects in rodents and inhibitory effects of the Na+/Ca++ ex- changer; and ATP, which has previously been reported to have helped 1 patient with AHC and can enhance ATPase pump – activity.4,17 21 In a subgroup of 4 patients who underwent VNS implantation in our center, we documented response to VNS in a detailed way. To do this, we used seizure counts and an AHC spell index (based on 3 major variables that determine the extent of attacks: severity, frequency, and duration9;seeappendixe-2 available from Dryad, doi.org/10.5061/dryad.qn3287b) calcu- lated for the 1-month baseline preceding VNS insertion and then at follow-up at the specific time points indicated below.

Status epilepticus fi Pie chart of seizure types of all patients with epilepsy. Values in parentheses We categorized SE as de ned in the literature. Refractory SE represent the number of patients with epilepsy with each type of localization persists despite a sufficient dose of benzodiazepines and at least 1 and percent of all. antiepileptic drug regardless of time.22 Superrefractory SE con- tinues for ≥24 hours after the use of anesthetic therapy, including cases that recur on weaning of the anesthetic agent.22 We also of these patients are shown in table 1. Characteristics of all 32 categorized SE as focal or generalized and documented its du- patients, seizures, epilepsy, and gene mutations are shown in fi ration, the medications needed to control it, whether subsequent table 2. Twenty-seven were classi ed on the basis of semiology regression occurred, and other clinical characteristics. and interictal EEG and 5 on the basis of ictal EEG recordings. Seizure types were as follows (figure 1). (1) Eighteen had focal Statistics seizures: 7 were localized to the , 2 to the temporal We used the Kruskal-Wallis, Fisher exact test, binomial, and lobe, 2 to the , and 2 to the ; 4 were paired and unpaired Student t tests as appropriate. bifocal (frontal and temporal in all 4); and 1 could not be localized. Of the 18, 15 had focal seizures with impaired awareness, 3 with preserved awareness, and 16 with focal to Data availability bilateral tonic-clonic seizures; 2 had gelastic seizures. (2) Eleven Anonymized data will be shared by request from any qualified had primary generalized seizures; all had primary generalized investigator. tonic-clonic seizures, 3 had myoclonic seizures, 2 had atonic seizures, and 2 had absence seizures. (3) Three had generalized Results tonic-clonic seizures of unknown onset (unknown if primary or secondary). Nineteen patients (59%) had drug-resistant epi- Epilepsy and seizures of various types occur lepsy; 7 patients achieved seizure freedom; and 6 patients had in AHC undetermined responses (figure e-1A available from Dryad, doi. Of the 51 patients, 32 had epilepsy starting at a mean age of 1.83 org/10.5061/dryad.qn3287b, shows the time course of patients’ ± 3.00 years (range 1 day–14 years, median 9 months) (figure 1 responses to medications). Seizures were often reported to and tables 1 and 2). Mean age at AHC spell onset was not occur in the context of AHC spells. Nonepileptic events were significantly different between patients with epilepsy and categorized according to the psychogenic nonepileptic seizures patients without epilepsy (p = 0.2595). However, 8 patients had classification system.23,24 One patient had spells of arm shaking onset of epilepsy before onset of AHC spells (average age at and repetitive stepping (documented by video-EEG); another onset of seizures 2.12 ± 2.74 months, average age at onset of had repetitive stepping only (clinically established by video re- AHC spells 7.75 ± 6.32 months; p = 0.0365). The characteristics cording); a third had arm weakness that the mother could abort

Neurology.org/N Neurology | Volume 93, Number 13 | September 24, 2019 e3 Copyright © 2019 American Academy of Neurology. Unauthorized reproduction of this article is prohibited. Table 1 Characteristics of 8 patients with onset of epilepsy before AHC spells

Age at Epilepsy onset/age at other AHC Developmental Patient last FU, y manifestations onset delay classification Description of epilepsy onset

4 7 DOL 1/12 mo Moderate Seizures began at DOL 1, left-sided stiffening, head deviated to the left, confirmed by EEG to be epileptic seizures

6 1.75 2 mo/3 mo Mild Left upper-extremity clonic jerking

10 20 1 mo/4 mo Severe GTC started at age 4 wk

11 8.5 8 mo/10 mo Moderate At 8 mo had seizure consisting of eye deviation to the left with and left arm tonic and clonic activity

19 7 DOL 1/3 mo Moderate During DOL 1, cyanotic and had a seizure characterized by eye deviation to the left with clonic shaking of left arm

22 7 4 mo/4 mo Moderate Right-sided clonic seizures started at 4 mo of age, occurred every 2 wk

29 3 2 mo/7 mo Profound Tonic-clonic seizures first noted at 2 mo of age

49 26 DOL 1/5 mo Mild Generalized seizures noted at DOL 1

Summary: Mean: Mean onset of epilepsy: 2.12 ± 2.74 Mild: 2 Focal semiology: 5 patients 10.03 ± mo (range DOL 1–8 mo) Moderate: 4 Generalized semiology: 3 patients 8.47 Mean onset of other AHC spells: Severe: 1 Range: 7.75 ± 6.32 mo (range 3–12 mo) Profound: 1 1.75–26

Abbreviations: AHC = alternating hemiplegia of childhood; DOL = day of life; FU = follow-up; GTC = generalized tonic-clonic.

by calling him out of them (documented by video-EEG); and these patients. Twenty-nine of 32 patients with epilepsy had a fourth had generalized tremors that the mother recognized as long-term video-EEG monitoring (>24 hours) of an average attention seeking (clinically established by video recording). duration of 6.96 ± 7.71 days (average 3.38 ± 2.66 monitorings per The first 2 patients were diagnosed as stereotypies; the last 2 patient, median 3, range 1–10). During monitoring, numerous were diagnosed as psychogenic nonepileptic seizures. AHC spells (hemiplegia, double hemiplegia, dystonia, autonomic dysfunction, myoclonus, and abnormal eye movements) were EEG epileptiform activity onset often lags recorded without any significant concurrent EEG changes. Five behind epileptic seizure onset patients had electrographic seizures captured on video-EEG: left Abnormalities included diffuse slowing in 9 patients, focal frontal in 1 patient, right temporal in 1 (figure 2E), right fron- slowing in 6, and epileptiform activity in 19 (figure 2 and figure totemporal in 1, and generalized spike and slow wave in 1 (figure e-1B available from Dryad, doi.org/10.5061/dryad.qn3287b). 2D). The fifth patient had 6 seizures captured on video-EEG: 1 Seven of the 19 patients had an initially normal EEG that later from the left frontotemporal region, 3 from the right temporal- became abnormal (lag 3.53 ± 4.65 years, p = 0.0484). This parasagittal area, and 2 from the left temporal region. An addi- supported the notion that the yield for abnormal and epilepti- tional patient was recorded in the postictal period after SE and form activity significantly increased with increasing number of demonstrated diffuse background slowing (figure 2F). EEGs performed. EEG was categorized into 4 categories: normal, diffuse slowing, localized slowing, and epileptiform (spikes, spike RAS are common and nonepileptic in nature slow waves, and sharp waves). Of our 32 patients (138 EEG RAS consisted of isolated lethargyanddrowsinesswithstaring studies; 4.31 ± 3.37 EEGs per patient, median 3, range 1–14), 19 and reduced responsiveness, usually lasting minutes to hours had at least 1 epileptiform EEG (5.26 ± 3.48 EEGs per patient, (table 3). Twenty-six of the 51 patients in this study had epi- range 1–14), and 3 had all of their EEGs as epileptiform (1.33 ± sodes of RAS. No significant difference was noted between the 0.57 EEGs per patient, range 1–2). Twenty-six had at least 1 number of patients with epilepsy who had RAS (16 of 32) and abnormal EEG (5.00 ± 3.37 EEGs per patient, range 1–14), and those without epilepsy who had RAS (10 of 19) (p = 1.000). 8 had all abnormal EEGs (3.38 ± 2.97 EEGs per patient, range Most episodes of RAS were not associated with other symp- 1–8). Six had all normal EEGs (1.33 ± 0.82 EEGs per patient, toms; however, in about one-third of RAS episodes, there were range 1–3). There were significantly fewer EEGs performed per also brief superimposed additional manifestations such as min- patient for the 6 patients with all normal EEGs compared to imal hemiplegia, dystonia, autonomic dysfunction, agitation, those 19 with at least 1 epileptiform EEG and to those 26 with at atonia of the neck, atonia of the body, pain, or abnormal eye least 1 abnormal EEG (p = 0.0126 and p = 0.0137, respectively). movements during part of the spell. Each of these occurred only Figure e-1B (available from Dryad, doi.org/10.5061/dryad. as a minor feature with the predominant symptom being re- qn3287b) shows the time course of the EEG findings for each of duced awareness. No symptom indicative of seizures was noted

e4 Neurology | Volume 93, Number 13 | September 24, 2019 Neurology.org/N Copyright © 2019 American Academy of Neurology. Unauthorized reproduction of this article is prohibited. Table 2 Characteristics of epilepsy in 32 patients with AHC

Characteristic Description (n)

Type of seizures Focal seizures with impaired awareness (15), focal seizures with preserved awareness (3), PGTS (11), focal to bilateral tonic-clonic (16), GTC with unknown onset (3), myoclonic (3), absence (2), atonic (2), febrile (2) In the patients with focal seizures, 2 had gelastic seizures, and 1 was described to have a Jacksonian motor march

93.8% (30/32) of patients had either GTC seizures or SE

Localization Frontal (7), frontal and temporal (4), temporal (2), occipital (2), parietal (2), primary generalized (11), unknown if primary or focal to bilateral (3)

Age at onset, y Mean 1.83 ± 3.00, range 0–14

Developmental delay In 32 patients with epilepsy: none (1, 3%), mild (8, 25%), moderate (6, 19%), severe (15, 47%), profound (2, 6%)

In 19 patients without epilepsy: none (1, 5%), mild (8, 42%), moderate (6, 32%) severe (4, 21%), profound (0, 0%)

Seizure frequency In 11 patients with drug-resistant focal seizures: range 4× per day—once per year, median 2–3 times per month

In 6 patients with drug-resistant primary GTC seizures: range once per week–once every few months, median 2–3 times per month

In 2 patients with absence of seizures: several times per day in 1, several times per month in the other

In 2 patients with drug-resistant GTC seizures of unknown onset: 2–3 times per week in 1, 2 times per month in 1

Myoclonic and atonic seizure frequency ranged between once per week and 3 times in the lifetime

Response to treatment 19 were drug resistant, 7 achieved seizure freedom, 6 were undetermined

Interictal EEG findings Normal EEGs: 19% (6/32) have only ever had normal EEGs, 72% (23/32) has at least 1 normal EEG

Diffuse slowing: 3% (1/32) have only had EEGs with diffuse slowing, 31% (10/32) had at least 1 EEG with diffuse slowing

Focal slowing: 0% (0/32) have only had EEGs with focal slowing, 22% (7/32) had at least 1 EEG with focal slowing

Epileptiform/paroxysmal activity: 9% (3/32) have had all EEGs with epileptiform/paroxysmal activity, 59% (19/32) had at least 1 EEG with epileptiform/paroxysmal activity

Neuroradiologic studies MRI: normal in 72% (23/32), abnormal in 28% (9/32)

MRS: abnormal in 80% of those tested (1/5 normal and 4/5 abnormal), MRS was normal in 2/2 patients with no prior epilepsy or SE

Video-EEG ictal and postictal Clinical seizures ictal recordings: 5 patients (5/32, 16%), focal onset in 4 and generalized in 1 recordings Clinical seizures postictal recordings: 1 patient with diffuse slowing of background

RAS captured: 28 in 10/32 patients, no EEG change in any

Genotype-phenotype No. with epilepsy/ Mean age at epilepsy Drug resistant, SE, Mean age at first Range age at correlation total (%) onset (y) n(%) n(%) SE, y first SE, y

E815K 3/5 (60) 2.18 ± 2.55 2/3 (66) 2/3 1.5 ± 2.12 Neonate–7 (66)

D801N 5/9 (56) 4.07 ± 6.11 2/5 (40) 3/5 8.33 ± 6.81 3–24 (60)

Abbreviations: AHC = alternating hemiplegia of childhood; GTC = generalized tonic-clonic; MRS = magnetic resonance spectroscopy; PGTC = primary generalized tonic-clonic; RAS = reduced awareness spells; SE = status epilepticus. Other less frequent mutations seen (total 20 different ATP1A3 mutations) included G89D, G775C, Y768H, A333T, P336S, Q851R, G947R, D923Y, C609Y, E828K, L326R, R756H, L839P, V589F, V589F, A320T, E828K, and L839P. The numbers of patients with each were less than the above 2 mutations and thus were not enough to meaningfully perform additional correlations.

during such spells. Altogether, 10 (8 with epilepsy, 2 without cerebellar atrophy in 4, nonspecific abnormalities in 4, and epilepsy) patients of the 26 had 28 RAS captured on video-EEG, a postictal (within 24 hours after SE) hyperintense fluid- all without EEG correlate (table 3 for details). attenuated inversion recovery signal in the right parietal cortex in 1 (figure 3B) (details of all patients in appendix e-3 Neuroradiologic abnormalities occur in available from Dryad, doi.org/10.5061/dryad.qn3287b). severe cases This patient also had residual long-term left Of the 32 patients with epilepsy, 23 had normal MRIs and 9 after this episode of SE (see section on SE below). Four of had abnormalities: diffuse usually more frontal and the 7 patients who had MRS had an abnormal MRS study,

Neurology.org/N Neurology | Volume 93, Number 13 | September 24, 2019 e5 Copyright © 2019 American Academy of Neurology. Unauthorized reproduction of this article is prohibited. Figure 2 EEGs of 5 patients with alternating hemiplegia of childhood with epilepsy

(A) Frontal interictal. Left frontal spike activity with predominantly left frontal slowing in patient 4 at the age of 4.5 years. (B) Temporal interictal. Left temporal spikes in patient 27 at the age of 16 years. (C) Occipital interictal. Left occipital spikes in sleep in patient 26 during sleep. (D) Generalized spike wave discharges in patient 10 with concurrent myoclonic jerks. (E) Ictal EEG during status epilepticus (SE). EEG in patient 4 at the age of 5 years during SE showing an ongoing rhythmic sharp electrographic seizure activity with phase reversals over T6. (F) Postictal slowing in patient 21 several hours after an episode of generalized tonic-clonic SE at the age of 24 years. (A) and (E) are for the same patient.

and all 4 had a history of SE and epilepsy (details presented n = 4, dose range 2 to 4 mg/kg/d; ATP n = 2, doses 1.5 mg/ in appendix e-3 available from Dryad, doi.org/10.5061/ kg/d; and amiloride n = 3, dose range 0.2 to 0.3 mg/kg/d. dryad.qn3287b). Three had normal MRS: 1 patient had None of these agents were associated with a favorable re- a history of SE, and 2 did not have epilepsy. Three patients sponse for seizures or hemiplegia. Of the 4 patients who had PET scans, 2 of which were normal and 1 showed received dextromethorphan, 1 patient had a decrease in the hypometabolism in the right in a patient with number of dystonia spells from 7 per day (10-minute du- known temporal lobe epilepsy. ration each) to 4 per week (10-minute duration each). In addition, pulse steroids (30 mg/kg/d × 3 d/mo) were used in 2 patients. One 8-year-old patient reported hemiplegia Epilepsy in AHC is often drug resistant but can frequency to be reduced to one-fourth the original frequency respond to VNS and from hours to minutes in duration. This patient has Of 22 antiepileptic drugs, we could not observe any de- continued to receive monthly pulse steroids for the past 2 finitive specific pattern of better response with certain years. The other patient did not tolerate treatment due to medications or classes of medications (table e-2 available behavioral side effects. Two of 6 patients benefited from the from Dryad, doi.org/10.5061/dryad.qn3287b). The 19 drug-resistant patients were on 2.95 ± 1.21 (range 1–5) ketogenic diet. One patient experienced a reduction in sei- medications as of the last follow-up (mean age when last zure and AHC spell frequency (baseline: seizures once per seen 11.40 ± 10.53 years, range 1.5–26 years). For the 11 month, hemiplegia 2 times per month, dystonia 3 times per patients on off-label repurposed drugs, the number of day; at the 4-month follow-up: no seizures, no hemiplegia, patients and doses used were as follows: dextromethorphan dystonia 2 times per week), and 1 patient became more vocal e6 Neurology | Volume 93, Number 13 | September 24, 2019 Neurology.org/N Copyright © 2019 American Academy of Neurology. Unauthorized reproduction of this article is prohibited. Table 3 Characteristics of RAS

Patients Additional superimposed Spells recorded with manifestations (total on video-EEG/ RAS manifestation spells, n patients, n) patients, n Frequency Duration

Reduced awareness alone as the only 8 None 19/6 Few times 10 s–30 min manifestation in all RAS of that patient a day–few times in life

Reduced awareness alone in some spells and in 18a Reduced awareness with 0/0 Few times per 10 s–3min other spells reduced awareness with other abnormal eye movements (3) day–few times manifestations per week

Reduced awareness with 1/1 Few times per 5–10 min agitation (2) week

Reduced awareness with 1/1 Multiple times 10 s–10 min autonomic changes (4) per day

Reduced awareness with 3/1 Few times per 10 dystonia (6) month seconds–30 min

Reduced awareness with pain 4/1 Few times per 10 s (2: head in 1; feet, eyes, and day–few times stomach in 1) per week

Abbreviation: RAS = reduced awareness spells. a Eighteen of 26 patients had both types of RAS: reduced awareness with no other accompanying manifestations and, in the same patient at other times, reduced awareness as the main manifestation of the spell with accompanying minor manifestations. The rows to the right detail the most common accompanying manifestations that occurred during RAS and the number of patients who had them. Many of these patients had multiple types of accom- panying manifestations. Some of the 18 patients had additional accompanying manifestations not listed in these rows, including myoclonus, repetitive speech, enuresis, mouthing movements, and vision loss. and alert during the 6 months of follow-up on the diet without are shown in table 4. One had 8 episodes; 5 had 3 episodes; 3 a change in seizure or spell frequency. All 4 patients who had the had 2 episodes; and 3 had 1 episode. All episodes were re- VNS device inserted in our center reported a >50% reduction in fractory and 3 were superrefractory. Seven had focal to bi- seizure frequency (mean follow-up, 22.2 ± 5.42 months, range lateral convulsive SE; 3 had primary generalized convulsive 9–36 months). Pre-VNS seizure frequency at baseline was 99.5 SE; 1 had nonconvulsive SE of focal etiology; and 1 had both ± 59.2 per month; at 1 month, 8.50 ± 4.97; at 6 months, 4.25 ± focal to bilateral convulsive SE and nonconvulsive SE of focal 2.72; and last follow-up, 1.50 ± 0.96 (p = 0.025). VNS was also etiology. In 5 of 32 episodes, there was an initial recognizable associated with a modest reduction in the spell index score for focal impaired awareness seizure phase before generalization. plegic attacks (baseline before VNS 7.63 ± 0.55, 1 month 6.5 ± Patients required 2 to 5 medications to halt SE: diazepam, 0.20, 6 months 5.38 ± 0.8, last follow-up [12–29 months] 4.88 ± lacosamide, topiramate, phenobarbital, midazolam, lor- 0.90; p = 0.036). The change in the spell index score for dys- azepam, levetiracetam, fosphenytoin, and/or propofol. Nine tonic attacks did not achieve statistical significance (baseline SE episodes in 5 patients required intubation. Four SE epi- beforeVNS6.13±1.05,1month5.5±1.02,6months3.63± sodes in 3 different patients were immediately preceded by 0.97, last follow-up 3.63 ± 1.13; p = 0.167). One patient had nonepileptic AHC spells. Two other SE episodes were pre- episodes of atonia associated with output current settings cipitated by attempts to taper phenobarbital in 1 patient and >1 mA that resolved by reducing the output current. Of the 2 levetiracetam in another. Four of 12 patients had regression patients in whom the VNS was inserted elsewhere, 1 patient after SE. Regression was reversible in 2 of the 4 (patients 15 reported >50% reduction in seizure and AHC spell frequency, and 18) but not in the other 2 (patients 21 and 51). Patient 15 and the other did not report improvement. In both patients, the had weakness on the left side for 12 hours following SE and VNS stimulation was discontinued; in the first due to an in- was poorly coordinated for 2 to 3 days, but then fully re- ability to reimplant after the first implantation became non- covered. Patient 18, at age 18 months, lost the ability to walk functional and in the second due to lack of efficacy. after 4 hours of SE and then regained this ability 1 month later. She then, that same year, regressed in swallowing and fine SE is often refractory, recurrent, and followed motor movements after 1.5 hours of SE and regained these in by regression 1 month. Then at age 4 years, after 1.5 hours of SE, she lost High frequency, drug resistance, and developmental re- her ability to comprehend speech and her 30 words of sign gression were noteworthy characteristics of SE (table 4). language and only regained these 6 months later. Patient 21 Twelve patients had 32 episodes of SE, ranging from 30 lost the ability to feed herself and the ability to walk and minutes to 4.5 days (mean age 5.58 ± 6.07 years, range developed diffuse left-sided greater than right-sided weakness. neonate–24 years, median 3.50 years). Characteristics of SE By 3 months, she recovered all these except for residual

Neurology.org/N Neurology | Volume 93, Number 13 | September 24, 2019 e7 Copyright © 2019 American Academy of Neurology. Unauthorized reproduction of this article is prohibited. Figure 3 MRIs of 3 patients with alternating hemiplegia of childhood with epilepsy

(A) Frontal atrophy and cerebellar at- rophy in patient 8 at the age of 22 years. (B) Increased intensity in the gray matter on fluid-attenuated in- version recovery in the right medial parietal area on MRI performed within 24 hours after an episode of status epilepticus in patient 21 at the age of 8 years. Postictal EEG at the time of MRI revealed right hemi- sphere voltage suppression, de- creased sleep spindles, and periods of high-voltage delta with sharp con- tours. (C) Magnetic resonance spec- troscopy (MRS) of patient 32 at age 2 years in the hippocampus demon- strating an increased lactate (Lac) peak and decreased N-acetyl aspar- tate (NAA) peak the day after a focal to bilateral tonic-clonic seizure that re- solved on its own and during an on- going dystonic spell. Highlighted squares on the right represent areas studied by MRS. Cr2 and Cr refer to peaks for creatine, Ins 1 for Myoinsi- tosol, and Cho for Choline.

weakness that persisted even at last follow-up 1 year later. The studies have reported a 15% to 77% range of occurrence of – fourth, patient 32, regressed, losing the previously established epilepsy in AHC.9,25 27 The observed prevalence in our study ability to sit up, roll over, reach for objects, play, smile, and take (62%) is on the upper end of this range, possibly due to food by mouth, none of which returned as of the latest follow- potential referral bias: patients with AHC may have been up after 18 months, at which time he was still being fed through referred to us because we are also a level 4 epilepsy center. fi a gastrostomy tube. For further details on regression after SE, Our ndings extend the knowledge about epilepsy in AHC by see appendix e-4 available from Dryad (doi.org/10.5061/ demonstrating that it can be focal or generalized, that it can dryad.qn3287b). These patients are reported in table 4 as also be multifocal, and that the anterior lobes are more af- patients 15, 18, 21, and 32. The MRI performed on patient 21 fected than the posterior lobes. This is consistent with the prior finding that cerebral hypometabolism in AHC is more within 24 hours after the SE episode is presented in figure 3B. severe in the frontal lobes.28,29 Of note, we observed a higher incidence of epilepsy and SE in the patients with E815K as Discussion previously reported.10,11,26 In addition, we demonstrated that in a significant subset of patients (8 of 32, 25%) with epilepsy, Seizure and epilepsy characteristics epileptic seizures and even SE can occur before other types of Epilepsy in our patients with AHC was characterized by AHC spells, with the epileptic seizures often starting in the multiple seizure types and localizations, drug resistance, and neonatal period. Physicians taking care of neonates and high frequency and high recurrence rate of refractory SE. Prior infants with new-onset epileptic seizures of unknown etiology e8 Neurology | Volume 93, Number 13 | September 24, 2019 Neurology.org/N Copyright © 2019 American Academy of Neurology. Unauthorized reproduction of this article is prohibited. Table 4 Characteristics of SE in AHC

Age at SE last Patient/ ATP1A3 Seizure types/ episodes, Age at SE follow- Duration of each Average sex mutation localization n episode, y up, y Regression episode duration

1/M E815K PGTC/primary 3 Neonate, 8 No At least 30 min, 1 h, 3 h 2 h generalized 1, 2

2/M E815K FBTC, FIA/ 3 3, 4, 7 9 No 45 min each 45 min frontal

4/F D801N FBTC, FIA, 3 3, 5, 5 7 No 2 h, 30 min, 30 min 1 h gelastic/frontal and temporal

7/M D801N FBTC, FPA/ 2 6, 7 11 No 30 min, 1 h 45 min occipital

8/M D810N FBTC, FIA/ 3 16, 23, 24 24 No 9 h, 1h, 1h 15 min 3.75 h frontal and temporal

9/M Y768H FIA, FPA, FBTC, 1a 13 23 No 30 min 30 min nonconvulsive SE/frontal

11/F L326R FBTC, FIA/ 2 6, 7 8.5 No 48 h, 24 h 36 h temporal

15/F D923Y PGTC/primary 2 3, 5 6 Yes; motor and language 180 min, 30 min 105 min generalized with full recovery

18/F Negative PGTC, 3 1.5, 1.5, 4 6 Yes; motor and language 4 h, 1.5 h, 1.5 h 140 min myoclonic, with full recovery atonic/primary generalized

21/F Negative FBTC, FIA, 1 8 8 Yes; motor and language 4.5 d 4.5 d febrile/parietal without full recovery

27/M Negative FBTC, FIA, tonic/ 11316No 36h 36h frontal and temporal

32/M V589F FBTC/frontal 8 0.5, 0.66, 4 Yes; motor (gross motor 30 min, 40 min, 45 36 min and temporal 1.25, 1.25, and feeding) interaction, min, 30 min, 30 min, 1.5, 2, 2, 2.5 and speech without 30 min, 45 min, 40 min recovery

Abbreviations: AHC = alternating hemiplegia of childhood; FIA = focal seizures with impaired awareness, FPA = focal seizures with preserved awareness, FBTC = focal to bilateral tonic-clonic seizures, PGTC = primary generalized tonic-clonic seizures; SE = status epilepticus. a Patient 9 had, FIA, FPA, and FBTC followed by nonconvulsive SE.

should consider AHC as a potential cause and be vigilant other spells not previously described or emphasized, partic- about recognizing the development of its symptoms. ularly RAS, underscores the fact that physicians should be vigilant for various semiologies with obvious implications on EEG and video-EEG studies therapy. For example, one of the patients was investigated in We identified a range of EEG abnormalities in patients with another center for epilepsy surgery candidacy with subdural AHC with epilepsy, including focal discharges and generalized electrode monitoring before the diagnosis of AHC was made. spike slow wave discharges. Our observations indicate that This also includes not only multiple types of focal seizures and while most patients had abnormal and epileptiform EEGs, various types of generalized seizures but also RAS, stereoty- many did not initially manifest those abnormalities. In addi- pies, and psychogenic nonepileptic seizure episodes that are tion, 6 patients had only normal EEGs. This is likely due to the distinct from other known paroxysmal events of AHC. This is significantly fewer number of EEGs performed in these the first study to explicitly define and describe RAS as a dis- patients (1.33 ± 0.82) compared to patients with epileptiform tinct and common type of spells in AHC that often mimics EEGs (5.26 ± 3.48). Physicians need to be aware of these seizures. It is possible that the underlying pathophysiologic observations, with a low threshold to repeat EEGs if epileptic mechanism of RAS is similar to that of confusional migraine30 seizures are still suspected. In addition, our observation that and that this involves spreading depression.4 The fact that patients with AHC can have a variety of seizure types and both confusional migraine and AHC animal models are

Neurology.org/N Neurology | Volume 93, Number 13 | September 24, 2019 e9 Copyright © 2019 American Academy of Neurology. Unauthorized reproduction of this article is prohibited. predisposed to spreading depression supports the hypothesis patients, which is similar to previous reports of 33% (3 of 9) that RAS are migrainous in nature.4,30 Physicians should be and 32% (6 of 19).11,25 Two of the 4 did not fully recover from aware that RAS and nonepileptic events are not uncommon in this regression. Of note is that recent data have highlighted that AHC. In such cases, video-EEG monitoring can be helpful to there is very frequently a delay in the initiation of therapy for SE distinguish various seizure and spell types. in children.46 This would be particularly troublesome for patients with AHC who are predisposed to regression after SE. Neuroradiologic studies In addition, many of our patients had SE start in the context of We note that our results are consistent with the previous prolonged AHC spells or after stress or during antiepileptic observations of usually normal MRI, except for occasional ce- drug taper. Thus, physicians should have a low threshold for 25–29,31–38 rebral and cerebellar atrophy noted in ≈10% of cases. admitting patients if there is a prolonged AHC spell with as- However, our findings extend prior findings with a number of sociated seizures because this could be followed by SE. novel observations. First, we observed that 1 patient had, on an MRI done within 24 hours after SE (patient 21), postictal Correlations with underlying pathophysiology changes of focal increased T2 signal that correlated with the The multiple types of seizures, drug resistance, and increased epileptic focus documented on ictal EEG and with the long- predisposition to SE are consistent with the recently uncovered term post-SE deficit of contralateral hemiplegia. Second, our pathophysiology of AHC. ATP1A3 is diffusely expressed in findings extended prior MRS observations by demonstrating brain neurons, predominantly in GABAergic interneruons.47 In increased lactate peaks postictally in 2 patients, suggesting the mouse model carrying the D801N mutation, there is in- anaerobic metabolism (novel). We also observed an increased creased excitability of hippocampal CA1 pyramidal cells, which choline peak in another patient, suggesting increased cellular is easily demonstrable with high-frequency, but not with low- turnover (confirms a prior observation).39 Finally, the reduced frequency, Schaffer collateral stimulation.4,5 This is due largely N-acetyl aspartate peak seen in the 4 of 5 patients who expe- to reduced firing rate of fast-spiking interneurons.5 It appears rienced prior SE and in 0 of 2 who had not had prior SE that the brain of patients with AHC not only has increased suggests that neuronal loss may be associated with a history of pyramidal cell excitability but also is often unable to stop on- SE (novel). These observations support further investigation of going seizure activity once initiated, presumably leading to drug a potential relationship between SE and MRS changes and resistance and to SE, because of the following: the ATP1A3 of MRS as a potential biomarker in AHC. sodium ATPase pump is specifically important in maintaining the sodium potassium gradient during periods of Therapies and observed responses rapid firing,5,48 and fast-spiking interneurons are also important Our findings are consistent with prior studies that have noted in such situations.5 drug resistance of epilepsy in AHC.9,10 We found in our cohort that as many as 37% (19 of 51 patients) had drug-resistant Limitations epilepsy. We did not observe any apparently more effective Patients in this study may have been specifically referred to antiepileptic medication(s). In addition, our data, albeit limited, our center because they may have had more severe disease. In did not support that ATP or amiloride had definitive antisei- addition, not all patients underwent all tests. We have also zure or other beneficial effects in AHC. For ATP therapy, our commented above on the limitations and exploratory nature doses were much lower, due to cost issues, than those pre- of our drug response data. Nonetheless, this study has the viously used in the prior case report (25 mg/kg/d).17 Our advantage of a hypothesis-driven, comprehensive analysis of observations of a potentially beneficial effect of dextro- the available epilepsy-related manifestations and data of an methorphan warrant further investigation. VNS therapy informative cohort of consecutive patients with AHC. appeared to help seizures and plegic spells in AHC. Our patient who appeared to respond to steroid therapy is the second such Epilepsy in AHC can be generalized or localization related. case in the literature.40 The other case, however, was treated Onset of epileptic seizures is usually after, but may be before, with continuous steroids rather than pulse steroids. Our data the onset of AHC spells. For seizures of focal origin, frontal on the above agents should be interpreted very cautiously due lobes and, less frequently, temporal lobes tend to be more to their open-label and observational nature. In the literature, 2 commonly affected than posterior lobes, and multiple lobes studies and 3 case reports have described the use of the keto- may be involved. EEG can be initially normal and then become – genic diet in AHC.41 45 All 7 treated patients were reported to epileptiform. Epilepsy in AHC is also characterized by drug respond to the diet for paroxysmal spells, 1 of whom was also resistance and by high risk of refractory SE that can occur in the reported to have epilepsy and noted disappearance of seizures. context of attempts of drug withdrawal, stress, or prolonged In our study, 2 of 6 patients appeared to benefit from diet. episodes of AHC spells. Regression after SE is not uncommon. We also found that distinct RAS, which do not have EEG Status epilepticus correlate, are likely of nonepileptic pathophysiology and occur Overall, high frequency (39% of patients with epilepsy had SE), commonly. Our findings are consistent with the recently high recurrence rate (75% of patients with SE had multiple characterized mechanisms of inhibitory-excitatory imbalance episodes), drug resistance, and possible regression characterize leading to enhanced neuronal excitability in AHC animal SE in AHC. Regression was noted in 4 of 12 (33%) of our models. e10 Neurology | Volume 93, Number 13 | September 24, 2019 Neurology.org/N Copyright © 2019 American Academy of Neurology. Unauthorized reproduction of this article is prohibited. Acknowledgment The authors thank all members of the Duke AHC Appendix (continued)

Multidisciplinary Program and Dr. Brain Smith for his Name Location Role Contribution assistance with the statistical analysis. Arsen Duke Author Role in the acquisition of Hunanyan, University, data; revised the Study funding PhD Durham, manuscript for This study was supported by Duke Fund Nos. 4410161 and NC intellectual content 3912247 and by a donation by the Cure AHC Founda- Mohmad Duke Corresponding Planned, designed, and tion (M.A.M.). Mikati, MD University, Author conceptualized the Durham, overall study; NC supervised all aspects of Disclosure the study; interpreted the data; contributed J. Uchitel, A. Helseth, L. Prange, M. McLean, R. Ghusayni, M. heavily to drafting, Sachdev, and A. Hunanyan report no disclosures relevant to editing and revision of the manuscript for the manuscript. M. Mikati reports receiving compensation for intellectual content a letter of support, presented to Xenon Corp to submit to the Food and Drug Administration, supporting orphan disease designation for AHC and approval of studies aimed at getting References Food and Drug Administration approval of flunarizine for 1. Masoud M, Prange L, Wuchich J, Hunanyan A, Mikati MA. Diagnosis and treatment of alternating hemiplegia of childhood. Curr Treat Options Neurol 2017;19:8. AHC. Go to Neurology.org/N for full disclosures. 2. Heinzen EL, Swoboda KJ, Hitomi Y, et al. De novo mutations in ATP1A3 cause alternating hemiplegia of childhood. Nat Genet 2012;44:1030–1034. 3. Heinzen EL, Arzimanoglou A, Brashear A, et al. Distinct neurological disorders with Publication history ATP1A3 mutations. Lancet Neurol 2014;13:503–514. Received by Neurology August 28, 2018. Accepted in final form 4. Hunanyan AS, Fainberg NA, Linabarger M, et al. Knock-in mouse model of alter- May 1, 2019. nating hemiplegia of childhood: behavioral and electrophysiologic characterization. Epilepsia 2015;56:82–93. 5. Hunanyan AS, Helseth AR, Abdelnour E, et al. Mechanisms of increased hippocampal excitability in the Mashl+/− mouse model of Na+/K+ -ATPase dysfunction. Epilepsia 2018;59:1455–1468. 6. Kansagra S, Mikati MA, Vigevano F. Alternating hemiplegia of childhood. Handb Clin Appendix Authors Neurol 2013;112:821–826. 7. Masoud M, Gordon K, Hall A, et al. Motor function domains in alternating hemiplegia Name Location Role Contribution of childhood. Dev Med Child Neurol 2017;59:822–828. 8. Viollet L, Glusman G, Murphy KJ, et al. Alternating hemiplegia of childhood: ret- Julie Duke Author Contributed to design rospective genetic study and genotype-phenotype correlations in 187 subjects from Uchitel University, and conceptualization of the US AHCF registry. PLoS One 2015;10:e0127045. Durham, the study; had the major 9. Panagiotakaki E, Gobbi G, Neville B, et al. Evidence of a non-progressive course of NC role in analysis and alternating hemiplegia of childhood: study of a large cohort of children and adults. acquisition of the data; Brain J Neurol 2010;133(pt 12):3598–3610. interpreted the data; 10. Panagiotakaki E, De Grandis E, Stagnaro M, et al. Clinical profile of patients with performed the statistical ATP1A3 mutations in alternating hemiplegia of childhood: a study of 155 patients. analysis; drafted the Orphanet J Rare Dis 2015;10:123. manuscript for 11. Sasaki M, Ishii A, Saito Y, et al. Genotype-phenotype correlations in alternating intellectual content; hemiplegia of childhood. Neurology 2014;82:482–490. contributed heavily to 12. Fisher RS, Acevedo C, Arzimanoglou A, et al. ILAE official report: a practical clinical the revision of the definition of epilepsy. Epilepsia 2014;55:475–482. manuscript for 13. Kwan P, Arzimanoglou A, Berg AT, et al. Definition of drug resistant epilepsy: intellectual content consensus proposal by the Ad Hoc Task Force of the ILAE Commission on Thera- peutic Strategies. Epilepsia 2010;51:1069–1077. Ashley Duke Author Major role in the 14. Fisher RS, Cross JH, French JA, et al. Operational classification of seizure types by the Helseth, University, acquisition of data; International League against epilepsy: position paper of the ILAE Commission for MD, PhD Durham, interpreted the data; Classification and Terminology. Epilepsia 2017;58:522–530. NC revised the manuscript 15. Katz G, Lazcano-Ponce E. Intellectual disability: definition, etiological factors, clas- for intellectual content sification, diagnosis, treatment and prognosis. Salud Publica Mex 2008;50(suppl 2): s132–141. Lyndsey Duke Author Major role in the 16. Ream MA, Mikati MA. Clinical utility of genetic testing in pediatric drug-resistant Prange, University, acquisition of data; epilepsy: a pilot study. Epilepsy Behav 2014;37:241–248. CPNP Durham, revised the manuscript 17. Ju J, Hirose S, Shi XY, Ishii A, Hu LY, Zou LP. Treatment with oral ATP decreases NC for intellectual content alternating hemiplegia of childhood with de novo ATP1A3 mutation. Orphanet J Rare Dis 2016;11:55. Melissa Duke Author Major role in the 18. Ueda K, Serajee F, Huq AM. Clinical benefit of NMDA receptor antagonists in McLean, University, acquisition of data; – BSBA Durham, revised the manuscript a patient with ATP1A2 gene mutation. Pediatrics 2018;141(suppl 5):S390 S394. NC for intellectual content 19. Ali A, Ahmad FJ, Pillai KK, Vohora D. Evidence of the antiepileptic potential of amiloride with neuropharmacological benefits in rodent models of epilepsy and be- – Ryan Duke Author Role in the acquisition of havior. Epilepsy Behav 2004;5:322 328. ff Ghusayni, University, data; revised the 20. Zhao X, Gorin FA, Berman RF, Lyeth BG. Di erential hippocampal protection when MD Durham, manuscript for blocking intracellular sodium and calcium entry during traumatic brain injury in rats. – NC intellectual content J Neurotrauma 2008;25:1195 1205. 21. Beaug´eLA, DiPolo R. Sidedness of the ATP-Na+-K+ interactions with the Na+ pump Monisha Duke Author Role in the acquisition of in squid axons. Biochim Biophys Acta 1979;553:495–500. Sachdev, University, data; revised the 22. Dubey D, Kalita J, Misra UK. Status epilepticus: refractory and super-refractory. MD Durham, manuscript for Neurol India 2017;65(suppl):S12–S17. NC intellectual content 23. LaFrance WC, Baker GA, Duncan R, Goldstein LH, Reuber M. Minimum require- ments for the diagnosis of psychogenic nonepileptic seizures: a staged approach:

Neurology.org/N Neurology | Volume 93, Number 13 | September 24, 2019 e11 Copyright © 2019 American Academy of Neurology. Unauthorized reproduction of this article is prohibited. a report from the International League against Epilepsy Nonepileptic Seizures Task 36. Kanavakis E, Xaidara A, Papathanasiou-Klontza D, Papadimitriou A, Velentza S, Force. Epilepsia 2013;54:2005–2018. Youroukos S. Alternating hemiplegia of childhood: a syndrome inherited with an 24. Popkirov S, Jungilligens J, Gr¨onheitW, Wellmer J. Diagnosing psychogenic non- autosomal dominant trait. Dev Med Child Neurol 2003;45:833–836. epileptic seizures: video-EEG monitoring, suggestive seizure induction and diagnostic 37. Yamashita S, Hamano SI, Tanaka M, Yoshinari S, Minamitani M, Hayakawa M. Single certainty. Epilepsy Behav 2017;73:54–58. photon emission computed tomography findings in a case of alternating hemiplegia of 25. Saito Y, Inui T, Sakakibara T, Sugai K, Sakuma H, Sasaki M. Evolution of hemiplegic childhood in relation to migraine [in Japanese]. No To Hattatsu 2005;37:413–418. attacks and epileptic seizures in alternating hemiplegia of childhood. Epilepsy Res 38. Pavlidis E, Uldall P, Madsen CG, et al. Alternating hemiplegia of childhood and 2010;90:248–258. a pathogenic variant of ATP1A3: a case report and pathophysiological considerations. – 26. Yang X, Gao H, Zhang J, et al. ATP1A3 mutations and genotype-phenotype corre- Epileptic Disord 2017;19:226 230. lation of alternating hemiplegia of childhood in Chinese patients. PLoS One 2014;9: 39. Giacanelli M, Petrucci A, Lispi L, et al. ATP1A3 mutant patient with alternating e97274. hemiplegia of childhood and brain spectroscopic abnormalities. J Neurol Sci 2017; – 27. Mikati MA, Kramer U, Zupanc ML, Shanahan RJ. Alternating hemiplegia of child- 379:36 38. hood: clinical manifestations and long-term outcome. Pediatr Neurol 2000;23: 40. Wong VCN, Kwong AKY. ATP1A3 mutation in a Chinese girl with alternating – 134–141. hemiplegia of childhood: potential target of treatment? Brain Dev 2015;37:907 910. 28. Sasaki M, Sakuma H, Fukushima A, Yamada K, Ohnishi T, Matsuda H. Abnormal 41. Pisciotta L, Gherzi M, Stagnaro M, et al. Alternating hemiplegia of childhood: pharmacological treatment of 30 Italian patients. Brain Dev 2017;39:521–528. cerebral glucose metabolism in alternating hemiplegia of childhood. Brain Dev 2009; 42. Roubergue A, Philibert B, Gautier A, et al. Excellent response to a ketogenic diet in 31:20–26. a patient with alternating hemiplegia of childhood. JIMD Rep 2015;15:7–12. 29. Mikati M, Fischman A. Positron emission topography in children with alternating 43. Vila-Pueyo M, R, Raspall-Chaure M, et al. Clinical and genetic analysis in hemiplegia of childhood. In: Alternating Hemiplegia of Childhood: International alternating hemiplegia of childhood: ten new patients from Southern Europe. J Neurol Review of Child Neurology: New York: Raven Press; 1995. Sci 2014;344:37–42. 30. Pacheva I, Ivanov I. Acute confusional migraine: is it a distinct form of migraine? Int J 44. Ulate-Campos A, Fons C, Artuch R, et al. Alternating hemiplegia of childhood with – Clin Pract 2013;67:250 256. a de novo mutation in ATP1A3 and changes in SLC2A1 responsive to a ketogenic 31. Saito Y, Sakuragawa N, Sasaki M, Sugai K, Hashimoto T. A case of alternating diet. Pediatr Neurol 2014;50:377–379. – hemiplegia of childhood with cerebellar atrophy. Pediatr Neurol 1998;19:65 68. 45. Schirinzi T, Graziola F, Cusmai R, et al. ATP1A3-related epileptic encephalopathy 32. Sasaki M, Ishii A, Saito Y, Hirose S. Progressive brain atrophy in alternating hemi- responding to ketogenic diet. Brain Dev 2018;40:433–438. – plegia of childhood. Mov Disord Clin Pract 2017;4:406 411. 46. S´anchezFern´andez I, Ga´ınza-Lein M, Abend NS, et al. Factors associated with 33. Sweney MT, Silver K, Gerard-Blanluet M, et al. Alternating hemiplegia of childhood: treatment delays in pediatric refractory convulsive status epilepticus. Neurology 2018; early characteristics and evolution of a neurodevelopmental syndrome. Pediatrics 90:e1692–e1701. 2009;123:e534–e541. 47. Bøttger P, Tracz Z, Heuck A, Nissen P, Romero-Ramos M, Lykke-Hartmann K. 34. Shiota N, Shimono M, Tomioka S, et al. A boy with nystagmus, refractory dystonia Distribution of Na/K-ATPase alpha 3 isoform, a sodium-potassium P-type pump and apneic attack due to alternating hemiplegia of childhood [in Japanese]. No To associated with rapid-onset of dystonia parkinsonism (RDP) in the adult mouse brain. Hattatsu 2007;39:285–288. J Comp Neurol 2011;519:376–404. 35. Zhang Y, Sun W, Qin J, Jiang Y, Wu X. Clinical characteristics of alternating hemi- 48. Simmons CQ, Thompson CH, Cawthon BE, et al. Direct evidence of impaired plegia of childhood in 13 patients [in Chinese]. Zhonghua Er Ke Za Zhi 2003;41: neuronal Na/K-ATPase pumpnction in alternating hemiplegia of childhood. Neu- 680–683. robiol Dis 2018;115:29–38.

e12 Neurology | Volume 93, Number 13 | September 24, 2019 Neurology.org/N Copyright © 2019 American Academy of Neurology. Unauthorized reproduction of this article is prohibited. The epileptology of alternating hemiplegia of childhood Julie Uchitel, Ashley Helseth, Lyndsey Prange, et al. Neurology published online September 4, 2019 DOI 10.1212/WNL.0000000000008159

This information is current as of September 4, 2019

Updated Information & including high resolution figures, can be found at: Services http://n.neurology.org/content/early/2019/09/04/WNL.0000000000008 159.full Subspecialty Collections This article, along with others on similar topics, appears in the following collection(s): All Epilepsy/Seizures http://n.neurology.org/cgi/collection/all_epilepsy_seizures All Genetics http://n.neurology.org/cgi/collection/all_genetics All Pediatric http://n.neurology.org/cgi/collection/all_pediatric MRI http://n.neurology.org/cgi/collection/mri MRS http://n.neurology.org/cgi/collection/mrs Permissions & Licensing Information about reproducing this article in parts (figures,tables) or in its entirety can be found online at: http://www.neurology.org/about/about_the_journal#permissions Reprints Information about ordering reprints can be found online: http://n.neurology.org/subscribers/advertise

Neurology ® is the official journal of the American Academy of Neurology. Published continuously since 1951, it is now a weekly with 48 issues per year. Copyright © 2019 American Academy of Neurology. All rights reserved. Print ISSN: 0028-3878. Online ISSN: 1526-632X.