CLINICAL ARTICLE J Neurosurg 128:277–286, 2018

Efficacy and safety of corpus callosotomy after vagal stimulation in patients with drug-resistant

Jennifer Hong, MD,1 Atman Desai, MD,2 Vijay M. Thadani, MD,3 and David W. Roberts, MD1,3

1Section of , Department of Surgery, 3Department of Neurology, Dartmouth-Hitchcock Medical Center, Lebanon, New Hampshire; and 2Department of Neurosurgery, Stanford University School of Medicine, Palo Alto, California

OBJECTIVE Vagal nerve stimulation (VNS) and corpus callosotomy (CC) have both been shown to be of benefit in the treatment of medically refractory epilepsy. Recent case series have reviewed the efficacy of VNS in patients who have undergone CC, with encouraging results. There are few data, however, on the use of CC following VNS therapy. METHODS The records of all patients at the authors’ center who underwent CC following VNS between 1998 and 2015 were reviewed. Patient baseline characteristics, operative details, and postoperative outcomes were analyzed. RESULTS Ten patients met inclusion criteria. The median follow-up was 72 months, with a minimum follow-up of 12 months (range 12–109 months). The mean time between VNS and CC was 53.7 months. The most common reason for CC was progression of after VNS. Seven patients had anterior CC, and 3 patients returned to the operat- ing room for a completion of the procedure. All patients had a decrease in the rate of falls and drop seizures; 7 patients experienced elimination of drop seizures. Nine patients had an Engel Class III outcome, and 1 patient had a Class IV outcome. There were 3 immediate postoperative complications and 1 delayed complication. One patient developed pneumonia, 1 developed transient mutism, and 1 had persistent weakness in the nondominant foot. One patient pre- sented with a wound infection. CONCLUSIONS The authors demonstrate that CC can help reduce seizures in patients with medically refractory epi- lepsy following VNS, particularly with respect to drop attacks. https://thejns.org/doi/abs/10.3171/2016.10.JNS161841 KEY WORDS atonic ; corpus callosotomy; drug-resistant epilepsy; vagal nerve stimulation

agal nerve stimulation (VNS) and corpus calloso- be patients with secondarily generalized seizures or mul- tomy (CC) are palliative surgical procedures that tiple seizure types.28,34 In the absence of Level I evidence can reduce seizure frequency, attenuate seizure comparing VNS to CC, the experience and preferences of Vseverity, and improve quality of life for patients with non- individual neurologists and surgeons heavily influence the localizing or unresectable drug-resistant epilepsy (DRE). recommendation for either VNS or CC in patients with This is in contrast to focal resection, where the goal is DRE who present for palliative surgery. complete control of seizures. Patient selection criteria Corpus callosotomy, which was first described by Van for VNS and for CC are not strictly defined. In general, Wagenen and Herren in the 1940s, predates by many de- VNS is recommended for patients with nonlocalizable cades the FDA approval of VNS for epilepsy in 1997.41 As or unresectable seizure disorders or those who either are a result, much of the literature describing the efficacy of unable to tolerate or choose not to undergo , CC for DRE comes from patients who did not have the op- and is FDA approved for children older than 12 years and tion of VNS. The mechanisms of action of each treatment adults with multifocal epilepsy and multiple seizure types. are very different. Callosotomy is a disconnection proce- Patients who predominantly have drop attacks are typi- dure wherein fibers that allow seizures to propagate and cally thought to be very good candidates for CC, as may secondarily generalize are severed.13,51 In contrast, VNS is

ABBREVIATIONS CC = corpus callosotomy; CPS = complex partial seizures; DRE = drug-resistant epilepsy; EEG = ; GTC = generalized tonic- clonic; LGS = Lennox-Gastaut syndrome; LITT = laser interstitial thermal therapy; SPS = simple partial seizures; SRS = stereotactic radiosurgery; VNS = vagal nerve stimu- lation. SUBMITTED July 14, 2016. ACCEPTED October 4, 2016. INCLUDE WHEN CITING Published online March 3, 2017; DOI: 10.3171/2016.10.JNS161841.

©AANS 2018, except where prohibited by US copyright law J Neurosurg Volume 128 • January 2018 277

Unauthenticated | Downloaded 09/30/21 06:21 PM UTC J. Hong et al. a form of neuromodulation, in which continuous, intermit- in the direction of rapid cycling, until stimulation occurred tent, low-intensity electrical stimulation of the vagal nerve every 1.1 minute for 14 seconds. Patients underwent a full results in afferent spread of action potentials to subcortical year of VNS therapy before they were considered to have nuclei and the cortex; this is thought to attenuate the dura- had an adequate trial. tion and frequency of seizures.49 It is plausible that the 2 approaches are complementary Operative Technique and can work synergistically in particularly refractory pa- All CCs were performed after induction of general tients. Based on data from the VNS registry, 57% of pa- anesthesia. Patients were usually positioned supine with tients with DRE who underwent VNS after CC reported a 20 the head neutral and held in 3-point fixation by using a > 50% reduction in seizure frequency. The potential ben- Mayfield clamp. The scalp was partially shaved, and a efit of CC in patients who have already undergone VNS is transverse incision was made anterior to the coronal su- not well described. Here we review our institutional expe- ture, with two-thirds of the incision placed right of mid- rience with patients who elected to undergo CC after VNS line. A pneumatic-powered perforator and craniotome to test the hypothesis that in patients with seizures who (or, earlier in the series, a 5-cm trephine) were used to have undergone VNS without satisfactory response, CC elevate a 5-cm-long × 4-cm-wide bone flap, of which two- may produce meaningful further benefit. thirds were to the right of midline. The was opened in a curvilinear fashion based usually on the right Methods side of the superior sagittal sinus. Using a microsurgical Participants and Study Design technique, dissection was continued down the interhemi- Following approval by the Dartmouth Institutional Re- spheric fissure, with exposure of the at view Board, hospital records and a prospectively main- its base. A Greenberg self-retaining retractor was placed. tained epilepsy database were queried for all patients who The CC was performed extraventricularly by using a underwent CC after implantation of a VNS between 1999 microdissector, commonly with the assistance of micro- and 2015. The Current Procedural Terminology codes for suction, beginning in the midline at the junction of the both VNS (64568, 64573) and CC (61541) and the ICD-9 posterior genu and the anterior body. The dissection was code for epilepsy (345.XX) were used to identify eligible continued down to the midline cleft formed by the leaves patients. Patients were included in the study only if the of the septum pellucidum, and the existing or potential VNS surgery predated the CC surgery. cavum septum pellucidum was entered. Dissection was then continued anteriorly around the genu and posteriorly Clinical Variables and Outcomes Assessment to the narrowed isthmus of the posterior body. A ligat- ing clip was placed at the most posterior extent of the Clinical data were retrieved through retrospective CC. Hemostasis was obtained and the dura was closed chart review. Data available included age, sex, presenting with 4-0 Vicryl suture. The bone flap was replaced with symptoms, duration of seizures, seizure subtypes, medi- titanium hardware. The galea and skin were then reap- cations, VNS parameters, clinical examination, opera- proximated. Blood loss was generally < 100 ml, and the tive notes, postoperative hospital course, and outpatient time of surgery was between 2.5 and 4 hours. Barring follow-up. complications, discharge was most often on postoperative All patients were evaluated in a multidisciplinary sur- Day 3 or 4. gical epilepsy conference prior to the CC. Routine pre- Completion CCs were undertaken in a similar manner, surgical workup included history and physical examina- usually with a second, more posterior craniotomy. If the tion, long-term noninvasive video electroencephalography patient had undergone prior intracranial EEG or resection, (EEG), and imaging including MRI as well as ictal and every effort was made to anticipate a potential CC at the interictal SPECT in some patients. The recommendation first operation, or, if the initial surgery was performed at for CC was made based on persistence of drop attacks or another institution, to incorporate the previous incision secondarily generalized seizures as well as failure of VNS and bone flap into the CC surgery. This surgery is facili- therapy, defined as a < 50% reduction in seizure frequency tated by the deeper extending posterior falx and smaller with maximal titration of stimulation. residual callosum. Patients were seen in clinic for routine follow-up ap- Protocol for VNS Titration proximately 4 weeks after discharge from the hospital. All patients who underwent VNS implantation at our institution were evaluated at a multidisciplinary epilepsy Results conference. Postoperatively, adjustments were made to VNS stimulation parameters every 6 weeks as tolerated, Patient Demographic Data and then every 3 months when response to therapy began We identified 10 patients in our epilepsy registry who to plateau, or sooner if side effects became pronounced and had undergone CC following VNS. Eight patients under- intolerable. The frequency of 30 cycles per second (range went VNS placement at our institution. The median age 20–30), and pulse width 250 or 500 μsec, was usually held at the time of consideration for CC was 37 years (range constant. The current was titrated up from 0.25 mA by 8–60 years). Full-scale IQ ranged from low average to ex- 0.25 mA at each visit. After patients reached 1.5 or 1.75 tremely low. There were 4 female and 6 male patients. The mA, which was the maximum tolerated in this group of median age of seizure onset was 8.5 years (range 1–20 patients, the duty cycle was increased in a systematic way years). All patients underwent 3 or more medications trials

278 J Neurosurg Volume 128 • January 2018

Unauthenticated | Downloaded 09/30/21 06:21 PM UTC Callosotomy after vagal nerve stimulation

TABLE 1. Demographic and clinical data in 10 patients with DRE Age Med Med Age at Sz Trials Trials Case (yrs), Onset Sz Prior to Prior to EEG Imaging Full-Scale No. Sex (yrs) Types* VNS CC Findings Results Etiology IQ 1 40, F 6 GTC, CPS Carbamazepine, gaba- Carbamazepine, Bifrontal spikes MRI: normal; SPECT: Posttrau- 81 pentin, clonazepam tiagabine, foci in rt basal matic Mysoline ganglia, insula, & pst lat frontal lobe 2 60, F 11 CPS, Phenytoin, valpro- Pregabalin, rufin- Possible LGS MRI: bilat band Juvenile 58 atonic ate, levetiracetam, amide heterotopia myo- zonisamide, oxcar- clonic bazepine epilepsy 3 42, M 20 GTC, CPS, Valproate, lamotrigine, Valproate, topira- Multifocal MRI: bilat hippocam- Viral en- 59 SPS topiramate, clonaz- mate, lamotri­ pal atrophy cepha- epam gine, primidone litis 4 31, M 7 CPS, GTC, Carbamazepine Carbamazepine, Multifocal MRI: extensive lt Cortical 58 SPS lamotrigine polymicrogyria dysplasia 5 39, M 10 GTC, Lamotrigine, Lamotrigine, Bilat fast sharp MRI: pst lt frontal Idiopathic Low nor- atonic zonisamide, loraz- zonisamide, waves encephalomalacia mal epam lorazepam 6 24, M 3 GTC, CPS, Phenytoin, lamotrigine, Phenytoin, Polymorphic MRI: normal LGS Severe ab- clonazepam lamotrigine, slowing, gen- retar- sence, clonazepam eralized spike dation atonic & wave bilat 7 35, F 15 GTC, CPS, Carbamazepine, la- Carbamazepine, Bilat spikes, MRI: extensive gray Bilat 88 vocal- motrigine, valproate lamotrigine, predominance matter heteroto- periven- ization, rufinamide for lt side pia; SPECT: foci in tricular SPS pst lt frontal lobe, heteroto- medial rt temporal pias 8 32, M 4 Absence, Carbamazepine, val- Phenobarbital, Bifrontal sharp MRI: normal Idiopathic 73 GTC, proate, felbamate, valproate, theta evolving atonic phenobarbital, rufinamide, into bilat gabapentin felbamate generalized spike-wave discharge 9 40, F 16 GTC, ab- Lamotrigine, phe- Valproate, topira- Slow back- MRI: normal Idiopathic 61 sence, nytoin, zonisamide, mate, phenytoin, ground, CPS valproate zonisamide, generalized lamotrigine sharp waves 10 8, M 1 Atonic, ACTH, levetiracetam, Felbamate, leveti- Hypsarrhythmia, MRI: normal LGS Severe GTC, prednisone, gaba- racetam, valpro- generalized retar- CPS pentin, clobazam, ate, tiagabine, slowing, dation clonazepam, pheno- rufinamide, predominantly barbital, phenytoin, perampanel, frontal diazepam, peram- zonisamide, panel, zonisamide, topiramate topiramate Median 8.5 ACTH = adrenocorticotropic hormone; Med = medication; pst = posterior; Sz = seizure. * Listed according to frequency. prior to initial VNS placement, 3 patients underwent more and bilateral EEG abnormalities. Cortical dysplasia was than 5 medication trials, and 1 patient had 12 medication identified in 2 patients, and 1 patient each had juvenile trials. The cause of epilepsy was unidentified in 3 patients: myoclonic epilepsy, posttraumatic epilepsy, and a history 1 had an area of encephalomalacia that was not epilepto- of viral encephalitis (Table 1). Two patients had Lennox- genic, and the other 2 had normal results on MRI studies Gastaut syndrome (LGS).

J Neurosurg Volume 128 • January 2018 279

Unauthenticated | Downloaded 09/30/21 06:21 PM UTC J. Hong et al.

TABLE 2. Surgical management in 10 patients with DRE who underwent CC following VNS Age Duration of VNS Current Age Time Btwn Clinical Other Case at VNS Sz Prior to Settings Prior VNS at CC VNS & CC FU Surgical No. (yrs) VNS (yrs) to CC* Settings* (yrs) (mos) (mos) Procedures 1 27 21 0.75 mA/30 sec/108 Off 29 24 106 Previous resec- sec/25 Hz tion; intracra- nial EEG study 2 49 38 2.25 mA/21 sec/300 1.75 mA/30 sec/300 54 60 79 None sec/30 Hz sec/20 Hz 3 26 6 1.25 mA/30 sec/300 1.0 mA/21 sec/180 32 (ant 2/3); 35 108 92 None sec/30 Hz sec/30 Hz (completion) 4 18 11 1.5 mA/14 sec/108 1.5 mA/14 sec/120 23 71 99 Intracranial EEG sec/25 Hz sec/30 Hz study 5 26 16 1.0 mA/14 sec/148 1.25 mA/14 sec/120 28 28 65 Limited resection sec/30 Hz sec/25 Hz at another institution 6 12 9 2.5 mA/30 sec/148 3.0 mA/20 sec/66 15 (ant 2/3); 15 45 109 None sec/10 Hz sec/10 Hz (completion) 7 23 8 Off due to hoarse- Explanted 31 108 33 None ness 8 35 31 0.75 mA/21 sec/148 Off; magnet 0.5 38 31 51 None sec/25 Hz mA/30 sec 9 34 18 1.5 mA/21 sec/180 1.5 mA/14 sec/108 37 32 39 None sec/25 Hz sec/20 Hz 10 5 4 3.0 mA/20 sec/148 3.0 mA/20 sec/148 7 (ant 2/3); 8 30 12 None sec/25 Hz sec/25 Hz (completion) Mean (SD) 25.5 (12.39) 16.2 (11.15) 29.8 (10.21) 53.7 (32.32) 68.5 (33.79) Median 26 13.5 31 38.5 72 Ant = anterior; FU = follow-up. * Values for VNS settings are expressed as voltage (mA)/time on (sec)/time off (sec)/cycling (Hz).

The EEG and Imaging Findings resection, but failure to localize a resectable focus led to The EEG and imaging findings are summarized in a recommendation for CC. One patient had a previous re- Table 1. Briefly, the EEG studies in all patients showed section at another hospital several years before presenting multifocal or generalized epileptiform abnormalities, but for evaluation at our institution. Three patients underwent there was no predominant pattern. Hypsarrhythmia was completion of their initial anterior two-thirds CC. seen in 2 patients. The MR images were negative for any The median age at VNS was 26 years (range 5–49 lesion in 5 patients (50%). Of the remaining patients, MRI years). The median duration of seizures prior to VNS was demonstrated bilateral findings in 3 patients and unilat- 13.5 years (range 4–38 years). The median age at CC was eral lesions in the other 2. Of the 2 patients with focal 31 years (range 8–54 years), resulting in an average time MRI findings, one patient had polymicrogyria in eloquent between VNS and CC of 53.7 months (SD 32.32 months). cortex that was unresectable, and the other patient had an The median clinical follow-up after CC was 72 months area of encephalomalacia that did not appear to be epilep- (range 12–109 months). This is summarized in Table 2. togenic based on intracranial EEG. Two patients under- Four patients had complications following CC. One went ictal and interictal SPECT studies; in both patients patient had persistent weakness in his left foot, although multiple regions of hyperperfusion were identified on ictal he was ambulating independently at last follow-up. One imaging. patient developed transient hemiparesis that resolved. One patient had a wound infection following intracranial moni- Neurosurgical Interventions toring and CC. One patient had pneumonia. Three patients All patients with epilepsy who were surgical candidates were discharged to rehabilitation following CC. The aver- were reviewed at a multidisciplinary epilepsy conference. age hospital length of stay was 4.1 days (Table 3). A recommendation for CC following VNS was based on failure to improve with VNS therapy or a significant aton- Seizure Outcomes ic seizure history. Several patients had additional surger- At a median follow-up of 72 months, 9 of 10 patients ies. Two patients underwent intracranial electrode inves- had Engel Class III outcomes (Table 3). One patient had a tigational studies at our institution with the goal of focal Class IV outcome. With respect to seizure frequency, half

280 J Neurosurg Volume 128 • January 2018

Unauthenticated | Downloaded 09/30/21 06:21 PM UTC Callosotomy after vagal nerve stimulation Home Rehab Home Home Home Rehab Home Home Home Home Home Home Rehab Disposition 4 6 3 3 4 7 4 2 4 4 7 3 5 LOS (days) 4 (2–7) Hospital 4.1 (1.52) - hemiparesis paresis Complications None Pneumonia None None None Postop stroke, stroke, Postop infection Wound None None None None None hemi Transient III III IV III III III III III III III III III III Class Engel Engel of No. Meds Unchanged Unchanged Increased Unchanged Increased Unchanged Unchanged Decreased Unchanged Increased Unchanged Decreased Decreased Falls Atonic Sz Frequency or Absent Absent Absent Decreased >50% No change Absent Unchanged Absent Unchanged Absent Unchanged Absent Decreased >50% Sz Frequency of drop szs, which recurred No change Decreased <50% Increased No change Initial elimination Decreased >50% Decreased >50% Decreased >50% Decreased >50% Decreased <50% Decreased >50% Decreased >50% Decreased >50% 0 0 0 3 7 0 0 0 0 0 1 10 30 0 (0–30) 0.4 (0.96) at Lastat FU Atonic Szs or Falls/Moor 2 2 1 1 28 20 20 40 56 84 at 112 100 100 Szs/Mo Last FU 23.77 (28.93) 23.77 20 (1 to >500)20 (1 7 1 4 7 0 4 7 2 5 10 10 30 30 4 (0–30) Atonic Szs 6.33 (9.68) Prior to CC or Falls/Moor 8 5 28 28 20 40 40 40 84 112 112 100 CC >500 Prior to Szs/Mo 40.55 (35.45) 28 (5 to >500)28 (5 8 24 24 40 480 NA NA NA VNS >500 Prior to Szs/Mo Unknown Unknown Unknown Unknown 24 (8 to >500) (8 24 115.2 (204.24) 115.2 No. Case (range) Indicates completion of anterior two-thirds CC. 8 9 7 4 5 6 2 3 1 3* 6* 10 10* Median Mean (SD) TABLE 3. SeizureTABLE outcomes and operative complications patients in 10 with DRE who underwent CC following VNS = lengthLOS of stay; NA = not applicable; rehab = rehabilitation. *

J Neurosurg Volume 128 • January 2018 281

Unauthenticated | Downloaded 09/30/21 06:21 PM UTC J. Hong et al. of patients reported a > 50% decrease in seizure frequency. ment because it does not require craniotomy or prolonged Two patients had a < 50% decrease in seizure frequency, inpatient hospitalization. If these patients continue to have 2 patients had no change, and 1 patient had an increase debilitating seizures despite adequate VNS therapy and in seizure frequency. With respect to atonic seizures and medication titration, the question arises whether a second falls, 7 patients had complete elimination of this seizure surgical procedure would be beneficial. subtype following CC; 2 patients had a > 50% reduction; and 1 patient had no history of atonic seizures and did not Is There an Additional Benefit to CC in Patients Previously develop them following CC. The number of medications Treated With VNS? was reduced in 3 patients, unchanged in 5 patients, and increased in 2 patients. Recently, Guillamón et al. and Arya et al. have reported small case series with patients who were treated with CC Three patients initially underwent an anterior two- 1,15 thirds CC that was ultimately completed. Seizure out- following VNS in the setting of severe DRE. In the 3 comes following the first surgery were Engel Class III in patients reviewed by Guillamón et al., a 66.6%, 97.1%, these patients. The first patient had elimination of daytime and 98.3% decrease in seizure frequency was reported; 1 seizures and a brief period of respite from atonic seizures; patient suffered a wound infection. Arya et al. found that unfortunately, the atonic seizures recurred 4 months after their group of 7 patients had widely varying responses, from complete seizure freedom to a seizure frequency anterior CC, and he developed side effects from his antiep- reduction of 15%. In reviewing seizure outcomes for 58 ileptic drugs that resulted in down-titration of his dosage pediatric patients after CC, a larger group of 22 pediatric and increasing seizure frequency. He underwent comple- patients who underwent CC following VNS was reported tion of his CC 3 years after his initial procedure and had by Kasasbeh et al.; however, selection criteria for pursu- a decrease in seizure frequency of > 50%, with complete ing additional surgery were not explicitly stated.21 Seizure resolution of atonic seizures. The other 2 patients were outcomes were reported after CC, but the initial response pediatric patients with LGS who had brief reductions in to VNS in these patients is unknown. all seizure types, with elimination of atonic seizures. Both In our 10-patient cohort, 9 patients had an Engel Class patients had a return of drop seizures within 2 months of III outcome following CC, at a median clinical follow-up having an anterior two-thirds CC and proceeded to com- of 72 months. Although Class III is generally considered pletion of the CC within 1 year. Both patients reported to be a suboptimal outcome for surgically treated epilepsy, decreased frequency of atonic seizures after completion analysis of this particular group of patients is susceptible of their CC; 1 patient had elimination of atonic seizures, to floor effects because their underlying disease is severe and 1 patient had a decrease of > 50%. All patients with and the goal of surgery is palliative. After we subclassified completed CC had Engel Class III outcomes. outcomes to reflect a reduction in seizure frequency, half of our patients reported a > 50% decrease in seizures—an out- come that was not achieved with long-term VNS therapy. Discussion Outcomes appear to be much more favorable for CC Vagal nerve stimulation and CC are palliative proce- following VNS if atonic seizures or falls are specifically dures that can improve seizure frequency and quality of considered. Of 9 patients with atonic seizures prior to CC, life in patients who have nonlocalizing DRE. Patients who 7 had complete resolution of atonic seizures, 1 had a sig- present for surgical intervention are typically difficult to nificant decrease from 30 per month to 3 per month (Case treat, and they often have debilitating seizures of mul- 6* in Table 3), and 1 had a decrease from 5 per month to tiple types as well as significant medical and psychiatric 1 per month (Case 2 in Table 3). Because falls can be a comorbidities. The decision to proceed to either VNS or significant source of morbidity with epilepsy, an elimina- CC is complex. A CC is an intracranial procedure, with tion or significant reduction of atonic seizures and falls is a infrequent but potentially serious neurological complica- meaningful goal. One potential implication of this result is tions, that appears to be particularly effective for reducing that CC might be pursued prior to VNS for a patient with atonic seizures.2,6,26,27,31,43,46 The response to CC is immedi- LGS. In individuals with LGS, VNS reduced seizures > ate, and patients who have undergone CC require no long- 50% for 8 of 15 patients who had previously undergone term maintenance of a device. CC, but atonic seizures did not respond.4,22 Therefore, of- In comparison, VNS is a less invasive, often outpatient fering CC before VNS could expedite treatment of atonic procedure, which has been shown to shorten seizures and seizures in these patients, who could continue to have dan- the postictal period, and to provide nonseizure-related gerous falls even after VNS. benefits with regard to mood, alertness, and quality of Whether the observed benefit of CC following VNS life.10,16,17,29 Patients with VNS implants often require sev- is synergistic is impossible to determine from this series. eral rounds of programming over weeks or months to It is possible that these patients would have had similar achieve maximum therapeutic effect. In addition, patients outcomes following CC even if they had never undergone are committed to routine surgical VNS generator revisions VNS. The retrospective nature of our analysis limits any every 5–10 years, for the duration of their therapy. conclusions regarding whether seizure outcomes follow- Both treatments are well tolerated. Long-term follow- ing VNS and CC are better than the expected summed up of patients with both CC and VNS demonstrate durable benefits for each procedure performed independently. In seizure control and low morbidity, with high rates of pa- patients who have had CC prior to VNS, 57% were re- tient satisfaction.35,45,47 Typically, patients, their families, sponders (> 50% decrease in seizures). Compared with and treating physicians may prefer VNS as an initial treat- patients who had no prior procedures, this response rate

282 J Neurosurg Volume 128 • January 2018

Unauthenticated | Downloaded 09/30/21 06:21 PM UTC Callosotomy after vagal nerve stimulation is lower (57% vs 65%); however, the patients who did have Guillamón et al. selected patients who had suffered previous procedures were typically more complicated falls as candidates for CC, whereas Arya et al. had only 2 cases, with longer durations of DRE, which confounds the patients of 7 in their series who had documented drop at- comparison. Overall, it appears that using VNS after CC tacks. This may account for the discrepancy in the seizure in patients with DRE does not result in a greater response response rates after CC between the 2 series. Comparisons rate compared with patients who have VNS alone. In pa- between the efficacy of VNS and CC for patients who have tients with LGS, VNS reduced seizures > 50% for 8 of atonic seizures have generally favored CC. A recent meta- 15 patients who had previously undergone CC. All seizure analysis determined that patients who undergo CC are 1.5 types responded to VNS, with the exception of atonic sei- times more likely to have a > 50% reduction in seizures zures.4,22 than are those treated with VNS,40 and another meta-anal- ysis found that in patients with LGS, those who underwent Patient Selection CC had significantly better outcomes for atonic seizures than those who received VNS (80% of patients vs 26.3% There are no clear criteria for determining who among reported a > 50% reduction in drop attacks; p < 0.05), but the growing cohort of patients with DRE and VNS is a there was no difference between the 2 procedures with re- good candidate for CC. Aside from the absolute require- spect to other seizure types.25 On the other hand, Cukiert ment of being able to tolerate major surgery, indications et al. reported that VNS was more effective than CC at for CC following VNS are incompletely understood. Gen- reducing myoclonic seizures.8 erally, patients must have debilitating seizures that con- These pooled analyses suggest that patients who have tinue to be refractory to their current therapy, but there had VNS but who continue to experience persistent drop appears to be no consensus on what constitutes a medical attacks are likely to derive significant benefit from CC. failure of VNS therapy. Guillamón et al. defined failure as That is reflected in our patient cohort as well. There may persistent falls, presumably from atonic seizures; in con- be other prognostic factors in this patient population; how- trast, Arya et al. used the criterion of inability to achieve ever, larger, prospectively followed cohorts would be nec- a > 50% decrease in seizure frequency with VNS, which essary to tease out significant variables. has been suggested by an international consortium.48,50 In both series, the reported final stimulation parameters for patients were what would be considered maximal. The Completion of Initial Anterior Two-Thirds CC VNS output currents were up to 3 mA (range 1.7–3 mA), Three patients underwent completion of the CC in our with frequency of 30 cycles per second (range 20–30), and series. The first was an adult patient with mental retarda- pulse width 250 or 500 μsec. The maximum time with tion who underwent an anterior CC with good response, VNS on was 30 seconds (range 7–30 seconds), and time but who suffered recurrence of atonic seizures and decid- with it off ranged from 21 seconds to 5 minutes. Patients ed in favor of completion. The remaining 2 patients were in both series had completed at least 1 year of VNS prior children with LGS, who had persistent atonic seizures and to CC, which is generally sufficient time for optimization falls despite anterior CC. of programming. Although it is difficult to predict an individual patient’s Our experience recapitulates these programming pa- response to CC, in the case of young patients with LGS in rameters; interestingly, 3 patients had their VNS turned whom VNS fails, an argument can be made for an initial off or explanted at last follow-up. One of these patients had complete CC rather than a staged procedure. These patients her device turned off prior to CC due to hoarseness; an- are generally less highly functioning at baseline and less other patient had her VNS turned on and off sporadically likely to be seriously affected by the cognitive problems because she saw no improvement in seizure control; and sometimes seen with disconnection syndromes, which are the last patient had very low stimulation intensities prior thought to be less likely by limiting CC to the anterior por- to CC, due to development of hoarseness and coughing. tions of the structure. Furthermore, a single surgery avoids After CC, the latter 2 patients elected to have their VNS the risks of an additional treatment with general anesthetic turned off or explanted because seizure control improved and surgery. Finally, in our patients, completion of the CC postoperatively. In other words, CC may be considered reduced the rate of atonic seizures and falls beyond what as a salvage option in patients who are nonresponders to was accomplished by an anterior two-thirds CC. Delaying VNS due to side effects, as opposed to an additive therapy. complete CC may harm this group of patients because it Seven patients in our series continued to use their VNS exposes them to greater risk of injury during the interim after CC. Two of these patients were small children di- period. Indeed, due to persistent falls, both patients with agnosed with LGS who did not demonstrate any adverse LGS returned to the operating room for completion CC reactions to VNS, although seizure control was minimal within 1 year of their first operation. An augmented im- even with high stimulation intensities (2.5 mA and 3.0 provement in seizure control attained by completion of a mA). Five patients reported modest improvement of sei- prior anterior CC is well described,12,21,24,38,39,44 and may be zure control with VNS, particularly those with complex unnecessarily postponed in a staged approach. partial seizures (CPS) and simple partial seizures (SPS), Kasasbeh et al. recently reported that in pediatric pa- in whom the magnet was useful for breaking a seizure or tients, up-front complete CC was superior to anterior CC averting generalization, and they therefore elected to con- for control of all seizure types except for SPS.21 They also tinue with stimulation. The primary indications for CC in found that patients who underwent up-front complete CC these patients were persistence of falls and atonic seizures were less likely to have subsequent VNS implantation despite VNS. compared with patients with anterior CC, although this

J Neurosurg Volume 128 • January 2018 283

Unauthenticated | Downloaded 09/30/21 06:21 PM UTC J. Hong et al. difference was not statistically significant. On the other been good candidates for responsive neurostimulation in hand, complete CC was associated with a greater number lieu of CC.18,32 of postoperative complications (15%) requiring surgical intervention, including subgaleal fluid collection, bone flap osteomyelitis, and epidural abscess. No patients with Postoperative Complications anterior CC suffered complications that required return to Reported types of complications from CC are many, the operating room. Among our patients, the most serious but they are infrequent and usually mild. They may be complication—permanent foot weakness—occurred in the considered as general surgical risks (e.g., wound infection), setting of completing a previous anterior CC. Given this or neurological deficits, particularly those associated with evidence, Kasasbeh et al. recommend that up-front com- hemispheric disconnection. The great majority of neu- plete CC be considered in selected patients, especially pe- rological complications, which include hemiparesis, are diatric patients, who have greater neural plasticity and may transient.3 A recent review of CC discovered 43 instances possibly develop normal interhemispheric communication of acute disconnection syndrome in a cohort of 289 pa- despite surgery.21,23 In retrospect, our pediatric patients tients (14.8%). 28 Disconnection syndromes occur more fre- may have benefited more from initial complete CC rather quently with complete CC rather than anterior two-thirds than a staged procedure. CC. Bower et al. believe that patients with better cognitive Avoidance of increased operative complications is a function and older age are at higher risk for this complica- potential benefit of minimally invasive options for com- tion.6 pletion of CC, although clinical experience to date is too We report a single permanent neurological deficit as a small to understand fully the potential role of these op- result of CC in our series—that of weakness of the left foot tions in this surgery. Current options include stereotactic following completion of CC. The 3 other complications radiofrequency lesioning,30,33 stereotactic radiosurgery seen in our series were minor and did not result in perma- (SRS), 5,7,9,30,33,36,42 or MRI-guided laser interstitial ablation nent injury. One patient had a transient hemiparesis con- (more commonly known as laser interstitial thermal ther- sistent with supplementary motor area syndrome, another apy [LITT]).19 These approaches are newer and have been patient had a superficial wound infection that responded to practiced in small numbers of patients but have promising oral antibiotics, and a third patient developed a pneumonia preliminary outcomes. Stereotactic lesioning by any le- that prolonged his hospital stay but resolved with intrave- sioning technique eliminates the need for a larger craniot- nous antibiotics. In general, with modern image guidance omy, although assurance of the completeness of resection and microsurgical technique, CC is an operation with low may be less. The SRS experience includes a single series rates of serious neurological complications. Any discus- of 8 pediatric patients who experienced decreased rates of sion of the risks of surgery must be balanced by the po- atonic and generalized tonic-clonic (GTC) seizures fol- tential benefit; 9 of the 10 patients in our cohort reported lowing Gamma Knife SRS.11 Two patients suffered tran- worthwhile improvement following CC, including the pa- sient headaches, but no long-term radiation complications tient who had a stroke. This is consistent with previous were observed over a 12-year follow-up period. A LITT- studies assessing quality of life following CC, in which mediated CC has been reported in 4 patients,19,37 with good satisfaction rates in families ranged from 76% to –88%.14,52 outcome reported in 1 patient.19 Overall complication rates for LITT are as high as 22.4%, but this is likely to improve as surgeons become more expert with this technology.37 Limitations of the Study Our experience with these minimally invasive approaches Our study is primarily limited by its retrospective for callosal section at our institution is limited to a single nature, making both reporter bias and selection bias for (non-VNS) radiosurgical case. surgery significant concerns. In addition, we have a small cohort, which makes sampling bias likely. We have un- Evaluation in the Future dertaken a detailed review of a relatively large number of A possible benefit of sequential CC in patients who patients; however, patients who undergo both VNS and have previously received VNS is the ability to use invasive CC are rare. Because seizure frequency was self-reported methods to study patients who may have had inadequate and assessed in a nonblinded manner, there is a possibil- or outdated workups for localization of seizure onset. Be- ity that responses to VNS and CC were over- or underre- cause CC requires a craniotomy, patients who earlier re- ported. In addition, medication trials were permitted over fused intracranial electrode monitoring may be afforded the study period, and several new antiepileptic drugs were an opportunity to have this done with little additional sur- introduced into our patient population. Any response to gical risk, because the patient would only require a single surgical intervention in these patients is confounded by si- craniotomy for both intracranial monitoring and CC. multaneous or subsequent medication adjustments. Intracranial monitoring in anticipation of CC may re- A comparison with a cohort of matched patients who veal some patients to be good candidates for resective did not undergo medication adjustments with the same surgery, and if the seizure onset region is eloquent or oth- operations may help to separate the effects of medica- erwise unresectable, they may be candidates for newer mo- tion from surgery. Likewise, a case-control study with a dalities of therapy such as responsive neurostimulation. In matched cohort of patients who underwent VNS but no our group, one patient had extensive polymicrogyria over CC might be helpful to distinguish CC-related improve- eloquent cortex that could not be safely resected, and an- ment in seizure frequency from non–CC-related changes other had bihippocampal disease; these patients may have such as prolonged VNS usage.

284 J Neurosurg Volume 128 • January 2018

Unauthenticated | Downloaded 09/30/21 06:21 PM UTC Callosotomy after vagal nerve stimulation

Conclusions BC, Barkley GL, et al: Two-year seizure reduction in adults with medically intractable partial onset epilepsy treated with Corpus callosotomy can help reduce seizure frequency responsive neurostimulation: final results of the RNS System in patients who have previously undergone VNS, particu- Pivotal trial. Epilepsia 55:432–441, 2014 larly with respect to drop seizures. Corpus callosotomy is 19. Ho AL, Miller KJ, Cartmell S, Inoyama K, Fisher RS, Halp- safe and well tolerated in these patients. ern CH: Stereotactic laser ablation of the splenium for intrac- table epilepsy. Epilepsy Behav Case Rep 5:23–26, 2016 20. Karceski S: and Lennox-Gastaut References syndrome: a review of the literature and data from the VNS patient registry. CNS Spectr 6:766–770, 2001 1. Arya R, Greiner HM, Horn PS, Turner M, Holland KD, 21. Kasasbeh AS, Smyth MD, Steger-May K, Jalilian L, Bertrand Mangano FT: Corpus callosotomy for childhood-onset drug- M, Limbrick DD: Outcomes after anterior or complete cor- resistant epilepsy unresponsive to vagus nerve stimulation. pus callosotomy in children. Neurosurgery 74:17–28, 2014 Pediatr Neurol 51:800–805, 2014 22. Katagiri M, Iida K, Kagawa K, Hashizume A, Ishikawa N, 2. Asadi-Pooya AA, Sharan A, Nei M, Sperling MR: Corpus Hanaya R, et al: Combined surgical intervention with vagus callosotomy. Epilepsy Behav 13:271–278, 2008 nerve stimulation following corpus callosotomy in patients 3. Behrens E, Schramm J, Zentner J, König R: Surgical and Acta Neurochir (Wien) neurological complications in a series of 708 epilepsy sur- with Lennox-Gastaut syndrome. 158: gery procedures. Neurosurgery 41:1–10, 1997 1005–1012, 2016 4. Ben-Menachem E, Hellström K, Waldton C, Augustinsson 23. Kawai K, Shimizu H, Yagishita A, Maehara T, Tamagawa LE: Evaluation of refractory epilepsy treated with vagus K: Clinical outcomes after corpus callosotomy in patients nerve stimulation for up to 5 years. Neurology 52:1265– with bihemispheric malformations of cortical development. J 1267, 1999 Neurosurg 101 (1 Suppl):7–15, 2004 5. Bodaghabadi M, Bitaraf MA, Aran S, Alikhani M, Ashrafian 24. Kim DS, Yang KH, Kim TG, Chang JH, Chang JW, Choi JU, H, Zahiri A, et al: Corpus callosotomy with Gamma Knife et al: The surgical effect of callosotomy in the treatment of radiosurgery for a case of intractable generalised epilepsy. intractable seizure. Yonsei Med J 45:233–240, 2004 Epileptic Disord 13:202–208, 2011 25. Lancman G, Virk M, Shao H, Mazumdar M, Greenfield JP, 6. Bower RS, Wirrell E, Nwojo M, Wetjen NM, Marsh WR, Weinstein S, et al: Vagus nerve stimulation vs. corpus cal- Meyer FB: Seizure outcomes after corpus callosotomy for losotomy in the treatment of Lennox-Gastaut syndrome: a drop attacks. Neurosurgery 73:993–1000, 2013 meta-analysis. Seizure 22:3–8, 2013 7. Celis MA, Moreno-Jiménez S, Lárraga-Gutiérrez JM, Alon- 26. Lin JS, Lew SM, Marcuccilli CJ, Mueller WM, Matthews so-Vanegas MA, García-Garduño OA, Martínez-Juárez IE, AE, Koop JI, et al: Corpus callosotomy in multistage epilep- et al: Corpus callosotomy using conformal stereotactic radio- sy surgery in the pediatric population. J Neurosurg Pediatr surgery. Childs Nerv Syst 23:917–920, 2007 7:189–200, 2011 8. Cukiert A, Cukiert CM, Burattini JA, Lima AM, Forster 27. Maehara T, Shimizu H: Surgical outcome of corpus calloso- CR, Baise C, et al: Long-term outcome after callosotomy or tomy in patients with drop attacks. Epilepsia 42:67–71, 2001 vagus nerve stimulation in consecutive prospective cohorts of 28. Malmgren K, Rydenhag B, Hallböök T: Reappraisal of cor- children with Lennox-Gastaut or Lennox-like syndrome and pus callosotomy. Curr Opin Neurol 28:175–181, 2015 non-specific MRI findings. Seizure 22:396–400, 2013 29. Malow BA, Edwards J, Marzec M, Sagher O, Ross D, Fromes 9. Eder HG, Feichtinger M, Pieper T, Kurschel S, Schroettner G: Vagus nerve stimulation reduces daytime sleepiness in O: Gamma Knife radiosurgery for callosotomy in children epilepsy patients. Neurology 57:879–884, 2001 with drug-resistant epilepsy. Childs Nerv Syst 22:1012–1017, 30. Marino Júnior R, Gronich G: Corpus callosum stimulation 2006 and stereotactic callosotomy in the management of refractory 10. Ergene E, Behr PK, Shih JJ: Quality-of-life assessment in pa- generalized epilepsy. Preliminary communication. Arq Neu- tients treated with vagus nerve stimulation. Epilepsy Behav ropsiquiatr 47:320–325, 1989 2:284–287, 2001 31. McInerney J, Siegel AM, Nordgren RE, Williamson PD, 11. Feichtinger M, Schröttner O, Eder H, Holthausen H, Pieper Thadani VM, Jobst B, et al: Long-term seizure outcome fol- T, Unger F, et al: Efficacy and safety of radiosurgical calloso- lowing corpus callosotomy in children. Stereotact Funct tomy: a retrospective analysis. Epilepsia 47:1184–1191, 2006 Neurosurg 73:79–83, 1999 12. Fuiks KS, Wyler AR, Hermann BP, Somes G: Seizure out- 32. Morrell MJ: Responsive cortical stimulation for the treat- come from anterior and complete corpus callosotomy. J Neu- ment of medically intractable partial epilepsy. Neurology rosurg 74:573–578, 1991 77:1295–1304, 2011 13. Gates JR, Rosenfeld WE, Maxwell RE, Lyons RE: Response 33. Nádvornik P, Krupa P, Chrastina J, Smrcka V, Novák Z, Zbo- of multiple seizure types to corpus callosum section. Epilep- rilová E: Circular stereotactic callosotomy: a preliminary re- sia 28:28–34, 1987 port. Technical note. Acta Neurochir (Wien) 139:359–360, 14. Gilliam F, Wyllie E, Kotagal P, Geckler C, Rusyniak G: Pa- 1997 rental assessment of functional outcome after corpus calloso- 34. Nordgren RE, Reeves AG, Viguera AC, Roberts DW: Cor- tomy. Epilepsia 37:753–757, 1996 pus callosotomy for intractable seizures in the pediatric age 15. Guillamón E, Miró J, Gutiérrez A, Conde R, Falip M, Jaraba group. Arch Neurol 48:364–372, 1991 S, et al: Combination of corpus callosotomy and vagus nerve 35. Passamonti C, Zamponi N, Foschi N, Trignani R, Luzi M, stimulation in the treatment of refractory epilepsy. Eur Neu- Cesaroni E, et al: Long-term seizure and behavioral out- rol 71:65–74, 2014 comes after corpus callosotomy. Epilepsy Behav 41:23–29, 16. Handforth A, DeGiorgio CM, Schachter SC, Uthman BM, 2014 Naritoku DK, Tecoma ES, et al: Vagus nerve stimulation 36. Pendl G, Eder HG, Schroettner O, Leber KA: Corpus cal- therapy for partial-onset seizures: a randomized active-con- losotomy with radiosurgery. Neurosurgery 45:303–308, trol trial. Neurology 51:48–55, 1998 1999 17. Harden CL, Pulver MC, Ravdin LD, Nikolov B, Halper JP, 37. Pruitt R, Gamble A, Black K, Schulder M, Mehta AD: Com- Labar DR: A pilot study of mood in epilepsy patients treated plication avoidance in laser interstitial thermal therapy: les- with vagus nerve stimulation. Epilepsy Behav 1:93–99, 2000 sons learned. J Neurosurg [epub ahead of print June 3, 2016. 18. Heck CN, King-Stephens D, Massey AD, Nair DR, Jobst DOI: 10.3171/2016.3.JNS152147]

J Neurosurg Volume 128 • January 2018 285

Unauthenticated | Downloaded 09/30/21 06:21 PM UTC J. Hong et al.

38. Reutens DC, Bye AM, Hopkins IJ, Danks A, Somerville E, 49. Vonck K, Van Laere K, Dedeurwaerdere S, Caemaert J, De Walsh J, et al: Corpus callosotomy for intractable epilepsy: Reuck J, Boon P: The mechanism of action of vagus nerve seizure outcome and prognostic factors. Epilepsia 34:904– stimulation for refractory epilepsy: the current status. J Clin 909, 1993 Neurophysiol 18:394–401, 2001 39. Roberts DW, Reeves AG, Norgren RE: The role of posterior 50. Wheeler M, De Herdt V, Vonck K, Gilbert K, Manem S, callosotomy in patients with suboptimal response to anterior Mackenzie T, et al: Efficacy of vagus nerve stimulation for callosotomy, in Reeves AG, Roberts DW (eds): Epilepsy and refractory epilepsy among patient subgroups: a re-analysis the Corpus Callosum 2. New York: Plenum Press, 1995, pp using the Engel classification. Seizure 20:331–335, 2011 183–190 51. Wilson DH, Reeves A, Gazzaniga M: Division of the corpus 40. Rolston JD, Englot DJ, Wang DD, Garcia PA, Chang EF: callosum for uncontrollable epilepsy. Neurology 28:649– Corpus callosotomy versus vagus nerve stimulation for atonic 653, 1978 seizures and drop attacks: a systematic review. Epilepsy Be- 52. Yang TF, Wong TT, Kwan SY, Chang KP, Lee YC, Hsu TC: hav 51:13–17, 2015 Quality of life and life satisfaction in families after a child 41. Rosenfeld WE, Roberts DW: Tonic and atonic seizures: has undergone corpus callosotomy. Epilepsia 37:76–80, 1996 what’s next—VNS or callosotomy? Epilepsia 50 (Suppl 8):25–30, 2009 42. Smyth MD, Klein EE, Dodson WE, Mansur DB: Radiosur- gical posterior corpus callosotomy in a child with Lennox- Gastaut syndrome. Case report. J Neurosurg 106 (4 Sup- Disclosures pl):312–315, 2007 43. Spencer DD, Spencer SS: Corpus callosotomy in the treatment The authors report no conflict of interest concerning the materi- of medically intractable secondarily generalized seizures of als or methods used in this study or the findings specified in this children. Cleve Clin J Med 56 Suppl Pt 1:S69–S83, 1989 paper. 44. Spencer SS, Spencer DD, Sass K, Westerveld M, Katz A, Mattson R: Anterior, total, and two-stage corpus callosum Author Contributions section: differential and incremental seizure responses. Epi- Conception and design: Roberts, Hong, Desai. Acquisition of data: lepsia 34:561–567, 1993 Hong. Analysis and interpretation of data: Hong, Desai. Draft- 45. Stigsdotter-Broman L, Olsson I, Flink R, Rydenhag B, ing the article: Roberts, Hong, Thadani. Critically revising the Malmgren K: Long-term follow-up after callosotomy—a article: all authors. Reviewed submitted version of manuscript: all prospective, population based, observational study. Epilepsia authors. Study supervision: Roberts. 55:316–321, 2014 46. Sunaga S, Shimizu H, Sugano H: Long-term follow-up of Supplemental Information seizure outcomes after corpus callosotomy. Seizure 18:124– Previous Presentations 128, 2009 47. Uthman BM, Reichl AM, Dean JC, Eisenschenk S, Gilmore Portions of this work were presented in poster form at the AANS R, Reid S, et al: Effectiveness of vagus nerve stimulation in Annual Scientific Meeting, which was held in Washington, DC, in epilepsy patients: a 12-year observation. Neurology 63:1124– May 2015. 1126, 2004 48. Vonck K, Thadani V, Gilbert K, Dedeurwaerdere S, De Correspondence Groote L, De Herdt V, et al: Vagus nerve stimulation for David W. Roberts, Section of Neurosurgery, Department of Sur- refractory epilepsy: a transatlantic experience. J Clin Neuro- gery, Dartmouth-Hitchcock Medical Center, One Medical Center physiol 21:283–289, 2004 Dr., Lebanon, NH 03756. email: [email protected].

286 J Neurosurg Volume 128 • January 2018

Unauthenticated | Downloaded 09/30/21 06:21 PM UTC