Fatal Status Epilepticus After Elective Revision Cranioplasty: Case Report and Literature Review

Introduction

Cranioplasty is a reconstructive neurosurgical procedure performed to correct skull defects, most commonly after decompressive craniectomy for traumatic brain injury or malignant ischemic stroke [1]. The benefits of reconstructive cranioplasty include restoration of cranial neuroprotection, cosmetic improvement, and potentially enhanced neurological function and psychological health [2,3]. Given the wide range of indications for decompressive craniectomy, including trauma, stroke, hemorrhage, tumor, and infection, cranioplasty has become a common neurosurgical procedure worldwide [1,4]. Although cranioplasty is frequently performed and is technically straightforward, reported complication rates indicate that it is not a low-risk procedure and may carry a higher risk profile than many other neurosurgical interventions [5]. Cranioplasty-related complication rates of 16.4% to 34% have been reported [2,5–8], compared with rates of 2% to 5% for other elective neurosurgical procedures [6,7,9]. Commonly reported complications include infection, wound dehiscence, bone resorption, intracranial hemorrhage, hydrocephalus, and seizures [1,2,4]. Several studies have identified seizures as the most frequent complication after cranioplasty, with incidences ranging from 8% to 33%; the majority of these events are self-limiting [2,10–12]. Nevertheless, self-limiting postoperative seizures have been associated with an increased risk of worse long-term functional outcomes; therefore, seizure prevention is desirable [2,13]. When seizures persist for more than 5 min, or when at least 2 seizures occur within a 5-min interval, the condition is regarded as status epilepticus (SE) and is associated with substantially worse outcomes [14,15]. The incidence of SE after cranioplasty and other neurosurgical procedures is not well characterized in the literature [6,16]. Reports of SE following neurosurgical procedures often include both suspected and confirmed cases, which may misrepresent the true incidence [17,18]. Additionally, few studies have examined differences in complications and outcomes between primary reconstructive cranioplasty and revision procedures, including differences related to seizure incidence. Here, we describe a confirmed case of SE after elective revision cranioplasty that ultimately resulted in patient death.

Case Report

A fit and healthy 76-year-old man underwent an elective revision cranioplasty to correct symptomatic subsidence of an acrylic plate. His medical history was unremarkable, with no prior seizure disorder. Ten years earlier, he had undergone resection of an intraosseous left temporal epidermoid tumor, accompanied by implantation of a noncustom titanium mesh plate (TiMesh, Medtronic; Figure 1A). Six years after the initial operation (4 years prior to the events in this case), he returned with symptomatic tumor recurrence and underwent an uncomplicated revision procedure involving removal of the titanium mesh, tumor resection, and repair of the skull defect with a custom 90-mm–diameter acrylic plate and a preformed craniotomy guide template (AnatomicsAcrylic, Anatomics; Figure 1B, 1C). Seizures were not associated with either procedure. Subsequent surveillance imaging demonstrated no evidence of tumor recurrence; however, progressive subsidence of the acrylic plate developed, clinically characterized by pain and deformity (Figure 1D), for which corrective surgery was offered.

Under general anesthesia, the acrylic cranioplasty was removed and replaced with a new custom titanium mesh cranioplasty (AnatomicsTitanium, Anatomics). Prophylactic antibiotics for surgical site infection were administered at induction, but antiepileptic drugs (AEDs) were not given. Upon flap elevation, the acrylic plate demonstrated mobility due to multiple fractured locking fixation plates. The implant was carefully released from adherent underlying soft tissue neodura that had grown through perforations manufactured in the plate. Complete release was achieved by extradural dissection of all soft tissue beneath and around the implant. Small dural defects were identified; the largest defect was repaired with a pericranial graft, and all remaining defects were primarily closed using 4-0 Nurolon suture (Ethicon). An overlapping custom titanium cranioplasty plate (AnatomicsTitanium, Anatomics) was positioned satisfactorily and secured with multiple 4-mm screws (TiMesh, Medtronic). A subgaleal wound drain was inserted, hemostasis was confirmed, and the wound was closed in layers before application of a wound dressing and crepe bandage. The subgaleal drain was placed on negative pressure at the conclusion of the procedure.

The patient was transferred to the recovery room and routinely extubated. He was initially neurologically intact; however, within 30 min he experienced 2 right-sided focal seizures with secondary generalization. He was administered a loading dose of 1 g intravenous levetiracetam and 2 mg intravenous midazolam, with seizure resolution, but he remained in a postictal state. Within 1 h, a third and more prolonged seizure episode occurred; he was reintubated for seizure control. Computed tomography (CT) of the brain demonstrated 2 small foci of intraparenchymal hemorrhage within the left posterior parietal lobe, a small volume of intraventricular blood layering within both occipital horns, and generalized fullness of the left cerebral hemisphere secondary to cerebral edema (Figure 2A). Scheduled levetiracetam and lacosamide were commenced; the patient remained intubated for seizure control. An electroencephalogram (EEG) obtained 2 days after surgery demonstrated nonconvulsive SE, and phenytoin was added as a third agent. A subsequent EEG demonstrated cessation of focal seizure activity, after which the patient was desedated and successfully extubated.

Four days after extubation, the patient remained agitated and was anesthetized to facilitate medical imaging. He showed no evidence of clinical seizure activity at that time. Magnetic resonance imaging (MRI) of the brain further characterized the previously identified intracranial hemorrhages (Figure 2B) but demonstrated no evidence of ischemia or other surgical complications; the cerebral edema had resolved. Multiple subsequent attempts at extubation after the MRI were unsuccessful. Repeat EEG demonstrated nonconvulsive SE despite maximal antiepileptic therapy with levetiracetam, lacosamide, phenytoin, and additional topiramate. Burst suppression was ultimately achieved only with a barbiturate-induced coma. The focal area of cerebral hemorrhage had resolved, according to CT images obtained 2 weeks after surgery (Figure 2C). Over the following 2 weeks, repeat EEGs consistently demonstrated substantially abnormal activity with ongoing nonconvulsive SE, and a final MRI performed on postoperative day 23 showed only expected involutional changes at the operative site (Figure 2D). Multidrug antiepileptic therapy was maintained throughout this period. Given the inability to control seizures without prolonged general anesthesia, the goals of care were shifted to comfort measures. The patient was extubated 28 days after cranioplasty and died shortly thereafter.

Discussion

Although seizures are a recognized and common complication after cranioplasty, SE remains a relatively rare event [6]. Multiple studies have investigated risk factors and predictors associated with postoperative seizures in patients undergoing cranioplasty. Frequently reported factors include male sex, age greater than 60 years, trauma-related indications for decompressive craniectomy, and the timing of cranioplasty after decompressive craniectomy, for which optimal timing remains controversial [2,4,7,11,13,19–21]. An interval exceeding 6 months between craniectomy and cranioplasty has been associated with a substantial increase in early post-cranioplasty seizures, defined as seizures occurring within 7 days of the procedure [4,19]. These factors have been widely reported for primary cranioplasty; however, limited data are available concerning revision procedures. Nevertheless, a greater number of prior neurosurgical interventions has been associated with increased risks of major and minor perioperative complications, including return to the operating room [22]. Accordingly, risks associated with revision cranioplasty should be considered higher than those associated with an index procedure (Table 1) [6,23–27].

Postoperative seizures may be categorized as early or late; late seizures are defined as those occurring more than 7 days after surgery [2]. Post-cranioplasty seizures have been associated with worse outcomes, and available evidence suggests that postoperative seizures can serve as a predictor of mortality [2,4,13], as demonstrated in our case. Early seizures have been suggested to arise from diffuse cerebral edema and/or ischemia, with subsequent release of excitatory and toxic cellular components [2]. In the present case, postoperative imaging demonstrated cerebral edema after seizure onset; however, this finding had resolved by the time of SE onset. Despite resolution of cerebral edema and hemorrhagic burden, the patient progressed to SE. This condition is associated with high rates of morbidity and mortality [28,29] and often requires induction of a therapeutic coma when refractory to first- and second-line anticonvulsant therapy [15,30], as occurred in the present case. Several mechanistic hypotheses have been proposed regarding SE pathophysiology, including impairment of seizure-terminating mechanisms and activation of processes that sustain prolonged seizure activity [28]. In refractory SE, underlying pathological drivers of seizure activity and structural brain changes, such as neuronal loss, inflammation, or disruption of the blood–brain barrier, have been proposed as contributors to pharmacoresistance [28,31].

The causes of SE in our patient, and the preceding sequelae, may be related to perioperative changes and initial insults occurring during the early postoperative period. The intracranial hemorrhage and cerebral edema observed after surgery may have developed as expansion injuries of the brain [32]. The cranioplasty defect measured 93×82 mm, and the size of the defect may have increased susceptibility to expansion injury. Manipulation of the neodura to release it from acrylic perforations may have further increased this risk. Expansion injuries have classically been described in the context of primary cranioplasty, rather than in patients undergoing revision procedures. In primary cranioplasty, the brain is often exposed to atmospheric pressure for a prolonged period, such that cranioplasty results in pressure changes that promote brain expansion to fill the defect. This mechanism is less likely in patients with a preexisting cranioplasty, in whom no large skull defect is present. In the present case, the cranioplasty plate was depressed, which may have reflected underlying intracranial pressure alterations and could have predisposed the patient to postoperative seizures in a manner similar to previously reported cases. Additionally, the use of a standard negative-pressure subgaleal drain may have contributed to the risk of expansion injury. Other postulated causes of post-cranioplasty seizures (eg, cerebral edema resulting from impaired autoregulation and reperfusion injury [33]) appear less relevant in our patient, given that the initial pathology was related to tumor resection rather than stroke or trauma, and no large preexisting skull defect was present before cranioplasty. Another notable feature in this case was subsequent resolution of the initial radiological findings. CT scans performed on the day after revision cranioplasty demonstrated interval resolution of cerebral edema; further imaging showed improvement in the hemorrhagic burden, whereas prior reports indicated that cerebral edema was often malignant and progressive rather than resolving [34]. Cerebral edema may have contributed to the initial postoperative seizures but had radiologically resolved by the time of SE onset, suggesting that it was less likely to be the primary driver of the refractory seizures. An additional atypical feature was successful extubation 3 days after the procedure, followed by re-anesthetization on postoperative day 7 to facilitate a semi-elective MRI study. One possibility is that exposure to general anesthetic agents exacerbated seizure activity [22] and, in the context of an already seizure-prone state, contributed to refractoriness. Alternatively, recurrent and refractory seizures may have developed irrespective of re-anesthetization for imaging.

Cranioplasty is an extradural operation; therefore, seizure activity would be expected to occur less frequently than after intradural, and particularly intra-axial, neurosurgical procedures. In the present case, the neodura was very thin posteriorly and the brain was exposed, suggesting that seizure risk approached the level associated with intradural surgery. Additionally, the development of postoperative intracerebral hemorrhages may have increased seizure risk to a level comparable with that observed after intra-axial procedures. Given that this was a revision cranioplasty, the risk of postoperative seizures may have been further increased. Prophylactic AED use has been shown to substantially reduce the risk of post-cranioplasty seizures [35–37]; thus, the role of perioperative seizure prophylaxis for all cranioplasty procedures warrants strong consideration, particularly in higher-risk settings, as illustrated by the present case. Although not predictable, administration of seizure prophylaxis at induction for the revision procedure may have prevented the eventual development of fatal SE.

When SE develops after cranioplasty, standard management includes stepwise escalation of AED therapy until seizure control is achieved [4]. Sedation, commonly with propofol and midazolam, is used for airway protection and treatment of seizures refractory to AED therapy. Additional reported approaches for post-cranioplasty SE include infusion of dexmedetomidine [27], removal of the cranioplasty [25], and insertion of a vagal nerve stimulator [26] (Table 1). All reported patients who underwent these interventions survived; therefore, these options should be considered when conventional treatment strategies have been exhausted.

Conclusions

Post-cranioplasty seizures have a relatively high incidence, remain incompletely understood, may be refractory despite escalation of multiple AEDs, and can be associated with substantial morbidity and mortality. Our report highlights features associated with the development of nonconvulsive SE, proposes mechanisms that may predispose to seizure development, and offers strategies to reduce seizure incidence. We hypothesize that a combination of revision cranioplasty, intracranial pressure alterations, exposure of the cerebral cortex due to dural disruption, negative-pressure subgaleal drainage, and postoperative intracerebral hemorrhage contributed to the development of refractory SE. Prophylactic perioperative AED use is recommended for cranioplasty procedures considered to carry a high seizure risk and for all revision cranioplasty operations. Once SE has developed, consideration should be given to infusion of dexmedetomidine, removal of the cranioplasty, and insertion of a vagal nerve stimulator, as reported in prior literature.