Malignant Cerebral Edema After Cranioplasty: A Case Report and Literature Insights

Introduction

Decompressive craniectomy (DC) is standard of care in patients with severe traumatic brain injury (TBI). Specifically, DC is indicated, although not always clinically appropriate, in the setting of acute subdural hematoma, parenchymal hemorrhage with mass effect, elevated intracranial pressure (ICP) refractory to medical management, and, less frequently, closed TBI with diffuse brain swelling [1]. After a period of decompression, patients undergo cranioplasty to restore the bony defect. Without replacement of the bone flap, there is risk of injury to underlying brain tissue and development of sinking skin flap syndrome (SSFS), also known as syndrome of the trephined or paradoxical herniation. This condition is characterized by sunken skin above the bone defect, with concomitant neurologic decline or plateau. Radiographically, SSFS typically presents with paradoxical midline shift or herniation [2]. Cranioplasty is indicated as the treatment for this condition, and restoration of the bony defect also restores the brain’s ability to regulate intracranial pressure [3].

When compared to a decompressive craniectomy, a procedure often performed in a high-acuity setting as a life-saving intervention, cranioplasty is often thought of as a benign operation with minimal risk to the patient. Complications such as local infection, epidural hematoma, and flap resorption are the most frequently documented [4]. Patient status also contributes to possible complications. The presence of SSFS or a VP shunt are both thought to increase the morbidity of the operation by increasing the incidence of intracranial fluid collections and infections [5,6]. Malignant cerebral edema, also termed pseudohypoxic brain swelling, is a devastating and often-fatal consequence of an otherwise uneventful cranioplasty. Cases of malignant cerebral edema are increasingly reported in the neurosurgical literature.

Case Report

We present the case of a 41-year-old man who suffered a gunshot wound (GSW) to the right frontal lobe. Initial evaluation in the ED showed hemorrhagic brain matter present at the superior frontal bone and an exit wound at the parietal bone, with a GCS of 11 (E3, V2, M6), and left-sided hemiparesis. He intermittently followed commands with his right upper extremity (RUE). Vital signs showed hypertension with bradycardia, consistent with elevated ICP. Following rapid-sequence intubation, a CT head scan demonstrated right-sided mixed intraparenchymal and subarachnoid hemorrhage with a leftward midline shift of 5 mm and pneumocephalus along the path of the penetrating GSW with retained ballistic fragments (Figure 1).

Given the severity of his injuries and concern for impending herniation, he was taken to the OR, where he underwent emergency right-sided frontoparietal decompressive craniectomy (DC), subdural hematoma evacuation, removal of necrotic and herniating brain matter, and external ventricular drain (EVD) placement (Figure 2). The dura was repaired with Integra DuraGen Dural Graft Matrix. Post-operatively, his hospital course was complicated by post-traumatic hydrocephalus 2 weeks after EVD removal (Figure 3A), which necessitated placement of a programmable Medtronic Strata ventriculoperitoneal (VP) shunt set to 1.5. Shunt placement was delayed 9 days after the initial development of hydrocephalus due to concern for meningitis in the setting of repeated fevers of unknown origin. After CSF cultures returned negative and the patient had completed a 2-week course of cefepime and vancomycin with resolution of fevers, the shunt was placed. A week after placement, the VP shunt was adjusted from 1.5 to 0.5 given persistent expanded ventricles on a CT head scan (Figure 3B). After 1 month, the shunt was programmed to 1.5 due to concern for possible over-shunting (Figure 3C). The patient remained hospitalized for the 3 months following the initial injury, where he gradually regained cognitive and motor abilities. By postoperative day (POD) 67, he had persistent LUE paralysis, LLE weakness, and required enteral nutrition due to severe oropharyngeal dysphagia. He followed commands and was able to converse spontaneously with the use of a tracheostomy due to vocal cord paralysis.

On POD 83, the patient had a GCS of 15 and returned to the OR for elective PEEK (polyetheretherketone) cranioplasty. Cranioplasty had been scheduled for POD 73 but was postponed due to a PEEK implant shipping delay. A pre-operative CT head scan showed radiographic evidence of sinking skin flap syndrome, with an 8-mm leftward midline shift (Figure 4). Continuous monitoring throughout the operation displayed a transient drop in blood pressure to 73/54 mmHg at the time of induction. In the middle of the case, there were brief episodes of bradycardia to the 40s. Otherwise, the operation was without incident, with no noted periods of hypercapnia or neurologic changes. The skin flap separated off the DuraGen with minimal manipulation and no adhesions. A subgaleal drain was placed prior to wound closure. At the time of extubation, the patient failed to regain consciousness. Neurologic examination revealed absent brainstem reflexes. An immediate postoperative CT scan demonstrated global diffuse cerebral edema and hypoperfusion (Figure 5). Three days later, he was declared brain dead. An autopsy performed ruled out other causes of death, including any cardiovascular events. On autopsy, the brain exhibited evidence of prolonged ventilatory support and hypoxic-ischemic changes.

Discussion

A careful review of the literature demonstrated many cases of MCE following an otherwise uneventful cranioplasty. MCE, also termed pseudohypoxic brain swelling, was first described in 2003 [7]. The condition is defined by early post-operative neurologic deterioration with concomitant bilateral imaging changes suggestive of edema, such as hypoattenuation. The authors who first reported MCE determined that anoxemic and ischemic hypoxia were unlikely to be the underlying causes [7].

This is consistent with symptom evolution despite adequate intra-operative mean arterial pressure, such as in the case presented here. This report and several others suggest that negative-pressure suction drainage plays an important role in the pathophysiology of MCE [7–15], likely due to interference with ICP auto-regulation. A study of pressure regulation in patients in the 12 hours before and 24 hours after cranioplasty showed significantly increased ICP once the calvarium was reconstructed [3]. This suggests that there is a period of acute pressure elevation following cranioplasty. The same study found significant differences in ICP between supine and seated positions in the week following cranioplasty [3], with participants experiencing an average pressure increase of 7.4±3.6 mmHg while supine. No significant change in pressure regulation was observed in the seated position. To date, there are no reports of cranioplasty performed in the seated position. This may be an important consideration given the increasing evidence that cranioplasty leads to transient ICP dysregulation, which is likely worsened by use of suction drainage. Additionally, it is thought that the presence of a VP shunt at the time of cranioplasty increases the risk of complications; specifically, the development of epidural and subdural fluid collections [5]. In the same study, ligation of the shunt prior to cranioplasty eliminated the risk of fluid collection. However, no data exist to suggest that the presence of a VP shunt at the time of cranioplasty increases the risk MCE. This gap in the literature is likely due to a limited number of cases where MCE develops and even fewer that occur with a VP shunt in place. It is reasonable to suggest that VP shunting increases the risk of MCE, especially given how many of the documented cases developed in the setting of negative-pressure suction drainage.

Negative-pressure suction drainage is standard of care following cranioplasty. If suction drainage and positioning alone caused MCE, it would be reasonable to expect a much higher incidence of the condition. Prior case reports have suggested that the presence of sunken skin flap syndrome elevates the risk of MCE [8,10,11,13,15–18]. The presence of a VP shunt is a known risk factor for development of SSFS [5]. As mentioned previously, SSFS increases the morbidity of cranioplasties, in part due to the intra-operative manipulation necessary to remove the skin flap from the underlying dura, resulting in trauma to the tissue [5]. The removal of the sunken skin flap and replacement of the bone fragment leave a larger potential space for post-operative fluid collections than would otherwise exist in a patient without SSFS. However, cranioplasty is often cited as the definitive management of SSFS [2,19]. Of the 22 cases we reviewed, 18 had SSFS prior to undergoing cranioplasty (Table 1). Some authors have postulated that delay of cranioplasty beyond 3 months after a decompressive craniectomy can contribute to evolution of SSFS and increase the risk of MCE [23,24]. While several reported cases of MCE delayed cranioplasty beyond this recommendation, 8 of 22 reviewed cases, including the case reported here, fell within the optimal timing interval. The combination of SSFS, a patent VP shunt, and the placement of a post-operative negative-pressure suction drain likely contributed to the intra-operative death of our patient.

Conclusions

Continued research is needed to better understand the pathophysiology of MCE so that future occurrences may be prevented. Based on the case reported in this paper and a review of the literature, a combination of negative-pressure suction drainage, an existing VP shunt, sunken skin flap syndrome, and delayed time to cranioplasty likely play a significant role in the evolution of MCE. Additionally, operative positioning may contribute to development of this condition. However, more investigation is needed to elucidate the exact effect of positioning on intracranial pressure in the OR. We urge neurosurgeons to consider the likelihood of MCE in their cranioplasty cases and to adapt surgical planning accordingly.