Decompressive Craniectomy for Traumatic Brain Injury

Introduction

Traumatic brain injury (TBI) is a significant cause of death and disability, accounting for an estimated 235,000 hospitalizations and 50,000 deaths annually in the United States alone. While neurosurgeons play an integral role in the management of TBI, much controversy exists regarding the use of surgical therapies, especially decompressive craniectomy. There have been few randomized controlled trials assessing the efficacy of decompressive craniectomy to non-surgical management for the treatment of TBI in adults ^{, ,} and its utility remains controversial. The goal of this chapter to is to summarize the data pertaining to the use decompressive craniectomy for TBI and to discuss various aspects of this procedure, including outcomes, techniques, and complications. The goal of this chapter is to summarize the data pertaining to the use decompressive craniectomy for TBI and to discuss the various issues associated with this procedure, including outcomes, techniques, and complications.

Historical Perspective

Evidence of surgical decompression performed to treat TBI goes as far back as Ancient Egypt and Ancient Greece, and the indications for its use included TBI, epilepsy, headache, and mental illness. The concept of using surgical decompression to treat elevated intracranial pressure (ICP) was introduced to modern neurosurgery by Kocher and Cushing ^, at the beginning of the 20th century. Discussion of surgical decompression for TBI in the literature did not become widespread, however, until the late 1960s and 1970s after a series of studies examining surgical decompression for various TBI-related entities. At that time, the benefit of surgical decompression was not clear. Jamieson and Yelland reported encouraging results for decompression of traumatic epidural hematomas, with a mortality rate of just 16% in a series of 167 patients. Surgical techniques used in their series ranged from subtemporal and suboccipital decompressions to “local” craniectomies, trephines, and osteoplastic craniotomies. Subsequent review of surgical management of traumatic subdural hematomas by the same authors, however, showed worse outcomes. In a series of 555 surgically managed subdural hematomas, 317 of the patients underwent a decompressive craniectomy, with an associated 43% mortality rate. Similarly, surgically managed intraparenchymal hematomas were associated with a 24% mortality rate, although the outcomes associated with various surgical techniques were not clearly described.

In the late 1960s and early 1970s, several groups began to study specific decompressive techniques for the treatment of severe TBI. Kjellberg et al. reported using a bifrontal decompressive craniectomy with duraplasty for intractable cerebral edema in 73 patients, 50 of whom had sustained TBI. They reported an 18% survival rate (22% among TBI patients), as well as an additional 16 patients who showed neurological improvement postoperatively but died as a result of other medical complications. Similarly, Venes et al. reported 13 TBI patients with intractable cerebral edema treated with wide decompressive bifrontal craniectomies. Although the mortality rate was only 31%, only 1 patient (a 2½-year-old child with moderately severe TBI) returned to normal neurological function postoperatively with the exception of a seizure disorder. A wide decompressive hemicraniectomy (DHC) with durotomy was described by Ransohoff et al. for the treatment of subdural hematoma. They reported a 40% survival rate, and 28% of patients returning to normal activities. These outcomes were much improved over an 85% mortality rate among TBI patients treated with burr holes or routine craniectomies. A follow-up study by the same group, however, showed much worse outcomes for traumatic subdural hematomas treated with decompressive craniectomy, with only a 10% survival rate. The authors attributed this difference to the presence of concomitant brain stem or subcortical injury. Britt et al. reported a cohort of 42 patients who underwent decompression of an acute traumatic subdural hematoma, 34 of whom had their bone flaps left off because of cerebral edema. Although Britt et al. did not separate outcomes based on craniectomy or craniotomy, they did report an overall 55% mortality rate (only 36% mortality within 30 days of decompression) and stated that their results led them to standardize the use of a large DHC for the treatment of traumatic subdural hematomas. Because poor outcomes from subdural and epidural hematomas were partially attributed to underlying contusions, Yamaura et al. studied the use of DHC for severe traumatic contusions. Twenty-four percent of their patients were converted to bilateral hemicraniectomies if the “unilateral opening did not seem sufficient in its decompressive effect.” They reported an overall 36% mortality rate and advocated for the use of bilateral hemicraniectomies. In radiographic follow-up of patients who had undergone hemicraniectomy, Morantz et al. confirmed the benefits of DHC on the resolution of midline shift and removal of hematomas. In addition to functional outcomes, Hase et al. examined the effect of decompressive surgery on the elasticity (change in pressure divided by change in volume) of the brain and found that decompression decreases elasticity, thus reducing pressure variations with changes in intracerebral volume.

In an attempt to further understand the pathophysiology underlying posttraumatic cerebral edema and the role of surgical decompression, several groups have developed experimental models. Moody et al. studied a model of TBI using epidural balloon compression in dogs. Ten dogs received no surgical decompression, and all died within 12 hours of injury. Autopsies revealed pontine hemorrhages. A second group of dogs underwent a DHC. Although none of the surgically treated dogs regained consciousness, decompression resulted in less pontine injury. However, the hemicraniectomy also resulted in hemorrhagic infarction, necrosis, and edema at the site of the craniectomy. Cooper et al. studied surgical decompression in a dog model of cold-injury-induced cerebral edema. Hemicraniectomy lowered ICP but resulted in significantly greater cerebral edema. This effect was attributed to possible reduction in the interstitial pressure within the brain after decompression, resulting in a greater hydrostatic pressure gradient between the intravascular and interstitial spaces.

Overall, these studies demonstrate an increasing interest in the utility of decompressive craniectomy for TBI and provide evidence for a possible benefit regarding mortality rate after surgical decompression in severe TBI but questioned whether the morbidity and quality of life attained are justifiable. Importantly, the complex nature of TBI was recognized, and many authors acknowledged the need for a systems-based approach to the management of TBI that encompasses more than just the surgical aspects of management. In addition, the Glasgow Coma Scale (GCS) and the Glasgow Outcome Scale (GOS) were developed, allowing for more uniform characterization and generalization of TBI patients across centers. For the next several decades, however, few studies were published regarding the use of decompressive craniectomy in TBI.

Current Evidence

More recently, there has been renewed interested in defining the use and benefits of decompressive craniectomy, with an ever-increasing number of publications on the topic over the past 20 years. Gower et al. began to reexplore decompressive craniectomy by reporting 10 patients treated with salvage subtemporal decompression among a series of 115 severe TBI patients. They demonstrated a mortality rate of 40% among decompressive craniectomy patients compared to 82% mortality among patients treated medically (with pentobarbital-induced coma). They also showed a 34% reduction in ICP after decompressive craniectomy. Gaab et al. applied both unilateral and bilateral frontoparietal temporal craniectomies and dural enlargement in patients with medically refractory cerebral hypertension and showed only a 13.5% mortality rate, with 78% achieving a GOS score of 4 or 5.

Polin et al. compared 35 patients with malignant posttraumatic cerebral edema treated with bifrontal decompressive craniectomy to matched controls from the Traumatic Coma Data Bank. They reported 23% mortality and 37% good outcomes (GOS score 4 or 5), a significant improvement compared to matched controls. They also demonstrated significantly lower postoperative ICP compared to controls who did not have surgery. In addition, Polin et al. identified young age and early timing of decompressive craniectomy (within 48 hour of injury) as possible favorable prognostic factors. Kleist-Welch Guerra et al. likewise demonstrated promising results in a prospective study of 57 patients with severe TBI and medically refractory cerebral edema, 31 of whom were treated with a unilateral craniectomy and the remainder with bilateral craniectomies. Kleist-Welch Guerra et al. reported 19% mortality and 58% favorable outcomes and advocated for the use of decompressive craniectomy before barbiturate-induced coma. In a retrospective review of DHC for severe TBI, Münch et al. reported higher mortality (52%) but similar favorable outcomes (41%). Similarly to Polin et al., they identified young age and early DHC as favorable prognostic factors. They also demonstrated various improvements in computed tomography (CT) characteristics after DHC, including decreased midline shift and increased visibility of the mesencephalic cisterns. Taylor et al. conducted the first randomized controlled trial of decompressive craniectomy, comparing early bilateral temporal craniectomy without durotomy to medical management for the treatment of elevated ICP in head-injured children. That study found that craniectomy led to greater reductions in ICP and noted a trend toward improved functional outcomes. This study was followed by a series of investigations into the role of decompressive craniectomy in the treatment of TBI, including indications, techniques, and prognostic factors, all of which have culminated in two large randomized-controlled trials of surgical decompression for TBI. ^,

The Decompressive Craniectomy in Diffuse Traumatic Brain Injury (DECRA) trial, published in 2011, was a multicenter randomized trial that compared bifrontal craniectomy to standard medical management for traumatic cerebral edema. Medically refractory cerebral edema was defined as ICP >20 mmHg for >15 minutes in a given hour despite standard medical interventions and external ventricular drainage. Overall, bifrontal craniectomy reduced ICP, days spent on mechanical ventilatory support, and the days spent in an intensive care unit. However, patients treated with bifrontal craniectomy actually had worse functional outcomes based on the extended GOS. Of note, the craniectomy group had higher numbers of patients with non-reactive pupils, and these differences in functional outcome were negated when the outcomes were adjusted for pupillary function. Several reasons for this lack of efficacy have been postulated, including the bifrontal technique itself and improper indication for surgery.

More recently, the multicenter randomized RESCUEicp trial also compared decompressive craniectomy to medical management, but had several important differences compared to the DECRA trial. Surgical decompression could be with either a bifrontal craniectomy or a decompressive hemicraniectomy. In addition, this trial had a higher threshold for randomization – ICP had to be >25 mmHg for >1 hour despite maximal medical management and external ventricular drainage. This study found a significant reduction in mortality at 6 months (27% versus 49%) as well as an increase in patients in a vegetative state or with severe disability. Although the primary outcome was the GOS-extended (GOSE) at 6 months, patients were followed and assessed at 12 months as a secondary outcome. By 12 months, favorable outcomes occurred in 45% of the surgical group as compared with 32% in the medical group, suggesting that longer follow-up is needed in future studies to determine the full benefit of decompressive craniectomy.

Indications and Timing

The most common indication for decompressive craniectomy in the setting of TBI has been salvage therapy for medically refractory cerebral hypeertension. ^{, , , , , , ,} In this setting, ICP monitoring is utilized to guide medical management. Typical medical therapy protocols include a combination of head-of-bed elevation, cerebrospinal fluid (CSF) drainage, sedation, hyperventilation, paralysis, hyperosmolar therapy, and barbiturate-induced coma. Several groups, however, have also employed decompressive craniectomy at the time of initial hematoma evacuation based on the intraoperative finding of cerebral swelling. ^{, , , , , ,} Both indications have been supported by TBI management guidelines, ^, but one of the chief controversies in the use of decompressive craniectomy for TBI is the most appropriate timing of decompression after injury. Several studies have compared the use of early (typically within 24 hours of injury and in conjunction with hematoma evacuation) to late (more than 24 to 48 hours after injury, typically to treat medically refractory cerebral hypertension). Patients within the early and late groups, of course, are distinct populations and cannot be generalized to one another. As such, comparisons of decompressive craniectomy for each group have had mixed results, with some studies reporting superior outcomes after early decompressive craniectomy, ^{, ,} some reporting worse outcomes, and others reporting no difference. ^, Coplin et al. sought to evaluate the benefit of early decompressive craniectomy at the time of initial hematoma evacuation by comparing TBI patients with traumatic mass lesions who underwent either craniotomy or craniectomy. They found no significant differences between the two groups, despite worse injuries in the craniectomy group. Similarly, Woertgen et al. compared craniotomy to craniectomy for treatment of acute traumatic subdural hematoma and also found no significant difference in outcomes. As Coplin et al. also reported, the craniectomy patients in the Woertgen et al. study were found to have worse injuries intraoperatively, prompting surgeons to leave off the bone flap. The fact that both studies found similar outcomes despite worse injuries in the craniectomy groups suggests that early decompressive craniectomy, or possibly even prophylactic decompressive craniectomy, at the time of initial hematoma evacuation may in fact provide benefit.

Prognostic Factors

Several studies have examined various potential prognostic factors in an attempt to better define the patient population that will benefit most from decompressive craniectomy. The two most common, and most easily obtained, factors studied have been age and preoperative GCS score. Early studies excluded older patients, with age cut-offs as low as 30 years, but several studies ultimately evaluated age as a prognostic factor. Several studies have reported a correlation between age and outcomes, ^{, , , , , , , , ,} which is not surprising, given that age has been shown to correlate with outcome after TBI in general. Other decompressive craniectomy studies, however, have found no correlation. ^, De Bonis et al. recently summarized the decompressive craniectomy literature with regard to age as a prognostic factor, citing studies that showed a correlation with age as well as studies that did not, and concluded that there are no strong data to support an effect of age on outcome and that the age cut-offs reported in many studies are arbitrary. At our institution, we typically do not include age in making the decision whether to proceed with decompressive craniectomy. However, on the basis of our own data suggesting that old age does correlate with poor outcomes after decompressive craniectomy for TBI, we do temper our expectations of outcomes in older patients.

GCS score is, of course, another commonly used prognostic factor in TBI, ^, and several decompressive craniectomy studies have shown a positive correlation between preoperative GCS score and GOS score. ^{, , , , ,} The correlation of GCS score with outcome is likely complicated, with evidence that the motor score alone may be more prognostic than the total GCS score. In addition, confounding factors such as the use of alcohol and other drugs as well as the timing of any acute changes in GCS score must be considered to accurately interpret a preoperative GCS score. Response of ICP to decompression has also been shown to be a possible postoperative prognostic factor, with patients who continue to have high ICP after decompressive craniectomy being more likely to have poor functional outcomes. ^,