There is insufficient evidence to support a strong, moderate strength or weak recommendation for urgent surgical decompression/craniectomy vs a period of observation prior to decompression.

In severely injured penetrating traumatic brain injury (pTBI) patients, we recommend performing a craniectomy or craniotomy early when indicated to reduce the likelihood of death or poor outcome.

In the absence of direct scientific evidence, EXPERT CONSENSUS concluded that:

• Patients with pTBI who are medically salvageable with a surgical indication should undergo urgent surgery to reduce morbidity and mortality. Surgical planning should be individualized and the availability of resources needed for safe surgery must be ensured. (100% Consensus)

No evidence or expert opinion supported distinct recommendations based on patient gender, age, wounding mechanism, or military vs. civilian context.

Nearly all pTBI patients have an open wound and possibly high ICP and require surgery as soon as feasible. It is generally accepted that this surgery should occur emergently or at least urgently in most circumstances. For instance, patients with craniocerebral TBI due to GSW or blast fragment injuries generally require debridement and dural repair. Patients impaled by sharp foreign bodies usually require emergent surgery to remove the foreign body, and to debride and close the wound. (Rosenfeld et al., 2015) Many factors can delay surgery, however. The surgery for pTBI patients may require mobilization of personnel and resources not readily available. These resources may not be present on site or after hours. Importantly, delays to reach a neurosurgeon or an advanced (Level I) trauma center are common for casualties in war and disaster zones and in remote or resource-poor environments. This therefore raises the question of whether surgical exploration, debridement, decompression and repair of dura can be delayed if the injury is not immediately life threatening. Would such a delay be inconsequential or could it lead to harms related to CSF fistulas, infection rate and ultimately quality of life? One must also consider that timing of surgery also depends on the general status of the patient, hemodynamic stability, other injuries, as well as physician judgment. Severe life-threatening polytrauma is more likely in the context of multiple gunshot wounds or blast injuries. Blast trauma includes injury from the blast energy, penetrating injury from projectiles, blunt injuries from secondary impacts and burns. These complex and challenging injuries affecting multiple body regions demand multidisciplinary care led by a general trauma surgeon. In pTBI, delays in surgery resulting from initial judgements of a poor prognosis may also induce unnecessary harm if the initial determination of poor prognosis is later, reversed. Mass casualty events in military or civilian hospitals can also limit the availability of operating rooms and staff. The trauma surgeon triages the urgency of the patients for the OR. It is often advantageous for the patient to have damage control laparotomy or orthopedic surgery for limb trauma and craniotomy/craniectomy performed simultaneously by two teams, thus conforming to the principles of damage control neurosurgery.(Moore et al., 2016) With all of this in mind, a literature review was conducted to compare outcomes of early versus delayed therapy and six relevant studies were identified as providing relevant evidence.

The prior penetrating head injury guidelines addressed "Surgical Management of Penetrating Brain Injury."("Surgical management of penetrating brain injury," 2001b) Timing of surgery was addressed by the evaluation of four civilian studies.(Clark et al., 1986; Hubschmann et al., 1979; Nagib et al., 1986; Raimondi & Samuelson, 1970). None of these four studies was carried forward into these updated guidelines. There was a paucity of evidence related to timing of surgical intervention in military environments.

It was recommended as an option:

The first edition noted that it was difficult to differentiate the effects on outcome of the timing of surgery from those resulting from rapid field triage and resuscitation. It was also noted that although authors often state that early surgery is desirable, no solid evidence supports this practice. In the first edition, identified patients undergoing late surgery were markedly different from those treated early preventing solid conclusions about the actual timing of surgery.

Six studies were determined to meet quality and other criteria for inclusion in the evidentiary base.

Wei et al provided a non-randomized, retrospective cohort study, which identified in civilian/military patients in China with cranial gunshot wounds (GSW) and blast injuries.(Wei et al., 2013) The authors identified 56 patients from 1966 until May 2009 and performed a retrospective analysis of outcomes. Surgical intervention in this cohort included debridement, hematoma evacuation, and decompressive craniectomy and the authors noted this was performed as early as possible. Surgical intervention was performed within 8 hours of injury in 39 patients (69.6%) and beyond 8 hours of injury in 15 (8.9%). Two of the 56 patients died (one extensive brain injury, 1 multiple organ failure). The authors report that patients who underwent surgical intervention within 8 hours of injury were more likely to recover (undefined) than those who received debridement after 8 hours (82.1% vs. 26.7%, p<0.001).

Helling et al provided a second non-randomized retrospective study performed in the United States (2 urban trauma centers in Kansas City, Missouri) examined the association of early surgery (ES), late surgery (LS), and no surgery (NS) with outcomes in a population of patients (N=89) with cranial penetrating GSW. (Helling et al., 1992) Early and late surgery were defined as less than 24 hours and greater than 24 hours, respectively. Reasons provided for delayed surgery were an initial determination of poor prognosis or that of minor injury. The study period spanned 1987 to 1989 and was retrospective by design. Twenty-seven patients underwent early surgery, six underwent late surgery, and 56 underwent no surgery due to clinical findings and presumptive poor prognosis as determined by local neurosurgeons. The authors report a statistically significant difference in mortality for the groups (ES [N=10, 37%], LS [N=0, 0%], NS [N=46, 82%], p<0.001). It should be noted that the mean Glasgow Coma Scale (GCS) score of the groups differed (ES [7.86 +/- 4.72], LS [12.17 +/- 4.10], and NS [5.59 +/- 4.42]). Patients in the NS group were the most severely injured and were the least severely injured in the LS group. There were fewer patients with GCS scores 3 and 4 in the ES group (N=11/27) compared with the NS Group (N=37/56). However, there were more survivors in the ES Group (N=4/11) compared with the NS group (N=1/37) amongst those patients with GCS 3 or 4 (p=0.007). No patients in the late surgery group died; however, the mean GCS of this cohort was much higher at presentation. There was one infection (brain abscess) and this occurred in a patient who underwent late surgery.

A third non-randomized, retrospective study from Ecker et al examined the impact of early versus delayed decompressive bilateral or bicompartmental craniectomy in a U.S. military population (N=33).(Ecker et al., 2011) The median age of the population was 24 years (range 19 to 46 years) and median GCS 5 (range 3 to 14). Nineteen patients underwent bifrontal decompression (58%), eight underwent bi-hemispheric decompression (24%), and six underwent both supra- and infratentorial decompression (18%). Eighty-eight percent of injuries in this cohort were due to blast injury and 12% from GSW. All but three patients underwent early decompression prior to evacuation. Poor outcome was defined as Glasgow Outcome Score (GOS) 1 to 3 at 6 months and at 1 to 5 years. At 6 months, 16 patients in the ES group had poor outcomes as compared with none in the LS group. Among good outcomes at 6 months, three (18%) patients underwent delayed surgery and 14 (82%) underwent early surgery (p=0.10). Among patients for whom GOS were available at 1 to 5 years, 12 patients had a poor outcome and 18 had good outcome. Of the patients with poor outcome, 100% underwent early surgery (N=12). Of those patients with good outcome 16 (89%) underwent early surgery and two (11%) underwent delayed surgery (p=0.36). Importantly, no time from injury to surgery were defined for early or delayed surgery by the authors.

Dolachee et al provide a fourth prospective and retrospective, non-randomized study of civilian injuries in Iraq (N=39). The investigators examined the association of surgical timing with outcomes.(Dolachee et al., 2019) Time to surgery from injury was defined as less than 5 hours (N=10), 5 to 8 hours (N=22), and 9 hours or more (N=5). Surgical intervention included craniotomy or craniectomy, debridement, hemostasis, and watertight dural repair. The authors report good outcome for five patients (50%) in the less than 5 hours cohort, 14 (63%) in the 5 to 8 hours cohort, and one (20%) in the 9 hours or more cohort. Poor outcomes were reported in five patients (50%) in the less than 5 hours cohort, eight (36%) in the 5 to 8 hours cohort, and four (80%) in the 9 hours or more cohort (p=0.2). It should be noted that no clear definition was provided for good or poor outcome, although GOS was referenced.

In a fifth non-randomized, retrospective study, Siccardi et al examined time to surgery from injury in a population of civilians (N=44) in Italy with GSW.(Siccardi et al., 1991) Time intervals from injury to surgery were within 3 hours, within 6 hours, and greater than 6 hours. Surgical intervention included debridement of necrosis, removal of in-driven fragments, evacuation of hematoma, hemostasis, and watertight dural closure. Of patients who underwent surgery within 3 hours (N=2), one patient survived and one patient died (50% mortality). Of patients who underwent surgery within 6 hours (N=33), 19 survived and 14 died (mortality 42.4%). Of patients who underwent surgery greater than 6 hours from time of injury (N=9), six survived and three died (mortality 33.3%).

Sertbas et al is the sixth non-randomized, retrospective study included as evidence. It examined the impact of duration of time between injury and surgery on morbidity and mortality in a military/civilian (civil war) population in Libya with craniospinal GSW.(Sertbas & Karatay, 2020) Data for morbidity was reported in aggregate for cranial and spinal GSW and could not be included. The association of time to surgery on mortality for patients with cranial GSW (N=35) was examined. Surgical intervention included duraplasty (N=29, 82%), decompressive craniectomy (N=21, 60%), hematoma evacuation (N=20, 57%), and bone fragment removal (N=20, 57%). The mortality for the overall cohort was 22.8% (N=8). Time intervals from injury to surgery which were examined were less than 24 hours, 24 to 72 hours, greater than 72 hours. Mortalities by time from injury to surgery were: one in less than 24 hours, three in 24 to 72 hours, and four in greater than 72 hours. For patients with cranial GSW the authors report a significantly lower mortality when operated on within 24 hours (p=0.044).

In summary, the six studies identified consisted of GSW and blast injury,(Dolachee et al., 2019; Ecker et al., 2011; Wei et al., 2013) and GSW alone (Helling et al., 1992; Sertbas & Karatay, 2020; Siccardi et al., 1991). Three of these studies are civilian (Dolachee et al., 2019; Helling et al., 1992; Siccardi et al., 1991), two are mixed civilian and military,(Sertbas & Karatay, 2020; Wei et al., 2013) and one is military (Ecker et al., 2011). Military and civilian trauma systems are different particularly in the prehospital resuscitation and evacuation systems, and therefore are not equivalent when assessing medical evidence. As a result, generalizations on outcome may not apply. The decision for early versus late surgery in these six reviewed articles varies. Some patients with severe pTBI were deemed to have a poor prognosis and were relegated to late or no surgery.(Helling et al., 1992) This skews results in favor of performing early surgery for better outcomes. Patients with higher mean GCS were relegated to the late surgery group and this also skews the results.(Helling et al., 1992) In the military setting, delays in casualties reaching neurosurgical care also resulted in late surgery being performed. The timing of surgery in these six articles varies considerably. What is characterized as early surgery in these six articles ranges from <3 hours to <24 hours, and in Ecker et al is undefined.(Ecker et al., 2011)

All six of these studies were judged to be at high risk of bias. Together they provide very low strength of evidence that early surgical intervention may reduce the likelihood of death/other poor outcome in more severely injured patients compared with delayed surgery and highlight a lack of agreement for what timeframe constitutes early versus delayed. Associations with other outcomes could not be determined due to the methods of reporting data (not by time to surgery) in the included studies. These six non-randomized observational studies had a high risk of evidentiary bias because of variability in the cohort size and composition, the non-standardized definition of early surgery and some patients being relegated to have no surgery. Surgical techniques were also not standardized. Studies conducted in war zones and low and middle income countries, (Dolachee et al., 2019; Sertbas & Karatay, 2020; Wei et al., 2013) cannot be easily compared to studies conducted within high income civilian trauma systems. (Ecker et al., 2011; Helling et al., 1992; Siccardi et al., 1991) It is unsurprising that no randomized trials of the timing of surgery in pTBI have been identified in the literature given the emergent nature of this issue. Moreover, important differences between characteristics of patients treated early vs. late are undoubtedly important confounds.

Though not taken formally as evidence, we additionally noted other relevant studies that merit discussion. Shackelford et al reported a retrospective series of 486 military combat casualties from Iraq who had combat-related brain injury and had a craniectomy, and found that postoperative mortality was significantly lower when craniectomy was initiated within 5.33 hours of injury.(Shackelford et al., 2018) 209 (43%) of patients had pTBI (168[35%] had GSW, 239[49%] explosives, and 79[16%] other). This 5.33 hour time point informed our consensus statement on the timing of surgery. These results have also encouraged the aggressive surgical management of combat casualties with pTBI. Aarabi et al in a series of 964 patients from the Iran-Iraq war, reported that the incidence of infection was lower when surgery was carried out in the first 24 hours rather than no surgery or evacuation to a more distant base hospital with further delay.(Aarabi et al., 1998)

For patients with closed or blunt TBI, early evacuation of surgical lesions has become standard practice. Evidence based guidelines have been published for the surgical management of TBI. It is recommended that patients with acute subdural hematoma (ASDH) and acute epidural hematoma (EDH), where surgery is indicated, be operated as soon as possible. (Bullock, Chesnut, et al., 2006a, 2006b; Bullock, Chesnut, Ghajar, Gordon, Hartl, Newell, Servadei, Walters, et al., 2006; Bullock, Chesnut, et al., 2006c, 2006d) However, these guidelines do not include pTBI. Moreover, the 'four hour rule' was developed from a series of 82 patients with closed TBI and ASDH. The mortality increased to 90% compared with 30% if surgery was delayed >4 hours. Functional survival was achieved in 65% of patients (GOS ≥ 4) in patients having surgery <4 hours. (Seelig et al., 1981) However, management paradigms for closed TBI are not directly applicable to patients with pTBI.

Patients with craniocerebral pTBI are usually treated acutely as there is inherently an open wound, which requires early debridement and closure; there is often significant brain swelling with pathological midline or trans-tentorial/foramen magnum shifts; CSF leaks require acute repair to prevent infection; and early decompression of large or expanding hematomas may be indicated. Indeed, the high profile case of the life-threatening cerebral GSW injury to Congresswoman Gabrielle Giffords in the USA in 2011 - who was operated about 38 minutes after arrival at the hospital, has drawn attention to putative benefits of urgent and aggressive surgery in patients with severe craniocerebral GSW.(Lin et al., 2012; Rosenfeld et al., 2015) Decompression/craniectomy is indicated in patients with pTBI and brain swelling/hematoma. Craniotomy may be performed for more localized trauma or where brain swelling/hematoma is not significant or not likely to worsen.

Current military casualty care positions surgical facilities within the area of combat operations such that early surgery (within 3 to 4 hours) is undertaken. The injured soldier, once stabilized, is then evacuated to a facility with a higher level of care. Delays in transporting the injured soldier for neurosurgery may occur due to various operational constraints. These may include damage control general surgery performed in a forward hospital where there is no neurosurgery, or restrictions in aeromedical transport opportunity or availability. Civilian trauma systems have evolved to transport a casualty rapidly to a trauma center where early surgery including neurosurgery can be achieved.

Based on the included studies reviewed and Shackleford et al, it is recommended surgery be performed acutely, or within 5.33 hours.(Shackelford et al., 2018) The Committee on Surgical Combat Casualty Care (CoSCCC) Position Statement: Neurosurgical Capability for the optimal management of brain injury during deployed operations was developed by the US military for the care of combat casualties, but also was designed to apply to civilian trauma systems. The recommendation is that provision should be made for casualties to access neurosurgical care as soon as possible, but within 5 hours of injury.(Gurney et al., 2023) Earlier surgery is recommended to prevent CSF leakage and chances of infection (Also refer to KQ 35).

If the clinical picture is unclear, initial local debridement, closure and observation with an ICP monitor if indicated +/- multimodal monitoring is recommended to guide treatment. Decompressive craniectomy may be indicated if the patient deteriorates. If there is no mass lesion and no evidence of elevated ICP, but there are bony fractures that require repair, it is recommended these be addressed as a typical depressed skull fracture to prevent infection and neurologic worsening.

Human brain is a privileged organ, which is immersed in CSF and protected by scalp, skull and pachy-/leptomeninges. pTBI violates this natural integrity, which should be repaired as soon as possible to prevent infection. Early surgical management, preferably within five hours after trauma improves functional outcome and reduces mortality. We also recommend 5 hours as a reasonable cutoff time for early surgery in future research based on the Shackelford et al investigation.(Shackelford et al., 2018) Patients with pTBI who do not require urgent decompressive surgery may have surgery >5 hours from the time of injury, where there is systemic constraint or no neurosurgeon available in the primary center.

There does appear to be equipoise in two scenarios regarding timing of surgical intervention: 1) presence of an open wound and retained foreign body in a neurologically intact patient, and 2) the patient who presents with no intracranial pathology (for example, from a tangential penetrating injury), and the need for wound washout. In both of these scenarios, prompt but not emergent interventions seems reasonable. There is not adequate data to provide a "safety cutoff" time. In these cases it would be recommended to intervene as soon as safe and clinically feasible.

While the literature that exists on this topic is very limited, it is recognized that a prospective randomized controlled trial on surgical timing would be not only difficult but also likely unethical. As such, careful and detailed prospective nonrandomized observational studies could help with timing of surgery in pTBI patients, being military or civilian.

To improve the quality of the pTBI evidence for early versus late surgery, there is a need to establish a widely agreed upon definition of early versus late surgery. As improved classification of pTBI emerges over time it is possible that patients with different characteristics (severity and types of pTBI patients based on mechanism of injury, clinical status, imaging, laboratory investigations and serum biomarkers) will reveal that some patients require surgery more urgently than others. There is a need to define surgical techniques with more precision and to better describe trauma system factors including extent of prehospital resuscitation. Relevant studies would be assisted by standardized outcome assessments following pTBI including GOSE, Modified Rankin Score, Quality of Life instruments, epilepsy status, infection, PROMs (patient reported outcome measures) and duration of follow-up.