There is insufficient evidence to support a strong, moderate strength or weak recommendation for surgery or endovascular intervention to reduce the risk of traumatic intracranial arterial aneurysm (TICA) rupture and improve outcome.

Endovascular treatment of TICA, which often require re-treatment, is recommended as a means of reducing the risk of aneurysmal rupture and the associated high rate of morbidity and mortality. There is, however, insufficient evidence to support superiority of endovascular aneurysm occlusion over open surgical parent vessel sacrifice despite a theoretical reduced risk of iatrogenic stroke.

Open surgery may be considered in the setting of TICA resulting from pTBI to reduce the chance of rupture and improve outcome.

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

• A multi-disciplinary approach to vascular sequelae of pTBI (TICA, A-V fistula, vasospasm, etc.) for patients with traumatic vascular injury(ies) is preferred when feasible.(100% consensus)
• It is recommended that TICAs which increase in size be treated with either open or endovascular means because of a presumed elevated risk for rupture (87.5% consensus).
• It is recommended that larger TICAs should be more strongly considered for treatment because of a presumed elevated risk for rupture. There is insufficient evidence or expert consensus to recommend a size threshold, however. (80.5% consensus)
• It is recommended that TICAs treated endovascularly should be reassessed with cerebral angiography within 8 weeks of initial treatment because of a risk of treatment failure (i.e. aneurysm recurrence or re-growth) for which further therapeutic procedures may be necessary. (95% consensus)
• In certain circumstances it may be permissible for TICAs to undergo a short period of conservative management in hopes of achieving spontaneous regression. Close observation is required if such an approach is attempted. Reasons for observation may include the lack of available neuroendovascular expertise, or a perceived unfavorable risk:benefit ratio inherent to treatment (95% consensus).

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

Penetrating traumatic brain injury (pTBI) often injures the cerebral vasculature.1-19 In fact, the high frequency of complex neurovascular injuries might be considered a defining feature of pTBI and a key characteristic distinguishing pTBI from blunt TBI. Damage (including complete or partial transection) or irritation to the arterial wall can be associated with the formation of TICA.4 These "pseudo aneurysms" lack an aneurysmal wall and are contained only by clot in the subarachnoid space, distinguishing them from saccular or fusiform aneurysms. Moreover, TICAs are highly unstable and unpredictable, and are associated with delayed development, growth, rupture, and in some cases spontaneous healing.1,4,6,10,17-19 Other cerebrovascular pathologies associated with pTBI include traumatic cerebral vasospasm, arterio-venous fistulas and complete occlusion of an arterial branch causing stroke.1,14,19,20 Historically there has been uncertainty as to when and how to screen for traumatic aneurysms, and if, in what circumstances and by what means they should be treated.1,4,9,13,19 Prior to the advent of specific endovascular techniques and technologies, open surgery following conventional angiography was the main treatment available.2-4,19 Open surgical treatment of TICAs almost always requires sacrifice of the parent artery and is technically challenging as the vessel typically bleeds profusely as efforts are made to expose the lesion. This can, of course, cause a debilitating stroke if the artery supplies eloquent cortex.2,4 Endovascular treatments offer a less invasive approach to certain lesions, but may be hampered by the rapid need for surgical decompression prior to intervention, or, in certain circumstances, the need for systemic anti-platelet or anti-thrombotic medications.1,9,11,18 The possibility of preserving the parent vessel is a strong theoretical advantage of endovascular treatment of TICAs as compared with open surgery. Based on our current understanding of TICAs, we believe that they are as likely to hemorrhage as a ruptured, untreated saccular aneurysm with a similar rate of mortality.

The prior pTBI guidelines addressed "Vascular Complications of Penetrating Brain Injury." Recommendations created in this edition were based on 8 articles defined as Class III.2-4,16,19,21-23 Of these articles, 4 addressed the surgical management of traumatic aneurysms2-4,19, two addressed the incidence of subarachnoid hemorrhage in pTBI21,23, one addressed the incidence of traumatic arteriovenous injury following pTBI, and one addressed incidence of SAH and vasospasm following pTBI22. Based on the rigor and focus of the current guidelines along with the specific focus of the key question asked, three of the prior included studies met criteria for inclusion in the current effort2-4.

Background

Traumatic cerebrovascular injury is central to the management of pTBI. The incidence of TICAs resulting from pTBI in larger studies ranges from 3-40%.1,4,6,16 Contrasting with spontaneous saccular aneurysms that have a dome comprised of connective tissue (adventitia), TICAs lack such a dome. These traumatic "false" aneurysms are generally confined instead by organized blood clot within the subarachnoid space which is adjacent to the area of vessel damage.4,10 This results in an unpredictable natural history characterized by delayed formation and presentation, spontaneous healing, interval growth, and delayed rupture.1,4,10,17-19 TICAs rupture has been associated with significant morbidity and mortality. In one study, delayed rupture was associated with a nearly 50% mortality rate.1

Physicians treating traumatic TICAs have had to balance the risk of rupture with the risks of treatment. Historically, TICAs were treated with open surgery, either at the time of initial presentation and surgical intervention, or in a delayed fashion after follow up imaging based diagnoses.2,4,19 This has typically involved significant risk of intraoperative rupture and the near ubiquitous need for parent vessel sacrifice which can lead to an iatrogenic stroke - sometimes of eloquent brain.2,4,19

Technological advancements in endovascular materials, including detachable coils, aneurysm stents, and flow diverting devices, have made endovascular occlusion of TICAs a viable alternative. In contrast to the perceived impression that held that TICAs could not be coiled, experience and evidence has demonstrated that endovascular treatment is feasible.1,7,9-11,17,18,24-26

Given the inherent complexities and divergent expert opinions on the topic, the management of TICAs is a pTBI subtopic which is particularly in need of evidence and consensus-based guidance which seek to identify and disseminate best care practices. Here we present evidence informing the open surgical and endovascular treatment of TICA supplemented with expert consensus statements.

Open Surgical Management of Traumatic Aneurysms

Eight non-randomized studies (7 case series, and 1 case report) were determined to be of sufficient quality to be included in the evidentiary base. These studies provided evidence on the risk of TICA rupture following surgery compared with no surgery.1-7 All studies were rated high risk of bias due to study design. Injuries were due to high velocity missiles or stab wounds. To briefly summarize, a total of 1768 patients with penetrating head injury and vascular imaging were included in these studies. Within this population, a total of 129 TICAs were present in 110 patients (total incidence of traumatic aneurysms from pTBI 7.3%). A total of 65 patients received open surgery to address 74 aneurysms.

Bell et al evaluated 187 patients (162 penetrating) who received cerebral angiograms an average of 10 days after their primary injury.1 A total of 31 TICAs were diagnosed in 24 patients in this study (total of 64 vascular injuries in 48 patients). Nine of 31 TICAs in 5/24 patients were treated with primary clip occlusion, and 3/31 TICAs in 3/24 patients were treated with a combination of coil occlusion and clip ligation. No ruptures were observed in patients treated with primary or secondary clip reconstruction. Eight TICAs in 8 patients were not treated, with rupture occurring in one of those patients resulting in death. Overall GOS was noted to be 3.1 in this study, without a specific outcome attributed to the TICA subgroup.

Haddad et. Al (N=15) described findings associated with TICAs at presentation.2 14 of 15 patients in this study presented in the pre-CT era.4 Four patients presented with subdural hematoma, all attributed by the authors to aneurysm rupture. Twelve patients presented with intracerebral hemorrhage, 11 attributed to aneurysm rupture by the authors. Of specific note, 5 patients suffered deterioration of clinical examination in a delayed fashion. One patient had a delayed headache and seizure, one a new hemiparesis, and three a deterioration of consciousness. While suspicious for delayed rupture, no specific conclusion can be made. Two of the five patients had what was described as a "good" outcome, one with "poor" outcome, and two with death. In all cases, surgical intervention (11 trapped and excised, 1 coagulated, 1 treatment completed but not specified) appeared in this study to occur after hemorrhage findings associated with the aneurysm. Outcomes in this study were attributed to the condition of the patient at presentation, without association specifically to the traumatic aneurysm. However, 2/13 surgical patients died. There were no reported aneurysm ruptures after treatment.

Aarabi evaluated 255 patients with penetrating brain injury who received cerebral angiography an average of ~17 days after injury during the Iran-Iraq War.3 Eight of 255 patients presented with traumatic aneurysms. Two of 8 ruptured prior to treatment. 7 of 8 aneurysms were treated with surgery and clip exclusion of the aneurysm. Of those treated surgically, 2 patients died, and 5 recovered with varying degrees of outcome (confusing, as the study published what looks like the MRS outcome by score, but referred to it as the GCO or Glasgow Coma Outcome). There were no reported ruptures after surgical treatment. The patient not treated had a favorable long-term recovery at 47 weeks.

Kieck et. Al evaluated the incidence of vascular injury in 109 patients with cranial stab wounds with 79/109 receiving cerebral angiograms.5 A total of 11 TICAs were noted in 11 patients. Six of 11 patients received surgery described as "clip, trap, or excise", without specific attribution of type of surgery to a specific patient. The paper reports good outcomes in all six of these patients, however it does note a delayed hemorrhage in a patient receiving a prior craniotomy and aneurysm clip at day 98 from a delayed rupture from a second aneurysm confirmed at autopsy (0 or 16% rupture risk after surgery). In this case, 1/6 patients is suspicious for delayed rupture after initial treatment. A total of 5/11 aneurysms were not treated, with aneurysm rupture and death occurring in 4/5 patients (80%).

Du Trevou et. Al evaluated the incidence of vascular injury in 250 patients stabbed in the head, with 181 patients receiving cerebral angiography.6 A total of 21 patients harboring 24 TICAs were noted in this population. Of those patients, 16 patients received "surgery", however the type of surgery was not specifically specified. Of the 5 patients not receiving surgery, 2 patients refused with outcomes not recorded, 2 patients aneurysms resolved on follow up angiogram, and 1 patient died. Of note, no patients who were treated surgically suffered a re-hemorrhage.

Mansour et. Al evaluated 128 patients with penetrating brain injury, and included the 72 patients who received a CTA within 72 hours.7 A total of 37 arterial injuries were noted in 26 patients, with 11 TICAs noted in 8 patients. Of those patients, only one patient received surgery, 3 patients received coil occlusion, and 4 patients were observed. Outcomes in the observed population were noted to be unfavorable, with no reference to delayed hemorrhage in either the treated or untreated groups.

Other studies reported the quality of overall outcomes (e.g., good, fair, poor, unfavorable, died, did well, surgery successful) with surgery versus no surgery or reported Glasgow Outcome Scale score after surgery or at discharge (Table). Patients tended to do better with surgery compared with no surgery. However, individuals who did not have surgery may have been more gravely injured (likely not to survive anyway) or have smaller aneurysms (where watchful waiting may be an acceptable treatment) than individuals who underwent surgical clipping or other surgery.

Nathoo et. Al evaluated the outcomes in 597 patients with transcranial stab wounds, specifically focusing on 17 with trans-brainstem penetrating injury.8 Within this population, 1 patient was diagnosed with a delayed basilar bifurcation aneurysm. This patient received craniotomy and aneurysm clipping, but ultimately died. Autopsy revealed that the aneurysm was "well clipped", and did not mention delayed hemorrhage. A ruptured posterior communicating artery aneurysm was discovered at autopsy in a second patient. This patient, by report did receive a cerebral angiogram prior to death which was negative. It is not clear from this study whether or not the rupture was the cause of death. However, the aneurysm was untreated, with a 100% incidence of rupture.

Overall, these studies provide very low level evidence that surgery in the setting of TICA may reduce the risk of rupture and may lead to more favorable outcomes than no surgery.

B-Endovascular Management of Traumatic Aneurysms

A total of 9 studies were determined to be of sufficient quality for inclusion in the evidentiary base. Bell et al evaluated outcomes in patients with pTBI and TICAs treated with coil occlusion (N=12), and one additional treated with stent buttress coil occlusion.1 Most injuries were blast-related (67%), with gunshot wounds (25%) and knife wound (8%) the other mechanisms of injury. The anterior cerebral artery was the vessel most commonly affected (60% of patients) followed by the middle cerebral artery (25% of patients). In general, those TICAs noted to increase in size or that were treated in this study were larger than those that were not treated. Following initial TICA treatment, 5/12 patients (42%) patients experienced aneurysm recurrence and required re-treatment with coil or stent-coil (3/5) or clip (2/5). There were no aneurysm ruptures that occurred following coil occlusion in this study. This compares with one recurrence/retreatment (20%) in the five pTBI patients with TICA treated with a clip, and one rupture (12.5%) in the eight pTBI patients whose TICA was not treated. However, the size/severity/location of the TICA was likely a confounding factor in the specific treatment or lack of treatment provided (Table).

In another non-randomized study (N=2) from a Level 1 trauma center in Jerusalem, Cohen et. Al evaluated two patients with pTBI and TICAs were treated with endovascular coils.9 One patient (17 years) suffered a missile wound in a bus bombing and following treatment had a Glasgow Outcome Scale (GOS) score of 5; the other patient (5 years) was stabbed with an umbrella (transorbital-transcavernous) with a post-treatment GOS of 4 (Table).

In a third non-randomized study (N=2), Hemphill et. Al reported one adult patient with a nail gun injury and TICA that was treated with endovascular coils and recovered without infarction (callosomarginal territory).10 The other adult patient in this study experienced a TICA due to a gunshot wound to the head. Follow-up angiography at 2 months indicated an enlargement of the TICA, which was then treated with detachable coils, sparing the parent artery (Table).

Another study (Mansour et al.) reported that three patients with pTBI were treated with endovascular embolization for TICAs but did not mention the use of coils.7 One patient had a GOS of 6 at discharge, one with a GOS 4 at discharge, and one patient died in the hospital (Table). There is no mention of aneurysm re-rupture in this study.

Other reports of treatment of TICA in pTBI with endovascular coiling were limited to case reports. Results were mixed (Table).11, 17, 18, 24-26

These studies provide very low level evidence that coil embolization of TICA in the setting of pTBI may reduce the risk of further bleeding outcomes or need for retreatment but do not provide evidence for endovascular coiling compared with other treatments in the absence of potential confounding factors.

As has been previously discussed, penetrating brain injury is a highly lethal wounding mechanism that is very difficult to study with prospective, randomized trials. With this in mind, the body of literature in general, and on this subject in particular, remains largely retrospective, observational, or case report in nature (Class III or IV), and therefore does not support Level I or II recommendations. That said, it is the lethal and unpredictable nature of penetrating brain injury in general, and TICAs in specific, that support the nature of the above recommendations.

At the time of the creation of the prior pTBI guidelines in 2001, open surgical management of TICAs (when feasible) was the primary means of treatment.2-4 This was in part secondary to the early and somewhat rudimentary nature of the endovascular technologies available at that time as well as a concern that coils would not be contained by the TICA clot. Technologies, including the advent of aneurysm stents and flow diverters, have advanced considerably.

The purpose of treatment of TICAs is not to reverse the primary injury, but to help reduce the risk of additional brain injury, often catastrophic, from aneurysm rupture. Open surgical techniques remain a viable approach but can involve substantial hemorrhage (especially if the TICA is large and proximal) as well as the morbidity of stroke in the distribution of the parent artery if sacrifice is necessary. From an algorithmic perspective, this approach may be preferred at the time of initial surgical intervention for superficial and active arterial bleeding. However, endovascular occlusion in the acute setting is a viable alternative that may 1) reduce the risk, at the time of open surgical decompression, of catastrophic hemorrhage from large, skull base arterial injury, 2) reduce the need for return to the operating room for clip reconstruction of an aneurysm in an injured and inflamed cortical environment, 3) identify, with the necessary pre-treatment gold standard cerebral angiogram, arterial injury (TICA, arteriovenous fistula, etc) not visualized with CTA, 4) allow for treatment of arteriovenous fistula sub-optimally visualized on non-invasive imaging, and 5) provide a means of preserving the parent artery at the time of initial treatment (see Traumatic Vascular Injury algorithm).

Included Evidentiary Support for Open Surgical Management of TICA's

The Level IV recommendation supporting consideration for open surgical management of traumatic intracranial aneurysms is based on a total of 110 patients who harbored 129 TICAs in the studies referenced.1-8 Within this population, a total of 65 patients received surgery (described as "clip", "trap and excision", "coagulation", or "unspecified") as the first operation to treat a total of 72 TICAs. Another two patients received a surgical clip after initial coil occlusion, and one received coil occlusion of a previously clipped aneurysm. Outcomes in this group were described as GOS 1 or "death" in 11 (16.2%), GOS 2-3 or "poor" in 4 (5.9%), GOS 4-5 or "good" or "fair" in 45 (66.2%), and not specified in 8 (11.8%). There were no reported aneurysm re-ruptures after initial or secondary treatment in this group. In comparison, another 30 patients harboring 34 aneurysms were observed or received no treatment. Within this group, outcomes were described as death in 8 (26.7%), poor or unfavorable in 5 (16.7%), TICA resolved or outcome good in 13 (43.3%), and lost to follow up in 4 (13.3%). Concerning mortality, in 6/8 deaths, definitive rupture was noted as causal (20%).

Concerning TICAs, and while the study risk of bias is considered high given the study designs, there appears to be a uniform effect with respect to protection from aneurysm rupture following treatment with open surgery. This treatment effect was present at each outcome measure, but could not be verified by meta-analysis given the small population and variable reporting measures. It should be noted that surgical management in the referenced studies was described in a number of ways, and almost always included parent artery occlusion with coagulation or aneurysm clip, often with resection of the traumatic aneurysm. There is no mention of parent artery preservation in any of the studies referenced. Neurologic outcomes, when included, are almost always attributed to the degree of underlying brain injury at presentation in those treated. With this in mind, it should be remembered that open surgery carries a risk of stroke with parent vessel sacrifice. This may not be a concern if the artery supplies already injured brain. It is a particular concern if the vessel supplies uninjured, eloquent brain.

Included Evidentiary Support for Endovascular Management of TICAs

The Level IV recommendation supporting endovascular occlusion of intracranial traumatic aneurysms in pTBI to reduce delayed hemorrhage is based on the reported literature experience of 33 endovascularly treated aneurysms, with 25 treated with coil occlusion, (one with stent-coil), one coil re-treated with flow diversion, and 8 treated with parent artery sacrifice.1,7,9-11,17,18,24-26 The rate of re-rupture following initial treatment was 1/25 (4%). In this one case, the aneurysm was diagnosed at the time of patient presentation with a protruding foreign body (knife), the aneurysm was coiled with the microcatheter left in place, and continued bleeding after the knife was removed was immediately treated. Overall, while counterintuitive to coil a false aneurysm or blood clot, the data support this option as a means of preventing aneurysm rupture and preserving the parent artery. One study specifically noted a rate of parent artery preservation as high as 50%.1 While this offers an advantage over surgery, the aneurysm recurrence rate is relatively high, necessitating early post intervention follow up angiography (preferably within approximately 4-8 weeks). Aneurysm re-treatment may be necessary should recurrence be noted. In these cases, technology that includes the use of one of the many flow diverters currently available may be a viable option, as was demonstrated in the case report by Aljuboori et. al.11 Flow diversion as a primary means of treatment may also be considered, depending on the location of injury and the need to preserve the parent artery (i.e. Internal carotid, proximal MCA/PCA/ACA). As flow diversion mandates the extended use of dual antiplatelet therapy (3 or more months), the timing of this intervention must take into account the need for additional open surgical intervention.

Limitations

It should be noted that there remain relatively few published cases or studies of the open surgical and endovascular management of penetrating traumatic cerebrovascular injury. The level IV recommendation reflects this reality.

Traumatic Aneurysms

It is our opinion that one of the most important aspects of these new pTBI guidelines are those related to TICA. Even amongst recognized TBI experts there have been diverse opinions on screening for TICA and their need for treatment. This likely relates in large part to the fact that even for most neurotrauma experts these lesions are encountered infrequently. Our panelists represent many of the world's highest experts with the most experience with pTBI and TICA. The resounding opinion of our experts is that in most circumstances an aggressive effort to detect TICA is needed and that they should almost always be treated urgently when identified. Our experts recounted numerous disastrous outcomes from re-hemorrhage in cases not managed aggressively. While there may still be a role in select circumstances for observing these lesions in hopes they will regress, our panelists felt that such an approach should only be rarely attempted. Small aneurysms where treatment risks infarcting eloquent brain may be such a circumstance. If an initial conservative approach is attempted, frequent re-imaging is needed and treatment should be initiated if the lesion grows or does not regress.

Another key aspect of these 2^nd edition guidelines is ensuring practitioners are aware of evidence and experience related to coiling or other endovascular management strategies of pseudoaneurysms. The historic belief that pseudoaneurysms cannot be coiled is clearly incorrect. It is important that the feasibility and potential superiority of endovascular treatment is widely disseminated.

That said, while the available evidence supports screening for traumatic aneurysms (preferably with DSA) and treating them, it does not at this time support evidence-based recommendations concerning the timing of initial diagnostic assessment and treatment of pTBI related traumatic aneurysms, whether observation or non-treatment of some traumatic aneurysms is warranted and the conditions where this might apply, the size threshold for mandatory treatment, and the frequency of follow up diagnostic imaging in the setting of observed or treated traumatic aneurysms. With this in mind, a DELPHI consensus model (>80%) was obtained through dedicated, in-person survey of the multi-disciplinary subject matter expert team during a 5-day guidelines consensus conference. Aneurysm treatment, timing of follow up evaluation for treated aneurysms (7-21 days, see Key Questions 2-4), and conditions where observation may be warranted were addressed. While the diagnosis of pTBI associated vasospasm was addressed with evidence based recommendations, endovascular treatment was not addressed in this guideline.

We are aware of diverse management strategies for TICAs including some who currently favor conservative management of these lesions. It was, however, the strong opinion of our expert group that a failure to diagnose and treat these lesions can have catastrophic consequences. Our expert panel advocates, and the evidence supports, aggressively screening for and treating TICAs. These lesions require complex judgment individualized to specific patients and circumstances, however. Clinicians caring for pTBI patients must have both a high index of suspicion in pTBI, and a low threshold for treatment. In short, all traumatic aneurysms should be considered for treatment, and consensus was achieved amongst our panelists that large aneurysms and those noted to increase in size must be treated with particular urgency (80.5% and 87.5% Consensus respectively). A consensus recommendation for minimum size threshold for aneurysm treatment was not obtained. One study (Bell et al) identified the average size of aneurysms treated (4.72 mm), observed (2.21 mm), and ruptured (8.3 mm).1 While smaller aneurysms have been noted to heal on interval imaging, they have also been noted to grow, rupture, and result in death. It is with this in mind that the consensus recommendation for observation under certain specific conditions was made (95% Consensus). In this setting, early follow up cerebral angiography (within 4-8 weeks) should be considered.

It should be noted that aneurysm recurrence in the setting of endovascular occlusion of traumatic aneurysms was reported as high as 27%.1 This is a significant limitation of this type of intervention, and requires early follow up (4-8 weeks) cerebral angiography as noted above (95% consensus). Secondarily, it is also acknowledged that the full scope of endovascular treatment modalities, including aneurysm stents and flow diverters, may be considered too high risk in the acute setting secondary to the need for dual anti-platelet therapies with these interventions. These approaches may be best considered with delayed or recurrent traumatic aneurysm presentation.11

The addition of a recommendation to consider multi-disciplinary approach to TICAs arises from the understanding that the treatment of cerebrovascular pathology of any kind requires consideration of open and endovascular surgical strategies. This recommendation also reflects the fact that endovascular techniques have advanced considerably over the past 22 years, making management of these highly complex and unstable conditions safer. There was uniform consensus within the guidelines subject matter expert working group on this matter.

Given that endovascular coil occlusion of TICAs has proven feasible and efficacious and with the ability to preserve the parent artery, it is our view that endovascular treatment is now preferred over open surgery when feasible. Endovascular surgery is likely safer with less risk of blood loss and the probability of avoiding a stroke in the distribution of the parent artery. Repeat endovascular treatment is often needed, however. Endovascular treatment is not universally available, however, particularly in austere environments. There are also TICAs that cannot be treated or optimally treated with endovascular therapy. For this reason open surgery remains an important means of treating TICAs as indeed the alternative is a high risk of mortality.

Traumatic Vasospasm and AV-Fistula

The available evidence did not support the creation of specific evidence-based recommendations concerning the treatment of traumatic vasospasm. Studies included in the 1st edition guidelines addressed the incidence and outcomes of patients with vasospasm following pTBI, but these studies did not meet inclusion criteria for the current guidelines.22,23 An additional study that did not meet inclusion criteria (Armonda et. al) evaluated the outcomes in patients with traumatic vasospasm during wartime.20 While this study demonstrated worse outcomes in patients with traumatic vasospasm, endovascular treatment modalities did not improve outcomes. It is our opinion that screening for vasospasm following pTBI should be considered, and if diagnosed, critical care and perhaps endovascular interventions also considered.

Concerning traumatic arteriovenous fistula, these entities should be diagnosed and treated in a manner similar to TICAs. CTA is not an effective method of diagnosing dynamic pathologies, further supporting the recommendation for DSA following pTBI. The presence of an AV fistula in pTBI usually signifies a significant arterial injury that will require treatment.

Our understanding of how TICA characteristics like size and location influence risk of rupture and the risk:benefit of different treatment options is advancing, but these remain very important topics for further research and elucidation. Much remains insufficiently understood about TICA natural history. We acknowledge that if neurosurgeons adhere to our recommendations and treat pseudoaneurysm aggressively, it will impede our ability to understand their national history. In our opinion it would now be unethical to randomize patients to treatment versus no treatment or to open surgical versus endovascular treatment. Additional work in characteristics associated with coiled traumatic aneurysm re-treatment, techniques to and the risks and benefits of different endovascular and open surgical therapies would also be beneficial. High quality prospective studies which assess outcomes based on different aneurysm characteristics and treatments would be highly valuable contributions to the literature. A study comparing the incidence of iatrogenic stroke based on the mechanism of treatment would be very helpful. As well, the optimal timing of diagnosis and treatment, and characteristics associated with aneurysm healing, would be beneficial.