Blunt cerebrovascular injury in trauma patients: an under-recognised injury pattern at Auckland City Hospital

Injury to the carotid and vertebral arteries after blunt traumatic injury, or blunt cerebrovascular injury (BCVI), is rare but potentially life-threatening, and has been reported in 1–3% of blunt trauma patients. Untreated BCVI is associated with a substantial rate of stroke (8–25%) and stroke/death (35–55%);^1,2 medical treatment is associated with improved patient outcomes (stroke <5%, stroke/death <20%).^3–5 Most patients are asymptomatic at time of presentation; as such, computed tomographic angiography (CTA) has become widely adopted as the test of choice to screen at-risk patients for BCVI presence.^6,7 Early screening of at-risk patients facilitates rapid detection of asymptomatic lesions and initiation of medical therapy in appropriate candidates, which is associated with reduced stroke risk.^3–5

Our trauma centre is one of many hospitals throughout Aotearoa New Zealand and worldwide that have adopted a policy of identifying at-risk blunt trauma patients. Our initial protocol was successfully implemented a decade ago.⁸ Blunt trauma patients who are undergoing computed tomography (CT) scans of the head, cervical spine and/or torso are evaluated for presence of clinical and radiographic criteria associated with risk of BCVI; patients who meet these criteria undergo screening neck CTA. We recently sought to evaluate the efficacy of our protocol at identifying BCVI and preventing stroke and to determine the baseline incidence of BCVI in our trauma patient population.

Methods

We conducted a retrospective chart review of the New Zealand Trauma Registry (NTZR) inclusive of all adult (age >15) blunt trauma patients who presented initially to our trauma centre from 1 March 2023 to 31 August 2023. We identified all patients who had Abbreviated Injury Scale (AIS) codes consistent with injuries that met radiologic criteria to obtain screening neck CTA based on our current protocol, which included: acute infarction (brain stem infarction, cerebellar infarction due to traumatic vascular occlusion or cerebral infarction due to traumatic vascular occlusion), all diffuse axonal injuries, complex skull vault fractures, all skull base fractures, LeFort II and III fractures, mandible fractures, cervical spine fractures, cervical spine spinous ligament injuries and cervical spine facet subluxation/dislocation. We also included all patients with upper rib fractures (first, second or third rib), which have previously not been officially included in our protocol but often prompt our providers to request a CTA. Data collected for each patient included demographics, injury patterns and mechanisms, incidence and timing of CTA and outcomes including stroke occurrence and death.

Results

A total of 760 patients met inclusion criteria. We identified 95 patients who had AIS codes consistent with injuries requiring screening neck CTA. Of these, a total of 41 (43%) eventually underwent screening neck CTA, six of whom were positive for BCVI (Figure 1).

View Figure 1–3, Table 1–2.

CTA was obtained at the same time as the initial trauma CT in 26 patients and within 6 hours in an additional 10 patients, for a total of 36 patients (37.8%) in whom indicated screening CTA was conducted in a timely fashion. Four additional patients had screening CTAs conducted at 17 hours, 18 hours, 45 hours and 93 hours, respectively, and one patient had a CTA nearly a month after initial injury for other reasons, and a BCVI was incidentally noted and treated. Of the 41 patients who eventually underwent screening CTA, six had a definite or possible BCVI identified and treated; the overall rate of BCVI in patients who met screening criteria and underwent screening CTA was therefore 6/41 (14.6%). Of these, one had progression to posterior circulation stroke despite timely treatment, two had progression to death due to severe concomitant traumatic brain injury and three had no sequelae.

Fifty-four patients did not undergo screening CTA as indicated. Of these, one had a missed BCVI that progressed to symptomatic stroke at 63 hours, at which point CTA identified the injury. The remaining 53 patients who did not have a screening CTA did not have evidence in the available medical record of readmission for or treatment of subsequent stroke.

Rate of CTA performance was analysed by type of injury identified (Table 1). The most common types of injuries requiring screening CTA were upper rib fractures (46 patients) and skull base fractures (38 patients). Incidence of screening CTA varied from 88% for patients with complex skull fractures to only 25% for those with mandible fractures.

All patients who were known to have BCVI are listed in Table 2.

Discussion

BCVI is an uncommon but increasingly recognised injury pattern, and it requires prompt diagnosis and treatment to prevent serious complications including stroke and death. Progression from asymptomatic to symptomatic BCVI tends to happen early (most strokes occur between 12 and 72 hours, with a median time to stroke <48 hours).⁹ Strokes in the trauma patient population are particularly devastating as they result in higher years of life lost, higher quality years of life lost and increased cost of care compared with medical stroke patients.

Previously, the gold standard for diagnosing BCVI was with invasive and resource-intensive angiography. However, advancements in CT technology in recent years have allowed CTA to supplant angiography as the test of choice to diagnose BCVI. CTA is accurate, rapid and low risk for detection of BCVI.^6,7

As awareness of BCVI has increased, it has increasingly been identified to occur in association with a wide variety of injury patterns and mechanisms. Protocols have been developed to identify blunt trauma patients with injury mechanisms, injury patterns and clinical signs and symptoms who would benefit from CTA screening for BCVI. The so-called “extended Denver criteria” is currently the most widely studied and adopted.^10–12 This protocol utilises a standard set of clinical and radiographic findings to identify patients at risk for BCVI who would benefit from screening CTA (Figure 2).

Current trauma guidelines consistently endorse protocolised screening for BCVI,^7,12–14 such that it should be considered a standard of care. Increased awareness of and screening for BCVI, combined with adoption of high-quality CTA, has led to increased rates of injury as high as 7% in blunt trauma patients.¹⁵

Once identified on CTA, lesions can be graded on a standardised severity scale, such as the Biffl/Denver grading scale. While medical management should be based on clinical considerations, grading can facilitate prognostication and risk/benefit assessment regarding medical therapy, inform decision making regarding repeated imaging for surveillance and enable structured reporting for documentation, audit and research purposes.⁷

Early diagnosis allows for prompt medical therapy, which is associated with lower rates of stroke and death in asymptomatic non-haemorrhagic BCVI.^3–5 While there is no consensus regarding antiplatelet regimens vs anticoagulation, aspirin monotherapy is often recommended because it is a mainstay of preventing cerebrovascular events in similar non-traumatic conditions and has a low risk of clinically significant bleeding complications in most scenarios. Even patients with concomitant traumatic brain injury or solid organ injury can typically be safely managed with aspirin without increasing the risk of bleeding complications.^16,17

This audit assessed the efficacy of our current BCVI screening protocol by measuring completion of timely screening CTA (within 6 hours) and found that only 41% of patients who had injury patterns associated with BCVI had a timely CTA. Of the patients who underwent CTA, incidence of BCVI was 14%; extrapolating a similar incidence in those who met criteria but were not screened, it is possible at least eight additional BCVIs may have been missed during this study period. It is possible that the actual number of missed BCVIs might be lower if there is some selection bias leading to patients with injuries being more likely to be screened; however, we found no unifying pattern of injury associated with more likely BCVI occurrence. The most common injury type associated with BCVI was upper rib fractures, occurring in 46 of the 95 patients, but patients with these injuries had one of the lowest rates of timely CTA completion at only 28%. This is unsurprising, as until recently this criterion was not part of our formal screening protocol. Of the 46 patients with upper rib fractures, nine had other injuries that met criteria for screening CTA, and all of them received one. However, of the remaining 37 patients in which isolated rib fracture was the only screening criteria, only four (11%) had a timely screening CTA. Unfortunately, one patient who was not screened progressed to a possibly preventable stroke. Better protocol adherence and inclusion of expanded criteria (especially upper rib fractures) might have resulted in the capture of at least eight missed BCVIs and prevented one stroke during the study period.

We suspect there are several factors contributing to the suboptimal adherence rate to protocolised BCVI screening. Patients may have injuries that meet radiographic criteria but are subtle or otherwise not identified on initial review at the time initial CTs are conducted. Our analysis did not indicate any particular type of injury as more likely to be missed initially and cause CTA performance to be delayed. Logistical challenges inherent in organising an expeditious second trip to the CT scanner for a neck CTA once these patients have been moved to the operating room, intensive card unit or ward can be numerous, hence our finding of a subset of patients whose screening CTA was not completed until many hours after admission. Concerns regarding administration of a second contrast bolus to such patients can also lead providers to delay obtaining CTA. Additionally, patients may have initial trauma CT scans at other hospitals but then are transferred to a referring centre for definitive treatment once their injuries are identified; these patients would not undergo the same screening protocol in the emergency department (ED) CT scanner as patients who are initially evaluated in our ED, and as such are at risk of not being identified as high risk for BCVI and flagged to undergo screening through our usual process.

We have suggested a new screening protocol that is more widely encompassing of all clinical and radiographic patterns associated with BCVI (Figure 3). Our processes are being refined to improve protocol adherence, including widely publicising the new protocol throughout trauma patient care areas and holding education sessions for clinical staff who are involved in acute management of trauma patients, particularly in the radiology and EDs. We advise a goal of completing indicated BCVI screening within 6 hours of presentation.

Unfortunately, this difficulty with implementing screening protocols has been described by others and is one of the factors that has led some to advocate for universal screening of all trauma patients with concerns for significant head, face, neck and upper torso trauma. These patients are likely already undergoing CT scans with IV contrast, so the addition of a neck CTA is likely to achieve the highest rate of injury detection with minimal additional risk, and has previously been shown to be cost effective in some settings.^18–20 An additional widely cited drawback to screening protocols is that even the most robust screening protocols with 100% adherence still do not capture all BCVI patients. The extended Denver criteria has been found to miss at least 25–30% of patients who have BCVI.^15,18,19 Taking this into account, we estimate that the overall incidence of BCVI in our patient population may be as high as 2.4%, and that even with the most stringent criteria-based screening protocol we still might miss a handful of BCVI patients per year at our institution alone. We advocate that further analysis of local BCVI injury patterns should be conducted to determine whether universal screening might be the optimal management and cost-effective use of resources for our patient population.

Our study had broader implications for trauma patients across Aotearoa New Zealand. While a major trauma centre in Aotearoa New Zealand, our hospital sees just under 15% of all major trauma cases; 6/7 of all trauma patients have initial care in a wide array of other hospital settings throughout the country, many of which are not major trauma centres. Over 6 months at our hospital, 54 major trauma patients missed screening for BCVI and one had a missed stroke. If BCVI screening protocols are similarly under-utilised across Aotearoa New Zealand, there may be over 700 patients annually who do not undergo appropriate screening for BCVI, and an additional 15 patients with missed, and possibly preventable, strokes. While our hospital has taken steps to improve adherence to our BCVI screening protocol, we suggest there should be a national protocol for BCVI screening that all hospitals can adhere to and audit against.

This study has some limitations. It is a retrospective chart review and as such is limited by the availability and accuracy of the data in the medical record and trauma registry. The incidence of BCVI and stroke in the unscreened population is unknown, as many have no follow-up data in our medical record system, and may have sought follow-up care elsewhere. We are unable to know the true incidence of BCVI in our patient population and have attempted to estimate it based on the incidence in the portion of the population that has been screened. However, selection bias may mean that the true incidence rates between the two groups are not similar.

Conclusion

In recent years, universal screening for BCVI in all blunt trauma patients has been advocated by some, citing essentially zero missed injury rates, improved early diagnosis and treatment leading to decreased stroke and mortality rates, and even cost-effectiveness.^15,18,19 Given the limitations of protocolised identification of patients to be screened as discussed above, and the devastating sequelae of missed injury on patients, their whānau and our society, consideration should be given to integrating neck CTA into standard trauma CT protocols. Increased awareness of BCVI screening protocols and more systematic screening to facilitate early identification and treatment should be implemented across Aotearoa New Zealand.

An erratum has been published for this article.