Vertebral Artery Injury in Cervical Trauma

Vertebral artery injury is one half of blunt cerebrovascular injury (BCVI) - blunt trauma to the carotid or vertebral arteries. It matters because it is a potentially preventable cause of posterior-circulation (vertebrobasilar) stroke: the diagnosis is made on imaging, not symptoms, and early antithrombotic therapy markedly reduces both stroke and death (EAST: stroke odds ratio 0.20, mortality odds ratio 0.17). Detection rises sharply when a screening protocol is used (odds ratio ~4.7) and is far higher among high-risk cervical injuries than low-risk ones (odds ratio ~12.7). The whole topic is therefore an argument for structured early screening of the right cervical-trauma patients.

The vertebral artery has four segments: V1 (pre-foraminal) from the subclavian to C6; V2 (foraminal) ascending through the transverse foramina from C6 to C2; V3 (extraspinal) from C2, looping around the atlas to pierce the dura; and V4 (intradural), joining its fellow to form the basilar artery. The V2 segment's bony, enclosed course through the transverse foramina, and the mobile V3 loop at the craniocervical junction, are what make the artery vulnerable in cervical trauma: a fracture that breaches the foramen transversarium, a facet dislocation/subluxation, a distraction injury, or an upper cervical fracture can stretch, compress or tear the vessel — causing intimal injury, dissection, thrombosis, pseudoaneurysm or occlusion.

The reason an isolated unilateral VAI often does not cause a stroke is collateral flow: the two vertebral arteries unite to form the single basilar artery, and the posterior communicating arteries link the posterior (vertebrobasilar) to the anterior (carotid) circulation through the Circle of Willis - so the contralateral VA and the carotid system can still perfuse the territory if one VA occludes. This tolerance is conditional, and the exceptions are what make an injury dangerous: when the injured vessel is the dominant vertebral artery (one VA - often the left - is usually larger, and the contralateral may be hypoplastic or end in PICA without reaching the basilar), when both vertebral arteries are injured, or when thrombus embolises distally into the basilar/PICA territory rather than simply occluding locally. This is why an isolated, non-dominant grade IV occlusion is often tolerated on aspirin, whereas a dominant-artery or bilateral injury - or any embolic propagation - is far more dangerous, and why antithrombotics (which prevent embolism) matter even when flow looks preserved.

A high-riding vertebral artery is one whose transverse foramen sits so far medial and cranial within the C2 lateral mass that the bone corridor for a pars or transarticular screw is too thin to take one safely. It is defined on the preoperative CT by an isthmus thickness under 5 mm or an internal height of the isthmus under 2 mm, and on a series of 100 consecutive cervical CTs it was present in 32% of patients - 13 left-sided, 9 right-sided and 10 bilateral. That is not a rarity to be aware of; it is a variant you should expect to meet in one patient in three, and it is frequently unrecognised because nobody measured.

Why other sources say 10 per cent, or 18 to 23 per cent, for the same variant. The definition above is the most inclusive one in use, so it returns the highest prevalence, and the figures are not in conflict - they measure different things. Mandel's cadaveric study of 205 C2 vertebrae applies a single 5 mm floor to isthmus height and width and puts about 10 per cent at risk. Paramore reconstructed CT along the planned screw trajectory and judged 18 to 23 per cent unsuitable on at least one side. The criteria used here add an internal height under 2 mm, which captures the shallow-but-wide isthmus the other two miss, and reach 32 per cent. A wider net, not a more dangerous population. Whenever you quote a number, say which criterion produced it - and see /topics/occipitocervical-fusion for how the choice changes the construct you plan.

The same study found a fully formed arcuate foramen (ponticulus posticus) in 14%, with a further 24% partially formed. This matters for a different screw: the ponticulus can be mistaken on exposure for a broad C1 posterior arch, and a C1 lateral mass screw started on what is actually the roof of the arcuate foramen enters the artery. The two anomalies coexist on the same side in about 5%.

The V3 segment leaves the C2 transverse foramen, runs upwards and laterally to the C1 transverse foramen, then turns medially across the superior surface of the posterior arch of C1 in a groove before piercing the dura. That horizontal course is what a subperiosteal exposure of the C1 arch runs into.

A cadaveric study measuring the landmarks directly found the distance from the posterior tubercle of C1 to the groove cradling the vertebral artery to be 19.1 ± 3.8 mm. Taking the lower end of that spread, staying within roughly 15 mm of the midline on the C1 arch is the practical rule - and the variability is the point: a mean of 19 mm with a standard deviation of nearly 4 mm means the artery is appreciably closer to the midline in some patients than the average suggests. The authors' other warning is the more important one: blind dissection is what injures the artery. Work subperiosteally, stay on bone, and do not sweep laterally into the soft tissue above the arch looking for a plane.

Because BCVI is frequently clinically silent initially, screening is based on injury patterns and risk factors (modified Denver/Memphis-type criteria). Screen for VAI/BCVI with:

• Cervical spine fractures — especially through the foramen transversarium, C1–C3 fractures, and facet (sub)luxation/dislocation or fracture-subluxation;
• Severe head injury (low GCS), basilar skull fracture, Le Fort II/III facial fractures;
• Cervical bruit/thrill, expanding neck haematoma, seatbelt sign over the neck, or near-hanging;
• Focal neurology unexplained by brain imaging, or a stroke/TIA pattern.

A pragmatic rule (and the EAST practice guideline position): screen any patient whose injuries would otherwise prompt a CT of the neck or chest — liberal screening catches the silent majority.

CT angiography (CTA) of the neck is the screening modality of choice (widely available, fast); digital subtraction angiography (DSA) remains the reference standard and is used in selected/equivocal cases or for intervention. Injuries are graded by the Biffl/Denver scale (I–V), which guides treatment and prognosis, and follow-up imaging is often used to track lesion evolution.

The goal is stroke prevention. Antithrombotic therapy — aspirin or therapeutic anticoagulation (heparin) — is the mainstay and has proven safety in trauma patients; treatment is guided by the grade of injury and started as early as the overall injury burden safely allows (given the early stroke window). For isolated vertebral artery injury the stroke rate is low and aspirin is generally effective (particularly for grade I and IV). Endovascular intervention (stent, coil/occlusion) is reserved for selected lesions — e.g. an enlarging pseudoaneurysm (grade III) or a grade V injury (the EAST guideline recommends against routine stenting as an adjunct to antithrombotics for grade II–III). The choice of agent and timing is individualised by balancing stroke risk against bleeding risk (concomitant TBI, solid-organ injury, planned surgery).