VA occlusions do not always cause ischemic symptoms immediately after the trauma. Many reported cases of posttraumatic VA occlusion have demonstrated asymptomatic interludes of a few hours between the trauma and the onset of neurologic symptoms17,24). It has been suggested that intimal tears can lead to the aggregation of platelets on the exposed subintimal tissue and subsequent thrombosis, clot propagation and/or embolization to the cerebellum11,17,18). It is likely that our case went through same process after traumatic VA dissection, causing progressive thromboembolic event. VA dissection usually arises from an intimal tear, which allows blood under arterial pressure to enter between the layers of the wall of the artery, forming an intramural hematoma, the so-called false lumen7). The interludes are likely due to the slow progression of thrombosis or the gradual swelling of the vessel18,24). Woolsey and Chung23) described an autopsy finding from a patient with a left VA occlusion that was followed by fatal basilar artery occlusion. The autopsy revealed an organized thrombus in the left VA and a fresh thrombus that occluded the entire basilar artery23). The fresh thrombus in the basilar artery, which caused the brain stem infarction, seems to have gradually progressed from the left VA occlusion site after the initial insult.

We believe that the interludes in these cases represent the time period of thrombus propagation after which the infarction develops. Although some cases with longer interludes of up to several weeks or months have been reported, these cases are predominately cases of either unilateral or incomplete occlusion1,20). In many cases of bilateral VA occlusion, the interlude between the injury and the onset of ischemic symptoms occurs within hours3,19,24).

The 5-day interlude observed in our case was exceptionally long and represents an extremely rare occurrence in cases of bilateral posttraumatic VA occlusion.

Fink et al.5) reported no significant difference in the incidence of stroke between patients with occlusive VA injuries and non-occlusive injuries. We believe this finding suggests that VA occlusion itself is not the sole factor that determines vertebrobasilar insufficiency or infarction. The size of the contralateral VA, the completeness of the circle of Willis, and the robustness of the native pial collateral seem to be important factors affecting the development of infarction after VA occlusion5).

Six et al.21) reported a case of asymptomatic bilateral posttraumatic VA occlusion. Although both VAs were occluded in this case, angiography demonstrated reconstitution by the intramuscular collateral vessels of the thyrocervical trunk and the collaterals from the superficial occipital artery. Collateral circulation may derive from the anterior circulation via the posterior communicating artery or from anastomoses via the branches of the ECA, the thyrocervical trunk, the costocervical trunk or the muscular and spinal branches of more proximal VAs4).

We believe that these abundant collateral channels resulted in the delayed symptom onset observed in our case. Our patient showed collateral flows from the bilateral ECA via the muscular branch of the occipital artery and the ascending cervical artery on cerebral angiography (Fig. 5). Although our patient's collateral blood supplies were able to meet posterior circulation demands to some extent, these demands could not be met over the course of several days due to bilateral VA occlusion.

Aspects of patients' premorbid state, such as atherosclerosis and VA dominance, should also be considered4). Bilateral VA occlusion secondary to atherosclerosis is usually well tolerated, as this atherosclerosis is a slow process that allows for the development of collateral circulation18). Although our patient had diabetes mellitus, there was no evidence of cerebrovascular problems.

Our patient had infarctions in the left occipital lobe and the left cerebellar hemisphere. We suggest that the infarctions in the cerebellar hemisphere and the occipital lobe were mediated by different mechanisms. Our patient's left PICA was completely occluded, which explains why the infarction involved only the left hemisphere; whereas both PCAs were patent on cerebral angiography.

The cerebellar infarction due to PICA occlusion was likely the result of the direct progression of thrombus from the VA injury site, which blocked the blood supply to the left cerebellar hemisphere. However, the left occipital lobe infarction was likely the result of an embolus, rather than thrombus formation. Emboli in the basilar territory do not commonly cause extensive brain stem infarction because an embolus of sufficient size to traverse the VA would be unlikely to occlude the larger diameter basilar artery23). Such embolic fragments usually lodge in one or both PCAs23). We propose that the left occipital lobe lesion was likely due to an embolus from the VA occlusion site because there was no infarction in the brain stem.

Screening for VA injury is warranted for certain high-risk patients at the time of presentation due to the potentials for both acute and delayed development of neurological sequelae12,15,17). However, VA injury is not always accompanied by bone or spinal cord injury14). Neck manipulations by chiropractors and physiotherapists have also been reported to be associated with VA injury without cervical fracture4,20). The mechanism responsible for VA injury has been suggested to be stretching, tearing of the intima and media, or compression of the vessels18,20). Motor vehicle accidents are the leading cause of VA injury due to the major rotational component4,7). Rotational forces applied to the mobile portions of the cervicocephalic arteries are responsible for intimal tears20). Although bilateral VA dissection or occlusion is rare and found in only 0.1% of the patients with blunt trauma with angiography, its mechanism is likely to be same as unilateral one13). Galtés et al.7) described an autopsy finding from a patient with bilateral VA dissection after rotational neck injury, which is similar to unilateral one. Medhkour and Chan13) also mentioned chiropractic manipulation, motor vehicle crashes, or falls as the cause of bilateral VA injury. Chung et al.2) stated that VAs that are located on the lateral aspect of the cervical spine are more susceptible to rotational injuries, whereas the spinal cord is more likely to be injured by sagittal plane insults such as compression and hyperflexion.

The presence of cervical spine injuries requiring operative stabilizations has not been found to be an independent risk factor for VA injury12). In the current study, the severity of the spinal cord injury, as indicated by the neurologic deficits, was not associated with VA injury2). Because the severity of injury to the bone and/or spinal cord is not linearly correlated with the severity of VA injury, it is difficult to predict or screen for VA injury at the initial examination.

There are some controversies surrounding screening for VA injuries in patients with cervical trauma10,15), that stem from the uncertainties regarding improved outcomes and the benefits of treatment15). The guidelines of the American Association of Neurological Surgeons state that "there is insufficient evidence to recommend anticoagulation in asymptomatic patients", and the risk of adverse effects of anticoagulation therapy may be as high as 14%12). Although clinicians support either anticoagulation or antiplatelet therapy in symptomatic patients, the medical evidence surrounding this issue is inconclusive9). At present, there seems to be no consensus about the principles of treatment for VA injury, with the exception of individualization based on the specifics of the patient's VA injury, associated injuries, and risk of bleeding9).