Posterior Versus Anterior Circulation Infarction: How Different Are the Neurological Deficits?

Cerebrovascular disease, with its complex anatomy, and ischemia in different parts of the brain can produce the same neurological deficits. Because important decisions, such as revascularization, are often based on the association between symptoms and stenosis in a specific vascular territory, the clinician must be confident in the localization. Traditionally, the Oxfordshire Community Stroke Project classification used a simple clinical scheme with a high correspondence to radiological findings to distinguish posterior circulation infarction (PCI) from anterior circulation infarction (ACI).^1–4 The presence of classic brain stem and cerebellar symptoms was used to diagnose PCI. However, this approach may not accurately localize all ischemic strokes. A substantial proportion of PCIs may not be accurately classified only by symptoms/signs, because they lack typical clinical features. Misdiagnosis usually occurs in the initial phase of patient evaluation, which may lead to erroneous clinical decision-making.

In some studies, PCI simulated involvement of the anterior circulation.^5–8 Recent evidence from magnetic resonance imaging (MRI)-based studies demonstrated that unilateral limb weakness, unilateral limb numbness, nausea/vomiting, and headache are the major clinical neurological deficits in PCI rather than crossed paralysis, crossed sensory deficits, visual field disturbance, isolated vertigo, and dysphagia.^9–11 This suggests that the clinical manifestations of PCI and ACI are more alike than dissimilar. A comparison of the symptoms and signs between the 2 main stroke localizations within a large series of acute stroke patients has not been performed. We compared the distribution of clinical manifestations between PCI and ACI with lesions confirmed by MRI. In addition, we tried to determine the diagnostic value of specific symptoms/signs for PCI.

Given the limited information deriving from small case series with highly selected PCI patients, as well as the low performance rate of MRI in these studies, there have not been reliable clinical criteria for PCI. In the current study, we initially compared the frequency of neurological deficits between the 2 groups and found a similar distribution among the most common symptoms and signs between PCI and ACI. Second, we assessed the ability of a single neurological deficit to discriminate PCI from ACI, and we observed that despite some symptoms and signs having approximately 100% specificity for diagnosing PCI, symptoms/signs with a higher predictive value had a very low prevalence. These findings emphasize the fact that the clinical manifestations between PCI and ACI have a high degree of similarity.

In this study, we observed that the most common neurological deficits were homolateral paralysis, central facial/lingual palsy, and hemisensory deficits in PCI; and also, that these were the 3 most common in ACI. Previous studies, the New England Medical Center Posterior Circulation Registry, and a Qatar study presented within-study comparisons of the frequency of the neurological deficits in PCI cohorts; but, these studies did not compare PCI with ACI patients.^11,18 The frequency appeared to vary in some clinical manifestations in these previous studies, as compared with that in our study; however, homolateral paralysis and facial palsy were also found previously to be the most common signs in PCI patients. The divergent frequencies may partly caused by differences in the severity of vertebrobasilar ischemia among stroke populations¹¹ and the different definitions studies used when classifying the neurological deficits. In a series of articles from the New England Medical Center Posterior Circulation Registry, Caplan et al reported that unilateral limb weakness (38%) was the most frequent sign in the whole PCI cohort, hemiparesis and tetraparesis (62%) were the most common signs for patients with midbrain infarction, and hemiparesis was present in more than half of patients (50.6%), along with basilar artery occlusive disease.^11,19,20 Furthermore, in a study that analyzed the clinical manifestations of pontine infarction, pure motor hemiparesis or hemiplegia was found in a large proportion of patients instead of classical crossed syndromes, and cannot be distinguished from the internal capsule-corona radiata region infarctions by clinical signs alone.²¹ These findings are consistent with our view that PCI patients with homolateral paralysis are common in clinical practice.

Besides the nonspecificity of most clinical manifestations in PCI versus ACI, we should also emphasize that there are still certain symptoms/signs favoring a diagnosis of PCI, such as crossed motor/sensory deficits, oculomotor nerve palsy, visual field deficits, and Horner's syndrome, which are highly suggestive. Another point to consider is that these signs occurred infrequently (<10%); hence, it is difficult to identify most PCI patients based on these relatively rare clinical manifestations. One previous study with the observation that approximately 10% to 20% of patients with a diagnosis of presumed ACI actually have a PCI supports our contention.⁸ The misdiagnosis rate in the previous study might be underestimated because of the small number of patients included and the low sensitivity of computed tomography in detecting multiple infarctions involving the posterior fossa.

According to the literature, crossed signs such as ipsilateral motor and sensory cranial nerve signs or symptoms plus contralateral hemiplegia or hemianaesthesia are specific symptoms that point to brain stem involvement.²² However, in the present cohorts, crossed motor deficits occurred infrequently (13 patients; 1.1%), with only 1 patient diagnosed with ACI having MRI-confirmed bilateral middle cerebral artery infarction that happened suddenly after cholecystectomy. Crossed sensory deficits occurred in only 9 PCI patients (3.0%) and in 0 patients with ACI. The crossed sensory deficits correlated well with lesion locations involving the medulla, whereas crossed motor deficits were associated with lesions in medulla (n=8), pons (n=2), and midbrain (n=2). A study focusing on pure lateral medullary infarction reported that the rate of the crossed sensory deficits was high at approximately 26%,²³ but the correlation between infarct topography and the crossed motor/sensory signs in PCI patients were not well-described and need additional investigation.

Contrary to current belief, our results do not support the conventional wisdom that holds that disturbed consciousness is highly specific for PCI. Since the initial description of the clinical findings in a small group of select patients with brain stem infarction by Kubik and Adams,²⁴ disturbances of consciousness were considered to be an important feature of PCI. This was followed by studies of patients with specific vertebrobasilar lesions, including the midbrain, cerebellum, and basilar artery occlusive disease separately; the rate of disturbed consciousness was high, frequently >20%.^20,25–27 In the most extensive of these studies, Carol et al reported 20 patients with basilar artery occlusion and found 19 of them had various impairment of consciousness.²⁷ In contrast, Sato et al reported a much lower rate of disturbed consciousness in patients with PCI than with ACI (18% versus 41%; P<0.001); this was in accordance with our cohorts that disturbed consciousness accompanied with a corresponding lower Glasgow Coma Scale score occurred in 10.3% of PCI, which was less prevalent than that in ACI (18.6%).²⁸ Similarly, 2 recent studies showed that the prevalence of altered level of consciousness in PCI patients was much lower (<10% and 18%, respectively) than was suggested by previous studies.^11,18 These controversies might be caused by a different distribution of stroke subtypes or different age groups. At the least, the present study revealed that additional investigations are required to determine whether consciousness disturbance should be considered diagnostic for PCI.

In the present study, nausea/vomiting, vertigo, and headaches were found to be the symptoms that occur more frequently in patients with PCI. Because circulation to the inner ear arises from the vertebrobasilar system, vertebral-artery disease can cause vertigo that is usually accompanied by other brain stem or cerebellar symptoms.²⁹ A previous study emphasized that isolated vertigo is rarely attributable to vascular events and it was in accordance with our findings that only 4 patients with PCI had isolated vertigo.³⁰

We are aware of the limitations of the present study. First, to elaborate the clinical features of pure PCI and ACI, we excluded patients with multiple lesions involving both posterior circulation and anterior circulation without MRI confirmation; this may have resulted in a selection bias. Second, the delineation of clinical symptoms and signs was completely dependent on chart review, and thus, the authors were unable to confirm rare manifestations. Nonetheless, most of the medical charts within the present study contained high-quality data. Third, as a hospital-based study, subjects who experienced more severe or minor stroke were not admitted to the hospital, although this hospital is the largest one in the region. Given the large sample and statistical significance of the observed effects in the present study, these limitations are unlikely to affect the results.

In this observational study, we found that there was no apparent difference in the frequency of the most common symptoms/signs between PCI and ACI. Some neurological deficits were highly specific for diagnosing PCI, but their sensitivity suggests that symptoms or signs considered typical of PCI are uncommon. Inaccurate localization would be common if clinicians only relied on clinical symptoms/signs to differentiate PCI from ACI, and this observation suggests that neuroimaging findings, particularly MRI, are vital for accurate localization.

The authors would like to acknowledge their gratitude to Dr Matthew S. Siket and Dr Lorenzano Svetlana for their valuable comments on this manuscript.

This work was funded by Science and Technology Department of Sichuan Province of China (Project No. 2011SZ0202).