Causes and mechanisms of territorial and nonterritorial cerebellar infarcts in 115 consecutive patients.

BACKGROUND AND PURPOSE

Territorial cerebellar infarcts have mainly a thromboembolic mechanism. Cerebellar infarcts less than 2 cm in diameter have recently been reported as nonterritorial infarcts, but it is not clear whether they are low-flow or embolic infarcts. The aim of the present study was to compare the characteristics and causes of territorial and nonterritorial infarcts in a prospective series of 115 patients.

METHODS

We collected data from 115 consecutive patients with cerebellar infarcts (79 territorial and 36 nonterritorial [ie, less than 2 cm]), using magnetic resonance imaging (88 patients) and computed tomography.

RESULTS

Patients with territorial infarcts and those with nonterritorial infarcts had similar vascular risk factors and clinical presentations and an equal frequency of cardiac source of embolism (32% versus 42%; P = NS) and of large artery occlusive disease (23% versus 19%; P = NS). Occlusive lesions of large arteries at angiography occurred at the level of one cerebellar artery (5% versus 0%; P = NS) and proximal to the ostia of the cerebellar arteries (18% versus 19%; P = NS). Infarcts distal to occlusive lesions were subdivided into unilateral vertebral artery occlusive disease (presumed artery-to-artery embolic mechanism; 18% versus 5%; P = NS) and low-flow state distal to bilateral vertebral or basilar artery occlusion (presumed hemodynamic mechanism; 0% versus 14%; P = .004). Patients with nonterritorial infarcts had more frequent hypercoagulable state (17% versus 1.25%; odds ratio, 15.6 [95% confidence interval, 1.8 to 135]). For the remaining patients, the mechanism of the infarct was unknown (34% versus 22%; P = NS).

CONCLUSIONS

Cerebellar infarcts less than 2 cm in diameter (ie, nonterritorial) have the same high rate of embolic mechanism as territorial infarcts (47% versus 49%; P = NS), have more frequent hypercoagulable state, and sometimes have a hemodynamic mechanism.