Testing the validity of the lacunar hypothesis: the Northern Manhattan Stroke Study experience.

Backgrouncd/Objective Few studies have attempted to validate the “lacunar hypothesis.” The accuracy of identifying lacunar and other nonlacunar mechanisms of infarction will be increasingly important in evaluating potential stroke treatments. The aim of this study was to determine the value of lacunar syndromes in predicting radiologic lacunes and the value of clinicoradiologic lacunes in predicting “lacunar infarction” as final stroke mechanism.

Methods From 1990 to 1994, 591 patients with cerebral infarction, who were from northern Manhattan and over the age of 39, were prospectively examined. Data were collected on the admitting clinical syndrome (lacunar or nonlacunar) and brain imaging findings. Lacunar syndromes were categorized as pure motor hemiparesis (PMH), pure sensory syndrome (PSS), sensorimotor syndrome (SMS), ataxic-hemiparesis (A-H), and other lacunar syndromes. Brain imaging findings were classified as radiologic lacune or nonlacune. Positive predictive values, sensitivities, and specificities of lacunar syndromes for identifying radiologic lacunes were calculated. The final mechanism of infarction was determined after review of all the diagnostic tests and compared among the lacunar groups.

Results Lacunar syndromes occurred in 225 cases. PMH was the most common lacunar syndrome, accounting for 45%, SMS 20%, A-H 18%, and PSS 7%. Lacunar syndromes had an overall positive predictive value (PPV) of 87% for detecting radiologic lacune: PSS loo%, A-H 95%, SMS 87%, and PMH 79%. Among the 195 patients who presented with a lacunar syndrome and had this condition confirmed radiologically, 147 were classified as having a final diagnosis of lacunar mechanism of infarction (PPV = 75%). Atherosclerosis accounted for 17 (9%), cardioembolism 10 (5%), cryptogenic 17 (9%), and other unusual causes 4 (2%).