Transoesophageal echocardiography in patients without arterial and major cardiac sources of embolism: difference between stroke subtypes.

We studied the records of 175 consecutive patients referred to our neurologic ward between January 1994 and February 2000 with a diagnosis of ischaemic cerebrovascular disease (ICVD) (stroke or transient ischaemic attack - TIA) who underwent transoesophageal echocardiography (TEE). We excluded patients with large vessel disease, high-risk embolic cardiopathies and other rare causes of stroke. According to clinical and neuroimaging findings, patients were divided into two groups. The lacunar (LAC) group (69/175 (39.4\%)) and the nonlacunar (N-LAC) one (106/175 (60.6\%)). The control population consisted of 78 consecutive patients, referred to the echocardiography laboratory for TEE without history of ICVD and known heart disorders. Patent foramen ovale (PFO) frequency was significantly higher in case patients than in control subjects (55/175 (31.4\%) vs. 13/78 (16.6\%); p = 0.02). Among case patients, PFO was more prevalent in the N-LAC group than in the LAC one (43/106 (40.6\%) vs. 12/69 (17.4\%); p = 0.0005). A large degree of shunt occurred in 53.5\% of N-LAC patients and in 16.7\% of LAC ones (p = 0.04). Atrial septal aneurysm (ASA) was detected in 12\% of case patients and 1.3\% of control subjects (p = 0.003) and was more frequent in the N-LAC group than in the LAC one (16 vs. 5.8\%; p = 0.05). Mitral prolapse (MP) was present in 6/175 (3.4\%) ICVD patients (vs. 1/78 among controls) in most cases associated with myxomatous valve redundancy. Aortic arch atheromas (AA) were detected in 12\% of ICVD patients and in 10.2\% of controls. The frequency was 9.4\% in N-LAC and 15.9 in LAC. No complicated AA (plaque thickness >4 mm, ulcerated atheroma, superimposed thrombus) were detected. After multivariate analysis, PFO (OR = 3.8; 95\% CI = 2.7-7.9) and ASA (OR = 8.01; 95\% CI = 3.0-16.1) appeared to be independent predictors of ICVD. PFO (OR = 2.24; 95\% CI = 1.24-4.92) was also independently associated with N-LAC stroke subtype and its importance was even higher in younger patients. Our study provides further evidence that TEE is a helpful diagnostic tool in stroke patients without arterial and major cardiac sources of embolism. However, its utility differs according to type and localization of the ischaemic lesion being more relevant in patient with N-LAC infarctions.