T he advent of transesophageal echocardiography (TEE) has greatly improved the ability to detect cardiac sources of embolism. 1 This technique allows an easy and accurate evaluation of anatomic structure and function of the left atria1 appendage (LAA).2-4 Although it is well established that there is an association of LAA thrombi and left atria1 spontaneous echo contrast with a history of peripheral embolism,2 only recently has a study focused on LAA Doppler flow signals.3 Kortz et al3 reported a quadriphasic pattern of LAA flow in normal subjects without overlapptig of tachycardia-related waves: a diastolic forward flow just after mitral valve opening is followed by a diastolic backward flow due to LAA recoil; subsequently, a forward and a backward flow wave respectivCly due to LAA contraction and relaxation can be detected. . . . We report our observations on a new LAA flow pattern characterized by the presence of an additional systolic forward flow wave after LAA relaxation. After the first occasional observation of a systolic LAA flow wave, a study was undertaken to characterize this new tiding. TEE was prospectively performed in 62 consecutive patients with sinus rhythm. Reasons for the examinations were determination of source of embolism (62%), assessment of suspected endocarditis (lo%), detection of aortic and mitral valve disease (18%), and evaluation of valve prosthesis (10%). Ten patients were excluded from the sttidy because of inadequate representation of the LAA. The study group consisted of 12 patients (mean age 51 f 16 years) in whom the presence of LAA systolic forward flow was observed. Anesthesia of the hypopharynx was p&formed with 10% lidocaine spray. For sedation, patients were given a mean intravenous dose of 2 mg of midazolam. TEE was performed using a Sonos 1500 (Hewlett-Packard, Andover, Massachusetts), a Sonolayer SSH 140-A (Toshiba, Tokyo, Japan), or a Domier (Deutsche Aerospace, Munich, Germany) ultrasound system equipped with 5 MHz multiplane phased-array transducers. The LAA was mainly visualized in the longitudinal view and the probe adjusted to maximize its dimensions. Flow velocities were obtained by positioning the sample volume inside the left atria1 appendage at the point that offered the best alignment with its flow and avoided the noise signal due to wall motion. Furthermore, color Mand B-mode of the left atrial appendage flow were recorded. The presence of systolic forward flow after LAA relaxation was assumed when there were concomitant pulsed Doppler signal and color Band M-mode findings. Pulmonary venous velocity recordings were obtained with