BACKGROUND: Endovascularly retrieved clots are a potential resource for diagnosing stroke etiology, which may influence secondary prevention treatment. In this study we measured thrombin activity eluted by serially washing clots. METHODS: Clots were retrieved from 68 patients with acute ischemic stroke, freshly frozen and classified by standard criteria into proven atrial fibrillation (AF, 18 patients), atherosclerotic origin (AS, 15 patients), cryptogenic stroke (Cr, 17 patients) and other known causes (18 patients). Thawed clot samples were washed by transferring them into 1 mL buffer in seven hourly cycles and a fluorescent substrate assay was used to measure secreted thrombin activity. The clots were also examined histologically. Artificial fibrin and red blood cell-rich clots were similarly assayed for wash-eluted thrombin activity as an external control. RESULTS: Thrombin activity eluted from clots of AF origin decreased significantly with time in contrast to steady levels eluted from AS origin thrombi (P<0.0001 by repeated measures ANOVA). The Cr stroke group was indistinguishable from the AF group and differed statistically from the AS group (P=0.017 by repeated measures ANOVA). In artificial clots we found a biphasic activity pattern, with initially decreasing levels of eluted thrombin (AF pattern) and then, with continuing washes, steady eluted thrombin levels (AS pattern). CONCLUSIONS: An assay measuring the change in thrombin in clots retrieved during acute stroke endovascular thrombectomy procedures may serve as a diagnostic marker of the origin of the clot. The suggested mechanism for these differences may be the clot location before its retrieval, with high blood flow causing thrombin washout in atherosclerotic clots, in contrast to atrium appendage low blood flow retaining high thrombin levels.