Background and purpose: Homocysteine (Hcy) is closely associated with stroke. Despite the fact that Hcy has consistently been shown to predict development of recurrent stroke, prior studies on the association of Hcy and stroke subtypes have been inconclusive. Methods: Data from the Ege Stroke Registry were examined and 5-year follow-up data were analyzed. Multivariate survival analyses were undertaken using Cox proportional hazards models to determine the prognostic value of Hcy in different ischemic stroke subtypes. Results: Of the 9522 patients with stroke, 307 (27%) with hyperhomocysteinemia (hHcy) had recurrent stroke. Univariate Cox regression model showed that hHcy group was associated with recurrent stroke (crude hazard ratio [HR] 1.16; 95% CI 1.02-1.30). But there was no such association in multivariate regression models (adjusted HR 1.11; 95% CI .97-1.26). hHcy was not associated with any ischemic stroke subtypes at 5 years. Univariate Cox regression model showed that hHcy group was associated with overall cardiovascular events (crude HR 1.44; 95% CI 1.32-1.57). However, this association no longer existed in multivariate regression models (adjusted HR 1.01; 95% CI .93-1.12). Higher plasma Hcy group was significantly associated with higher mortality compared with normal plasma Hcy group (OR 1.83; 95% CI .45-2.32). Conclusions: Our results showed that elevated Hcy is not associated independently with stroke recurrence and overall cardiovascular events in patients with ischemic stroke. There was no association between the hHcy and stroke recurrence in the stroke subtypes within 5 years.