Subclinical hypothyroidism (SCH) is thought to have an influence on stroke outcomes. However, few reports demonstrate a favorable relationship between the two. We evaluated this association in acute ischemic stroke. From Jan 2005 to June 2008, 756 acute ischemic stroke patients were recruited within seven days of onset. The patients with overt hypothyroidism/hyperthyroidism or other medical conditions that may affect thyroid function were excluded. Thyroid stimulating hormone (TSH) and free thyroxine (FT4) levels were measured within two days. Patients were divided into two groups: the SCH group (TSH > 5.0 microU/mL and normal FT4 levels) and the control group. Stroke outcomes were assessed using two different criteria. In the first outcome model, favorable outcomes [I] were simply defined by modified Rankin Scale (mRS) scores (<or= 1), while the favorable outcomes [II] were defined as follows: a) a mRS score of 0, if the baseline National Institute of Health Stroke Scale (NIHSS) scores were < 8, b) a mRS score of 0 or 1, if the NIHSS scores were 8-14, c) a mRS score 0-2, if the NIHSS scores were >14. The changes in mRS scores and the proportion of patients with favorable outcomes [I] or [II] at the 30(th) and 90(th) day were compared between the two patient groups. Of the 756 patients, 31 (4.1%) were patients with SCH. More patients from the SCH group showed improvement in NIHSS scores on the 30(th) day compared to the control group (48.4% vs. 25.3%, p=.006). In addition, the proportion of patients who exhibited favorable outcomes [I] was significantly higher in the SCH group on the 90(th) day (74.2% vs. 55.3%, p=.027) and that trend was seen as early as the 30(th) day (p=.102). Similarly, the proportion of the patients with favorable outcomes [II] was significantly greater in the SCH group both on the 30(th) (29.0% vs. 14.6%, p=.039) and 90(th) day (58.0% vs. 31.0%, p=.003). We found that acute ischemic stroke patients with SCH at admission were more likely to show favorable functional outcomes than those without SCH. We can suggest preconditioning before the stroke combined with a reduced response to stress as a possible protective mechanism.