BackgroundRegulatory T cells (Treg) have been identified as a key modulator of neuroinflammation in stroke. However, little is known about the association of Treg subpopulations with clinical outcome in patients with acute ischemic stroke (AIS). MethodsPatients within 1 week from stroke onset were prospectively enrolled in this study. Healthy controls were sex-and age-matched 1:1 to AIS patients. The frequencies of Treg and Treg subsets were analyzed by flow cytometry and compared with nonstroke control. Univariate and multivariate logistic regression analysis was performed to investigate the prognostic value of Treg subsets in stroke outcomes. ResultsA total of 328 patients and 328 controls were included in the study. Compared with controls, patients with AIS had higher levels of Treg frequency and memory Treg (mTreg) frequency, but lower levels of naive Treg (nTreg) frequency and nTreg/mTreg ratio. One hundred twenty-six (38.4%) patients experienced unfavorable outcome (modified Rankin score 2-6). Multivariate regression analysis showed that nTreg/mTreg ratio was negatively associated with unfavorable 90-day outcome (the highest tertile versus the lowest tertile: odds ratio 0.13, 95% confidential interval [CI] 0.05-0.35). The risk estimation of unfavorable 90 day outcome can be significantly improved by adding nTreg/mTreg ratio to the conventional clinical parameters (continuous net reclassification improvement 91.26, 95% CI 69.04-113.5%, p < 0.001; integrated discrimination improvement 22.38, 95% CI 17.16-27.59%, p < 0.001). ConclusionThis study showed that patients with AIS had elevated Treg frequency and mTreg frequency, but reduced nTreg frequency and nTreg/mTreg ratio. Admission nTreg/mTreg ratio was an independent predictor of unfavorable 90 day outcome in AIS. However, large sample-size cohort studies are needed to confirm our findings.