Background and purpose: The relationship between neutrophil to lymphocyte ratio (NLR) and prognosis after acute ischemic stroke (AIS) remains controversial. The aim of this cohort study and systematic review was to ascertain the association of admission NLR with major clinical poor outcomes after AIS. Methods: We analyzed data from Chengdu stroke registry and performed a systematic review for previous literature. The outcomes were hemorrhagic transformation (HT), parenchymal hematoma (PH), symptomatic intracranial hemorrhage (sICH), 3-month death or disability (modified Rankin Scale >= 3), and 3-month death. Odds ratios (ORs) and 95% confidence intervals (CIs) of NLR as a continuous and categorical variable and poor outcomes were pooled separately. We also calculated the predictive accuracy of admission NLR in different outcomes. Results: We included 808 patients from registry database and 9563 patients from previous studies. Our registry data showed that NLR 5 was associated with HT (OR 2.03, 95%CI 1.19-3.46), PH (OR 2.54, 95%CI 1.20-5.35) and 3-month death (OR 5.55, 95%CI 1.41-21.89); meta-analysis with our data and other observational studies indicated that higher NLR was associated with HT (OR 1.99, 95% CI 1.45-2.73), sICH (OR 2.22, 95% CI 1.60-3.09), 3-month death or disability (OR 1.68, 95% CI 1.18-2.38), and 3-month death (OR 2.79, 95% CI 1.57, 4.94). NLR had the highest predictive accuracy for 3-month death. Conclusions: Higher NLR is positively associated with the risk of HT and 3-month death after stroke. Considering the limited predictive ability of a single biomarker, more studies should validate the role of NLR in prognostic models.