Background: Low IGF-I levels may be associated with the development of stroke; however, prospective data appear to be unavailable. Methods: This was a nested case-control study within a Danish follow-up study, including 57,053 men and women. Baseline data included circulating IGF-I, IGF-II, and IGF binding protein (IGFBP)-3 concentrations as well as lifestyle factors and medical history. We identified 254 cases with incident ischemic stroke and 254 gender- and age-matched controls. Results: Participants in the bottom quartiles of IGF-I and IGFBP-3 levels (median concentrations, 72 and 2937 ng/ml, respectively) were at increased risk of ischemic stroke, e.g. adjusted odds ratios (ORs) of 2.06 95% confidence interval (CI), 1.05–4.03 and 2.29 (95% CI, 1.17–4.49), respectively, when compared with participants in the top quartiles (median concentrations, 125 and 4835 ng/ml, respectively). A negative, although weaker, association was also found for IGF-II (adjusted OR 1.44, 95% CI 0.79–2.64) when comparing the bottom quartile with the top quartile. No substantial associations were seen for IGF-I and IGF-II when also adjusting for IGFBP-3; adjusting IGFBP-3 for IGF-I and -II had only a minor impact on the risk estimates. Conclusion: These findings give some support to the hypothesis that the IGF axis is involved in the pathogenesis of ischemic stroke.