The clinical relevance of high P-selectin levels in venous thrombosis is unknown. We prospectively followed 544 patients with first unprovoked venous thromboembolism (VTE) after cessation of anticoagulation and evaluated P-selectin as a risk factor of recurrent VTE. VTE recurred in 63 (12%) patients. Patients with recurrence had significantly higher P-selectin levels than those without (45.8 mg/dl +/- 16.4 vs. 40.1 mg/dl +/- 13.3; p = 0.006). After four years, the probability of recurrence was 20.6% (95% confidence interval [CI] 12.6-28.5) among patients with P-selectin values above the 75(th) percentile of the patient population and was 10.8% (95% CI 7.2-14.3) among patients with lower values (p = 0.046). Compared to patients with low P-selectin, adjusted risk of recurrence was 1.7-fold (95% CI 1.0-2.9, p = 0.045) increased among patients with P-selectin levels exceeding the 75(th) percentile. We conclude that high circulating P-selectin is a risk factor of recurrent VTE.