BACKGROUND: Observational studies have reported that patients transfused with red blood cells (RBCs) have a worse clinical outcome than untransfused patients and that storage age of RBCs at the time of transfusion may be an independent predictor of this adverse clinical outcome. STUDY DESIGN AND METHODS: Eight RBC concentrates in additive solution were studied over an 8-week storage period. The RBC supernatant was ultracentrifuged to concentrate microvesicles (RCMVs). RCMVs were studied by flow cytometry to identify phosphatidylserine (PS)-expressing RCMVs and in a thromboelastograph (TEG) using a modified assay to detect a thrombin facilitation effect. RESULTS: For all products, the percentage of RCMVs that exhibited PS expression on Day 1 was 50 +/- 13%, which increased with storage, and on Day 31 was 90 +/- 4%. After 31 days, four of the eight products showed a thrombin facilitation effect as evidenced by a shortening of the TEG reaction (R) time of 1.3 +/- 1.1 minutes, which persisted to Day 41. Data are the mean +/- 1 SD. This TEG R shortening effect was neutralized by annexin V. No such effect was observed on, or before, Day 21. CONCLUSIONS: Some stored RBCs release RCMVs, which express PS and are capable of facilitating thrombin generation in vitro. This provides a possible mechanism by which stored RBCs could promote adverse thrombotic or inflammatory effects.