In 2000, the American Red Cross (ARC) received reports of unusual transfusion reactions of unknown aetiology among patients receiving leucocyte-reduced (LR) red blood cell (RBC) units in multiple distribution regions. We evaluated potential risk factors of reactions among patients who received LR-RBC transfusions. A case-patient was defined as any patient with onset of back pain while receiving an LR-RBC transfusion from 1 January to 25 May 2000. Controls were chosen randomly and selected in a 1:3 case : control ratio from healthcare facilities in which case-patients were transfused. Product-specific risk factors of reactions were further determined through nested case-control study, procedural review of blood collection facility and quality-control-testing record review of product processing. Reaction incidence rates were determined through ARC blood product distribution data by region of blood collection and processing. There were 29 reactions detected in patients who received transfusions in 13 healthcare facilities in five states. Eighteen case-patients and 78 controls were included in the case-control study. In univariate analysis, case-patients were more likely than controls to have a haematologic malignancy, to have received the transfusion as an outpatient, to have received an RBC transfusion within the previous 3 months, to have received medication used to prevent reactions or to diminish their intensity upon transfusion (i.e. premedication) or to have received LR-RBC units prepared with the HemaSure r\LS System(HS) rather than two other filters used. In multivariate analysis limited to recipients of HS-filtered RBC units, transfusion premedication [adjusted odds ratio (AOR) = 7; 95% confidence interval (CI) 1.4-37; P = 0.02] and transfusion as an outpatient (AOR = 5; 95% CI 1.1-20; P = 0.03) were independently associated with reactions. The rate of reported transfusion reactions was 2.0 reactions per 10 000 RBC units distributed. A multistate cluster of transfusion reactions was significantly associated with leucocyte filtration of RBC units prepared with a specific product, the HS filter. The reactions also were independently associated with premedication and transfusion as an outpatient; these may be surrogates for an increased risk of reaction or for greater likelihood of detection. The mechanism for these reactions has not been elucidated. This cluster of reactions underscores the importance of surveillance efforts to detect adverse events after transfusion, particularly when new methods to modify blood products are introduced.