There is debate in the literature about the frequency and importance of delayed transfusion reactions. This uncertainty could reflect the endpoints used (clinical or serological) and the type of study (typically retrospective or case series). In this report we describe a prospective investigation to determine the frequency of alloimmunization post transfusion and whether the alloantibody production is a laboratory event or has clinical relevance. A total of 2490 patients were transfused 11,218 red cell concentrates. One or more blood samples were collected within 7 d post transfusion and screened for serological evidence of alloimmunization. If any antibody was detected the patient's post-transfusion sample was screened for biochemical evidence of haemolysis and the patient's chart reviewed for documentation of clinical signs of a transfusion reaction. Post transfusion alloimmunization occurred in 2.6% of the patients (95% CI 2.1-3.6%), who had no detectable alloantibody in pre-transfusion testing. For those 86 patients (3.5%) with alloantibodies detectable pretransfusion, 8.9% (95% CI 3.6-17.4%) developed additional aloantibodies. The most common alloantibodies detected were anti-Jka, anti-E and anti-K. Despite the high frequency of serological evidence of delayed transfusion reactions, only one patient (0.05%) had clinical evidence of a delayed haemolytic transfusion reaction (95% CI 0.0-0.27%). Serological evidence of a delayed transfusion reaction is common; however, these reactions rarely cause clinical symptoms.