Uses and sources of data on long-term survival after blood transfusion

Several public policy decisions in transfusion medicine require information on the long-term (> or =10-year) survival of transfused patients. This information is needed (1) to estimate the number of surviving transfusion recipients who have contracted a particular infection through transfusion, (2) to assess the cost-effectiveness of measures introduced to further improve the safety of the allogeneic blood supply, (3) to estimate the total anticipated number of transfusion-transmitted cases of disease when a novel transfusion-transmitted infection with a long incubation period emerges, and (4) determine the scope of any proposed lookback investigation in terms of the length of time that should be covered retrospectively by the lookback effort. Although the probabilities of survival of Olmsted County, MN, residents transfused in 1981 were often used previously when input data on long-term posttransfusion survival were needed in the United States, these data most likely do not reflect the survival of patients transfused in the 1990s. Recent data from Sweden, Northern England, and New York City suggest that the short-term (up to 5 years posttransfusion) probabilities of survival reported from Olmsted County may have to be reduced by up to 20% before they can be used for making public policy decisions in the future, and that probabilities of survival of 66%, 60%, and 47%, respectively, at 1, 2, and 5 years posttransfusion may reflect the life expectancy of subjects transfused in the 1990s. No empirical data on the 10-year probability of survival of such patients are currently available from population-based studies, but some data suggest that the 10-year survival of an unselected population transfused in 1988 to 1990 may be 40%. A population-based study that includes several US counties has to be undertaken to generate the information needed for public policy decisions in the future.