[A retrospective study of the practice of "look-back" procedures, on the incidence of HIV-1/2-positive blood donors and the risk of transfusion-associated HIV infection in public-community blood banks in Germany].

Objective: 'Look-back' investigations can reveal and confirm transfusion-transmitted infectious diseases and provide data for risk calculations of blood transfusions.

Design: In 1993 we distributed a questionnaire to all governmental and communal blood transfusion services in Germany. The questionnaire comprised questions about the methods, numbers and results of look-back investigations in case of HIV-1/2-positive blood donors with previous donations and in case of HIV-1/2-positive recipients of blood transfusions. The questionnaire was returned by almost all blood transfusion services (n = 75). One additional institution briefly informed us by telephone.

Setting: All governmental and communal blood transfusion services in Germany.

Patients: All recipients of blood or blood products in the years from 1985 till the end of 1992 who were treated in hospitals supplied by the transfusion services defined above.

Interventions: None.

Results: All blood transfusion services included have performed look-back studies since 1985. The methods used varied considerably. The interval of looking back mostly was sufficient. A main problem was the poor documentation in the medical records. The incidence of HIV-1/2-positive blood donations decreased from 11.6/100,000 in 1985 to 3.4/100,000 in 1992. Only 7 of 73 transfusion-transmitted HIV infections derived from transfusions after the introduction of HIV testing (October 1985). Since then the risk of transfusion-transmitted HIV infection can be calculated as 1/800,000 whole-blood donations of governmental and communal blood transfusion services.

Conclusions: Since the introduction of HIV testing the risk of transfusion-transmitted HIV infection in Germany has been very low, at a rather stable rate of 1/800,000. The data from the look-back studies confirm the previous estimations of the risk of transfusion-transmitted HIV infections, which was calculated by the HIV incidence in the donor population. Nevertheless there is a need for standardization of look-back investigations.