You might find this article form John Judd to be of interest: Judd, W.J., et al, The Evaluation of a Positive Direct Antiglobulin Test in Pretransfusion Testing, Transfusion 1980; 20:17-23 In this University of Michigan study, an analysis was performed of 879 samples with positive direct antiglobulin tests. Eluates were performed and 83 were reactive. Most were autoantibodies and a few contained penicillin/Keflin antibodies and a few contained passively acquired anti-A. In only 11 of the 879 cases allo-antibodies were detected in the eluate. After 14 days allo-antibodies detected in the eluate were also detected in the plasma in all but one patient sample that eluted anti-K 17 days post transfusion. The article states, "One of the six patients whose red blood cells eluted a transfusion-induced alloantibody, but in whom the eluted antibody was not detected in the serum by routine pretransfusion screening tests, had been transfused 17 days before a detailed serological evaluation of the DAT was performed (case 4, Table 3). The red blood cells from this patient eluted anti-Kell. This isolated instance does not warrant an extension of our definition of “recently transfused” to a post transfusion interval beyond 14 days, for to do so would only increase the number of samples requiring evaluation with very little corresponding gain in terms of significant serological findings." The University of Michigan established a 14-day cut-off for "recently transfused" when determining if an eluate needs to be performed. We have chosen 28 days as our "recent transfusion" cutoff to perform an eluate. If the DAT becomes positive within 28 days we will perform eluate, if greater than 28 days since transfusion we will not perform eluate. The article was written in 1980 and the automated testing methods for antibody detection widely in use today are likely to be more sensitive than those used in the study.