For a long time anti-thymocyte globulins (ATGs) have been known to contain HLA-specific antibodies. We have measured high amounts of HLA Class I antibodies in both ATG preparations, whereas significant HLA Class II antibodies were only found in thymoglobulin (1). Focosi and Boggi (2) speculated that antibodies in ATGs specific for HLA Class I and II antigens might include donor-specific antibodies (DSA) that could be involved in the development of antibody-mediated allograft rejection. However, we presume that most HLA-reactive antibodies contained in ATGs will bind to epitopes shared between different HLA molecules. In our recent study, we have used cells expressing antigens known to be bound by ATGs to deplete both ATG preparations from all identified humanspecific antibodies to prevent the isolation of already known antigens during our screening for new ATG antigens. To deplete ATGs of HLA antibodies, we have used cells expressing only one subtype of HLA-A, -B and -C antigen. This treatment effectively prevented the isolation of cells expressing HLA molecules, which might indicate that most HLA antibodies in ATGs are broadly crossreactive. Such ATG antibodies are more likely to bind to host cells and tissues and therefore should not act like DSA. However, evidence for ATG antibodies that specifically react with a certain HLA type has been reported, and this phenomenon certainly merits further investigation (3). In a previous study, we have found a high homogeneity between different batches of thymoglobulin (4). However, we have not specifically addressed whether different lots differ in their HLA-antibody profile as suggested by Focosi and Boggi (2). This could be effectively investigated by modifying bead-based HLA-antibody assays for the detection of rabbit antibodies. In addition, the use of anti-human immunoglobulin reagents that do not cross-react with antibodies from rabbits or other animals would prevent false positive results in human panel reactive antibody testing without the need for removal of xenogenic antibodies.Recently, it has been shown that CD107a not only is a marker for cytotoxic activity but also prevents cytotoxic cells from self-destruction (5). We agree that it would be interesting to investigate whether ATG interferes with such a protective role of CD107a, thereby suppressing graftspecific lymphocytes.