Here, the influence of T vs T and B cell depletion on the incidence of EBV-associated lymphoproliferative disorder (EBV-LPD) after bone marrow transplantation (BMT) from a matched unrelated donor (MUD) is analyzed. From 1982 to 1997 the soy bean agglutinin/sheep red blood cell (SBA/SRBC) method was used for T cell depletion. This technique is well established, but the use of SRBC has a risk of transmitting prions or viruses. Therefore, a new T cell depletion method was introduced, using CD2 and CD3 monoclonal antibodies (CD2/3 method) instead of SRBC. Unfortunately, this led to an unexpected high number of EBV-LPDs in patients receiving transplants from MUDs. SBA depletion was reintroduced and combined with the CD2/3 method (SBA/CD2/3) in this patient population, later replaced by B cell-specific (CD19 and CD22) antibodies (CD3/19/22 method). The number of T (x 10(5)/kg) and B (x 10(5)/kg) cells in the graft was 1.5 +/- 0.8 and 2 +/- 1 (T/B ratio 0.75), 2.2 +/- 2.0 and 41 +/- 21 (ratio 0.055), 5.0 +/- 0.0 and 2 +/- 1 (ratio 2.5), 2.5 +/- 1.2 and 10 +/- 6 (ratio 0.25) using the SBA/SRBC, CD2/3, SBA/CD2/3 and CD3/19/22 techniques, respectively. When B cell depletion was performed (SBA/SRBC, SBA/CD2/3, CD3/19/22) four out of 31 patients (13%) receiving a BMT from a MUD developed an EBV-LPD. Without B cell depletion (CD2/3) this occurred in five out of seven patients (71%) (P < 0.05). A T/B cell ratio in the graft of > or = 0.25 seems sufficient to significantly reduce the incidence of EBV-LPD after BMT from MUDs.