Platelet support of patients with hematological malignancies.

The most significant advances in platelet transfusion therapy for oncology-hematology patients can be summarized as follows: 1) prophylaxis versus treatment of hemorrhage. Usual practice is based on hemorrhage prophylaxis. Debate is still open on the transfusion trigger, which is traditionally set at 20 x 10(9)/L platelets: some authors suggest it could safely be decreased in stable patients to 10 or 5 x 10(9)/L platelets; 2) preparation of platelet concentrates. Platelets prepared from platelet-rich plasma or buffy-coats obtained from multiple bag donations should be used as the first-choice for all patients, while apheresis platelets, which have a significantly higher cost of production, should be reserved for patients refractory to random donor support. The final choice, however, of a prudent strategy must also consider logistic aspects, such as product availability, distance from site of production to site of use, etc; 3) leukocyte reduction. Filtration is the method of choice to prepare leukocyte-reduced platelets. Leukocyte-reduced platelets can be used to prevent transmission of CMV in selected patient groups for whom this is indicated. When leukocyte reduction is used for the prevention of NHFTR, it should be performed with fresh platelets and reserved for patients developing more than 1 reaction. Routine leukocyte reduction for all oncology-hematology patients cannot be recommended at this time, in the absence of definitive information on the cost-effectiveness of this approach; 4) quality control. Studies are under way to check whether evaluation of the swirling phenomenon, that is produced by good quality platelets when inspected with the naked eye against a strong light source is a useful and inexpensive test for quality control; 5) correction of refractoriness to random donor platelet support. Effective platelets for refractory patients can be obtained through HLA typing and/or platelet cross-matching. Although HLA typing can be very effective, cross-matching seems to be equally effective, simpler and less expensive.