In 1999, we implemented an automated platelet cross-matching (XM) programme to select compatible platelets from the local inventory for patients refractory to random donor platelets. In this study, we evaluated platelet count increments in 40 consecutive refractory patients (8.3% of 480 consecutive platelet recipients) given 569 cross-match-negative platelets between April 1999 and December 2001. XM was performed automatically with a commercially available immunoadherence assay. Pre-, 1- and 24-h post-transfusion platelet counts (mean +/- SD) for the 569 XM-negative platelet transfusions containing 302 +/- 71 x 109 platelets were 7.7 +/- 5.5, 32.0 +/- 21.0 and 16.8 +/- 15.5 x 109/l respectively. Increments were significantly higher (P < 0.05, t-test) than those observed in the same patients given 303 random platelet pools (dose = 318 +/- 52 x 109 platelets) during the month before refractoriness was detected, when pre-, 1- and 24-h post-transfusion counts were 7.0 +/- 8.6, 15.9 +/- 16.1 and 9.6 +/- 12.8 x 109/l respectively. The cost of the platelet XM disposable kit per transfusion to produce 1-h post-transfusion platelet count increments >10 x 109/l was euro 447. This programme enabled the rapid selection of effective platelets for refractory patients, from the local inventory.