Considering the possibility that a pooled random donor platelet concentrate could become contaminated by welding with a sterile connecting device, we undertook a study to determine the influence of pooling on the contamination rate. As a control group, apheresis platelets were examined. Bacteriological testing was done with a sensitive CO2 detecting culture system, the BacT/ Alert. Out of 1105 pooled platelet concentrates prepared by the buffy coat method, 15 (1.4%) were confirmed as contaminated, all with Staphylococcus epidermidis and two with a second bacterial species, i.e. Staphylococcus capitis and Propionibacterium acnes, respectively. Median detection time by the BacT/Alert was 23 h. Twelve pools of five units were contaminated, which is significantly more than the three contaminated pools of four units. On the other hand, the reuse of the welding wafers proved not be a risk factor for contamination. One welded tubing segment of a contaminated platelet concentrate failed the air leakage test, an incident which was 73 times more frequent than with the sterile platelet concentrates. We found five pooled platelet concentrates containing Staphylococci from which no bacteria could be grown from the individual buffy coats that had been pooled. We suggest the contamination here to have occurred after separation of the buffy coat from the whole blood, possibly during the welding process. Finally, none out of 378 apheresis platelet concentrates was contaminated. All our observations highlight the potential risk for contamination when making pooled platelet concentrates with a sterile connecting device. For this type of transfusion product, we advocate bacteriological screening of all units before release. The incubation time for the sterility test should, however, be limited to 36 h, if logistical problems with the availability of platelets are to be avoided.