Culture-based systems are currently the preferred means for bacterial screening of platelet (PLT) concentrates. Alternative bacterial detection techniques based on nucleic acid amplification have also been developed but these have yet to be fully evaluated. In this study we evaluate a novel 16S rDNA polymerase chain reaction (PCR) assay and compare its performance with automated culture. A total of 2050 time-expired, 176 fresh, and 400 initial-reactive PLT packs were tested by real-time PCR using broadly reactive 16S primers and a "universal" probe (TaqMan, Invitrogen). PLTs were also tested using a microbial detection system (BacT/ALERT, bioMérieux) under aerobic and anaerobic conditions. Seven of 2050 (0.34%) time-expired PLTs were found repeat reactive by PCR on the initial nucleic acid extract but none of these was confirmed positive on testing frozen second aliquots. BacT/ALERT testing also failed to confirm any time-expired PLTs positive on repeat testing, although 0.24% were reactive on the first test. Three of the 400 "initial-reactive" PLT packs were found by both PCR and BacT/ALERT to be contaminated (Escherichia coli, Listeria monocytogenes, and Streptococcus vestibularis identified) and 14 additional packs were confirmed positive by BacT/ALERT only. In 13 of these cases the contaminating organisms were identified as anaerobic skin or oral commensals and the remaining pack was contaminated with Streptococcus pneumoniae. These results demonstrate that the 16S PCR assay is less sensitive than BacT/ALERT and inappropriate for early testing of concentrates. However, rapid PCR assays such as this may be suitable for a strategy of late or prerelease testing.