Methods to detect bacterial contamination of blood products

Bacterial contamination of blood products presents an ongoing challenge to transfusion therapy. Contaminating bacteria are typically introduced in low numbers at the time of collection, but can proliferate during storage to reach >10‚ÄövÖœÄ colony forming units/mL (CFUhnL). The consequences of contamination include product wastage and transfusion-transmitted sepsis, but in Australia screening is conducted on only 1% of products by a slow and labour-intensive culture method. Accordingly, there is a need for a more practicable and rapid assay. The purposes of this study were to investigate the growth kinetics of bacteria in stored products, and to develop a rapid method to detect them. Pseudomonas, Staphylococcus and Yersinia species were inoculated into buffy coat platelet concentrates (PCs) and red cell concentrates (RCCs) at 10, 10¬¨‚⧠or 10¬¨‚â• CFU/mL and then stored at 22 ¬¨¬± 2¬¨‚àûC, agitated for PCs, or at 4 ¬¨¬± 2¬¨‚àûC, stationary for RCCs. Bacterial growth was monitored by plate count, and product spoilage (clumping or haemolysis) was noted. Even at the lowest inoculum, all bacterial species grew rapidly in PCs, although clumping was not observed until 3 to 5 days (>10‚ÄövÖ‚àëCFU/mL). In RCCs, all species were recovered after 36 days, but only P. fluorescens and Y. enterocolitica proliferated, reaching >10‚ÄövÖ‚àë CFU/mL before causing haemolysis. To detect bacterial contamination, we developed a PCR-based assay. A method using bead-beating and spin column purification extracted bacterial DNA from the densely cellular blood products in under 30 minutes. Using this template and primers targeting conserved regions of the bacterial 16S rRNA gene, a 1361-bp fragment was amplified from >30 bacterial species and strains tested at less than 10¬¨‚â• CFU/mL in all sample types. Examination of the PCR product using Southern blotting with specific oligonucleotide probes revealed that identification of the bacteria was also possible. Evaluation of this methodology using stored specimens from the kinetic study, showed that bacteria which had grown in blood products could be detected in 2 to 3 days for PCs and in 6 to 9 days for RCCs. Whilst these results matched our preferred time-frame, the sensitivity of detection was reduced in comparison to directly spiked samples. This reduced sensitivity may have been a consequence of long-term storage of the test samples. In conclusion, this study has shown that screening of blood products for bacteria using a universal PCR is feasible and practicable, and can be performed within a time frame required by the transfusion industry.