Evolution of surveillance methods for detection of bacterial contamination of platelets in a university hospital, 1991 through 2004
Version of Record online: 4 MAY 2006
Volume 46, Issue 5, pages 719–730, May 2006
How to Cite
Yomtovian, R. A., Palavecino, E. L., Dysktra, A. H., Downes, K. A., Morrissey, A. M., Bajaksouzian, S., Pokorny, M. A., Lazarus, H. M. and Jacobs, M. R. (2006), Evolution of surveillance methods for detection of bacterial contamination of platelets in a university hospital, 1991 through 2004. Transfusion, 46: 719–730. doi: 10.1111/j.1537-2995.2006.00790.x
- Issue online: 4 MAY 2006
- Version of Record online: 4 MAY 2006
- Received for publication July 15, 2005; revision received September 24, 2005, and accepted September 26, 2005.
BACKGROUND: Platelet (PLT) bacterial contamination (PBC) is the most common transfusion-associated infection. It is important to understand the impact of interventions addressing this problem.
STUDY DESIGN AND METHODS: PBC was studied by prospective (active) and transfusion-reaction triggered (passive) surveillance from July 1991 to December 2004. Active surveillance, utilized for 10 years, included bacterial culture of all or 4- and 5-day-old PLTs at issue and intermittent use of Gram stain, pH measurements, and early (24-hr) culture of single-donor plateletpheresis (SDP) units.
RESULTS: Active surveillance detected 38 instances of PBC, 7 in SDP units (1:2213) and 31 in random-donor PLT units (1:2090 units, p = 0.89; or 1:418 pools of 5 units, p < 0.001). Contaminants were coagulase-negative staphylococci (CONS; n = 27), Staphylococcus aureus (4), Bacillus cereus (1), Serratia marcescens (2), streptococci (2 S. bovis, 1 S. uberis), and CONS with viridans group streptococcus (1). Only one instance of contamination, caused by Pseudomonas aeruginosa, was detected by passive surveillance, with fatal outcome. Colony counts of contaminants ranged from 0.5 × 102 to 4 × 1011 colony-forming units per mL at time of issue. PBC was interdicted before transfusion in 6 cases through Gram stain screening. Transfusion reactions occurred in 13 of 32 recipients (41%), with 9 severe reactions (28%) and 3 deaths (9%). pH testing failed to detect 5 contaminated units and resulted in discard of nearly 2 percent of units, whereas culture of SDP units at 24 hours failed to identify a contaminated unit.
CONCLUSION: Improved active surveillance methods for detecting PBC are needed to improve the safety of PLT transfusions.