Abstract
Bacterial contamination has represented the second leading cause for transfusion- associated mortality. The French hemovigilance study indicated that in neutropenic and immune depressed patients the mortality is almost 100%. With the approval of two different systems to perform quality control tests on leukocyte-reduced platelets from bacterial contamination, we instituted a process to culture each platelet unit after a 24 hour incubation hold at 22±2C. After this hold, an aliquot (4 mL) was removed from each product bag and cultured aerobically (28% using the Pall BDS and 72% using the BacT/Alert system). The assignment was predetermined by collection site. Following 24 hour incubation at 35C, the cultures were reviewed and those platelet products with no growth/negative results were released. The BacT/Alert bottles remained under incubation until product outdate 5 days post collection. Between September 2003 and June 2004, 69,161 products were cultured. The reference laboratories performed on both the product and culture bottles/pouches a gram stain, and culture on agar plate and in tryptic soy broth, reading after five days if no growth. To resolve any questions regarding sampling volume as a contribution to lack of identification of true positives, larger volumes (20 mL) of product associated with all positive culture bottles (20% of samples sent to the reference laboratory) were examined but did not change the discordant negative yield versus true positive yield of organisms. There were 12 true positives (growth in the culture system which could be reproduced on subculture with the same organism growing from the product and culture system by an independent microbiology laboratory). There were 153 false positives or discordant negatives (growth in the culture system, which did not grow in the product but did repeat on subculture of the culture bottle/pouch). Bacterial species identified in products included Staph coagulase neg 5, Staph aureus 2, Bacillus species 1, Strep Group A beta hemolytic 1, Diphtheroids 1, Enteric gram negative bacillus 1, Corynebacterium species 1, Propionbacterium acnes 1. (One product had more than one organism.) The discordant negatives included additionally Comamonas acidivorans 1 and Micrococcus species 1; however, 95% of the organisms cultured were recognized skin contaminant organisms. There were 2.5 times more false positives with the Pall BDS system than the BacT/Alert. (The BDS system has been superseded by the eBDS system as of July). Comparatively the number of true positives remained within the 95% CI for the two systems. These results identify a true positive frequency of 1 in 5763 (0.00017%) [95% CI 5618–5911], but a false positive rate of 1 in 453 (0.00221%). There were 4 delayed positive cultures (bottle developed growth after the 24 hour reading time); however, none of the associated transfused products resulted in a reported reaction. Culturing platelet products has identified a number of organisms, primarily skin organisms. These data support the prevalence of true positive bacterial contamination in Platelets, Pheresis products as previous models have reported.
An improved device for sampling bacterial populations on blankets