Bacterial Contamination of Cellular Blood Components
By F. Enrique Alvarez and Benjamin Lichtiger, M.D., Ph.D.
Copyright 1993-2010 The University of Texas MD Anderson Cancer Center, Houston, Texas. All rights reserved.
Bacterial contamination of blood products and transfusion-transmitted sepsis have always been two of the most important and critical untoward events in the collection, processing, storage, and infusion of blood (1). Recently, numerous published articles and studies have reported a rise in the number of bacterially contaminated units in blood bank inventories (2-4). If true, the impact would be particularly critical on our transfusion-dependent patient population, given the unique immunosuppression and susceptibility to infection of cancer patients and the potentially fatal outcome of infection in such patients (5). Thus, we retrospectively reviewed all cellular blood products subjected to bacterial cultures at our institution between December 1989 and June 1993 to determine the incidence of bacterially contaminated units of packed red blood cells (PRBCs), random-donor platelet concentrates (RDPs), and single-donor platelet concentrates (SDPs); to assess the seriousness and impact of this issue in our unique environment; and to point out necessary corrective measures.
Our review covered all bacterial cultures performed on cellular blood components during the 43 months from December 1989 to June 1993. The cultured units were segregated into three main categories: PRBCs, RDPs, and SDPs. These were then subdivided into the following groups: Group I, PRBCs implicated in transfusion reactions; Group II, PRBCs checked out and returned to the transfusion service after our self-imposed 30-minute limit; Group III, PRBCs that expired on the shelf; Group IV, PRBCs randomly selected for quality control (QC); Group V, RDPs implicated in transfusion reactions; Group VI, RDPs randomly selected for QC; Group VII, SDPs implicated in transfusion reactions; and Group VIII, SDPs randomly selected for QC.
To better interpret the results of our study, we analyzed them individually by component type to give appropriate consideration to variables such as storage temperature and age of the component. A summary review of the records of all cellular blood components cultured during the study can be found in Table 1.
Table 1. Results of Study by Component Group
|Group||Category||Unit age in days||Storage temperature||No. units cultured||No. positive cultures (%)|
|I||PRBCs implicated in Trans Rx*||1-42||1-6 degrees C||159||5 (3.1%)|
|II||PRBCs out > 30 min||1-42||1-6 degrees C||155||None|
|III||PRBCS that expired on shelf||>42||1-6 degrees C||276||3 (1.1%)**|
|IV||PRBCs selected for QC||1-42||1-6 degrees C||36||None|
|V||RDPS implicated in Trans Rx||1-5||20-24 degrees C||309||12 (3.9%)***|
|VI||RDPs selected for QC||1-5||20-24 degrees C||214||3 (1.4%)|
|VII||SDPs implicated in Trans Rx||1-2||20-24 degrees C||43||2 (4.7%)|
|VIII||SDPs selected for QC||>5****||20-24 degrees C||54||None|
* Trans Rx, tranfusion reactions.
** Most units cultured were from autologous or directed donors.
*** Three pools of four RDP units each; low probability that all platelets in pool were initially contaminated.
**** Left past 5 days to allow bacterial growth, if present.
Of 102,593 PRBC units produced from December 1989 to June 1993, 626 (0.6%) were cultured. These 626 units belonged to either Group I, II, III (approximately half of Group III units were either autologous or directed), or IV. Among these four groups, only two, Groups I and III, showed any kind of bacterial growth upon culture. Of 159 PRBC units in Group I, five (3.1%) were positive for bacterial culture, with four instances of Bacillus and/or coagulase-negative Staphylococcus species (routinely, no further identification is indicated by our microbiology protocol) and one occurrence of Enterobacter agglomerans. Of 276 units in Group III, three (1.1%) were positive: two cases of gram-negative rods and one case of coagulase-negative Staphylococcus species. The rest of the PRBCs cultured were negative for bacterial contamination, bringing the total of cultured units showing the presence of bacterial growth to eight of 626 (1.3%).
The platelets reviewed in the study were divided into RDPs and SDPs. From a total of 95,005 RDP units and 13,641 SDP units produced, 523 (0.6%) and 97 (0.7%) units, respectively, were checked for bacterial contamination.
From the cluster of 309 RDP units in Group V, three pools of four RDP units each (12 units), or 3.9%, were found to be positive for bacterial growth, with three instances of coagulase-negative Staphylococcus species. Of a total of 214 units in Group VI, only three (1.4%) showed evidence of bacterial contamination, with two occurrences of coagulase-negative Staphylococcus species plus an unclassified gram-negative rod and one instance of coagulase-negative Staphylococcus species plus Pseudomonas paucimobilis (the latter was considered a contaminant of the inoculation process during culture set-up). In total, of 523 RDP units cultured (309 from Group V and 214 from Group VI), 2.9% (assuming 12 units contaminated in Group V) showed evidence of bacterial contamination.
Of a total of 43 units in Group VII, only two (4.7%) showed evidence of bacterial contamination. One unit grew gram-variable rods, and the other grew an Enterococcus species.
Of 54 units in Group VIII, none was positive for bacterial growth or signs of contamination. As a whole, of 97 SDP units cultured from 13,641 units produced, two units (2.1%) showed evidence of contamination.
Our study yielded interesting results, which, although comparable with recently reported levels of contamination (1-10% on average) (2,3), differ particularly in the type of organisms isolated. Table 2 summarizes, by group, the number of positive bacterial cultures in our study; notable for its absence is Yersinia enterocolitica, the leading cause of concern in most recent studies (2,5-7). Goldman and Blachjman (2) cited a steady increase in the number of deaths associated with bacterial sepsis between 1976 and 1988 and the presence of Y. enterocolitica as an insidious contaminant with devastating consequences. The concern over Yersinia is legitimized by its ability to proliferate at refrigerated temperatures (1-6 degrees Celsisus) and its predilection for an iron-rich environment, which makes stored blood a preferred growth medium. Also, studies by Arduino et al. (7) and Tipple et al. (8) agree in concluding that even low-level Yersinia contamination can result in a blood product containing high concentrations of endotoxin and bacteria after 2-3 weeks of storage, thus making it a very serious threat to the blood bank inventory. That no Y. enterocolitica was isolated in any of our cultures is significant and could be directly related to our donor selection protocol, which, following suggestions by the U.S. Food and Drug Administration's Blood Product Advisory Committee (9,10), includes questions dealing with recent conditions such as gastrointestinal ailments, diarrhea, and fever. Admittedly, this approach is only expected to identify 50% of the cases of Yersinia infection and eliminate 11% of donors (9,10); it also implies that occasionally blood donors may be in a state of incubation or recovery from bacteremia (asymptomatic) and thus able to slip through the donor selection process, placing a potentially contaminated unit into the blood bank's available inventory. In most of these cases, direct visual inspection of the unit does not reveal turbidity, icterus, or any other signs of contamination, even though, according to Glick et al. (11), this is not a reliable method for determining if contamination is present.
Table 2. Summary of Positive Cultures by Group
|Group||Category||Instances of organisms isolated|
|I||PRBCs implicated in Trans Rx*||4 coagulase-negative Staphylococcus|
|1 Enterobacter agglomerans**|
|III||PRBCs that expired on shelf||2 gram-negative rods|
|1 coagulase-negative Staphylococcus|
|V||RDPs implicated in Trans Rx||3 coagulase-negative Staphylococcus***|
|VI||RDPs selected for QC||2 coagulase-negative Staphylococcus and gram-negative rods (unclassified)|
|1 coagulase-negative Staphylococcus and Pseudomonas paucimobilis****|
|VII||SDPs implicated in Trans Rx||1 gram-variable rods|
* Trans Rx, transfusion reactions.
** Patient was septic and receiving massive antibiotic therapy before transfusion.
*** Three culture-positive pools composed of four RDP units each.
**** Considered a contaminant introduced during media inoculation process.
Other factors such as possible contamination during pooling of components, transport of units to the transfusion site, and procedures directly related to the bedside infusion of blood components must also be addressed. While the last two procedures are usually "off-limits" to our blood bank personnel as far as ensuring proper handling, the pooling of components is directly under our control, and new efforts have been directed toward procuring a more sterile product.
Manual pooling of RDPs can be a source of contamination when strict adherence to the use of sterile technique is not followed. A busy blood-distribution area may also aggravate the problem since time there may be precious. The advent of sterile connecting devices (SCDs) has had, in our case, considerable impact on our ability to maintain a smooth flow of blood components through our blood distribution area by allowing us to pool RDPs ahead of their time of release. This has in turn created a less stressful environment and also ensured the sterile handling of the product, thus reducing the possibility of contamination or transfusion-transmitted sepsis.
It should be noted that the highest percentage of positive cultures in our study (4.7%) involved SDP units that were 1-2 days old and were implicated in transfusion reactions. In contrast, none of the SDPs in the quality control group that were held for up to 7 days before culture showed any evidence of bacterial contamination. This suggests that the positive results were a consequence of contamination during collection or manipulation of the product by nursing personnel. Next, the unit testing procedure itself must be analyzed and put into perspective. Given the manipulation involved in sampling the units, and given that the possibility of contamination during inoculation and culturing cannot be completely discounted (2), a positive culture does not necessarily reflect direct contamination of the product from the original donor. It is notable that six of 18 positive products in our study grew gram-negative rods. The cold storage conditions at which some blood products are kept may favor the growth of environmental gram-negative rods such as P. paucimobilis and facilitate contamination from outside the collection bags. Thus, a potential control measure may be to disinfect the storage cabinets more frequently.
The patient population and the type of medication being administered can also have an impact on the morbidity and mortality of the potential bacterial contaminants. In the case of cancer patients requiring constant transfusions of platelets and RBCs, antibiotic therapy is also common and this may influence the clinical outcome associated with the transfusion of bacterially contaminated blood products.
Another aspect of our research seems to agree with previous findings by Hamill et al. (12) with regard to the amount of time a unit of PRBCs may stay outside the blood bank after release for infusion before being returned to the blood distribution area and safely placed back into the available inventory. Hamill et al. (12) have suggested that a 2-hour cutoff limit be adopted since no significant increase in bacterial proliferation seems to occur and since extension of the currently used 30-minute "grace period" would allow the recapture of most units being discarded while providing a reasonably safe margin. In our study, none of the units in Group II showed any evidence of bacterial contamination, and this supports the assertion of Hamill et al. (12).
We agree with Goldman and Blajchman (2) that more research and the implementation of more comprehensive preventive measures such as the extension of donor screenings; the avoidance of scarred phlebotomy sites, which may be difficult to disinfect; and the development of a rapid, reliable test for the detection of bacterial contaminants are needed. Furthermore, as mentioned by Hoppe (13), the addition of antibiotics to blood containers to suppress bacterial growth and the use of prestorage filtration devices to remove white blood cells (14-18) are avenues that still remain unexplored. But, in the age of cutbacks and increased demand for cost-cutting measures, expensive new technology must be weighed against its overall effectiveness and the impact it would have on the operational capabilities of the transfusion service.
In closing, and given our main concern of transfusion-transmitted sepsis, it should be noted that of the patients receiving culture-positive units from our study, none showed evidence of transfusion-related infection or any consequence thereof. The results of our study are comparatively low, but given the increased susceptibility of immunosuppressed cancer patients, more extensive monitoring of bacterial contamination must be implemented to identify the source of the contaminants.
- Braude AI, Sanford JP, Bartlett JE, Mallery OT. Effects of clinical significance of bacterial contaminants in transfused blood. J Lab Clin Med 39:902-916, 1952.
- Goldman M, Blajchman MA. Blood product-associated bacterial sepsis. Trans Med Rev 7:73-83, 1991.
- Honig CL, Bove JR. Transfusion-associated fatalities: Review of Bureau of Biologics Report 1976-1978. Transfusion 20:653-661, 1980.
- Braine HG, Kickler TS, Charanche P, et al. Bacterial sepsis secondary to platelet transfusion: An adverse effect of extended storage at room temperature. Transfusion 26:391-393, 1986.
- Sazama K. Report of 355 transfusion-associated deaths: 1976-1985. Transfusion 30:583-590, 1990.
- Aber RC. Transfusion-associated Yersinia enterocolitica. Transfusion 30:193-195, 1990.
- Arduino MJ, Bland LA, Tipple MA, et al. Growth and endotoxin production of Yersinia enterocolitica and Enterobacter agglomerans in packed erythrocytes. J Clin Microbiol 27:1483-1485, 1989.
- Tipple MA, Bland LA, Murphy JJ, et al. Sepsis associated with transfusion of red cells contaminated with Yersinia enterocolitica. Transfusion 30:207-213, 1990.
- Blood Products Advisory Committee. Prevention of transfusion of Yersinia infection (transcript). Bethesda: U.S. Food and Drug Administration, January 17-18, 1991.
- Blood Products Advisory Committee. Prevention of transmission of Yersinia infection (transcript). Bethesda: U.S. Food and Drug Administration, May 9, 1991.
- Glick MR, Ryder KW, Glick SJ, Woods JR. Unreliable visual estimation of the incidence and amount of turbidity, hemolysis, and icterus in serum from hospitalized patients. Clin Chem 35:837-839, 1989.
- Hamill TR, Hamill SG, Busch MP. Effects of room-temperature exposure on bacterial growth in stored red cells. Transfusion 30:302-306, 1990.
- Hoppe PA. Interim measures for detection of bacterially contaminated red cell components. Transfusion 32:199-201, 1992.
- Heal JM, Cohen HJ. Do white cells in stored blood components reduce the likelihood of posttransfusion bacterial sepsis? Transfusion 31:581-583, 1991.
- Hogman CF, Gong J, Hambraeus A, Johansson CS, Eriksson L. The role of white cells in the transmission of Yersinia enterocolitica in blood components. Transfusion 32:654-657, 1992.
- Kim DM, Brecher ME, Bland LA, et al. Prestorage removal of Yersinia enterocolitica from red cells with white cell-reduction filters. Transfusion 32:658-662, 1992.
- Wenz B, Burns ER, Freundlich LF. Prevention of growth of Yersinia enterocolitica in blood by polyester fiber filtration. Transfusion 32:663-666, 1992.
- Hogman CF, Gong J, Eriksson L, Hambraeus A, Johansson CS. White cells protect donor blood against bacterial contamination. Transfusion 31:620-626, 1991.