Reduction of the risk of bacterial contamination of blood components through diversion of the first part of the donation of blood and blood components

Introduction

The level of safety of transfusion therapy is now very high thanks to the combination of serological methods and genomic amplification used to screen for transmissible diseases and the meticulous care with which repeat, voluntary, unpaid donors are selected^1–4.

The main risk of transfusion-related infectious diseases is currently that of bacterial sepsis^1,3–5. The risk of bacterial contamination of blood components is, in fact, estimated to be about three orders of magnitude greater than that of post-transfusional HIV and HCV infections^5,6; the risk of bacterial sepsis causes by the transfusion of platelets is more than two orders of magnitude greater than the risk of the same viral infections^5,6.

According to a study published in 2004⁷, the year in which bacteriological screening tests for platelet units were introduced in the USA^3,8,9, the average prevalence of bacterial contamination in platelets from whole blood was 33.9/100,000 units, that of platelets from apheresis 51/100,000, while that of red cell concentrates was 2.6/100,000. The overall prevalence of bacterial contamination of units of cellular blood components was, therefore, about 1 in 3,000 donations (33.3/100,000)^7,10.

Table I reports the data on bacterial contamination of units of platelets and red blood cells derived from studies before 2003^11–19. More recent estimates indicate that bacterial culture tests on units of platelets are positive, and confirmed as such, in about 1 in 5,000 units (20/100,000)⁴. The risk of receiving platelet concentrates contaminated by bacteria is, therefore, considerably higher than the risk of post-transfusion infection by HIV, HCV, HBV and HTLV^3,4.

Table I.

Prevalence of bacterial contamination of cellular blood components recorded in studies before 2003

Type of blood component	Reference	Year of publication	Number of positive units	Number of units tested	Prevalence per 100,000 units
Platelets from whole blood	Barrett BB11	1993	1	4,272	23.4
	Yomtovian R12	1993	6	14,481	41.4
	Chiu EKW13	1994	10	21,503	46.5
	Blajchman MA14	1994	16	31,610	50.6
	Leiby DA15	1997	4	4,995	80.1
	Blajchman MA16	1997	7	10,065	69.5
Total			44	86,926	50.6
Platelets from apheresis	Blajchman MA17	1992	14	6,055	231.2
	Barrett BB11	1993	5	17,928	27.9
	Yomtovian R12	1993	0	2,476	0.0
	Dzieczkowski JS18	1995	1	5,197	19.2
	AuBuchon JP19	2002	1	2,678	37.3
Total			21	34,334	61.2
Red blood cell concentrates	Barrett BB11	1993	1	31,385	3.2
	Dzieczkowski JS18	1995	1	7,080	14.1
Total			2	38,465	5.2
Overall total			67	159,725	41.9

Prior to the introduction of bacteriological screening tests²⁰, transfusion-related sepsis occurred after approximately 1 in 25,000 transfusions of platelet concentrates and 1 in 250,000 transfusions of red cells^5,7,20.

In the USA, where about 14 million units of red cells are transfused each year, in the period from 1976 to 1998, there were reports of 26 deaths due to bacterial contamination of this blood component, with an incidence, therefore, of just over one case per year^21,22. More recently, still in the USA, analysis of data from 1998 to 2000 indicated that the incidence of fatal events was 1 case per 8,000,000 units of transfused red blood cells²³. At the other end of the scale, the incidence of fatal events reported in New Zealand is 1 for every 104,000 units of transfused red cells²¹.

According to North American studies, deaths following the transfusion of platelets are more frequent and occur after 1 in 50,000–500,000 transfusions^23,24.

Analysis of haemovigilance data from the United Kingdom, France and USA indicate that the mortality risk due to sepsis following transfusion of platelets is 1 case per 7,500–100,000 units transfused²¹.

The contamination of platelets is easier and more common than that of red cells because of the storage temperature of platelets (20–24 °C), which favours the proliferation of bacteria, unlike the storage temperature of red cells, which is 2–6 °C ^25,26. Fresh-frozen plasma, precisely because of its storage conditions^25,26, is rarely contaminated by bacteria; on the other hand, this product can be contaminated during the thawing process in inadequately sterilised equipment²¹.

The published mortality rates due to sepsis from bacterial contamination of transfused units of blood components vary in the different series: this is probably because of unrecognised and/or not reported cases and because of variations in the incidence in the context of different national epidemiological situations²¹.

The aim of this review is to consider the scientific evidence supporting the efficacy of diversion of the first part of the blood collected in preventing bacterial contamination of blood components.

Source of contamination of blood components and prevention strategies

It is not always easy to identify the source of bacterial contamination of blood components. The contamination can be due to endogenous causes, such as asymptomatic bacteraemia in the donor, including transient episodes or due to exogenous causes, mainly related to contamination by transient or commensal bacterial flora present on the skin of the donor’s arm or to contamination of the collection bags/kits and/or the blood component itself during the production process and/or storage^27,28. In these latter cases, prevention is based on applying the principles of “good manufacturing practice”, using sterile, closed collection systems and maintaining scrupulous hygiene of areas dedicated to the processing and storage of blood and blood components as well as the areas in which the materials used for the collection are stored.

Diversion of the first part of the blood collected is one of the interventions that can prevent the contamination of blood components caused by the introduction of skin bacteria into units of whole blood or blood components from apheresis at the time of venipuncture. Indeed, the major source of bacterial contamination is the donor’s skin and disinfection, even if carried out very carefully thus reducing the bacterial load present on the superficial layers of the skin, cannot act on the deeper layers. The introduction of a needle through the skin to collect blood causes, in 65% of all cases²⁹, the passage of skin fragments containing live microbes or the movement of microbes from small skin flaps produced by the needle itself into the collection bags³⁰. Furthermore, it is even more difficult to achieve satisfactory disinfection of a skin surface that has become irregular due to scars from previous donations²⁸.

The aim of diverting the first part of the blood collected into a collateral bag, used for the performance of tests of biological validation of the unit donated, is precisely that of preventing the initial flow of blood, containing microbes or skin fragments from the donor, from entering the collection bag and thus contaminating it^28,29,31,32.

Diversion of the first part of the blood

The diversion of the first 10–50 mL of donated blood, with the aim of preventing microbes or skin fragments from entering the collection bag, has been the object of numerous studies from 1995 onwards^33–47.

Olthuis et al. were the first, in 1995, to demonstrate that the first 10 mL of blood, collected at the beginning of a plasmapheresis procedure with an open system, were more frequently contaminated by bacteria than were the subsequently collected 10 mL³⁴; indeed, diversion of the first part of the donation guaranteed an 88% reduction in the percentage of bacterial contamination²⁸.

In the same year, using an experimental model, Figueroa showed that most of the bacteria (100 colony-forming units of Staphylococcus aureus) inoculated in the lumen of a collection needle entered the first 10 mL of fluid that it collected; this demonstrated, experimentally, that the diversion of the first 10 mL of blood collected could reduce the bacterial load of skin contaminants in the blood/blood component collected³⁵.

In 2000, Wagner et al., also using an in vitro experimental model, demonstrated that diversion of the first 21–42 mL of physiological saline or whole blood reduced the load of colonies of Staphylococcus aureus, deliberately inoculated on the collection surface, by about 1 log³⁶. This confirmed that a diversion system for the collection of samples of blood, before filling the collection bags, would be able to reduce the bacterial load in blood components subsequently produced from whole blood and, therefore, the incidence of sepsis due to cutaneous microbes.

In 2001, Cecile Bruneau’s group in France published the results of a multicentre study³⁷ on new collection bags, specifically modified by the manufacturer (Maco Pharma, Tourcoing, France), which added two consecutive satellite bags, each containing 15 mL, to the collection line with the aim of comparing bacterial contamination between the first and second 15 mL of blood collected. The authors demonstrated that diversion of the first 15 mL reduced the bacterial contamination of the units of blood by 72% and that using the first 30 mL of blood to carry out the tests of biological validation of the units further reduced the potential bacterial load within the bags of whole blood.

Again in France, in 2002, Schneider’s group obtained a 58% reduction in the contamination of platelet concentrates (PCs) from buffy coat (BC) pools by using diversion of the first part of the blood collected³⁸; contemporaneously, in the Netherlands, Bos et al. obtained a 53% reduction of contamination of the same blood component³⁹.

In the same country, the use of blood bags supplied with diversion bags showed that the diversion of the first 20 mL of blood reduced the bacterial contamination of PCs from BC pools from 1% to 0.4% (− 60%)⁴⁰.

A subsequent multicentre, prospective clinical study, whose results were published in 2002 by Dirk de Korte’s group from the Netherlands, demonstrated that the diversion of the first 10 mL of the donation enabled a statistically significant, 40% reduction in the percentage of contaminated whole blood units, compared to that of the standard collection without diversion⁴¹; consequently, there was also a statistically significant reduction (from 0.14% to 0.03% - p<0.02) in the frequency of staphylococcal contamination of units.

On the basis of these results in the Netherlands, since July 1, 2004, diversion of the first 20–30 mL of blood (used for validation tests of the units) has become obligatory for all whole blood and apheresis donations and this produced, on a national scale, a 59% reduction in bacterial contamination of PCs from BC pools already after 1 year⁴².

In 2006, the same Dutch group published the results of another prospective, multicentre clinical study carried out between 2002 and 2003, in which they evaluated the effect of skin disinfection techniques, diversion bags, and bacterial detection tests on the contamination of more than 110,000 PCs from BC pools⁴³. Diversion of the first 20–30 mL of blood was associated with a statistically significant reduction in the bacterial contamination of this blood component of 47% with the “old disinfection technique” and of 54% with the “new disinfection technique” (isopropylic alcohol 70%). Furthermore, the diversion caused a statistically significant reduction in contamination due to microbes of cutaneous origin.

More recent data from the same country, updated to 2007, but currently available only in abstract form, confirm the efficacy of the strategy adopted in 2004⁴⁴; indeed, before diversion was obligatory, the rate of bacterial contamination among 127,797 PCs from BC pools was 0.90%, while the rate in 147,021 PCs produced after July 1, 2004, had fallen to 0.46% (− 49%). The strategy of diversion also caused:

1. a 35% reduction in the units transfused and found, subsequently, to be positive in culture tests;

2. a reduction (from 40% to 25%) of the frequency of contamination of red cell concentrates produced from the units from which the PCs from BC pools positive for bacterial contamination originated. The bacterial strains on which the diversion had the greatest effect were diphtheroids and coagulase-negative staphylococci.

A prospective study on units of whole blood, carried out in the United Kingdom and published in 2004 by McDonald, showed that diversion of the first 20 mL of blood produced a 47% reduction in bacterial contamination⁴⁵.

A statistically significant reduction (p = 0.005) in bacterial contamination of PCs from whole blood was also found in a Canadian study by Robillard⁴⁶, who analysed the data produced by the local haemovigilance system. In Canada the diversion of the first 40 mL of blood, introduced in February 2003, caused, in 2004, a 90% reduction in the bacterial contamination of PCs from whole blood; this result was confirmed by the data updated to the following year, in which no PCs, from either whole blood or apheresis, were found to be contaminated by bacteria⁴⁷.

In a study recently performed in the USA on bacterial contamination of platelets from apheresis³³, it was found that the bacterial contamination of PCs produced by single-needle cell separators was 47% lower than that of PCs produced by double-needle separators. The lower rate of bacterial contamination of the PCs produced by the former is explained by the different configuration of the two types of cell separators. In the double-needle separators, the sample for tests of biological validation is taken from the reinfusion line and there is not, therefore, diversion of the first amount of blood taken; in contrast, in the single-needle separators, the initial sample (40–50 mL) for haematological and biochemical investigations is taken from the collection line, thus preventing this blood from contaminating the collection kit and blood component produced.

Table II summarises the main studies on diversion, some of which are only available in abstract form, the amount of blood diverted and the percentage reduction in bacterial contamination obtained in the blood components under analysis.

Table II.

Reduction of the percentage of bacterial contamination of blood components achieved with different volumes of diversion in studies from 1995 to 2007

Reference	Year of publication	Country	Production procedure/blood component analysed	Volume of diversion (mL)	Reduction of contamination of the blood components (%)
Olthuis H34,	1995	The Netherlands	Plasmapheresis	10	88
Bruneau C37	2001	France	Whole blood	15 (+15)	72
Schneider T38	2002	France	PC from BC pools	Not stated	58
Bos H39	2002	The Netherlands	PC from BC pools	10	53
Yedema T40	2003	The Netherlands	PC from BC pools	20	60
de Korte D41	2002	The Netherlands	Whole blood	10	40
McDonald CP45	2004	United Kingdom	Whole blood	20	47
Robillard P46,	2005	Canada	PC from whole blood	40	90
De Korte D44	2007	The Netherlands	PC from BC pools	20–30	49
Eder AF33	2007	USA	Platelets from apheresis	40–50	47

PC, platelet concentrates; BC, buffy coat.

Conclusions

Post-transfusion sepsis is the most common cause of death due to infectious agents^30,48.

According to data from SHOT (Serious Hazards Of Transfusion)⁴⁹, the haemovigilance system in the United Kingdom, in the period 1995–2006, there were 33 cases of post-transfusion bacterial sepsis, 29 of which (88%) were due to treatment with platelets (10 cases due to platelets from apheresis and 19 to platelets from pools) and four due to treatment with red cell concentrates. In the same period, there were no deaths due to post-transfusion infections with hepatitis viruses, HIV or HTLV-I, while eight deaths were caused by bacterial contamination of transfused platelets. This puts platelet transfusion-related sepsis alongside transfusion-related acute lung injury (TRALI) as the second most frequent cause of transfusion-related death, after post-transfusion graft-versus-host disease^48,49.

The frequency of clinical symptoms following transfusion of platelets contaminated by bacteria varies greatly; indeed, the reported range extends from a frequency of 1 case per 358 transfusions of platelets from whole blood¹³, reported in a study from 1994, to the considerably lower value of 1 case per 73,000 transfusions of platelets from single donors²⁴, reported in a study from 2001. This latter result is completely in line with the findings of the American Red Cross³³ that, between 2004 and 2006, the residual risk of septic reactions to platelets from apheresis was 1 case per 74,807 units, and the residual risk of fatal reactions was 1 case per 498,711 units distributed.

This broad variability depends on the criteria and methods used to collect the data and is also influenced by the method used to produce the PC (from single units or pools, from BC or from platelet-rich plasma, or from apheresis) and whether strategies to divert the first part of the blood collected had been introduced⁵⁰.

The technique of diverting the first 20–40 mL of blood at the beginning of the donation of whole blood or blood components from apheresis is relatively simple and has no effect on the donor, since this is the quantity of blood that is normally taken to carry out the tests necessary for the biological validation of the unit donated.

The data from the studies analysed demonstrate the fundamental role that this modification to the collection process can have on reducing the rate of bacterial contamination of blood components: indeed, the percentage of contaminated blood components is decreased by 40–90%. The range in the reduction is probably due to differences in criteria for collecting and analysing the data and in the production processes of the blood components examined in the various studies.

In the United Kingdom, the introduction of diversion was already suggested in the SHOT report of 2002 as one of the strategies to use to reduce the risk of bacterial contamination of blood components⁴⁹; in the same year, this measure was made obligatory in Germany⁵¹ and other countries, including France, the Netherlands and Canada, subsequently adopted this modification of the production process of blood components^5,46,52, with manufacturers of blood bags and apheresis collection kits adapting their products to meet the circumstances. In Italy, the legislation on transfusion matters was recently integrated by Legislative Decree 208 of November 9, 2007⁵³; this decree transposes Commission Directive 2005/62/EC of 30 September 2005, implementing Directive 2002/98/EC of the European Parliament and the Council as regards Community standards and specifications relating to a quality system for blood establishments, and sets out that “the procedures for collecting blood must minimise the risk of microbial contamination”.

The frequency of bacterial contamination of platelets is influenced by the technique used to produce them⁵¹; the rate of contamination of platelets from single donors or from apheresis is significantly lower than that of PCs produced from pools. The predominant source of micro-organisms responsible for the contamination of PCs is the skin flora^51,54. Most deaths are, however, caused by the transfusion of blood components contaminated by Gram-negative bacteria²³, predominantly of intestinal origin, responsible for asymptomatic bacteraemia in the donor, on which diversion, obviously, is not effective. The frequency of donations made during asymptomatic bacteraemia can be reduced by rigorously applying the selection criteria set out in the protocols for determining the suitability of donors of blood and blood components⁵⁵ and, during the selection process, encouraging donors to report any changes of health that occur in the period immediately after the donation⁵⁴.

Although the effect of diversion is limited to micro-organisms originating from the skin, such as coagulase-negative staphylococci, it is hoped that this strategy is applied extensively to both the collection of whole blood and apheresis collections^41,56,57.

In France⁵², a statistically significant reduction in the rate of bacterial contamination of blood components was achieved through the combined action of a series of strategies aimed at preventing or reducing the initial contamination of the blood component (better skin disinfection techniques, care with regards to the personal hygiene of the staff collecting the donations, diversion, active reporting by donors of any changes in health after the donation), and the proliferation of microbes within it (meticulous control of the storage temperature of blood components, universal leucoreduction, separation of whole blood after 2–20 hours, transfusion of platelets stored for less than 3 days), as well as through constant improvement in scientific knowledge on the subject (production of guidelines on the management of post-transfusion bacterial infections and promotion of specific research on the subject). The implementation of these preventive measures at various stages of the production process led, at a national level, to a significant reduction in the risk of post-transfusion bacterial infection and provided results equivalent to those of other countries that privileged the use of diagnostic tests to detect bacterial contamination. The limitation of such diagnostic tests is the considerable percentage (about 50%) of false negative cultures at 24–48 hours^52,58,59; these results are responsible for an estimated global residual risk of septic reactions of about 1/45,000 transfusions (2.2/100,000) and led, on January 29, 2008, to the premature interruption of the phase IV study, “PASSPORT” (“Post-Approval Surveillance Study of Platelet Outcomes, Release Tested”)⁶⁰, the outcome of an agreement between Gambro BCT (Lakewood, CO, USA) and Fenwal Inc. (Lake Zurich, IL, USA). “PASSPORT” had the aim of validating the clinical use of platelets from apheresis produced by cell separators from two companies [Trima Accel, COBE Spectra (Gambro BCT) and AMICUS (Fenwal)], stored for up to 7 days and transfused after having been found to be negative for bacterial contamination according to a screening test (BacT/ALERT - bioMérieux SA, Marcy l’Etoile, France).

In conclusion, a global strategy of minimising post-transfusion bacterial infections must include the adoption of multiple preventive measures, such as those analysed in an article by George Andreau⁵², as well as diagnostic tests and could also exploit pathogen inactivation methods, but remains dependent on appropriate use of transfusion therapy.