the effect of prestorage wbc reduction on the rates of febrile nonhemolytic transfusion reactions to...
TRANSCRIPT
10 TRANSFUSION
Volume 44, January 2004
Blackwell Science, LtdOxford, UKTRFTransfusion0041-11322003 American Association of Blood BanksJanuary 20044410Original Article
WBC REDUCTION REDUCES FNHTR TO PC AND RBCYAZER ET AL.
ABBREVIATIONS:
CHTS = Capital Health Transfusion Service;
FNHTR = febrile non-hemolytic transfusion reactions; PC =
platelet concentrate; PoUR = post-universal WBC reduction
group; Pr/PoUR = post-universal WBC reduction of PCs but pre-
universal WBC reduction of RBCs group; PrUR = pre-universal
WBC reduction group.
From the Department of Laboratory Medicine and Pathology,
University of Alberta; and Capital Health Transfusion Service,
Edmonton, Alberta, Canada.
Address reprint requests to:
Mark Yazer, MD, 4B1.34 Walter
Mackenzie Health Sciences Center, 8440-112 Street, Edmonton
Alberta T6G 2B7 Canada; e-mail: [email protected].
All figures in CDN Dollars. Cost per double collection bag
for RBC without filter = $6.20, quad collection bag for PC without
filter = $8.54, average = $7.37. Cost per bag for both RBC and PC
collections with in-line filter = $42. Average additional cost per
RBC and PC collection with in-line filter = $35. Cost of prestorage
WBC reduction per FNHTR prevented = [(# units transfused/
month
¥
average additional cost) / # FNHTR prevented by
prestorage WBC reduction]. Collection bag costs provided by
personal communication with Pall Medical, May 2003.
Received for publication March 18, 2003; revision received
May 28, 2003, and accepted May 29, 2003.
TRANSFUSION
2004;44:10-15.
T R A N S F U S I O N P R A C T I C E
The effect of prestorage WBC reduction on the rates of febrile nonhemolytic transfusion reactions to platelet concentrates
and RBC
Mark H. Yazer, Linda Podlosky, Gwen Clarke, and Susan M. Nahirniak
BACKGROUND:
Febrile non-hemolytic transfusion reactions (FNHTRs) are a common complication of platelet concentrate (PC) and RBC transfusions, usually ascribed to cytokines released by WBCs and perhaps the platelets themselves during storage. Prestorage WBC reduction should abrogate the accumulation of these cytokines reducing the number of FNHTRs.
STUDY DESIGN AND METHODS:
A retrospective analysis of FNHTR to PCs and RBCs before universal WBC reduction (PrUR) (July 1997-January 1998 for PCs, July 1997-July 1999 for RBCs) and after its introduction (PoUR) (February 1998-August 2001 for PC, August 1999-August 2001 for RBCs) was undertaken. All transfusion reactions were stratified based on component and date of reaction. Other adverse transfusion reactions were grouped into three periods: July 1997-January 1998, February 1998-July 1999, and August 1999-August 2001. A chi-square test was performed to determine the significance of the differences between groups.
RESULTS:
In the PrUR group, there were 231 FNHTRs in 70,396 RBC units transfused (0.33%) and 29 FNHTRs in 6502 PC units transfused (0.45% percent). In the PoUR group, there were 136 FNHTRs in 72,949 RBC units transfused (0.19%, p
<
0.001) and 56 FNHTRs in 50,555 PC units transfused (0.11%, p
<
0.001). Of the other adverse events, only TRALI reactions were sig-nificantly reduced.
CONCLUSION:
Prestorage WBC reduction significantly reduced the rate of FNHTRs to PCs and RBCs.
ransfusion support with platelet concentrates(PCs) and RBCs are an important adjunct in thetreatment of various malignancies, in surgicalpractice, and in the setting of BMT. Febrile
non-hemolytic transfusion reactions (FNHTR) are adverseevents related to the transfusion of allogeneic blood com-ponents, characterized by a rise in pretransfusion temper-ature of at least 1
∞
C without any other explanation andmay be accompanied by other signs of inflammationincluding rigors and chills.
1
The rate of FNHTR dependson the component. Published incidence rates rangefrom 0.12
2
to 0.5 percent
3
for non-WBC-reduced RBCsto between 1.7
4
and 31 percent
5
for non-WBC-reducedplatelets.
WBCs are considered an important cause of FNHTRfor at least two reasons: in stored PCs, the rate of WBC-
T
WBC REDUCTION REDUCES FNHTR TO PC AND RBC
Volume 44, January 2004
TRANSFUSION 11
derived pro-inflammatory cytokines increases withstorage
6-9
due to the active synthesis of cytokines
8
by thesecells. Transfused WBCs themselves may also be the targetsof preformed antibodies in the recipient, which can stim-ulate the production and release of cytokines. Therefore,a reduction in the number of WBCs to less than 5
¥
10
6
would be expected to mitigate these effects. Third-generation WBC-reduction filters have been shown toachieve a three-log reduction of WBCs without affectingthe metabolic or antigenic profile of the platelets.
6
The accumulation during storage of soluble cytokinesinvolved in mediating FNHTRs to transfused PCs was con-vincingly demonstrated by Heddle et al.
10
These experi-ments showed greater frequency and severity of febrilereactions to the plasma portion of poststorage WBC-reduced platelets compared to transfusion of plasma-reduced platelets. Because there were a small number ofresidual cases of FNHTR to the platelets themselves, animmune mechanism was hypothesized to be responsiblefor mediating some reactions. Another explanation for theresidual FNHTRs to PCs is the accumulation of platelet-derived pro-inflammatory cytokines. Soluble CD40 ligand(CD154) has been shown to accumulate during storage ofPC units and, at least in cultured human lung fibroblasts,induces the production of PGE
2
, a known mediator offever.
11
New data has demonstrated that when the storage-accumulated cytokines in poststorage WBC-reduced PCsare removed by washing, the rate of FNHTR is significantlylower compared to unwashed poststorage WBC-reducedPC.
12
Taken together, these data suggest a crucial rolefor soluble mediators of fever, which accumulate duringplatelet storage.
A recent study demonstrated a trend toward loweroverall and severe reaction rates to prestorage WBC-reduced platelets (both apheresis units and those preparedfrom random donor units) compared to platelets thatunderwent plasma reduction immediately before transfu-sion, although the differences between the two groups wasnot significant.
13
Interestingly, it appears that children alsoappear to benefit from prestorage WBC-reduced platelets,achieving low rates of FNHTRs similar to adults receivingthe same components.
14
In contrast to the accumulationof cytokines in stored platelets, levels of cytokines in RBCunits that have undergone prestorage WBC reduction donot increase during storage,
15
nor have increased cytokinesbeen implicated in FNHTRs,
16
suggesting a different mech-anism is responsible for FNHTRs to this component.Despite the lack of change in cytokine levels during stor-age, prestorage WBC-reduced RBCs have been associatedwith lower rates of FNHTR compared to poststorage WBC-reduced RBC units.
17
Poststorage WBC reduction has beenshown to be equally effective in removing WBCs comparedto prestorage WBC reduction and may even have the addedbenefit of removing certain activated complement frag-ments,
6
although it cannot abrogate the accumulation of
WBC-derived cytokines during storage. Conversely, plate-lets that have undergone prestorage WBC reduction havebeen shown to contain low levels of pro-inflammatorycytokines throughout storage
7
and have been associatedwith lower rates of FNHTR.
18
To examine the impact of universal prestorage WBCreduction, Dzik et al.
19
performed a large-scale random-ized study comparing outcomes in patients receivingprestorage WBC-reduced RBCs (and process WBC-reduced apheresis platelets) to patients receiving non-modified components. In this study, no difference in thelength of hospital stay, in-hospital mortality, and total hos-pital costs between these two groups was observed, norwere there any significant differences in secondary mea-sures of transfusion outcome. In terms of FNHTRs, a trendtoward reduced FNHTRs was observed in the group ofpatients who received WBC-reduced components,although it did not reach significance. This observationsuggests an important role for WBCs in mediating thesereactions and adds credence to our practice of prestorageWBC reduction in reducing FNHTRs.
In February 1998, prestorage WBC reduction of allplatelet components began in Canada. Universal WBCreduction for all blood components was instituted in July1999. Reduction of FNHTRs was cited as one of the rea-sons for including this processing step;
20
other consider-ations included beneficial effects on CMV transmissionand HLA alloimmunization rates.
20
We have performed aretrospective study comparing the rates of FNHTR before(PrUR: before universal WBC reduction) and after (PoUR:after universal WBC reduction) the introduction ofprestorage WBC reduction to both RBCs and platelets totest the hypothesis that the removal of WBCs before stor-age will both prevent the accumulation of cytokines dur-ing storage and will also remove the antigenic targets forpreformed anti-WBC, leading to a reduction in the num-ber of FNHTRs.
MATERIALS AND METHODS
Hospitals
The Capital Health Transfusion Service (CHTS, Edmonton,Alberta) receives allogeneic blood components, collectedfrom volunteer donors who meet Health Canada eligibilitycriteria from the Canadian Blood Services. In total, sevenhealth care facilities are included in the CHTS supportinga wide range of clinical programs including solid organtransplantation, neonatal, pediatric, and adult intensivecare units. The population is approximately 1.3 millionpeople.
Blood components
Platelets were prepared by the platelet-rich plasmamethod and, beginning in February 1998, filtered through
YAZER ET AL.
12 TRANSFUSION
Volume 44, January 2004
a Pall ATSPL filter (Pall Medical, Ottawa, Canada) asdirected by the manufacturer’s guidelines and stored for amaximum of 5 days on an elliptical shaker at room tem-perature. RBCs were collected into CP2D and stored at 4
∞
Cin AS-3 for a maximum of 42 days. After July 1999, all RBCunits were gravity filtered (Pall RCM1 filter, Pall Medical).
Patients
All patients who suffered a transfusion reaction from July1997 through August 2001 were identified by a searchof the blood bank information system database. Patientswere stratified by date of reaction and by componentreceived. Patients were not stratified by diagnosis, by in-patient or out-patient status at the time of the reaction, orby component modification (e.g., irradiated). The provi-sion of premedication was at the discretion of the orderingphysician.
Evaluation of transfusion reactions
A standard transfusion reaction investigation protocol isfollowed: all recognized transfusion reactions are reportedto the blood bank on a standard reporting form. The clin-ical information provided includes the patient’s pre- andposttransfusion vital signs and any other symptoms notedduring the reaction. Further consultation with the wardstaff, patient, and/or chart reviews are performed toresolve ambiguities. The remaining blood component anda posttransfusion blood sample from the patient are sentto the blood bank where a clerical check is performed inparallel with a DAT on the posttransfusion sample. Visualinspection of the plasma component of the posttransfu-sion sample is performed for evidence of hemolysis. At thediscretion of the blood bank physician, bacterial culturesof the remaining segments, bag, and patient may be per-formed. Once the investigation is complete, results arereviewed by transfusion service physicians. Reactionsare classified as febrile nonhemolytic, delayed hemolytic,acute hemolytic, TRALI, volume overload, septic,
hypotensive, or allergic and/or anaphylactic according tostandard definitions.
1
Reactions may also be classified asunrelated to transfusion.
Statistics
A chi-square test was performed to determine the signifi-cance of the differences between the PrUR and PoURgroups, as well as the differences in the rates of non-PC or RBC febrile reactions in the PrUR, Pr/PoUR (post-WBC reduction of PCs but pre-WBC reduction of RBCs),and PoUR groups.
RESULTS
RBCs
In the PrUR group (July 1997-July 1999), a total of 70,396RBC units were transfused. A total of 486 transfusion reac-tions were reported during that period, of which 231 wereclassified as FNHTR to RBCs (48%). The overall rate ofFNHTR to RBC was 0.33 percent. In the PoUR group(August 1999-August 2001), a total of 72,949 RBC unitswere transfused. A total of 398 transfusion reactions werereported during that period, of which 136 were classifiedas FNHTR to RBCs (34%, p
<
0.001). The overall rate ofFNHTR to RBCs was 0.19 percent (p
<
0.001) (Table 1). Thisrepresents a significant reduction in the rate of FNHTRafter institution of prestorage WBC reduction.
Platelets
In the PrUR group, a total of 6502 platelet units were trans-fused. A total of 165 transfusion reactions were reportedduring that period, of which 29 were classified as FNHTRto PCs (18%). The overall rate of FNHTR was 0.45 percent.In the PoUR group, a total of 50,555 platelet units weretransfused. A total of 695 transfusion reactions werereported during that period, of which 56 were classified asFNHTR to PCs (8%, p
<
0.001). The overall rate of FNHTRto PCs was 0.11 percent (p
<
0.001) (Table 1). This again
TABLE 1. Comparison of the number of RBC and PC units transfused during the PrUR and PoUR time periods and the number of transfusion reactions
PrUR RBC PoUR RBC PrUR PC PoUR PC
Dates July 1997-July 1999
August 1999-August 2001
July 1997-January 1998
February 1998-August 2001
RBCs transfused (number) 70,396 72,949 NA NA
PCs transfused (number) n/a n/a 6502 50,555
Total adverse reactions (number) 486 398 165 695
FNHTRs (RBC or PC) 231 136 29 56
Rate FNHTR (RBC or PC, percent) 0.33 0.19* 0.45 0.11*
Transfusion reactions (percent) 48 34* 18 8*
* p
<
0.001.
WBC REDUCTION REDUCES FNHTR TO PC AND RBC
Volume 44, January 2004
TRANSFUSION 13
represents a significant reduction in the rate of FNHTRafter prestorage WBC reduction of platelets.
Due to the overlapping post-universal WBC-reduction group for PCs and the pre-universal WBC-reduction group for RBCs, the rates of all other transfusionreactions were stratified as follows: pre-WBC reduction ofRBCs and PCs (PrUR, July 1997-January 1998), post-WBCreduction of PCs but pre-WBC reduction of RBCs (Pr/PoUR, February 1998-July 1999), and post-WBC reductionof RBCs and PCs (PoUR, August 1999-August 2001). Theresults are presented in Fig. 1.
DISCUSSION
In our small study, prestorage WBC reduction has effecteda significant decrease in the rate of FNHTRs to both RBC
and PC components. Before WBCreduction, our rates of FNHTRs to RBCswere within the range commonly citedfor this component, and although themagnitude of the decrease in the ratemight be significant, the absolutenumber of patients that benefited fromprestorage WBC reduction was small.The rate of FNHTR to PCs was consid-erably below the reported range.
Before the implementation of uni-versal prestorage WBC reduction, WBC-reduced blood components would havebeen administered at the discretionof the prescribing physician and wouldhave undergone poststorage WBCreduction at the bedside. Commonindications for such componentsincluded multiple and severe febrilereactions, and in any situation, in whichanti-HLA alloimmunization wouldhave negatively affected future care,such as patients requiring BMT. In thePrUR RBC and PC groups, a total of fiveFNHTRs (three to RBCs, two to PCs)were observed in four patients receivingbedside WBC-reduced components,which represents a small number ofreactions compared to the totalnumber of FNHTRs for each compo-nent. Our study is thus significant inthat the decreased rate of FNHTRs toboth PCs and RBCs is attributable tothose patients who would not otherwisehave received WBC-reduced compo-nents and suggests that prestorage WBCreduction can benefit all recipients.
There are several explanations forour observed paucity of FNHTR to PCs.
The CHTS does not have a universal premedication guide-line, which probably leads to widespread use of acetami-nophen and Benadryl in many patients. Although it isunlikely that these measures would mask the fever asso-ciated with bacterial contamination of a blood compo-nent, it is possible that they mask the 1
∞
C rise intemperature required to qualify as a FNHTR.
1
Alterna-tively, if a patient commenced a transfusion without hav-ing received premedication and began to suffer a febrilereaction, it is possible the ward staff may have interruptedthe transfusion to administer anti-inflammatory medica-tion and restarted the transfusion from the original bagwithout consultation with the blood bank.
The current study was a retrospective analysis of pre-recorded data and relied on the pathologists at each of theseven health care facilities of the CHTS to adhere to pub-
Fig. 1. The relative rates of adverse transfusion reactions excluding FNHTR to PCs and
RBCs during the entire study period. Due to the overlap in the PoUR PC group and the
PrUR RBC group, the rates of non-FNHTR to PC and RBC adverse reactions were
divided into three time periods: PrUR (July 1997-January 1998), Pr/PoUR (February
1998-July 1999), and PoUR (August 1999-August 2001). In each of these three groups,
there were some patients who were recorded in the blood bank electronic database as
suffering a transfusion reaction, but the nature of their reaction could not be deter-
mined. The number of such patients is as follows: PrUR (19), Pr/PoUR (17), and PoUR
(33). These patients were not included in the total number of transfusion reactions in
their respective time period. (
�
) PrUR, n = 78; (
�
) Pr/PsUR, n = 119; ( ) PoUR,
n = 235. *p < 0.05.
To
tal a
dve
rse
reac
tio
ns
(exc
lud
ing
FN
HT
R t
o P
Cs
and
RB
Cs)
(%
)
Adverse reaction
overload
VolumeAllergic
components
TRALI Unrelated Hemolytic Febrile
(plasma
only)
Delayed
YAZER ET AL.
14 TRANSFUSION
Volume 44, January 2004
lished criteria when classifying a transfusion reaction.
1
Incomplete clinical information at the time of reactionclassification or ambiguities in the clinical presentation(e.g., onset of fever in a neutropenic patient with a historyof fever, whose current fever commenced shortly afterreceipt of a blood component) may have artificiallyinflated the number of reactions deemed unrelated totransfusion and minimized the actual number of febrilereactions.
Because of the experimental design, we did notincrease the ward staff’s sensitivity toward transfusionreactions nor did we employ a sensitive, patient-centeredadverse-reaction checklist to scrutinize the symptoms ofeach patient who was suffering a transfusion reaction.Thus, the patient may not have ascribed the fever itself orthe other symptoms of inflammation that often accom-pany a FNHTR (if indeed they experienced them at all dueto the premedication) to the blood component, whichwould cause the reaction to pass unrecorded. Thus, thefailure of the patient to declare themselves with a reaction,or the ward staffs’ failure to recognize the blood com-ponent as the cause of the fever, would lead to under-reporting of this phenomenon. This explanation seemssomewhat unlikely in our region as evidenced by thesignificant number of transfusion reaction investigationsthat are performed on samples that are ultimately deemedto be unrelated to the transfusion event, however, the roleof under-reporting in reducing the number of adverseevents cannot be overlooked.
A perennial concern when new technology is intro-duced into a manufacturing process is the cost-benefitratio. On average, the CHTS transfuses approximately3000 RBC units and 1000 PC units per month. At an addi-tional cost of approximately $35 CDN per collection bag(with in-line filters), it costs approximately $35,000 CDNto prevent one FNHTR to RBCs and approximately $10,000CDN to prevent one FNHTR to PCs per month in the Cap-ital Health Region (see footnote). On the surface it wouldappear that the scale is firmly tipped toward excess expen-diture for a minimal improvement in patient care. How-ever, the decision to implement universal prestorage WBCreduction was not taken solely to reduce FNHTR but toalso take advantage of the other, albeit more controversial,benefits of this process, including reductions in CMVtransmission, immunomodulation, and platelet alloim-munization—none of which were evaluated in this study.Time and further research will tell if this investment inpatient care is worthwhile.
Our study demonstrated a significant decrease in therate of FNHTR to both PC and RBC units after the imple-mentation of prestorage WBC reduction. The questionsof cost effectiveness and clinical relevance remain unan-swered and will likely continue to be so until the otherpotential benefits of prestorage WBC reduction have beenmore fully elucidated.
REFERENCES
1. Brecher, ME, ed. Technical manual, 14th ed. Bethesda:
American Association of Blood Banks, 2002.
2. Uhlmann EJ, Isgriggs E, Wallhermfechtel M, Goodnough LT.
Prestorage universal WBC reduction of RBC units does not
affect the incidence of transfusion reactions. Transfusion
2001;41:997-1000.
3. Menitove JE, McElligott MC, Aster RH. Febrile transfusion
reaction: what blood component should be given next? Vox
Sang 1982;42:318-21.
4. Goodnough LT, Riddell J, Lazarus H, et al. Prevalence of
platelet transfusion reactions before and after implementa-
tion of leukocyte-depleted platelet concentrates by filtra-
tion. Vox Sang 1993;65:103-7.
5. Heddle NM, Klama LN, Griffith L, et al. A prospective study
to identify the risk factors associated with acute reactions to
platelet and red cell transfusions. Transfusion 1993;33:794-7.
6. Ferrer F, Rivera J, Corral J, et al. Evaluation of pooled platelet
concentrates using prestorage versus poststorage WBC
reduction: impact of filtration timing. Transfusion 2000;
40:781-8.
7. Aye MT, Palmer DS, Giulivi A, et al. Effect of filtration of
platelet concentrates on the accumulation of cytokines and
platelet release factors during storage. Transfusion 1995;
35:117-24.
8. Muylle L, Joos M, Wouters E, et al. Increased tumor necrosis
factor
a
(TNF
a
), interleukin 1, and interleukin 6 (IL-6) levels
in the plasma of stored platelet concentrates: relationship
between TNF
a
and IL-6 levels and febrile transfusion
reactions. Transfusion 1993;33:195-9.
9. Muylle L, Peetermans ME. Effect of prestorage leukocyte
removal on the cytokine levels in stored platelet concen-
trates. Vox Sang 1994;66:14-7.
10. Heddle NM, Klama L, Meyer R, et al. A randomized
controlled trial comparing plasma removal with white cell
reduction to prevent reactions to platelets. Transfusion
1999;39:231-8.
11. Phipps RP, Kaufman J, Blumberg N. Platelet derived CD154
(CD40 ligand) and febrile responses to transfusion. Lancet
2001;357:2023-4.
12. Vo TD, Cowles J, Heal JM, et al. Platelet washing to prevent
febrile reactions to leucocyte-reduced transfusions. Trans-
fusion Med 2001;11:45-7.
13. Heddle NM, Blajchman MA, Meyer RM, et al. A randomized
controlled trial comparing the frequency of acute reactions
to plasma-removed platelets and prestorage WBC-reduced
platelets. Transfusion 2002;42:556-66.
14. Couban S, Carruthers J, Andreou P, et al. Platelet
transfusions in children: results of a randomized,
prospective, cross-over trial of plasma removal and a
prospective audit of WBC reduction. Transfusion 2002;42:
753-8.
15. Shanwell A, Kristiansson M, Remberger M, et al. Generation
of cytokines in red cell concentrates during storage is
WBC REDUCTION REDUCES FNHTR TO PC AND RBC
Volume 44, January 2004
TRANSFUSION 15
prevented by white cell reduction. Transfusion 1997;37:678-
84.
16. Lin JS, Tzeng CH, Hao TC, et al. Cytokine release in febrile
non-haemolytic red cell transfusion reactions. Vox Sang
2002;82:156-60.
17. Federowicz I, Barrett BB, Andersen JW, et al. Charac-
terization of reactions after transfusion of cellular blood
components that are white cell reduced before storage.
Transfusion 1996;36:21-8.
18. Patterson BJ, Freedman J, Blanchette V, et al. Effect of
premedication guidelines and leukoreduction on the rate
of febrile nonhaemolytic platelet transfusion reactions.
Transfusion Med 2000;10:199-206.
19. Dzik WH, Anderson JK, O’Neill EM, et al. A prospective,
randomized clinical trial of universal WBC reduction.
Transfusion 2002;42:1114-22.
20. Sher GD. Leukoreduction of the blood supply.
Canadian Blood Services. PDF available at
http://www.bloodservices.ca/centreapps/internet/
uw_v502_mainengine.nsf/web/0E98B978BE
790BC585256AB10068D347?OpenDocument.