- Email: [email protected]
Leukocyte filtration and postoperative infections
j o u r n a l o f s u r g i c a l r e s e a r c h o c t o b e r 2 0 1 6 ( 2 0 5 ) 4 9 9 e5 0 9
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.JournalofSurgicalResearch.com
Leukocyte filtration and postoperative infections Seunghyug Kwon, BS, MPH,a Sungyub Lew, MD,a,b and Ronald S. Chamberlain, MD, MPA, FACSa,b,c,* a
Saint George’s University School of Medicine, Grenada, West Indies Department of Surgery, Saint Barnabas Medical Center, Livingston, New Jersey c Department of Surgery, New Jersey Medical School, Rutgers University, Newark, New Jersey b
article info
abstract
Article history:
Background: Leukocyte filtration has been hypothesized to reduce the risk of postoperative
Received 29 January 2016
infections by alleviating the immunosuppressive effect of whole blood. However, the
Received in revised form
literature regarding the clinical efficacy of leukocyte filtration remains conflicted. This
22 April 2016
meta-analysis investigates the impact of allogeneic and autologous leukocyte-filtered
Accepted 9 June 2016
blood transfusions on the incidence of postoperative infections in adult surgical patients.
Available online 27 June 2016
Methods: A comprehensive literature search of PubMed, Google Scholar, and Cochrane Central Registry of Controlled trials (1966-2016) was completed for all published
Keywords:
randomized controlled trials. Postoperative infections under “as-per-protocol” (APP) and
Leukocyte filtration
“intention-to-treat” (ITT), length of stay, and mortality were analyzed.
Postoperative infections
Results: Sixteen randomized controlled trials involving 6586 randomized (ITT) patients
Transfusion
(4615 APP patients) in various clinical settings were evaluated. The leukocyte-filtered blood
Transfusion-related
group demonstrated an overall 26% risk reduction in postoperative infections when
immunomodulation
analyzed by APP (relative risk [RR] ¼ 0.74; 95% confidence interval [CI, 0.60-0.92]; P ¼ 0.007)
White-blood-cell reduction
and a 22% risk reduction when analyzed by ITT (RR ¼ 0.78; 95% CI [0.65-0.94]; P ¼ 0.009).
Meta-analysis
Leukocyte-filtered blood was also associated with a significant reduction in length of stay (standardized difference of mean [SDM] ¼ 0.74; 95% CI [1.32 to 0.15]; P ¼ 0.014) and all-cause mortality (RR ¼ 0.74; 95% CI [0.57-0.95]; P ¼ 0.018). Conclusions: Leukocyte-filtered blood transfusions are associated with significantly lower postoperative infection rates in both the APP and ITT populations. Leukocyte filtration also shortens length of stay and decreases all-cause mortality in surgical patients and should be considered in all surgical patients. ª 2016 Elsevier Inc. All rights reserved.
Introduction The American Red Cross has reported that a total of 15 million red blood cell (RBC) units are transfused annually.1 Surgical patients account for two-thirds of these RBC transfusions.2 Though transfusions are essential in saving lives, multiple studies have reported numerous findings of adverse effects, ranging from the transmission of infectious disease to
deleterious transfusion-related immunomodulation (TRIM).3-5 With modern blood banking technology, the transmission of infectious diseases has been vigilantly managed and concerns are substantially minimized overall.6 Most transfusion-related morbidity and mortality consist of noninfectious complications, the greatest of concern being the immunosuppressive effect of TRIM.7 The immunosuppressive effect of transfusions were first demonstrated by
* Corresponding author. Department of Surgery, Saint Barnabas Medical Center, Rutgers University, New Jersey Medical School (NJMS), 94 Old Short Hills Rd, Livingston, NJ 07039. Tel.: þ1 973 322 5195; fax: þ1 973 322 2471. E-mail address: [email protected] (R.S. Chamberlain). 0022-4804/$ e see front matter ª 2016 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jss.2016.06.055
500
j o u r n a l o f s u r g i c a l r e s e a r c h o c t o b e r 2 0 1 6 ( 2 0 5 ) 4 9 9 e5 0 9
Opelz et al. 8 in a prospective study of 148 cadaver donor transplant recipients in which rejection rates were compared based on blood transfusion status. The result of this study revealed that renal allograft rejection was improved in recipients receiving 10 or more transfusions as opposed to those who did not receive transfusions (66% versus 29%, P < 0.001).8 The author inferred from these findings that blood transfusions result in an immunosuppressive effect of the transfused host’s immune system.8 Historically, the immunosuppressive effect of blood transfusions was an often used means to limit rejection in renal transplants until cyclosporine and other immunosuppressives were introduced.8 TRIM has also been observed in surgical patient groups in which an increased number of postoperative infections are observed in patients receiving RBC transfusions. A prospective observational study by Tartter et al. 9, involving 168 colorectal cancer patients, sought to identify perioperative determinants of infectious complications. Using multivariate analysis, these authors reported that 24 of 168 patients developed infectious complications and observed that blood transfusion was an independent significant risk factor for infection (P ¼ 0.01).9 A meta-analysis of 21 randomized controlled trials (RCTs) involving 131,512 patients by Hill et al. 10 (1992) demonstrated a three-fold increase in postoperative infections in transfused patients compared to nontransfused patients (OR ¼ 3.45; 95% confidence interval [CI, 1.42-15.15]; P < 0.05). The presence of leukocytes in the transfused products appears critical to produce the TRIM effect.11 A prospective study involving 80 orthopedic surgery patients by Innerhofer et al. 12 observed impaired T-cell-mediated immunity in patients who were transfused with leukocyte containing RBCs. These findings were continued by Lee et al. 13 who reported persistent donor leukocytes in humans for up to 1.5 y after blood transfusion. Unsurprisingly, various other benefits associated with leukocyte-filtered blood have also been reported, including prevention of recurrent febrile nonhemolytic transfusion reactions (FNHTR), alloimmunization in transplantation, and transmission of cytomegalovirus.14 European countries (i.e., Germany, the United Kingdom, and the Netherlands) and Canada have implemented a universal leukocyte filtration policy, reaping benefits of decreased FNHTR incidences and cost savings due to decreased length of hospital stay.15-17 Despite implementation in these countries, the association between leukocyte filtration and postoperative infection remains controversial.18-21 The most recent Cochrane review by SimancasRacines et al. 21 included 10 RCTs involving 3557 surgical and nonsurgical patients transfused with allogeneic blood and yielded no difference in infection risks (RR ¼ 0.76; 95% CI [0.58-1.00]; P ¼ 0.05) when leukocyte-filtered blood was used. Bilgin et al. 22 conducted an RCT involving 496 patients undergoing valve surgery and reported lower postoperative infection rates (33.8% versus 24.3%, P ¼ 0.032) with the use of leukocyte-filtered blood. Conversely, Nathens et al. 23 conducted an RCT involving 324 trauma patients and reported no difference in infection rates (36.0% versus 30.3%, P ¼ 0.321) with the use of leukocyte-filtered blood.
The current meta-analysis provides a comprehensive review of all published studies in surgical patients, in which leukocyte-filtered blood transfusions were provided to determine the impact on postoperative outcomes, including postoperative infection rates, length of stay, and mortality.
Materials and methods Study selection A comprehensive search of all published RCTs comparing patients receiving leukocyte-filtered and nonleukocytefiltered blood transfusions was conducted using PubMed, Google Scholar, Cochrane Central Registry of Controlled Trials (1966-2016). Additional citations were searched using references retrieved from prior publications (Fig. 1). The last search was conducted on January 10, 2016, and only articles conducted in English were considered. Keywords searched included combinations of “leuk(c)oreduced,” “leuk(c)odepleted,” “leuk(c)ocyte filtered,” “white cell reduced,” “leuk(c) ocyte reduced,” “leuk(c)ocyte depleted,” “leuk(c)ocyte depleting,” and “transfusions.” The inclusion criteria were limited to RCTs in adult surgical populations (>18 y), comparing leukocyte-filtered and nonleukocyte-filtered blood, and reporting the incidence of postoperative infections (surgical and nonsurgical site infections) with sample size. Studies which did not include postoperative infections as an outcome were excluded. This meta-analysis was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement.24
Data extraction Articles retrieved from searches were assessed for eligibility and data pertaining to patients, interventions, comparison groups, outcomes, and methodology were abstracted. The primary clinical outcome of interest was the incidence of postoperative infection. Secondary outcomes were hospital length of stay and all-cause mortality.
Statistical analysis For each trial, RR with a 95% CI for the incidence of postoperative infections and mortality was calculated. SDM with 95% CI was calculated for LOS. Meta-analysis of the pooled data was performed using Comparative Meta-analysis software version 3 (CMA v 3; Biostat, Englewood, NJ). For individual studies reporting zero events in any group, a continuity correction factor of 0.5 was adopted to calculate the RR and variance. Both fixed-effect and random-effect models were considered, depending on the heterogeneity of the included studies. To assess the heterogeneity between studies, both Cochrane’s Q statistic and I2 statistic were used. Heterogeneity was considered statistically significant when P < 0.05 or I2 > 50. If heterogeneity was observed, data were analyzed using a random-effect model. In the absence of heterogeneity, a fixed-effect model was assumed. For all outcomes, publication bias was the first qualitatively assessed using a funnel plot, and further quantitatively
k w o n e t a l l e u k o c y t e - fi l t e r e d b l o o d a n d p o s t o p e r a t i v e i n f e c t i o n s
501
Fig. 1 e Consolidated Standards of Reporting Trials diagram of the study selection process.
assessed with Egger’s and Begg’s tests. A two-tailed P < 0.05 was considered statistically significant. The presence of publication bias was further evaluated using Duval and Tweedie’s Trim and Fill to account for any missing studies. Subgroup analysis was performed based on leukocyte filtration before versus after storage, blood product type for control arm, allogeneic versus autologous blood source, clinical setting, and location of trial were performed to assess the variability within the trials.
Results Demographic characteristics of the studies A total of 16 RCTs were identified involving 4615 patients in the “as-per-protocol” (APP) population (studies published between 1992 and 2009).22,23,25-38 The leukocyte-filtered blood group included 2325 patients and the nonleukocyte-filtered blood group consisted of 2290 patients. The two major clinical settings were cardiac surgery (n ¼ 1949) and colorectal surgery (n ¼ 981). Additional clinical settings included trauma, orthopedic surgery, cancer surgery, and a combination of settings involving a total of 1685 patients (Table 1).
Postoperative infections: as-per-protocol Data on the incidence of postoperative infection were reported in all 16 RCTs involving 4615 patients (2325 patients receiving leukocyte-filtered blood and the 2290 patients
receiving nonleukocyte-filtered blood). Thirteen of the 16 trials reported a lower risk of postoperative infections with patients receiving leukocyte-filtered blood; 5 of which were statistically significant. A lower proportion of patients in the leukocyte-filtered group developed postoperative infections compared to the nonleukocyte-filtered blood (500/2325 [22%] versus 626/2290 [27%]). There was significant heterogeneity between trials (P < 0.01, I2 ¼ 70), and a random-effect model was used. Meta-analysis revealed a significantly lower risk of postoperative infections in the leukocyte-filtered group compared to the nonleukocyte-filtered group (RR ¼ 0.743; 95% CI [0.598-0.924]; P ¼ 0.007; Fig. 2). Subgroup analysis demonstrated that leukocyte-filtered blood had the greatest benefit for leukocyte filtration after storage or in allogeneic blood (RR ¼ 0.349; 95% CI [0.150-0.815]; P ¼ 0.015 and RR ¼ 0.716; 95% CI [0.562-0.912]; P ¼ 0.007, respectively). Significant benefits were also observed in the cardiac surgery setting and trials conducted in Europe (RR ¼ 0.738; 95% CI [0.616-0.884]; P ¼ 0.001 and RR ¼ 0.767; 95% CI [0.593-0.991]; P ¼ 0.042, respectively; Table 2).
Postoperative infections: intention-to-treat Data on the incidence of postoperative infections were reported in all 16 RCTs involving 6586 patients (3326 patients receiving leukocyte-filtered blood and the 3260 patients receiving nonleukocyte-filtered blood). Twelve studies reported a lower risk of postoperative infections with patients receiving leukocyte-filtered blood, 4 of which were statistically significant.22,23,25,26,28,31-36,38 A lower proportion of
502
j o u r n a l o f s u r g i c a l r e s e a r c h o c t o b e r 2 0 1 6 ( 2 0 5 ) 4 9 9 e5 0 9
Table 1 e Characteristics of all included randomized control trials of leukocyte-filtered blood transfusions (1966-2016). Author, y
Jensen, 1992
Clinical setting
Double blind
Multicenter
Leukocytefiltered arm (filter time)
Nonleukocytefiltered arm
Randomized patients (#)
Transfused patients (#)
Colorectal
N
N
Allo WB (A)
Allo WB
197
104
Colorectal cancer
N
Y
Allo RBCs (B)
BR allo RBCs
697
446
Cardiac
N
N
Allo RBCs (B n ¼ 305 or an n ¼ 303)
BR allo RBCs
914
861
Tartter, 1998
GI surgery
N
N
Allo WB (A)
Allo RBCs
221
59
Titlestad, 2001
Houbiers, 1994 van de Watering, 1998
Colorectal
Y
N
Allo RBCs (B)
BR allo RBCs
279
112
Wallis, 2002
Cardiac
N
N
Allo RBCs (B)
BR (n ¼ 204) or PR (n ¼ 198) allo RBCs
597
509
Bilgin, 2004
Cardiac
Y
Y
Allo RBCs (B)
BR allo RBCs
474
430
Nathens, 2006
Trauma
Y
N
Allo RBCs (B)
Allo RBCs
324
268
Lapierre, 2007
Cancer
N
N
Allo RBCs (B)
Allo RBCs
37
35
Frietsch, 2008
Total hip arthroplasty
Y
Y
Auto WB (B)
Auto WB
951
481
Colorectal
N
N
Allo WB (A)
BR allo RBCs
586
260
Multiple
Y
Y
Allo RBCs (B)
BR allo RBCs
1051
479
Efstathiou, 2003
Cardiac
N
N
Auto WB (A)
Auto WB
80
80
Nielsen, 1999
Trauma
N
N
Allo WB (B)
Allo WB
24
24
Cardiac
N
N
Allo WB (A)
Allo WB
104
69
Orthopedic
Y
N
Auto WB (B)
Auto WB
400
398
Jensen, 1996 van Hilten, 2004
Connery, 2005 Tasaki, 2009
A ¼ after; allo ¼ allogeneic; auto ¼ autologous; B ¼ before; BR ¼ buffy-coat reduced; N ¼ no; PR ¼ plasma reduced; RBC ¼ red blood cell; WB ¼ whole blood; Y ¼ yes.
patients in the leukocyte-filtered group developed postoperative infections compared to the nonleukocyte-filtered group (652/3326 [20%] versus 764/3260 [23%]). There was significant heterogeneity between trials (P < 0.01, I2 ¼ 64), and
a random-effect model was assumed. Meta-analysis revealed a significantly lower risk of postoperative infections in the leukocyte-filtered group compared to the nonleukocytefiltered group (RR ¼ 0.783; 95% CI [0.652-0.940]; P ¼ 0.009; Fig. 3).
Fig. 2 e Forest plot evaluating the risk of postoperative infections with the use of leukocyte-filtered blood (APP population). (Color version of figure is available online.)
Subgroup
Filter time
NLF type
Categories of subgroup
Before
ITT
APP
ITT
APP
ITT
APP
ITT
0.032
0.020
0.819
0.896
0.207
0.167
0.036
0.027
0.805
0.661
5444
0.893 (0.781-1.021)
0.904 (0.805-1.016)
0.097
0.091
1448
0.349 (0.150-0.815)
0.466 (0.269-0.806)
0.015
0.006
Buffy-coat reduced
722,25-28,33,37
2939
4400
0.760 (0.557-1.037)
0.812 (0.629-1.048)
0.083
0.110
Whole
629,31,32,35,36
1156
1626
0.649 (0.361-1.165)
0.656 (0.388-1.109)
0.147
0.116
362
362
0.796 (0.483-1.311)
0.796 (0.483-1.311)
0.370
0.370
334
393
0.598 (0.365-0.982)
0.725 (0.452-1.165)
0.042
0.184
959
1429
0.987 (0.639-1.527)
0.969 (0.745-1.261)
0.954
0.814
3656
5157
0.716 (0.562-0.912)
0.765 (0.621-0.942)
0.007
0.012
Autologous
Cardiac
Other (trauma, orthopedic, multiple) European USA Japan
323,30,34 1
26
329,31,35 1322,23,25-28,30,32-34,36-38 522,26,28,35,36
1949
2134
0.738 (0.616-0.884)
0.747 (0.624-0.894)
0.001
0.001
981
1725
0.447 (0.199-1.006)
0.563 (0.313-1.011)
0.052
0.054
1685
2727
0.954 (0.810-1.123)
0.976 (0.842-1.130)
0.571
0.743
3821
5792
0.767 (0.593-0.991)
0.811 (0.658-1.000)
0.042
0.050
323,34,36
396
396
0.642 (0.384-0.918)
0.641 (0.384-1.070)
0.089
0.089
131
398
398
0.656 (0.260-1.657)
0.656 (0.260-1.657)
0.372
0.372
525,32-34,37 623,27,29-31,38 1222,25-30,32,33,35,37,38
Difference between group P value
APP
1095
Colorectal
Country trial conducted
P value
3824
Allogeneic Clinical setting
Summary relative risk (95% confidence interval)
528,32,33,35,36
Plasma reduced
12
22,23,25-31,34,37,38
Number of patients
After
Red blood cells
Blood type
Number of studies (references)
APP ¼ as-per-protocol; ITT ¼ intention-to-treat; LF ¼ leukocyte filtration; NLF ¼ nonleukocyte filtered.
k w o n e t a l l e u k o c y t e - fi l t e r e d b l o o d a n d p o s t o p e r a t i v e i n f e c t i o n s
Table 2 e Subgroup analysis of leukocyte-filtered blood transfused groups on postoperative infections.
503
504
j o u r n a l o f s u r g i c a l r e s e a r c h o c t o b e r 2 0 1 6 ( 2 0 5 ) 4 9 9 e5 0 9
Fig. 3 e Forest plot evaluating the risk of postoperative infections with the use of leukocyte-filtered blood (intention-to-treat population). (Color version of figure is available online.)
Subgroup analysis demonstrated that leukocyte-filtered blood had the greatest benefit for leukocyte filtration after storage or allogeneic blood (RR ¼ 0.466; 95% CI [0.269-0.806]; P ¼ 0.006 and RR ¼ 0.765; 95% CI [0.621-0.942]; P ¼ 0.012, respectively). Significant benefits were also observed in the cardiac surgery setting (RR ¼ 0.747; 95% CI [0.624-0.894]; P ¼ 0.001; Table 2).
Length of stay Data on the length of stay were reported in 3 of the 16 RCTs involving 630 patients in the APP population (333 patients receiving leukocyte-filtered blood and the 297 patients receiving nonleukocyte-filtered blood).29,35,36 All studies reported shorter length of stay with leukocyte-filtered blood, however, only two reached statistical significance.29,35 There was significant heterogeneity between trials (P ¼ 0.001, I2 ¼ 86), and a random-effect model was assumed. Meta-analysis revealed a significant difference in mean length of stay between the leukocyte-filtered group and the nonleukocyte-filtered group (SDM ¼ 0.737; 95% CI [1.324 to 0.150]; P ¼ 0.014; Fig. 4).
All-cause mortality Data on the incidence of all-cause mortality were reported in 9 of the 16 RCTs for the APP population involving 2544 patients (1375 patients receiving leukocyte-filtered blood and the 1169 patients receiving nonleukocyte-filtered blood).22,23,25,27,28,30,33,36,38 Six studies reported a decrease in all-cause mortality with leukocyte-filtered blood, however, only one reached statistical significance.22,25,27,28,36,38 A lower
proportion of patients in the leukocyte-filtered group reported all-cause mortality compared to the nonleukocyte-filtered blood (95/1357 [7.0%] versus 121/1151 [11%]). There was no significant heterogeneity between trials (P ¼ 0.261, I2 ¼ 20.4), and a fixed-effect model was assumed. Meta-analysis revealed a significantly lower risk of death in the leukocytefiltered group compared to the nonleukocyte-filtered group (RR ¼ 0.735; 95% CI [0.570-0.949]; P ¼ 0.018; Fig. 5).
Subgroup analysis In the APP population, significant differences were observed among postoperative infection rates related to leukocyte filtration before or after storage (P ¼ 0.032) and in various clinical settings (P ¼ 0.036). No difference was observed for blood product type in the control arm (P ¼ 0.819), allogeneic versus autologous (P ¼ 0.207), and country in which the trial was conducted (P ¼ 0.805; Table 2). In the intention-to-treat population, significant differences were observed among postoperative infection rates related to leukocyte filtration before or after storage (P ¼ 0.02) and clinical setting (P ¼ 0.027). No difference was also observed for blood product type for control arm (P ¼ 0.896), allogeneic versus autologous (P ¼ 0.167), and country where trial was conducted (P ¼ 0.661; Table 2).
Publication bias Negative findings are less likely to be published, and consequently, over or underestimation of the outcome may occur.39 A qualitative and quantitative assessment of the publication
k w o n e t a l l e u k o c y t e - fi l t e r e d b l o o d a n d p o s t o p e r a t i v e i n f e c t i o n s
505
Fig. 4 e Forest plot evaluating the SDM for length of stay with the use of leukocyte-filtered blood (APP population). (Color version of figure is available online.)
bias was performed using a funnel plot and both Egger’s and Begg’s tests.40 Qualitative evidence of asymmetry was observed on the funnel plot (Fig. 6). While Begg’s test was nonsignificant (P ¼ 0.053), Egger’s test was suspected for publication bias, P ¼ 0.038. Duval and Tweedie’s method of Trim and Fill based on the random effect model and searching for studies on the right of the mean effect estimated 5 unpublished negative studies.41 The point estimate and 95% CI for the combined studies was 0.846 (0.764-0.947). Using Trim and Fill, the imputed point estimate was 0.911 (0.8251.006). After adjusting for countries with universal leukocyte filtration, publication bias was not observed by either Begg’s (P ¼ 0.07) or Egger’s tests (P ¼ 0.09).
Discussion In 2010, the Centers for Disease and Control estimated 51.4 million surgeries were performed in the United States (US).42 Approximately two million of these surgeries will require a
transfusion, averaging three transfusions per surgery.15 Although utilized for many reasons based on the belief that transfusion improves oxygen transport, blood transfusions also carry substantial adverse risks.43 The risks include a four-fold increase in infection, a one in 5000 chance of fatal transfusion-related acute lung injury, and an approximately 10% occurrence of autoantibodies.44 Leukocyte filtration has been hypothesized to decrease these complications and has been advocated as the standard for all blood transfusions; however, the efficacy of leukocyte filtration to alleviate the TRIM effect remains controversial.14,45 Several mechanisms have been proposed to explain the higher incidences of the noninfectious complications in transfused recipients, with TRIM being the most frequently supported.5 Soluble and cellular factors in leukocytes have been proposed as the primary mediator of the TRIM effect.46 The immunogenicity of cellular or soluble antigens associated with transfused blood components depends on the viability of donor dendritic antigen presenting cells and costimulatory signals for the presentation of the donor
Fig. 5 e Forest plot evaluating all-cause mortality with the use of leukocyte-filtered blood (APP population). (Color version of figure is available online.)
506
j o u r n a l o f s u r g i c a l r e s e a r c h o c t o b e r 2 0 1 6 ( 2 0 5 ) 4 9 9 e5 0 9
Fig. 6 e Funnel plot assessing publication bias (analyzing the effect of leukocyte-filtered blood on postoperative infection incidence in surgical patients).
antigens to recipient’s T-cells.5 When a multitude of antigens are presented, T-cell unresponsiveness and immune tolerance ensues due to either nonviable antigen presenting cells or absent costimulatory signals.47 Ottonello et al. 48 demonstrated the immunosuppression effect by comparing neutrophil function in postenonleukocyte-filtered RBC transfusion blood samples from eight healthy volunteers. After RBC transfusion, neutrophils in the recipient synthesized transforming growth factor b1, a cytokine exerting an overall immunosuppressive activity, which in turned caused an inhibition of neutrophil chemotaxis that lasted 8-15 d.48 Additional studies by Onorati et al. 49 examined proinflammatory cytokine release in 60 chronic obstructive pulmonary disease patients undergoing aortic valve surgery by randomly allocating 30 to a leukocyte-filtered group and the other to a nonleukocyte-filtered group. The leukocytefiltered group had a significant reduction in various proinflammatory markers, such as interleukin (IL)-6, IL-8, IL10, and tumor necrosis factor-a (P < 0.01).49 The postoperative infection risk reported in this meta-analysis is consistent with the results from previous studies when analyzed by APP. Fergusson et al. 20 reported a meta-analysis of 10 RCTs involving 2358 cardiac and colorectal surgery patients transfused with allogeneic blood and similarly resulted a significantly lower risk of postoperative infections (RR ¼ 0.60; 95% CI [0.38-0.93]; P < 0.05) with leukocyte-filtered blood in those transfused. Blumberg et al. 18 reported on 9 RCTs involving 3093 mostly cardiac and colorectal surgery patients who were transfused allogeneic blood and also noted lower postoperative infection risks (OR ¼ 0.522; 95% CI [0.332-0.821]; P < 0.05) with the use of leukocyte-filtered blood. In contrast, the most recent Cochrane review by Simancas-Racines et al. 21 involving 10 RCTs involving 3557 surgical and nonsurgical patients
transfused with allogeneic blood yielded no difference in infection risks (RR ¼ 0.76; 95% CI [0.58-1.00]; P ¼ 0.05) with the use of leukocyte-filtered blood. The difference between the conclusions of this report with Cochrane’s result may be explained in part by the inclusion or exclusion criteria and population analyzed. The Cochrane review excluded studies using autologous blood, patients receiving other blood components, and patients receiving RBCs with no indications.29,31,32,35-38 Five of 7 of these trials favored leukocyte-filtered blood.31,32,35,36,38 The current study was limited to surgical patients due to the lack of nonsurgical patient trials on infection outcomes. In addition, the Cochrane review included only one nonsurgical patient trial which involved HIV patients in their analysis, whose results may have limited generalizability.50 Collier et al. (2001) hypothesized that the compromised immune system of the population may have reacted differently to leukocytes, thus making any conclusion beyond HIV patient seem inappropriate.50 Finally, when we analyzed the Cochrane review data by excluding the nonsurgical trial, the results were similar to the current analysis (RR ¼ 0.71; 95% CI [0.53-0.95]; P ¼ 0.022). The Cochrane review also analyzed all-cause mortality reported in 10 RCTs involving 4060 surgical and nonsurgical patients transfused with allogeneic blood and seemingly noted no difference (RR ¼ 0.80; 95% CI [0.60-1.07]; P ¼ 0.13) with the use of leukocyte-filtered blood.21 Fever was the only outcome that was significantly decreased (RR ¼ 0.75; 95% CI [0.60-0.94], P ¼ 0.01) in the leukocyte-filtered blood group; however, only two RCTs involving 634 cardiac and cancer surgery patients experienced this outcome.51,21 Leukocyte-filtered blood may not benefit all patients, as Collier et al. (2001) have reported leukocyte filtration yielded no benefit in HIV patients, and thus the universal
k w o n e t a l l e u k o c y t e - fi l t e r e d b l o o d a n d p o s t o p e r a t i v e i n f e c t i o n s
implementation of leukocyte filtration is not warranted.50 Leukocyte filtration is estimated to cost an additional $20-30 US per unit of blood, which will add an additional $180 US million to the United States hospital budget.15 And yet in appropriate populations where leukocyte filtration is associated with lower hospital costs such as the surgical patients in this meta-analysis, the cost may well be justified.50-54 In an RCT by Tartter et al. 34, lower average hospital charges were reported in gastrointestinal surgery patients receiving filtered blood compared to nonfiltered blood ($33,954 US versus $41,002 US, respectively). While this meta-analysis focused primarily on postoperative infection risks, there are other potential benefits in using a universal leukocyte-filtered blood supply.16,17,55 In patients requiring a blood transfusion, leukocyte filtration has also been shown to decrease FNHTR. Paglino et al. 17 compared the 3-y incidence rate of FNHTR in RBC units transfused before and after the implementation of universal leukocyte filtration (85,321 RBC units prior and 60,048 transfused after) and reported FNHTR rates reduced by approximately 50% (0.35%-0.18%, P < 0.001). Similarly, Yazer et al. 16 compared the 3-y incidence rate of FNHTR in RBC units transfused before and after the implementation of universal leukocyte filtration (70,396 RBC units transfused prior and 72,949 transfused after) and reported significantly lower incidence rates of FNHTR following the implementation of universal leukocyte filtration (0.19% versus 0.33%, P < 0.001). Leukocyte filtration also has the potential to decrease transfusion-associated graft-versushost-disease (TA-GVHD).56 Williamson et al. (2007) used the United Kingdom hemovigilance scheme Serious Hazards of Transfusion database and compared the number of TA-GVHD reports among patients receiving RBC and platelet transfusions over a 4-y period before and after the implementation of universal leukocyte filtration (8407 RBCs and platelets prior and 17,384 transfused after). These authors reported a significant decrease in TA-GVHD following implementation of universal leukocyte filtration (11 versus 2 reports, P < 0.001).55 Although the results of this meta-analysis are significant, there are limitations to this study due to the variation and heterogeneity of the RCTs included. Significant heterogeneity was observed for the filter time and surgery type; thus, the effect of leukocyte filtration may be underestimated. A similar result was also observed with filtration time in the meta-analysis by Fergusson et al.20 That said, outcomes related to filtration timing in animal studies have shown that prestorage filtration is superior to poststorage filtration for the prevention of tumor growth and alloimmunization.57,58 A number of recent retrospective studies have also shown that prestorage leukocyte filtration ameliorates the effects of aging in blood as well.59,60 While the accumulation of toxic mediators in stored blood is a possibility, the clinical manifestation of these mediators remains a controversial topic. Further randomized studies are needed to verify and explain the mechanism of filtration time. In addition, the significant variation in the units transfused and the storage time of the filtered blood, both of which have been thought to be associated with greater adverse effects.23,27,61 Due to the lack of availability of information, the effect of transfusion amount and storage time could not be thoroughly investigated in this analysis. Additional sources of heterogeneity in the
507
analysis include surgeon-specific selection factors, such as a liberal or conservative transfusion policy. The severity of each patient’s illness is an additional significant underlying factor, which directly relates to different proportions of transfused patients and the frequency of postoperative infection. Some studies transfused additional blood components in both trial arms as well.32,38 There is also the concern of selection bias, as not all RCTs were double blind and the details of the randomization process of some RCTs were not well reported. In regard to length of stay data, one trial was converted from median to mean number of days.29,62,63 Mortality for van de Watering et al. 28 was calculated based on the mortality rate presented in the paper not raw numbers; thus, the actual mortality count may differ. At last, some trials were exposed to a greater risk of publication bias due to variation in policies governing implementation of universal leukocyte filtration.23,34,36
Conclusions This study demonstrates that leukocyte-filtered blood is a valuable means to improve postoperative outcomes by reducing the incidence of postoperative infections, especially among cardiac surgery patients, which also resulted in a decreased length of hospital stay and overall mortality. Leukocyte-filtered blood has the potential to play a promising role in limiting transfusion-related postoperative complications for surgical patients, which should translate into a significant decrease in both morbidity and mortality.
Acknowledgment Authors’ contributions: Conception and design was developed by S.K. and R.S.C. Data analysis and interpretation and final approval of manuscript were carried out by S.K., S.L., and R.S.C. Manuscript writing was done by S.K., S.L., and R.S.C.
Disclosure The authors indicated no potential conflicts of interest.
references
1. American Red Cross, Blood facts and Statistics. Available at: http://www.redcrossblood.org/learn-about-blood/bloodfacts-and-statistics. Accessed August 14, 2015. 2. Cuong J, Catlett C, Berenholtz S, Haut E. Change in use of allogeneic red blood cell transfusions among surgical patients over time. J Am Coll Surg. 2009;207:352e359. 3. Day JR, Taylor KM. The systemic inflammatory response syndrome and cardiopulmonary bypass. Int J Surg. 2005;3:129e140. 4. Bilgin YM, van de Watering LM, Brand A. Clinical effects of leucoreduction of blood transfusions. Neth J Med. 2011;69:441e450.
508
j o u r n a l o f s u r g i c a l r e s e a r c h o c t o b e r 2 0 1 6 ( 2 0 5 ) 4 9 9 e5 0 9
5. Vamvakas EC, Blajchman MA. Transfusion-related immunomodulation (TRIM): an update. Blood Rev. 2007;21:327e348. 6. American Red Cross, “History of blood transfusion.” Available at: http://www.redcrossblood.org/learn-about-blood/historyblood-transfusion. Accessed January 8, 2015. 7. Hendrickson JE, Hillyer CD. Noninfectious serious hazards of transfusion. Anesth Analg. 2009;108:759e769. 8. Opelz G, Sengar DP, Mickey MR, Terasaki PI. Effect of blood transfusions on subsequent kidney transplant. Transpl Proc. 1973;5:253e259. 9. Tartter P, Quintero S, Barron D. Perioperative blood transfusion associated with infectious complications after colorectal cancer operations. Am J Surg. 1986;152:479e482. 10. Hill GE, Frawley WH, Griffith KE, Forestner JE, Minei JP. Allogeneic blood transfusion increases the risk of postoperative bacterial infection: a meta-analysis. J Trauma. 2003;54:908e914. 11. Horimi T, Terasaki P, Chia D, Sasaki N. Factors influencing the paradoxical effect of transfusions on kidney transplants. Transplantation. 1983;35:320e323. 12. Innerhofer P, Luz G, Spotl L, et al. Immunologic changes after transfusion of autologous or allogeneic buffy coat-poor versus white cell-reduced blood to patients undergoing arthroplasty. I. Proliferative T-cell responses and the balance of helper and suppressor T cells. Transfus Complicat. 1999;39:1089e1096. 13. Lee TH, Paglieroni T, Ohto H, Holland PV, Busch MP. Survival of donor leukocyte subpopulations in immunocompetent transfusion recipients: frequent long-term microchimerism in severe trauma patients. Blood. 1999;93:3127e3139. 14. American Association of Blood Banks. “CIRCULAR OF information,” Circular of Informationdfor the use of blood and blood components. Bethesda, MD: American Association of Blood Banks; 2013. 15. Lannan KL, Sahler J, Spinelli SL, Phipps RP, Blumberg N. Transfusion immunomodulationethe case for leukoreduced and (perhaps) washed transfusions. Blood Cells. Mol. Dis. 2013;50:61e68. 16. Yazer MH, Podlosky L, Clarke G, Nahirniak SM. The effect of prestorage WBC reduction on the rates of febrile nonhemolytic transfusion reactions to platelet concentrates and RBC. Transfus Pract. 2004;44:10e15. 17. Paglino JC, Pomper GJ, Fisch GS, Champion MH, Snyder EL. Reduction of febrile but not allergic reactions to RBCs and platelets after conversion to universal prestorage leukoreduction. Transfus P. 2004;44:16e24. 18. Blumberg N, Zhao H, Wang H, Messing S, Heal JM, Lyman GH. The intention-to-treat principle in clinical trials and meta-analyses of leukoreduced blood transfusions in surgical patients. Transfusion. 2007;47:573e581. 19. Vamvakas EC. White-blood-cell-containing allogeneic blood transfusion and postoperative infection or mortality: an updated meta-analysis. Vox Sang. 2007;92:224e232. 20. Fergusson D, Khanna MP, Tinmouth A, He´bert PC. Transfusion of leukoreduced red blood cells may decrease postoperative infections: two meta-analyses of randomized controlled trials. Can J Anaesth. 2004;51:417e424. 21. Simancas-Racines D, Osorio D, Martı´-carvajal A, Arevalo-Rodriguez I. Leukoreduction for the prevention of adverse reactions from allogeneic blood transfusion. Cochrane Database Syst Rev. 2015. 22. Bilgin YM, van de Watering LMG, Eijsman L, et al. Double-blind, randomized controlled trial on the effect of leukocyte-depleted erythrocyte transfusions in cardiac valve surgery. Circulation. 2004;109:2755e2760. 23. Nathens AB, Nester TA, Rubenfeld GD, Nirula R, Gernsheimer TB. The effects of leukoreduced blood
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39. 40.
41.
transfusion on infection risk following injury: a randomized controlled trial. Shock. 2006;26:342e347. Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. Int J Surg. 2010;8:336e341. Titlestad IL, Ebbesen LS, Ainsworth AP, Lillevang ST, Qvist N, Georgsen J. Leukocyte-depletion of blood components does not significantly reduce the risk of infectious complications. Results of a double-blinded, randomized study. Int J Colorectal Dis. 2001;16:147e153. Wallis JP, Chapman CE, Orr KE, Clark SC, Forty JR. Effect of WBC reduction of transfused RBCs on postoperative infection rates in cardiac surgery. Transfusion. 2002;42:1127e1134. van Hilten JA, van de Watering LM, van Bockel JH, et al. Effects of transfusion with red cells filtered to remove leukocytes: randomised controlled trial in patients undergoing major surgery. BMJ. 2004;328:1281. van de Watering LM, Hermans J, Houbiers JG, et al. Beneficial effects of leukocyte depletion of transfused blood on postoperative complications in patients undergoing cardiac surgery: a randomized clinical trial. Circulation. 1998;97:562e568. Frietsch T, Karger R, Scholer M, et al. Leukodepletion of autologous whole blood has no impact on perioperative infection rate and length of hospital stay. Transfusion. 2008;48:2133e2142. Lapierre V, Aupe´rin A, Robinet E, et al. Immune modulation and microchimerism after unmodified versus leukoreduced allogeneic red blood cell transfusion in cancer patients: results of a randomized study. Transfusion. 2007;47:1691e1699. Tasaki T, Ohto H, Sasaki S, Kanno T, Igari T, Hoshi Y. Significance of pre-storage leucoreduction for autologous blood. Vox Sang. 2009;96:226e233. Jensen LS, Andersen AJ, Christiansen PM, et al. Postoperative infection and natural killer cell function following blood transfusion in patients undergoing elective colorectal surgery. Br J Surg. 1992;79:513e516. Jensen LS, Kissmeyer-Nielsen P, Wolff B, Qvist N. Randomised comparison of leukocyte-depleted versus buffy-coat-poor blood transfusion and complications after colorectal surgery. Lancet (London, England). 1996;348:841e845. Tartter PI, Mohandas K, Azar P, Endres J, Kaplan J, Spivack M. Randomized trial comparing packed red cell blood transfusion with and without leukocyte depletion for gastrointestinal surgery. Am J Surg. 1998;176:462e466. Efstathiou A, Vlachveis M, Tsonis G, Asteri T, Psarakis A, Fessatidis IT. Does leukodepletion during elective cardiac surgery really influence the overall clinical outcome? J Cardiovasc Surg (torino). 2003;44:197e204. Connery CP, Toumpoulis IK, Anagnostopoulos CE, et al. Does leukofiltration reduce pulmonary infections in CABG patients? A prospective, randomized study with early results and mid-term survival. Acta Cardiol. 2005;60:285e293. Houbiers JG, Brand A, van de Watering LM, et al. Randomised controlled trial comparing transfusion of leukocyte-depleted or buffy-coat-depleted blood in surgery for colorectal cancer. Lancet (London, England). 1994;344:573e578. Nielsen H, Hmmer J, Krarup A, et al. Prestorage leukocyte filtration may reduce leukocyte-derived bioactive substance accumulation in patients operated for burn trauma. Burns. 1999;25:162e170. Thornton A, Lee P. Publication bias in meta-analysis: its causes and consequences. J Clin Epidemiol. 2000;53:207e216. Peters J, Sutton AJ, Jones DR, Abrams KR, Rushton L. Comparison of two methods to detect publication bias in meta-analysis. JAMA. 2006;295:676e680. Duval S, Tweedie R. Trim and fill: a simple funnel-plot-based method of testing and adjusting for publication bias in meta-analysis. Biometrics. 2000;56:455e463.
k w o n e t a l l e u k o c y t e - fi l t e r e d b l o o d a n d p o s t o p e r a t i v e i n f e c t i o n s
42. Centers for Disease and Control. National hospital Discharge Survey, 2010; 2015. Available at: http://www.cdc.gov/nchs/ fastats/hospital.htm. Accessed February 2, 2016. 43. Pape A, Stein P, Horn O, Habler O. Clinical evidence of blood transfusion effectiveness. Blood Transfus. 2009;7:250e258. 44. Alter HJ, Klein HG. ASH 50th anniversary review the hazards of blood transfusion in historical perspective. Blood. 2016;112:2617e2627. 45. Angelbeck JH, Ortolano GA. Universal leukocyte reduction: is it appropriate medical practice or not? J Infus Nurs. 2005;28:273e281. 46. Vamvakas EC. Why have meta-analyses of randomized controlled trials of the association between non-white-bloodcell-reduced allogeneic blood transfusion and postoperative infection produced discordant results? Vox Sang. 2007;93:196e207. 47. Clarkson MR, Sayegh MH. T-cell costimulatory pathways in allograft rejection and tolerance. Transplantation. 2005;80:555e563. 48. Ottonello L, Ghio M, Contini P, et al. Nonleukoreduced red blood cell transfusion induces a sustained inhibition of neutrophil chemotaxis by stimulating in vivo production of transforming growth factor-??1 by neutrophils: role of the immunoglobulinlike transcript 1, sFasL, and sHLA-I. Transfusion. 2007;47:1395e1404. 49. Onorati F, Santini F, Mariscalco G, Bertolini P, Faggian G, Mazzucco A. Leukocyte filtration ameliorates the inflammatory response in patients with mild to moderate lung dysfunction. Ann Thorac Surg. 2011;92:111e121. 50. Collier AC, Kalish LA, Busch MP, et al. Leukocyte-reduced red blood cell transfusions in patients with anemia and human immunodeficiency virus infection: the Viral Activation Transfusion Study: a randomized controlled trial. JAMA. 2001;285:1592e1601. 51. Simancas-Racines D, Osorio D, Martı´-Carvajal AJ, Arevalo-Rodriguez I. Leukoreduction for the prevention of adverse reactions from allogeneic blood transfusion. Cochrane Database Syst Rev. 2015;12:CD009745. 52. Jensen LS, Grunnet N, Hanberg-Sorensen F, Jørgensen J. Cost-effectiveness of blood transfusion and white cell
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
509
reduction in elective colorectal surgery. Transfusion. 1995;35:719e722. Blumberg N, Heal JM, Kirkley SA, DiPersio JF, Rapoport AP, Rowe JM. Leukodepleted-ABO-identical blood components in the treatment of hematologic malignancies: a cost analysis. Am J Hematol. 1995;48:108e115. Postma MJ, van de Watering LM, et al. Cost-effectiveness of leukocyte depletion of red-cell transfusions for patients undergoing cardiac surgery. Vox Sang. 2003;84:65e67. Williamson LM, Stainsby D, Jones H, et al. The impact of universal leukodepletion of the blood supply on hemovigilance reports of posttransfusion purpura and transfusion-associated graft-versus-host disease. Transfusion. 2007;47:1455e1467. Bassuni WY, Blajchman MA, Al-Moshary MA. Why implement universal leukoreduction? Hematol Oncol Stem Cell Ther. 2008;1:106e123. Bordin J, Bardossy L, Blajchman M. Growth enhancement of established tumors by allogeneic blood transfusion in experimental animals and its amelioration by leukodepletion: the importance of the timing of the leukodepletion. Blood. 1994;84:344e348. Blajchman M, Bardossy L, Carmen R, Goldman M, Heddle N, Singal D. An animal model of allogeneic donor platelet refractoriness: the effect of the time of leukodepletion. Blood. 1992;79:1371e1375. Phelan HA, Gonzalez RP, Patel HD, et al. Prestorage leukoreduction ameliorates the effects of aging on banked blood. J Trauma. 2010;69:330e337. Phelan HA, Eastman AL, Aldy K, et al. Prestorage leukoreduction abrogates the detrimental effect of aging on packed red cells transfused after trauma: a prospective cohort study. Am J Surg. 2012;203:198e204. Wang D, Sun J, Solomon SB, Klein HG, Natanson C. Transfusion of older stored blood and risk of death: a meta-analysis. Transfusion. 2012;52:1184e1195. Hozo SP, Djulbegovic B, Hozo I. Estimating the mean and variance from the median, range, and the size of a sample. BMC Med Res Methodol. 2005;5:13. Hackshaw A. A Concise Guide to clinical trials. London, UK: BMJ Books; 2009.
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.JournalofSurgicalResearch.com
Leukocyte filtration and postoperative infections Seunghyug Kwon, BS, MPH,a Sungyub Lew, MD,a,b and Ronald S. Chamberlain, MD, MPA, FACSa,b,c,* a
Saint George’s University School of Medicine, Grenada, West Indies Department of Surgery, Saint Barnabas Medical Center, Livingston, New Jersey c Department of Surgery, New Jersey Medical School, Rutgers University, Newark, New Jersey b
article info
abstract
Article history:
Background: Leukocyte filtration has been hypothesized to reduce the risk of postoperative
Received 29 January 2016
infections by alleviating the immunosuppressive effect of whole blood. However, the
Received in revised form
literature regarding the clinical efficacy of leukocyte filtration remains conflicted. This
22 April 2016
meta-analysis investigates the impact of allogeneic and autologous leukocyte-filtered
Accepted 9 June 2016
blood transfusions on the incidence of postoperative infections in adult surgical patients.
Available online 27 June 2016
Methods: A comprehensive literature search of PubMed, Google Scholar, and Cochrane Central Registry of Controlled trials (1966-2016) was completed for all published
Keywords:
randomized controlled trials. Postoperative infections under “as-per-protocol” (APP) and
Leukocyte filtration
“intention-to-treat” (ITT), length of stay, and mortality were analyzed.
Postoperative infections
Results: Sixteen randomized controlled trials involving 6586 randomized (ITT) patients
Transfusion
(4615 APP patients) in various clinical settings were evaluated. The leukocyte-filtered blood
Transfusion-related
group demonstrated an overall 26% risk reduction in postoperative infections when
immunomodulation
analyzed by APP (relative risk [RR] ¼ 0.74; 95% confidence interval [CI, 0.60-0.92]; P ¼ 0.007)
White-blood-cell reduction
and a 22% risk reduction when analyzed by ITT (RR ¼ 0.78; 95% CI [0.65-0.94]; P ¼ 0.009).
Meta-analysis
Leukocyte-filtered blood was also associated with a significant reduction in length of stay (standardized difference of mean [SDM] ¼ 0.74; 95% CI [1.32 to 0.15]; P ¼ 0.014) and all-cause mortality (RR ¼ 0.74; 95% CI [0.57-0.95]; P ¼ 0.018). Conclusions: Leukocyte-filtered blood transfusions are associated with significantly lower postoperative infection rates in both the APP and ITT populations. Leukocyte filtration also shortens length of stay and decreases all-cause mortality in surgical patients and should be considered in all surgical patients. ª 2016 Elsevier Inc. All rights reserved.
Introduction The American Red Cross has reported that a total of 15 million red blood cell (RBC) units are transfused annually.1 Surgical patients account for two-thirds of these RBC transfusions.2 Though transfusions are essential in saving lives, multiple studies have reported numerous findings of adverse effects, ranging from the transmission of infectious disease to
deleterious transfusion-related immunomodulation (TRIM).3-5 With modern blood banking technology, the transmission of infectious diseases has been vigilantly managed and concerns are substantially minimized overall.6 Most transfusion-related morbidity and mortality consist of noninfectious complications, the greatest of concern being the immunosuppressive effect of TRIM.7 The immunosuppressive effect of transfusions were first demonstrated by
* Corresponding author. Department of Surgery, Saint Barnabas Medical Center, Rutgers University, New Jersey Medical School (NJMS), 94 Old Short Hills Rd, Livingston, NJ 07039. Tel.: þ1 973 322 5195; fax: þ1 973 322 2471. E-mail address: [email protected] (R.S. Chamberlain). 0022-4804/$ e see front matter ª 2016 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jss.2016.06.055
500
j o u r n a l o f s u r g i c a l r e s e a r c h o c t o b e r 2 0 1 6 ( 2 0 5 ) 4 9 9 e5 0 9
Opelz et al. 8 in a prospective study of 148 cadaver donor transplant recipients in which rejection rates were compared based on blood transfusion status. The result of this study revealed that renal allograft rejection was improved in recipients receiving 10 or more transfusions as opposed to those who did not receive transfusions (66% versus 29%, P < 0.001).8 The author inferred from these findings that blood transfusions result in an immunosuppressive effect of the transfused host’s immune system.8 Historically, the immunosuppressive effect of blood transfusions was an often used means to limit rejection in renal transplants until cyclosporine and other immunosuppressives were introduced.8 TRIM has also been observed in surgical patient groups in which an increased number of postoperative infections are observed in patients receiving RBC transfusions. A prospective observational study by Tartter et al. 9, involving 168 colorectal cancer patients, sought to identify perioperative determinants of infectious complications. Using multivariate analysis, these authors reported that 24 of 168 patients developed infectious complications and observed that blood transfusion was an independent significant risk factor for infection (P ¼ 0.01).9 A meta-analysis of 21 randomized controlled trials (RCTs) involving 131,512 patients by Hill et al. 10 (1992) demonstrated a three-fold increase in postoperative infections in transfused patients compared to nontransfused patients (OR ¼ 3.45; 95% confidence interval [CI, 1.42-15.15]; P < 0.05). The presence of leukocytes in the transfused products appears critical to produce the TRIM effect.11 A prospective study involving 80 orthopedic surgery patients by Innerhofer et al. 12 observed impaired T-cell-mediated immunity in patients who were transfused with leukocyte containing RBCs. These findings were continued by Lee et al. 13 who reported persistent donor leukocytes in humans for up to 1.5 y after blood transfusion. Unsurprisingly, various other benefits associated with leukocyte-filtered blood have also been reported, including prevention of recurrent febrile nonhemolytic transfusion reactions (FNHTR), alloimmunization in transplantation, and transmission of cytomegalovirus.14 European countries (i.e., Germany, the United Kingdom, and the Netherlands) and Canada have implemented a universal leukocyte filtration policy, reaping benefits of decreased FNHTR incidences and cost savings due to decreased length of hospital stay.15-17 Despite implementation in these countries, the association between leukocyte filtration and postoperative infection remains controversial.18-21 The most recent Cochrane review by SimancasRacines et al. 21 included 10 RCTs involving 3557 surgical and nonsurgical patients transfused with allogeneic blood and yielded no difference in infection risks (RR ¼ 0.76; 95% CI [0.58-1.00]; P ¼ 0.05) when leukocyte-filtered blood was used. Bilgin et al. 22 conducted an RCT involving 496 patients undergoing valve surgery and reported lower postoperative infection rates (33.8% versus 24.3%, P ¼ 0.032) with the use of leukocyte-filtered blood. Conversely, Nathens et al. 23 conducted an RCT involving 324 trauma patients and reported no difference in infection rates (36.0% versus 30.3%, P ¼ 0.321) with the use of leukocyte-filtered blood.
The current meta-analysis provides a comprehensive review of all published studies in surgical patients, in which leukocyte-filtered blood transfusions were provided to determine the impact on postoperative outcomes, including postoperative infection rates, length of stay, and mortality.
Materials and methods Study selection A comprehensive search of all published RCTs comparing patients receiving leukocyte-filtered and nonleukocytefiltered blood transfusions was conducted using PubMed, Google Scholar, Cochrane Central Registry of Controlled Trials (1966-2016). Additional citations were searched using references retrieved from prior publications (Fig. 1). The last search was conducted on January 10, 2016, and only articles conducted in English were considered. Keywords searched included combinations of “leuk(c)oreduced,” “leuk(c)odepleted,” “leuk(c)ocyte filtered,” “white cell reduced,” “leuk(c) ocyte reduced,” “leuk(c)ocyte depleted,” “leuk(c)ocyte depleting,” and “transfusions.” The inclusion criteria were limited to RCTs in adult surgical populations (>18 y), comparing leukocyte-filtered and nonleukocyte-filtered blood, and reporting the incidence of postoperative infections (surgical and nonsurgical site infections) with sample size. Studies which did not include postoperative infections as an outcome were excluded. This meta-analysis was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement.24
Data extraction Articles retrieved from searches were assessed for eligibility and data pertaining to patients, interventions, comparison groups, outcomes, and methodology were abstracted. The primary clinical outcome of interest was the incidence of postoperative infection. Secondary outcomes were hospital length of stay and all-cause mortality.
Statistical analysis For each trial, RR with a 95% CI for the incidence of postoperative infections and mortality was calculated. SDM with 95% CI was calculated for LOS. Meta-analysis of the pooled data was performed using Comparative Meta-analysis software version 3 (CMA v 3; Biostat, Englewood, NJ). For individual studies reporting zero events in any group, a continuity correction factor of 0.5 was adopted to calculate the RR and variance. Both fixed-effect and random-effect models were considered, depending on the heterogeneity of the included studies. To assess the heterogeneity between studies, both Cochrane’s Q statistic and I2 statistic were used. Heterogeneity was considered statistically significant when P < 0.05 or I2 > 50. If heterogeneity was observed, data were analyzed using a random-effect model. In the absence of heterogeneity, a fixed-effect model was assumed. For all outcomes, publication bias was the first qualitatively assessed using a funnel plot, and further quantitatively
k w o n e t a l l e u k o c y t e - fi l t e r e d b l o o d a n d p o s t o p e r a t i v e i n f e c t i o n s
501
Fig. 1 e Consolidated Standards of Reporting Trials diagram of the study selection process.
assessed with Egger’s and Begg’s tests. A two-tailed P < 0.05 was considered statistically significant. The presence of publication bias was further evaluated using Duval and Tweedie’s Trim and Fill to account for any missing studies. Subgroup analysis was performed based on leukocyte filtration before versus after storage, blood product type for control arm, allogeneic versus autologous blood source, clinical setting, and location of trial were performed to assess the variability within the trials.
Results Demographic characteristics of the studies A total of 16 RCTs were identified involving 4615 patients in the “as-per-protocol” (APP) population (studies published between 1992 and 2009).22,23,25-38 The leukocyte-filtered blood group included 2325 patients and the nonleukocyte-filtered blood group consisted of 2290 patients. The two major clinical settings were cardiac surgery (n ¼ 1949) and colorectal surgery (n ¼ 981). Additional clinical settings included trauma, orthopedic surgery, cancer surgery, and a combination of settings involving a total of 1685 patients (Table 1).
Postoperative infections: as-per-protocol Data on the incidence of postoperative infection were reported in all 16 RCTs involving 4615 patients (2325 patients receiving leukocyte-filtered blood and the 2290 patients
receiving nonleukocyte-filtered blood). Thirteen of the 16 trials reported a lower risk of postoperative infections with patients receiving leukocyte-filtered blood; 5 of which were statistically significant. A lower proportion of patients in the leukocyte-filtered group developed postoperative infections compared to the nonleukocyte-filtered blood (500/2325 [22%] versus 626/2290 [27%]). There was significant heterogeneity between trials (P < 0.01, I2 ¼ 70), and a random-effect model was used. Meta-analysis revealed a significantly lower risk of postoperative infections in the leukocyte-filtered group compared to the nonleukocyte-filtered group (RR ¼ 0.743; 95% CI [0.598-0.924]; P ¼ 0.007; Fig. 2). Subgroup analysis demonstrated that leukocyte-filtered blood had the greatest benefit for leukocyte filtration after storage or in allogeneic blood (RR ¼ 0.349; 95% CI [0.150-0.815]; P ¼ 0.015 and RR ¼ 0.716; 95% CI [0.562-0.912]; P ¼ 0.007, respectively). Significant benefits were also observed in the cardiac surgery setting and trials conducted in Europe (RR ¼ 0.738; 95% CI [0.616-0.884]; P ¼ 0.001 and RR ¼ 0.767; 95% CI [0.593-0.991]; P ¼ 0.042, respectively; Table 2).
Postoperative infections: intention-to-treat Data on the incidence of postoperative infections were reported in all 16 RCTs involving 6586 patients (3326 patients receiving leukocyte-filtered blood and the 3260 patients receiving nonleukocyte-filtered blood). Twelve studies reported a lower risk of postoperative infections with patients receiving leukocyte-filtered blood, 4 of which were statistically significant.22,23,25,26,28,31-36,38 A lower proportion of
502
j o u r n a l o f s u r g i c a l r e s e a r c h o c t o b e r 2 0 1 6 ( 2 0 5 ) 4 9 9 e5 0 9
Table 1 e Characteristics of all included randomized control trials of leukocyte-filtered blood transfusions (1966-2016). Author, y
Jensen, 1992
Clinical setting
Double blind
Multicenter
Leukocytefiltered arm (filter time)
Nonleukocytefiltered arm
Randomized patients (#)
Transfused patients (#)
Colorectal
N
N
Allo WB (A)
Allo WB
197
104
Colorectal cancer
N
Y
Allo RBCs (B)
BR allo RBCs
697
446
Cardiac
N
N
Allo RBCs (B n ¼ 305 or an n ¼ 303)
BR allo RBCs
914
861
Tartter, 1998
GI surgery
N
N
Allo WB (A)
Allo RBCs
221
59
Titlestad, 2001
Houbiers, 1994 van de Watering, 1998
Colorectal
Y
N
Allo RBCs (B)
BR allo RBCs
279
112
Wallis, 2002
Cardiac
N
N
Allo RBCs (B)
BR (n ¼ 204) or PR (n ¼ 198) allo RBCs
597
509
Bilgin, 2004
Cardiac
Y
Y
Allo RBCs (B)
BR allo RBCs
474
430
Nathens, 2006
Trauma
Y
N
Allo RBCs (B)
Allo RBCs
324
268
Lapierre, 2007
Cancer
N
N
Allo RBCs (B)
Allo RBCs
37
35
Frietsch, 2008
Total hip arthroplasty
Y
Y
Auto WB (B)
Auto WB
951
481
Colorectal
N
N
Allo WB (A)
BR allo RBCs
586
260
Multiple
Y
Y
Allo RBCs (B)
BR allo RBCs
1051
479
Efstathiou, 2003
Cardiac
N
N
Auto WB (A)
Auto WB
80
80
Nielsen, 1999
Trauma
N
N
Allo WB (B)
Allo WB
24
24
Cardiac
N
N
Allo WB (A)
Allo WB
104
69
Orthopedic
Y
N
Auto WB (B)
Auto WB
400
398
Jensen, 1996 van Hilten, 2004
Connery, 2005 Tasaki, 2009
A ¼ after; allo ¼ allogeneic; auto ¼ autologous; B ¼ before; BR ¼ buffy-coat reduced; N ¼ no; PR ¼ plasma reduced; RBC ¼ red blood cell; WB ¼ whole blood; Y ¼ yes.
patients in the leukocyte-filtered group developed postoperative infections compared to the nonleukocyte-filtered group (652/3326 [20%] versus 764/3260 [23%]). There was significant heterogeneity between trials (P < 0.01, I2 ¼ 64), and
a random-effect model was assumed. Meta-analysis revealed a significantly lower risk of postoperative infections in the leukocyte-filtered group compared to the nonleukocytefiltered group (RR ¼ 0.783; 95% CI [0.652-0.940]; P ¼ 0.009; Fig. 3).
Fig. 2 e Forest plot evaluating the risk of postoperative infections with the use of leukocyte-filtered blood (APP population). (Color version of figure is available online.)
Subgroup
Filter time
NLF type
Categories of subgroup
Before
ITT
APP
ITT
APP
ITT
APP
ITT
0.032
0.020
0.819
0.896
0.207
0.167
0.036
0.027
0.805
0.661
5444
0.893 (0.781-1.021)
0.904 (0.805-1.016)
0.097
0.091
1448
0.349 (0.150-0.815)
0.466 (0.269-0.806)
0.015
0.006
Buffy-coat reduced
722,25-28,33,37
2939
4400
0.760 (0.557-1.037)
0.812 (0.629-1.048)
0.083
0.110
Whole
629,31,32,35,36
1156
1626
0.649 (0.361-1.165)
0.656 (0.388-1.109)
0.147
0.116
362
362
0.796 (0.483-1.311)
0.796 (0.483-1.311)
0.370
0.370
334
393
0.598 (0.365-0.982)
0.725 (0.452-1.165)
0.042
0.184
959
1429
0.987 (0.639-1.527)
0.969 (0.745-1.261)
0.954
0.814
3656
5157
0.716 (0.562-0.912)
0.765 (0.621-0.942)
0.007
0.012
Autologous
Cardiac
Other (trauma, orthopedic, multiple) European USA Japan
323,30,34 1
26
329,31,35 1322,23,25-28,30,32-34,36-38 522,26,28,35,36
1949
2134
0.738 (0.616-0.884)
0.747 (0.624-0.894)
0.001
0.001
981
1725
0.447 (0.199-1.006)
0.563 (0.313-1.011)
0.052
0.054
1685
2727
0.954 (0.810-1.123)
0.976 (0.842-1.130)
0.571
0.743
3821
5792
0.767 (0.593-0.991)
0.811 (0.658-1.000)
0.042
0.050
323,34,36
396
396
0.642 (0.384-0.918)
0.641 (0.384-1.070)
0.089
0.089
131
398
398
0.656 (0.260-1.657)
0.656 (0.260-1.657)
0.372
0.372
525,32-34,37 623,27,29-31,38 1222,25-30,32,33,35,37,38
Difference between group P value
APP
1095
Colorectal
Country trial conducted
P value
3824
Allogeneic Clinical setting
Summary relative risk (95% confidence interval)
528,32,33,35,36
Plasma reduced
12
22,23,25-31,34,37,38
Number of patients
After
Red blood cells
Blood type
Number of studies (references)
APP ¼ as-per-protocol; ITT ¼ intention-to-treat; LF ¼ leukocyte filtration; NLF ¼ nonleukocyte filtered.
k w o n e t a l l e u k o c y t e - fi l t e r e d b l o o d a n d p o s t o p e r a t i v e i n f e c t i o n s
Table 2 e Subgroup analysis of leukocyte-filtered blood transfused groups on postoperative infections.
503
504
j o u r n a l o f s u r g i c a l r e s e a r c h o c t o b e r 2 0 1 6 ( 2 0 5 ) 4 9 9 e5 0 9
Fig. 3 e Forest plot evaluating the risk of postoperative infections with the use of leukocyte-filtered blood (intention-to-treat population). (Color version of figure is available online.)
Subgroup analysis demonstrated that leukocyte-filtered blood had the greatest benefit for leukocyte filtration after storage or allogeneic blood (RR ¼ 0.466; 95% CI [0.269-0.806]; P ¼ 0.006 and RR ¼ 0.765; 95% CI [0.621-0.942]; P ¼ 0.012, respectively). Significant benefits were also observed in the cardiac surgery setting (RR ¼ 0.747; 95% CI [0.624-0.894]; P ¼ 0.001; Table 2).
Length of stay Data on the length of stay were reported in 3 of the 16 RCTs involving 630 patients in the APP population (333 patients receiving leukocyte-filtered blood and the 297 patients receiving nonleukocyte-filtered blood).29,35,36 All studies reported shorter length of stay with leukocyte-filtered blood, however, only two reached statistical significance.29,35 There was significant heterogeneity between trials (P ¼ 0.001, I2 ¼ 86), and a random-effect model was assumed. Meta-analysis revealed a significant difference in mean length of stay between the leukocyte-filtered group and the nonleukocyte-filtered group (SDM ¼ 0.737; 95% CI [1.324 to 0.150]; P ¼ 0.014; Fig. 4).
All-cause mortality Data on the incidence of all-cause mortality were reported in 9 of the 16 RCTs for the APP population involving 2544 patients (1375 patients receiving leukocyte-filtered blood and the 1169 patients receiving nonleukocyte-filtered blood).22,23,25,27,28,30,33,36,38 Six studies reported a decrease in all-cause mortality with leukocyte-filtered blood, however, only one reached statistical significance.22,25,27,28,36,38 A lower
proportion of patients in the leukocyte-filtered group reported all-cause mortality compared to the nonleukocyte-filtered blood (95/1357 [7.0%] versus 121/1151 [11%]). There was no significant heterogeneity between trials (P ¼ 0.261, I2 ¼ 20.4), and a fixed-effect model was assumed. Meta-analysis revealed a significantly lower risk of death in the leukocytefiltered group compared to the nonleukocyte-filtered group (RR ¼ 0.735; 95% CI [0.570-0.949]; P ¼ 0.018; Fig. 5).
Subgroup analysis In the APP population, significant differences were observed among postoperative infection rates related to leukocyte filtration before or after storage (P ¼ 0.032) and in various clinical settings (P ¼ 0.036). No difference was observed for blood product type in the control arm (P ¼ 0.819), allogeneic versus autologous (P ¼ 0.207), and country in which the trial was conducted (P ¼ 0.805; Table 2). In the intention-to-treat population, significant differences were observed among postoperative infection rates related to leukocyte filtration before or after storage (P ¼ 0.02) and clinical setting (P ¼ 0.027). No difference was also observed for blood product type for control arm (P ¼ 0.896), allogeneic versus autologous (P ¼ 0.167), and country where trial was conducted (P ¼ 0.661; Table 2).
Publication bias Negative findings are less likely to be published, and consequently, over or underestimation of the outcome may occur.39 A qualitative and quantitative assessment of the publication
k w o n e t a l l e u k o c y t e - fi l t e r e d b l o o d a n d p o s t o p e r a t i v e i n f e c t i o n s
505
Fig. 4 e Forest plot evaluating the SDM for length of stay with the use of leukocyte-filtered blood (APP population). (Color version of figure is available online.)
bias was performed using a funnel plot and both Egger’s and Begg’s tests.40 Qualitative evidence of asymmetry was observed on the funnel plot (Fig. 6). While Begg’s test was nonsignificant (P ¼ 0.053), Egger’s test was suspected for publication bias, P ¼ 0.038. Duval and Tweedie’s method of Trim and Fill based on the random effect model and searching for studies on the right of the mean effect estimated 5 unpublished negative studies.41 The point estimate and 95% CI for the combined studies was 0.846 (0.764-0.947). Using Trim and Fill, the imputed point estimate was 0.911 (0.8251.006). After adjusting for countries with universal leukocyte filtration, publication bias was not observed by either Begg’s (P ¼ 0.07) or Egger’s tests (P ¼ 0.09).
Discussion In 2010, the Centers for Disease and Control estimated 51.4 million surgeries were performed in the United States (US).42 Approximately two million of these surgeries will require a
transfusion, averaging three transfusions per surgery.15 Although utilized for many reasons based on the belief that transfusion improves oxygen transport, blood transfusions also carry substantial adverse risks.43 The risks include a four-fold increase in infection, a one in 5000 chance of fatal transfusion-related acute lung injury, and an approximately 10% occurrence of autoantibodies.44 Leukocyte filtration has been hypothesized to decrease these complications and has been advocated as the standard for all blood transfusions; however, the efficacy of leukocyte filtration to alleviate the TRIM effect remains controversial.14,45 Several mechanisms have been proposed to explain the higher incidences of the noninfectious complications in transfused recipients, with TRIM being the most frequently supported.5 Soluble and cellular factors in leukocytes have been proposed as the primary mediator of the TRIM effect.46 The immunogenicity of cellular or soluble antigens associated with transfused blood components depends on the viability of donor dendritic antigen presenting cells and costimulatory signals for the presentation of the donor
Fig. 5 e Forest plot evaluating all-cause mortality with the use of leukocyte-filtered blood (APP population). (Color version of figure is available online.)
506
j o u r n a l o f s u r g i c a l r e s e a r c h o c t o b e r 2 0 1 6 ( 2 0 5 ) 4 9 9 e5 0 9
Fig. 6 e Funnel plot assessing publication bias (analyzing the effect of leukocyte-filtered blood on postoperative infection incidence in surgical patients).
antigens to recipient’s T-cells.5 When a multitude of antigens are presented, T-cell unresponsiveness and immune tolerance ensues due to either nonviable antigen presenting cells or absent costimulatory signals.47 Ottonello et al. 48 demonstrated the immunosuppression effect by comparing neutrophil function in postenonleukocyte-filtered RBC transfusion blood samples from eight healthy volunteers. After RBC transfusion, neutrophils in the recipient synthesized transforming growth factor b1, a cytokine exerting an overall immunosuppressive activity, which in turned caused an inhibition of neutrophil chemotaxis that lasted 8-15 d.48 Additional studies by Onorati et al. 49 examined proinflammatory cytokine release in 60 chronic obstructive pulmonary disease patients undergoing aortic valve surgery by randomly allocating 30 to a leukocyte-filtered group and the other to a nonleukocyte-filtered group. The leukocytefiltered group had a significant reduction in various proinflammatory markers, such as interleukin (IL)-6, IL-8, IL10, and tumor necrosis factor-a (P < 0.01).49 The postoperative infection risk reported in this meta-analysis is consistent with the results from previous studies when analyzed by APP. Fergusson et al. 20 reported a meta-analysis of 10 RCTs involving 2358 cardiac and colorectal surgery patients transfused with allogeneic blood and similarly resulted a significantly lower risk of postoperative infections (RR ¼ 0.60; 95% CI [0.38-0.93]; P < 0.05) with leukocyte-filtered blood in those transfused. Blumberg et al. 18 reported on 9 RCTs involving 3093 mostly cardiac and colorectal surgery patients who were transfused allogeneic blood and also noted lower postoperative infection risks (OR ¼ 0.522; 95% CI [0.332-0.821]; P < 0.05) with the use of leukocyte-filtered blood. In contrast, the most recent Cochrane review by Simancas-Racines et al. 21 involving 10 RCTs involving 3557 surgical and nonsurgical patients
transfused with allogeneic blood yielded no difference in infection risks (RR ¼ 0.76; 95% CI [0.58-1.00]; P ¼ 0.05) with the use of leukocyte-filtered blood. The difference between the conclusions of this report with Cochrane’s result may be explained in part by the inclusion or exclusion criteria and population analyzed. The Cochrane review excluded studies using autologous blood, patients receiving other blood components, and patients receiving RBCs with no indications.29,31,32,35-38 Five of 7 of these trials favored leukocyte-filtered blood.31,32,35,36,38 The current study was limited to surgical patients due to the lack of nonsurgical patient trials on infection outcomes. In addition, the Cochrane review included only one nonsurgical patient trial which involved HIV patients in their analysis, whose results may have limited generalizability.50 Collier et al. (2001) hypothesized that the compromised immune system of the population may have reacted differently to leukocytes, thus making any conclusion beyond HIV patient seem inappropriate.50 Finally, when we analyzed the Cochrane review data by excluding the nonsurgical trial, the results were similar to the current analysis (RR ¼ 0.71; 95% CI [0.53-0.95]; P ¼ 0.022). The Cochrane review also analyzed all-cause mortality reported in 10 RCTs involving 4060 surgical and nonsurgical patients transfused with allogeneic blood and seemingly noted no difference (RR ¼ 0.80; 95% CI [0.60-1.07]; P ¼ 0.13) with the use of leukocyte-filtered blood.21 Fever was the only outcome that was significantly decreased (RR ¼ 0.75; 95% CI [0.60-0.94], P ¼ 0.01) in the leukocyte-filtered blood group; however, only two RCTs involving 634 cardiac and cancer surgery patients experienced this outcome.51,21 Leukocyte-filtered blood may not benefit all patients, as Collier et al. (2001) have reported leukocyte filtration yielded no benefit in HIV patients, and thus the universal
k w o n e t a l l e u k o c y t e - fi l t e r e d b l o o d a n d p o s t o p e r a t i v e i n f e c t i o n s
implementation of leukocyte filtration is not warranted.50 Leukocyte filtration is estimated to cost an additional $20-30 US per unit of blood, which will add an additional $180 US million to the United States hospital budget.15 And yet in appropriate populations where leukocyte filtration is associated with lower hospital costs such as the surgical patients in this meta-analysis, the cost may well be justified.50-54 In an RCT by Tartter et al. 34, lower average hospital charges were reported in gastrointestinal surgery patients receiving filtered blood compared to nonfiltered blood ($33,954 US versus $41,002 US, respectively). While this meta-analysis focused primarily on postoperative infection risks, there are other potential benefits in using a universal leukocyte-filtered blood supply.16,17,55 In patients requiring a blood transfusion, leukocyte filtration has also been shown to decrease FNHTR. Paglino et al. 17 compared the 3-y incidence rate of FNHTR in RBC units transfused before and after the implementation of universal leukocyte filtration (85,321 RBC units prior and 60,048 transfused after) and reported FNHTR rates reduced by approximately 50% (0.35%-0.18%, P < 0.001). Similarly, Yazer et al. 16 compared the 3-y incidence rate of FNHTR in RBC units transfused before and after the implementation of universal leukocyte filtration (70,396 RBC units transfused prior and 72,949 transfused after) and reported significantly lower incidence rates of FNHTR following the implementation of universal leukocyte filtration (0.19% versus 0.33%, P < 0.001). Leukocyte filtration also has the potential to decrease transfusion-associated graft-versushost-disease (TA-GVHD).56 Williamson et al. (2007) used the United Kingdom hemovigilance scheme Serious Hazards of Transfusion database and compared the number of TA-GVHD reports among patients receiving RBC and platelet transfusions over a 4-y period before and after the implementation of universal leukocyte filtration (8407 RBCs and platelets prior and 17,384 transfused after). These authors reported a significant decrease in TA-GVHD following implementation of universal leukocyte filtration (11 versus 2 reports, P < 0.001).55 Although the results of this meta-analysis are significant, there are limitations to this study due to the variation and heterogeneity of the RCTs included. Significant heterogeneity was observed for the filter time and surgery type; thus, the effect of leukocyte filtration may be underestimated. A similar result was also observed with filtration time in the meta-analysis by Fergusson et al.20 That said, outcomes related to filtration timing in animal studies have shown that prestorage filtration is superior to poststorage filtration for the prevention of tumor growth and alloimmunization.57,58 A number of recent retrospective studies have also shown that prestorage leukocyte filtration ameliorates the effects of aging in blood as well.59,60 While the accumulation of toxic mediators in stored blood is a possibility, the clinical manifestation of these mediators remains a controversial topic. Further randomized studies are needed to verify and explain the mechanism of filtration time. In addition, the significant variation in the units transfused and the storage time of the filtered blood, both of which have been thought to be associated with greater adverse effects.23,27,61 Due to the lack of availability of information, the effect of transfusion amount and storage time could not be thoroughly investigated in this analysis. Additional sources of heterogeneity in the
507
analysis include surgeon-specific selection factors, such as a liberal or conservative transfusion policy. The severity of each patient’s illness is an additional significant underlying factor, which directly relates to different proportions of transfused patients and the frequency of postoperative infection. Some studies transfused additional blood components in both trial arms as well.32,38 There is also the concern of selection bias, as not all RCTs were double blind and the details of the randomization process of some RCTs were not well reported. In regard to length of stay data, one trial was converted from median to mean number of days.29,62,63 Mortality for van de Watering et al. 28 was calculated based on the mortality rate presented in the paper not raw numbers; thus, the actual mortality count may differ. At last, some trials were exposed to a greater risk of publication bias due to variation in policies governing implementation of universal leukocyte filtration.23,34,36
Conclusions This study demonstrates that leukocyte-filtered blood is a valuable means to improve postoperative outcomes by reducing the incidence of postoperative infections, especially among cardiac surgery patients, which also resulted in a decreased length of hospital stay and overall mortality. Leukocyte-filtered blood has the potential to play a promising role in limiting transfusion-related postoperative complications for surgical patients, which should translate into a significant decrease in both morbidity and mortality.
Acknowledgment Authors’ contributions: Conception and design was developed by S.K. and R.S.C. Data analysis and interpretation and final approval of manuscript were carried out by S.K., S.L., and R.S.C. Manuscript writing was done by S.K., S.L., and R.S.C.
Disclosure The authors indicated no potential conflicts of interest.
references
1. American Red Cross, Blood facts and Statistics. Available at: http://www.redcrossblood.org/learn-about-blood/bloodfacts-and-statistics. Accessed August 14, 2015. 2. Cuong J, Catlett C, Berenholtz S, Haut E. Change in use of allogeneic red blood cell transfusions among surgical patients over time. J Am Coll Surg. 2009;207:352e359. 3. Day JR, Taylor KM. The systemic inflammatory response syndrome and cardiopulmonary bypass. Int J Surg. 2005;3:129e140. 4. Bilgin YM, van de Watering LM, Brand A. Clinical effects of leucoreduction of blood transfusions. Neth J Med. 2011;69:441e450.
508
j o u r n a l o f s u r g i c a l r e s e a r c h o c t o b e r 2 0 1 6 ( 2 0 5 ) 4 9 9 e5 0 9
5. Vamvakas EC, Blajchman MA. Transfusion-related immunomodulation (TRIM): an update. Blood Rev. 2007;21:327e348. 6. American Red Cross, “History of blood transfusion.” Available at: http://www.redcrossblood.org/learn-about-blood/historyblood-transfusion. Accessed January 8, 2015. 7. Hendrickson JE, Hillyer CD. Noninfectious serious hazards of transfusion. Anesth Analg. 2009;108:759e769. 8. Opelz G, Sengar DP, Mickey MR, Terasaki PI. Effect of blood transfusions on subsequent kidney transplant. Transpl Proc. 1973;5:253e259. 9. Tartter P, Quintero S, Barron D. Perioperative blood transfusion associated with infectious complications after colorectal cancer operations. Am J Surg. 1986;152:479e482. 10. Hill GE, Frawley WH, Griffith KE, Forestner JE, Minei JP. Allogeneic blood transfusion increases the risk of postoperative bacterial infection: a meta-analysis. J Trauma. 2003;54:908e914. 11. Horimi T, Terasaki P, Chia D, Sasaki N. Factors influencing the paradoxical effect of transfusions on kidney transplants. Transplantation. 1983;35:320e323. 12. Innerhofer P, Luz G, Spotl L, et al. Immunologic changes after transfusion of autologous or allogeneic buffy coat-poor versus white cell-reduced blood to patients undergoing arthroplasty. I. Proliferative T-cell responses and the balance of helper and suppressor T cells. Transfus Complicat. 1999;39:1089e1096. 13. Lee TH, Paglieroni T, Ohto H, Holland PV, Busch MP. Survival of donor leukocyte subpopulations in immunocompetent transfusion recipients: frequent long-term microchimerism in severe trauma patients. Blood. 1999;93:3127e3139. 14. American Association of Blood Banks. “CIRCULAR OF information,” Circular of Informationdfor the use of blood and blood components. Bethesda, MD: American Association of Blood Banks; 2013. 15. Lannan KL, Sahler J, Spinelli SL, Phipps RP, Blumberg N. Transfusion immunomodulationethe case for leukoreduced and (perhaps) washed transfusions. Blood Cells. Mol. Dis. 2013;50:61e68. 16. Yazer MH, Podlosky L, Clarke G, Nahirniak SM. The effect of prestorage WBC reduction on the rates of febrile nonhemolytic transfusion reactions to platelet concentrates and RBC. Transfus Pract. 2004;44:10e15. 17. Paglino JC, Pomper GJ, Fisch GS, Champion MH, Snyder EL. Reduction of febrile but not allergic reactions to RBCs and platelets after conversion to universal prestorage leukoreduction. Transfus P. 2004;44:16e24. 18. Blumberg N, Zhao H, Wang H, Messing S, Heal JM, Lyman GH. The intention-to-treat principle in clinical trials and meta-analyses of leukoreduced blood transfusions in surgical patients. Transfusion. 2007;47:573e581. 19. Vamvakas EC. White-blood-cell-containing allogeneic blood transfusion and postoperative infection or mortality: an updated meta-analysis. Vox Sang. 2007;92:224e232. 20. Fergusson D, Khanna MP, Tinmouth A, He´bert PC. Transfusion of leukoreduced red blood cells may decrease postoperative infections: two meta-analyses of randomized controlled trials. Can J Anaesth. 2004;51:417e424. 21. Simancas-Racines D, Osorio D, Martı´-carvajal A, Arevalo-Rodriguez I. Leukoreduction for the prevention of adverse reactions from allogeneic blood transfusion. Cochrane Database Syst Rev. 2015. 22. Bilgin YM, van de Watering LMG, Eijsman L, et al. Double-blind, randomized controlled trial on the effect of leukocyte-depleted erythrocyte transfusions in cardiac valve surgery. Circulation. 2004;109:2755e2760. 23. Nathens AB, Nester TA, Rubenfeld GD, Nirula R, Gernsheimer TB. The effects of leukoreduced blood
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39. 40.
41.
transfusion on infection risk following injury: a randomized controlled trial. Shock. 2006;26:342e347. Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. Int J Surg. 2010;8:336e341. Titlestad IL, Ebbesen LS, Ainsworth AP, Lillevang ST, Qvist N, Georgsen J. Leukocyte-depletion of blood components does not significantly reduce the risk of infectious complications. Results of a double-blinded, randomized study. Int J Colorectal Dis. 2001;16:147e153. Wallis JP, Chapman CE, Orr KE, Clark SC, Forty JR. Effect of WBC reduction of transfused RBCs on postoperative infection rates in cardiac surgery. Transfusion. 2002;42:1127e1134. van Hilten JA, van de Watering LM, van Bockel JH, et al. Effects of transfusion with red cells filtered to remove leukocytes: randomised controlled trial in patients undergoing major surgery. BMJ. 2004;328:1281. van de Watering LM, Hermans J, Houbiers JG, et al. Beneficial effects of leukocyte depletion of transfused blood on postoperative complications in patients undergoing cardiac surgery: a randomized clinical trial. Circulation. 1998;97:562e568. Frietsch T, Karger R, Scholer M, et al. Leukodepletion of autologous whole blood has no impact on perioperative infection rate and length of hospital stay. Transfusion. 2008;48:2133e2142. Lapierre V, Aupe´rin A, Robinet E, et al. Immune modulation and microchimerism after unmodified versus leukoreduced allogeneic red blood cell transfusion in cancer patients: results of a randomized study. Transfusion. 2007;47:1691e1699. Tasaki T, Ohto H, Sasaki S, Kanno T, Igari T, Hoshi Y. Significance of pre-storage leucoreduction for autologous blood. Vox Sang. 2009;96:226e233. Jensen LS, Andersen AJ, Christiansen PM, et al. Postoperative infection and natural killer cell function following blood transfusion in patients undergoing elective colorectal surgery. Br J Surg. 1992;79:513e516. Jensen LS, Kissmeyer-Nielsen P, Wolff B, Qvist N. Randomised comparison of leukocyte-depleted versus buffy-coat-poor blood transfusion and complications after colorectal surgery. Lancet (London, England). 1996;348:841e845. Tartter PI, Mohandas K, Azar P, Endres J, Kaplan J, Spivack M. Randomized trial comparing packed red cell blood transfusion with and without leukocyte depletion for gastrointestinal surgery. Am J Surg. 1998;176:462e466. Efstathiou A, Vlachveis M, Tsonis G, Asteri T, Psarakis A, Fessatidis IT. Does leukodepletion during elective cardiac surgery really influence the overall clinical outcome? J Cardiovasc Surg (torino). 2003;44:197e204. Connery CP, Toumpoulis IK, Anagnostopoulos CE, et al. Does leukofiltration reduce pulmonary infections in CABG patients? A prospective, randomized study with early results and mid-term survival. Acta Cardiol. 2005;60:285e293. Houbiers JG, Brand A, van de Watering LM, et al. Randomised controlled trial comparing transfusion of leukocyte-depleted or buffy-coat-depleted blood in surgery for colorectal cancer. Lancet (London, England). 1994;344:573e578. Nielsen H, Hmmer J, Krarup A, et al. Prestorage leukocyte filtration may reduce leukocyte-derived bioactive substance accumulation in patients operated for burn trauma. Burns. 1999;25:162e170. Thornton A, Lee P. Publication bias in meta-analysis: its causes and consequences. J Clin Epidemiol. 2000;53:207e216. Peters J, Sutton AJ, Jones DR, Abrams KR, Rushton L. Comparison of two methods to detect publication bias in meta-analysis. JAMA. 2006;295:676e680. Duval S, Tweedie R. Trim and fill: a simple funnel-plot-based method of testing and adjusting for publication bias in meta-analysis. Biometrics. 2000;56:455e463.
k w o n e t a l l e u k o c y t e - fi l t e r e d b l o o d a n d p o s t o p e r a t i v e i n f e c t i o n s
42. Centers for Disease and Control. National hospital Discharge Survey, 2010; 2015. Available at: http://www.cdc.gov/nchs/ fastats/hospital.htm. Accessed February 2, 2016. 43. Pape A, Stein P, Horn O, Habler O. Clinical evidence of blood transfusion effectiveness. Blood Transfus. 2009;7:250e258. 44. Alter HJ, Klein HG. ASH 50th anniversary review the hazards of blood transfusion in historical perspective. Blood. 2016;112:2617e2627. 45. Angelbeck JH, Ortolano GA. Universal leukocyte reduction: is it appropriate medical practice or not? J Infus Nurs. 2005;28:273e281. 46. Vamvakas EC. Why have meta-analyses of randomized controlled trials of the association between non-white-bloodcell-reduced allogeneic blood transfusion and postoperative infection produced discordant results? Vox Sang. 2007;93:196e207. 47. Clarkson MR, Sayegh MH. T-cell costimulatory pathways in allograft rejection and tolerance. Transplantation. 2005;80:555e563. 48. Ottonello L, Ghio M, Contini P, et al. Nonleukoreduced red blood cell transfusion induces a sustained inhibition of neutrophil chemotaxis by stimulating in vivo production of transforming growth factor-??1 by neutrophils: role of the immunoglobulinlike transcript 1, sFasL, and sHLA-I. Transfusion. 2007;47:1395e1404. 49. Onorati F, Santini F, Mariscalco G, Bertolini P, Faggian G, Mazzucco A. Leukocyte filtration ameliorates the inflammatory response in patients with mild to moderate lung dysfunction. Ann Thorac Surg. 2011;92:111e121. 50. Collier AC, Kalish LA, Busch MP, et al. Leukocyte-reduced red blood cell transfusions in patients with anemia and human immunodeficiency virus infection: the Viral Activation Transfusion Study: a randomized controlled trial. JAMA. 2001;285:1592e1601. 51. Simancas-Racines D, Osorio D, Martı´-Carvajal AJ, Arevalo-Rodriguez I. Leukoreduction for the prevention of adverse reactions from allogeneic blood transfusion. Cochrane Database Syst Rev. 2015;12:CD009745. 52. Jensen LS, Grunnet N, Hanberg-Sorensen F, Jørgensen J. Cost-effectiveness of blood transfusion and white cell
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
509
reduction in elective colorectal surgery. Transfusion. 1995;35:719e722. Blumberg N, Heal JM, Kirkley SA, DiPersio JF, Rapoport AP, Rowe JM. Leukodepleted-ABO-identical blood components in the treatment of hematologic malignancies: a cost analysis. Am J Hematol. 1995;48:108e115. Postma MJ, van de Watering LM, et al. Cost-effectiveness of leukocyte depletion of red-cell transfusions for patients undergoing cardiac surgery. Vox Sang. 2003;84:65e67. Williamson LM, Stainsby D, Jones H, et al. The impact of universal leukodepletion of the blood supply on hemovigilance reports of posttransfusion purpura and transfusion-associated graft-versus-host disease. Transfusion. 2007;47:1455e1467. Bassuni WY, Blajchman MA, Al-Moshary MA. Why implement universal leukoreduction? Hematol Oncol Stem Cell Ther. 2008;1:106e123. Bordin J, Bardossy L, Blajchman M. Growth enhancement of established tumors by allogeneic blood transfusion in experimental animals and its amelioration by leukodepletion: the importance of the timing of the leukodepletion. Blood. 1994;84:344e348. Blajchman M, Bardossy L, Carmen R, Goldman M, Heddle N, Singal D. An animal model of allogeneic donor platelet refractoriness: the effect of the time of leukodepletion. Blood. 1992;79:1371e1375. Phelan HA, Gonzalez RP, Patel HD, et al. Prestorage leukoreduction ameliorates the effects of aging on banked blood. J Trauma. 2010;69:330e337. Phelan HA, Eastman AL, Aldy K, et al. Prestorage leukoreduction abrogates the detrimental effect of aging on packed red cells transfused after trauma: a prospective cohort study. Am J Surg. 2012;203:198e204. Wang D, Sun J, Solomon SB, Klein HG, Natanson C. Transfusion of older stored blood and risk of death: a meta-analysis. Transfusion. 2012;52:1184e1195. Hozo SP, Djulbegovic B, Hozo I. Estimating the mean and variance from the median, range, and the size of a sample. BMC Med Res Methodol. 2005;5:13. Hackshaw A. A Concise Guide to clinical trials. London, UK: BMJ Books; 2009.