The Journal of Arthroplasty 28 (2013) 1463–1472

Contents lists available at SciVerse ScienceDirect

The Journal of Arthroplasty

j ourna l homepage: www.arth rop lasty journa l .o rg

The Use of Preoperative Erythropoiesis-Stimulating Agents (ESAs) in Patients WhoUnderwent Knee or Hip ArthroplastyA Meta-Analysis of Randomized Clinical Trials

Khalid Alsaleh, MD a, Ghazi S. Alotaibi, MBBS a, Hind S. Almodaimegh, PharmD b,Aamer A. Aleem, MD a, C. Tom Kouroukis, MD c

a Department of Medicine, Division of Hematology/Oncology, College of Medicine & King Khalid University Hospital, King Saud University, Riyadh, Saudi Arabiab King Saud Bin Abdulaziz University, Riyadh, Saudi Arabiac McMaster University, Ontario, Canada

This study was supported by a grant from the CollegDeanship of Scientific Research, King Saud University, R

The Conflict of Interest statement associated with thidx.doi.org/10.1016/j.arth.2013.01.024.

Reprint requests: Khalid Alsaleh, MD, DepartmeHematology/Oncology, College of Medicine & King KhBox: 2394372, King Saud University, Riyadh, Saudi Arab

0883-5403/2809-0004$36.00/0 – see front matter © 20http://dx.doi.org/10.1016/j.arth.2013.01.024

a b s t r a c t

a r t i c l e i n f o

Article history:Received 13 December 2012Accepted 25 January 2013

Keywords:arthroplastybloodtransfusionerythropoietinESAs

Erythropoiesis-stimulating agents (ESAs) have been used in orthopedic patients to reduce allogeneic bloodtransfusion (ABT). The purpose of this systematic review of randomized clinical trials is to evaluate theefficacy of preoperative administration of ESAs on hemoglobin level at discharge and frequency of ABT inpatients undergoing hip or knee surgery. Pooled results of 26 trials with 3560 participants showed that theuse of preoperative ESAs reduced ABT in patients undergoing hip or knee surgery [RR: 0.48, 95% CI: 0.38 to0.60, P b 0.00001]. Hemoglobin mean difference between ESA and control groups was 7.16 (g/L) [95% CI of4.73 to 9.59, P = 0.00001]. There was no difference in the risk of developing thromboembolism between ESAand control groups [RD: 0, 95 % CI: −1%–2%, P = 0.95]. ESAs offer an alternative blood conservation methodto avoid ABT in patients undergoing hip or knee surgery.

e of Medicine Research Center,iyadh, Saudi Arabia.s article can be found at http://

nt of Medicine, Division ofalid University Hospital, P.O.ia.

13 Elsevier Inc. All rights reserved.

© 2013 Elsevier Inc. All rights reserved.

Despite advances in transfusion medicine, allogeneic bloodtransfusion (ABT) remains a global issue because of the associatedrisks, resources utilization and safety concerns [1–3]. The need tosearch for a safer and more cost-effective alternative method becamean area of research interest for over two decades. Several perioper-ative blood salvage methods have been attempted in order to avoid orminimize ABT. Some of the therapeutic strategies that were usedinclude: using transfusion algorithms, preoperative autologous blooddonation, increasing the hemoglobin and red cell mass by pharma-cological therapy using iron and ESA and the use of fibrinolyticinhibitors to prevent bleeding [3–5]. However, all of these methodshave advantages and disadvantages.

The use of erythropoiesis stimulating agents (ESAs) has beeninvestigated as a blood transfusion conserving method with the aimto reduce the need for ABT. A mounting body of evidence suggests

that ESAs can be used as a safer alternative to ABT [5]. However, itsuse in this indication has not been widely adopted. An internationalsurvey conducted in 1999 revealed that ESAs were used in less than10% of surveyed hospitals except in Canada and Japan where theywere used in 13% and 51% of hospitals, respectively [6]. Postulatedbarriers to the wide use of ESA to minimize ABT are cost and safetyconcerns, and possible unfamiliarity with such an importantintervention. There remains an uncertainty regarding the cost-effectiveness of ESAs use in joint arthroplasty [7]. Another concernhas been the possibility that ESA can increase the risk of thrombo-embolic events [8].

Whatever the circumstances, there is certainly a gap between theevidence and the practice. This gap made it necessary to evaluate thestrength of the current and updated evidence to support or refute theuse of ESAs in orthopedic surgery. In 1998, Laupacis et al [9]conducted a meta-analysis to examine the safety and efficacy of ESAin minimizing the need for allogeneic blood transfusion in cardiac andorthopedic surgery. They concluded that ESA reduced the need forperioperative allogeneic transfusion. Our aim of this review is toupdate and evaluate the strength of the current evidence regardingthe efficacy and safety of ESAs in reducing the need for allogeneicblood transfusion, increasing hemoglobin level, and the frequency ofthromboembolic events in elective orthopedic surgeries. The cost ofESAs will not be discussed because the evidence evaluating its cost-effectiveness is scarce.

1464 K. Alsaleh et al. / The Journal of Arthroplasty 28 (2013) 1463–1472

Methods

Patient Population

This systematic review included all adult patients (N 18 years old)who underwent elective hip or knee arthroplasty. All types of hip orknee arthroplasty were included whether unilateral or bilateral,primary or revision arthroplasty.

Intervention and Comparison

The intervention was the preoperative administration of ESAs byany route, either alone or in combinationwith other interventions. Thecohorts were compared with those who received intravenous iron orplacebo, or had preoperative autologous blood donation (PAD).

Literature Search Strategy and Study Selection

We systematically searched the published and unpublishedmedical literature for randomized clinical trials (RCTs), comparingthe use of ESAs with any other intervention in patients whounderwent hip or knee surgeries. The Cochrane library, MEDLINE,EMBASE, the Database of Abstracts of Reviews of Effectiveness(DARE), Science Citation Index Expanded, Cumulative Index toNursing and Allied Health Literature (CINAHL), and ProQuest’sDissertations & theses (PQDT) databases were searched frominception through June 2012 without language restriction orlimitation on study design used. The following searched terms andkeywords were used as Medical Subject Heading terms and/orkeywords: “ Erythropoiesis-Stimulating Agents,” “ Erythropoietin,”“ Epo,” “ Epoetin alfa,” “ Procrit,” “ Epogen,” “ Epoetin beta,”“ NeoRecormon,” “ Darbepoetin alfa,” “ Mircera,” AND “replacementOR arthroplasty OR total AND ] Hip OR knee[”. We used Duggan’s

Table 1Characteristics of Included Studies.

Author Intervention Route Total Dose U, (IU/kg) Da

Ahmed, 2009 EPO & ANHD S.C 1200Akosy, 2001 EPO & PAD – 1500Avall, 2003 EPO & PAD S.C 50,000 IUc

Beris 1993 EPO & PAD S.C 900–1080Bezwada, 2003 EPO & PAD S.C 2400Biesma 1993 EPO & PAD S.C 3000Buljan, 2012 EPO & PAD IV 1200De Preeb, 1997 EPO & PAD S.C 900Gombotz, 2000 EPO & ANHD S.C 600 ± (600 if needed)Goodnough 1989 EPO & PAD IV 3600 2Goodnough, 1994 Epo& PAD IV 900, 1800, 3600Kikuchi, 1997 EPO& PAD S.C & IV 36,000 IU (IV, TKA)c

54,000 IU (IV, THA)c

48,000 IU (SC, TKA)c

72,000 IU (SC, THA)c

14 TK

Mastuda, 2001 EPO & PAD S.C 12,000–24,000 IUb

Mercuriali 1993 EPO & PAD IV 1800–3600Mercuriali, 1997 EPO & PAD IV & S.C 800 IU (S.C), 1800 (IV)Price 1996 EPO & PAD IV 3600Stowell, 1999 EPO & PAD S.C 2400EPO AloneCanadian Group, 1993 EPO S.C 2700, 4200Deutsch, 2006 EPO S.C 40,000c

Faris ,1996 EPO S.C (1500, 4500)Feagan, 2000 EPO S.C 160,000, 80,000c

Keating, 2007 EPO S.C 2400Na, 2011 EPO S.C 3000 (±3000)c OnOlijhoek, 2001 EPO S.C 1800Rosencher, 2005 EPO S.C 80,000 (±40,000)c

Weber, 2005 EPO S.C 120,000c

ANHD: Acute Normovolumic Hemodilution. PAD: Pre-Operative Autologous Donation.a Hematocrit converted to hemoglobin by multiplying by 3.3, a blank indicates that transb Orthopedic + other GI and vascular symptoms.c Not weight based dose.

filter to combine search results to achieve higher sensitivity ofcitations of clinical trials [10]. The terms used were : “effect:,”“trial,” “random,” “control,” “experimental,” “double blind,” “com-par,” “matched,” “blind,” “examine,” “study,” “comparative study,”and “randomized controlled trial,”. The search terms were limited tothe adult age group (N18 years old). To search for unpublished datafrom proceedings, the reviewers screened major orthopedic andhematology societies for abstracts from their annual meetings.Reference lists of retrieved review articles were hand searched tofind studies not identified from the electronic search. Onlyrandomized trials that described patients who received ESAs priorto surgery were included. There was no limitation on the blinding ofthe study design used (i.e. used placebo or open-label controls).There were no restrictions on language, so trials in English and non-English journals were included. We did not restrict our search tostudies which concomitantly used ESAs with other intervention (i.e.with IV iron or autologous blood transfusion or donation) to avoidexcluding trials that reported subgroup data on patients whounderwent hip or knee procedures. Abstracts were allowed if theyincluded adequate data for the analysis. Trials that includedchildren, surgeries other than hip or knee procedures, and non-randomized clinical trials were excluded.

Data Collection

A standardized abstraction form was developed and tested beforedata collection by GO & HM. Data from the studies were indepen-dently abstracted by two reviewers (GO & HM). Abstracted dataincluded data on patient and study characteristics, treatment in-terventions and outcomes. Any discrepancies were solved bydiscussion between the reviewers. In case the data were not providedin the text, we contacted the study’s primary authors for any missingdata or performed direct estimation from the graphs.

ys Preop Iron Transfusion threshold (Hgb: g L11)a Quality/GRADE

28 Oral – High14 Oral 80 Low21 Oral 85 Moderate28 Oral – Moderate21 I.V 80 Moderate21 Oral – Moderate17 Oral 80 Moderate21 I.V – High14 I.V 80 Moderate5–35 Oral – Low21 oral – Moderate

A & 21 THA Oral 70 Low

7–14 – – Low21 Oral , I.V – High21 IV Low21 Oral – High21 Oral – Moderate

10 Oral 90 High14 Oral – Moderate10 Oral 90 High28 Oral – Moderate21 Oral 80 ModerateOR Day IV 69 Low14 IV & Oral – High21 Oral 70 Low21 Oral – Moderate

fusion threshold not given.

Table 2Quality of Included Studies Using the “Risk of Bias” Assessment Tool From the Cochrane Handbook.

StudyAdequate Sequence

GenerationAllocation

Concealment? Blinding?Incomplete Outcome

Data AddressedFree of Selective

Outcome ReportingFree of

Other Bias?

1 Ahmed, 2009 Yes Yes Yes Yes Yes Yes2 Akosy, 2001 Unclear Unclear No Yes Yes Yes3 Avall, 2003 Unclear Unclear No No Yes Yes4 Beris, 1993 Yes Yes No Yes Yes Yes5 Bezwada, 2003 Unclear Unclear No Yes Yes Yes6 Biesma, 1993 Unclear Unclear Unclear Yes Yes Yes7 Buljan, 2012 Unclear Unclear No Yes Yes Yes8 Canadian Gp, 1993 Yes Yes Yes Yes Yes Yes9 DePree, 1997 Yes Yes Yes Yes Yes Yes10 Deutsch, 2006 Unclear Unclear No No Yes Yes11 Faris, 1996 Unclear Unclear Yes Yes Yes Yes12 Feagan, 2000 Yes Yes Yes Yes Yes No13 Gombotz, 2000 Unclear Unclear No Yes Yes Yes14 Goodnough, 1989 Unclear Unclear No Yes Yes No15 Goodnough, 1994 Unclear Unclear Yes Yes Yes Yes16 Keating, 2007 Yes Yes No Yes Yes Yes17 Kikouchi 1997 Unclear Unclear No Yes Yes Yes18 Mastuda, 2001 Unclear Unclear No Yes Yes Yes19 Mercuriali, 1993 Yes Yes Yes Yes Yes Yes20 Mercuriali, 1997 Unclear Unclear No Yes Yes Yes21 Na, 2011 Yes Yes No Yes Yes Yes22 Olijhoek, 2001 Yes Unclear Yes Yes Yes Yes23 Price, 1996 Yes Unclear Yes Yes Yes Yes24 Rosencher, 2005 Yes Yes No Yes Yes No25 Stowell, 1999 Yes Yes No Yes Yes Yes26 Weber, 2005 Yes Yes No Yes Yes Yes

Fig. 1.Overall VTE incident rate among both erythropoietinwith preoperative autologous blood donation (PAD) and erythropoietin alone groups expressed as risk differencewith 95%confidence interval.

1465K. Alsaleh et al. / The Journal of Arthroplasty 28 (2013) 1463–1472

Table 3Detailed Studies Characteristics and GRADE Grading for Included Studies.

Reference Pts (n)

Intervention

Grade and QualityESA Control

Ahmed et al, 2009 40 Perioperative erythropoietin (600 IU/kg up to a maximum dose of 40,000 IU/injection at28 and 14 days prior to surgery) with co-administration of pure iron oral medication

Autologous blood transfusion High(No important study limitations, NO IMPORTANT INCONSISTENCIES,DIRECT, NO important imprecision Undetected publication bias)++++

Akosy et al, 2001 40 Perioperative erythropoietin (300 U/kg twice a week) for 2 weeks preop and one dosewas given after operation.

Placebo Low(Serious study limitations, No important inconsistencies, Direct,Imprecision, Undetected publication bias) ++

Avall et al, 2003 38 Erythropoietin (10,000 U SQ twice weekly preop × 1 week) Autologous blood transfusion Moderate(Serious study limitations, No important inconsistencies, Direct,No important imprecision, Undetected publication bias)+++

Beris et al, 1993 101 Erythropoietin 150–180 U/kg (10,000 U) SC given three times per week4 & 2 weeks before surgery

Autologous blood donation Moderate(Serious study limitations, No important inconsistencies, Direct,No important imprecision, Undetected publication bias)+++

Bezwada et al, 2003 240 Erythropoietin 600 IU/kg subcut on Days 21, 14, 7, and oneday prior to surgery

Autologous blood donation Moderate(Serious study limitations, No important inconsistencies, Direct,No important imprecision, Undetected publication bias)+++

Biesma, 1993 40 Erythropoietin 500 U/kg of body weight twice a week for the 3 weeks before surgery ABD Moderate(Serious study limitations, No important inconsistencies, Direct,No important imprecision, Undetected publication bias)+++

Buljan et al, 2012 93 Erythropoietin 15,000 IU was administered intravenouslytwice a week or 30,000 IU once a week (total 90,000 IU)Plus iron

Iron Moderate(Serious study limitations, No important inconsistencies, Direct,No important imprecision, Undetected publication bias)+++

Canadian Gp, 1993 208 Two regimens:1. 14 days of erythropoietin (300 U/kg to a maximum of 30,000 U)OR2. Erythropoietin for the 9 days after placebo.

Placebo High(No important study limitations No important inconsistencies DirectNo important imprecision Undetected publication bias)++++

De Pree et al, 1997 50 Erythropoietin (141 U/kg) six times subcutaneously over a 14-day period(between Days −21 and −7).N.B: patients with body weight b60 kg received 8000 U (148.1 ±9.6 U/kg);those with body weight of 60 to 80 kg, 10,000 U (149.2 ±12.6 U/kg); and thosewith body weight of N80 kg, 12,000 U (133.2 ±10.3 U/kg)

Placebo High(No important study limitations No important inconsistenciesDirectNo important imprecision Undetected publication bias)++++

Deutsch et al, 2006 52 Erythropoietin alfa 40,000 U at preoperative days 14 and 7 Autologous blood donation Moderate(Serious study limitations, No important inconsistencies, Direct,No important imprecision, Undetected publication bias)+++

Faris et al , 1996 200 Two regimens:1. Erythropoietin (300 U/kg) dailyOR2. Erythropoietin 100 U/kg dailyA total of fifteen doses were given subcutaneously, beginning ten days before the operationand extending through the fourth postoperative day

Placebo High(No important study limitationsNo important inconsistenciesDirectNo important imprecisionUndetected publication bias)++++

1466K.A

lsalehet

al./The

JournalofArthroplasty

28(2013)

1463–1472

Feagan, 2000 201 Four weekly doses of erythropoietin alfa 40, 000 U (high-dose) or 20, 000 U (low dose) Placebo Moderate(No important study limitations, No important inconsistencies,Direct, No important imprecision, Strongly suspected publication bias)+++

Gombotz, 2000 60 Erythropoietin 600 U/kg s.c. on day 14 and, if needed, on day 7 before surgery Autologous blood donation Moderate(Serious study limitations, No important inconsistencies, Direct,No important imprecision, Undetected publication bias)+++

Goodnough, 1989 47 Erythropoietin 600 IU/kg twice weekly for 21 days. Placebo Low(Serious study limitations, No important inconsistencies, Direct,No important imprecision, Possible publication bias)++

Goodnough, 1994 116 28 (18) patients received EPO at 600 U/kg. 30 (26) patients received EPO at 300 U/kg, 29(24) patients received EPO at 150 U/kg.

Placebo Moderate(Serious study limitations, No important inconsistencies, Direct,No important imprecision, Undetected publication bias)+++

Keating, 2007 287 Erythropoietin 600 IU/kg once weekly for 4 doses Autologous blood donation Moderate(Serious study limitations, No important inconsistencies, Direct,No important imprecision, Undetected publication bias)+++

Kikuchi 120 2 groups: Erythropoietin 6000 U ( three times a week for 2 weeks in those undergoingtotal knee arthroplasty (TKA), and for 3 weeks in those undergoing total hip arthroplasty(THA)). 24,000 U Epo was given subcutaneously to second group S once a week(for 2 weeks for TKA and for 3 weeks for THA).

Autologous blood donation Low(Serious study limitations, No important inconsistencies, Direct,Imprecision, Undetected publication bias)++

Mastuda, 2001 73 Three dosing regimens:Erythropoietin 1200 U subcut 1 week before phlebotomy and 1 week afterErythropoietin 1200 U subcut 1 week before phlebotomyErythropoietin 1200 U subcut weekly 2 weeks before phlebotomy

Autologous blood donation Low(Serious study limitations, No important inconsistencies,Direct, Imprecision, Undetected publication bias)++

Mercuriali, 1993 50 Erythropoietin 300 IU/kg & 600 IU/kg on day 1 then every 3–4 days for 21 days(total of six doses)

Placebo (saline solution) High(No important study limitationsNo important inconsistenciesDirectNo important imprecisionUndetected publication bias)++++

Mercuriali, 1997 22 + 12 Total Epoetin alfa s.c. dose of 800 IU/kg was compared with a total i.v. dose of 1,800 IU/kg(i.v. bolus of Epoetin alfa 200 IU/kg immediately followed by an s.c. dose of 100 IU/kg.For the following 5 visits, s.c. Epoetin alfa 100 IU/kg was administered. Group 2: i.v.Epoetin alfa 300 IU/kg twice a week for 3 weeks at each visit (total Epoetin alfa dose of 1800IU/kg).

Iron Low(Serious study limitations, No important inconsistency, DirectImprecision, Undetected publication bias)++

Na, 2011 108 Erythropoietin 3000 IU subcutaneously during the operation and during the postoperativeperiod if the hemoglobin level was 70–80 g/L.

Placebo Low(Serious study limitations, No important inconsistency, Direct,No important imprecision,Strongly suspected publication bias)++

Olijhoek, 2001 110 Erythropoietin alfa (600 IU/kg) subcutaneously on Days 1 and 8 of the 14-day study Placebo HighNo important study limitationsNo important inconsistenciesDirectNo important imprecision

(continued on next page)

1467K.A

lsalehet

al./The

Journalof

Arthroplasty

28(2013)

1463–1472

Table3(con

tinu

ed)

Referenc

ePts(n

)

Interven

tion

Grade

andQua

lity

ESA

Control

Und

etectedpu

blicationbias

++++

Price,

1996

204

Erythr

opoietin

600IU

/kgIV

21da

yspreo

p×

6do

ses

Placeb

oHigh

Noim

portan

tstud

ylim

itations

Noim

portan

tinco

nsistenc

ies

Direct

Noim

portan

tim

precision

Und

etectedpu

blicationbias

++++

Rosenc

her,20

0593

Erythr

opoietin

40,000

UI/wee

ksc

untilthe

yreache

damax

imal

Htof

40%

Autolog

ousbloo

ddo

nation

Low

(Serious

stud

ylim

itations

,Noim

portan

tinco

nsistenc

y,Direct,

Noim

portan

tim

precision,

Strong

lysu

spectedpu

blicationbias)

++

Stow

ell,19

9949

0W

eeklydo

sesof

subc

utan

eous

erythr

opoietin

alfa

(600

U/kg)

onpreo

perative

Day

s−

21,−

14,a

nd−

7,an

don

theda

yof

surgery.

Autolog

ousbloo

ddo

nation

Mod

erate

(Serious

stud

ylim

itations

,Noim

portan

tinco

nsistenc

ies,

Direct,Noim

portan

tim

precision,

Und

etectedpu

blicationbias)

+++

Web

er,2

005

704

Erythr

opoietin

alfa

40,0

00IU

subc

utan

eous

lyon

cewee

klyfor3wee

ksbe

fore

surgeryan

don

theda

yof

surgery

Iron

Mod

erate

(Serious

stud

ylim

itations

,Noim

portan

tinco

nsistenc

ies,Direct,

Noim

portan

tim

precision,

Und

etectedpu

blicationbias)

+++

1468 K. Alsaleh et al. / The Journal of Arthroplasty 28 (2013) 1463–1472

Quality Assessment

The studies retrieved by the search strategy were reviewed by tworeviewers (GO and KS). Relevant studies were selected and criticallyappraised using the criteria recommended by the internationalguidelines [11]. We evaluated the quality of the included trials usingboth Cochrane Collaboration’s tool for assessing risk of bias inrandomized trials [12] and the Grading of RecommendationsAssessment, Development and Evaluation (GRADE) Working Groupgrading scheme [13].

Outcomes

Included studies had to report at least one of the pre-specifiedprimary outcomes of interest. These included the effect of preoper-ative ESA administration on the level of hemoglobin at discharge oron day 7 post operatively, the number of patients requiringallogeneic blood transfusion, and the incident rate of thrombosis orvenous-thromboembolism(VTE) as a result of ESAs use. A pre-planned subgroup analysis for the studies used ESA vs. ESA withpreoperative autologous blood donation (PAD) since the need fortransfusion and level of hemoglobin might be substantially differentbetween the two groups. A sensitivity analysis based on the studies’quality is planned to be done to explore the cause of heterogeneity incase it is found.

Statistical Analysis

The statistical analysis was conducted using the ReviewManager software (RevMan, version 5.1: from The CochraneCollaboration, http://www.cochrane.org). For the meta-analysis,the estimated effect size and variation were expressed as relativerisks (RR) and their respective 95% confidence intervals (CI) foreach study and for the pooled studies. The standard deviation wasobtained from the paper whenever provided or estimated from thestudy graphs, whenever the error bars were provided with nolabeling, we contacted the primary author of the study, if thisfailed we calculated the standard deviation from the data providedin the article using the equation [Standard Deviation = Standarderror of the mean/√sample number[. Heterogeneity was alsoassessed using the Cochrane chi-square (Q) and the correspondingI2 score to examine any possible heterogeneity between studies. ACochrane’s Q P b 0.10 and I2 N 50% were considered to showsignificant heterogeneity. The random effects model assumptionwas used to adjust for within and between study heterogeneity.Forest plots were used to illustrate the individual studies, theirfinal pooled effect size, and each individual study’s weight (whichis based on the inverse of variance). Continuous data, such asphysiologic parameters (e.g. hemoglobin)were reported as mean.We summarized continuous data as mean difference with 95%confidence intervals. We converted hemoglobin (Hg) from hemat-ocrit (Hct) if only Hct was given, by using a standard publishedequation (Hg [g/dl] = Hct [%]/3) [14]. A funnel plot was used todetect publication bias. A preplanned subgroup analysis for thetrials with and without preoperative autologous blood donation(PAD) intervention was conducted to see if this modified theoutcomes of using ESAs.

Results

Twenty-six trials met the pre-specified inclusion criteria [15–40].The 26 trials comprised 3560 patients who were scheduled for anytype of hip or knee arthroplasty. Total hip arthroplasty was morecommon than knee arthroplasty. Seventeen trials, comprising 1666patients, used ESA alone, and 9 trials, comprising 1897 patients, usedESAs in addition to preoperative autologous blood donation (PAD).

Fig. 2.Hemoglobin (Hg) level changes at discharge (or last recorded) in both erythropoietin with preoperative autologous blood donation (PAD) and erythropoietin alone expressedas mean difference with 95% confidence interval. Some Hg levels were estimated from Hct according to the equation mentioned.

1469K. Alsaleh et al. / The Journal of Arthroplasty 28 (2013) 1463–1472

Table 1 shows a summary of the characteristics of the trials, dosingregimen of ESAs, and GRADE score. Cochrane risk of bias assessmentfor each study is included in Table 2 and Fig. 1 (supplementary), moredetailed information about the studies and GRADE grading is includedin Table 3 (supplementary).

Primary Efficacy Outcomes

Transfusion RequirementsTwenty-five of the included trials measured the percentage of

patients who needed allogeneic blood transfusion (ABT). Our resultsshowed that preoperative administration of ESAs was associated with astatistically significant reduction in the number of patients requiringABT [RR=0.48, 95% Confidence Interval [CI]: 0.38 to 0.60, P b 0.00001)(Fig. 2)]. There was significant heterogeneity among overall studies butdriven by significant heterogeneity in the studies testing ESA alone.One trial [30] did not report any events in either the intervention orcontrol arm and thus did not contribute to the aggregated estimate. Asensitivity analysis was done to explore the difference by omitting thetrials with low quality, and heterogeneity was reduced (I2 = 27%)when 2 trials [23,39] in the ESA alone arm were excluded.

Hemoglobin at Discharge (or Last Measured)Results from 21 studies that reported postoperative hemoglobin

level (12 in the ESA & PAD group, and 9 from ESA alone group)showed that the mean difference in hemoglobin level between theESA (both alone and with PAD) and control groups was 7.16 g/L[95% CI of 4.73 to 9.59, P = 0.00001] (Fig. 3). ESAs resulted in higherhemoglobin levels at discharge when compared to the controlgroup. Heterogeneity was high (I2 = 88%), a pre-planned sensitivity

analysis to exclude studies with low quality failed to reduce theheterogeneity significantly.

The funnel plot for studies comparing hemoglobin difference atdischarge (Fig. 4, supplementary) shows that there may be publica-tion bias, but there are other possible explanations for the asymmetrylike the quality of the reporting studies as most were small studieswith low to moderate quality.

Primary Safety Outcomes

Thromboembolic EventsTo allow for the inclusion of trials which reported zero occurrence

of the event, risk difference was utilized. The low rate of events mightbe explained by the aggressive use of VTE prophylaxis and theinclusion of relatively well patients in most of the trials which mayhave allowed for early mobilization and a reduction in the risk ofthromboembolic disease. There was no difference in the risk ofdeveloping venous-thromboembolismbetween the ESA group and thecontrol groups [RD 0, 95 % CI: −1%–2%, P = 0.95; I2 = 0%] (Fig. 5).

Discussion

In this systematic review, the preoperative use of erythropoiesis-stimulating agents (ESAs) in patients undergoing hip or knee surgerywas examined. ESAs, given with or without PAD, significantly reducedthe need for ABT. Our results are similar to a previous meta-analysisthat looked into the role of ESA in preventing perioperativetransfusion [9]. The frequency of transfusion in the control groupwas extremely variable and ranged from zero to 100%. There has beenso much effort invested in finding an alternative to ABT, especially in

Fig. 3.Number of patients who required allogeneic blood transfusion in both erythropoietin with preoperative autologous blood donation and erythropoietin alone expressed as riskratio with 95% confidence interval.

Fig. 4. Funnel plot to detect publication bias in studies comparing hemoglobindifference at discharge. Erythropoietin with preoperative autologous blood donation(PAD) studies are black boxes. Erythropoietin alone studies are red diamond shaped.Asymmetry indicates bias mostly found in studies reported with use of EPO alone.

1470 K. Alsaleh et al. / The Journal of Arthroplasty 28 (2013) 1463–1472

orthopedic surgery, where blood loss occurs quite frequently. Manyblood management strategies have been proposed hoping to avoid orminimize the need for ABT. One of the methods used to reduce theneed for ABT is PAD. However, PAD itself has its own drawbacks thatnecessitate the search for an easier alternative. One of the mostimportant drawbacks of PAD is that autologous blood is moreexpensive than allogeneic blood. The higher cost includes the extratime and attention required by the autologous donor; the extraadministrative and clerical requirements; and the need for specialhandling like separate storage, etc. [41]. Therefore, the use of ESAsseems to be an optimal strategy for a paradigm shift in perioperativeblood management in patients undergoing hip or knee surgery. Ourreview demonstrates that ESAs are an effective strategy for reducingABT requirements, and also in maintaining hemoglobin levelspostoperatively. There was no evidence of increased efficacy if PADwas combined with ESA which might suggest that PAD might not benecessary with an ESA. Significant heterogeneity was seen in theincluded studies in this review, mostly attributed to the differentscheduling patterns and transfusion thresholds. Another possibleexplanation is the presence of possible publication bias especially inthe trials which utilized ESAs alone (Fig. 4).

The concern of the risk of thrombosis associated with ESAs wasraised in a couple of trials where the incidence of VTE was higher inthe group that received preoperative ESA [15,24]. This concern limitedthe wide use of ESAs in the preoperative period to reduce the need forABT. However, the association between ESAs and the development ofany thromboembolic event was never proven [42]. Our study showedno significant difference in VTE incidence between ESA and controlgroups. Information about the use of VTE prophylaxis was notprovided in any of the included studies. Orthopedic surgeries,

especially hip and knee surgeries, are associated with higher risk ofthrombosis and that should be considered when assessing the risk ofthrombosis regardless of the medication administered.

Fig. 5. Cochrane risk of bias assessment for included studies.

1471K. Alsaleh et al. / The Journal of Arthroplasty 28 (2013) 1463–1472

Another major contributor to the gap between evidence andpractice is related to the perceived lack of cost-effectiveness of ESAsuse in this indication. In 1998, the Canadian Coordinating Office forHealth Technology Assessment performed an economic evaluationof erythropoietin use in cardiac and orthopedic surgery. In theirreport they calculated the total costs of erythropoietin for theCanadian health care to be 5.9 million dollars annually in order tohave a potential benefit of 0.12 life year gained for the totalpopulation [43]. Another study showed no acceptable benefit on thebasis of cost-effectiveness [44]. However, both analyses were donein Canada based on the mid-nineties prices. The analyses also didnot include other factors like patients anxiety from transfusionwhich could change the analysis value. This review suggests thatESAs are able to reduce the risk of ABT by half which could result ina significant reduction of the blood transfusion utilization and costsince orthopedic procedures represent a major source of bloodproduct utilization [45]. Perioperative anemia is prevalent inpatients undergoing major orthopedic surgery and is associatedwith adverse clinical outcomes which may be attenuated with animproved hemoglobin level [46]. Hence, to balance between costand much needed intervention, indication criteria should beestablished for the preoperative use of ESAs in orthopedic patientsundergoing hip or knee surgeries. Some literature suggest Hb levelbetween 10 and 13 gm/dL as a trigger for preoperative ESAs [42,47].Another appropriate indication criterion could include expectedblood loss [47].

Our meta-analysis has some limitations. One major limitation isthe lack of detailed information about the study results relevant to ouroutcomes, which might have contributed to the variation in thestudies and significant heterogeneity in some of the outcomes orexplicit standardized triggers for transfusions. Most reports ofindividual studies included only summary results, such as means orgraphs. Lack of detailed information about hemoglobin level fromindividual studies precluded us from including some of the trials inthe comparison of effect. Many investigators were contacted forinformation but no response was received.

There are also other limitations that may have affected ourfindings including the lack of double blind randomized controlledmethodology. Moreover, many studies were underpowered to detecta clinically significant difference between groups (with wideconfidence intervals around estimates). Finally, future studies areneeded to address the optimal dose of ESA needed to prevent patient'sexposure to either allogeneic or autologous blood or the use of ESAversus other alternatives like fibrinolytic inhibitors.

Conclusion

Erythropoietin improves postoperative hemoglobin levels anddecreases the need for allogeneic blood transfusion in patientsundergoing hip or knee surgery.

Acknowledgment

This study was supported by the College of Medicine ResearchCenter, Deanship of Scientific Research, King Saud University, Riyadh,Saudi Arabia.

References

1. Goodnough L, Brecher M, Kanter M, et al. Transfusion medicine—blood transfusion.N Engl J Med 1999;340(6):438.

2. Goodnough L, Monk T, Andriole G. Erythropoietin therapy. N Engl J Med 1997;336(13):933.

3. Tobias JD. Strategies for minimizing blood loss in orthopedic surgery. SeminHematol 2004;41(1 Suppl 1):145.

4. Horstmann W, Ettema H, Verheyen C. Dutch orthopedic blood managementsurveys 2002 and 2007: an increasing use of blood-saving measures. Arch OrthopTrauma Surg 2010;130(1):55.

1472 K. Alsaleh et al. / The Journal of Arthroplasty 28 (2013) 1463–1472

5. Rashiq S, Jamieson-Lega K, Komarinski C, et al. Allogeneic blood transfusionreduction by risk-based protocol in total joint arthroplasty. Can J Anaesth 2010;57(4):343.

6. Fergusson D, Blair A, Henry D, et al. Technologies to minimize blood transfusion incardiac and orthopedic surgery. Results of a practice variation survey in ninecountries. International Study of Peri-operative Transfusion (ISPOT) Investigators.Int J Technol Assess Health Care 1999;15(4):717.

7. Green W, Toy P, Bozic K. Cost minimization analysis of preoperative erythropoi-etin vs autologous and allogeneic blood donation in total joint arthroplasty.J Arthroplasty 2010;25(1):93.

8. Fergusson D, Hébert P. The health(y) cost of erythropoietin in orthopedic surgery.Can J Anaesth 2005;52(4):347.

9. Laupacis A, Fergusson D. Erythropoietin to minimize perioperative bloodtransfusion: a systematic review of randomized trials. The International Study ofPeri-operative Transfusion (ISPOT) Investigators. Transfus Med 1998;8(4):309.

10. McKibbon K, Wilczynski N, Haynes R. Retrieving randomized controlled trialsfrom Medline: a comparison of 38 published search filters. Health Info Libr J 2009;26(3):187.

11. Scottish Intercollegiate Guidelines Network (SIGN). Critical appraisal: notes andchecklists. 2012.

12. Cochrane handbook for systematic reviews of interventions version 5.1.0 [updatedMarch 2011]. In: Higgins JPT, Green S, editors. The Cochrane Collaboration; 2011.Available from: www.cochrane-handbook.org.

13. Guyatt G, Oxman A, Vist G, et al. GRADE: an emerging consensus on rating quality ofevidence and strength of recommendations. BMJ 2008;336(7650):924.

14. Quintó L, Aponte J, Menéndez C, et al. Relationship between haemoglobin andhaematocrit in the definition of anaemia. Trop Med Int Health 2006;11(8):1295.

15. Ahmed SA, Abdulsattar M. Two-shoot erythropoietin injection coupled with acutenormovolumic hemodilution as allogenic blood transfusion sparing strategy. KaserEl Aini Med J 2009;15(4):35.

16. Avall A, Hyllner M, Bengtson J, et al. Recombinant human erythropoietin inpreoperative autologous blood donation did not influence the haemoglobinrecovery after surgery. Acta Anaesthesiol Scand 2003;47(6):687.

17. Beris P, Mermillod B, Levy G, et al. Recombinant human erythropoietin as adjuvanttreatment for autologous blood donation: a prospective study. Vox Sang 1993;65(3):212.

18. Bezwada H, Nazarian D, Henry D, et al. Preoperative use of recombinant humanerythropoietin before total joint arthroplasty. J Bone Joint Surg Am 2003;9:1795.

19. Biesma D, Kraaijenhagen R, Marx J, et al. The efficacy of subcutaneous recombinanthuman erythropoietin in the correction of phlebotomy-induced anemia inautologous blood donors. Transfusion 1993;33(10):825.

20. Buljan M, Nemet D, Golubic-Cepulic B, et al. Two different dosing regimens ofhuman recombinant erythropoietin beta during preoperative autologous blooddonation in patients having hip arthroplasty. Int Orthop 2012;36(4):703.

21. Canadian Orthopedic Perioperative Erythropoietin Study Group. Effectiveness ofperioperative recombinant human erythropoietin in elective hip replacement.Lancet 1993;341(8855):1227.

22. de Pree Mermillod B, Hoffmeyer P, Beris P. Recombinant human erythropoietin asadjuvant treatment for autologous blood donation in elective surgery with largeblood needs (Nor = 5 units): a randomized study. Transfusion 1997;37(7):708.

23. Deutsch A, Spaulding J, Marcus R. Preoperative epoetin alfa vs autologous blooddonation in primary total knee arthroplasty. J Arthroplasty 2006;21(5):628.

24. Faris P, Ritter M, Abels R. The effects of recombinant human erythropoietin onperioperative transfusion requirements in patients having a major orthopaedicoperation.The American Erythropoietin Study Group. J Bone Joint Surg Am 1996;78(1):62.

25. Feagan B, Wong C, Kirkley A, et al. Erythropoietin with iron supplementation toprevent allogeneic blood transfusion in total hip joint arthroplasty. A randomized,controlled trial. Ann Intern Med 2000;133(11):845.

26. Gombotz H, Gries M, Sipurzynski S, et al. Preoperative treatment with recombinanthuman erythropoietin or predeposit of autologous blood in women undergoingprimary hip replacement. Acta Anaesthesiol Scand 2000;44(6):737.

27. Goodnough L, Rudnick S, Price T, et al. Increased preoperative collection ofautologous blood with recombinant human erythropoietin therapy. N Engl J Med1989;321(17):1163.

28. Goodnough L, Price T, Friedman K, et al. A phase III trial of recombinant humanerythropoietin therapy in nonanemic orthopedic patients subjected to aggressiveremoval of blood for autologous use: dose, response, toxicity, and efficacy.Transfusion 1994;34(1):66.

29. Keating E, Callaghan J, Ranawat A, et al. A randomized, parallel-group, open-labeltrial of recombinant human erythropoietin vs preoperative autologous donation inprimary total joint arthroplasty: effect on postoperative vigor and handgripstrength. J Arthroplasty 2007;22(3):325.

30. Kikuchi H, Tan A, Nonaka T, et al. Comparison of intravenous and subcutaneouserythropoietin therapy for preoperative acquisition of blood for autologoustransfusion in patients undergoing total arthroplasty. J Orthop Sci 1997;2(2):84.

31. Matsuda S, KondoM, Mashima T, et al. Recombinant human erythropoietin therapyfor autologous blood donation in rheumatoid arthritis patients undergoing total hipor knee arthroplasty. Orthopedics 2001;24(1):41.

32. Mercuriali F, Zanella A, Barosi G, et al. Use of erythropoietin to increase the volumeof autologous blood donated by orthopedic patients. Transfusion 1993;33(1):55.

33. Mercuriali F, Inghilleri G, Biffi E, et al. Comparison between intravenous andsubcutaneous recombinant human erythropoietin (Epoetin alfa) administration inpresurgical autologous blood donation in anemic rheumatoid arthritis patientsundergoing major orthopedic surgery. Voxsanguinis 1997;72(2):93.

34. Na H, Shin S, Hwang J, et al. Effects of intravenous iron combined with low-doserecombinant human erythropoietin on transfusion requirements in iron-deficientpatients undergoing bilateral total knee replacement arthroplasty. Transfusion 2011;51(1):118.

35. Olijhoek G, Megens J, Musto P, et al. Role of oral versus IV iron supplementation inthe erythropoietic response to rHuEPO: a randomized, placebo-controlled trial.Transfusion 2001;41(7):957.

36. Price T, Goodnough L, Vogler W, et al. The effect of recombinant humanerythropoietin on the efficacy of autologous blood donation in patients with lowhematocrits: a multicenter, randomized, double-blind, controlled trial. Transfusion1996;36(1):29.

37. Rosencher N, Poisson D, Albi A, et al. Two injections of erythropoietin correctmoderate anemia in most patients awaiting orthopedic surgery. Can J Anaesth2005;52(2):160.

38. Stowell C, Chandler H, Jové M, et al. An open-label, randomized study to comparethe safety and efficacy of perioperative epoetin alfa with preoperative autologousblood donation in total joint arthroplasty. Orthopedics 1999;22(1 Suppl):s105.

39. Weber E, Slappendel R, Hémon Y, et al. Effects of epoetin alfa on blood transfusionsand postoperative recovery in orthopaedic surgery: the European Epoetin AlfaSurgery Trial (EEST). Eur J Anaesthesiol 2005;22(4):249.

40. Aksoy M, Tokgozoglu A. Erythropoietin for autologous blood donation in total hiparthroplasty patients. Arch Orthop Trauma Surg 2001;121(3):162.

41. Silvergleid A, Waltham M, Kleinman S. Preoperative autologous blood donation2012. Available online [accessed August 25, 2012]. http://www.sign.ac.uk/methodology/checklists.html.

42. van Erve Ruud HGP, Wiekenkamp Alexander C. Available from, In: Kochhar Puneet,editor. Transfusion reduction in orthopedic surgery, blood transfusion in clinicalpractice; 2012ISBN: 978-953-51-0343-1, InTech. http://www.intechopen.com/books/blood-transfusion-in-clinical-practice/transfusion-reduction-inorthopedic-surgery.

43. Otten N. Economic evaluation of erythropoietin use in surgery. Ottawa: CanadianCoordinating Office for Health Technology Assessment (CCOHTA); 1998.

44. Coyle D, Lee KM, Fergusson DA, et al. Economic analysis of erythropoietin use inorthopaedic surgery. Transfus Med 1999;9(1):21.

45. Feagan B, Wong C, Lau C, et al. Transfusion practice in elective orthopaedic surgery.Transfus Med 2001;11(2):87.

46. Spahn D. Anemia and patient blood management in hip and knee surgery: asystematic review of the literature. Anesthesiology 2010;113(2):482.

47. Dubois RW, Lim D, Hebert P, et al. The development of indications for thepreoperative use of recombinant erythropoietin. Can J Surg 1998;41(5):351.