Rev Esp Anestesiol Reanim.

2015;62(5):253---264

Revista Española de Anestesiología

y Reanimación
www.elsevier.es/redar

ORIGINAL ARTICLE

Efficiency of tranexamic acid in perioperative blood

loss in hip arthroplasty: A systematic literature review
and meta-analysis夽,夽夽,夽夽夽
C.E. Pinzón-Florez a,∗ , K.M. Vélez Cañas b , D.M. Díaz Quijano a

a
Grupo de Investigación Clínica, Facultad de Medicina, Universidad de La Sabana, Chía, Cundinamarca, Colombia
b
Programa de Anestesiología, Facultad de Medicina, Universidad del Rosario, Bogotá, DC, Colombia

Received 3 June 2014; accepted 1 October 2014

Available online 17 April 2015

KEYWORDS Abstract
Tranexamic acid; Introduction and objective: Tranexamic acid (TXA) is an antifibrinolytic drug used to reduce
Hip arthroplasty; bleeding in mortality risk situations such as trauma. Our objective was to conduct a systematic
Antifibrinolytic; literature review to evaluate the effectiveness and safety of TXA in reducing bleeding in hip
Total blood loss; arthroplasty.
Allogeneic Materials and methods: A systematic literature review and meta-analysis of primary studies
transfusion similar to controlled trials was performed. Literature was searched in MEDLINE, Embase,
Cochrane, LILACS, SciELO and Google Scholar. The review was proposed and undertaken by
two reviewers and the inclusion criteria were: (a) patients undergoing arthroplasty for pri-
mary unilateral hip replacement; (b) comparison of a treatment group with TXA to a control
group that received a placebo or no treatment at all, and (c) outcome measures, total blood
loss, number of patients receiving allogeneic transfusion and/or incidence of thromboembolic
complications. The search was restricted to studies published from 1966 to June 2013.
Results: A total of 16 studies with 246 patients were retrieved for this review. The total blood
loss outcome evidenced a weighted mean difference in favor of TXA vs. controls undergoing hip
arthroplasty (−0.45 [P < 0.001, 95% CI −0.65 to −0.24]). Weighted relative risk was estimated
for the allogeneic transfusion requirement outcome, showing a trend in favor the TXA arm, with
fewer patients requiring allogeneic transfusion in hip surgery (0.8 [P < 0.02, 95% CI 0.57---1.11]);
however, this trend was not statistically significant.

夽 Please cite this article as: Pinzón-Florez CE, Vélez Cañas KM, Díaz Quijano DM. Efectividad del ácido tranexámico en las pérdidas

sanguíneas perioperatorias en la artroplastia de cadera: revisión sistemática de la literatura y metaanálisis. Rev Esp Anestesiol Reanim.
2015;62:253---264.
夽夽 This study is part of the doctoral thesis in anesthesiology submitted by Dr. Karina María Vélez.
夽夽夽 This article is part of the Anaesthesiology and Resuscitation Continuing Medical Education Program. An evaluation of the questions on

this article can be made through the Internet by accessing the Education Section of the following web page: www.elsevier.es/redar.
∗ Corresponding author.

E-mail address: cepinzon@gmail.com (C.E. Pinzón-Florez).

2341-1929/© 2014 Sociedad Española de Anestesiología, Reanimación y Terapéutica del Dolor. Published by Elsevier España, S.L.U. All rights
reserved.
254 C.E. Pinzón-Florez et al.

Limitations: There is a noticeable difference in methods for quantifying total blood loss across
the studies reviewed. The need for transfusion outcomes is probably not significant taking into
account the number of events in the TXA group.
Conclusions: TXA can be routinely used to reduce intra- and post-operative blood loss in primary
hip arthroplasty.
© 2014 Sociedad Española de Anestesiología, Reanimación y Terapéutica del Dolor. Published
by Elsevier España, S.L.U. All rights reserved.

PALABRAS CLAVE Efectividad del ácido tranexámico en las pérdidas sanguíneas perioperatorias en la
Ácido tranexámico; artroplastia de cadera: revisión sistemática de la literatura y metaanálisis
Artroplastia
de cadera; Resumen
Antifibrinolítico; Introducción y objetivo: El ácido tranexámico (ATX) es un fármaco antifibrinolítico utilizado
Pérdida de sangre para la reducción del sangrado en situaciones de riesgo de mortalidad como es el trauma. El
total; objetivo es evaluar la efectividad y seguridad del ATX en la reducción del sangrado operatorio
Transfusión alogénica en la artroplastia de cadera mediante una revisión sistemática de la literatura.
Material y métodos: Se realizó una revisión sistemática de la literatura y un metaanálisis de
estudios primarios tipo ensayos clínicos controlados. La búsqueda de la literatura fue realizada
en MEDLINE, Embase, Cochrane, LILACS, SciELO y Google Scholar. La revisión fue propuesta y
realizada por 2 revisores y los criterios de inclusión fueron: a) pacientes sometidos a artroplastia
de cadera unilateral primaria; b) comparación de un grupo de tratamiento con ATX frente a un
grupo de control que recibió un placebo o ningún tratamiento en absoluto; c) las medidas de
resultado: pérdida sanguínea total, número de pacientes sometidos a transfusión alogénica y/o
incidencia de complicaciones tromboembólicas. La búsqueda de estudios se realizó desde el
año 1966 a junio de 2013.
Resultados: Un total de 16 estudios fueron contemplados para esta revisión, con 246 pacientes
incluidos. Se evidenció en el desenlace de pérdida sanguínea total un efecto ponderado de
diferencia de promedios a favor de TXA vs. control de los pacientes sometidos a artroplastia
de cadera (−0.45 [p < 0,001, IC 95% −0,65 a −0,24]). Se estimó el riesgo relativo ponderado
para el desenlace de requerimiento de transfusión alogénica, el cual evidenció una tendencia
a favorecer al brazo de ATX con un número de pacientes que requieren menos transfusión
alogénica en cirugía de cadera (0,8 [p < 0,02, IC 95% 0,57 a 1,11]), sin embargo, esta tendencia
no fue estadísticamente significativa.
Limitaciones: Existe una diferencia notable en los métodos de medición para la cuantificación
de las pérdidas sanguíneas totales entre los estudios. El desenlace de necesidad de transfusión
no es significativo, probablemente por el número de eventos en el grupo de ATX.
Conclusiones: El ATX se puede considerar de uso rutinario en la artroplastia primaria de cadera
para reducir la pérdida de sangre intra y posquirúrgica.
© 2014 Sociedad Española de Anestesiología, Reanimación y Terapéutica del Dolor. Publicado
por Elsevier España, S.L.U. Todos los derechos reservados.

Introduction to the surgical wound. TXA reduces risks, and consequently

Hip arthroplasty is an increasingly common surgical proce-
dure that is associated with major blood loss,1---9 which in
turn increases the risk of transfusion.5,9,10 Blood loss often
leads to serious postoperative anemia,11 which puts patients
at risk of cardiopulmonary events and transfusion reactions,
and increases health care costs.5,12 Allogenic transfusion,
meanwhile, can also increase the risk of postoperative
infection.3,10,13,14
Tranexamic acid (TXA) is a synthetic amino acid6---8
that blocks lysine binding sites on plasminogen molecules,
thereby inhibiting plasminogen to plasmin conversion.10---15
TXA can be administered intravenously or applied topically
bleeding, and is effective in major orthopedic interven-
tions such as spinal surgery11 and knee arthroplasty.15,17---19
TXA, however, has the same drawbacks as other antifibri-
nolytics, namely, it can increase the risk of thromboembolic
complications such as deep vein thrombosis (DVT), acute
myocardial infarction, or pulmonary thromboembolism.
We performed a meta-analysis of randomized controlled
trials to evaluate the efficacy of TXA in hip arthroplasty.
Study outcomes were: total blood loss, number of patients
receiving allogenic transfusion, and incidence of DVT.
We hypothesized that TXA in hip arthroplasty significantly
reduces blood loss and the need for allogenic blood transfu-
sion, without increasing incidence of DVT.
Efficacy of tranexamic acid in hip arthroplasty 255

Materials and methods Inclusion criteria

We performed a systematic review of randomized clinical
trials and studies in order to evaluate the effectiveness of
TXA in reducing intraoperative bleeding in patients under-
going hip arthroplasty We searched for primary studies in
all available scientific databases using MeSH terms and All
Fields for each topic in the PICOT list of research questions.
Following this, each study retrieved was evaluated using
the Cochrane Collaboration’s tool for assessing risk of bias.
Once the evidence had been evaluated, the relevant infor-
mation was extracted and a meta-analysis was performed
to evaluate the combined effect of TXA on the ‘‘need for
transfusion’’ outcome. Research protocol (Appendix A).
Population
Adults (aged between 18 and 60 years) indicated for hip
arthroplasty, ASA I---III.
Type of surgery
Studies evaluating TXA vs placebo, dummy treatment, no
treatment, standard treatment.
Outcomes
(a) Studies evaluating the efficacy of TXA in the man-
agement of intraoperative bleeding in hip arthroplasty
surgery.
(b) More than 10 patients included in the study.
(c) Randomization and allocation concealment method
described in the materials and methods section.
Exclusion criteria
(a) Non-randomized study.
(b) Control group consists of placebo or dummy treatment,
no treatment, standard treatment.12
(c) Initial TXA dose not reported.
Search procedure
The studies to be included in this meta-analysis were
selected from a search made of all available databases using
the same search strategy. We entered a combination of key
words and filters recommended in the Clinical Study Cate-
gories in the PubMed Clinical Queries feature in database
search engines. We made a manual search of related ref-
erences and key authors. We did not restrict our search by
language.
Search terms used were:

Efficacy in reducing bleeding and/or need for allogenic
transfusion in patients undergoing hip arthroplasty. We also
evaluated the safety outcome, estimating the number of
adverse effects in terms of frequency and severity of pro-
thrombotic events.
--- #1 ‘‘tranexamic acid’’ [MeSH Terms].
--- #2 ‘‘hip’’ [MeSH Terms] AND ‘‘surgery’’ [Subheading].
--- #3 ‘‘surgical procedures, operative’’ [MeSH Terms].
Databases searched (Table 1):

Studies - Cochrane Anesthesiology group.

- Cochrane Controlled Trials Register: Cochrane Library.
Published and in-press randomized controlled studies. - Science Citation Index (1981---2013).

Table 1 Search strategy.

Database Number Search string
of studies
PubMed 58 (‘‘Tranexamic acid’’ [MeSH Terms] OR (‘‘tranexamic’’ [All Fields] AND ‘‘acid’’ [All Fields]) OR
‘‘tranexamic acid’’ [All Fields]) AND ((‘‘hip’’ [MeSH Terms] OR ‘‘hip’’ [All Fields]) AND
(‘‘surgery’’ [Subheading] OR ‘‘surgery’’ [All Fields] OR ‘‘surgical procedures, operative’’
[MeSH Terms] OR (‘‘surgical’’ [All Fields] AND ‘‘procedures’’ [All Fields] AND ‘‘operative’’ [All
Fields]) OR ‘‘operative surgical procedures’’ [All Fields] OR ‘‘surgery’’ [All Fields] OR
‘‘general surgery’’ [MeSH Terms] OR (‘‘general’’ [All Fields] AND ‘‘surgery’’ [All Fields]) OR
‘‘general surgery’’ [All Fields]))
EMBASE 37 (‘‘Tranexamic acid’’ [MeSH Terms] OR (‘‘tranexamic’’ [All Fields] AND ‘‘acid’’ [All Fields]) OR
‘‘tranexamic acid’’ [All Fields]) AND ((‘‘hip [MeSH Terms] OR ‘‘hip’’ [All Fields]) AND
(‘‘surgery [Subheading] OR ‘‘surgery’’ [All Fields] OR ‘‘surgical procedures, operative’’ [MeSH
Terms] OR (‘‘surgical’’ [All Fields] AND ‘‘procedures’’ [All Fields] AND ‘‘operative [All Fields])
OR ‘‘operative surgical procedures’’ [All Fields] OR ‘‘surgery’’ [All Fields] OR ‘‘general
surgery’’ [MeSH Terms] OR (‘‘general’’ [All Fields] AND ‘‘surgery’’ [All Fields]) OR ‘‘general
surgery’’ [All Fields]))
LILACS 48 Tranexamic acid AND hip surgery
Cochrane 25 Tranexamic acid
TRIP Database 60 ‘‘Tranexamic acid AND hip surgery’’, by quality
Google Scholar 1970 Tranexamic acid AND hip surgery
256
- MEDLINE (1966---2013).
- CINAHL (1982---2013).
- SIGLE (1980---2013).
- LILACS (1982---2011).
- SciELO (2005---2013).
- Center of Reviews and Dissemination of United Kingdom.
- Clinical Trials.gov.
- Google Scholar (1990---2013).
Study selection
We revised the abstracts of articles retrieved from the
search to eliminate any that were irrelevant. Relevant arti-
cles were then checked by two independent reviewers to
ensure they met the inclusion criteria. Differences were
settled by consensus of both reviewers.13
Data extraction
Studies meeting our criteria were processed and the data
extracted. Data were extracted independently by two
researchers, and the results were then re-checked for con-
sistency with the data collection form. The following data
were included: number of patients and their characteristics
(age, sex), type of surgical procedure, type of anesthesia
and outcomes (presence or absence of the study event, pres-
ence or absence of adverse events).13
Evaluation of study quality
The methodological quality of the studies was evaluated
by two independent researchers who were blinded to the
author and the journal where the article was published.
Methodological quality was evaluated in terms of random-
ization, allocation concealment, blinding, evaluation and
follow-up. Differences were settled by consensus. Method-
ological quality was evaluated numerically, based on the
Cochrane Collaboration’s risk of bias tool.13
Effect measurements
In studies reporting the proportion of patients with signifi-
cant improvement in blood loss and/or need for transfusion,
the overall relative risk (RR) was calculated using the fixed
effects model.
Reduction in intra- and postoperative blood loss was
determined by measuring the mean change and compar-
ing it between the two groups with standard deviation and
confidence intervals (CI).
The frequency of adverse effects was compared with the
proportional difference in thrombotic events. The associa-
tion was measured using RR as a simple estimator, with its
respective CI.
Incomplete data
The following actions were taken, depending on the type
of missing data: (1) we contacted the author in an attempt
to obtain the missing data; (2) we limited the analysis to
the available data: if no response was obtained from the
C.E. Pinzón-Florez et al.
author, the data was eliminated from our analysis, and (3) if
some studies reported odds ratios and other RR, these were
converted to the estimator used in the studies.14
Heterogeneity
Heterogeneity and inconsistency were evaluated in four
ways: comparison of study methods, participants and inter-
ventions (methodological heterogeneity), comparison of
patient types (clinical heterogeneity), visual evaluation
of the forest plot, statistics and CI. The p value for hetero-
geneity was underestimated to avoid false negatives (type II
error) when a limited number of studies were evaluated,
or in the case of studies with small cohorts. The degree
of inconsistency between studies was evaluated using the
I2 index. Data were analyzed following the fixed effects
model.14,15
Summary measures
According to the type of outcome, we established frequency
measurements, plotted the main outcomes onto a graph,
and compared populations. We calculated estimations and
the 95% CI of risk ratios in terms of weighted RR and weighted
mean difference.14
Unit of analysis
Each randomized patient was taken as the unit of analysis
to evaluate the different studies.
Synthesis of outcomes
Articles were classified according to the type of study con-
ducted and the level of evidence. The results were then
pooled quantitatively according to the measured outcome.
The results of the studies were pooled quantitatively
according to the measured outcome. Prior to this, they were
classified by type of study and clinical evidence.
Results
Our search terms retrieved a total of 2198 studies. Of these,
the titles and/or abstracts of 2152 did not meet our inclu-
sion criteria, and were excluded. We also excluded duplicate
studies. Of the remaining 46 studies, 12 did not include a
control group, three were not randomized, and in 13 stud-
ies the method used to evaluate outcomes was unclear. All
these studies were excluded (Fig. 1). This gave a total of 16
studies included in the review. All the eligible studies were
written in English.
None of the 16 studies were triple blind, 14 were dou-
ble blind,16---27,30,31 one was patient-blind,28 and blinding
was unclear in the last29 (Table 2). Three studies28,29,31
used computer-based randomization techniques, and a
further three used a sealed envelope system.23,28,29 In
one study, patients were randomized by an individ-
ual/pharmacist/resident physician not involved with either
the study or the study patients26 (Table 2). In four studies,
 Efficacy of tranexamic acid in hip arthroplasty
Table 2 Characteristics of studies reviewed.
Reference Number Intervention Fixation DVT prophylaxis DVT screening Type of Blood Patients Route of Bias risk
of subjects anesthesia transfusion lost to adminis-
protocol follow up tration
Claeys et al., 40 TXA 10 mg/kg IV 15 min Hybrid Low-molecular- Medical Regional Hb < 85 g/l 0 IV Low
200725 before surgery weight inspection or Hct < 27%
heparin and
ultrasound
Placebo (saline solution)
Ido et al., 40 TXA 1 g IV immediately Cement None Medical Not reported None IV High
200030 before and 3 h after surgery inspection
Control (none)
Garneti and 50 TXA 10 mg/kg IV on Cement Mechanical Ultrasound or Regional None 0 IV Low
Field et al., anesthesia induction venography
200417
Placebo (NSS)
Niskanen and 36 TXA 10 mg/kg IV Cement Low-molecular- Medical Regional Hct 0.28---0.30 1 IV Low
Korkala immediately before and 8 weight heparin inspection or general
et al., 200524 and 16 h after surgery + mechanical
Placebo (NSS)
Benoni et al., 40 TXA 10 mg/kg IV Cement Low-molecular- Medical Regional None 2 IV Low
200123 immediately before surgery weight inspection or general
heparin
Placebo (NSS)
Benoni et al., 39 TXA 10 mg/kg IV at the end Cement Low-molecular- Medical Regional None 1 IV Low
200018 of surgery and 3 h after weight inspection or general
heparin
Placebo (NSS)
Ekbäck et al., 40 TXA 10 mg/kg IV Cement Low-molecular- Ultrasound Regional Hct < 27% IV Low
200022 immediately before surgery, weight
then 1 mg/kg/h infusion for heparin
10 h + 10 mg/kg bolus 3 h
after surgery
Placebo (NSS)
Husted et al., 40 TXA 10 mg/kg IV Cement or Low-molecular- Medical Regional Hb 0 IV Low
200320 immediately before surgery, hybrid weight inspection decrease > 25%
then 1 mg/kg/h infusion for heparin and clinical
10 h symptoms
Placebo (NSS)
Johansson 100 TXA 15 mg/kg IV Cement Low-molecular- Medical Regional Hb < 90 g/l 19 IV Low
et al., 200521 immediately before surgery weight inspection
heparin
Placebo (NSS)

257
 258
Table 2 (Continued )

Reference Number Intervention Fixation DVT prophylaxis DVT screening Type of Blood Patients Route of Bias risk
of subjects anesthesia transfusion lost to adminis-
protocol follow up tration
Lemay et al., 39 TXA 10 mg/kg IV Cement or Low-molecular- Medical Regional Hb < 70 g/l 1 IV Low
200426 immediately before surgery, no cement weight heparin inspection (adjusted
then 1 mg/kg/h infusion + mechanical and by case).
until wound closure ultrasound Hb < 90 g/l
(older
patients with
comorbidi-
ties)
Placebo (NSS)
Yamasaki 40 TXA 1 mg/kg IV immediately No cement None Medical Regional None 0 IV Low
et al., 200428 before surgery inspection
Placebo (none)
Kazemi et al., 64 15 mg/kg slow 5 min Cement Enoxaparin Medical Regional Using 0 IV Low
201016 pre-surgery or hybrid inspection approved
blood loss
formula
Placebo (NSS)
Malhotra et al., 50 15 mg/kg 15 min before Cement Low-molecular- Ultrasound Regional Hb decrease 0 IV Low
201119 surgery or hybrid weight over
hep- 25% + clinical
arin + mechanical symptoms
Placebo (NSS)
McConnell 44 10 mg/kg bolus at start Cement Mechanical + early Not reported Regional Changes in 0 IV Low
et al., 201129 of surgery or hybrid mobiliza- hematocrit
tion + acetylsalicylic method
acid
Imai et al., 117 (1) 1 g TXA 10 min before Cement Mechanical Full body Regional Changes in --- IV Low
201227 skin closure; (2) 10 before or hybrid + enoxaparin tomography or general hematocrit
skin closure and 6 h later; at 7 days method
(3) 10 min before surgery; post-surgery
(4) 10 min before and 6 h

C.E. Pinzón-Florez et al.

after surgery, respectively
Vijay et al., 90 IV bolus 500 mg TA via 50 ml Cement Low-molecular- General Hb 0 IV Moderate
201332 syringe for 10---15 min or hybrid weight decrease > 25%
before incision, followed heparin and clinical
by infusion of 1 mg/kg/h symptoms
dissolved in 1 L of NSS until
end of surgery
 Efficacy of tranexamic acid in hip arthroplasty
Table 3 Characteristics of patients.
Study, year Number of patients Distribution by sex (% men) Average age, years (SD or range)
TXA Control TXA Control TXA Control
Ekbäck et al., 200022 20 20 45 55 66.4 (9.0) 65.6 (8.8)
Benoni et al., 200018 20 19 30 58 69.5 (10) 68 (10)
Benoni et al., 200123 18 20 50 50 66 (9.5) 68 (9.4)
Husted et al., 200320 20 20 35 30 65 67
Yamasaki et al., 200428 20 20 95 90 55.5 (14.2) 61.2 (6.9)
Garneti and Field, 200417 25 25 N/R 69.6 (11.99) 67.6 (11.4)
Lemay et al., 200426 20 19 60 68 59.7 (10.3) 53.6 (12.8)
Johansson et al., 200521 47 53 53 53 69 (7) 69 (7)
Niskanen and Korkala, 200524 19 20 32 35 66 (9.1) 65 (8.2)
Claeys et al., 200725 20 20 25 35 73 (8) 68 (11)
Kazemi et al., 201016 32 32 72 63 46.6 (16.2) 45.4 (17.2)
Malhotra et al., 201119 25 25 40 48 52.6 (39---72) 54.7 (40---71)
McConnell et al., 201129 22 22 32 41 N/R N/R
Imai et al., 2012,27 T1c 24 22 17 23 64.4 (49---74) 60.2 (54---72)
Imai et al., 2012,27 T2d 20 22 20 23 62.2 (50---85) 60.2 (54---72)
Imai et al., 2012,27 T3e 25 22 16 23 63.3 (47---80) 60.2 (54---72)
Imai et al., 2012,27 T4f 26 22 15 23 63.3 (48---81) 60.2 (54---72)
Vijay et al., 201332 45 45 22 22 48.8 (32.6---65) 49.3 (29.8---68.8)

Study, year Mean weight, kg (SD or range) Total blood loss (ml) p Transfusion required p
TXA Control TXA Control TXA Control
Ekbäck et al., 200022 81.1 (13) 76.8 (13) 1130 (400) 1770 (523) <0.001 1 1 N/R
Benoni et al., 200018 76 (14) 77 (11) 990 (210)a 950 (162.5)a N/R 9 15 0.05
Benoni et al., 200123 79 (16) 78 (17) 759 (275.8b ) 996 (398b ) 0.03 2 10 0.23
Husted et al., 200320 N/R 54.5 (8.7) 814 (1323.8b ) 1231 (1692.7b ) 0.001 2 7 0.04
Yamasaki et al., 200428 52.6 (11.5) 74.7 (14.1) 1350 (477) 1667 (401) <0.05 0 0 N/R
Garneti and Field, 200417 N/R 78 (13) 1443 (809) 1340 (665) 0.822 16 14 N/R
Lemay et al., 200426 80.1 (15.7) 82 (14) 1308 (462) 1469 (405) N/R 0 8 0.0012
Johansson et al., 200521 81 (16) 72 (16) 969 (434) 1324 (577) <0.001 8 23 0.009
Niskanen and Korkala, 200524 80 (19) 69.9 (11.1) 792 (390.1b ) 1102 (485.2b ) 0.03 5 8 0.3
Claeys et al., 200725 76 (15) 81.9 (62---98) 801 (244) 1038 (289) 0.013 1 6 <0.05
Kazemi et al., 201016 72.1 (10.4) N/R 1024 (544) 1399 (587) 0.024 N/R N/R
Malhotra et al., 201119 80.2 (58---100) 56.3 (47---80) 410 (300---510) 615 (515---750) <0.5 6 18 N/R
McConnell et al., 201129 N/R 930 (399b ) 1200 (469b ) 0.02 0 0 N/R
Imai et al., 2012,27 T1c 55.9 (44---86) 56.3 (47---80) 649, 566 1026g N/R 0 0 N/R
Imai et al., 2012,27 T2d 54.1 (40---76) 56.3 (47---80) 388, 418g N/R --- --- N/R
Imai et al., 2012,27 T3e 54.1 (41---70) 56.3 (47---80) N/R --- --- N/R
Imai et al., 2012,27 T4f 53.8 (44---73) 56.3 (47---80) N/R --- --- N/R
Vijay et al., 201332 --- --- <0.001 7 18 0.023
a Sum of median perioperative and drainage measurements.
b CI 95%.
c Average.
d Standard deviation (SD).
e Tranexamic acid (TXA).
f Patients received 1 g TXA 10 min before sx closure, 10 min before sx closure and 6 h later, 10 min before sx, 10 min before sx and 6 h later, respectively.
g Intra sx + post sx.

259
 260
Table 4 Quality evaluation of studies reviewed.
No. Author Year Randomization Allocation Blinding of Incomplete Selective Other Risk of bias Evidence
sequence concealment participants, outcome outcome sources in individual grading
generation researchers and data reporting of bias studies
outcome assessors
blinded
1 Claeys 2007 Low risk Low risk Low risk No (97% Not Unclear Low High quality
et al.25 follow-up) reported
2 Ido et al.30 2000 Low risk Moderate risk Low risk Yes (87% Not Unclear Moderate Moderate
follow-up) reported quality
3 Garneti and 2004 Low risk Low risk Low risk 99% Not Not Low High quality
Field17 follow-up reported identified
4 Niskanen 2005 Low risk Low risk Low risk 81% Not Unclear Low High quality
and follow-up reported
Korkala24
5 Benoni 2001 Low risk Low risk Low risk Not Unclear Low High quality
et al.23 reported
6 Benoni 2000 Low risk Low risk Low risk Not Unclear Low High quality
et al.18 reported
7 Ekbäck 2000 Low risk Low risk Low risk Unclear Low High quality
et al.22
8 Husted 2003 Low risk Moderate risk Low risk Not Not Low High quality
et al.20 reported identified
9 Johansson 2005 Low risk Low risk Low risk Not Not Low High quality
et al.21 reported identified
10 Lemay 2004 Low risk Moderate risk Low risk Not Not Moderate Moderate
et al.26 reported identified quality
11 Yamasaki 2004 Low risk Low risk Low risk Not Not Low High quality
et al.28 reported identified
12 Kazemi 2010 Low risk Low risk Low risk Not Unclear Low High quality
et al.16 reported
13 Malhotra 2011 Low risk Low risk Low risk Not Unclear Low High quality
et al.19 reported

C.E. Pinzón-Florez et al.

14 McConnell 2011 Low risk Low risk Low risk Not Not Low High quality
et al.29 reported identified
15 Imai et al.27 2012 Low risk Low risk Low risk Not Not Low High quality
reported identified
16 Vijay 2013 High risk High risk Low risk Reported Not Moderate Moderate
et al.32 identified quality
Efficacy of tranexamic acid in hip arthroplasty 261

Potentially relevant studies to −0.24) (Fig. 4). This means that in patients undergoing
retrieved from search No.=2198 hip arthroplasty, blood loss was lower in the TXA group than
Medline=58 in controls, and this difference was statistically significant.
EMBASE=37 The studies reviewed had a moderate level of heterogeneity
LILACS =48 (p = 0.006, I2 = 58).
ScIELO=12
Pooled RR for the number of hip arthroplasty patients
Cochrane=25
Google Scholar=1958 receiving allogenic blood transfusion was 0.8 (p < 0.74, CI
95%, 0.58---1.11) (Fig. 5). There were no significant differ-
Studies excluded by title: ences in terns of allogenic blood transfusion requirements in
No.=2133
patients undergoing hip arthroplasty. The studies reviewed
Total abstracts evaluated: were not heterogeneous (p = 0.74, I2 = 0.000).
No.=65 Pooled RR for the number of hip arthroplasty patients
developing DVT was 1.15 (p = 0.73, CI 95%, 0.5---2.6). This
Studies excluded by abstracts: indicates that incidence of TXA-associated DVT was not
No.=19 higher in patients undergoing hip arthroplasty. The studies
reviewed were not heterogeneous (p = 0.72, I2 = 0.0) (Fig. 6).
Full text studies evaluated: As far as complications are concerned, in the TXA
No.=46 groups 12 patients developed DVT, one developed pulmonary
Studies excluded by full text: thromboembolism, there were three cases of wound infec-
No.=30 tion, nine of wound bruising, one of chest infection, and
none of myocardial infarction. In the control group there
Total full text studies included:
were seven cases of DVT, three of wound infection, and four
No.=16 of wound bruising.

Figure 1 Study selection flowchart.

Discussion

the randomization method was unclear.22,23,27,30 The charac-
teristics of the studies and of the patients are summarized in
Tables 2 and 3, respectively. On the quality scale, 13 studies
met Cochrane criteria for ‘‘low risk of bias’’, and three were
assessed as having a moderate risk of bias26,30,31 (Table 4).
The TXA administration method was identical in all the
studies reviewed, i.e., TXA was administered intravenously
in all 16 studies.16---32 In Figs. 2 and 3, the potential publica-
tion bias in studies of patients undergoing hip arthroplasty
has been represented as a funnel plot. Fig. 2 shows only
minimal asymmetry and a few atypical values. This indi-
cates a slight publication bias for the blood loss outcome.
Fig. 3 shows moderate asymmetry, also indicating a slight
publication bias for the allogenic transfusion outcome.
The combined weighted mean difference for total blood
loss in hip arthroplasty was −0.45 (p < 0.001, CI 95%, −0.65
In order to reduce potential bias as far as possible, we
first conducted a primary analysis solely on randomized
controlled trials, even though some of these were not
double-blind, and in three the allocation concealment
method was unclear. Our results show that TXA can signif-
icantly reduce intra- and postoperative blood loss. These
findings are similar to those reported by Sukeik et al.33
However, the benefit of TXA in the ‘‘need for allogenic trans-
fusion’’ outcome was unclear. Comparing the total blood loss
and ‘‘occult’’ blood loss data, it can be seen that occult
blood loss is an important issue in hip arthroplasty.23 Never-
theless, only two of the studies reviewed reported outcomes
and methods used for measuring occult blood loss.
TXA did not increase incidence of DVT, pulmonary
embolism, or hospital stay. However, as no systematic, val-
idated protocol was established to measure this outcome,

0.0

0.1
Standard error

0.2

0.3

0.4

–2.0 –1.5 –1.0 –0.5 –0.0 –0.5 –1.0 1.5 2.0

Std diff in means

Figure 2 Funnel plot of standard errors by weighted mean average.

262 C.E. Pinzón-Florez et al.

0.0

0.5

Standard error 1.0

1.5

2.0

–4 –3 –2 –1 0 1 2 3 4
Log odds ratio

Figure 3 Funnel plot of standard errors by log odds ratio.

Model Author/Year Statistics for each study Standardized mean difference and CI 95% Weights (fixed effects model)
Std diff in Standard Lower Upper Relative Relative Relative Relative
means error Variance limit limit Z value p value weights weights weights weights
Ekcback 2000 –0.086 0.316 0.100 –0.706 0.534 –0.272 0.786 7.27 7.75
Benomi 2000 0.212 0.321 0.103 –0.417 0.842 0.661 0.509 7.05 7.58
Benomi 2001 –0.686 0.334 0.112 –1.341 –0.030 –2.051 0.040 6.51 7.16
Hustetd 2003 –0.275 0.318 0.101 –0.897 0.348 –0.864 0.388 7.21 7.70
Yamasaki 2004 –0.719 0.326 0.106 –1.359 –0.080 –2.205 0.027 6.83 7.41
Garneti 2004 –0.139 0.283 0.080 –0.416 0.694 0.491 0.623 9.07 9.02
Lemay 2004 –0.370 0.323 0.104 –1.003 0.263 –1.145 0.252 6.97 7.52
Johannson 2005 –0.690 0.206 0.043 –1.094 –0.285 –3.344 0.001 17.11 13.09
Niskanen 2005 –0.702 0.330 0.109 –1.349 –0.055 –2.127 0.033 6.68 7.29
Claeys 2007 –0.886 0.331 0.110 –1.536 –0.237 –2.674 0.007 6.63 7.25
Kazemi 2010 –0.664 0.257 0.066 –1.168 –0.161 –2.587 0.010 11.03 10.22
McConnell 2011 –0.620 0.309 0.095 –1.225 –0.015 –2.009 0.045 7.64 8.02
Fixed –0.462 0.085 0.007 –0.629 –0.295 –5.420 0.000
Random –0.453 0.104 0.011 –0.657 –0.248 –4.332 0.000

–2.00 –1.00 –0.00 –1.00 –2.00

In favor TXA In favor placebo

Figure 4 Meta-analysis of the ‘‘total blood loss in surgery’’ outcome.

Relative
Study
RR (CI 95%) weight (%)

Ekba?ck 1.00 (0.15, 6.48) 2.92

Benoni 1.49 (0.58, 3.85) 7.98

Benoni 1.39 (0.35, 5.53) 4.28

Husted 0.54 (0.19, 1.53) 11.88

Yamasaki 1.00 (0.07, 14.95) 1.46

Garneti 0.78 (0.19, 3.16) 5.73

Lemay 0.14 (0.02, 1.00) 12.24

Johansson 1.25 (0.48, 3.21) 9.76

Niskanen 0.63 (0.25, 1.58) 12.96

Claeys 0.32 (0.07, 1.35) 10.26

Malhotra 1.00 (0.36, 2.76) 8.75

McConnell 1.00 (0.07, 15.04) 1.46

Imai 0.92 (0.06, 13.87) 1.52

Bhavani S 1.14 (0.41, 3.17) 8.83

Overall (I-squared=0.0%, P=.749) 0.80 (0.58, 1.11) 100.00

.1 1.1

Figure 5 Meta-analysis of the ‘‘need for allogenic transfusion’’ outcome.

Efficacy of tranexamic acid in hip arthroplasty 263

Relative weight
Study RR ( 95%) (%)

Ekbäck G 2000 1.00 (0.07, 14.95) 10.84

Benoni G, 2000 1.00 (0.27, 3.69) 32.52

Claeys MA, 2007 6.33 (0.35, 114.81) 5.69

Kazemi SM, 2010 0.34 (0.01, 8.13) 16.02

Imai N, 2012 0.86 (0.19, 3.90) 34.93

Overall (I-squared=0.0%, P=.720) 1.15 (0.51, 2.60) 100.00

.1 1 10

Figure 6 Meta-analysis of the ‘‘safety, deep vein thrombosis’’ outcome.

and patients with adverse events were not followed-up, the
statistical power of the reported rate of adverse events is
low and the data unreliable.
A randomized controlled study with enough statistical
power to detect a difference in effect size for total blood
loss of 0.5 standard deviations (assuming a power of 80%
and statistical significance of 0.05%) would need 68 patients
in each arm. In all the individual studies included in our
meta-analysis, however, the cohorts were small. This could,
therefore, lead to a type II error when establishing a causal
relationship. This shows the importance of meta-analysis
techniques, since in our study we reviewed 16 separate stud-
ies with a total of 718 patients. This increased the power of
the meta-analysis to assess different outcomes, and there-
fore, to reach a conclusion.
Limitations
The methods used to quantify total blood loss differed
greatly between studies; this could have given rise to bias
when compiling the meta-analysis. However, blood loss in
the TXA group follows a downward trend in relation to other
therapeutic strategies. Estimates of RR and CI 95% for the
‘‘need for transfusion’’ outcome was not statistically sig-
nificant, probably due to the number of event in the TXA
group. We have presented the I2 value for each outcome
analyzed. It should be noted that the data are to a certain
extent heterogeneous, and should therefore be interpreted
with caution. TXA does not significantly increase the appar-
ent risk of thromboembolic complications; we were unable
to perform a meta-analysis of this outcome because infor-
mation on this topic was incomplete in some of the studies
reviewed.
Conclusion
In conclusion, TXA reduces total blood loss in patients under-
going hip arthroplasty. This can have a significant effect on
clinical practice and on the cost of managing these patients.
Further studies related to the outcomes evaluated here,
with larger sample sizes, are needed to confirm the safety
of TXA in hip arthroplasty.
Conflict of interest
The authors declare they have no conflict of interest.
Acknowledgments
We would like to thank the University of Rosario for allowing
us to undertake this study.
Appendix A.
The systematic review protocol is held in the digital
library of the University of Rosario, available from: http://
repository.urosario.edu.co
References
1. Zuckerman JD. Hip fracture. N Engl J Med. 1996;334:1519---25.
2. Parker J, Pryor G. Hip fracture management. Oxford: Blackwell
Scientific Publication; 2000.
3. Lane A, Crosby ET. Blood management for hip reconstruction
surgery. Orthop Clin North Am. 2009;40:417---25.
4. Wong CJ, Vandervoort MK, Vandervoort SL, Donner A, Zou G,
MacDonald JK, et al. A cluster-randomized controlled trial of a
blood conservation algorithm in patients undergoing total hip
joint arthroplasty. Transfusion. 2007;47:832---41.
5. Pola E, Papaleo P, Santoliguido A, Gasparini G, Aulisa L, de Santis
E. Clinical factors associated with an increased risk of perioper-
ative blood transfusion in nonanemic patients undergoing total
hip arthroplasty. J Bone Joint Surg Am. 2004;86:57---61.
6. Serious Hazards of Transfusion (SHOT). The serious hazards of
transfusion steering group. Annual report 2006; 2014. Available
at: http://www.shotuk.org/SHOT report 2006.pdf [accessed
16.02.14].
264
7. Klein HG, Spahn DR, Carson JL. Red blood cell transfusion in
clinical practice. Lancet. 2007;370:415---26.
8. Carson JL, Altman DG, Duff A, Noveck H, Weinstein MP, Son-
nenberg FA, et al. Risk of bacterial infection associated with
allogeneic blood transfusion among patients undergoing hip
fracture repair. Transfusion (Paris). 1999;39:694---700.
9. Bierbaum BE, Callaghan JJ, Galante JO, Rubash HE, Tooms RE,
Welch RB. An analysis of blood management in patients hav-
ing a total hip or knee arthroplasty. J Bone Joint Surg Am.
1999;81:2---10.
10. Rosencher N, Kerkkamp HE, Macheras G, Munuera LM,
Menichella G, Barton DM, et al., OSTHEO Investigation. Ortho-
pedic Surgery Transfusion Hemoglobin European Overview
(OSTHEO) study: blood management in elective knee and hip
arthroplasty in Europe. Transfusion (Paris). 2003;43:459---69.
11. Jones CA, Voaklander DC, Johnston DWC, Suarez-Almazor ME.
The effect of age on pain, function, and quality of life after total
hip and knee arthroplasty. Arch Intern Med. 2001;161:454---60.
12. Jadad AR, Moore RA, Carroll D, Jenkinson C, Reynolds JM,
Gavaghan DJ, et al. Assessing the quality of reports of ran-
domized clinical trials: is blinding necessary? Control Clin Trials.
1996;17:1---2.
13. Berger VW, Alperson SY. A general framework for the evaluation
of clinical trial quality. Rev Recent Clin Trials. 2009;4:79---88.
14. Higgins JPT, Green S, editors. Cochrane handbook for systematic
reviews of interventions. Version 5.1.0. The Cochrane Collab-
oration; 2011 [cited 24.11.13], Available at: www.cochrane-
handbook.org [accessed March 2011].
15. Hedges LV, Olkin I. Statistical methods for meta-analysis.
Orlando: Academic Press; 1985. p. 108---38.
16. Kazemi SM, Mosaffa F, Eajazi A, Kaffashi M, Daftari Besheli L,
Zanganeh RF. The effect of tranexamic acid on reducing blood
loss in cementless total hip arthroplasty under epidural anes-
thesia. Orthopedics. 2010;33:17.
17. Garneti N, Field J. Bone bleeding during total hip arthro-
plasty after administration of tranexamic acid. J Arthroplast.
2004;19:488---92.
18. Benoni G, Lethagen S, Nilsson P, Fredin H. Tranexamic acid,
given at the end of the operation, does not reduce post-
operative blood loss in hip arthroplasty. Acta Orthop Scand.
2000;71:250---4.
19. Malhotra R, Kumar V, Garg B. The use of tranexamic acid to
reduce blood loss in primary cementless total hip arthroplasty.
Eur J Orthop Surg. 2011;21:101---4.
20. Husted H, Blønd L, Sonne-Holme S, Holm G, Jacobsen TW,
Gebuhr P. Tranexamic acid reduces blood loss and blood
C.E. Pinzón-Florez et al.
transfusions in primary total hip arthroplasty: a prospective ran-
domized double-blind study in 40 patients. Acta Orthop Scand.
2003;74:665---9.
21. Johansson T, Pettersson L, Lisander B. Tranexamic acid in total
hip arthroplasty saves blood and money: a randomized, double-
blind study in 100 patients. Acta Orthop. 2005;76:314---9.
22. Ekbäck G, Axelsson K, Ryttberg L, Edlund B, Kjellberg J, Weck-
ström J, et al. Tranexamic acid reduces blood loss in total hip
replacement surgery. Anesth Analg. 2000;91:1124---30.
23. Benoni G, Fredin H, Knebel R, Nilsson P. Blood conservation
with tranexamic acid in total hip arthroplasty. A randomized,
double-blind study in 40 primary operations. Acta Orthop Scand.
2001;72:442---8.
24. Niskanen RO, Korkala OL. Tranexamic acid reduces blood loss in
cemented hip arthroplasty: a randomized, double-blind study of
39 patients with osteoarthritis. Acta Orthop. 2005;76:829---32.
25. Claeys MA, Vermeersch N, Haentjens P. Reduction of blood loss
with tranexamic acid in primary total hip replacement surgery.
Acta Chir Belg. 2007;107:397---401.
26. Lemay E, Guay J, Côté C, Roy A. Tranexamic acid reduces the
need for allogenic red blood cell transfusions in patients under-
going total hip replacement. Can J Anesth. 2004;51:31---7.
27. Imai N, Dohmae Y, Suda K, Miyasaka D, Ito T, Endo N. Tranexamic
acid for reduction of blood loss during total hip arthroplasty. J
Arthroplast. 2012;27:1838---43.
28. Yamasaki S, Masuhara K, Fuji T. Tranexamic acid reduces blood
loss after cementless total hip arthroplasty --- prospective ran-
domized study in 40 cases. Int Orthop. 2004;28:69---73.
29. McConnell JS, Shewale S, Munro NA, Shah K, Deakin AH,
Kinninmonth AW. Reduction of blood loss in primary hip arthro-
plasty with tranexamic acid or fibrin spray. Acta Orthop.
2011;82:660---3.
30. Ido K, Neo M, Asada Y, Kondo K, Morita T, Sakamoto T, et al.
Reduction of blood loss using tranexamic acid in total knee and
hip arthroplasties. Arch Orthop Trauma Surg. 2000;120:518---20.
31. Zufferey P, Merguiol F, Laporte S, Decousus H, Mismetti
P, Auboyer C, et al. Do antifibrinolytics reduce allogeneic
blood transfusion in orthropedic surgery? Anesthesiology.
2006;105:1034---46.
32. Vijay BS, Bedi V, Mitra S, Das B. Role of tranexamic acid in reduc-
ing postoperative blood loss and transfusion requirement in
patients undergoing hip and femoral surgeries. Saudi J Anaesth.
2013;7:29---32.
33. Sukeik M, Alshryda S, Haddad FS, Mason JM. Systematic review
and meta-analysis of the use of tranexamic acid in total hip
replacement. J Bone Joint Surg Br. 2011;93:39---46.