Combined versus single application of tranexamic acid in total knee and hip arthroplasty: A meta-analysis of randomized controlled trials

Combined versus single application of tranexamic acid in total knee and hip arthroplasty: A meta-analysis of randomized controlled trials

International Journal of Surgery 43 (2017) 171e180 Contents lists available at ScienceDirect International Journal of Surgery journal homepage: www...

936KB Sizes 0 Downloads 45 Views

International Journal of Surgery 43 (2017) 171e180

Contents lists available at ScienceDirect

International Journal of Surgery journal homepage: www.journal-surgery.net

Review

Combined versus single application of tranexamic acid in total knee and hip arthroplasty: A meta-analysis of randomized controlled trials M.M. Peng Zhang 1, M.M. Jifeng Li 1, M.M. Xiao Wang* Department of Orthopedics, Huaihe Hospital, Henan University, Henan, China

h i g h l i g h t s  To compare the efficacy of the combined application of TXA for patients with THA and TKA.  Only high quality studies were selected.  Combined administration of TXA in total knee and hip arthroplasty was associated with significantly reduced total blood loss.

a r t i c l e i n f o

a b s t r a c t

Article history: Received 17 February 2017 Received in revised form 7 May 2017 Accepted 29 May 2017

Objective: To compare the efficacy and safety of the combined application of both intravenous and topical tranexamic acid versus the single use of either application in patients with total knee and hip arthroplasty. Methods: Potentially relevant studies were identified from electronic databases including Medline, PubMed, Embase, ScienceDirect and the Cochrane Library. Patients undergoing primary total knee and hip arthroplasty were included in our studies, with an experimental group that received combined intravenous and topical application of tranexamic acid and a control group that received a single application of tranexamic acid or normal saline. The primary outcomes were total blood loss, hemoglobin decline and transfusion requirements. The secondary outcomes were length of stay, operation time and tranexamic acid-related adverse effects, such as superficial infection, deep vein thrombosis or pulmonary embolism. Modified Jadad scores were used to assess the quality of the included randomized controlled trials (RCTs). The data was pooled using RevMan 5.3. After testing for heterogeneity across studies, the data were aggregated using random-effects modeling when appropriate. We have registered the trial at http://www.researchregistry.com. Results: Six RCTs that included 704 patients met the inclusion criteria. The present meta-analysis indicated significant differences existed in the total blood loss (MD ¼ 134.65, 95% CI: 191.66 to 77.64, P < 0.0001), postoperative hemoglobin level (MD ¼ 0.74, 95% CI: 0.39 to 1.10, P < 0.0001), drainage volume (MD ¼ 40.19, 95% CI: 55.95 to 24.43, P < 0.00001) and transfusion rate (RD ¼ 0.07, 95% CI: 0.11 to 0.03, P ¼ 0.0004) between groups. Conclusion: Combined administration of tranexamic acid in total knee and hip arthroplasty was associated with significantly reduced total blood loss, postoperative hemoglobin decline, drainage volume, and transfusion requirements. Based on the limitations of current meta-analysis, well-designed, highquality RCTs with long-term follow-up are still required. © 2017 IJS Publishing Group Ltd. Published by Elsevier Ltd. All rights reserved.

Keywords: Tranexamic acid Total knee and hip arthroplasty Blood loss Meta-analysis

1. Introduction Total knee arthroplasty and hip arthroplasty (TKA and THA) are

* Corresponding author. E-mail address: [email protected] (M.M. Xiao Wang). 1 Peng Zhang and Jifeng Li equally contributed to this study.

major orthopedic surgeries used to treat degenerative arthritis and traumatic conditions such as displaced femoral neck fractures. More than 700 thousand TKAs are estimated to be performed annually in the United States. However, substantial perioperative blood loss has been associated with patient dissatisfaction and systemic diseases, especially in elderly individuals [1,2]. Many methods have been used to manage blood loss including tourniquet

http://dx.doi.org/10.1016/j.ijsu.2017.05.065 1743-9191/© 2017 IJS Publishing Group Ltd. Published by Elsevier Ltd. All rights reserved.

172

M.M. Peng Zhang et al. / International Journal of Surgery 43 (2017) 171e180

application, blood transfusion, administration of hemostatic agents and autologous donation [3]. Allogenic blood transfusion may increase the risk of adverse events, such as virus infections, immunologically mediated diseases and cardiovascular dysfunction, resulting in a financial burden and potentially life-threatening effects on patients [4]. Recently, the use of tranexamic acid in total knee or hip arthroplasty has become popularized in orthopedics. Tranexamic acid is a synthetic analog of an amino acid whose biological activity inhibitis plasminogen from dissolving clots [5]. In previous studies, intravenous administration and topical application of tranexamic acid were reported to be associated with reduced perioperative blood loss and transfusion units. Furthermore, meta-analyses of high-quality randomized control trials (RCTs) indicated that tranexamic acid was effective and safe for the management of blood loss in patients with total knee or hip arthroplasty [6,7]. Despite this previous research, whether the combined application of tranexamic acid is superior to a single use remains unclear due to a lack of published studies and the inclusion of small sample sizes. Therefore, we performed the present systematic review and meta-analysis to evaluate the efficiency and safety of the combined application of intravenous and topical tranexamic acid compared to the single use of either application in patient with total knee and hip arthroplasty. We included only high quality RCTs that compared the efficacy and safety of combined application of intravenous and topical tranexamic acid with the single use of either application in patients with total knee and hip arthroplasty, in which the experimental group received combined intravenous and topical

application of tranexamic acid and the control group received a single application of tranexamic acid or normal saline. 2. Methods 2.1. Search strategy Potentially relevant studies were identified from electronic databases including Medline (1966e2017.5), PubMed (1966e2017.5), Embase (1980e2017.5), ScienceDirect (1985e2017.5) and the Cochrane Library. The following key words were used in combination with the Boolean operators AND or OR: “total knee replacement OR arthroplasty”, “total hip replacement OR arthroplasty” and ‘‘tranexamic acid’’, ‘‘blood loss’’ or ‘‘blood transfusion’’. The bibliographies of the retrieved trials and other relevant publications were cross-referenced to identify additional articles. We placed no restrictions on the publication language. The search process was performed as presented in Fig. 1. We have registered the trial at http://www.researchregistry.com. 2.2. Inclusion criteria and study selection Tranexamic acid is an antifibrinolytic agent that inhibits fibrinolysis by reversibly blocking the lysine-binding sites of plasminogen; this agent is commonly used in orthopedic surgery to decrease intraoperative blood loss and transfusion amounts. The bioavailability of tranexamic acid is 34% and its half-life is 3.1 h. Studies were considered for inclusion if they met the following

Fig. 1. Search results and the selection procedure.

M.M. Peng Zhang et al. / International Journal of Surgery 43 (2017) 171e180

criteria: 1) published RCTs; 2) included a patient population that underwent total knee or hip arthroplasty, with an experimental group that received combined intravenous and topical application of tranexamic acid and a control group that received a single application of tranexamic acid or normal saline; and 3) reported surgical outcomes including hemoglobin decline or postoperative hemoglobin level, blood loss, drainage volume, transfusion requirements, length of stay, and operation time as well as surgeryrelated adverse effects, such as wound infection, deep vein thrombosis and pulmonary embolism. Studies were excluded from the present meta-analysis if they included incomplete data. All analyses were based on previous published studies, thus no ethical approval nor patient consent were required. 2.3. Date extraction Two reviewers independently scanned the abstracts of the potentially included studies. Subsequently, the full text of the studies that met the inclusion criteria was screened, and a final decision was made. Disagreement was resolved by consulting a senior reviewer. Two of the authors independently extracted data from the included studies. Corresponding authors were consulted for details if the data were incomplete. The following data were extracted and recorded in a spreadsheet: first author name, publication year, baseline characteristics, intervention procedures, samples size, transfusion trigger, and outcome parameters. Other relevant data were also extracted from individual studies. The primary outcomes were total blood loss, hemoglobin decline and transfusion requirements. The secondary outcomes were the length of stay, operation time and tranexamic acid-related adverse effects, such as superficial infection, deep vein thrombosis or pulmonary embolism. 2.4. Assessment of methodological quality A quality assessment of each randomized trial was performed by two reviewers based on the Cochrane Handbook for Systematic Reviews of Interventions. Disagreement was resolved by consulting a senior reviewer. We created a “risk of bias’’ table that included the following elements: random sequence generation, allocation concealment, blinding, incomplete outcome data, free of selective reporting and other bias [6,7]. Publication bias is a tendency on average to produce results that appear significant, because negative or near neutral results are almost never published. Publication bias may exist in all meta-analyses. Selective reporting is a strong bias that prevents correct conclusions arising from hypothesis tests, this bias is a specific form of selection bias whereby only interesting or relevant examples are cited. Therefore, the meta-analysis results should be considered appropriate. 2.5. Data analysis and statistical methods The data were pooled using RevMan 5.3 (The Cochrane Collaboration, Oxford, UK). After extracting the data from the included studies, we exported the means, SDs and sample sizes of groups into RevMan 5.3 to determine the heterogeneity. Statistical heterogeneity was assessed based on the P and I2 values using the standard Chi-square test. When I2  50% or P < 0.1, significant heterogeneity was indicated and a random-effects model was applied for the meta-analysis. Otherwise, a fixed-effects model was used. Dichotomous outcomes (i.e., transfusion requirements) were expressed as risk differences (RDs) with 95% confidence intervals (CIs). For continuous outcomes (i.e., transfusion requirements), mean differences (MDs) and 95% confidence intervals (CIs) were calculated. A subgroup analysis was conducted when significant

173

heterogeneity was detected to find the source if possible. Various surgical procedure may causes significant heterogeneity. Therefore, we only included studies with the same surgical procedure to find the source of heterogeneity. 3. Results 3.1. Search result A total of 468 studies were identified through the initial search. By scanning the abstracts, 462 reports that did not meet inclusion criteria were excluded from the current meta-analysis. No gray literature was included. Finally, six RCTs [8e13] published between 2014 and 2016 were included in the present meta-analysis; these studies included 352 patients in the experimental groups and 352 patients in the control groups. All included studies were indexed in PubMed and published in English. 3.2. Study characteristics The sample sizes ranged from 80 to 184 patients. Only studies that included patients with end-stage knee arthritis or osteonecrosis of the femoral head were included in the present metaanalysis. In these studies, the experimental groups received combined intravenous and topical tranexamic acid and the control groups received a single application of tranexamic acid or normal saline. The characteristics of the included studies are reported in Table 1. Statistically similar baseline characteristics were observed between groups. 3.3. Risk of bias within studies The Cochrane Handbook for Systematic Review of Interventions was consulted to assess risk of bias of the RCTs. All RCTs provided clear inclusion and exclusion criteria and described their randomization methodology, and four studies [9,11e13] described the use of computer-generated randomization. Three studies [8,10,13] reported allocation concealment by closed envelope or other techniques. Double blinding was reported in four RCTs [8,11e13]. An intentioneto-treat analysis was not performed in any of the RCTs; therefore a potential risk of type II statistical error existed. No unclear bias due to incomplete outcome data or selective outcome reporting was identified in the RCTs. The methodological quality assessment is summarized in Table 2. Each risk of bias item is presented as the percentage across all included studies, which indicates the proportion of different levels of risk of bias for each item (Table 3). 3.4. Outcomes for meta-analysis Total blood loss. Four articles [8e10,12] reported the outcomes of total blood loss following the operation. A random-effects model was used because significant heterogeneity was found among the studies (c2 ¼ 6.02, df ¼ 3, I2 ¼ 50%, P ¼ 0.11). The pooled results demonstrated that total blood loss was significantly higher in the control groups than in the experimental groups (MD ¼ 134.65, 95% CI: 191.66 to 77.64, P < 0.0001; Fig. 2). Postoperative hemoglobin level. Two studies [9,11] reported the outcomes of postoperative hemoglobin level. A fixed-effects model was used because no significant heterogeneity was found among the studies (c2 ¼ 0.9, df ¼ 1, I2 ¼ 0%, P ¼ 0.34). The pooled results demonstrated that the postoperative hemoglobin level was significantly higher in the experimental groups than in the control groups (MD ¼ 0.74, 95% CI: 0.39 to 1.10, P < 0.0001; Fig. 3). Hemoglobin decline. Three studies [8,9,12] reported the

174

M.M. Peng Zhang et al. / International Journal of Surgery 43 (2017) 171e180

Table 1 Cohort characteristics. Studies

Study design

Patient

Intervention

Cases Mean Female Surgical TXA intervention age patient (E/C)

methods

Huang 2014 [8]

RCT (method unreported) and controlled 92/ blind parallel study with a follow up of 1e3 92 months

(E/C) (E/C) 65.4/ 64.7

55/62

TKA

Lin 2015 [9]

40/ RCT (computer generated) and controlled nonblind parallel study with a follow up of 3 40 months

70.7/ 71.0

30/33

59/ Jain 2015 RCT (computer generated) and controlled [12] blind parallel study with a unreported follow 60 up

68.3/ 70.0

39/36

Xie 2015 [10]

70/ RCT (method unreported) and controlled nonblind parallel study with a follow up of 3 70 months

60.5/ 59.5

48/50

Karaaslan 2015 [11] Zeng 2016 [12]

RCT (computer generated) and controlled blind parallel study with a follow up of 2 years RCT (computer generated) and controlled blind parallel study with a follow up of 6 months

41/ 40

65.9/ 65.6

32/35

50/ 50

53.6/ 54

21/26

Comparison Prophylactic antithrombotic

E:1.5 g topical injection þ1.5 g LMWH, 6000IU i.v. C:1.5 g i.v.

1.total blood loss 2 hemoglobin decline 3.transfusion rate 4.dragine volume TKA E: 1 g topical injectionþ1 g i.v. Rivaroxaban,10 mg 1.total blood C:1 g topical injection loss 2 hemoglobin decline 3.transfusion rate 4.dragine volume 1.total blood TKA E:pre-op.15 mg/kg i.v.þpost- Aspirin,75 mg loss op.10 mg/kg i.v.þ2 g topical 2.hemoglobin injection decline C:pre-op.15 mg/kg i.v.þpost3.transfusion op.10 mg/kg i.v rate THA E: 1 g topical injectionþ2 g i.v. Enoxaparin,6000IU 1.total blood C:1.5 g i.v. loss 2.transfusion rate Bilateral E:15 mg/kg topical LMWH,4000IU 1.transfusion TKA injection þ10 mg/kg i.v. rate THA

outcomes of hemoglobin decline following the operation. A random-effects model was used because significant heterogeneity existed among these studies (c2 ¼ 16.16, df ¼ 2, I2 ¼ 88%, P ¼ 0.0003). The pooled results demonstrated that the hemoglobin decline was significantly higher in control groups than in the experimental groups (MD ¼ 0.44, 95% CI: 0.79 to 0.09, P ¼ 0.01; Fig. 4). Transfusion rate. The transfusion rates were reported in six studies [8e12]. A fixed-effects model was applied because no significant heterogeneity was found among these studies (c2 ¼ 5.66, df ¼ 5, I2 ¼ 12%, P ¼ 0.34). A significant difference was detected in the transfusion rate between the two groups (RD ¼ 0.07, 95% CI: 0.11 to 0.03, P ¼ 0.0004; Fig. 5). Drainage volume. The drainage volume was provided in three studies [8,9]. A fixed-effects model was used because no significant heterogeneity was found among these studies (c2 ¼ 5.63, df ¼ 2, I2 ¼ 64%, P ¼ 0.06). The drainage volume was significantly higher in control groups than in the experimental groups (MD ¼ 40.19, 95% CI: 55.95 to 24.43, P < 0.00001; Fig. 6). Operation time. The operation time was reported in four studies [8e11]. A random-effects model was used because significant heterogeneity was found among the pooled data (c2 ¼ 12.79, df ¼ 3, I2 ¼ 77%, P ¼ 0.006). No significance difference in the operation time was observed between the two groups. (MD ¼ 0.22, 95% CI: 4.65 to 5.08, P ¼ 0.09; Fig. 7). Length of hospital stay (LOS). Three studies reported the lengths of the hospital stays for the groups [8,10,11]. A fixed-effects model was used because no significant heterogeneity was

E: 1 g topical injectionþ 15 mg/ LMWH,2000IU kg i.v. C: 15 mg/kg i.v.

1.transfusion rate 2.dragine volume

Main outcome measures

1.P 2.P 3.P 4.P

¼ ¼ ¼ ¼

0.067 0.031 0.028 0.011

1.P 2.P 3.P 4.P

< < < <

0.001 0.001 0.001 0.001

1.P ¼ 0.001 2.P ¼ 0.001 3.P ¼ 0.036

1.P ¼ 0.001 2.P ¼ 0.016

1.P < 0.05

1.P < 0.05 2.P < 0.05

identified in the pooled results (c2 ¼ 3.87, df ¼ 2, I2 ¼ 48%, P ¼ 0.14). No significant difference in the LOS was observed between the two groups. (MD ¼ 0.09, 95% CI: 0.23 to 0.04, P ¼ 0.17; Fig. 8). Superficial infection. The superficial infection incidence was reported in three studies [8e10]. A fixed-effects model was used because no significant heterogeneity was found among these studies (c2 ¼ 0.29, df ¼ 2, I2 ¼ 0%, P ¼ 0.86). No significant difference in the incidence of superficial infection was found between the two groups (RD ¼ 0.00, 95% CI: 0.03 to 0.02, P ¼ 0.68; Fig. 9). Deep vein thrombosis. Five articles [8e12] reported the incidence of deep vein thrombosis following joint replacement. A fixed-effects model was used due to the low significant heterogeneity among these studies (c2 ¼ 1.92, df ¼ 4, I2 ¼ 0%, P ¼ 0.75). No significant difference was found between the groups (RD ¼ 0.00, 95% CI: 0.02 to 0.02, P ¼ 1.00; Fig. 10). Pulmonary embolism. Pulmonary embolism was reported in five studies [8e12]. A fixed-effects model was used because no significant heterogeneity was found among the studies (c2 ¼ 0.00, df ¼ 4, I2 ¼ 0%, P ¼ 1.00). No significant difference was found in the pulmonary embolism incidence between the two groups (RD ¼ 0.00, 95% CI: 0.01 to 0.01, P ¼ 1.00). Subgroup analysis for total blood loss. A subgroup analysis was performed to assess total blood loss. Only studies in which unilateral TKA was performed were included. A fixed-effects model was used because no significant heterogeneity was found among these groups (c2 ¼ 4.18, df ¼ 2, I2 ¼ 52%, P ¼ 0.12). A significant difference was observed between the two groups (MD ¼ 157.96, 95% CI: 206.31 to 109.61, P < 0.00001; Fig. 11).

M.M. Peng Zhang et al. / International Journal of Surgery 43 (2017) 171e180

175

Table 2 Methodological quality of the randomized controlled trials.

4. Discussion To the best of our knowledge, this study is the first metaanalysis to evaluate the efficiency and safety of a combined application of tranexamic acid versus a single application in patients with total knee and hip arthroplasty in prospective randomized control trials. The most important finding of the meta-analysis was that the combined application of intravenous and intraarticular tranexamic acid in patients with TKA and THA was associated with a significantly reduced postoperative hemoglobin decline, transfusion requirements and drainage volume compared to the single application. Moreover, no increased risk of the incidence of infection, deep vein thrombosis and or pulmonary embolism was

identified. The combined application of tranexamic acid was not associated with a prolonged operation time or length of stay. Several high-quality RCTs have confirmed improved outcomes in patients with TKA and THA who have been administered this agent [14e16]. However, blood loss is often more critical among elderly individuals. Tranexamic acid can be applied by various routes including intravenous, intraarticular, oral and intramuscular. To achieve the maximum plasma concentration, tranexamic acid is administered for approximately 2 h via the oral, 30 min via the intramuscular, and 5e15 min via the intravenous routes [17]. The present meta-analysis indicated that combined application of tranexamic acid was associated with significantly reduced total blood loss compared with a single application of tranexamic acid.

176

M.M. Peng Zhang et al. / International Journal of Surgery 43 (2017) 171e180

Table 3 Risk of bias.

Experimental Study or Subgroup

Mean 867

374

Jain 2015

385.68

Lin 2015

578.7

Huang 2014

Xie 2015

Control

SD Total 92

Mean

Mean Difference

IV, Fixed, 95% CI

285

92

16.5%

182.5

59 590.69 191.1

60

33.8% -205.01 [-272.14, -137.88]

246.9

40

40

14.9%

-126.40 [-227.63, -25.17]

70

34.8%

-101.28 [-167.46, -35.10]

776.75 188.95

Total (95% CI)

957

Mean Difference

SD Total Weight

705.1 213.9

70 878.03

210

261

IV, Fixed, 95% CI

-90.00 [-186.08, 6.08]

262 100.0% -138.23 [-177.27, -99.19]

Heterogeneity: Chi² = 6.02, df = 3 (P = 0.11); I² = 50%

-200

Test for overall effect: Z = 6.94 (P < 0.00001)

-100

0

Favours [experimental]

100

200

Favours [control]

Fig. 2. Forest plot diagram showing effect of combination TXA on total blood loss.

Experimental Study or Subgroup

Mean

Karaaslan 2015

11.36 1.52

Lin 2015 Total (95% CI)

12.2

Control

SD Total Mean

1.2

Mean Difference

SD Total Weight

IV, Fixed, 95% CI

41 10.41 0.97

40

41.3%

0.95 [0.40, 1.50]

40

40

58.7%

0.60 [0.14, 1.06]

80 100.0%

0.74 [0.39, 1.10]

11.6

81

0.9

Heterogeneity: Chi² = 0.90, df = 1 (P = 0.34); I² = 0% Test for overall effect: Z = 4.10 (P < 0.0001)

Mean Difference IV, Fixed, 95% CI

-4

-2

Favours [experimental]

0

2

4

Favours [control]

Fig. 3. Forest plot diagram showing effect of combination TXA on postoperative hemoglobin level.

Furthermore, the postoperative hemoglobin levels were significantly higher in the experimental group than in the control groups. A subgroup analysis performed for total blood loss; after the exclusion of THA and bilateral TKA from the analysis showed similar results. TKA without antifibrinolytics was associated with blood loss ranging from 761 to 1784 [18e21] ml and 7.7%e18.93% [22e25] of these patients required a transfusion to relieve anemia. However, transfusion was considered undesirable due to its associated risks of various adverse reactions [4]. Transfusion is required less frequently following either systemic or topical tranexamic acid administration [26,27]. The current meta-analysis showed that the

combined application of tranexamic acid was associated with a further significant reduction in the transfusion requirements and hemoglobin decline compared to a single application in patients with total knee and hip replacement. However, future research is needed to validate this result because we did not focus the study on inapparent postoperative blood loss. Long bedridden periods and operation times may increase the expenses associated with operations. More importantly, adverse events, such as, hypostatic pneumonia, deep vein thrombosis and pulmonary embolism are associated with increased morbidity and mortality [21]. Early weight-bearing and rehabilitation have been shown to contribute to better functional outcomes following knee

M.M. Peng Zhang et al. / International Journal of Surgery 43 (2017) 171e180

Experimental Study or Subgroup

Mean

Control

SD Total Mean

Mean Difference

SD Total Weight

2.56 0.53

92

2.73 0.55

92

35.9%

-0.17 [-0.33, -0.01]

Jain 2015

1.14

0.5

59

1.82 0.61

60

34.3%

-0.68 [-0.88, -0.48]

Lin 2015

0.8

0.7

40

40

29.8%

-0.50 [-0.81, -0.19]

192 100.0%

-0.44 [-0.79, -0.09]

Total (95% CI)

0.7

191

Mean Difference

IV, Random, 95% CI

Huang 2014

1.3

177

IV, Random, 95% CI

Heterogeneity: Tau² = 0.08; Chi² = 16.16, df = 2 (P = 0.0003); I² = 88%

-1

Test for overall effect: Z = 2.49 (P = 0.01)

-0.5

0

Favours [experimental]

0.5

1

Favours [control]

Fig. 4. Forest plot diagram showing effect of combination TXA on hemoglobin decline.

Experimental Study or Subgroup

Control

Risk Difference

Total Events Total Weight

Events

Huang 2014

6

92

10

92

26.1%

-0.04 [-0.12, 0.04]

Jain 2015

1

59

4

60

16.9%

-0.05 [-0.12, 0.02]

Karaaslan 2015

7

41

13

40

11.5%

-0.15 [-0.34, 0.03]

Lin 2015

0

40

1

40

11.4%

-0.03 [-0.09, 0.04]

Xie 2015

0

70

3

70

19.9%

-0.04 [-0.10, 0.01]

Zeng 2016

1

50

8

50

14.2%

-0.14 [-0.25, -0.03]

352 100.0%

-0.07 [-0.11, -0.03]

Total (95% CI)

352 15

Total events

Risk Difference

M-H, Fixed, 95% CI

M-H, Fixed, 95% CI

39

Heterogeneity: Chi² = 5.66, df = 5 (P = 0.34); I² = 12%

-0.5

Test for overall effect: Z = 3.52 (P = 0.0004)

-0.25

0

Favours [experimental]

0.25

0.5

Favours [control]

Fig. 5. Forest plot diagram showing effect of combination TXA on transfusion rate.

Experimental Study or Subgroup

Mean

Control

SD Total Mean

Mean Difference

SD Total Weight

328

104

92

364

86

92

32.7%

Lin 2015

56.3

34.6

40 110.9

61.3

40

52.2% -54.60 [-76.41, -32.79]

50 126.8 91.91

50

15.2%

Total (95% CI)

127.2 113.52

182

IV, Fixed, 95% CI

-36.00 [-63.58, -8.42]

Huang 2014 Zeng 2016

Mean Difference

IV, Fixed, 95% CI

0.40 [-40.09, 40.89]

182 100.0% -40.19 [-55.95, -24.43]

Heterogeneity: Chi² = 5.63, df = 2 (P = 0.06); I² = 64% Test for overall effect: Z = 5.00 (P < 0.00001)

-200

-100

Favours [experimental]

0

100

200

Favours [control]

Fig. 6. Forest plot diagram showing effect of combination TXA on drainage volume.

and hip arthroplasty. The present meta-analysis indicated that the combined administration of tranexamic acid was not associated with prolonged bedridden durations or operation times following total knee and hip arthroplasty compared to single application. Infection is relatively rare after TKA but can be devastating in terms of morbidity and costs [28]. The current meta-analysis showed no significant difference in the incidence of infection, which was 2/202 in the experimental groups and 3/202 in the control groups. The overall infection incidence was 1.24%, which was in accordance with previous studies reporting an incidence ranging between 1% and 3% [29]. Trials with large sample sizes and

high quality are required to further explore the correlation between infection and tranexamic acid application. Deep vein thrombosis has been identified as a common complication that may develop into PE and even result in death following major orthopedic surgery [30]. All participants received routine prophylactic antithrombotic therapy. Previous studies have reported a higher risk of developing deep vein thrombosis and pulmonary embolism when tranexamic acid is utilized [31]. This finding may be due to the tendency of tranexamic acid, which is an antifibrinolytic agent, to increase the risk of clotting [5]. Although the combined application of tranexamic acid could be likely to

178

M.M. Peng Zhang et al. / International Journal of Surgery 43 (2017) 171e180

Experimental Study or Subgroup

Mean

Control

SD Total Mean

Mean Difference

SD Total Weight

Huang 2014

76

9.8

92

72

8.8

92

30.8%

Karaaslan 2015

75

18

41

85

15

40

19.5% -10.00 [-17.21, -2.79]

Lin 2015

101.9

12.9

40 100.4

10.3

40

24.7%

1.50 [-3.62, 6.62]

Xie 2015

67.04 14.92

70 64.81 15.07

70

25.1%

2.23 [-2.74, 7.20]

242 100.0%

0.22 [-4.65, 5.08]

Total (95% CI)

243

Mean Difference IV, Random, 95% CI

IV, Random, 95% CI 4.00 [1.31, 6.69]

Heterogeneity: Tau² = 18.13; Chi² = 12.79, df = 3 (P = 0.005); I² = 77%

-20

Test for overall effect: Z = 0.09 (P = 0.93)

-10

0

Favours [experimental]

10

20

Favours [control]

Fig. 7. Forest plot diagram showing effect of combination TXA on operation time.

Experimental Study or Subgroup

Control

Mean Difference

SD Total Weight

SD Total Mean

Mean

Huang 2014

6.9

0.9

92

7.2

0.8

92

30.1% -0.30 [-0.55, -0.05]

Karaaslan 2015

4.2

0.4

41

4.2

0.4

40

60.1%

0.00 [-0.17, 0.17]

4.39 1.28

70

4.43 1.33

70

9.8%

-0.04 [-0.47, 0.39]

Xie 2015

IV, Fixed, 95% CI

202 100.0% -0.09 [-0.23, 0.04]

203

Total (95% CI)

Mean Difference

IV, Fixed, 95% CI

Heterogeneity: Chi² = 3.87, df = 2 (P = 0.14); I² = 48%

-1

Test for overall effect: Z = 1.37 (P = 0.17)

-0.5

0

Favours [experimental]

0.5

1

Favours [control]

Fig. 8. Forest plot diagram showing effect of combination TXA on length of stay.

Experimental Study or Subgroup

Events

Control

Risk Difference

Total Events Total Weight

Huang 2014

1

92

1

92

45.5%

0.00 [-0.03, 0.03]

Lin 2015

0

40

0

40

19.8%

0.00 [-0.05, 0.05]

Xie 2015

1

70

2

70

34.7%

-0.01 [-0.06, 0.03]

202 100.0%

-0.00 [-0.03, 0.02]

202

Total (95% CI) Total events

Risk Difference M-H, Fixed, 95% CI

M-H, Fixed, 95% CI

3

2

Heterogeneity: Chi² = 0.29, df = 2 (P = 0.86); I² = 0%

-0.1

Test for overall effect: Z = 0.41 (P = 0.68)

-0.05

Favours [experimental]

0

0.05

0.1

Favours [control]

Fig. 9. Forest plot diagram showing effect of combination TXA on risk of superficial infection.

result in the formation of a thrombus, no significant difference was found between the groups in terms of the incidence of deep vein thrombosis or pulmonary embolism in the present analysis. However, large-scale trials are needed to further investigate and to confirm the observed outcomes. Several potential limitations of this study should be noted. (1) Only six RCTs were included, and the sample size was relatively small. (2) Some outcome parameters such as the visual pain score (VAS) and range of motion (ROM), were not fully described and could not be included in the meta-analysis. (3) Due to the limited number of included studies, subgroup analyses were not performed for hemoglobin decline and operation time; therefore, we could not determine the sources of heterogeneity. (4) Short-term follow-up may lead to the underestimation of complications. (5) Publication

bias is an inherent weakness that exists in all meta-analyses. Despite the aforementioned limitations, this study is the first meta-analysis to pool the results from randomized controlled trials to evaluate the efficiency and safety of the combined application of tranexamic acid compared to a single application in patients with total knee and hip arthroplasty. Furthermore, our meta-analysis also provided implications for future research about optimal dose, appropriate application and adverse effects associated with tranexamic acid. 5. Conclusions Combined administration of tranexamic acid in total knee and hip arthroplasty was associated with significantly reduced total

M.M. Peng Zhang et al. / International Journal of Surgery 43 (2017) 171e180

Experimental Study or Subgroup

Events

Control

Risk Difference

Total Events Total Weight

0

92

1

92

30.5%

-0.01 [-0.04, 0.02]

Jain 2015

0

59

1

60

19.7%

-0.02 [-0.06, 0.03]

Karaaslan 2015

1

41

0

40

13.4%

0.02 [-0.04, 0.09]

Lin 2015

0

40

0

40

13.2%

0.00 [-0.05, 0.05]

Xie 2015

2

70

1

70

23.2%

0.01 [-0.03, 0.06]

302 100.0%

-0.00 [-0.02, 0.02]

302

3

Total events

Risk Difference

M-H, Fixed, 95% CI

Huang 2014

Total (95% CI)

179

M-H, Fixed, 95% CI

3

Heterogeneity: Chi² = 1.92, df = 4 (P = 0.75); I² = 0%

-0.1

Test for overall effect: Z = 0.00 (P = 1.00)

-0.05

0

Favours [experimental]

0.05

0.1

Favours [control]

Fig. 10. Forest plot diagram showing effect of combination TXA on risk of Deep Vein Thrombosis.

Experimental Study or Subgroup Huang 2014

Mean 867

Control

SD Total 374

Jain 2015

385.68 182.5

Lin 2015

578.7 246.9

Total (95% CI)

92

Mean 957

Mean Difference

SD Total Weight 285

92

25.3%

59 590.69 191.1

60

51.9% -205.01 [-272.14, -137.88]

40

40

22.8%

705.1 213.9

191

Mean Difference

IV, Fixed, 95% CI

IV, Fixed, 95% CI

-90.00 [-186.08, 6.08] -126.40 [-227.63, -25.17]

192 100.0% -157.96 [-206.31, -109.61]

Heterogeneity: Chi² = 4.18, df = 2 (P = 0.12); I² = 52%

-200

Test for overall effect: Z = 6.40 (P < 0.00001)

-100

Favours [experimental]

0

100

200

Favours [control]

Fig. 11. Subgroup analysis for total blood loss.

blood loss, postoperative hemoglobin decline, drainage volume, and transfusion requirements. Based on the limitations of the current meta-analysis, well-designed, high-quality RCTs with longterm follow-up are still required. Conflicts of interest Each author certifies that he has no commercial associations that might pose a conflict of interest with the submitted article. Sources of funding None. Ethical approval All analyses were based on previous published studies, thus no ethical approval and patient consent are required. Research registry number-UIN reviewregistry112. Trail registry number-ISRCTN None.

Author contribution Peng Zhang and Ji-feng Li: data collections, revised the manuscript and writing. Xiao Wang: study design. Guarantor Xiao Wang. Acknowledgments None. References [1] T.P. Sculco, Global blood management in orthopaedic surgery, Clin. Orthop. Relat. Res. (357) (1998) 43e49. [2] P.M. Mannucci, M. Levi, Prevention and treatment of major blood loss, N. Engl. J. Med. 356 (22) (2007) 2301e2311. [3] F. Conteduca, F. Massai, R. Iorio, E. Zanzotto, D. Luzon, A. Ferretti, Blood loss in computer-assisted mobile bearing total knee arthroplasty. A comparison of computer-assisted surgery with a conventional technique, Int. Orthop. 33 (6) (2009) 1609e1613. [4] M.G. Bilgili, E. Ercin, G. Peker, C. Kural, S.H. Basaran, A. Duramaz, C. Avkan, Efficiency and cost analysis of cell saver auto transfusion system in total knee arthroplasty, Balkan Med. J. 31 (2) (2014) 149e153. [5] G. Benoni, S. Lethagen, H. Fredin, The effect of tranexamic acid on local and plasma fibrinolysis during total knee arthroplasty, Thrombosis Res. 85 (3) (1997) 195e206. [6] J. Cid, M. Lozano, Tranexamic acid reduces allogeneic red cell transfusions in patients undergoing total knee arthroplasty: results of a meta-analysis of randomized controlled trials, Transfusion 45 (8) (2005) 1302e1307. [7] X. Xu, S. Xiong, Z. Wang, X. Li, W. Liu, Topical administration of tranexamic

180

[8]

[9]

[10]

[11]

[12]

[13]

[14]

[15]

[16]

[17]

[18]

M.M. Peng Zhang et al. / International Journal of Surgery 43 (2017) 171e180 acid in total hip arthroplasty: a meta-analysis of Randomized Controlled Trials, Drug Discov. Ther. 9 (3) (2015) 173e177. Z. Huang, J. Ma, B. Shen, F. Pei, Combination of intravenous and topical application of tranexamic acid in primary total knee arthroplasty: a prospective randomized controlled trial, J. Arthroplast. 29 (12) (2014) 2342e2346. S.Y. Lin, C.H. Chen, Y.C. Fu, P.J. Huang, J.K. Chang, H.T. Huang, The efficacy of combined use of intraarticular and intravenous tranexamic acid on reducing blood loss and transfusion rate in total knee arthroplasty, J. Arthroplast. 30 (5) (2015) 776e780. J. Xie, J. Ma, C. Yue, P. Kang, F. Pei, Combined use of intravenous and topical tranexamic acid following cementless total hip arthroplasty: a randomised clinical trial, Hip Int. J. Clin. Exp. Res. Hip Pathol. Ther. (2015) 0. F. Karaaslan, S. Karaoglu, M.U. Mermerkaya, A. Baktir, Reducing blood loss in simultaneous bilateral total knee arthroplasty: combined intravenous-intraarticular tranexamic acid administration. A prospective randomized controlled trial, Knee 22 (2) (2015) 131e135. N.P. Jain, P.P. Nisthane, N.A. Shah, Combined administration of systemic and topical tranexamic acid for total knee arthroplasty: can it Be a better regimen and yet Safe? A randomized controlled trial, J. Arthroplast. 31 (2) (2016 Feb) 542e547. Z. Yi, S. Bin, Y. Jing, Z. Zongke, K. Pengde, P. Fuxing, Tranexamic acid administration in primary total hip arthroplasty: a randomized controlled trial of intravenous combined with topical versus single-dose intravenous administration, J. bone Jt. Surg. Am. Vol. 98 (12) (2016) 983e991. S. Alshryda, M. Sukeik, P. Sarda, J. Blenkinsopp, F.S. Haddad, J.M. Mason, A systematic review and meta-analysis of the topical administration of tranexamic acid in total hip and knee replacement, bone & Jt. J. 96-B (8) (2014) 1005e1015. H. Wang, B. Shen, Y. Zeng, Comparison of topical versus intravenous tranexamic acid in primary total knee arthroplasty: a meta-analysis of randomized controlled and prospective cohort trials, Knee 21 (6) (2014) 987e993. Z. Wei, M. Liu, The effectiveness and safety of tranexamic acid in total hip or knee arthroplasty: a meta-analysis of 2720 cases, Transfus. Med. 25 (3) (2015) 151e162. H.K. Ronday, J.M. Te Koppele, R.A. Greenwald, S.A. Moak, J.A. De Roos, B.A. Dijkmans, F.C. Breedveld, J.H. Verheijen, Tranexamic acid, an inhibitor of plasminogen activation, reduces urinary collagen cross-link excretion in both experimental and rheumatoid arthritis, Br. J. Rheumatol. 37 (1) (1998) 34e38. G. Benoni, H. Fredin, Fibrinolytic inhibition with tranexamic acid reduces blood loss and blood transfusion after knee arthroplasty: a prospective, randomised, double-blind study of 86 patients, J. bone Jt. Surg. Br. Vol. 78 (3) (1996) 434e440.

[19] M.A. Camarasa, G. Olle, M. Serra-Prat, A. Martin, M. Sanchez, P. Ricos, A. Perez, L. Opisso, Efficacy of aminocaproic, tranexamic acids in the control of bleeding during total knee replacement: a randomized clinical trial, Br. J. Anaesth. 96 (5) (2006) 576e582. [20] S. Hiippala, L. Strid, M. Wennerstrand, V. Arvela, S. Mantyla, J. Ylinen, H. Niemela, Tranexamic acid (Cyklokapron) reduces perioperative blood loss associated with total knee arthroplasty, Br. J. Anaesth. 74 (5) (1995) 534e537. [21] M. Veien, J.V. Sorensen, F. Madsen, P. Juelsgaard, Tranexamic acid given intraoperatively reduces blood loss after total knee replacement: a randomized, controlled study, Acta Anaesthesiol. Scand. 46 (10) (2002) 1206e1211. [22] T. Danninger, R. Rasul, J. Poeran, O. Stundner, M. Mazumdar, P.M. Fleischut, L. Poultsides, S.G. Memtsoudis, Blood transfusions in total hip and knee arthroplasty: an analysis of outcomes, Sci. World J. 2014 (2014) 623460. [23] N.B. Frisch, N.M. Wessell, M.A. Charters, S. Yu, J.J. Jeffries, C.D. Silverton, Predictors and complications of blood transfusion in total hip and knee arthroplasty, J. Arthroplast. 29 (9 Suppl) (2014) 189e192. [24] A.K. Klika, T.J. Small, A. Saleh, C.R. Szubski, A.L. Chandran Pillai, W.K. Barsoum, Primary total knee arthroplasty allogenic transfusion trends, length of stay, and complications: nationwide inpatient sample 2000-2009, J. Arthroplast. 29 (11) (2014) 2070e2077. [25] H. Yoshihara, D. Yoneoka, National trends in the utilization of blood transfusions in total hip and knee arthroplasty, J. Arthroplast. 29 (10) (2014) 1932e1937. [26] G. Digas, I. Koutsogiannis, G. Meletiadis, E. Antonopoulou, V. Karamoulas, C. Bikos, Intra-articular injection of tranexamic acid reduce blood loss in cemented total knee arthroplasty, Eur. J. Orthop. Surg. Traumatol. Orthop. Traumatol. 25 (7) (2015) 1181e1188. [27] M. Pitta, M. Zawadsky, R. Verstraete, A. Rubinstein, Intravenous administration of tranexamic acid effectively reduces blood loss in primary total knee arthroplasty in a 610-patient consecutive case series, Transfusion 56 (2) (2016) 466e471. [28] M. de Dios, J. Cordero-Ampuero, Risk factors for infection in total knee artrhoplasty, including previously unreported intraoperative fracture and deep venous thrombosis, Rev.ista espanola de cirugia ortopedica y traumatologia 59 (1) (2015) 36e43. [29] A. Soriano, G. Bori, S. Garcia-Ramiro, J.C. Martinez-Pastor, T. Miana, C. Codina, F. Macule, M. Basora, J.A. Martinez, J. Riba, et al., Timing of antibiotic prophylaxis for primary total knee arthroplasty performed during ischemia, Clin. Infect. Dis. Off. Publ. Infect. Dis. Soc. Am. 46 (7) (2008) 1009e1014. [30] C. Wang, Z. Han, T. Zhang, J.X. Ma, X. Jiang, Y. Wang, X.L. Ma, The efficacy of a thrombin-based hemostatic agent in primary total knee arthroplasty: a metaanalysis, J. Orthop. Surg. Res. 9 (2014) 90. [31] R. Raveendran, J. Wong, Tranexamic acid: more evidence for its use in joint replacement surgery, Transfusion 54 (1) (2014) 2e3.