Tranexamic acid in primary total knee arthroplasty: a systematic review and meta-analysis of blood loss, transfusion, and deep vein thrombosis

Abstract

Background

Tranexamic acid (TXA) has been widely used in primary total knee arthroplasty (TKA) to reduce perioperative blood loss, but the pooled evidence regarding blood loss, transfusion, and deep vein thrombosis (DVT) remains to be further clarified. The purpose of this meta-analysis was to evaluate the efficacy and safety of TXA in primary TKA.

Methods

PubMed, Embase, Web of Science, the Cochrane Library, and relevant additional sources were searched to identify randomized controlled trials (RCTs) comparing TXA with a control regimen in primary unilateral TKA. The primary outcomes were blood loss, transfusion, and DVT. Continuous outcomes were pooled as mean differences (MDs) with 95% confidence intervals (CIs), and dichotomous outcomes were pooled as risk ratios (RRs) or odds ratios (ORs) with 95% CIs. Risk of bias was assessed using the Cochrane RoB 2 tool. Random-effects models were applied for meta-analysis. For the blood-loss outcome, an additional robustness analysis was performed using an empirical Bayes random-effects model with Knapp–Hartung standard error adjustment to evaluate whether the result was sensitive to model assumptions, given that methodological choices in random-effects meta-analysis may influence statistical significance when substantial heterogeneity is present.

Results

A total of 17 studies were included in the systematic review and meta-analysis. TXA was associated with significantly reduced blood loss compared with control (MD = -423.28 mL, 95% CI -536.90 to -309.66; P < 0.001), although substantial heterogeneity was observed (I² = 93.44%). Leave-one-out sensitivity analysis showed that the pooled effect remained significant after exclusion of individual studies; exclusion of one influential study reduced heterogeneity to 81.88%. The Knapp–Hartung-adjusted robustness analysis yielded a similar significant effect estimate. TXA also significantly reduced transfusion risk (RR = 0.55, 95% CI 0.45 to 0.68; P < 0.001), with no significant heterogeneity (I² = 0.0%). No significant difference was found between the TXA and control groups in the incidence of DVT (OR = 0.75, 95% CI 0.47 to 1.22; P = 0.25), and heterogeneity was low (I² = 0.0%). Funnel plot inspection suggested possible asymmetry for blood loss and transfusion; however, Egger’s test was not significant for blood loss (P = 0.842), whereas significant small-study effects were detected for transfusion (P = 0.002).

Conclusions

The current evidence indicates that TXA in primary TKA significantly reduces blood loss and transfusion requirements with no detected increase in DVT risk in the available randomized evidence. The blood loss analysis showed substantial heterogeneity, whereas the transfusion and DVT analyses were more consistent across studies.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12893-026-04060-2.

Background

Total knee arthroplasty (TKA) is an effective and widely performed procedure for patients with end-stage knee disease, providing substantial pain relief and functional improvement [1]. However, perioperative blood loss remains an important concern in primary TKA and may contribute to postoperative anemia, allogeneic transfusion, prolonged hospitalization, and increased healthcare burden [2, 3]. Because transfusion is associated with additional costs and potential complications, strategies to minimize perioperative blood loss have become an important component of contemporary blood management in TKA [2, 4].

Tranexamic acid (TXA), a synthetic lysine analogue, inhibits fibrinolysis by blocking the binding of plasminogen and plasmin to fibrin, thereby reducing perioperative bleeding [3]. Over the past decade, TXA has been administered through different routes in TKA, including intravenous (IV), intra-articular (IA), topical, and oral regimens, and has become one of the most commonly used blood-conservation measures in arthroplasty practice [5]. Randomized controlled trials and meta-analyses have consistently suggested that TXA can reduce perioperative blood loss and transfusion requirements in primary TKA without a clear increase in thromboembolic complications [4, 6, 7].

Despite the growing body of literature, several issues remain relevant. First, the reported blood-loss outcomes are not entirely uniform across studies, with authors variably reporting total blood loss, estimated blood loss, postoperative bleeding, or drainage volume [6]. Second, TXA regimens differ substantially with regard to route, dose, timing, and number of administrations, which may contribute to heterogeneity across trials [6, 8]. Third, although the efficacy of TXA in reducing blood loss is widely accepted, the pooled evidence focusing simultaneously on blood loss, transfusion, and deep vein thrombosis (DVT) in primary TKA still warrants careful synthesis, particularly when only randomized evidence is considered.

From a clinical perspective, blood loss, transfusion, and DVT represent three highly relevant outcomes in primary TKA. Blood loss reflects the direct hemostatic efficacy of TXA, transfusion is one of the most clinically meaningful downstream outcomes, and DVT remains a key safety concern whenever antifibrinolytic agents are used perioperatively [2, 4, 6]. A focused synthesis of these outcomes may therefore provide practical evidence for perioperative decision-making and for the interpretation of different TXA strategies used in routine arthroplasty care.

Therefore, the purpose of the present systematic review and meta-analysis was to evaluate the efficacy and safety of TXA in primary TKA. Specifically, we aimed to compare TXA with control regimens with respect to blood loss, transfusion, and DVT in randomized controlled trials.

Methods

Search strategy

This systematic review and meta-analysis was performed in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement [9]. PubMed, Embase, Web of Science, and the Cochrane Library were systematically searched for relevant studies published from 1 January 2014 to 1 April 2026 to identify randomized controlled trials evaluating tranexamic acid (TXA) in primary total knee arthroplasty (TKA). Additional records were identified through manual screening of the reference lists of relevant studies and reviews.

The search strategy combined terms related to TKA and TXA, including “total knee arthroplasty”, “total knee replacement”, “TKA”, “TKR”, “tranexamic acid”, and “TXA”. The search strategy was adapted for each database as appropriate. A representative search strategy for PubMed was as follows: (“total knee arthroplasty” OR “total knee replacement” OR TKA OR TKR) AND (“tranexamic acid” OR TXA). Only studies published in English were considered.

Eligibility criteria

Studies were included if they met the following criteria: (1) the study design was a randomized controlled trial (RCT); (2) the study population consisted of patients undergoing primary unilateral TKA; (3) the intervention group received TXA by any route of administration; (4) the control group received placebo, saline, or no TXA; and (5) at least one of the following outcomes was reported: blood loss, transfusion, or deep vein thrombosis (DVT).

Studies were excluded if they met any of the following criteria: (1) revision TKA, bilateral TKA, or non-primary procedures; (2) no control group; (3) insufficient extractable outcome data; (4) reviews, editorials, case reports, or conference abstracts; or (5) duplicate publications or overlapping patient populations.

Study selection

Two reviewers independently screened the titles and abstracts of all identified records. Potentially eligible studies were then assessed by full-text review. Disagreements were resolved by discussion and consensus. When necessary, a third reviewer was consulted to resolve discrepancies.

Data extraction

Two reviewers independently extracted data from the included studies using a predefined data extraction form. The following information was collected: first author, publication year, country, study design, sample size, TXA regimen, control regimen, thromboprophylaxis, follow-up, and outcome data. The primary outcomes of the present meta-analysis were blood loss, transfusion, and DVT.

For continuous outcomes, mean values, standard deviations, and sample sizes were extracted. For dichotomous outcomes, the numbers of events and total patients in each group were extracted. In multi-arm trials, pairwise comparisons were handled separately for meta-analysis, whereas the studies were presented as single trials in the characteristics table.

Risk of bias assessment

The risk of bias of the included RCTs was assessed independently by two reviewers using the Cochrane Risk of Bias 2 (RoB 2) tool [10]. The following domains were evaluated: bias arising from the randomization process, bias due to deviations from intended interventions, bias due to missing outcome data, bias in measurement of the outcome, and bias in selection of the reported result. Each domain was judged as “Low”, “Some concerns”, or “High”, and an overall risk-of-bias judgment was assigned for each study. Any disagreement between reviewers was resolved by discussion.

Statistical analysis

Statistical analyses were performed using Stata 17.0 (StataCorp, College Station, TX, USA). Continuous outcomes were pooled as mean differences (MDs) with 95% confidence intervals (CIs), whereas dichotomous outcomes were pooled as risk ratios (RRs) or odds ratios (ORs) with 95% CIs. Blood loss was analyzed as a continuous outcome, transfusion as a dichotomous outcome using RR, and DVT as a dichotomous outcome using OR.

Because clinical and methodological heterogeneity across studies was anticipated, random-effects models were applied for all pooled analyses. For the blood-loss outcome, an additional robustness analysis was performed using an empirical Bayes random-effects model with Knapp–Hartung standard error adjustment to evaluate whether the result was sensitive to model assumptions, given that methodological choices in random-effects meta-analysis may influence statistical significance when substantial heterogeneity is present [11]. Statistical heterogeneity was assessed using Cochran’s Q test and the I² statistic [12]. An I² value greater than 50% was considered to indicate substantial heterogeneity.

Sensitivity analysis was performed for blood loss using a leave-one-out approach, in which the meta-analysis was repeated after sequential omission of each study to assess the robustness of the pooled estimate and to identify influential studies. Publication bias and small-study effects were assessed by visual inspection of funnel plots and by Egger’s regression test when appropriate [13]. Due to sparse events, formal assessment of publication bias was not performed for DVT. A two-sided P value < 0.05 was considered statistically significant.

Results

Study selection

A total of 1123 records were identified through database searching. After removal of 401 duplicates, 722 records remained for title and abstract screening. Of these, 652 were excluded after initial screening, and 70 full-text articles were assessed for eligibility. After full-text review, 53 reports were excluded for predefined reasons, and 17 studies were finally included in the systematic review and meta-analysis. Multi-arm trials contributed multiple pairwise comparisons to the quantitative synthesis when appropriate. The study selection process is shown in Fig. 1.

Fig. 1.

PRISMA flow diagram

Study characteristics

The main characteristics of the included studies are summarized in Table 1. All included studies were randomized controlled trials conducted in different countries, including China, Turkey, Brazil, Serbia, Greece, Spain, Nigeria, and New Zealand. The TXA regimens varied across studies with respect to route of administration, dose, and timing, and included intravenous, intra-articular, topical, local, or combined strategies. Control groups generally received placebo, saline, routine hemostasis, or no TXA. The outcomes extracted for the present meta-analysis included blood loss, transfusion, and DVT, although not all studies reported all three outcomes. Detailed perioperative protocols, including TXA dose, timing, number of administrations, transfusion criteria, DVT assessment, thromboprophylaxis, and major exclusion criteria, are summarized in Supplementary Table 2.

Table 1.

Characteristics of the included studies

Study	Country	Design	Sample size	TXA regimen	Control	Outcomes	Thromboprophylaxis	Follow-up
Ugbeye 2025 [14]	Nigeria	RCT	46/52	IV TXA	Placebo/control	BL, TR, DVT	NR	NR
Ye 2024 [15]	China	RCT	75/75/75	IA TXA	Placebo/control	BL, TR, DVT	LMWH; rivaroxaban	14 days
Jovanovic 2024 [16]	Serbia	RCT	48/48	IV TXA	Placebo/control	BL, TR	Nadroparin calcium	15–19 days
Arslan 2018 [17]	Turkey	RCT	25/25/25/25	IA TXA	Placebo/control	BL, TR	NR	NR
Digas 2015 [18]	Greece	RCT	30/30/30	IV and IA TXA	Control	BL, TR, DVT	NR	NR
Sahin 2019 [19]	Turkey	RCT	33/34	Local/IA TXA	Control	TR, DVT	NR	NR
Liu 2015 [20]	China	RCT	75/75/74	IV TXA	Control	TR, DVT	NR	NR
Wang 2015 [21]	China	RCT	30/30	IA TXA	Placebo	BL, TR, DVT	NR	NR
Yang 2014 [22]	China	RCT	40/40	IA TXA	Placebo	TR, DVT	NR	NR
Aguilera 2015 [23]	Spain	RCT	50/50/50	IV and topical TXA	Control	BL, TR	NR	NR
Volquind 2016 [24]	Brazil	RCT	32/30	IV TXA	Saline/placebo	Postop bleeding, DVT	UFH	30 days
Shen 2015 [25]	China	RCT	41/40	IV TXA	Control	BL, TR, DVT	NR	NR
Stowers 2017 [26]	New Zealand	RCT	60/51/23	IA and IV TXA	Placebo	BL, TR, DVT	NR	NR
Ugurlu 2016 [27]	Turkey	RCT	40/42/41	IV and topical TXA	Control	TR, DVT	Enoxaparin	30 days
Saglam 2025 [28]	Turkey	RCT	25/25	IV TXA	Control	BL	NR	NR
Carvalho Jr. 2015 [29]	Brazil	RCT	42/38/40	Topical TXA	Betadine only	BL, TR, DVT	Enoxaparin	12 weeks
Yen 2017 [30]	China	RCT	31/32/30	IV and topical/IA TXA	Placebo	BL, TR, DVT	Rivaroxaban	3 months

Abbreviations: BL blood loss, DVT deep vein thrombosis, IA intra-articular, IV intravenous, NR not reported, RCT randomized controlled trial, TR transfusion, TXA tranexamic acid, UFH unfractionated heparin

Risk of bias

Risk of bias was assessed using the Cochrane RoB 2 tool. Overall, most included trials were judged as having some concerns. Bias in measurement of the outcome was generally rated as low risk, whereas concerns most commonly arose from the randomization process, deviations from intended interventions, and selection of the reported result. No study was judged to be at high overall risk of bias. The overall risk-of-bias assessment is summarized in Fig. 2, and the study-level judgments are presented in Supplementary Fig. 1.

Fig. 2.

Summary plot of risk of bias for the included randomized controlled trials

Blood loss

A total of 10 pairwise comparisons were included in the primary meta-analysis of blood loss. The definitions of blood-loss outcomes varied across studies, and the specific outcome used or excluded from the quantitative blood-loss analysis is summarized in Supplementary Table 3. TXA was associated with significantly reduced blood loss compared with control (MD = -423.28 mL, 95% CI -536.90 to -309.66; P < 0.001). However, substantial heterogeneity was observed across studies (I² = 93.44%, Cochran Q = 138.54, P < 0.001), and therefore the pooled estimate was interpreted using a random-effects model. The pooled effect estimates for blood loss are presented in Fig. 3.

Fig. 3.

Forest plot of blood loss

Transfusion

A total of 13 pairwise comparisons were included in the pooled analysis of transfusion. TXA significantly reduced the risk of transfusion compared with control (RR = 0.55, 95% CI 0.45 to 0.68; P < 0.001). No relevant between-study heterogeneity was detected (I² = 0.0%, Cochran Q P = 0.685), indicating high consistency of the effect estimate across studies. The pooled analysis of transfusion is presented in Fig. 4.

Fig. 4.

Forest plot of transfusion

Deep vein thrombosis

A total of 7 pairwise comparisons were included in the pooled analysis of DVT after exclusion of double-zero comparisons from the effect estimation model. No significant difference was found between the TXA and control groups in the incidence of DVT (OR = 0.75, 95% CI 0.47 to 1.22; P = 0.25). Heterogeneity was not significant (I² = 0.0%, Cochran Q P = 0.930). The forest plot for DVT is shown in Fig. 5.

Fig. 5.

Forest plot of DVT

Sensitivity analysis

Because the blood loss analysis demonstrated substantial heterogeneity, a leave-one-out sensitivity analysis was performed. The pooled effect remained statistically significant after omission of any single study, indicating that the overall conclusion was robust. Exclusion of Saglam_2025_IV produced the greatest reduction in heterogeneity, decreasing I² from 93.44% to 81.88%, while the pooled effect remained significant (MD = -373.97 mL, 95% CI -451.60 to -296.35). In contrast, exclusion of other influential studies, including Aguilera_2015_IV, Wang_2015_IA, and Shen_2015_IV, did not materially reduce heterogeneity. The leave-one-out sensitivity analysis plot is presented in Supplementary Fig. 2, and the corresponding numerical results are summarized in Supplementary Table 1. Additional model-based robustness analysis using an empirical Bayes random-effects model with Knapp–Hartung standard error adjustment showed a similar pooled estimate, and the reduction in blood loss remained statistically significant (MD = -423.26 mL, 95% CI -553.28 to -293.23; P < 0.001; Supplementary Table 4).

Publication bias

Visual inspection of the funnel plot for blood loss suggested possible asymmetry. However, Egger’s test did not indicate significant small-study effects (P = 0.842), suggesting no clear statistical evidence of publication bias for this outcome.

For transfusion, the funnel plot also appeared asymmetric, and Egger’s test demonstrated significant small-study effects (P = 0.002), suggesting the possibility of publication bias or other small-study effects. Because the DVT analysis included sparse events and a limited number of informative comparisons, formal publication bias testing was not performed for this outcome. Funnel plots for blood loss and transfusion are presented in Supplementary Figs. 3 and 4, respectively. The main pooled results of the meta-analysis are summarized in Table 2.

Table 2.

Summary of meta-analysis results

Outcome	Studies (comparisons)	Effect measure	Model	Pooled effect (95% CI)	P value	I² (%)	Heterogeneity P value	Sensitivity analysis	Publication bias
Blood loss	10 [10]	MD (mL)	Random-effects	-423.28 (-536.90 to -309.66)	< 0.001	93.44	< 0.001	Excluding Saglam_2025_IV reduced heterogeneity (MD -373.97, 95% CI -451.60 to -296.35; I²=81.88%)	Funnel plot showed visual asymmetry; Egger’s test p = 0.842
Transfusion	13 [13]	RR	Random-effects	0.55 (0.45 to 0.68)	< 0.001	0.00	0.685	Not performed	Funnel plot showed visual asymmetry; Egger’s test p = 0.002
DVT	7 [7]	OR	Random-effects	0.75 (0.47 to 1.22)	0.25	0.00	0.930	Not performed	Not assessed due to sparse events

Abbreviations: CI confidence interval, DVT deep vein thrombosis, MD mean difference, OR odds ratio, RR risk ratio

Discussion

The most important finding of the present meta-analysis is that TXA in primary TKA significantly reduced blood loss and transfusion requirements with no detected increase in DVT risk in the available randomized evidence. Specifically, TXA was associated with a marked reduction in blood loss and a significantly lower transfusion risk compared with control. In contrast, no statistically significant difference was observed in the incidence of DVT between the two groups. These findings support the efficacy of TXA as a blood-conservation strategy in primary TKA while also suggesting an acceptable thromboembolic safety profile [3, 4].

Reduction of perioperative blood loss remains one of the principal goals of blood management in TKA. Considerable blood loss may occur during and after primary TKA because of surgical trauma, fibrinolytic activation, and, in some settings, tourniquet-related effects [2, 31]. In the current study, the pooled estimate showed that TXA reduced blood loss by more than 400 mL, which is clinically meaningful and consistent with the pharmacological action of TXA as an antifibrinolytic agent [3, 6]. This result is also in agreement with previous randomized trials and meta-analyses demonstrating that TXA can effectively reduce perioperative bleeding in arthroplasty [3, 4, 7].

The transfusion analysis provided another clinically important finding. Although blood loss is a useful efficacy outcome, transfusion is generally regarded as a more directly relevant clinical endpoint because it reflects the downstream consequences of perioperative bleeding and is closely related to resource utilization and transfusion-related risks [2, 32]. In the present meta-analysis, TXA significantly reduced the risk of transfusion, and the heterogeneity for this outcome was negligible. Compared with the blood loss analysis, the transfusion result was therefore more consistent across the included trials and may be considered particularly robust. This finding is also in line with earlier evidence that TXA reduces allogeneic transfusion in TKA [4].

Contemporary evidence also indicates that perioperative transfusion after arthroplasty is influenced by multiple patient- and procedure-related factors, including preoperative hemoglobin level, operative time, intraoperative blood loss, body mass index, and comorbidity burden [33, 34]. Therefore, the transfusion benefit observed in the present meta-analysis should be interpreted within the broader context of modern patient blood management, in which TXA represents one component of a multimodal strategy rather than an isolated determinant of transfusion risk.

Regarding safety, no significant difference in DVT incidence was found between the TXA and control groups. This observation is clinically relevant because concerns about thromboembolic risk remain one of the most common reasons for caution in the perioperative use of TXA. However, as in previous meta-analyses and randomized studies, the available evidence did not demonstrate an increased risk of postoperative venous thromboembolism associated with TXA in primary TKA [4, 6, 35]. It should nevertheless be noted that DVT was a relatively infrequent event, and the number of informative comparisons for this outcome was smaller than that for blood loss or transfusion. Therefore, although the present findings are reassuring, the safety conclusion should still be interpreted in the context of limited event numbers. In the DVT analysis, double-zero studies were excluded from the effect estimation model because they did not contribute directly to the relative effect estimate. However, this approach may reduce the amount of safety information considered, especially when thromboembolic events are rare. Therefore, the absence of a statistically significant increase in DVT should be interpreted cautiously and should not be regarded as definitive evidence of equivalent thromboembolic safety across all patient populations.

One notable issue in our meta-analysis was the substantial heterogeneity observed in the blood loss outcome. Several factors may explain this finding. First, the definition of blood loss was not fully consistent across studies; although the primary analysis focused on the most comparable continuous blood-loss measures, the underlying definitions across trials were not entirely uniform. Second, the TXA protocols varied substantially with regard to route of administration, dose, and timing. Third, some included trials were multi-arm studies, and the selection of pairwise comparisons may also have contributed to variation. Finally, perioperative management strategies, including tourniquet use, drainage protocols, and thromboprophylaxis, differed across studies. These considerations likely account for a substantial proportion of the observed heterogeneity and are consistent with the complexity reported in previous TKA meta-analyses [6, 8].

The leave-one-out sensitivity analysis further clarified the heterogeneity pattern. Exclusion of Saglam_2025_IV led to the greatest reduction in heterogeneity, while the pooled effect remained statistically significant. This suggests that the overall finding for blood loss was not dependent on any single study, although Saglam_2025_IV was an important contributor to between-study variability. In contrast, omission of other potentially influential studies did not materially change the heterogeneity estimate. Taken together, these findings support the robustness of the direction of effect while also highlighting the need for cautious interpretation of the exact magnitude of blood-loss reduction.

Publication bias assessment also deserves comment. For blood loss, the funnel plot showed some visual asymmetry, but Egger’s test was not significant, indicating no clear statistical evidence of small-study effects. For transfusion, however, funnel-plot asymmetry was more apparent and Egger’s test was significant, suggesting possible publication bias or other small-study effects. This does not invalidate the overall transfusion result, particularly given the consistency and low heterogeneity across studies, but it indicates that the pooled estimate may be influenced to some extent by study size or selective dissemination of results. This issue should therefore be acknowledged when interpreting the strength of the evidence [13].

The present study has several strengths. First, only randomized controlled trials were included, which improves the overall internal validity of the evidence base. Second, the analysis focused specifically on primary TKA, thereby reducing clinical heterogeneity related to different surgical populations. Third, the three outcomes selected for quantitative synthesis—blood loss, transfusion, and DVT—represent clinically meaningful measures of efficacy and safety. Fourth, sensitivity analysis and publication bias assessment were performed, allowing a more comprehensive interpretation of the findings.

Several limitations should also be acknowledged. First, substantial heterogeneity was present in the blood loss analysis. Second, TXA regimens were not uniform across studies, particularly regarding dose, route, and timing of administration. Moreover, pooling heterogeneous blood-loss measures remains methodologically challenging, and future evidence syntheses may benefit from harmonization approaches such as MCID-based normalization, which can transform heterogeneous outcomes into a common clinically interpretable scale before quantitative synthesis [36]. Third, not all included studies contributed data to all three outcomes, and the number of studies available for DVT was relatively limited. Fourth, some outcomes, especially DVT, were affected by sparse events. Fifth, although the risk-of-bias assessment did not identify studies at high overall risk, most trials were judged to have some concerns, particularly in the domains of randomization process and selection of the reported result. Finally, the current meta-analysis did not evaluate some potentially relevant outcomes, such as hemoglobin drop, drainage volume, pain, function, or length of stay, because the predefined focus of the study was blood loss, transfusion, and DVT.

In conclusion, the present meta-analysis indicates that TXA in primary TKA effectively reduces blood loss and transfusion requirements without a detected increase in DVT risk. Although substantial heterogeneity was observed in the blood-loss analysis, the direction of effect remained consistent across leave-one-out sensitivity analysis and additional model-based robustness analysis using Knapp–Hartung adjustment. Therefore, the exact magnitude of the pooled MD should be interpreted cautiously, but the overall direction and clinical relevance of the TXA effect appear methodologically robust.

Conclusions

The current evidence indicates that TXA in primary TKA significantly reduces blood loss and transfusion requirements compared with control, without a detected increase in DVT risk. The pooled analyses for transfusion and DVT were consistent across studies, whereas the blood loss analysis showed substantial heterogeneity. Overall, TXA appears to be an effective and safe blood-conservation strategy in primary TKA. Further high-quality randomized trials with standardized TXA regimens and uniform outcome definitions are warranted to strengthen the evidence base and refine clinical recommendations.

Abbreviations

BL

Blood loss

CI

Confidence interval

DVT

Deep vein thrombosis

IA

Intra-articular

IV

Intravenous

MD

Mean difference

OR

Odds ratio

RCT

Randomized controlled trial

RR

Risk ratio

TKA

Total knee arthroplasty

TR

Transfusion

TXA

Tranexamic acid

Authors' contributions

J. W. Feng drafted the manuscript and completed data analysis. S. H. Liang, L. Y. Yang, X. Chen and M. Zhang were responsible for literature screening and data extraction. Y. Jin and J. C. Huang conceived and designed the study, revised the manuscript critically and provided funding support. All authors read and approved the final manuscript.

Funding

This work and its publication were supported by the following grants: Henan Provincial and Ministry Co-construction Project (SBGJ202102031); National Natural Science Foundation of China (82002840); Key Scientific and Technological Projects in Henan Province (LHGJ20240046); Key Research Projects of Henan Provincial Colleges and Universities (26A320039); and the Project of the Center for School Planning, Construction and Development of the Ministry of Education (CSDP26LF8A409). The funders had no role in the study design, literature search, study selection, data extraction, statistical analysis, interpretation of the results, preparation of the manuscript, or the decision to submit the manuscript for publication.

Data availability

All data generated or analyzed during this study are included in this published article and its supplementary materials. The datasets used for the meta-analysis were extracted from previously published studies.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.