Antiplatelets match anticoagulants for VTE prevention after joint arthroplasty: a meta-analysis of randomized clinical trials

Abstract

Purpose

• Venous thromboembolism (VTE), encompassing deep vein thrombosis (DVT) and pulmonary embolism (PE), remains a major risk after joint arthroplasty. This study compares antiplatelet agents versus anticoagulants for VTE prevention in adults undergoing joint arthroplasty.

Methods

• The review process was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. Systematic searches were conducted in PubMed, EMBASE, Web of Science, and the Cochrane Central Register of Controlled Trials. Randomized controlled trials (RCTs) comparing antiplatelets with anticoagulants for VTE prevention in adults undergoing joint arthroplasty were included. Data were pooled using a random effects meta-analysis to account for heterogeneity.

Results

• Nineteen RCTs involving 30,290 patients met the inclusion criteria. No significant difference was found between antiplatelet agents and anticoagulants in preventing VTE (risk ratio (RR): 1.08; 95% CI: 0.81–1.44), DVT (RR: 1.12; 95% CI: 0.83–1.51), or PE (RR: 1.26; 95% CI: 0.94–1.69). However, in studies conducted after 2015 and those incorporating mechanical prophylaxis, the antiplatelet group had higher rates of VTE, DVT, and PE. Notably, female patients on antiplatelets had lower VTE rates, suggesting potential sex-specific benefits. Regarding safety, antiplatelet agents were associated with fewer bleeding complications compared to anticoagulants.

Conclusion

• Antiplatelet agents are as effective as anticoagulants in preventing VTE, DVT, and PE after joint arthroplasty and may offer a better safety profile with fewer bleeding complications. Despite some heterogeneity and potential biases, the evidence supports considering antiplatelet agents, particularly aspirin, as a cost-effective and reliable option for VTE prophylaxis in this patient population.

Introduction

Joint arthroplasty, particularly hip and knee replacements, is an effective intervention for alleviating pain and improving joint function in patients with conditions such as osteoarthritis (1). However, postoperative asymptomatic venous thromboembolism (VTE), including deep vein thrombosis (DVT) and pulmonary embolism (PE), is a significant risk, with an incidence as high as 60% within 90 days without prophylaxis (2, 3). VTE increases morbidity, hospitalization, and healthcare costs and can be life-threatening if emboli reach vital organs such as the lungs (4). To prevent VTE, anticoagulants such as rivaroxaban, low-molecular-weight heparin (LMWH), and warfarin are commonly used (5, 6, 7), though they carry a risk of bleeding and require careful management (8, 9).

Recently, antiplatelet agents such as aspirin have emerged as simpler, potentially safer alternatives for VTE prevention post-joint arthroplasty (10). By inhibiting platelet aggregation, these drugs reduce the likelihood of clot formation. Studies suggest that aspirin alone offers efficacy comparable to traditional anticoagulants for VTE prevention, with the added advantages of lower cost, wider availability, and easier administration (10, 11, 12).

Evidence increasingly supports the use of antiplatelet therapy, particularly aspirin, for VTE prevention after joint arthroplasty, potentially signaling a shift in clinical practice. Although anticoagulants remain essential, aspirin’s comparable effectiveness, lower cost, and ease of use make it an appealing option, especially where cost and healthcare access are concerns. However, uncertainties persist regarding the optimal prophylactic strategy, as studies show varied results. This meta-analysis aims to systematically evaluate the effectiveness and safety of antiplatelet agents versus anticoagulants in VTE prevention, pooling data from RCTs. By examining potential heterogeneity sources, we seek to provide clearer guidance for clinical practice and improve patient care in joint arthroplasty.

Methods

This meta-analysis was performed according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines and has been reported in line with the AMSTAR (Assessing the methodological quality of systematic reviews) Guideline (13, 14). The protocol for this meta-analysis was registered on PROSPERO (Registration No: CRD 42024553061).

Inclusion criteria

The inclusion criteria are i) RCTs evaluating the efficacy and safety of antiplatelets and anticoagulants for the prevention of VTE in adults (≥18 years old) undergoing joint (hip or knee) arthroplasty; ii) studies reporting outcomes including VTE, DVT, PE, and associated complications; and iii) trials that implemented a combined VTE prevention strategy with antiplatelets as one of the medications (e.g. initial treatment with LMWH followed by a longer course of antiplatelets) to reflect current clinical practice.

Exclusion criteria

The exclusion criteria are i) Studies involving pediatric populations (patients under 18 years of age); ii) non-randomized study designs, including case–control studies, cohort studies, case reports, and abstracts from conference proceedings; iii) studies where full-text access is unavailable or restricted; and iv) studies that do not incorporate the use of antiplatelet agents or anticoagulants as part of their intervention or comparison strategy.

Data sources and searches

Two authors searched PubMed, EMBASE, Web of Science, and the Cochrane Central Register of Controlled Trials from the inception dates to June 1, 2024, using the keywords (‘hip arthroplasty’ or ‘knee arthroplasty’), and ‘antiplatelet’ and ‘anticoagulant’ and (‘venous thromboembolism’ or ‘pulmonary embolism’ or ‘deep vein thrombosis’ or ‘complication’) in humans. No language restrictions were applied during the search.

Study selection

Following the elimination of duplicates, two separate researchers conducted screenings of all titles and abstracts, determining their inclusion status. Upon identifying eligible studies, they acquired full texts for additional evaluation. In instances where there was disagreement between the two researchers regarding a study’s eligibility, and a consensus could not be reached, the senior researcher made the final decision following a group discussion.

Data extraction and quality assessment

Two researchers independently extracted data using a standard form, capturing the first author, publication year, country, study type, sample size, age, VTE prophylaxis regimens, follow-up, and outcomes. Discrepancies were resolved by consensus. For trials with multiple groups or factorial designs, only relevant data were extracted. The researchers assessed the quality of RCTs using the Cochrane risk-of-bias criteria (15), which classify trials as low, high, or unclear risk based on seven factors: randomization, allocation concealment, participant and personnel blinding, outcome assessment blinding, incomplete data, selective reporting, and other biases (e.g. sponsorship or baseline differences). Trials were rated as high, moderate, or low quality based on these criteria. The GRADE approach was used to evaluate the overall evidence quality, considering limitations, inconsistency, imprecision, indirectness, and publication bias (16).

Data synthesis and analysis

The meta-analysis was conducted using Stata (v17;StataCorp, USA, 2021) and R (v4.4.0, Austria, R Core Team). Heterogeneity was assessed with the Q test and I² statistic, and a random effects model was applied to account for potential heterogeneity. Relative risks (RRs) with 95% confidence intervals (CIs) assessed count outcomes. For multiple comparisons within the same study, we followed the Cochrane Handbook’s guidance to combine groups into a single comparison (17). Predetermined characteristics, including geographic location, joint type, aspirin dosage, sex, and age, were used for heterogeneity assessment through stratified analysis. In addition, post hoc analyses considered publication year, funding status, trial quality, VTE type (symptomatic vs screening), thrombus location (proximal vs distal), and use of mechanical prophylaxis. Publication bias was examined with Begg funnel plots and Egger tests. Sensitivity analysis was performed by excluding the most significant trial.

Results

Study identification

The initial search yielded 982 potentially relevant trials (Fig. 1). After screening the titles and abstracts, 35 articles were chosen for full-text evaluation. Out of these, 15 were excluded. Ultimately, 20 RCTs (18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37) met the eligibility criteria. Among them, two studies (32, 33) originated from the same research team, with one reporting only mortality and the other focusing on VTE-related complications. Therefore, they were considered as a single study when calculating the total number of included trials. For the overall sample size, the study with the larger cohort was used. As a result, 19 studies were included in the meta-analysis (Supplementary Table 1 (see section on Supplementary materials given at the end of the article)).

Figure 1.

Flow diagram for search and selection of included studies.

Study characteristics

Among the included studies, two were conducted by Sidhu et al. (32, 33). Notably, the 2023 study (33) encompassed a larger cohort for mortality outcomes compared to their 2024 study (32). Since both studies involved the same patient population, we only extracted mortality data from the 2023 study to avoid duplication. Ultimately, our analysis included a total of 30,290 participants, with 17,501 in the antiplatelet group and 12,789 in the anticoagulant group. Nine studies exclusively evaluated patients undergoing hip (18, 19, 20, 21, 26, 29, 30, 31, 35), seven evaluated patients undergoing knee (22, 23, 25, 27, 34, 36, 37), and four evaluated patients undergoing either hip or knee (24, 28, 32, 33). Nineteen studies reported VTE (18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 34, 35, 36, 37). Seventeen studies specifically reported DVT (18, 19, 20, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 34, 35, 37), and fourteen studies specifically reported PE (18, 19, 20, 23, 24, 25, 28, 29, 30, 31, 32, 34, 35, 37).

Quality of trials

As shown in Fig. 2, thirteen trials had a low quality (18, 21, 23, 24, 27, 28, 29, 30, 31, 32, 33, 34, 35), four trials had a moderate quality (22, 26, 36, 37), and three trials had a high quality (19, 20, 25). The primary sources of bias were related to the blinding of participants and personnel (due to the difference in administration routes between LMWH administered via subcutaneous injections, and aspirin administered orally) and other biases (e.g. the study from Westrich et al. received funding from a pharmaceutical company (34)).

Figure 2.

Risk of bias assessments (18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 34, 35, 36, 37).

Primary VTE outcomes

A total of 19 studies reported VTE outcomes (18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 34, 35, 36, 37). However, the study by Hongnaparak et al. (25) was excluded from the meta-analysis because no events occurred in either group, making data pooling infeasible. After combining the results of the remaining studies (18, 19, 20, 21, 22, 23, 24, 26, 27, 28, 29, 30, 31, 32, 34, 35, 36, 37), there was no statistically significant difference in the risk of VTE between patients receiving antiplatelet therapy and those receiving anticoagulants following joint replacement surgery (RR: 1.08; 95% CI: 0.81–1.44, P = 0.149) (Fig. 3). The analysis revealed moderate heterogeneity among the included trials (I² = 63.4%; P < 0.001), which could not be attributed to any of the assessed study-level characteristics (Fig. 4). When the largest trial, which contributed the most data and utilized a non-inferiority design, was excluded (32), the pooled risk ratio remained unchanged (RR: 1.02; 95% CI: 0.77–1.34, P = 0.903), and heterogeneity decreased from moderate to low levels (I² = 40.2%; P = 0.044).

Figure 3.

Effectiveness of antiplatelets compared with anticoagulants on venous thromboembolism in patients undergoing joint arthroplasty (18, 19, 20, 21, 22, 23, 24, 26, 27, 28, 29, 30, 31, 32, 34, 35, 36, 37).

Figure 4.

Subgroup for venous thromboembolism according to study-level characteristics (18, 19, 20, 21, 22, 23, 24, 26, 27, 28, 29, 30, 31, 32, 34, 35, 36, 37).

Similarly, the pooled risk of DVT among patients treated with antiplatelet agents was not significantly different from those treated with anticoagulants across 16 trials (18, 19, 20, 22, 23, 24, 26, 27, 28, 29, 30, 31, 32, 34, 35, 37) (RR: 1.12; 95% CI: 0.83–1.51, P = 0.459, Fig. 5), with moderate heterogeneity observed (I² = 55.3%). The incidence of PE also showed no significant difference between the antiplatelet and anticoagulant groups in 11 trials (18, 19, 20, 23, 24, 28, 29, 31, 32, 34, 35) (RR: 1.26; 95% CI: 0.94–1.69, P = 0.121, Fig. 6), with no observed heterogeneity (I² = 0%).

Figure 5.

Effectiveness of antiplatelets compared with anticoagulants on deep vein thrombosis in patients undergoing joint arthroplasty (18, 19, 20, 22, 23, 24, 26, 27, 28, 29, 30, 31, 32, 34, 35, 37).

Figure 6.

Effectiveness of antiplatelets compared with anticoagulants on pulmonary embolism in patients undergoing joint arthroplasty (18, 19, 20, 23, 24, 28, 29, 31, 32, 34, 35).

Complications

Reports of complications included gastrointestinal bleeding (five studies (20, 30, 31, 34, 36)), wound hematoma (four studies (23, 29, 31, 35)), wound bleeding (three studies (20, 23, 36)), wound infection (six studies (19, 20, 22, 23, 31, 32)), other wound complications (three studies (22, 27, 30)), major bleeding (four studies (19, 20, 31, 32)), minor bleeding ((20, 23, 31)), any bleeding (one study (19)), and mortality (four studies (19, 20, 22, 33)). As shown in Fig. 7, the pooled results revealed no statistically significant differences in complications between patients using antiplatelet agents and those using anticoagulants following joint arthroplasty, with the exception of subcutaneous hematoma. Notably, while wound infection exhibited low heterogeneity (I² = 38.1%) and subcutaneous hematoma displayed moderate heterogeneity (I² = 53.6%), the other studies demonstrated homogeneity (I² < 25%). Of particular interest, subcutaneous hematoma presented a pooled RR of 0.512 (95% CI: 0.313–0.839; P = 0.008), indicating a significantly lower risk associated with antiplatelet agents compared to anticoagulants. This finding underscores the potential advantage of antiplatelet therapy in reducing the incidence of subcutaneous hematoma in patients undergoing joint arthroplasty.

Figure 7.

Summary of the pooled effect of antiplatelets compared with anticoagulants on complications in patients undergoing joint arthroplasty (19, 20, 22, 23, 27, 29, 30, 31, 32, 33, 34, 35, 36).

Subgroup analysis

The subgroup analysis revealed that, in studies conducted after 2015 and those where both groups received mechanical prophylaxis, the incidence of VTE (Fig. 4), DVT (Fig. 8), and PE (Fig. 9) was higher in the antiplatelet group compared to the anticoagulant group. In addition, in studies with moderate quality, the antiplatelet group exhibited a higher incidence of both DVT and VTE. Interestingly, female patients treated with antiplatelet therapy had a lower incidence of VTE compared to those receiving anticoagulants. For other subgroup analyses, there were no significant differences observed between the two groups.

Figure 8.

Subgroup for deep vein thrombosis according to study-level characteristics (18, 19, 20, 22, 23, 24, 26, 27, 28, 29, 30, 31, 32, 34, 35, 37).

Figure 9.

Subgroup for pulmonary embolism according to study-level characteristics (18, 19, 20, 23, 24, 28, 29, 31, 32, 34, 35).

Publication bias

In the comparisons of VTE, DVT, and PE across studies with 10 or more included trials, funnel plots indicated that data points were largely clustered near the top, suggesting smaller standard errors and relatively stable results across larger sample sizes (Fig. 10). However, the funnel plots for DVT and PE showed some data points deviating from the center line, indicating potential publication bias or other systematic errors. This was further supported by Egger’s regression test, which showed significant asymmetry for VTE (P = 0.01) and DVT (P = 0.035), suggesting the presence of publication bias. In contrast, the results for PE (P = 0.255) did not show significant bias, implying that the findings for PE may be more reliable.

Figure 10.

Funnel plot of the included studies (venous thromboembolism (A), deep vein thrombosis (B), and pulmonary embolism (C)).

GRADE ratings

We assessed the quality of evidence for pooled analyses that included 10 or more studies using the GRADE approach (Table 1). The quality of evidence for VTE and DVT was rated as moderate, while the evidence for PE was rated as high.

Table 1.

Antiplatelet compared to anticoagulant for VTE prophylaxis after joint arthroplasty.

Outcomes	Anticipated absolute effects*			Relative effect, RR (95%CI)	Number of		Certainty of evidence	GRADE†
	Risk with anticoagulant	Risk with antiplatelet	95% CI		Participants	Studies
VTE	39 per 1,000	42 per 1,000	32–56	1.08 (0.81–1.44)	18,520	18 RCTs	⨁⨁⨁◯	Moderate‡,§
DVT	30 per 1,000	33 per 1,000	25–45	1.12 (0.83–1.51)	18,266	16 RCTs	⨁⨁⨁◯	Moderate‡,||
PE	9 per 1,000	11 per 1,000	9–15	1.26 (0.94–1.69)	17,560	11 RCTs	⨁⨁⨁⨁	High‡

CI, confidence interval; RR, risk ratio.

^*The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).

^† GRADE Working Group grades of evidence – High certainty: we are very confident that the true effect lies close to that of the estimate of the effect; Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different; Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect; Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

^‡ Concerns about bias in the domains of allocation concealment.

^§ I² = 63.4%.

^|| I² = 55.3%.

Discussion

Our meta-analysis found no significant difference in the risk of VTE, including DVT and PE, between patients receiving antiplatelet agents and those treated with anticoagulants following joint arthroplasty, consistent with previous studies (38, 39). Moderate heterogeneity, particularly for VTE and DVT, may be due to variations in study design, patient populations, or intervention protocols. Subgroup analyses showed that patients treated after 2015 and those also receiving mechanical prophylaxis had higher VTE, DVT, and PE rates with antiplatelet therapy, suggesting that changing clinical practices might influence these outcomes. In addition, moderate-quality studies reported higher incidences of VTE and DVT in the antiplatelet group, highlighting a possible link between study quality and observed heterogeneity. In gender-based subgroups, female patients on antiplatelet therapy had a lower incidence of VTE, which aligns with findings that women exhibit higher platelet reactivity (40). Excluding the largest trial with a non-inferiority design reduced VTE heterogeneity while maintaining consistent overall findings, underscoring the impact of large-scale trials on meta-analytic results and the need to consider study design and quality in future research. Most included studies did not clearly report the implementation of double blinding, and given the surgical nature of the interventions, performance bias may be unavoidable. In addition, the lack of explicit reporting on blinded outcome assessment in several studies may introduce detection bias, particularly for subjective outcomes such as functional scores and pain.

While these findings offer valuable insights, the reliability of our conclusions is influenced by evidence quality and potential publication bias. Funnel plots and statistical tests showed asymmetry for VTE and DVT, suggesting possible selective publication of studies with positive results, which may overestimate intervention effectiveness. In contrast, no significant publication bias was found for PE, indicating more robust evidence for this outcome. The GRADE assessment rated the evidence for VTE and DVT as moderate, meaning that while the data are reasonably reliable, further research could change our conclusions. This aligns with the observed publication bias, indicating that caution is needed when interpreting these results. For PE, the evidence was rated as high quality, consistent with the absence of publication bias, suggesting a reliable estimate. This combination of high-quality evidence and lack of bias strengthens confidence in the conclusion that antiplatelet agents and anticoagulants are similarly effective in preventing PE after joint arthroplasty.

Our study’s findings diverge significantly from those reported by Meng et al. (41) and Zheng et al. (42). Meng et al. concentrated on the comparison between aspirin and LMWH in knee replacement patients, ultimately concluding that aspirin elevates the risk of VTE (41). Similarly, Zheng et al. analyzed the efficacy of aspirin versus oral anticoagulants across various orthopedic surgeries, determining that aspirin was less effective (42). Matharu et al. also explored the efficacy of aspirin relative to other medications for VTE prevention (38), but their study is marred by substantial shortcomings, including inaccuracies in the reported event counts and total sample sizes during subgroup analyses. In contrast, our study offers a more comprehensive and rigorous comparison of antiplatelet and anticoagulant therapies for VTE prevention in joint arthroplasty, representing the most extensive meta-analysis on this subject to date. With a total sample size of 30,290 patients, six times larger than that of Matharu et al., we thoroughly examined potential sources of heterogeneity across multiple levels, thereby providing a more robust and reliable analysis of the comparative efficacy of these therapeutic strategies. Notably, our findings suggest that aspirin’s efficacy is generally comparable to that of both LMWH and oral anticoagulants across various surgical contexts.

Aspirin may lead to gastrointestinal-related adverse reactions (43), somewhat limiting its clinical utility. Nevertheless, our study did not find that aspirin increased the incidence of related complications. The cost of treatment is a critical consideration, with the average cost of enoxaparin therapy being approximately 84 times that of aspirin treatment (23). Aspirin represents a cost-effective option, which both public and private healthcare systems may adopt to reduce costs while achieving comparable impact. Furthermore, aspirin does not require monitoring, can be used in patients with renal insufficiency, and does not increase the risk of bleeding during surgery, with reversal possible through platelet transfusion. Compared with using enoxaparin, the use of aspirin along with mechanical prophylaxis saves $2,628 per person (44).

Limitations

Our study also has limitations. Despite efforts to account for variability, moderate heterogeneity remained in the analyses for VTE and DVT, which may affect the generalizability of our findings. Potential publication bias was identified, raising concerns about the accuracy of reported effects, particularly for VTE and DVT. The exclusion of certain data and the variability in study quality across the included trials could also influence the reliability of the pooled results.

Conclusion

In conclusion, our comprehensive meta-analysis demonstrates that antiplatelet agents, particularly aspirin, offer a comparable efficacy to anticoagulants, such as LMWH and oral anticoagulants, in preventing VTE, DVT, and PE following joint arthroplasty. With a robust sample size and thorough examination of potential sources of heterogeneity, our findings provide strong evidence supporting the use of antiplatelet therapy as a cost-effective and practical option. Despite moderate heterogeneity and potential publication bias in some outcomes, the overall quality of the evidence remains reliable. These insights contribute significantly to the current understanding of VTE prophylaxis, suggesting that aspirin can be an effective and economical alternative in clinical practice, while also highlighting the need for further research to refine these conclusions and address specific complications. However, the optimal duration of prophylaxis remains variable across studies and should be individualized based on patient risk factors and current clinical guidelines.

ICMJE Statement of Interest

The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the work reported.

Funding Statement

This work did not receive any specific grant from any funding agency in the public, commercial, or not-for-profit sector.

Author contribution statement

LZ, HL, and YR were responsible for the study conception, protocol design, and overall supervision. HL and YG independently performed the literature search, study selection, data extraction, and quality assessment. YS conducted the statistical analysis and contributed to the interpretation of results. LZ and HL drafted the initial version of the manuscript. CC and YR critically revised the manuscript for important intellectual content and are the corresponding authors who oversaw the final approval of the version to be submitted. All authors read and approved the final manuscript.