Aspirin Is as Effective and Safe as Oral Anticoagulants for Venous Thromboembolism Prophylaxis After Joint Arthroplasty

Abstract

Background:

Joint arthroplasty effectively treats osteoarthritis, providing pain relief and improving function, but postoperative venous thromboembolism (VTE) remains a common complication. This study therefore assessed the effectiveness and safety of aspirin compared with oral anticoagulants (OACs) for VTE prophylaxis after joint arthroplasty.

Methods:

A systematic review and meta-analysis was performed by searching PubMed, Embase, the Web of Science, and the Cochrane Library for randomized controlled trials (RCTs) up to May 14, 2024, that compared the effect of aspirin versus OACs on VTE prophylaxis in adults undergoing joint arthroplasty. Data extraction followed the PRISMA guidelines. Two independent researchers conducted the literature searches and data extraction. A random-effects model was used to estimate effects. The primary outcome was the incidence of VTE, including deep vein thrombosis (DVT) and pulmonary embolism (PE); secondary outcomes included bleeding, wound complications, and mortality.

Results:

The meta-analysis included 11 RCTs with a total of 4,717 participants (55.1% female) from several continents. The relative risk (RR) of VTE following joint arthroplasty was 1.11 (95% confidence interval [CI], 0.93 to 1.32) for aspirin compared with OACs. Similar results were observed for DVT (RR, 1.12; 95% CI, 0.90 to 1.40) and PE (RR, 1.18; 95% CI, 0.51 to 2.71). There were no significant differences in the risks of bleeding, wound complications, or mortality between patients receiving aspirin and those receiving OACs. Subgroup analyses considering factors such as study region, type of joint surgery, type of VTE detection, year of publication, use of mechanical VTE prophylaxis, aspirin dose, type of OAC comparator, study quality, and funding also found no significant differences in VTE incidence between aspirin and OACs. The overall quality of evidence for VTE and DVT outcomes was high.

Conclusions:

Based on high-quality evidence from RCTs, aspirin is as effective and safe as OACs in preventing VTE, including DVT and PE, after joint arthroplasty, without increasing complications.

Level of Evidence:

Therapeutic Level I. See Instructions for Authors for a complete description of levels of evidence.

Joint arthroplasty is a definitive treatment for conditions such as late-stage osteoarthritis; it markedly alleviates pain, improves quality of life, and preserves joint function to the greatest extent possible. However, postoperative venous thromboembolism (VTE) is a typical complication. Reports indicate that without effective prophylactic measures, the incidence of VTE within 90 days after joint arthroplasty can be as high as 60%¹. VTE development increases treatment costs and hospital stay and may lead to potentially life-threatening embolization of vital organs².

Current literature confirms the efficacy of factor Xa inhibitors (e.g., rivaroxaban), low-molecular-weight heparin (LMWH), and warfarin in preventing postoperative VTE; however, there is no consensus on the optimal chemoprophylactic agent^3,4. Existing medications reduce the incidence of VTE; however, excessive intervention may increase the risks of postoperative bleeding, wound infection, and internal hemorrhage⁵. In addition, LMWH requires long-term subcutaneous injections, which are inconvenient for patients after discharge. Oral anticoagulants such as rivaroxaban are expensive, and although warfarin is relatively affordable, patients receiving long-term warfarin therapy require regular monitoring of the international normalized ratio (INR).

Previous studies suggest that aspirin alone, as a chemoprophylactic agent for VTE after joint arthroplasty, is not less effective than other chemical prophylactics^5,6. Similarly, Matharu et al. revealed no significant differences between aspirin and other agents in the prevention of VTE, including deep vein thrombosis (DVT) and pulmonary embolism (PE), and in the rates of bleeding, infection, and mortality compared with alternative medications⁷. Moreover, aspirin administration is markedly less expensive compared with oral anticoagulants^8,9. The American College of Chest Physicians guidelines state that aspirin can be used as a chemoprophylactic agent for VTE after hip and knee arthroplasty¹⁰, although using aspirin as the sole prophylactic agent for VTE after joint arthroplasty is relatively rare in current clinical practice. However, no meta-analysis has compared the efficacy of aspirin with that of oral anticoagulants (OACs). Therefore, we used that evidence-based medicine method to investigate the clinical effectiveness of aspirin compared with OACs in preventing VTE after joint arthroplasty.

Materials and Methods

This study followed the principles outlined in the Handbook of the Cochrane Collaboration^11,12 and the guidelines established by the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) statement¹³. The protocol for this meta-analysis was registered in PROSPERO.

Inclusion Criteria

The inclusion criteria were (1) a randomized controlled trial (RCT) evaluating the efficacy and/or safety of aspirin and OACs for preventing VTE in adults (≥18 years old) undergoing primary and/or revision hip arthroplasty and/or knee arthroplasty; (2) reporting of outcomes including VTE, DVT, PE, and/or associated complications; (3) use of a combined VTE prevention strategy including aspirin as 1 of the medications in the aspirin arm (initial treatment with LMWH or an oral anticoagulant followed by a longer course of aspirin), to reflect current clinical practice; and (4) administration of the medication for VTE prophylaxis lasting ≥14 days.

Search Strategy

Two of the authors performed the search in PubMed, Embase, the Web of Science, and the Cochrane Central Register of Controlled Trials from the inception date of each database to May 14, 2024, using the keywords (“hip arthroplasty” or “knee arthroplasty”), “aspirin” and “anticoagulant,” and (“venous thromboembolism” or “pulmonary embolism” or “deep vein thrombosis” or “complication”) in humans. Detailed search strategies are reported in the Appendix. No language restrictions were applied in the search.

Study Selection

After duplicates were removed, 2 independent researchers reviewed all titles and abstracts and evaluated the inclusion and exclusion criteria. For studies deemed potentially eligible, the full text was retrieved for additional screening. In addition, we reviewed the reference lists of all included articles to identify other potentially eligible studies. If the 2 researchers disagreed on a study’s eligibility and could not reach a consensus, the senior researcher made the final decision following a group discussion.

Data Extraction

Two researchers independently extracted relevant data from the selected trials using a standard form. The data included the first author’s name, publication year, country, study type, sample size, age, VTE prophylactic regimen, follow-up, and outcomes. Disagreements were resolved by consensus. For trials with >2 groups or with factorial designs, only the necessary information and data were extracted.

Assessment of Risk of Bias and Quality of Evidence

Two researchers independently assessed the quality of the RCTs using the Cochrane risk-of-bias criteria, grading each item as having low, high, or unclear risk^11,14. The overall quality of each trial was then categorized as low (high risk in randomization and/or allocation concealment), high (low risk in both randomization and allocation concealment, with other criteria low or unclear), or moderate (not meeting high- or low-quality criteria). The GRADE (Grading of Recommendations Assessment, Development, and Evaluation) approach was also used to assess the overall quality of evidence based on study limitations, inconsistency, imprecision, indirectness, and publication bias¹⁵.

Data Synthesis

A meta-analysis was performed using Stata (version 17; StataCorp) and R (version 4.4.0; R Core Team). Heterogeneity was evaluated using the Q test and calculation of the I² value. Given the potential heterogeneity among the trials, analyses utilized a random-effects model. The relative risk (RR) with a corresponding 95% confidence interval (CI) was used for assessing count outcomes. When dealing with multiple correlated comparisons in the same experiment, we combined groups to create a single pairwise comparison (e.g., if a study tested aspirin and multiple OACs, the OAC groups were combined), in accordance with the guidance provided by the Cochrane Handbook¹². Subgroup analyses were performed on the basis of predetermined study-level characteristics (including study region, type of joint surgery, VTE identification type, publication year, use of mechanical VTE prophylaxis, aspirin dose, type of OAC comparator, study quality, and study funding) that could have introduced heterogeneity into the results. Potential publication bias was evaluated using established methods, including Begg funnel plots and Egger regression symmetry tests. We also performed a sensitivity analysis by excluding individual studies.

Results

Study Identification

The initial search yielded 731 potentially relevant articles (Fig. 1). Following the screening of titles and abstracts, 24 articles were chosen for full-text review. Of these, 13 were excluded (see Appendix). The remaining 11 RCTs met the criteria for inclusion in the meta-analysis (Table I).

Fig. 1.

Flow diagram for the search and selection of included studies.

TABLE I.

Characteristics of the Included Studies^*

Study	Country	Participants	Duration	DVT Diagnostic Methods	Aspirin	OAC	Use of Mechanical VTE Prophylaxis	Mean Age (yr)		Females		Total No. of Subjects		Min. Follow-up	Funding	Outcomes
								Aspirin	OAC	Aspirin	OAC	Aspirin	OAC
Anderson, 201816	Canada	Primary and revision TKA and THA	2013-2016	Venous ultrasonography, lung scan, and CT pulmonary angiography	10 mg rivaroxaban qd for 5 d + 81 mg aspirin qd for 30 d (hip) or 9 d (knee)	10 mg rivaroxaban qd for 5 d + additional 30 d (hip) or 9 d (knee)	Per local policy (none, IPC, graduated stockings, or both)	62.9	62.7	903	884	1,707	1,717	90 d	NS	VTE, DVT, PE, death, myocardial infarction, stroke, wound infection, bleeding
Colleoni, 201717	Brazil	Primary TKA	NS	Venous ultrasonography	150 mg aspirin bid for 14 d	10 mg rivaroxaban qd for 14 d	NS	71.21	67.11	13	14	14	18	4 wk	NS	DVT, wound complication, systemic complication, hemoglobin and hematocrit levels
Hongnaparak, 202118	Thailand	Primary TKA	2016-2018	Venous ultrasonography or CT	300 mg aspirin qd for 14 d	10 mg rivaroxaban qd for 14 d	NS	68.15	70.5	13	17	20	20	14 d	Yes	DVT, PE, wound complication
Jatoi, 202219	Pakistan	Primary THA	NS	Venous ultrasonography	2,000 U enoxaparin qd for 5 d + 100 mg aspirin qd for 16 d	2,000 U enoxaparin qd for 5 d + 10 mg rivaroxaban qd for 16 d	NS	58.87	59.44	42	25	80	80	10 mo	NS	DVT
Jiang, 201420	China	Primary TKA	2012-2013	Venous ultrasonography	100 mg aspirin qd for 14 d	5,000 U LMWH qd for 5 d + 10 mg qd rivaroxaban for 9 d	IPC + thromboembolic deterrent stocking	65.1	63.8	55	56	60	60	6 wk	NS	DVT, VTE, blood loss, wound complication, cost
Lotke, 199621	U.S.A.	Primary and revision THA and TKA	1989-1990	Venograms and ventilation perfusion scans	325 mg aspirin bid for 6 wk	Warfarin, maintaining prothrombin time of 1.2-1.5 INR, for 6 wk	NS	66.4	67.1	100	92	166	146	6 mo	NS	DVT, PE, wound complication
Salzman, 197123	England	Hip arthroplasty	NS	Lung perfusion scan and pulmonary angiography (no venogram or fibrinogen uptake test used for DVT detection)	600 mg aspirin bid for 21-35 d	Warfarin (dosage adjusted for prothrombin time) for 21-35 d	NS	51	48	22	22	43	43	NA	NS	DVT, PE, VTE, blood loss, wound complication
Ren, 202122	China	Primary THA	2019-2020	Venous ultrasonography	100 mg aspirin bid for 5 wk	10 mg rivaroxaban qd for 5 wk	IPC	54.5	50	21	25	34	36	90 d	Yes	VTE, DVT, PE, bleeding events, blood test, HHS, complication
Woolson, 199124	U.S.A.	Primary and revision THA	1986-1989	Venography or venous ultrasonography	650 mg aspirin bid	Warfarin, maintaining prothrombin time 1.2-1.3 INR	IPC + graduated stockings	62.3	67.6	36	38	72	69	90 d	NS	DVT, PE, wound complication
Zhou, 202325	China	Primary TKA	2020-2021	Venous ultrasonography	100 mg aspirin qd for 30 d	10 mg rivaroxaban qd for 30 d	NS	66.4	64.8	34	33	60	60	90 d	NS	VTE, PE, blood loss, hematologic parameters, blood transfusion rate, bleeding complication
Zou, 201426	China	Primary TKA	2011-2013	Venous ultrasonography	100 mg aspirin qd for 14 d	10 mg rivaroxaban qd for 14 d	NS	62.7	63.5	82	70	110	102	4 wk	NS	DVT, PE, wound complication, blood loss

^*DVT = deep vein thrombosis, OAC = direct oral anticoagulant, VTE = venous thromboembolism, THA = total hip arthroplasty, TKA = total knee arthroplasty, CT = computed tomographic, qd = 4× daily, IPC = intermittent pneumatic compression, NS = not specified, PE = pulmonary embolism, bid = 2× daily, LMWH = low-molecular-weight heparin, HHS = Harris hip score.

Study Characteristics

The 11 RCTs^16-26 included 4,717 participants (2,366 in the aspirin group and 2,351 in the control group), of whom 2,597 (55.1%) were female. The mean age of the participants was 62.8 years. Nine trials were open-label^17,19-26, and the remaining 2 were double-blinded^16,18. Six studies were conducted in Asia^{18-20,22,25,26}; 4 in the Americas^16,17,21,24; and 1 in Europe²³. Four studies reported exclusively on patients undergoing hip arthroplasty^19,22-24, 5 reported exclusively on patients undergoing knee arthroplasty^{17,18,20,25,26}, and the remaining 2 included both procedures^16,21.

Study Quality

The quality of the included studies was assessed using the Cochrane risk-of-bias tool. Five studies were of low quality^20-24, 4 were of moderate quality^17,19,25,26, and 2 were of high quality^16,18. The primary sources of bias were the blinding of participants and personnel (Fig. 2).

Fig. 2.

Risk-of-bias summary: the review authors’ judgments regarding each risk-of-bias item for each included study.

Primary VTE Outcomes

In the entire cohort^16,17,19-26, the risk of VTE following joint replacement surgery was not significantly different in patients treated with aspirin compared with OACs (RR, 1.11; 95% CI, 0.93 to 1.32; p = 0.255) (Fig. 3). There was no heterogeneity among the included studies (I² = 0%, p = 0.455). After excluding the trial with the greatest weight, which contributed most of the data to the analysis and was based on a noninferiority design²¹, the pooled result remained consistent (RR, 1.25; 95% CI, 0.89 to 1.75; I² = 0%; p = 0.202). The pooled risk of DVT (9 studies^{16,17,19-24,26}) among the patients treated with aspirin was not significantly different compared with that of the OAC group (RR, 1.12; 95% CI, 0.90 to 1.40; I² = 4.4%; p = 0.299) (Fig. 4). Similarly, there was no difference in the risk of PE (3 studies^16,23,24) in the aspirin group compared with the OAC group (RR, 1.18; 95% CI, 0.51 to 2.71; I² = 0%; p = 0.700) (Fig. 4).

Fig. 3.

Effectiveness of aspirin compared with oral anticoagulants with respect to the rate of venous thromboembolism (deep vein thrombosis plus pulmonary embolism) in randomized controlled trials of patients undergoing joint arthroplasty.

Fig. 4.

Effectiveness of aspirin (ASA) compared with oral anticoagulants with respect to the rate of venous thromboembolism (VTE) in randomized clinical trials of patients undergoing joint arthroplasty, grouped according to study-level characteristics. DVT = deep vein thrombosis, PE = pulmonary embolism, RA = rivaroxaban, IPC = intermittent pneumatic compression.

Complications

Reports of complications included gastrointestinal bleeding (3 studies^22,23,25), wound hematoma (2 studies^23,24), wound bleeding (1 study²⁵), wound infection (3 studies^16,17,23), other wound complications (4 studies^17,18,20,22), major bleeding (2 studies^16,23), minor bleeding (1 study²³), any bleeding (1 study¹⁶), and mortality (2 studies^16,17). The pooled results showed no significant differences in complications between patients using aspirin and those using OACs following joint arthroplasty (Fig. 5).

Fig. 5.

Safety of aspirin compared with oral anticoagulants with respect to the rate of complications in randomized controlled trials of patients undergoing joint arthroplasty. GI = gastrointestinal.

Subgroups

We conducted subgroup analyses that considered several potential confounding variables to gain a more comprehensive understanding of our primary results. These variables included the country where the trial was conducted, type of joint surgery (knee arthroplasty, hip arthroplasty, or combined), VTE identification type (detected by routine screening or symptomatic), publication year, use of mechanical VTE prophylaxis (yes or no), aspirin dose (low or high), number of VTE events (<20 or ≥20), type of OAC comparator (rivaroxaban or warfarin), study quality (low, moderate, or high), and funding (yes or no). There was no indication that any of these study-level characteristics affected the difference in VTE between aspirin and OACs (Fig. 4).

Publication Bias

For VTE, the only comparison that included ≥10 studies, visual inspection of the Begg funnel plot revealed symmetry (Fig. 6). This finding was supported by the Egger regression symmetry test, which also indicated no significant evidence of publication bias (p = 0.598).

Fig. 6.

Funnel plot for publication bias in studies reporting VTE rates. se = standard error.

Sensitivity Analysis

The results were not significantly affected by the deletion of any of the individual studies.

GRADE Ratings

The GRADE ratings for the 2 pooled analyses that included ≥5 studies, VTE and DVT, indicated that the quality of evidence was high for both (Table II).

TABLE II.

GRADE Summary of Findings^*

Outcome	Anticipated Absolute Effect†, per 1,000		Relative Risk (95% CI)	No. of Participants (RCTs)	GRADE Certainty of the Evidence
	Risk with Direct Oral Anticoagulant	Risk with Aspirin (95% CI)
VTE	58	64 (54 to 76)	1.11 (0.93 to 1.32)	4,677 (10)	High‡
DVT	54	61 (49 to 76)	1.12 (0.90 to 1.40)	4,557 (9)	High‡

^*GRADE = GRADE Working Group grade of evidence, CI = confidence interval, RCT = randomized controlled trial, VTE = venous thromboembolism, DVT = deep vein thrombosis. High certainty: we are very confident that the true effect lies close to that of the estimate of the effect.

^†The risk in the aspirin 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).

^‡Concerns existed about bias in the domain of allocation concealment.

Discussion

The meta-analyses of the efficacy and safety of aspirin versus OACs administered to prevent VTE following joint arthroplasty included 11 RCTs. The primary outcome analysis revealed no significant difference in the incidence of VTE in the aspirin group compared with the OAC group (RR, 1.11; 95% CI, 0.93 to 1.32), and no differences in DVT and PE. Previous meta-analyses, such as that by Zheng et al., reported a higher incidence of VTE with aspirin than with OACs in patients undergoing total joint replacement or hip surgery²⁷, which likely differs from our results because the high heterogeneity of the included studies led to biased results. Our findings are consistent with those of Matharu et al. and Singjie et al.^7,28. However, Singjie et al. included patients undergoing major orthopaedic lower-limb surgery, without a subgroup analysis of joint arthroplasty²⁸. Matharu et al. included patients who underwent joint arthroplasty but compared aspirin with other oral and injectable anticoagulants⁷. The comparison of injectable medications and oral aspirin is challenging to anonymize, leading to lower study quality. In contrast, our primary analysis compared the efficacy of aspirin and OACs in patients who underwent joint arthroplasty. The comparison among oral medications facilitated blinding, resulting in higher-quality evidence. Matharu et al. did include a subgroup analysis of aspirin versus rivaroxaban (3 studies, 3,855 participants)⁷; however, our study included 11 RCTs with 4,717 patients, providing a broader range of studies and a larger sample size, and conducted a comprehensive analysis of aspirin versus OACs.

Additionally, we thoroughly assessed the methodological quality of the included studies. We conducted a rigorous analysis by systematically exploring potential sources of heterogeneity across various study-level characteristics and testing for evidence of effect modification. We also compared bleeding, wound complications, and mortality between the aspirin and OAC groups, and found no significant differences. These adverse-event findings were based on data from only a few studies, and some estimates were imprecise; therefore, these results should be interpreted with caution. Furthermore, the RCTs exhibited heterogeneity in the study populations (particularly hip and/or knee arthroplasty). However, formal subgroup analyses confirmed that study-level characteristics, including study region, type of joint surgery, VTE identification type, publication year, use of mechanical VTE prophylaxis, aspirin dose, study quality, and funding, did not influence the main findings. The quality of evidence for both VTE and DVT was rated as high.

Our results align with those of large observational cohort studies, which have shown that aspirin is effective in preventing VTE and offers efficacy and safety comparable to, or slightly better than, other common anticoagulants^6,29. Thus, the current evidence supports the use of aspirin for VTE prophylaxis after joint replacement surgery. The American College of Chest Physicians guidelines has assigned a grade of 1B to their recommendation for the use of aspirin as a postoperative anticoagulant¹⁰, and aspirin has become an essential agent for DVT prevention following joint replacement surgery. Compared with other anticoagulants, aspirin is cost-effective^8,9, does not require coagulation monitoring, and is convenient to use. Aspirin can cause gastrointestinal side effects³⁰, which have limited its clinical use. However, our study did not observe an increased incidence of complications associated with aspirin use. Furthermore, the treatment cost is a crucial factor to consider. The daily cost of 20 mg of rivaroxaban is more than 300 times that of 325 mg of aspirin³¹. Aspirin is an affordable option preferred by public and private health-care systems to reduce costs while achieving similar outcomes. It can be used in patients with renal impairment, does not increase the risk of major bleeding intraoperatively, and can be reversed with platelet transfusions.

Strengths and Limitations

Previous studies have been conducted on the use of aspirin for preventing VTE, including comparisons of LMWH with aspirin and aspirin with a placebo. However, no meta-analyses have compared aspirin with OACs. Our pooled results showed essentially no heterogeneity, with the highest I² value being 4.4%, and the conclusions were robust in all sensitivity analyses and subanalyses. The GRADE evidence level was high, indicating that it provides meaningful guidance for clinical practice. We used a thorough and rigorous search strategy encompassing multiple databases without language restrictions. By exclusively including RCTs, we eliminated the selection bias associated with observational studies. We also performed a thorough evaluation of the methodological quality of the included studies. We conducted a thorough analysis by systematically investigating the potential sources of heterogeneity across various study-level characteristics and testing for evidence of effect modification. Notably, our analyses included only patients undergoing joint arthroplasty, addressing the limitations we have noted in previous trials and systematic reviews.

Despite the high level of evidence, however, this study has some limitations, most of which are inherent to meta-analyses. First, only 2 of the included studies reported blinding of both participants and outcome assessors, 5 did not specify the blinding status of at least 1 of these, and the remaining 4 failed to blind at least 1 of these, resulting in a notable risk of bias. Second, differences in the patient inclusion and exclusion criteria used in the included studies may have led to heterogeneity in the study populations. Third, differences in drug dosage and duration, reported outcomes (such as the results of routine VTE screening versus symptomatic VTE), use of mechanical VTE prophylaxis, and follow-up times across the included studies could have resulted in biased outcomes. These factors were evaluated in the sensitivity analyses; however, they may still have resulted in biased estimates. Fourth, the reporting of adverse events appeared to be selective, as these data were reported in only a few of the included studies, potentially leading to a loss of power in demonstrating associations that did exist.

Conclusions

On the basis of existing evidence from RCTs, no significant differences were found in the clinical efficacy of aspirin compared with OACs with respect to the prevention of VTE, including DVT and PE, after joint arthroplasty. Additionally, aspirin did not increase the incidence of complications. The body of evidence was of high quality. Nonetheless, more large-scale, well-designed RCTs would be beneficial to strengthen these findings.

Appendix

Supporting material provided by the authors is posted with the online version of this article as a data supplement at jbjs.org (http://links.lww.com/JBJS/I368).