Efficacy and safety of rivaroxaban versus low-molecular-weight heparin for the prevention of symptomatic in-hospital venous thrombosis following primary total hip and knee arthroplasty: a large-sample, multicenter, retrospective study Authors: Yuanping Liao, Jie Deng, Xiaotao Shi, Xiao Wang, Honglue Tan, Guorui Cao Published: 2026 ABSTRACT AbstractPurposeTo evaluate the efficacy and safety of rivaroxaban and low molecular weight heparin (LMWH) for the prevention of postoperative venous thrombosis (VTE) following primary unilateral total hip arthroplasty (THA) and total knee arthroplasty (TKA).MethodsA total of 11,686 patients who received rivaroxaban and LMWH (6,110 cases of THA and 5,576 cases of TKA) between January 2014 and November 2017 in 26 large teaching hospitals were enrolled. The patients were divided into two groups according to the postoperative use of anticoagulant drugs, including 4,591 cases in the rivaroxaban group (2,288 cases of THA and 2,303 cases of TKA) and 7,095 cases in the LMWH group (3,822 cases of THA and 3,273 cases of TKA). The incidence of deep vein thrombosis (DVT) in the lower extremities, and pulmonary embolism (PE), were analyzed as indicators of efficacy. The incidences of postoperative bleeding events, total blood loss (TBL), and transfusion rate, were considered as safety indicators.ResultsThe overall incidence of DVT was 0.40% (47/11,686) in patients undergoing primary THA/TKA. The incidence of DVT in the rivaroxaban group was 0.74% (34/4591); this was significantly higher than that in the LMWH group (0.18%; 13/7095) (OR 4.07, 95% CI 2.14–7.71;p < 0.001). In the THA subgroup, the incidence of DVT in the rivaroxaban group was 0.52% (12/2288 cases); this was significantly higher than that in the LMWH group 0.10% (0.10%; 4/3822) (p = 0.002). In the TKA subgroup, the incidence of DVT was 0.96% (22/2303 cases) in the rivaroxaban group and 0.27% (9/3273 cases) in the LMWH group (p < 0.001). Only one patient treated with LMWH after TKA was diagnosed with symptomatic PE in this study. There was no significance difference in terms of safety between rivaroxaban and LMWH (p > 0.05) in terms of the incidence of bleeding events and TBL. However, in the TKA subgroup, TBL was significantly higher in the rivaroxaban group than in the LMWH group (915 ± 514 ml vs. 872 ± 487 ml, p = 0.030). With regards to transfusion rate, patients in the rivaroxaban group had a significantly higher need for blood transfusion than those in the LMWH group (OR 3.30,95% CI 2.88–3.78;p < 0.001). Subgroup analysis revealed that the transfusion rates in the THA and TKA sub-groups were both higher in the rivaroxaban group (14.8%, 14.2%) than those in the LMWH group (6.5%, 3.0%) (p < 0.001).ConclusionsFor patients following primary unilateral THA/TKA, LMWH is superior to rivaroxaban in terms of the prevention of symptomatic in-hospital DVT‚ and the control of TBL and transfusion rates. However, the efficacy of these drugs for controlling bleeding events was similar when compared between the two groups.Keywords: Rivaroxaban, Low molecular weight heparin, Total hip arthroplasty, Total knee arthroplasty, Symptomatic thrombosis Introduction Total hip arthroplasty (THA) and total knee arthroplasty (TKA) are important treatments for end-stage hip and knee disease and can provide significant pain relief and improve a patient’s quality-of-life [1]. However, postoperative venous thromboembolism (VTE) is a common complication that can arise from the procedures, including lower extremity deep vein thrombosis (DVT) and pulmonary embolism (PE). The occurrence of VTE can delay early recovery, is associated with an increased risk of thrombus dislodgement and increased mortality, and also occupies healthcare resources for a longer period of time, thus leading to a huge economic burden for the health system [2, 3]. This is particularly the case for patients with symptomatic thrombosis due to the patients complexity of this condition. In 2012, the American College of Chest Physicians published related guidelines that clearly stated that drugs, including low molecular weight heparin (LMWH), rivaroxaban, aspirin, and fondaparinux, are effective for VTE prophylaxis following THA/TKA (grade 1B) [4]. Another study reported that these medications resulted in sufficient prophylaxis, a significantly reduced risk of VTE, with a symptomatic VTE rate of only 4.3% [5]. However, there is a clear lack of a uniform consensus relating to optimal pharmacological prophylaxis, as various drugs remain controversial in terms of their efficacy and safety. Balancing the relationship between efficacy and safety when selecting anticoagulant drugs remains a significant challenge for clinicians. Currently, thromboprophylaxis regimens for THA and TKA are generally determined according to the experience of the treating clinician. The most commonly used drugs are LMWH and rivaroxaban. Over recent decades, LMWH has been widely used in clinical practice and can be effective in reducing the risk of postoperative thrombosis. Consequently, LMWH has become the standard anticoagulant for the clinical prevention of VTE in some hospitals [6]. However, LMWH is not easy to administer and requires subcutaneous injection, thus increasing the workload of nurses and reducing patient compliance. In addition, there is a significant lack of studies relating to the safety of LMWH during the postoperative phase. Over recent years, rivaroxaban has gained popularity in an increasing number of countries and regions, largely due to the fact that it can be managed with convenience in the clinic. Moreover, large-scale clinical studies have showed that rivaroxaban has greater benefits when compared to LMWH in terms of preventing the incidence of VTE and the safety of controlling postoperative bleeding following THA/TKA [7–10]. However, some studies also reported that rivaroxaban did not provide a greater therapeutic advantage than LMWH in the prevention of VTE. Furthermore, in terms of safety, rivaroxaban is known to increase the risk of postoperative complications, including bleeding [11, 12]. Overall, the preferable choice between these two anticoagulants remains controversial. In addition, some prospective studies reported that the manner in which LMWH is administered can affect patient compliance [11] and may induce clinical distress, thus creating uncertainty when selecting between the two thromboprophylactic drugs. Due to the lack of large, multi-center studies, the selection of LMWH or rivaroxaban for postoperative thromboprophylaxis in cases of THA/TKA remains inconclusive. At present, there is a notable lack of large data analyses comparing the efficacy and safety of these two anticoagulants following primary THA/TKA, especially in Asian regions. Previous studies mainly focused on a single center and featured a small number of samples. The present study aims to address this clear gap in the literature. The primary objective of this study was to compare the efficacy of rivaroxaban and LMWH for postoperative symptomatic in-hospital VTE in patients following primary unilateral THA/TKA. Our secondary objective was to evaluate the safety of these two drugs, including the incidence of postoperative bleeding events, total blood loss (TBL), and transfusion rate. Materials and methods Research design and objectives This retrospective cohort study was a sub-study conducted within a large prospective observational program sponsored by the Chinese Health Ministry, which attempted to investigate the efficacy and safety of hip and knee arthroplasty. We include patients who underwent primary THA and TKA between January 2014 and November 2017 in 26 large teaching hospitals. Ethical approval was obtained from the Biomedical Ethics Committee. The inclusion criteria were patients ≥ 18 years-of-age who underwent primary unilateral THA/TKA and able to provide informed consent. Patients were excluded if they had predetermined high risk factors for bleeding (such as a gastrointestinal ulcer or had suffered hemorrhagic stroke), coagulation dysfunction, hepatic or renal impairment (such as levels of aspartate aminotransferase or alanine aminotransferase that were twice the upper limit of the normal range; a creatinine clearance < 30 ml/min), or if they were on anticoagulant therapy or had incomplete medical records. Patients who underwent concurrent bilateral THA/TKA or staged bilateral THA/TKA were also excluded (Fig. 1). Process of sample screening and outcome. THA, total hip arthroplasty; TKA, total knee arthroplasty; LMWH, low-molecular-weight heparin A total of 11,686 patients were enrolled in this study, of which 6,110 underwent THA and 5,576 underwent TKA. Patients were divided into two groups according to the anticoagulant drugs they received postoperatively. Patients in the rivaroxaban group (4591 patients) were treated with oral rivaroxaban (10 mg; Xarelto, Bayer, Germany) while patients in the LMWH group (7095 patients) were treated with subcutaneous injections of LMWH (4000U; Clexane; Sanofi-Aventis, France). The application time was 6–8 h postoperatively, repeated once a day for 14 days.All patients were suggested to received physical thromboprophylaxis measures.Physical prophylaxis included the exercises of ankle pump and knee extension when awakening from anesthesia, and the application of intermittent pneumatic compression device early postoperatively. Subgroups of the two groups were analyzed according to the type of surgery (THA/TKA). In total, 2288 patients were treated with rivaroxaban and 3822 with LMWH in the THA group. In contrast, 2303 patients were treated with rivaroxaban and 3273 with LMWH in the TKA group. Assessment of efficacy outcomes The main efficacy indicators were the incidence of symptomatic in-hospital DVT and PE within 14 days. The doctors and nurses screened all enrolled patients daily for clinical signs of VTE. All patients underwent diagnostic Doppler ultrasound of both lower extremities on day 3 postoperatively. If a patient had symptoms of DVT, such as pain, swelling, and localized elevation of skin temperature in the affected extremity, then direct ultrasound or lower extremity venography were performed. Standardized compression ultrasound was used to confirm the presence of DVT. Since there is no gold standard for the diagnosis of symptomatic VTE, the diagnosis was based on clinical suspicion confirmed by venography or ultrasonography combined with impedance volume tracing [1, 13]. When PE was suspected on the basis of clinical features, such as respiration, chest pain, or tachycardia, confirmation was achieved by perfusion/ventilation imaging, pulmonary angiography, or spiral CT by a senior imaging physician. Assessment of safety outcomes The main safety indicators were the incidence of postoperative bleeding events, TBL and transfusion rate. Bleeding events that occurred postoperatively were assessed based on standard definitions. Events categorized as hemorrhage from organs and requiring surgical intervention were defined as a major bleeding event, while incisional bleeding or hematomas were defined as a minor bleeding event [1]. Total blood loss(TBL) was calculated according to the Nadler and Gross formula [14]. TBL = Patient’s blood volume (PBV) × (Hctpre-Hctpost)/ Hctave. (PBV = k1 × height (m)3 + k2 × body weight (kg) + k3 (male k1 = 0.3669, k2 = 0.03219, k3 = 0.6041; female k1 = 0.3561, k2 = 0.03308, k3 = 0.1833. Hctpre = initial preoperative Hct level, Hctpost = postoperative Hct level on the third postoperative day, and Hctave = mean of Hctpre and Hctpost). If allogeneic transfusion was performed, then the TBL was equal to the loss calculated from the change in Hct plus the volume transfuse [14]. Statistical analysis All analyses were performed using SPSS version 27.0 (SPSS Inc., USA). Continuous variables are presented as mean ± standard deviation, and categorical variables as number (percentage). The normality of continuous variables was assessed using the Shapiro–Wilk test. Between-group comparisons were performed using the independent-samples t test, Wilcoxon Mann-Whitney U test, Pearson‘s chi-square test, or Fisher’s exact test, as appropriate. All reported p-values are two-sided and were not adjusted for multiple comparisons. For key outcomes, effect estimates (e.g., odds ratios, mean differences) are reported with their 95% confidence intervals (CIs). A p-value < 0.05 was considered statistically significant. Data handling and missing values Analyses were conducted on the available case data. Data for the primary efficacy and safety outcomes (i.e., symptomatic in-hospital DVT, PE, bleeding events, and transfusion) were complete for all patients, with no missing values. For derived continuous variables (e.g., total blood loss), cases with missing prerequisite laboratory values (e.g., hematocrit) were excluded from the corresponding analysis (i.e., a complete-case analysis was applied). No imputation was performed for missing values. Results Study population and demographics Of the 11,686 patients included in this study, 4,591 patients received rivaroxaban postoperatively (THA2288, TKA2303) and 7,095 patients received LMWH postoperatively (THA3822, TKA3273). The general preoperative characteristics of the two groups were assessed and there were no significant differences between the two groups in terms of baseline characteristics (Table 1). Table 1Baseline characteristics of the patientsVariablesTHATKARivaroxban(n = 2288)LMWH(n = 3822) p valueRivaroxaban(n = 2303)LMWH(n = 3273) p valueDemographic characteristics Age, years ( x ± s)58.0 ± 13.858.3 ± 13.70.51266.2 ± 8.666.6 ± 8.70.055 Female gender (n, %)1283(56.1%)2052(53.7%)0.0701831(79.5%)2542(77.7%)0.100 BMI(x ± s,kg/m2)23.8 ± 3.423.7 ± 3.50.40125.8 ± 3.625.6 ± 3.60.072Primary joint disease0.1650.517 OA(n, %)716 (31.3%)1302 (34.1%)2132(92.3%)3048(93.1%) ONFH(n, %)721 (31.5%)1142 (29.9%)-- DDH(n, %)346 (15.1%)592 (15.5%)-- Fracture(n, %)313 (13.7%)499 (13.1%)-- AS(n, %)43 (1.9%)52 (1.4%)-- RA(n, %)61 (2.7%)83 (2.2%)157 (6.8%)201 (6.1%) Others(n, %)88 (3.8%)152 (4.0%)14 (0.6%)24 (0.7%)Comorbidities(n,%) Hypertension305 (13.3%)502 (13.1%)0.827615(26.7%)800 (24.4%)0.056 Diabetes93 (4.1%)172 (4.5%)0.419202 (8.8%)242 (7.4%)0.061 Coronary heart disease28 (1.2%)54 (1.4%)0.53482 (3.6%)95 (2.9%)0.168Surgical site(n,%)0.2930.758 Left1117(48.8%)1919(50.2%)1139(49.5%)1605(49.0%) Right1171(51.2)1903(49.8%)1164(50.5%)1668(51%)THA total hip arthroplasty, TKA total knee arthroplasty, LMWH low-molecular-weight heparin, BMI body mass index, OA osteoarthritis, ONFH osteonecrosis of femoral head, DDH developmental dysplasia of the hip, AS ankylosing spondylitis, RA rheumatoid arthritis Baseline characteristics of the patients THA total hip arthroplasty, TKA total knee arthroplasty, LMWH low-molecular-weight heparin, BMI body mass index, OA osteoarthritis, ONFH osteonecrosis of femoral head, DDH developmental dysplasia of the hip, AS ankylosing spondylitis, RA rheumatoid arthritis Efficacy outcomes DVT rates varied among patients receiving different patterns of pharmacological thromboprophylaxis (Fig. 2). The overall incidence of DVT in the lower extremities after primary THA/TKA was 0.40% (47/11686); 0.74% (34/4591) in the rivaroxaban group compared to 0.18% (13/7095) in the LMWH group (OR = 4.07, 95% CI: 2.14 to 7.71; p < 0.001; Table 2). THA subgroup analysis showed that the incidence of DVT was 0.52% (12/2288) in the rivaroxaban group and 0.10% (4/3822) in the LMWH group (OR = 5.03, 95% CI: 1.62 to 15.62; p = 0.002). TKA subgroup analysis showed that the incidence of DVT was 0.96% (22/2303) in the rivaroxaban group and 0.27% (9/3273) in the LMWH group (OR = 3.50, 95% CI: 1.61 to 7.61; p < 0.001)). One patient in the LMWH group developed symptomatic PE and underwent TKA (Table 3). Rivaroxaban vs. LMWH: DVT & Bleeding Incident (Total/THA/TKA Subgroups). THA, total hip arthroplasty; TKA, total knee arthroplasty; LMWH, low-molecular-weight heparin. LMWH: low-molecular-weight heparin; DVT, deep vein thrombosis; THA, total hip arthroplasty; TKA, total knee arthroplasty; OR, odds ratio; CI, confidence interval; ***p<0.001, **p<0.01 Table 2Primary efficacy and safety outcomesVariablesRivaroxaban(n = 4591)LMWH(n = 7095)χ2/F值p valueEffect Size (95% CI)DVT, no. (%)34(0.74%)13(0.18%)21.6<0.001*OR = 4.07 (2.14–7.71)PE, no.-1--Bleeding incident9(0.20%)16(0.23%)0.1130.736OR = 0.87 (0.38–1.97)Major bleeding event670.2570.612Minor bleeding event390.5160.473TBL(ml)987 ± 556971 ± 5431.6970.317MD = 15.1 (-14.4–44.6)Transfusion rate, no. (%)666(14.5%)347(4.9%)325<0.001*OR = 3.30 (2.88–3.78)LMWH low-molecular-weight heparin, VTE venous thrombosis, DVT deep venous thrombosi, PE pulmonary embolism TBL total blood loss Primary efficacy and safety outcomes LMWH low-molecular-weight heparin, VTE venous thrombosis, DVT deep venous thrombosi, PE pulmonary embolism TBL total blood loss Table 3Comparison of the efficacy of anticoagulants under THA and TKAVariablesTHATKARivaroxban(n = 2288)LMWH(n = 3822) p valueEffect Size (95% CI)Rivaroxaban(n = 2303)LMWH(n = 3273) p valueEffect Size(95% CI)Incidence of VTEDVT, no.(%)12(0.52%)4(0.10%)0.002*OR = 5.03(1.62–15.62)22 (0.96%)9(0.27%)<0.001*OR = 3.50 (1.61–7.61)PE, no.----1-THA total hip arthroplasty, TKA total knee arthroplasty, LMWH low-molecular-weight heparin, VTE venous thrombosis, DVT deep venous thrombosis, PE pulmonary embolism Comparison of the efficacy of anticoagulants under THA and TKA THA total hip arthroplasty, TKA total knee arthroplasty, LMWH low-molecular-weight heparin, VTE venous thrombosis, DVT deep venous thrombosis, PE pulmonary embolism Safety outcomes The overall incidence of postoperative bleeding events in this study was 0.21% (25/11686); there was no significant difference between the two groups(OR = 0.87, 95% CI: 0.38 to 1.97; p = 0.736, Fig. 2)). TBL was higher in the rivaroxaban group than in the LMWH group, although this was not significant (MD = 15.1 ml, 95% CI: -14.4 to 44.6 ml; p = 0.317). TKA subgroup analysis showed that TBL was significantly higher in the rivaroxaban group (915 ml ± 514 ml) than in the LMWH group (872 ml ± 487 ml)(MD = 42.6 ml, 95% CI: 4.0 to 81.2 ml; p = 0.030); however, this was not significant in the THA group (MD = 33.7 ml, 95% CI: -10.6 to 78.0 ml; p = 0.145). With regards to the transfusion rate, 14.5% and 4.9% of patients in the rivaroxaban group and LMWH group required a blood transfusion, respectively (OR = 3.30, 95% CI: 2.88 to 3.78; p < 0.001)). Subgroup analysis showed that the transfusion rate for both THA and TKA was higher in the rivaroxaban group (14.8%, 14.2%) than in the LMWH group (6.5%, 3.0%) (OR = 2.50, 95% CI: 2.10 to 2.97; p < 0.001;OR = 5.32, 95% CI: 4.22 to 6.71; p < 0.001)). These results are summarized in Tables 2 and 4. Table 4Comparison of the safety of anticoagulants under THA and TKATHATKARivaroxban(n = 2288)LMWH(n = 3822)P valueEffect Size (95% CI)Rivaroxaban(n = 2303)LMWH(n = 3273)P valueEffect Size (95% CI)TBL1077 ± 5931044 ± 5700.140MD = 33.7(-10.6-78.0)ml915 ± 514872 ± 4870.030*MD = 42.6 (4.0 -81.2)transfusion rate338(14.8%)248(6.5%)<0.001*OR = 2.50 (2.10–2.97)328(14.2%)99(3.0%)<0.001*OR = 5.32 (4.22–6.71)Bleeding incident5(0.21%)7(0.18%)0.9974(0.17%)9(0.27%)0.440Major bleeding320.562351.000Minor bleeding250.925140.608THA total hip arthroplasty, TKA total knee arthroplasty, LMWH low-molecular-weight heparin, TBL total blood loss, PBV parenchymal blood volume Comparison of the safety of anticoagulants under THA and TKA THA total hip arthroplasty, TKA total knee arthroplasty, LMWH low-molecular-weight heparin, TBL total blood loss, PBV parenchymal blood volume Sensitivity analysis To quantify the robustness of our primary findings to potential unmeasured confounding, we calculated E-values. For the significantly increased risk of symptomatic in-hospital DVT associated with rivaroxaban (OR = 4.07, 95% CI: 2.14 to 7.71), the E-value was 7.3 for the point estimate, and 3.3 for the lower confidence limit. Regarding the increased risk of transfusion (OR = 3.30, 95% CI: 2.88 to 3.78), the corresponding E-values were 5.8 and 5.1, respectively. In the TKA subgroup, the markedly higher transfusion risk (OR = 5.32, 95% CI: 4.22 to 6.71) had an E-value of 9.0 for the point estimate and 7.1 for the lower CI limit. These results indicate that an unmeasured confounder would need to be strongly associated with both the choice of anticoagulant and the outcome, with risk ratios ranging from 3.3 to 9.0, to fully explain away these observed associations, suggesting that the findings are robust to potential confounding. Discussion Lower extremity arthroplasty, predominantly THA and TKA, carries the risk of postoperative VTE. However, the selection of postoperative medications to prevent VTE remains a key postoperative issue. Although the American College of Chest Physicians guidelines help clinicians with formulating preventive strategies, the best options following primary THA/TKA still remains limited. In addition, due to the lack of large national studies, it has become a major challenge for orthopedic surgeons to select drugs for the prevention of VTE and avoid associated postoperative complications. Our study was a large study involving multiple teaching hospitals across the country, and used a large sample size to directly compare the differences in efficacy and safety between rivaroxaban and LMWH for the prevention of VTE in primary THA/TKA. We concluded that the incidence of postoperative VTE following arthroplasty under standardized prophylaxis was very low (< 1%), and that LMWH was more effective than rivaroxaban in preventing symptomatic in-hospital VTE, with rivaroxaban use associated with a more than four-fold increase in the odds of DVT (OR = 4.07, 95% CI: 2.14 to 7.71).This result differed from previous studies in which several large trials reported that the incidence of VTE in 5,077 patients treated with rivaroxaban after hip and knee arthroplasty was 3.3%. This was lower than the 6.1% observed in the LMWH group of patients, thus demonstrating that the efficacy of rivaroxaban for preventing postoperative VTE in patients with THA/TKA was superior to that of LMWH across the board [7, 8]. However, the most important finding of our current study was that LMWH was superior to rivaroxaban in controlling the incidence of symptomatic in-hospital DVT. This was supported by three factors: First, in a large multicenter retrospective study, Perez et al. evaluated VTE prevention in 72,670 arthroplasty patients and found that LMWH was more advantageous in preventing the incidence of postoperative VTE when compared to rivaroxaban [15], a finding that was consistent with our present study. Furthermore, several previous studies showed that LMWH as effective as rivaroxaban in preventing VTE [16, 17].This suggests that rivaroxaban is not as absolutely superior to low molecular heparin sodium as has been described in the literature. Thus, our findings were justified, although controversy remains. Second, the context of our present study was different from that of previous studies. We only analyzed the efficacy of the two drugs in preventing symptomatic in-hospital VTE under primary unilateral arthroplasty, and excluded patients with simultaneous bilateral, revision, and asymptomatic VTE. Furthermore, we only monitored thromboprophylaxis over a 14-day period in Asian patients. The application of a long follow-up visit can be prone to the loss of data. Previous studies, which included asymptomatic VTE patients in their analysis, did not consider the incidence of patients with symptomatic VTE as the primary efficacy comparison. In addition, these studies chose to consider venous thrombosis for 30 days, or even longer, after pharmacological prophylaxis [2, 7, 8, 10, 18–20]. However, due to the different nature of the two drugs, patients with subcutaneous LMWH may have difficulty completing the entire anticoagulation regimen, this may result in a higher rate of blood clots in patients treated with LMWH and lead to a possible clinical underestimation of the efficacy of LMWH. Finally, the low incidence of VTE could have compromised the true efficacy of thromboprophylaxis. The results of a previous high-quality randomized clinical trial and observational study showed that the incidence of symptomatic DVT after THA and TKA was only 0.26% and 0.63%, respectively [13]. The incidence of symptomatic VTE in the present study was also only 0.40%, thus suggesting that VTE was uncommon. This made it difficult for the previous study to obtain sufficient statistics to identify a difference between the two drugs. In the present study, on the other hand, there were more patients using LMWH, both in the whole cohort and in the subgroup analysis, and although this was a choice of prevention taken by the physicians based on their clinical judgment, the results were more in favor of the LMWH group. However, our data did show that LMWH is more favored by clinicians in Asia. Further, high-quality studies are now needed to confirm the true efficacy of these two drugs. Postoperative bleeding events were the main safety indictors in the present study. We found no significant difference between the two drug groups (both in terms of major and minor bleeding, OR = 0.87, 95% CI: 0.38 to 1.97). This were consistent with previous studies [18, 21]. This finding was also consistent with the results of the four RECORD trials [7–10]. There were also some inconsistencies between the reported bleeding events of some earlier studies and the present study. However, these earlier studies involved small and single-center investigations with limited clinical statistical evidence [16]. In addition to our statistics relating to the incidence of bleeding events, we also used TBL and transfusion rates to evaluate the safety of these medications. Our study showed that TBL was higher with rivaroxaban than that with LMWH, and this difference was significant in the TKA subgroup (mean difference = 42.6 ml, 95% CI: 4.0 to 81.2 ml). The rational explanation for this result is that LMWH can effectively inhibit the activation of thrombin and ensure the normal flow of blood, thus reducing postoperative blood loss without increasing the risk of bleeding [22]. A previous study reported that rivaroxaban may induce increased levels of bleeding in drug prophylaxis [23]. However, there is a clear lack of high-quality studies comparing TBL between these two drugs. Excessive postoperative blood loss is known to increase the surgical risk of arthroplasty and the incidence of complications [24]; therefore, we performed our analysis to provide reference guidelines for doctors and patients. Because rivaroxaban is associated with a higher risk of bleeding, there is also a significantly greater need for transfusion when using it (Total OR = 3.30,95% CI: 2.88 to 3.78;TKA, OR = 5.32, 95% CI: 4.22 to 6.71). A previous meta-analysis that included eight studies and a total of 13,384 patients, concluded that there was a higher transfusion requirement with rivaroxaban compared to other anticoagulants(p = 0.03) [18].It is also worth noting that the high rate of blood transfusion in our study was due to the lack of a consensus on standardized blood management in China at that time. We believe that there may be a greater advantage when utilizing LMWH in the postoperative management of patients with bleeding tendencies because pharmacopeial monographs provide specifications to ensure the efficacy of LMWH in relation to blood loss [25]. However, rivaroxaban lacks an effective clinical reversal agent in the event of hemorrhage. This study also had some limitations that need to be considered.Although residual confounding from unmeasured factors cannot be ruled out in an observational study, a sensitivity analysis using E-values suggests our primary conclusions are robust. To fully explain away the increased risk of symptomatic DVT associated with rivaroxaban (OR = 4.07), an unmeasured confounder would need to be associated with both treatment allocation and the outcome by risk ratios of 7.3-fold each. This indicates that our key findings are unlikely to be negated by plausible levels of unmeasured confounding. First, because of the inherent limitations of data acquired from a large sample, multicenter, and nationwide study, there was a clear lack of clinically relevant data, including surgical procedures, implant information, and postoperative medications for rehabilitation. In addition, although standard inclusion and exclusion criteria were used, we could not adjust for all confounding factors. Second, We focused only on the incidence of in-hospital venous thromboembolism (VTE) within 14 days postoperatively, as this is the primary period of concern for VTE prevention. However, we overlooked VTE events occurring after discharge (out-of-hospital VTE), which may contribute to a higher overall incidence. Future studies will extend the follow-up period and conduct comprehensive data collection to gain a more thorough understanding of VTE incidence in both in-hospital and out-of-hospital settings. Third, we only studied the efficacy of rivaroxaban and LMWH in preventing postoperative symptomatic in-hospital VTE. Our findings may not be applicable to other thromboprophylactic agents because of the different mechanisms by which these agents prevent VTE. Our current study determined the precise incidence of symptomatic in-hospital VTE after THA and TKA with rivaroxaban and LMWH. To our knowledge, this represents the first multicenter and large sample study in Asia to directly compare the efficacy and safety of rivaroxaban with LMWH in the prevention of symptomatic in-hospital VTE following primary unilateral arthroplasty. We utilized a more standard diagnostic approach and the fact that our patients were recruited in-hospital ensured the authenticity of the data. In the present study, the incidence of symptomatic in-hospital VTE was higher for rivaroxaban. In this regard, we cautiously recommend LMWH as the drug choice for thromboprophylaxis, although this needs to be confirmed by high-quality prospective studies in the future. The choice of prophylactic drugs still remains controversial, and there is a clear need to improve the thrombotic risk stratification system to understand how we might better decide which type of venous thromboembolic drugs are most appropriate. Recommending an individualized prophylactic regimen that balances efficacy and safety may be the way forward. Conclusion In this large-scale, multicenter retrospective study of Asian patients undergoing primary THA or TKA, LMWH demonstrated a superior risk-benefit profile compared to rivaroxaban for in-hospital thromboprophylaxis. Specifically, LMWH was significantly more effective in preventing symptomatic in-hospital DVT, with rivaroxaban associated with a > 4-fold increased odds (OR = 4.07, 95% CI: 2.14 to 7.71). Regarding safety, while major bleeding rates were comparable, the use of rivaroxaban was linked to a significantly higher transfusion requirement (OR = 3.30, 95% CI: 2.88 to 3.78), a risk that was especially pronounced following TKA (OR = 5.32, 95% CI: 4.22 to 6.71). The differential impact on blood loss between THA and TKA subgroups further suggests that the type of arthroplasty is an important consideration in anticoagulant selection. These findings, supported by sensitivity analyses for unmeasured confounding, provide robust real-world evidence that challenges the assumed universal superiority of direct oral anticoagulants in this setting. For perioperative thromboprophylaxis in primary joint arthroplasty—particularly in TKA patients—LMWH appears to offer a more favorable balance of efficacy (preventing VTE) and safety (conserving blood). We therefore recommend that LMWH be considered the preferred agent in clinical protocols for this population, pending confirmation from prospective, randomized trials designed to assess patient-important outcomes like transfusion and clinically relevant blood loss. Acknowledgments Statement on the Use of Generative AI and AI-Assisted Technologies During the preparation and revision of this work, the authors did not use any generative artificial intelligence (AI) or AI-assisted technologies in the writing process. All aspects of research conception, study design, data analysis, interpretation of results, and manuscript composition were conducted solely by the authors, who take full responsibility for the entire content of this publication. Authors’ contributions Guorui Cao and Honglue Tan contributed to the design of the study. Yuanping Liao and Guorui Cao were responsible for data collection and supervision of the study. Xiaotao Shi, Xiao Wang and Jie Deng performed statistical analysis. Yuanping Liao wrote the manuscript. The author s read and approved the final manuscript. Funding This research was funded by Scientific and technological research in Henan province (242102310025), Post-doctoral research funding project in Henan .Province (365154) and Start-up funds for doctoral research (2024BS02). Data availability The datasets generated and/or analyzed during the current study are not publicly available due data sharing was not included in the informed consent process, but are available from the corresponding author on reasonable request. Declarations Ethics approval and consent to participate Ethical approval was obtained from the Clinical Trials and Biomedical Ethics Committee IRB (2012-268) of West China Hospital(Sichuan University).All the participants was provided written informed consent Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Honglue Tan and Guorui Cao contributed equally to this work. Contributor Information Honglue Tan, Email: hnlc.love@163.com. Guorui Cao, Email: 13688172272@163.com. References Associated Data Data Availability Statement The datasets generated and/or analyzed during the current study are not publicly available due data sharing was not included in the informed consent process, but are available from the corresponding author on reasonable request.