Author：Anhua Long • Lihai Zhang • Yingze Zhang •Baoguo Jiang • Zhi Mao • Hongda Li •Shanbao Zhang • Zongyan Xie • Peifu Tang

Abstract Thromboprophylaxis with rivaroxaban has proved effective and safe in patients undergoing hip and knee replacement surgery. As it is unclear whether it is also effective and safe in fracture patients,the aim of the present study was to evaluate the efficacy and safety of rivaroxaban in patients with lower limb fractures.We peformed a retrospective cohort study of 2,050 consecutive patients treated for lower limb fractures at our trauma center, comparing rates of venous thromboembolism (VTE), bleeding and surgical complications, and the length of hospital stay for 608 patients who received rivaroxaban and 717 who received a low-molecular-weight heparin (LMWH). Rates of symptomatic VTE were 4.9 and 8.6 % in the rivaroxaban and LMWH groups, respectively (p = 0.008), and distal VTE rates were 1.8 and 5.7 %, respectively (p = 0.036). The incidence of major bleeding events in the rivaroxaban group was also lower than in the LMWH group (0.2 vs 0.6 %),but the difference between the groups was not statistically significant. The mean length of hospital stay was significantly shorter in the rivaroxaban group (12.2 vs 13.1 days,respectively; p = 0.016). This retrospective cohort study is thefirst report documenting the the first report documenting the efficacy and safety of rivaroxaban in patients with lower extremity fractures.In comparison with LMWH, rivaroxaban reduced the incidence of VTE by 45 % without increasing the risk of bleeding. However,prospective, randomized controlled trials comparing rivaroxaban and LMWH are needed to confirm our findings.

Keywords Deep vein thrombosis,Venous thromboembolism,Major bleeding events,Thromboprophylaxis,Rivaroxaban,Low-molecularweight heparin

Introduction

The risk of venous thromboembolism (VTE)is extremely high inpatients withtrauma.Venography studies have shown that in patients with trauma who do not receive prophylactic anticoagulanttherapy, the incidenceof VTE is ashigh as 58% .Although it has been reported that the incidence of VTE in Asian populations is lower than in European and US populations,a recent mena-analysis of 22 studies showed that the incidence of postoperative VTE in Asian patients who underwent major orthopedic surgery was as high as 31.7%.Consequently,current guidelines all recommend that thrombosis prevention should be routinely performed in trauma patients at high risk of thrombosis,and that chemoprophylaxis is the most effective way to prevent VTE. Low-molecular-weightheparins (LMWHs)have been recommended as the first choice therapy for prevention of thrombosis in patients undergoing major orthopedic surgery for the past two decades. Recently,the oraldirect factor Xainhibitor rivaroxaban was varoxaban was developed and has been foud be a safe,effective andconvenient method to prevent VTE.Four large-scale,phaseⅢclinical trials all showed that rivaroxaban has superior efficacyto enoxaparinin reducing the incidence of VTE after hip and knee replacement surgery.In comparison with enoxaparin,rivaroxaban significantly reduced the  incidence of  VTE (RR 0.41; 95% CI 0.20–0.83) without increasing the srisk of bleeding (RR 1.73; 95% CI 0.94–3.17). Rivaroxaban is currently recommended for thromboprophylaxis after hip and knee replacement surgery,for which it is started at 6–10h after surgery and administered continuously until up to 35 days after surgery.

Thus far,all phase III clinical trials of rivaroxaban have been conducted in patients undergoing hip and knee replacement and none has been carried out in trauma patients. Evidence-based medicine studies of rivaroxaban, such as efficacy and safety trials,have not been performed in trauma patients,especially those with lower extremity  fractures.Consequently,there is no stronge vidence supporting the use of rivaroxaban in these patients.

This study summarizes our clinical experience with rivaroxaban thromboprophylaxis inpatients with lower extremity fractures treated at the Department of Orthopedics andTrauma atourhospital. As part ofthe study, a comparison of the efficacy and safety of LMWH and rivaroxaban for thromboprophylaxis  was undertaken. The study protocol was approved by the Ethics Committee of Chinese PLA  General  Hospital.  Due to its retrospective nature and the fact that the patient data were anonymous, informed consent was not requested from patients.

Methods

Patients

The studywas a retrospective cohort investigation. Data were obtained on patients with lower extremity fractures who were admitted to the Department of Orthopedics and Trauma at the PLA General  Hospital between January 2009 and December 2012. Eligible patients were thoseaged 18 years or older who had  lower extremity fractures,including proximal femoral fractures, femoral shaft fractures, distal femur fractures,patella fractures,proximal tibial fractures,tibia and fibula fractures,distal tibial fractures,and ankleand foot fractures.Patients who were admitted but did not undergo surgery for various reasons or who werediagnosed withany type ofVTE at admission were excluded fromthe study.

Diagnosis

After admission to hospital,laboratory investigations such as routine blood tests, biochemistry and coagulation tests, and compression ultrasound (CUS) were performed. Laboratory parameters were routinely re-measuredat Day 1, Day3 and Day7 postoperatively.Lower extremity vascular ultrasound or angiography examinations  were repeated inpatients who suffered from symptoms ofsuspected deep vein thrombosis (DVT) such as lower limb swelling,local tenderness and  unexplained fever.A diagnosis of DVT was made based on the results of ultrasound and angiography. Lung ventilation/perfusion scanning or spiral CT was performed in patients with chest pain,berathing difficulties and other symptoms of suspected pulmonary  embolism (PE).  A diagnosis of PE was based on the results of the examinations.  

Thromboprophylaxis regimens

All patients admitted to hospital received routine drug and/or physical preventive measures according to current guidelines.From January 2009 to October 2010,a LMWH (nadroparin;Fraxiparine,GlaxoSmithKline) was used as the first choice for thromboprophylaxis and was administered  subcutaneously once daily at a dose volume of 0.2–0.4 mL from the evening of the day the patients were admitted to hospital to the evening of the day before surgery.Subsequently, the LMWH regimen wa sresumed at least 8h after surgeryand continued until up to 35 days after surgery.

Since November 2010,  oral  administration  of  10mg rivaroxaban once daily has been  recommended in patients with lower extremity  fractures at our hospital.   Rivaroxaban was administered from the evening of the day the patients were admitted to hospital to the evening of the day before surgery. Thereafter  it was resumed at least 8h after surgery and continued until up to 35days after surgery.

Both agents  were  used  during  the  period  November2010 to December 2012. Therefore,two large  cohorts of unselected, consecutive patients receiving   thrombopro phylaxis with either LMWH or  rivaroxaban were investigated in our study. Based on the different regimens for prevention of thrombosis,  the patients were divided into three groups: rivaroxaban group, LMWH group, and a non- drug prevention group.

Efficacyand safety outcomes

The primary efficacy outcome was all VTE events that occurred inpatients within 3months after surgery.Secondary outcomes included the incidences of DVT,PE, and proximal thrombosis (thrombosis in the popliteal veinand above), and the incidence of single distal embolization within 3months after surgery.

The primary safety outcome was the incidence of major bleeding events occurring from the day thromboprophylaxis was commenced to 2 days after with drawal of the drug. Major bleeding events were defined as fatal bleeding,bleeding in criticalsites such as retroperitoneal,intracranial and spinal cord bleeding,bleeding leading toreoperation, significant bleedingin extra-surgical  sites leading to a 20g/Lor greater fall in hemoglobin, or a transfusion requirement of more than  two units of whole blood/ red blood cells.The secondary safety outcome was the incidence of minor bleeding events,which were defined as incision hematomas,gastrointestinal bleeding,and other bleedingevents that were not major events  and required clinical treatment.

Other outcomes evaluated included the average duration of hospitalization, and the transfusion of more or less than three units of blood during hospitalization.

Statisticalanalysis

Student’s t test was used toanalyze continuous data for comparisons of thebaseline characteristics of the  various patients groups. Pearson’s Chi  square test,Fisher’s exact test andanalysis ofvariance (ANOVA)were used for inter-group or inner-group comparisons of constituent ratios. Logistic regression analysis was used for 2-category data.Multivariate analysis was performed by the stepwise maximum likelihood method, and95 % confidence intervals(CIs) and odds ratio (OR) values were used to describe the hazard ratio.A p valueless than 0.05was considered statistically significant. All statisticalprocedures  were accomplishedwith SPSS 19.0.

Results

From January 2009 to December 2012, 2,050 patientsmet the inclusion criteria; 608 patients (29.7%) received rivaroxaban as the primary drug for prevention  of thrombosis,717 (35.0 %) received LMWH, and 725 (35.3 %) did not receive any drugs. All patients were followed until 3 months postoperatively,except for  106  (5.2 %)  with whom contact waslost during the follow-up period.Although the medianage of thepatients and the numbers with multiple fractures were lower  in  the rivaroxaban group than in the LMWH group, and these differences and the types of anesthesia and surgery performed  were significantly different between the two groups,there were no significant differences between the groups in terms of gender,body mass index(BMI),number of cases with open fractures,VTE history, fracture type, and ASA score (Table 1). However, the  mean  duration  of thrombopro-phylaxiswas significantly longer in the rivaroxaban group (15.5 days,95% CI 14.29–16.71vs 10.8 days,95 % CI  10.08–11.44;p<0.001; Table1).

Table 1  Patients characteristics and type of VTE prophylaxis in all patientsadmitted for lower extremity fracturesbetween January 2009 andDecember 2012

Efficacy outcomes

The incidences of all VTEs in the rivaroxaban and LMWH groups were4.9 % (30/608)and 8.6 %(62/717), respectively, and the difference between the groups was statistically significant (p=0.008).No patient in the rivaroxaban group experienced PE,but twopatients (0.3 %) in LMWH group had postoperative PE;however, the difference between the two groups for this endpoint was not statistically significant (p=0.192).The mediantimes for diagnosis of VTE after surgery were 18 days in rivaroxaban group and 21 days in LMWH group;this difference was not statistically significant(p=0.432;Table2).

The incidence of DVT in the rivaroxaban group was significantly lower than in LMWH group (4.9vs 8.6 %; p=0.008). Although the  incidences of proximal thrombosis in the two groups were not significantly different (0.9 vs 2.9 %; p=0.123),the incidences of single distal venous thrombosis were significantly different  (1.8 vs5.7 %;p =0.036) suggesting that rivaroxaban reduced the incidence of DVT mainly by reducing the occurrence of single distal venous thrombosis(Table 2).

Safety outcomes

The incidences of both major andminor bleeding events in the two groups were not significantly different. Only one case of major bleeding occurred in the rivaroxaban group and this patient died of gastrointestinal bleeding at Day 7 postoperatively. Inthe LMWH group, four patients developed major bleeding events that resulted in a reduction of hemoglobin of 20 g/L or more,all whom recovered after symptomatictreatment such as red blood cell transfusions. In the rivaroxaban group, five patients suffered from postoperative wound hematomas and one suffered from gastrointestinal bleeding. In comparison,there were12 casesof postoperative wound hematomasand three of gastrointestinal bleedingin the LMWH group. However, there were no significant differences in the incidence of either major bleeding events or minorbleeding events between the two groups (Table2).

Of the patients in the rivaroxaban group, 144(23.7 %) received transfusions of more than three unitsof red blood cells during the perioperative period,as compared with 172 (24.0 %)in the LMWH group, but this difference was not statistically significant (p=0.897). The average length of hospital stays in the rivaroxaban and LMWH groups were12.2 and13.1 days,respectively,and this difference was statistically significant (p=0.016) (Table2).

Univariateand multivariate analyses

Univariateanalysis of the risk of VTE showed that advancedage, female sex, a prior history of VTE,fractures of the proximal hip,ASA scores of grade III-IV, the use of general anesthesia,and intramedullary fixation were all risk factors for VTE,and that treatment with rivaroxaban could significantly reduce the risk of VTE (OR0.55, 95 % CI 0.35–0.86; p= 0.009).However, multivariate analysis showed that only the proximal hip joint fracture site was an independent  risk  factor  for  VTE  (OR 1.19, 95 % CI 1.04–1.35;p =0.008),the risk of which could besignificantly reduced by rivaroxaban treatmentnt(OR 0.56, 95% CI 0.34–0.91;p =0.009) (Table 3).

Discussion

The results of this retrospective cohort study in patients with lower extremity fractures indicate that rivaroxaban reduced the incidence of VTE by 45% in comparison with LMWH the rapy.These results are consistent with the results of 2 phase III clinical trials ofpatients who underwent total hipor knee replacement.In these trials, comparisons of rivaroxaban and enoxaparin showed that after knee replacement,rivaroxaban was associated with a more than 50%  reduction in the incidence of VTE.Currently, there are no large-scale,comparative studies of rivaroxaban and LMWHs in trauma patients.Therefore, we can only horizontally compare the results of our study with the results of studies in patients who underwent total hip or  knee replacement.In our study, rivaroxaban reduced the incidence of VTE versus LMWH mainly by reducing the incidence of singledistal venous thrombosis,as the incidences of proximal vein thrombosisin the rivaroxaban and LMWH groups were not statistically different.Since single distal venous thrombosis has a significantly lower risk of PE than proximal vein thrombosis,the   benefit for patients of lowering    the incidence of single distal venous  thrombosis is currently controversial.

In the multivariate regression analysis, factors such as age, gender, BMI,prior VTE history, fracture site,ASA classification, anesthesia and  surgery type, and the drug used for thromboprophylaxis were comprehensively analyzed. The results showed that rivaroxaban significantly reduced the incidence of VTE versus LMWH with a risk ratio of 0.56 (95%CI 0.34–0.91;p=0.020), and in this regard, our results were similar to those of the studies  conductedin patients  who underwent  total hipand knee replacement.While our retrospective cohort study compared the efficacy and safety of rivaroxaban with LMWH inpatients with lower extremity fractures, further prospective, large-scale,multicenter, randomized, controlled studies will be needed to provide ahigher level evidence.

In our study, the incidences of VTE inpatients treated with rivaroxaban and LMWH for thromboprophylaxis were significantly higher than in patients who did not  receive any drugfor prophylaxis(7.1 vs 1.1%; p = 0.000).This reason for this is that therisk of VTE varies between patients with lower extremity fractures,and factors such as age, gender, surgerytype, and anesthesia type affect the clinician’s choice of anticoagulant measures.Thus, our result ssuggest that routine thromboprophylaxis for all patients with lower extremity  fractures is inappropriate, and isonly applicable for patients with lower extremity fractures who are at high risk of thrombosis. Multivariate analysis showed that the fracture site was anindependent risk factor of VTE  in patients with  lower extremity fractures. Therefore,prophylactic measures for patients with lower extremity fractures should differ according to the fracture site.

The incidences of major bleeding events were very low in all groups in our study. Only one patient in the rivaroxaban group and four patients in the LMWH group suffered from major bleeding events. Because this was an an observational study, specific interventions for   patients were not pre-determined. The clinicians chose the appropriate anticoagulant drug (or no anticoagulant drug) according to the patients’ specific conditions,and decided  whether or not to use hemostatics in those with bleeding.Therefore, the incidence of major bleeding events in our study was less than that reported in  similar studies. In addition, the incidence of minor bleeding events was also lower. However, the incidences of total bleeding events, major bleeding events, and minor bleeding events in the rivaroxaban group were not statistically different from those in LMWH group,which suggests that the safety of rivaroxaban in patients with lower extremity fractures is comparable to that of LMWHs widely used in clinical practice.

Whereas LMWHs need to be injected subcutaneously due to their pharmaceutical properties, which may result in poor compliance by patients after discharge and difficulty in achieving the full course of anticoagulation, rivaroxaban is administered orally. Its greater convenience and the fact that laborstory monitoring and dose adjustments for the elderly are not required, may lead to improved compliance with long-term anticoagulant  therapy after discharge. Therefore,patients maybe more likely to continue to taking rivaroxaban to prevent VTE in accordance with  their clinician’s instructions,and may be able  to be discharged earlier.In the present study,the mean duration of hospital stay was significantly shorter in the rivaroxaban group than in the LMWH group (12.2 vs 13.1 days, respectively;p =0.016).And we observed that the duration of thromboprophylaxis in rivaroxaban group was clearly longer than in LMWH group,this result reflected oral medication may improve the compliance of thromboprophylaxis after discharge.

Our study has several limitations. Firstly,its retrospective nature limited the randomization and control of patients. Consequently, patients in the rivaroxaban and LMWH groups exhibited some significant differences in baseline demographic characteristics.The median age of the patients and the number  with multiple fractures in the rivaroxaban group were lower than forpatients in the LMWH group, and patients in the twogroups were also significantly different interms of the types of anesthesia and surgery. Therefore,a multivariate analysis of all these factors was performed which indicated that the fracture site was anindependent risk factor fort he occurrence of VTE in this study and that the risk could be significantly reduced  by rivaroxaban administration. Secondly, although the longest duration of follow-up in this study was 3 months after surgery,the patients didnot undergo routine venous ultrasonography or venographyafter discharge and only underwent ultrasound examination of the veins of the lower extremities for suspected VTE at follow-up visits. Therefore, ourresults might have underestimated the real incidence of VTE.Thirdly, we could not determine whether the VTEs were symptomatic or asymptomatic because this information had not been correctly recorded.In addition, due to the limitations of the retrospective design,the medication compliance of patients en rolledin the study was not able to beassessed, and we could not controlfor the use of mechanical thromboprophylaxis  measures after discharge.Despite these shortcomings,our study is the first to report the results of alarge-scale   investigation of the efficacy and safety of rivaroxabanin patients with lower extremity fractures.Our postoperative follow-updata over a period of up to 3months indicated that rivaroxaban had superior efficacy to LMWH in preventing VTE without increasing the risk of bleeding.

Conclusion

This non-randomized, retrospective cohort studyis the first large-scale investigation of the efficacy andsafety of rivaroxaban inpatients with lower extremityfractures. In comparison with LMWH, rivaroxaban reduced the incidence of VTE incidenceby 45 % without increasing the risk of bleeding.

参考文献（略）