Abstract
Introduction
Venous thromboembolism (VTE) is a potentially fatal complication of hip arthroplasty and knee arthroplasty. The National Institute for Health and Care Excellence recommend rivaroxaban for VTE prevention. Amid concerns over bleeding complications, the modified thromboprophylaxis policy of Chelsea and Westminster Hospital (CWH; London, UK) advises enoxaparin given after surgery in the inpatient setting followed by rivaroxaban upon hospital discharge. This retrospective study investigated the efficacy and safety of rivaroxaban in this novel, modified venous-prophylaxis regimen in a surgical orthopaedic cohort at CWH.
Methods
A total of 479 patients who received modified thromboprophylaxis treatment at CWH after hip arthroplasty or knee arthroplasty between April 2013 and October 2014 formed the study cohort. Seven outcomes based on efficacy and safety while undergoing treatment with rivaroxaban were investigated: symptomatic deep-vein thrombosis (DVT), pulmonary embolism (PE), death, stroke, myocardial infarction (MI), major bleeding episodes (MBEs) and non-major bleeding episodes (NMBEs). Median follow-up was 369 days. Fisher’s exact and Mann–Whitney U-tests were adopted to identify associations with these outcomes.
Results
Prevalence of symptomatic PE, DVT, death, stroke and MI during treatment was zero. One (0.2%) MBE and nine (1.9%) NMBEs occurred. The MBE (a wound haematoma) required a return to theatre for aspiration. Off-treatment VTEs occurred in four (0.8%) patients after completion of a course of rivaroxaban, and were associated with known risk factors.
Conclusions
Rivaroxaban is an effective and safe anticoagulant for thromboprophylaxis after hip arthroplasty or knee arthroplasty if used in a modified regimen involving enoxaparin administered in the inpatient setting followed by rivaroxaban upon hospital discharge.
Keywords: Venous thromboembolism, Replacement, Arthroplasty, Rivaroxaban, Knee, Hip
Venous thromboembolism (VTE) is a potentially fatal complication of hip arthroplasty and knee arthroplasty. VTE commonly manifests as deep-vein thrombosis (DVT) and pulmonary embolism (PE).1 Despite adequate thromboprophylaxis with low-molecular-weight heparin (LMWH) or direct/indirect inhibitors of factor Xa or factor IIa, a meta-analysis involving 44,844 cases found that ≈0.5% and ≈1% of patients after hip arthroplasty and knee arthroplasty, respectively, suffered symptomatic VTE before hospital discharge.2 This prevalence increases ≤35 days after surgery, reaching 1.25% for symptomatic DVT and 0.55% for PE on LMWH after major orthopaedic surgery.1
Rivaroxaban is a direct inhibitor of factor Xa and is given via the oral route. Rivaroxaban was introduced into guidelines set by the UK National Institute for Health and Care Excellence (NICE) in 2009. Rivaroxaban is given for VTE prophylaxis after hip replacement surgery or knee replacement surgery in adults for 35 days or 14 days, respectively.3 NICE recommended rivaroxaban after four large multicentre phase-III randomised controlled trials (RCTs), RECORD 1–4, showed rivaroxaban to be superior to enoxaparin with regard to efficacy.4–7 The composite of DVT, PE and mortality occurred less frequently in patients taking rivaroxaban compared with enoxaparin: 1.1% vs 3.7% after hip arthroplasty4 and 9.6% vs 18.9% after knee arthroplasty.6 Furthermore, an extended duration of rivaroxaban was more efficacious than short-course enoxaparin,5 as was a once-daily regimen of rivaroxaban compared with a twice-daily regimen of enoxaparin.7
RCTs present their own limitations with respect to evaluation of the clinical impact of rivaroxaban upon VTE prevention. For example, stringent criteria for patient selection can produce study samples that represent the intended population for the drug inaccurately, thereby compromising the applicability of RCT results to routine practice.8 This scenario calls for regular studies to validate their conclusions in real-life clinical contexts. In the UK, observational studies have primarily investigated rivaroxaban regimens that commence on or within 24 hours of the day of surgery, and have revealed a high prevalence of bleeding and haemorrhagic wound complications relative to LMWH.9–11 Consequently, the modified thromboprophylaxis policy of our institution hospital (Chelsea and Westminster Hospital (CWH), London, UK) advises an initial inpatient course of LMWH (s.c.) postoperatively, then switching to rivaroxaban (p.o.) upon hospital discharge. Our single-centre study aimed to investigate the efficacy and safety of rivaroxaban thromboprophylaxis after hip arthroplasty and knee arthroplasty when used as part of this novel, modified anticoagulation regimen in a cohort of surgical orthopaedic patients in the UK.
Methods
This retrospective cohort study was conducted in a 470-bed teaching hospital (CWH) in accordance with the UK Good Clinical Practice code (clinical governance CAPP 924). All adult patients dispensed rivaroxaban (10mg) after elective or emergency, primary or revision, hip arthroplasty or knee arthroplasty between 1 April 2013 and 1 October 2014, were included. Patients not fulfilling these criteria were excluded from the study. Patients were identified using the dispensing records for rivaroxaban of the CWH pharmacy.
Prevalence of seven primary outcomes during rivaroxaban treatment were investigated: imaging-confirmed symptomatic pulmonary embolism (PE), imaging-confirmed symptomatic DVT, myocardial infarction (MI), stroke, death, major bleeding episodes (MBEs), and non-major bleeding episodes (NMBEs). “Major bleeding” was defined in accordance with recommendations from the Control of Anticoagulation Subcommittee of the International Society on Thrombosis and Haemostasis as a bleed that was fatal, occurred in a critical area or organ, resulted in a fall in haemoglobin of ≥20g/l, required at least two units of blood transfusion,or a surgical-site bleed necessitating return to theatre for open, arthroscopic or endovascular intervention.12 “Non-major bleeding” encompassed all other types of bleeding. Secondary outcomes comprised the prevalence of off-treatment PE and DVT after completion of rivaroxaban courses until the end of follow-up.
Demographic data and details of inpatient admission were obtained from electronic records. Data comprised comorbidities, medication history, operative records, blood results, and thromboprophylaxis management. The follow-up period began on the day of hospital discharge (day-1 of the rivaroxaban course) and ended on 1 February 2015 for all patients. It entailed review of all correspondence from orthopaedic clinics, discharge summaries from readmissions, and imaging reports of five investigations if they occurred after rivaroxaban had been initiated: computed tomography of the pulmonary arteries, ultrasound Doppler of the lower limbs, ventilation–perfusion scans, echocardiography and endoscopy of the gastrointestinal tract.
Guidelines for VTE prophylaxis from CWH recommend mechanical thromboprophylaxis in combination with one of two pharmacological options after elective replacement of the hip or knee. Option one is rivaroxaban (10mg, once daily (OD)) to start 6–10 hours postoperatively and to continue for 35 days after hip arthroplasty or 14 days after knee arthroplasty. Option two is enoxaparin (40mg OD) to start 6–12 hours postoperatively and to continue for the remainder of hospital admission. In the latter, enoxaparin is replaced by rivaroxaban (10mg OD) upon hospital discharge to achieve a cumulative duration of anticoagulation of 35 days after hip arthroplasty and 14 days after knee arthroplasty. This particular regimen was chosen by CWH to promote greater compliance with postoperative anticoagulation after the introduction of rivaroxaban into NICE guidelines. The regimen of enoxaparin (s.c.) during inpatient stay was familiar to the clinical team, and so minimised confusion in hospital, whereas switching to rivaroxaban (p.o.) upon hospital discharge meant patients could self-administer rivaroxaban conveniently. The guideline for emergency hip replacement comprised mechanical prophylaxis and enoxaparin (40mg OD) only continuing for 28 days after surgery.
Statistical analyses
SPSS v22.0 (IBM, Armonk, NY, USA) was used for data analyses. Fisher’s exact test was used to identify associations between categorical variables and outcomes in view of small numbers. The Mann–Whitney U-test was used to compare continuous variables and outcomes. Data are presented as medians accompanied by interquartile ranges. p<0.05 (two sided) was considered significant.
Results
Four-hundred and seventy-nine patients formed the study cohort (Table 1 and Fig 1). Median follow-up was 369 (range, 116–756) days. All patients (including emergency cases) followed option 2 of the thromboprophylaxis protocol of CWH (Fig 2a and 2b). Routine assessment of VTE and bleeding risk for 477 patients upon hospital admission revealed that 464 had one or more risk factor for VTE and 109 had one or more risk factor for bleeding.
Table 1.
Patient characteristics, procedural information and blood results in the inpatient setting
| General | ||
| Female sex – n (%) | 293 (61.2) | |
| Age – years median (IQR) | 70 (61–75) | |
| Elective – n (%) | 446 (93.1) | |
| Duration of stay – days median (IQR) | 5 (4–7) | |
| Procedure – n (%) | ||
| Primary hip arthroplasty Primary knee arthroplasty Revision hip arthroplasty Revision knee arthroplasty |
247 (51.6) 205 (42.8) 13 (2.7) 14 (2.9) |
|
| Medical history – n (%) | ||
| Hypertension Diabetes mellitus Ischaemic heart disease COPD Atrial fibrillation Previous VTE Chronic kidney disease Congestive cardiac failure |
222 (46.3) 50 (10.4) 25 (5.2) 15 (3.1) 11 (2.3) 8 (1.7) 3 (0.6) 1 (0.2) |
|
| Medication history – n (%) | ||
| ACEI/ARB Statin Gastric protection CCB Aspirin Thiazide Clopidogrel Anti-retroviral Anticoagulation |
149 (31.1) 141 (29.4) 126 (26.3) 100 (20.9) 48 (10.0) 41 (8.6) 6 (1.3) 6 (1.3) 0 (0.0) |
|
| Blood results upon hospital discharge – median (IQR) | ||
| Haemoglobin – g/l Platelets – ×109/l PT – s APTT – s Urea – mmol/l Creatinine – μmol/l eGFR – ml/min/1.73 m2 Bilirubin – μmol/l Alanine aminotransferase – iU/l |
104 (94.0–113.0) 231 (184.8–289.5) 10.4 (10.1–10.8) 25.2 (23.6–26.8) 4.65 (3.6–6.1) 68 (60.0–80.8) 84 (70.0–91.0) 11 (8.0–15.0) 20 (15.0–30.0) |
|
ACEI = angiotensin-converting enzyme inhibitor; ARB = angiotensin receptor blocker; IQR = interquartile range; COPD = chronic obstructive pulmonary disease; PT= prothrombin time; APTT = activated partial thromboplastin time; eGFR = estimated glomerular filtration rate; CCB = calcium channel blocker
Figure 1.
STROBE28 flow diagram showing the number of patients identified, included and analysed during the study period
Figure 2.
(a) Compliance and time until first postoperative administration of anti-embolism stockings (AES) and enoxaparin in the inpatient setting. Horizontal axis denotes the type of thromboprophylaxis in the inpatient setting. Vertical axis denotes the percentage of patients administered thromboprophylaxis at the end of the following time points postoperatively: <12 hours, 12–24 hours, >24 hours, never. (b) Postoperative anticoagulation course. Duration of enoxaparin in the inpatient setting and rivaroxaban in the outpatient setting, start day of rivaroxaban and total duration of anticoagulation. Values are expressed as the median and interquartile range.
Primary outcomes
VTE, death, stroke or MI did not occur during rivaroxaban treatment. One (0.2%) MBE and nine (1.9%) NMBEs took place on or within 48 hours of stopping rivaroxaban. The MBE (a wound haematoma) occurred 7 days into the rivaroxaban course and was associated with a reduction in haemoglobin level by 15g/l. Management entailed return to theatre for aspiration of 50ml of blood-stained fluid. The patient recovered and was discharged with no further complications. The nine NMBEs resolved without the need for surgical intervention. Three NMBEs were related to the surgical site (two cases of wound haematoma and one case of wound oozing). The remaining six NMBEs comprised five cases of leg bruising and one case of per vaginam bleeding. The latter was attributed to pre-existing gynaecological disease and follow-up was conducted by the gynaecology team. Bleeding events were associated with a lower activated partial thromboplastin time (APTT) upon hospital discharge (p=0.028) only. Associations with ischaemic heart disease (p=0.091), reduced estimated glomerular filtration rate (p=0.064) and, therefore, higher levels of urea (p=0.094) and creatinine (p=0.079) could not be excluded (Supplementary Table 2). None of the patients in whom bleeding events occurred had risk factors for bleeding at risk assessment upon hospital admission.
Supplementary Table 2.
Factors associated with bleeding events on treatment or within 48 hours of stopping rivaroxaban (significant p-values are emboldened)
| Bleed (n=10) | No bleed (n=469) | p | |
| Female sex – n (%) | 6 (60) | 287 (61.2) | 1.000 |
| Age – years median(IQR) | 75 (15) | 69 (14) | 0.123 |
| Elective – n (%) | 10 (100) | 436 (93) | 1.000 |
| Hip arthroplasty* – n (%) | 3 (30) | 257 (54.8) | 0.198 |
| Duration of stay – days median (IQR) | 5 (2) | 5 (3) | 0.384 |
| Medical history | |||
| Hypertension – n (%) | 4 (40) | 218 (46.5) | 0.758 |
| Diabetes – n (%) | 0 (0) | 50 (10.7) | 0.609 |
| IHD – n (%) | 2 (20) | 23 (4.9) | 0.091 |
| COPD – n (%) | 1 (10) | 14 (3.0) | 0.275 |
| AF – n (%) | 0 (0) | 11 (2.3) | 1.000 |
| Previous VTE – n (%) | 1 (10) | 7 (1.5) | 0.156 |
| CKD – n (%) | 0 (0) | 3 (0.6) | 1.000 |
| CCF – n (%) | 0 (0) | 1 (0.2) | 1.000 |
| Drug history | |||
| ACEI/ARB – n (%) | 3 (30) | 146 (31.1) | 1.000 |
| Statin – n (%) | 4 (40) | 137 (29.2) | 0.490 |
| Gastric protection – n (%) | 3 (30) | 123 (26.3) | 0.727 |
| CCB – n (%) | 1 (10) | 99 (21.1) | 0.696 |
| Aspirin – n (%) | 2 (20) | 46 (9.8) | 0.264 |
| Thiazide – n (%) | 2 (20) | 39 (8.3) | 0.208 |
| Clopidogrel – n (%) | 0 (0) | 6 (1.3) | 1.000 |
| Anti-retroviral – n (%) | 1 (10) | 5 (1.1) | 0.119 |
| Anticoagulation – n (%) | 0 (0) | 0 (0) | n/a |
| Blood results upon hospital discharge | |||
| Haemoglobin – g/l, median (IQR) | 101.5 (50.55) | 104 (19) | 0.904 |
| Platelets – ×109/l, median (IQR) | 245 (120.75) | 231 (104.5) | 0.668 |
| PT – s, median (IQR) | 10.7 (0.6) | 10.4 (0.7) | 0.275 |
| APTT –s, median (IQR) | 23.9 (3.2) | 25.2 (3.2) | 0.028 |
| Urea – mmol/l, median (IQR) | 6.15 (4.33) | 4.6 (2.4) | 0.094 |
| Creatinine – μmol/l, median (IQR) | 83.5 (38.5) | 68 (20) | 0.075 |
| eGFR – ml/min/1.73 m2, median (IQR) | 75.5 (32.75) | 84 (21) | 0.064 |
| Bilirubin – μmol/l, median (IQR) | 10 (7) | 11 (7) | 0.999 |
| ALT – iU/l, median (IQR) | 24 (21.25) | 20 (15) | 0.584 |
| Inpatient thromboprophylaxis | |||
| AES <12 hours, – n (%) | 1 (10) | 31 (6.6) | 0.503 |
| AES 12–24 hours, – n (%) | 3 (30) | 189 (40.3) | 0.747 |
| AES >24 hours – n (%) | 1 (10) | 48 (10.4) | 1.000 |
| AES not administered – n (%) | 5 (50) | 201 (42.9) | 0.751 |
| Enoxaparin <12 hours – n (%) | 9 (90) | 392 (83.6) | 1.000 |
| Enoxaparin 12–24 hours – n (%) | 1 (10) | 30 (6.4) | 0.491 |
| Enoxaparin >24 hours – n (%) | 0 (0) | 46 (9.8) | 0.609 |
| Enoxaparin not administered – n (%) | 0 (0) | 1 (0.2) | 1.000 |
| Enoxaparin duration – days median (IQR) | 5 (1) | 4 (3) | 0.575 |
| Outpatient thromboprophylaxis | |||
| Rivaroxaban start dayo – days median (IQR) | 6 (1) | 6 (4) | 0.653 |
| Rivaroxaban duration (hip) – days median (IQR) | 30*** | 31 (3) | 0.510 |
| Rivaroxaban duration (knee) – days median (IQR) | 9 (2) | 10 (3) | 0.278 |
| Total anticoagulation | |||
| Total anticoagulation (hip) – days median (IQR) | 35*** | 35 (1) | 0.653 |
| Total anticoagulation (knee) – days median (IQR) | 14 (2) | 14 (1) | 0.110 |
*Composite of primary and revision hip arthroplasties. **n=1 patient so the value is absolute rather than median, no IQR. ***n=3 patients so the IQR was not available. oNumber of days after surgery that rivaroxaban was commenced.
IHD = ischaemic heart disease; COPD = chronic obstructive pulmonary disease; AF = atrial fibrillation; VTE = venous thromboembolism; CKD = chronic kidney disease; CCF = congestive cardiac failure; ACEI/ARB = angiotensin-converting enzyme inhibitor/angiotensin receptor blocker; CCB = calcium channel blocker; PT = prothrombin time; APTT = activated partial thromboplastin time; eGFR = estimated glomerular filtration rate; ALT = alanine aminotransferase; AES = anti-embolism stocking; IQR = interquartile range
Secondary outcomes
After completion of the rivaroxaban course until the end of follow-up, four (0.8%) off-treatment VTE events were observed (Supplementary Table 3). They consisted of three PEs (0.6%) and one DVT (0.2%) on days 2, 41, 67 and 5, respectively, after stopping rivaroxaban. The PE on day-41 resulted in death. These VTE events were associated with a history of atrial fibrillation (p=0.003), chronic kidney disease (p=0.025), high platelet count (p=0.029), prolonged stay in hospital (p=0.041), long treatment course of enoxaparin (p=0.034) and later start day for rivaroxaban (p=0.04), whereas an association with previous VTE could not be excluded (p=0.065). Despite comorbidities, no patients were taking anticoagulation before hospital admission. Three of these patients did not receive anti-embolism stockings during inpatient admission, but all four were administered enoxaparin within 12 hours of their surgical procedure. One stroke occurred ≈60 days after rivaroxaban completion against a background of atrial fibrillation and previous VTE, and this patient died later. Risk assessment upon hospital admission showed that patients in whom these secondary outcomes occurred had one or two risk factors for VTE.
Supplementary Table 3.
Factors associated with off-treatment VTE events after completion of a course of rivaroxaban (significant p-values are emboldened)
| VTE (n=4) | No VTE (n=475) | p | |
|---|---|---|---|
| Female sex – n (%) | 2 (50) | 291 (61.3) | 0.644 |
| Age – years median (IQR) | 72.5 (25) | 70 (14) | 0.308 |
| Elective – n (%) | 3 (75) | 443 (93.3) | 0.249 |
| Hip arthroplasty* – n (%) | 1 (25) | 259 (54.5) | 0.336 |
| Duration of stay – days median (IQR) | 9.5 (12) | 5 (3) | 0.041 |
| Medical history | |||
| Hypertension – n (%) | 2 (50) | 220 (46.3) | 1.000 |
| Diabetes mellitus – n (%) | 2 (50) | 48 (10.1) | 0.056 |
| IHD – n (%) | 1 (25) | 24 (5.1) | 0.194 |
| COPD – n (%) | 0 (0) | 15 (3.2) | 1.000 |
| AF – n (%) | 2 (50) | 9 (1.9) | 0.003 |
| Previous VTE – n (%) | 1 (25) | 7 (1.5) | 0.065 |
| CKD – n (%) | 1 (25) | 2 (0.4) | 0.025 |
| CCF – n (%) | 0 (0) | 1 (0.2) | 1.000 |
| Drug history | |||
| ACEI/ARB – n (%) | 1 (25) | 148 (31.2) | 1.000 |
| Statin – n (%) | 2 (50) | 139 (29.3) | 0.585 |
| Gastric protection – n (%) | 2 (50) | 124 (26.1) | 0.284 |
| CCB – n (%) | 1 (25) | 99 (20.8) | 1.000 |
| Aspirin – n (%) | 1 (25) | 47 (9.9) | 0.245 |
| Thiazide – n (%) | 0 (0) | 41 (8.6) | 1.000 |
| Clopidogrel – n (%) | 0 (0) | 6 (1.3) | 1.000 |
| Anti-retroviral – n (%) | 0 (0) | 6 (1.3) | 1.000 |
| Anticoagulation – n (%) | 0 (0) | 0 (0) | n/a |
| Blood results upon hospital discharge | |||
| Haemoglobin – g/l, median (IQR) | 108 (30) | 104 (19) | 0.833 |
| Platelets – ×109/l, median(IQR) | 307 (2.8) | 231 (104) | 0.029 |
| PT – s, median (IQR) | 10.65 (0.85) | 10.4 (0.7) | 0.635 |
| APTT –s, median (IQR) | 23.5 (3.13) | 25.2 (3.2) | 0.272 |
| Urea – mmol/l, median (IQR) | 9.05 (5.35) | 4.60 (2.38) | 0.06 |
| Creatinine – μmol/l, median (IQR) | 103 (71.25) | 68 (20) | 0.212 |
| eGFR – ml/min/1.73 m2, median (IQR) | 68 (54) | 84 (21) | 0.378 |
| Bilirubin – μmol/L median (IQR) | 16.5 (19) | 11 (7) | 0.229 |
| ALT – iU/L median (IQR) | 23 (19.5) | 20 (15) | 0.667 |
| Inpatient thromboprophylaxis | |||
| AES <12 hours – n (%) | 0 (0) | 32 (6.7) | 1.000 |
| AES 12-24 hours – n (%) | 1 (25) | 191 (40.1) | 0.653 |
| AES >24 hours – n (%) | 0 (0) | 49 (10.2) | 1.000 |
| AES not administered – n (%) | 3 (75) | 203 (42.7) | 0.319 |
| Enoxaparin <12 hours – n (%) | 4 (100) | 397 (83.6) | 0.100 |
| Enoxaparin 12–24 hours – n (%) | 0 (0) | 31 (6.5) | 1.000 |
| Enoxaparin >24 hours – n (%) | 0 (0) | 46 (9.7) | 1.000 |
| Enoxaparin not administered – n (%) | 0 (0) | 1 (0.2) | 1.000 |
| Enoxaparin duration – median days (IQR) | 9.5 (10) | 4 (3) | 0.034 |
| Outpatient thromboprophylaxis | |||
| Rivaroxaban start dayo –day median (IQR) | 10.5 (10) | 6 (5) | 0.04 |
| Rivaroxaban duration (hip) – days (IQR) | 16** | 31 (3) | 0.1 |
| Rivaroxaban duration (knee) – days median (IQR) | 11*** | 10 (3) | 0.933 |
| Total anticoagulation | |||
| Total anticoagulation (hip) –days(IQR) | 33** | 35(1) | 0.118 |
| Total anticoagulation (knee) – days median (IQR) | 15*** | 14 (1) | 0.333 |
*Composite of primary and revision hip arthroplasty. **n=1 patients so the value is absolute rather than median, no IQR. ***n=3 patients so the IQR was not available. oNumber of days after surgery that rivaroxaban was commenced.
IHD = ischaemic heart disease; COPD = chronic obstructive pulmonary disease; AF = atrial fibrillation; VTE = venous thromboembolism; CKD = chronic kidney disease; CCF = congestive cardiac failure; ACEI/ARB = angiotensin-converting enzyme inhibitor/angiotensin receptor blocker; CCB = calcium channel blocker; PT = prothrombin time; APTT = activated partial thromboplastin time; eGFR = estimated glomerular filtration rate; ALT = alanine aminotransferase; AES = anti-embolism stocking; IQR = interquartile range
Discussion
Our study is the first in the UK to investigate the efficacy and safety of rivaroxaban in a modified thromboprophylaxis regimen involving enoxaparin given in the inpatient setting followed by rivaroxaban given in the outpatient setting after hip arthroplasty or knee arthroplasty.
We found no cases of symptomatic VTE (composite of PE and DVT) or death during rivaroxaban treatment. Of ten (2.1%) bleeding events that occurred during or within 48 hours of stopping rivaroxaban, only one (0.2%) was classified as a MBE and necessitated a return to theatre.
Pooled analyses of the four RECORD phase-III trials4–7 quoted a prevalence of a composite of symptomatic VTE and all-cause mortality during rivaroxaban treatment to be 0.5%13 to 0.6%.14–16 Bleeding events during or within 48 hours of cessation of rivaroxaban occurred at a prevalence of 7.02%,17 with MBEs comprising 0.3–0.5%.13,14,17 NMBEs took place in 5.6% of patients.13
After the RECORD RCTs, various studies focused on the efficacy of anticoagulation protocols consisting exclusively of rivaroxaban after hip arthroplasty or knee arthroplasty. In a single-centre prospective study of 206 patients administered rivaroxaban (10mg) after knee replacement in 2010 in the UK, 1.5% of patients developed symptomatic VTE and 1.9% of patients returned to theatre as a result of wound oozing or haematoma.11 In another single-centre cohort in Dresden (Germany), the prevalence of in-hospital symptomatic VTE was 2.1% among 1,043 patients receiving rivaroxaban (10mg) venous prophylaxis after hip replacement or knee replacement between 2010 and 2011.18 Prevalence of MBE was 2.9%, which included 0.38% of patients who needed to return to theatre.
Other UK studies have primarily investigated the bleeding risk associated with rivaroxaban if commenced within 24 hours of the surgical procedure, while also reporting on the prevalence of VTE as a secondary outcome. A multi-centre prospective study involving 2,762 patients in the UK taking rivaroxaban after hip arthroplasty or knee arthroplasty identified a prevalence of MBE of 1.05% and a prevalence of wound complications of 3.84%, which included (but was not exclusive to) wound haematoma. Symptomatic VTE at 90 days occurred in 0.72%of cases.19 In another UK centre of 559 patients taking rivaroxaban compared with 489 patients taking enoxaparin after hip replacement or knee replacement, nine compared with zero patients had to return to theatre as a result of wound haematoma, respectively. PE occurred in 0.9% of patients taking rivaroxaban.9 An observational study of 202 patients across two UK centres receiving rivaroxaban after elective hip arthroplasty or knee arthroplasty elicited four patients (2.0%) who had a NMBE in addition to three patients who sustained haematomas (1.5%) that required a return to theatre. No other MBEs, PEs or deaths occurred, but 1% of patients suffered DVT.10 All of these studies, in which rivaroxaban was commenced in the immediate postoperative period, demonstrated a higher prevalence of bleeding and return to theatre than that obtained with our modified anticoagulation regimen. Therefore, a prospective study is warranted to compare these two protocols directly.
Our secondary outcome, off-treatment VTE events, occurred in patients with known risk factors, including atrial fibrillation and previous VTE. Major orthopaedic surgery alone conveys a high risk of VTE,1 and our findings highlight the importance of caring for patients with additional risk factors for thromboembolism on an individual basis. Thus, attention is drawn to the poor compliance with anti-embolism stockings observed in the present study as a preventable aetiological factor in these patients. Furthermore, The ROCKET-AF20 and EINSTEIN-extension21 studies have shown the efficacy of rivaroxaban (20mg) as another solution for patients with atrial fibrillation or a history of VTE, respectively, and have attained NICE approval.22,23 Longer durations of hospital stay are also known to contribute to VTE by reducing patient mobility,24,25 observations that are consistent with the findings from our cohort. Due to the nature of the CWH guidelines for thromboprophylaxis, these longer durations of admission translated to anticoagulation regimens comprising proportionally more enoxaparin relative to rivaroxaban, which provides one explanation why VTE was also associated with longer courses of enoxaparin and later start dates of rivaroxaban. Though they themselves may be a possible cause, the confounding impact exerted by reduced ambulation in these patients negates the clinical validity of independent conclusions that can be drawn. Finally, chronic kidney disease and higher platelet counts upon hospital discharge were also found in VTE patients. Chronic kidney disease is a well-cited aetiological factor,26 whereas increased platelet counts can contribute to thrombosis through an increase in blood viscosity (one of the three constituents of Virchow’s triad). This phenomenon is seen in chronic myeloproliferative disorders, including essential thrombocythaemia.27 A larger study cohort is needed to confirm the relevance of the associations identified in our study, particularly for bleeding events (in which the association with lower APTT values shown in our cohort contradicts those from other studies).
To appreciate fully the clinical importance of the comparably low prevalence of symptomatic VTE and bleeding in our cohort, the seven main limitations of the present study must be considered carefully. First, the study focused on the efficacy and safety of rivaroxaban only in a modified thromboprophylaxis regimen that comprised enoxaparin and rivaroxaban. Only primary outcomes that occurred after hospital discharge were measured because this was when rivaroxaban was commenced. Prevalence of outcomes during inpatient stay (while patients were on enoxaparin) was not measured. Second, data on thromboprophylaxis in the inpatient setting were extracted from electronic prescribing records. Such records detail the prevalence of administration of prescribed medications, but such treatments may have been administered without prescription. This phenomenon could account for the low prevalence of use of anti-embolism stockings seen in our study. Third, all patients were prescribed and dispensed rivaroxaban, but the prevalence of patient compliance was not determined. Fourth, the end of follow-up was fixed at 1 February 2015, so the duration of follow-up varied between patients (which could have accounted for some of the unavailable clinic letters). Fifth, readmissions to our hospital were a key component of follow-up, but we did not take into account presentations to (or investigations at) other hospitals, which was a potential source of missed outcomes. Sixth, patients received enoxaparin and rivaroxaban at different points of their thromboprophylaxis course, so attributing the causes of adverse events to a specific treatment was difficult. Finally, statistical associations were found using appropriate tests, but the clinical relevance of such findings (and the importance of subsequent conclusions) were difficult to determine because of the small numbers of outcomes observed.
Many of these limitations were secondary to the retrospective nature of this study. Consequently, the sources used for data collection were not tailored specifically towards the outcomes of this study.
Conclusions
Our retrospective cohort study showed a modified thromboprophylaxis regimen of rivaroxaban taken in the outpatient setting following enoxaparin taken in the inpatient setting to be an effective and safe option after hip arthroplasty or knee arthroplasty. Prevalence of on-treatment symptomatic VTE, mortality and bleeding was lower than that reported in similar UK studies of rivaroxaban-exclusive thromboprophylaxis courses. Our findings call for greater vigilance in ensuring compliance with mechanical thromboprophylaxis if contraindications are absent, particularly in patients with pre-existing risk factors who are likely to require individualised thromboprophylaxis regimens.
Reference
- 1.Falck-Ytter Y, Francis CW, Johanson NA et al. Prevention of VTE in orthopedic surgery patients: Antithrombotic Therapy and Prevention of Thrombosis, 9th ed: American College of Chest Physicians Evidence-Based Clinical Practice Guidelines. Chest 2012; : e278S– 325S. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Januel JM, Chen G, Ruffieux C et al. Symptomatic in-hospital deep vein thrombosis and pulmonary embolism following hip and knee arthroplasty among patients receiving recommended prophylaxis: a systematic review. JAMA 2012; : 294–303. [DOI] [PubMed] [Google Scholar]
- 3.NICE guidance Rivaroxaban for the prevention of venous thromboembolism after total hip or total knee replacement in adults (TA 170). Issued April 2009 www.nice.org.uk/guidance/ta170 (cited May 2015).
- 4.Eriksson BI, Borris LC, Friedman RJ et al. Rivaroxaban versus enoxaparin for thromboprophylaxis after hip arthroplasty. N Engl J Med 2008; : 2,765–2,775. [DOI] [PubMed] [Google Scholar]
- 5.Kakkar AK, Brenner B, Dahl OE et al. Extended duration rivaroxaban versus short-term enoxaparin for the prevention of venous thromboembolism after total hip arthroplasty: a double-blind, randomised controlled trial. Lancet 2008; : 31–39. [DOI] [PubMed] [Google Scholar]
- 6.Lassen MR, Ageno W, Borris LC et al. Rivaroxaban versus enoxaparin for thromboprophylaxis after total knee arthroplasty. N Engl J Med 2008; : 2,776–2,786. [DOI] [PubMed] [Google Scholar]
- 7.Turpie AG, Lassen MR, Davidson BL et al. Rivaroxaban versus enoxaparin for thromboprophylaxis after total knee arthroplasty (RECORD4): a randomised trial. Lancet 2009; : 1,673–1,680. [DOI] [PubMed] [Google Scholar]
- 8.Rothwell PM. External validity of randomised controlled trials: “to whom do the results of this trial apply?” Lancet 2005; : 82–93. [DOI] [PubMed] [Google Scholar]
- 9.Jensen CD, Steval A, Partington PF, Reed MR, Muller SD. Return to theatre following total hip and knee replacement, before and after the introduction of rivaroxaban: a retrospective cohort study. J Bone Joint Surg Br 2011; : 91–95. [DOI] [PubMed] [Google Scholar]
- 10.Sindali K, Rose B, Soueid H et al. Elective hip and knee arthroplasty and the effect of rivaroxaban and enoxaparin thromboprophylaxis on wound healing. Eur J Orthop Surg Traumatol 2013; : 481–486. [DOI] [PubMed] [Google Scholar]
- 11.Rath NK, Goodson MW, White SP, Forster MC. The use of rivaroxaban for chemical thromboprophylaxis following total knee replacement. Knee 2013; : 397–400. [DOI] [PubMed] [Google Scholar]
- 12.Schulman S, Anger SU, Bergqvist D, et al. Definition of major bleeding in clinical investigations of antihemostatic medicinal products in surgical patients. J Thromb Haemost 2010; : 202–204. [DOI] [PubMed] [Google Scholar]
- 13.Eriksson BI, Kakkar AK, Turpie AG et al. Oral rivaroxaban for the prevention of symptomatic venous thromboembolism after elective hip and knee replacement. J Bone Joint Surg Br 2009; : 636–644. [DOI] [PubMed] [Google Scholar]
- 14.Huisman MV, Quinlan DJ, Dahl OE, Schulman S. Enoxaparin versus dabigatran or rivaroxaban for thromboprophylaxis after hip or knee arthroplasty: Results of separate pooled analyses of phase III multicenter randomized trials. Circ Cardiovasc Qual Outcomes 2010; : 652–660. [DOI] [PubMed] [Google Scholar]
- 15.Turpie AG, Lassen MR, Eriksson BI et al. Rivaroxaban for the prevention of venous thromboembolism after hip or knee arthroplasty. Pooled analysis of four studies. Thromb Haemost 2011; : 444–53. [DOI] [PubMed] [Google Scholar]
- 16.Kwong LM. Therapeutic potential of rivaroxaban in the prevention of venous thromboembolism following hip and knee replacement surgery: a review of clinical trial data. Vasc Heal Risk Manag 2011; : 461– 466. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Lassen MR, Gent M, Kakkar AK et al. The effects of rivaroxaban on the complications of surgery after total hip or knee replacement: results from the RECORD programme. J Bone Joint Surg Br 2012; : 1,573–1,578. [DOI] [PubMed] [Google Scholar]
- 18.Beyer-Westendorf J, Lutzner J, Donath L et al. Efficacy and safety of rivaroxaban or fondaparinux thromboprophylaxis in major orthopedic surgery: findings from the ORTHO-TEP registry. J Thromb Haemost 2012; : 2,045–52. [DOI] [PubMed] [Google Scholar]
- 19.Jameson SS, Rymaszewska M, Hui AC et al. Wound complications following rivaroxaban administration: a multicenter comparison with low-molecular-weight heparins for thromboprophylaxis in lower limb arthroplasty. J Bone Joint Surg Am 2012; : 1,554–1,558. [DOI] [PubMed] [Google Scholar]
- 20.Patel MR, Mahaffey KW, Garg J, Pan G, Singer DE, Hacke W, et al. Rivaroxaban versus warfarin in nonvalvular atrial fibrillation. N Engl J Med [Internet]. 2011; : 883–91. [DOI] [PubMed] [Google Scholar]
- 21.Romualdi E, Donadini MP, Ageno W. Oral rivaroxaban after symptomatic venous thromboembolism: the continued treatment study (EINSTEIN-extension study). Expert Rev Cardiovasc Ther 2011; : 841–844. [DOI] [PubMed] [Google Scholar]
- 22.NICE guidance Rivaroxaban for the prevention of stroke and systemic embolism in people with atrial fibrillation (TA 256). Issued May 2012 www.nice.org.uk/guidance/ta256( cited May 2015) .
- 23.NICE guidance Rivaroxaban for the treatment of deep vein thrombosis and prevention of recurrent deep vein thrombosis and pulmonary embolism (TA 261). Issued July 2012 www.nice.org.uk/guidance/ta261( cited May 2015).
- 24.Samama CM, Ravaud P, Parent F et al. Epidemiology of venous thromboembolism after lower limb arthroplasty: the FOTO study. J Thromb Haemost 2007; : 2,360–2,367. [DOI] [PubMed] [Google Scholar]
- 25.Chandrasekaran S, Ariaretnam SK, Tsung J, Dickison D. Early mobilization after total knee replacement reduces the incidence of deep venous thrombosis. ANZ J Surg 2009; : 526–529. [DOI] [PubMed] [Google Scholar]
- 26.Wattanakit K, Cushman M, Stehman-Breen C, Heckbert SR, Folsom AR. Chronic kidney disease increases risk for venous thromboembolism. J Am Soc Nephrol 2008; : 135–140. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Landolfi R, Marchioli R, Patrono C. Mechanisms of bleeding and thrombosis in myeloproliferative disorders. Thromb Haemost 1997; : 617–621. [PubMed] [Google Scholar]
- 28.Vandenbroucke JP, Von Elm E, Altman DG et al. Strengthening the Reporting of Observational Studies in Epidemiology (STROBE): explanation and elaboration. PLoS Med 2007; : 1,628–1,654. [DOI] [PMC free article] [PubMed] [Google Scholar]