Background
Patients undergoing total hip or total knee replacement (THR, TKR) are at risk of developing venous thromboembolic events and therefore, thromboprophylaxis is recommended [
1]. Compared with low-molecular-weight heparin (LMWH) administered subcutaneously, oral anticoagulants such as dabigatran etexilate (hereafter described as dabigatran) offer important practical advantages. Dabigatran, a low-molecular-weight, reversible thrombin inhibitor, is recommended for primary prevention of venous thromboembolism (VTE) in THR and TKR [
1] and has been approved for this indication in more than 100 countries.
Dabigatran is 80 % renally excreted and its terminal half-life is approximately 11–17 h [
2,
3]. The European Medicines Agency recommends dabigatran at a dose of 220 mg once daily for the primary prevention of VTE in patients undergoing orthopaedic surgery [
3]. For patients with moderate renal impairment, and for those aged ≥75 years or receiving concomitant amiodarone or quinidine, a reduced dose of 150 mg once daily is recommended [
3]. Both doses of dabigatran should be initiated 1–4 h after surgery with a half dose to mitigate the bleeding risk in the vulnerable post-operative phase.
The efficacy and safety of dabigatran in the primary prevention of VTE after elective THR or TKR was demonstrated in the phase 3 studies, RE-MODEL [
4], RE-MOBILIZE [
5], RE-NOVATE [
6] and RE-NOVATE II [
7]. More than 10,000 patients were randomised in these trials and approximately 6400 received dabigatran. The pooled analysis of the RE-MODEL, RE-MOBILIZE and RE-NOVATE trials showed that dabigatran had similar efficacy to enoxaparin in the prevention of major VTE and VTE-related mortality after knee or hip replacement [
8]. The results of the RE-NOVATE II trial were consistent with these findings [
7]. No statistically significant differences in the incidence of major bleeding events (MBEs) between treatment groups were found, and no safety concerns regarding elevations in liver function tests or acute coronary syndrome (ACS) events were identified [
4,
6‐
9].
This report provides the results from a large, real-world, observational study of dabigatran in patients undergoing THR or TKR in a routine clinical setting. The purpose of this study was to assess the safety and efficacy of dabigatran 220 mg once daily in all patients who received the drug, as well as in protocol-defined subgroups of patients who had a potentially increased risk of bleeding and/or VTE (i.e. patients with special comorbidities or comedication). This is the first study to provide an insight into the outcomes following use of dabigatran for the prevention of VTE in a real-world orthopaedic setting.
Discussion
Results from this observational study showed that there is no increased risk of bleeding with dabigatran 220 mg once daily in patients undergoing THR or TKR (irrespective of the presence or absence of protocol-defined risk factors for bleeding) in a real-world clinical setting compared with the phase 3 clinical trials [
4,
6,
7]. A history of VTE, a known risk factor for subsequent VTE [
12], was the only protocol-defined characteristic associated with an increased primary efficacy event rate (sVTE or death). A higher event rate in patients with this risk factor was observed for those undergoing TKR (7.96 %) but not for those undergoing THR (0 %).
The American College of Chest Physicians (ACCP) guidelines estimated that, from 0 to 35 days, the cumulative post-operative rate of nonfatal sVTE in patients undergoing major orthopaedic surgery is 4.3 % in the absence of prophylaxis and 1.8 % with an LMWH [
1]. The incidence of the primary endpoint in this study, which included fatal as well as nonfatal VTE events, was 1.04 % and therefore lower than the ACCP estimates when patients were treated with an LMWH. Although several factors may have influenced the incidence of sVTE events and bleeding events, it is worth noting that characteristics of the patients included in this observational study are broadly aligned with previously published real-world data sets [
13‐
15].
The primary efficacy endpoint is not directly comparable between this observational study and the phase 3 trials. In the observational study the primary endpoint was sVTE and all-cause death during the treatment period and the rate was primarily driven by the occurrence of symptomatic distal DVTs. The primary endpoint in the phase 3 studies, RE-MODEL, RE-NOVATE and RE-NOVATE II, was total VTE and all-cause mortality, and included events observed during routine venography.
To explore the data more fully, we performed a post-hoc analysis to facilitate closer comparison between this observational study and the phase 3 results. To lessen the differences introduced by the study designs, only patients ≤75 years of age, from outside the United States and Canada, treated with dabigatran 220 mg once daily with documented information on medication start date and CrCl >50 mL/min, were included. Based on medical judgement, the patient populations were similar in the observational study and the phase 3 TKR and THR trials of dabigatran, apart from a lower incidence of CAD at baseline and a lower concomitant use of NSAIDs in the observational, real-world setting study. Analysis in these comparable populations (4999 patients from the observational study and 2118 from the phase 3 trials) showed that safety and efficacy of dabigatran in this observational study were as favourable as those obtained in the phase 3 dabigatran clinical trials. There was a similar or lower risk of an MBE in this observational study compared with the phase 3 trials in this post-hoc comparison: 0.62 % (95 % CI 0.36, 1.01) for THR and 0.74 % (95 % CI 0.44, 1.17) for TKR in the observational study compared with 0.65 % (95 % CI 0.24, 1.41) in RE-NOVATE and 0.57 % (95 % CI 0.16, 1.46) in RE-NOVATE II for THR, and 1.20 % (95 % CI 0.44, 2.58) in RE-MODEL for TKR. In the same analysis, the incidence of the composite of sVTE and all-cause mortality in the observational study was 0.50 % (95 % CI 0.27, 0.86) for THR and 1.40 % (95 % CI 0.97, 1.96) for TKR. This was intermediate between the results from RE-NOVATE (1.20 % [95 % CI 0.60, 2.13]) and RE-NOVATE II (0.14 % [95 % CI 0.00, 0.08]) for THR, and higher than the results from RE-MODEL (0.20 % [95 % CI 0.01, 1.10]) for TKR. Differences in data collection methods (especially central adjudication of events vs local investigator assessment), patient populations and definitions of efficacy outcome events in the observational study and the phase 3 trials should be considered when interpreting the results from this post-hoc analysis.
Contrary to the protocol recommendation and to ACCP recommendations, a subset of sites performed routine ultrasound examinations to detect VTE; asymptomatic as well as sVTE events could therefore have been reported. We noticed in a stratified analysis by country that the incidence of sVTE and all-cause mortality was higher in France (2.15 %), where duplex sonography is routine clinical practice in some institutions [
16], than in other countries (0–1.37 %). Within France, the incidence was highest for patients undergoing knee surgery (4.38 % vs 0.67 % for hip surgery patients). A further post-hoc analysis showed that the incidence of sVTE and all-cause mortality was lower when patients from sites that routinely performed ultrasound examinations were excluded (
n = 405); the incidence of sVTE and all-cause mortality decreased to 0.86 % (95 % CI 0.62, 1.16) in the remaining 4887 patients (1.12 % in TKR patients [95 % CI 0.74, 1.63]; and 0.60 % in THR patients [95 % CI 0.34, 1.00]). There was a significant effect of diagnosis using routine ultrasound on the reported incidence of sVTE and all-cause mortality (odds ratio 3.83 for sites using vs not using routine ultrasound [95 % CI 1.96, 6.99;
p = 0.0002]). This difference was driven by a lower incidence of patients with distal symptomatic DVT. Diagnosis using routine ultrasound may therefore explain the higher event rates in TKR patients compared with the rate obtained in phase 3 clinical trials.
It is unclear whether the low incidence of symptomatic VTE and all-cause mortality in THR patients with a history of VTE (0 % compared with 0.56 % in those without a history of VTE) has a medical explanation. It may be a ‘chance’ finding due to sample size. Only 73 patients with a history of VTE underwent hip surgery, so even if the risk increase (due to history of VTE) were as high as that observed for the knee surgery group, we would only expect two to three patients with an event in the THR patients with history of VTE. The use of duplex sonography of the calf in some study centres may also have contributed to the relatively greater difference in the incidence of the primary efficacy outcome between the TKR patients with and without a history of VTE (7.96 % vs 1.27 %). Asymptomatic VTE could be interpreted as symptomatic VTE in knee surgery because of calf haematoma and oedema, but in hip surgery, there is no calf haematoma and thus no possibility to attribute an asymptomatic DVT to a symptomatic event.
Further potential limitations of this observational study include a lack of control group, diagnosis based only on the investigator’s clinical judgement (which could have resulted in false positive or negative findings) and a possible bias in the selection of patients. The expected sample size ≥400 patients was not reached in all subgroups, therefore, the power to detect significant differences in rates on a descriptive basis was decreased, particularly for the CAD and CHF groups. However, the point estimates did not indicate any relevant increase in risk in these groups. A formal comparison of outcomes between THR and TKR surgery groups was not considered worthwhile because of significant differences in patient characteristics and the expectation of different outcomes based on the body of previously published findings.
It is important to note that the favourable AE profile in this observational study was comparable with that obtained in the dabigatran phase 3 trials. No increased ACS signal was detected with dabigatran compared with enoxaparin during or after treatment in a pooled analysis of pivotal phase 3 dabigatran trials [
9]. Results from this observational study confirm the low incidence of ACS during dabigatran treatment in a real-world orthopaedic setting.
Competing interests
Nadia Rosencher has received personal fees and fees paid to employing institution from Bayer, Sanofi, GSK, Boehringer Ingelheim, Pfizer and Bristol-Myers Squibb, outside the submitted work. Simon Frostick has received grants from Biomet Inc. and fees paid to employing institution from Boehringer Ingelheim and Biomet Inc, outside the submitted work. Charles M. Samama reports personal fees and nonfinancial support from Boehringer Ingelheim, Bayer and Daichi and personal fees from Bristol-Myers Squibb. Martin Feuring, Martina Brueckmann and Eva Kleine are employees of Boehringer Ingelheim. Andreas Clemens was an employee of Boehringer Ingelheim at the time of the study and is currently a full time employee at Novartis Pharma AG.
Authors’ contributions
All of the authors have contributed to the design of the study or the acquisition, analysis, or interpretation of the data; drafting or reviewing and revising the manuscript content; and approving the final version.