Highlights

  • Dose adjustment of direct oral anticoagulants (DOACs) is not required in the setting of acute pulmonary embolism (PE) treatment according to the manufacturer’s labelling

  • Very few data exist currently on the rate and clinical implications of non-recommended DOAC dose prescription in real-like practice

  • Among 656 patients discharge with DOACs, the rate of inappropriate DOAC dose prescription was 4.3% (n = 28) in our study

  • The rate of the primary endpoint of all-cause death, recurrent venous thromboembolism, major bleeding, and chronic venous thromboembolism was significantly higher in the non-recommended DOAC dose group (25% vs. 6.1%, respectively, p < 0.001), mainly driven by an excess of bleeding

  • Implementation of further strategies, such as correcting bleeding risk factors, or implanting an IVC filter in patients with a very high bleeding risk, might be safer than reducing the DOAC dose

Introduction

Pulmonary embolism (PE) is a relatively common disease, with an incidence ranging from 60 to 112 per 100,000 inhabitants of the United States [1], and is the third leading cardiovascular cause of death [2]. The mainstay treatment is anticoagulation, adapted to patient contraindications, to prevent early death and recurrent venous thromboembolism (VTE) [3]. For over 5 decades, PE antithrombotic treatment strategy included the use of a parenteral anticoagulant in combination with a vitamin K antagonist (VKA) until the target International Normalized Ratio of 2.0–3.0 was reached. The results of randomized trials using direct oral anticoagulants (DOACs) for the treatment of acute PE indicate that these agents are non-inferior in terms of efficacy at preventing recurrence of VTE and also safer, particularly in terms of major bleeding, than the conventional heparin/VKA regimen [4,5,6,7]. The 2016 ACCP guidelines recommend henceforth DOACs rather than VKA to treat acute VTE [3].

Contrary to stroke prevention in atrial fibrillation (AF) [8], dose adjustment is not required in the setting of acute PE treatment according to the manufacturer’s labelling, beyond the contraindication in patients with a creatinine clearance (CrCl) < 30 mL/min. In PE studies, the patients rate with criteria for reducing dosage according to AF treatment labelling range from 10 to 15% for patients aged > 75 years and was 7.9% for patients with renal function impairment (i.e., CrCl 30–45 mL/min) [4, 6]. The frequency of prescription of non-recommended DOAC doses, and its impact on outcomes, is poorly described in real-life practice [9,10,11,12].

Therefore, the present study aimed to investigate patterns of prescription of non-recommended DOAC doses for the treatment of acute PE, and the associated risk of 6-month adverse events.

Materials and methods

Study design

We designed a prospective, observational study, based on a multicenter registry including 11 departments in 5 centers, and covering 7 different clinical specialties [cardiology (n = 5 departments), vascular medicine (n = 1), internal medicine/geriatrics (n = 1), respiratory diseases (n = 1), oncology (n = 1), critical care (n = 1), and orthopedic surgery (n = 1)] in the region of Burgundy Franche Comte, France. For this study, only patients treated with DOACs were considered. The study period encompassed September 2012 through April 2017. We chose September 2012 as the starting point of this study, corresponding to the launch date of rivaroxaban, the first DOAC to be approved on the French market for the dedicated indication of acute PE treatment. Apixaban was authorized in this indication from April 2015 onwards.

Inclusion criteria were all patients aged 18 years or older with confirmed PE by computed tomography pulmonary angiography (CTPA), or ventilation–perfusion (V–Q) scan, discharge from hospital with a DOAC therapy. Patients were considered to receive the recommended therapy if DOACs were prescribed as follows: rivaroxaban 15 mg twice daily for 21 days (initial dosing) and then 20 mg once daily (long-term dosing); or apixaban 10 mg twice daily for 7 days and then 5 mg twice daily (rivaroxaban and apixaban were the only DOACs available on the French market in the indication of PE treatment at the time of this study).

Data collection

The study was conducted in accordance with the Declaration of Helsinki. Institutional Review Board approval was obtained. Informed consent was obtained from each study participant.

Patients were identified by requests to the hospital informatics database (Programme de Médicalisation des Systèmes d’Information, PMSI) using the International Classification of Disease (ICD-10) PE diagnosis codes (code I260 and I269) [13]. Research physicians performed a review of medical records at each site and abstracted data using a standard data collection form. Technicians entered data into a dedicated database for de-identification of personal information and for partial data validation. The research technicians were also asked to ensure that recruitment was consecutive.

Baseline characteristics including socio-demographic data, clinical data, imaging findings, and biological data were recorded prospectively. In-hospital adverse events including death, causes of death, recurrent VTE and bleeding were also recorded. Six months after the diagnosis of PE, patients returned to the hospital for a follow-up visit to record clinical status, anticoagulation treatments and outcomes (i.e., death, cause of death, recurrent VTE, and bleeding since hospital discharge). If the patient failed to attend the hospital visit, the following sequential procedure was followed: telephone interview with the patient or the patient’s family, consultation of the hospitalization records since discharge, telephone contact with the patient’s general practitioner or other treating physician, and finally, consultation of the national death registry. Six-month follow-up was 100% complete.

Clinical endpoint and definitions of outcomes

The primary endpoint was a composite of death from any cause, recurrent VTE, major bleeding, and chronic thromboembolic pulmonary hypertension (CTEPH). Prespecified secondary endpoints included each component of the primary end point (i.e., all-cause death, recurrent VTE, major bleeding, and CTEPH). All adverse events were adjudicated by an independent committee.

Recurrent VTE includes PE confirmed by visualization of new filling defects on V–Q scan or CTPA scan, or diagnosis of DVT on compression ultrasonography. Bleeding events were classified according to the International Society of Thrombosis and Haemostasis criteria for major bleeding [14]. CTEPH was defined according to the European Society of Cardiology (ESC) guidelines with a mean pulmonary artery pressure ≥ 25 mmHg with pulmonary arterial wedge pressure ≤ 15 mmHg on cardiac catheterization, and pulmonary artery obstruction seen by V–Q scan or CTPA [15].

Statistical analysis

Normally-distributed variables are expressed as mean ± SD and categorical variables as number and percentage (%). Unadjusted differences were compared using the Chi square test for categorical variables and Student’s t test for continuous variables. Predictors of non-recommended DOAC dose prescription were identified using multivariate logistic regression, including all baseline characteristics with a p value < 0.10 by univariate analysis. Results are reported as odds ratios (OR) with associated 95% confidence interval (CI) and p value. We compared follow-up outcomes between recommended and non-recommended DOAC dose groups by constructing a multivariate Cox models adjusted for any baseline characteristics and in-hospital course with a p value < 0.10 by univariate analysis. Results are reported as hazard ratios (HR) with associated 95% CI and p value. The Kaplan–Meier method was used to draw survival curves at 6 months. All multivariate models included site of care as a random effect to account for patient clustering within sites. Statistical tests were two-sided and considered significant if they yielded a p value < 0.05. Analyses were performed using SAS, version 9.4 (SAS Institute Inc., Cary, NC, USA).

Results

In total, 1155 patients were admitted to the participating hospitals with an objectively confirmed diagnosis of acute PE between 09/2012 and 04/2017. Overall, 71 patients (6.1%) died during the hospital stay. Among the survivors, 656 patients (60.5%) were treated with DOACs [rivaroxaban: 614 patients (93.6%), apixaban 42 patients (6.4%)] and composed the study population.

The baseline characteristics of the study population are shown in Table 1. Mean age was 63.6 ± 17.9 years, 46.9% were males. Twenty patients (3.0%) had high-risk PE, 143 (21.8%) had intermediate-high-risk PE, 188 (28.6%) had intermediate-low-risk PE, and 305 (46.5%) had low-risk PE as defined by the ESC guidelines [16]. A total of 628 patients (95.7%) were treated with the recommended DOAC dose, and 28 patients (4.3%) had non-recommended DOAC dosing. Patients with and without the recommended DOAC dose differed significantly in terms of pre-existing medical conditions, and in relevant clinical, echographic and laboratory parameters (Table 1). Patients with non-recommended DOAC dose were significantly older, and more frequently presented coronary artery disease (CAD) and severe renal insufficiency as defined by a creatinine clearance below 30 mL/min. New York Heart Association functional status and arterial oxyhemoglobin saturation were also poorer in this group of patients. Creatinine clearance was lower and there was more frequent right ventricular dysfunction at admission in the non-recommended DOAC dose group. These findings resulted in a significantly higher sPESI score in the group of patients treated with non-recommended DOAC dose (Table 1). The rate of prior bleeding did not differ between both groups.

Table 1 Baseline characteristics of the study population stratified by prescription of recommended or non-recommended direct oral anticoagulant dose for the treatment of acute pulmonary embolism
Full size table

During the hospital course, reperfusion therapy including thrombolysis and surgical embolectomy, and extracorporeal membrane oxygenation and inferior vena cava filter implantation were not significantly different between recommended and non-recommended DOACs dose groups; in-hospital adverse events were also similar (Table 2).

Table 2 In-hospital course stratified by stratified by prescription of recommended or non-recommended direct oral anticoagulant dose for the treatment of acute pulmonary embolism
Full size table

At discharge, all the non-recommended DOAC dose prescriptions were under-dosed, as compared to the drug labelling. Five patients on rivaroxaban (0.8%) and seven patients on apixaban (16.6%) did not receive any recommended initial dose. Among the five patients who did not receive an initial dose of rivaroxaban, four were directly treated with the appropriate long-term dose (i.e., rivaroxaban 20 mg once daily) and one patient received a lower long-term dose (i.e., 15 mg once daily). Among the seven patients who did not received an initial dose of apixaban four patients received directly a non-recommended dose (i.e., 2.5 mg twice daily) and three patients received a recommended long-term dose (i.e., 5 mg twice daily). Regarding the long-term dose, 17 patients on rivaroxaban (2.7%) were prescribed 15 mg once daily, including one patient who did not receive an initial dose of rivaroxaban; 4 patients (9.5%) who initially did not receive an initial dose, were also treated with a non-recommended long-term dose of 2.5 mg twice daily. The rate of concomitant antiplatelet prescription was higher in the non-recommended DOAC dose group as compared to the recommended group (Table 2).

By multivariate analysis, age > 70 years (OR 2.9; 95% CI 1.2–7.4), a history of CAD (OR 2.7; 95% CI 1.05–6.7), creatinine clearance < 50 mL/min (OR 2.5; 95% CI 1.05–6.1), and concomitant aspirin therapy (OR 4.1; 95% CI 1.2–13.5) were independent factors associated with non-recommended DOAC prescription (C-statistic: 0.82; Hosmer Lemeshow test: 0.50) (Fig. 1).

Fig. 1
figure 1

Clinical factors independently associated with the prescription of non-recommended direct oral anticoagulant dose for the treatment of acute pulmonary embolism. DOAC direct oral anticoagulant, CAD coronary artery disease, CrCl creatinine clearance

Full size image

Time from PE diagnosis to follow-up was 171 ± 11 days in the compliant group and 168 ± 9 days in the non-compliant group (p = 0.87). During the 6 months of follow up, 16 patients died: 1 disseminated cancer, 1 fatal PE, 1 major bleeding, 5 cardiovascular deaths and 8 from other causes. Ten patients presented recurrent VTE; 11 patients had one or more major bleeding events (4 intracranial, 3 gastrointestinal, 3 muscular, 2 pericardial and 1 urological); 12 patients presented CTEPH.

The primary composite endpoint occurred in 7 patients (25%) in the non-recommended dose group and in 38 patients (6.1%) in the recommended dose group (OR 3.19; 95% CI 1.16–8.70; p = 0.02) (Table 3). The corresponding Kaplan–Meier curves representing 6-month event free survival estimates are shown in Fig. 2. The higher primary endpoint rate observed in the non-recommended dose group was driven by a significantly higher rate of major bleeding (7.1 vs. 1.4%; p = 0.008), and a non-significant trend toward higher rate of death (7.1 vs. 2.2%, p = 0.23), recurrent VTE (3.6 vs. 1.4%, p = 0.31), and CTEPH (7.1 vs. 1.6%, p = 0.32) (Table 3).

Table 3 Adjusted outcomes at 6-month follow-up stratified by the prescription of recommended or non-recommended direct oral anticoagulant dose for the treatment of acute pulmonary embolism
Full size table
Fig. 2
figure 2

Kaplan–Meier curves of 6-month survival free from the combined primary end-point in the recommended and non-recommended direct oral anticoagulant dose prescription groups (as per manufacturer’s labelling) for the treatment of acute pulmonary embolism. HR hazard ratio, CI confidence interval

Full size image

Discussion

The present study, from a prospective multicenter multidisciplinary registry, indicates that 4.3% of patients treated with DOACs for the treatment of acute PE did not receive an appropriate dose at discharge. Older age, CAD on aspirin, and renal function impairment were factors associated with off-label DOAC dose prescription by providers. The rate of adverse events at 6 months after PE diagnosis was significantly higher in the group of patients inappropriately treated with DOACs as compared to those who received a DOAC dose in line with the manufacturer’s labeling. This higher rate of adverse events was driven by a significantly higher rate of major bleeding and by numerically albeit non-significantly higher rates of death, recurrent VTE, and CTEPH.

The rate of non-recommended DOAC dose prescription observed in our study was lower than previously reported, namely from 10 to 18% in other PE studies [9, 11, 12]. In the setting of AF, data from a large U.S. administrative database showed that 43% of patients with a renal indication for DOAC dose reduction were potentially overdosed and 13.3% of patients with no renal indication for DOAC dose reduction were potentially underdosed [17]. The inappropriate-low-dose rates ranged between 19.7 and 26.7% based on the CrCl, age, and weight criteria for dose adaptation in a registry including 1689 AF patients treated with DOACs [18].

Our results regarding outcomes in the non-recommended DOAC dose group are in line with those recently reported from the RIETE registry including 1511 patients, with an 18% prescription rate of off-label DOAC doses. In this registry, the authors showed a significantly increased rate of recurrent VTE (HR 10.5; 95% CI 1.28–85.9) associated with a numerically but non-significantly higher rate of bleeding and death in inappropriate DOAC dose patients [9]. Tran et al. reported likewise in 26 of 261 PE patients inappropriately treated with DOACs (10%), no death, 3 recurrent VTE (11.5%), 3 major bleeding (11.5%), and 3 clinically relevant non-major bleeding (11.5%) [11]. Regarding the relationship between non-recommended DOAC dose according to manufacturer’s labelling and outcomes in the setting of AF, Yao et al. observed an increased risk of bleeding in patients with renal impairment and no reduction of the DOAC dose (HR 2.19; 95% CI 1.07–4.46) and an increased risk of stroke in patients with normal renal function who received a reduced DOAC dose (HR 4.87; 95% CI 1.30–18.26) [17].

The factors we identified as related to prescription of off-label DOAC doses (old age, renal function impairment, concomitant antiplatelet therapy) are well known to increase the risk of bleeding. In the study that developed the HAS–BLED score, old age was significantly related to high bleeding risk (OR 2.66, 95% CI 1.33–5.32), as was renal function impairment (OR 2.86; 95% CI 1.33–6.18) [19]. Aspirin use was shown to be a risk factor for major bleeding in association with warfarin (HR 1.83; 95% CI 1.72–1.96) [20]. Our findings emphasize the provider’s concerns about DOAC prescription and bleeding risk, at least in some cases. However, we observed that attempts to limit the effect of DOAC by reducing the dose did not lead to a reduction in the adjusted risk of bleeding. Rather, treatable bleeding risk factors such as concomitant antiplatelet therapy should be identified and corrected, and as a last resort, refraining from prescribing anticoagulants and implanting an inferior vena cava (IVC) filter in very high bleeding risk patients could be alternative options [21]. The prescription of VKA rather than DOACs to reduce the risk of bleeding is likely an inappropriate strategy. In pivotal randomized trials of DOACs versus VKA for the treatment of acute PE, DOACs were associated with significantly less bleeding [4,5,6,7]. Moreover, DOACs were safer than VKA with respect to intracranial hemorrhage in patients receiving concomitant antiplatelet therapy in a recently published meta-analysis (HR 0.38; 95% CI 0.26–0.56) [22].

In total, our results underscore the difficulty of providing treatment in compliance with guidelines and manufacturer’s drug labeling,and present an important opportunity for multidisciplinary quality improvement to improve patient safety. The implementation of an on-site multidisciplinary pulmonary embolism response team, which engages multiple specialists to deliver evidence-based, organized, and efficient care to PE patients, could improve compliance with PE guidelines and patient safety [23].

Strength and limitations

This study was based on a large prospective multicenter registry. The outcomes were adjudicated, and follow-up was 100% complete. Moreover, we actively recruited patients through a range of possible ports of entry of such patients into the hospital, in a range of different medical specialties likely to encounter patients with PE.

However, our study has some limitations that deserve to be underlined. We do not have any individual data regarding the reasons motivating provider’s decisions to prescribe a treatment that was not in accordance with the drug labelling. The rate of inappropriate DOAC dose prescription was low and the event rates were low in our registry; therefore, the findings should be viewed as hypothesis-generating and need to be confirmed by future studies.

Conclusion

To conclude, the rate of empiric dose reduction of DOACs was low in our real-life registry but was associated with 6-month adverse events, especially in high-risk bleeding patients. Implementation of further strategies, such as correcting bleeding risk factors, or implanting an IVC filter in patients with a very high-risk of bleeding, could be safer than reduction of the DOAC dose. More robust data are warranted to confirm these results.