Introduction
Venous thromboembolism (VTE) is one of the most common acute cardiovascular disorders,
along with acute coronary syndromes and stroke.[1] Annual incidence rates are estimated at 39 to 115 per 100,000 people for pulmonary
embolism (PE) and 53 to 162 per 100,000 people for deep venous thrombosis (DVT).[2]
[3] The management of VTE substantially changed after the introduction of direct oral
anticoagulants (DOACs), which were shown to be at least as effective but safer than
vitamin K antagonists (VKAs).[4]
The initial 3 to 6 months after a VTE event is usually considered to be a period of
high risk for VTE recurrence,[5]
[6] and it is generally covered by anticoagulant treatment that is often discontinued
thereafter. The discontinuation of anticoagulant treatment is associated with an increased
risk of VTE recurrence. The risk of bleeding on anticoagulant treatment remains relatively
stable. Consequently, assessment of the benefits and risks of anticoagulant treatment
extended beyond 6 months remains rather complex.[5]
“Edoxaban Treatment in routiNe clinical prActice in patients with Venous ThromboEmbolism
in Europe” (ETNA-VTE Europe) is a noninterventional study of unselected patients with an acute
VTE that occurred in 2 weeks prior to the study inclusion and who were treated with
edoxaban for up to 18 months.[7]
[8]
[9] In the present analysis, we sought to assess the annualized rates of (1) recurrent
VTE, (2) bleeding events, and (3) VTE-relevant clinical events. The results were assessed
in relation to the actual duration of edoxaban treatment received. In addition, we
compared the results with the results of data obtained in the Prevention of Thrombolic
Events–European Registry in Venous Thromboembolism (PREFER-VTE) disease registry and
the edoxaban versus warfarin for the treatment of symptomatic venous thromboembolism
(HOKUSAI-VTE) clinical study.
Methods
ETNA-VTE Europe is a prospective, noninterventional, post-authorization safety study
conducted in hospitals and outpatient clinics in eight European countries (Austria,
Belgium, Germany, Ireland, Italy, the Netherlands, Switzerland, and the United Kingdom).
The study design,[8] characteristics of the patient population,[9] and benefits and risks of edoxaban after the first 3 months of treatment[7] have been previously described. Approval was obtained from the responsible Ethics
Committees or Institutional Review Boards, and compliance with the Declaration of
Helsinki was ensured throughout the study. Additionally, the study design was discussed
and approved by the European Medicines Agency (EMA). Patients provided written informed
consent prior to their inclusion in the study.
Patient Population
Consecutive patients with symptomatic or incidental first episode or recurrent proximal
or distal DVT and/or PE that had occurred ≤2 weeks prior to enrollment and for whom
a decision to administer edoxaban was made at the discretion of the treating physician
were included in the study. The dose of edoxaban (30 or 60 mg) administered and the
duration of treatment were left to the discretion of the treating physician. Any concomitant
treatment was allowed, and changes to medications were unrestricted. Exclusion criteria
were a lack of written informed consent and participation in a simultaneous interventional
study.
Data Collection
Data were collected from medical records and telephone calls between the patients
and their respective treating physicians. Since heparin lead-in is required prior
to edoxaban treatment as per Summary of Product Characteristics (SmPC) guidelines,
the baseline was defined as the first day of heparin/fondaparinux administration after
the index acute VTE event. The first data entry was completed within 2 weeks after
the VTE event. Patient data were collected and documented throughout 18 months of
the study during each visit at months 1, 3, 6, 12, and 18. At the last patient out
of each country, the last known patient status was documented for each patient enrolled
in the respective country. Also, the final assessment data, including recurrence of
symptomatic VTE (defined as any PE-only, DVT-only, and PE with DVT that occurred after
the index event), bleeding, and death (all-cause, VTE-related, or CV-related), were
documented at any follow-up data collection point or at premature study termination.
Major bleeding was defined as fatal or symptomatic bleeding in a critical area/organ
or causing a ≥2 g/dL fall in hemoglobin and/or ≥6.0% fall in hematocrit; clinically
relevant nonmajor bleeding (CRNMB) was defined as requiring medical attention but
not fulfilling major bleeding criteria; minor bleeding was defined as any other bleeding
that did not fulfil the criteria for major bleeding/CRNMB.[8]
[9] Key outcome variables, such as recurrent VTE, bleeding events (major bleeding and
CRNMB), and deaths, were adjudicated by an independent committee.[8]
Statistics
Patient Disposition
All patients with signed informed consent with trustworthy data comprised all documented
patient sets (APS). The baseline analysis set (BAS) comprised APS after excluding
patients based on the exclusion criteria. Finally, the full analysis set (FAS) included
all BAS patients except those without any follow-up data (including the last assessment).
All analyses presented in this paper were conducted using data from FAS.
For categorical variables, the number of patients and corresponding percentages are
presented, and medians and interquartile ranges (IQRs) are shown. Imputation was performed
for specific incomplete data related to the exposure and outcome events. Imputation
followed a worst-case approach, i.e., for events, the earliest possible occurrence
was assumed and for exposure, the longest possible intake (to not underestimate the
rate of events that occurred on edoxaban). For all other data, no imputation was performed,
and data was analyzed and presented as if they were recorded in the database.
Outcome rates were calculated as the number of patients with the specific event divided
by the number of patients in the analysis group. Annualized outcome rates (on edoxaban)
were calculated as the number of patients with at least one event (on edoxaban) within
18 months divided by the sum of all individual patient time (on edoxaban) until the
first event or censoring within 18 months multiplied by 100 (years). Outcomes were
also analyzed during the whole observational period, considering on-treatment events
as those that occurred during edoxaban treatment or up to 3 days after its discontinuation.
The statistical analyses were performed using SAS version 9.4 and were descriptive
and exploratory in nature.
Comparison of ETNA-VTE with PREFER-VTE and HOKUSAI-VTE
In this study, we compared the results from the ETNA-VTE study with PREFER-VTE and
HOKUSAI-VTE clinical studies. PREFER-VTE was a prospective, observational, multicenter
registry conducted in seven European countries to assess the characteristics and management
of patients with VTE with a follow-up of 12 months from baseline. HOKUSAI-VTE was
a multinational, multicenter, randomized, double-blinded, double-dummy, parallel-group,
phase-three study on VTE patients with 1-year follow-up.
As both PREFER-VTE and HOKUSAI-VTE had an observation period of 1 year, for the comparison
with ETNA-VTE, the corresponding summary statistics, frequencies, and rates, respectively,
were taken after 12 months. The 95% CIs were also provided. These results were compared
with the results from the PREFER-VTE and HOKUSAI-VTE registries. These comparisons
were purely descriptive/explorative, and there was no joint model that compared the
different treatments directly. Furthermore, although every effort was made to align
the endpoint definitions of the ETNA-VTE as much as possible to the endpoint definitions
of the PREFER-VTE and HOKUSAI-VTE, the compared endpoints were highly harmonized but
not necessarily identical.
Results
A total of 2,809 patients (APS) were enrolled in ETNA-VTE Europe, of which 165 (5.9%)
were excluded for the reasons shown in [Fig. 1], resulting in 2,644 patients in the FAS.
Fig. 1 Patient disposition. Enrolled patients (with informed consent form [ICF] signed and
trustworthy data) may have been excluded from the baseline analysis set (BAS) for
one or more of the criteria stated. aTreatment of acute venous thromboembolism (VTE) (either by edoxaban or by heparin
lead-in followed by edoxaban) starts 1 day after the baseline data collection point
or earlier and heparin lead-in (if given) lasts maximum 30 days.
Patient and Treatment Characteristics
Patients had a median age of 65.0 years, and 46.6% were female ([Table 1]). Hypertension was the most frequent comorbid condition (43.5%), followed by dyslipidemia
(20.2%), chronic venous insufficiency (11.4%), diabetes (11.3%), and chronic obstructive
pulmonary disease (COPD; 6.7%). Overall, 22.8% patients had a history of VTE. Prolonged
immobilization (15.5%) and history of major surgery or trauma (13.8%) were the most
frequent risk factors for VTE.
Table 1
Patient demographics and clinical characteristics by type of VTE
|
ETNA-VTE overall
(N = 2,644)
|
DVT only
(N = 1,540)
|
PE with DVT
(N = 520)
|
PE only
(N = 584)
|
|
Age (years)
|
65.0 (52.0, 76.0)
|
64.0 (51.0, 75.0)
|
65.0 (53.0, 76.0)
|
66.0 (54.0, 76.0)
|
|
<65 years old, n (%)
|
1,312 (49.6)
|
789 (51.2)
|
256 (49.2)
|
267 (45.7)
|
|
≥65 and <75 years old, n (%)
|
593 (22.4)
|
330 (21.4)
|
117 (22.5)
|
146 (25.0)
|
|
≥75 years old, n (%)
|
739 (28.0)
|
421 (27.3)
|
147 (28.3)
|
171 (29.3)
|
|
Female (gender), n (%)
|
1,231 (46.6)
|
696 (45.2)
|
241 (46.3)
|
294 (50.3)
|
|
Weight (kg), n (%)
|
80.0 (70.0, 92.0)
|
80.0 (70.0, 90.0)
|
82.0 (71.0, 94.0)
|
80.0 (69.0, 95.0)
|
|
Body mass index (kg/m2)
|
27.20 (24.5, 30.5)
|
27.10 (24.6, 30.1)
|
27.50 (24.6, 31.1)
|
27.20 (24.2, 31.6)
|
|
Frailty (as judged by investigator), n (%)
|
326 (12.8)
|
176 (12.0)
|
77 (15.2)
|
73 (13.0)
|
|
Diabetes mellitus, n (%)
|
294 (11.3)
|
166 (11.0)
|
60 (11.6)
|
68 (11.7)
|
|
Dyslipidemia, n (%)
|
510 (20.2)
|
281 (19.4)
|
114 (22.7)
|
115 (20.3)
|
|
Hypertension, n (%)
|
1130 (43.5)
|
618 (41.1)
|
245 (47.6)
|
267 (46.0)
|
|
Chronic venous insufficiency, n (%)
|
291 (11.4)
|
210 (14.2)
|
50 (9.8)
|
31 (5.5)
|
|
COPD, n (%)
|
175 (6.7)
|
84 (5.6)
|
28 (5.4)
|
63 (11.0)
|
|
Renal impairment
|
955 (37.2)
|
529 (35.9)
|
197 (38.0)
|
229 (39.6)
|
|
History of VTEa, n (%)
|
|
|
|
|
|
No former history
|
2,035 (77.2)
|
1,166 (75.9)
|
409 (78.8)
|
460 (79.4)
|
|
DVT only
|
426 (16.2)
|
309 (20.1)
|
62 (11.9)
|
55 (9.5)
|
|
PE only
|
98 (3.7)
|
27 (1.8)
|
19 (3.7)
|
52 (9.0)
|
|
PE with DVT
|
103 (3.9)
|
47 (3.1)
|
36 (6.9)
|
20 (3.5)
|
|
Puerperium, n (%)
|
9 (0.3)
|
6 (0.4)
|
0 (0.0)
|
3 (0.5)
|
|
Prolonged immobilization, n (%)
|
401 (15.5)
|
230 (15.3)
|
95 (18.6)
|
76 (13.2)
|
|
>5 days in bed, n (%)
|
218 (8.5)
|
127 (8.5)
|
41 (8.0)
|
50 (8.7)
|
|
History of major surgery or trauma, n (%)
|
359 (13.8)
|
199 (13.2)
|
74 (14.3)
|
86 (15.0)
|
|
Known thrombophilic conditions, n (%)
|
111 (4.6)
|
74 (5.3)
|
23 (4.9)
|
14 (2.6)
|
Abbreviations: COPD, chronic obstructive pulmonary disease; DVT, deep venous thrombosis;
FU, follow-up; PE, pulmonary embolism; VTE, venous thromboembolism.
Note: aPercentage calculation can sum to >100% because subjects can fall in more than one
category.
Among all 2,644 patients, 58.2% (n = 1540) of the patients had a DVT (only) as the index event, with the remaining suffering
a PE only (n = 584; 22.1%) or PE with DVT (n = 520 [19.7%]; [Table 1]). Patient characteristics were generally similar in patients across groups with
minor differences in age, weight, BMI, and history of hypertension. Chronic venous
insufficiency (14.2% vs. 9.8% vs. 5.5%) and a history of DVT only (20.1% vs. 11.9%
vs. 9.5%) were more common in patients with a DVT only, COPD (11.0% vs. 5.6% vs. 5.4%)
was common in PE only, whereas prior PE with DVT (6.9% vs. 3.5% and 3.1% respectively)
was more prevalent in patients with PE with DVT.
Out of the total patients (n = 2,644), 1,985 patients did not discontinue edoxaban treatment permanently at 6
months, and 612 patients discontinued treatment permanently ([Supplementary Table S1], available in the online version). Patients with permanent discontinuation were
younger (58.6 ± 17.6 years vs. 64.0 ± 15.2 years), had less comorbidity burden, and
had less frequent formal history of VTE (86.6% vs. 73.9%). However, it has a higher
prevalence of VTE-specific causes (like prolonged immobilization, more than 5 days
in bed, or a history of major surgery or trauma). The major reason for the permanent
discontinuation of edoxaban was that the treatment was no longer necessary in most
patients, followed by the occurrence of adverse drug reaction (ADR) or a clinical
event ([Supplementary Table S2], available in the online version).
Regarding the dose of edoxaban at the initiation of treatment, 87.7% of the patients
received the 60-mg dose, and 12.3% received the 30-mg dose of edoxaban. The median
treatment duration was 50.6 weeks (IQR: 23.4–77.7) ([Table 2]). Treatment was ongoing at month 3 in 89.4% of the patients, 70.2% at month 6, and
48.5% at month 12. A total of 27.2% of patients remained on treatment at the end of
the study period (18 months). Treatment was significantly longer in patients with
PE with DVT than in patients with DVT only.
Table 2
Actual duration of treatment with edoxaban
|
ETNA-VTE overall
(N = 2,644)
|
DVT only
(N = 1,540)
|
PE with DVT
(N = 520)
|
PE only
(N = 584)
|
|
Patient on edoxaban
|
|
Month 1 ongoing, n (%)
|
2,527 (96.3)
|
1,484 (96.9)
|
497 (96.9)
|
546 (94.5)
|
|
Month 3 ongoing, n (%)
|
2,346 (89.4)
|
1,354 (88.4)
|
479 (93.4)
|
513 (88.8)
|
|
Month 6 ongoing, n (%)
|
1,842 (70.2)
|
994 (64.9)
|
421 (82.1)
|
427 (73.9)
|
|
Month 12 ongoing, n (%)
|
1,272 (48.5)
|
689 (45.0)
|
288 (56.1)
|
295 (51.0)
|
|
Month 18 ongoing, n (%)
|
713 (27.2)
|
377 (24.6)
|
153 (29.8)
|
183 (31.7)
|
|
Subjects off edoxaban at 18 months
|
1,910 (72.8)
|
1,155 (75.4)
|
360 (70.2)
|
395 (68.3)
|
|
Median exposure, weeks
|
50.6 (23.4, 77.7)
|
41.4 (16.3, 77.6)
|
53.7 (28.6, 77.7)
|
52.9 (25.3, 78.0)
|
Abbreviations: DVT, deep venous thrombosis; PE, pulmonary embolism; VTE, venous thromboembolism.
The results from ETNA-VTE patients (FAS; n = 2,644) were compared with PREFER-VTE patients treated with VKA (n = 917) and non-vitamin K antagonist oral anticoagulant (NOAC) (n = 501). The gender distribution and medical history in the ETNA-VTE population were
largely comparable to the PREFER-VTE population ([Supplementary Table S3], available in the online version). The rate of chronic venous insufficiency was
higher in the PREFER-VTE VKA population compared with ETNA-VTE (17.7% vs. 11.4%).
More subjects in ETNA-VTE were documented with DVT only (58.2%) than in PREFER-VTE
VKA (50.0%).
Additionally, we compared results from ETNA-VTE edoxaban patients (FAS; n = 2,644) with HOKUSAI-VTE edoxaban (n = 2,391), and warfarin patients (n = 2,398), only from Europe indicated to treat recurrent VTE.
Compared with the HOKUSAI-VTE population treated with edoxaban, patients in ETNA-VTE
were older (median age of 65.0 vs. 58.0 years), were less often males (53.4% vs. 60.5%),
and had higher rates of chronic venous insufficiency (11.4% vs. 2.3%) ([Supplementary Table S4], available in the online version). Nominally, more subjects underwent prolonged
immobilization in ETNA-VTE (15.5%) than in HOKUSAI-VTE (5.9%). In contrast, there
were more subjects with a history of major surgery or trauma in HOKUSAI-VTE than there
were in ETNA-VTE (17.1% vs. 13.8%, respectively). More ETNA-VTE subjects had a history
of VTE than in HOKUSAI-VTE edoxaban, with higher rates seen for DVT only in particular
(16.2% vs. 7.6%).
VTE Recurrence
The cumulative incidence of recurrent VTEs is displayed in the upper panel of [Fig. 2]. Out of 100 patients with recurrent VTE events, 37 occurred while being on edoxaban
treatment. Recurrence was higher in patients with PE with or without DVT than in patients
with DVT only ([Fig. 2], lower panel). Overall, 100 patients (3.8%) suffered from recurrent VTE events ([Table 3]), corresponding to an annualized rate of 2.7%/year. Of these, 37 (1.4%) had a recurrent
VTE during edoxaban treatment (annualized rate: 1.6%/year). When comparing the VTE
recurrence every 6 months in patients who continued with anticoagulant treatment for
18 months, there is an increase in the VTE recurrence (annualized rate baseline to
6 months: 0.4%/year; 6 to 12 months: 1.2%/year; 12 to 18 months: 2.8%/year), which
is driven by DVT only ([Supplementary Table S5], available in the online version). Comparing VTE recurrence from month 6 to month
18 in patients who had permanently discontinued edoxaban within the first 6 months
compared with events on edoxaban in patients who had not permanently discontinued
edoxaban within the first 6 months showed higher annual rates in patients who had
permanent discontinuation (2.6%/year vs. 1.4%/year at a hazard ratio of 1.9 [95% CI
0.96, 3.79]; [Supplementary Table S6], available in the online version).
Table 3
Overall and on-treatment clinical effectiveness of edoxaban during the 18-month study
period
|
ETNA-VTE overall
(N = 2,644)
|
DVT only
(N = 1,540)
|
PE with DVT
(N = 520)
|
PE only
(N = 584)
|
|
Number of subjects with at least one event—overall
|
|
Recurrent VTEa, n (%) [95% CI]
|
100 (3.8) [3.09, 4.58]
|
54 (3.5) [2.64, 4.55]
|
22 (4.2) [2.67, 6.34]
|
24 (4.1) [2.65, 6.05]
|
|
Recurrent DVT onlya, n (%) [95% CI]
|
61 (2.3) [1.77, 2.95]
|
43 (2.8) [2.03, 3.74]
|
11 (2.1) [1.06, 3.75]
|
7 (1.2) [0.48, 2.45]
|
|
Recurrent PE with DVTa, n (%) [95% CI]
|
7 (0.3) [0.11, 0.54]
|
2 (0.1) [0.02, 0.47]
|
5 (1.0) [0.31, 2.23]
|
0 (0.0) [0.00, 0.63]
|
|
Recurrent PE onlya, n (%) [95% CI]
|
31 (1.2) [0.80, 1.66]
|
9 (0.6) [0.27, 1.11]
|
7 (1.3) [0.54, 2.75]
|
15 (2.6) [1.44, 4.20]
|
|
Stroke, n (%) [95% CI]
|
26 (1.0) [0.64, 1.44]
|
18 (1.2) [0.69, 1.84]
|
2 (0.4) [0.05, 1.38]
|
6 (1.0) [0.38, 2.22]
|
|
Systemic embolic event, n (%) [95% CI]
|
2 (0.1) [0.01, 0.27]
|
0 (0.0) [0.00, 0.24]
|
1 (0.2) [0.00, 1.07]
|
1 (0.2) [0.00, 0.95]
|
|
Hospitalization related to CV, n (%) [95% CI]
|
253 (9.6) [8.47, 10.75]
|
125 (8.1) [6.80, 9.59]
|
59 (11.3) [8.75, 14.39]
|
69 (11.8) [9.31, 14.71]
|
|
Number of subjects with at least one event—on edoxaban treatment
|
|
Recurrent VTEa, n (%) [95% CI]
|
37 (1.4) [0.99, 1.92]
|
22 (1.4) [0.90, 2.15]
|
10 (1.9) [0.93, 3.51]
|
5 (0.9) [0.28, 1.99]
|
|
Recurrent DVT onlya, n (%) [95% CI]
|
27 (1.0) [0.67, 1.48]
|
19 (1.2) [0.74, 1.92]
|
6 (1.2) [0.42, 2.49]
|
2 (0.3) [0.04, 1.23]
|
|
Recurrent PE with DVTa, n (%) [95% CI]
|
1 (0.0) [0.00, 0.21]
|
0 (0.0) [0.00, 0.24]
|
1 (0.2) [0.00, 1.07]
|
0 (0.0) [0.00, 0.63]
|
|
Recurrent PE onlya, n (%) [95% CI]
|
9 (0.3) [0.16, 0.65]
|
3 (0.2) [0.04, 0.57]
|
4 (0.8) [0.21, 1.96]
|
2 (0.3) [0.04, 1.23]
|
|
Stroke, n (%) [95% CI]
|
15 (0.6) [0.32, 0.93]
|
8 (0.5) [0.22, 1.02]
|
2 (0.4) [0.05, 1.38]
|
5 (0.9) [0.28, 1.99]
|
|
Systemic embolic event, n (%) [95% CI]
|
1 (0.0) [0.00, 0.21]
|
0 (0.0) [0.00, 0.24]
|
1 (0.2) [0.00, 1.07]
|
0 (0.0) [0.00, 0.63]
|
|
Hospitalization related to CV, n (%) [95% CI]
|
167 (6.3) [5.42, 7.31]
|
80 (5.2) [4.14, 6.42]
|
46 (8.8) [6.55, 11.62]
|
41 (7.0) [5.08, 9.40]
|
Abbreviations: CI, confidence interval; CV, cardiovascular; DVT, deep venous thrombosis;
PE, pulmonary embolism; VTE, venous thromboembolism.
Note: aAdjudicated events.
Fig. 2 Venous thromboembolism (VTE) recurrence during the 18-month study period. Upper panel: 100 VTE recurrences in 2,644 patients of the full analysis set (FAS), 37 recurrences
on-treatment. Lower panel: VTE recurrence by type (deep vein thrombosis [DVT] only, pulmonary embolism [PE]
only, PE with DVT).
Assessing the rates of VTE recurrence in patients on edoxaban based on an initial
dose of 30/60 mg from baseline to 18 months showed a higher incidence of recurrent
VTE (2.1% vs. 1.3%; annualized rate 1.5%/year vs. 0.9%/year) in patients receiving
30 mg than patients with 60 mg ([Supplementary Table S7], available in the online version). Also, during the first 6 months, the VTE recurrence
rate was slightly higher (annualized rate: 1.1%/year vs. 0.5%/year) in patients with
30 mg compared with 60 mg ([Supplementary Table S8], available in the online version).
The rate of recurrent symptomatic VTE within the first 12 months of the study was
lower in ETNA-VTE overall compared with the PREFER-VTE VKA overall and PREFER-VTE
NOAC (2.4% vs. 3.3% vs. 3.4%; [Supplementary Table S9], available in the online version). Similarly, on edoxaban treatment, the rates of
recurrent VTE were lower in ETNA-VTE than in PREFER-VTE during VKA treatment and NOAC
treatment (1.0% vs. 2.3% vs. 2.8%).
In the HOKUSAI-VTE study, the overall rate of recurrent VTE in patients treated with
edoxaban and warfarin within the first 12 months was comparable with the rate documented
in ETNA-VTE overall (2.8% vs. 3.3% vs. 2.4%, respectively) ([Supplementary Table S10], available in the online version). No differences were observed during the edoxaban
treatment (1.3% in HOKUSAI-VTE-edoxaban vs. 1.8% in HOKUSAI-VTE-warfarin vs. 1.0%
in ETNA-VTE-edoxaban).
Bleeding Events
During the 18-month study period, 305 patients (11.5%) suffered from bleeding events
(any) while on edoxaban treatment (annualized rate 13.5%/year) ([Table 4]). Among them, 37 patients (1.4%) had major bleeding at an annualized rate of 1.5%/year.
Assessing the severity of bleeding events, 9.0% were classified as major, 19.5% as
CRNMB, and 71.4% as minor. The majority of events (71.5%) were spontaneous, while
28.5% had a potentially transient, identifiable cause.
Table 4
On-treatment safety of edoxaban during the 18-month study period
|
Number of subjects with at least one event—on edoxaban
|
ETNA-VTE overall
(N = 2,644)
|
DVT only
(N = 1,540)
|
PE with DVT
(N = 520)
|
PE only
(N = 584)
|
|
Any bleeding, n (%) [95% CI]
|
305 (11.5) [10.34, 12.81]
|
159 (10.3) [8.85, 11.95]
|
70 (13.5) [10.65, 16.70]
|
76 (13.0) [10.39, 16.02]
|
|
Major bleedinga, n (%) [95% CI]
|
37 (1.4) [0.99, 1.92]
|
19 (1.2) [0.74, 1.92]
|
10 (1.9) [0.93, 3.51]
|
8 (1.4) [0.59, 2.68]
|
|
Hemorrhagic stroke, n (%) [95% CI]
|
3 (0.1) [0.02, 0.33]
|
0 (0.0) [0.00, 0.24]
|
1 (0.2) [0.00, 1.07]
|
2 (0.3) [0.04, 1.23]
|
|
CRNMBa, n (%) [95% CI]
|
63 (2.4) [1.84, 3.04]
|
33 (2.1) [1.48, 3.00]
|
18 (3.5) [2.06, 5.42]
|
12 (2.1) [1.07, 3.56]
|
|
Subjects with any bleeding events, n (%)
|
305
|
159
|
70
|
76
|
|
Total number with any bleeding events, n
|
420
|
198
|
116
|
106
|
|
Severity of the bleeding event
|
|
Major, n (%)
|
38 (9.0)
|
19 (9.6)
|
11 (9.5)
|
8 (7.5)
|
|
Clinically relevant nonmajor, n (%)
|
82 (19.5)
|
37 (18.7)
|
31 (26.7)
|
14 (13.2)
|
|
Minor, n (%)
|
300 (71.4)
|
142 (71.7)
|
74 (63.8)
|
84 (79.2)
|
|
Spontaneous or provoked bleeding
|
|
Spontaneous, n (%)
|
288 (71.5)
|
146 (77.7)
|
76 (66.7)
|
65 (65.0)
|
|
Provoked, n (%)
|
115 (28.5)
|
42 (22.3)
|
38 (33.3)
|
35 (35.0)
|
|
Unknown, n (%)
|
18
|
10
|
2
|
6
|
Abbreviations: CI, confidence interval; CRNMB, clinically relevant non-major bleeding;
CV, cardiovascular; DVT, deep venous thrombosis; PE, pulmonary embolism; VTE, venous
thromboembolism.
Note: aAdjudicated events.
In 18 months of follow-up of FAS overall, the annualized rate of any bleeding events
and at least one major bleeding event was 10.3%/year and 1.2%/year, respectively.
Bleeding events tended to be more common in the first 3 months ([Fig. 3], upper panel) and tended to occur more often in patients who had PE with DVT. When comparing the
bleeding events every 6 months in patients who continued with anticoagulant treatment
for 18 months, there is a decrease over time (annualized rate of any bleeding: baseline
to 6 months: 6.8%/year; 6 to 12 months: 2.2%/year; 12 to 18 months: 2.7%/year; [Supplementary Table S5], available in the online version).
Fig. 3 Major bleeding events during the 18-month study period. Upper panel: 45 major bleedings in 2,644 patients of the full analysis set (FAS), 37 major bleedings
on treatment. Lower panel: Major bleeding by type (deep vein thrombosis [DVT] only, pulmonary embolism [PE]
only, PE with DVT).
Assessing the bleeding rates in patients on edoxaban based on a dose of 30/60 mg from
baseline to 18 months showed a higher incidence of any bleeding events (13.0% vs.
11.3%; annualized rate 9.3%/year vs. 7.8%/year) and major bleeding events (3.0% vs.
1.2%; annualized rate 2.2%/year vs. 0.8%/year) in patients receiving 30 mg than patients
with 60 mg ([Supplementary Table S7], available in the online version). Also, during the first 6 months, the any bleeding
rate was higher (annualized rate:10.1%/year vs. 8.2%/year) in patients with 30 mg
compared with 60 mg ([Supplementary Table S8], available in the online version).
In the PREFER-VTE study, major bleeding events within the first 12 months were reported
in seven patients (0.8%) receiving VKA and in two patients (0.4%) receiving NOAC,
which is lower compared with the rate reported in ETNA-VTE (1.3%) ([Supplementary Table S11], available in the online version). The rates of CRNMB were comparable (2.3% in ETNA-VTE
vs. 2.6% in PREFER-VTE VKA and 2.0% PREFER-VTE NOAC).
Within the first 12 months of edoxaban treatment, the rates of hemorrhagic stroke
(0.1% in ETNA-VTE) and major bleeding (1.3% in ETNA-VTE) were not substantially different
compared with HOKUSAI-VTE edoxaban subjects (0.1 and 1.2%, respectively) ([Supplementary Table S12], available in the online version). There was a reduced rate of CRNMB observed in
ETNA-VTE subjects (2.3%) compared with the rates documented in HOKUSAI-VTE subjects
on edoxaban (6.3%) and warfarin (7.0%).
Death
The cumulative incidence of all-cause deaths was low in the overall FAS population
([Fig. 4]). A total of 95 deaths (3.6%) were reported, corresponding to an annualized rate
of 2.6%/year. A total of 50 deaths (1.9%) occurred during edoxaban treatment (annualized
rate 2.1%/year), and 45 (1.7%) occurred after edoxaban discontinuation ([Table 5]). Overall, 71 deaths were unrelated to CV or VTE causes. On edoxaban treatment,
only one VTE-related death was reported during the first month of the study (annualized
rate 0.0%/year), and 11 (0.4%) CV-related death was reported (annualized rate 0.5%/year).
Fig. 4 All-cause death during the 18-month study period. Upper panel: 95 deaths in 2,644 patients of the full analysis set (FAS), 50 deaths on treatment.
Lower panel: All-cause death by type (deep vein thrombosis [DVT] only, pulmonary embolism [PE]
only, PE with DVT).
Table 5
Overall and on-treatment mortality during the 18-month study period
|
ETNA-VTE overall
(N = 2,644)
|
DVT only
(N = 1,540)
|
PE with DVT
(N = 520)
|
PE only
(N = 584)
|
|
Number of subjects died—overall
|
|
All-cause deatha, n (%) [95% CI]
|
95 (3.6) [2.92, 4.37]
|
51 (3.3) [2.48, 4.33]
|
19 (3.7) [2.21, 5.65]
|
25 (4.3) [2.79, 6.25]
|
|
CV-related deatha, n (%) [95% CI]
|
23 (0.9) [0.55, 1.30]
|
12 (0.8) [0.40, 1.36]
|
5 (1.0) [0.31, 2.23]
|
6 (1.0) [0.38, 2.22]
|
|
VTE-related deatha, n (%) [95% CI]
|
1 (0.0) [0.00, 0.21]
|
0 (0.0) [0.00, 0.24]
|
0 (0.0) [0.00, 0.71]
|
1 (0.2) [0.00, 0.95]
|
|
Number of subjects died on edoxaban treatment
|
|
All-cause deatha, n (%) [95% CI]
|
50 (1.9) [1.41, 2.49]
|
27 (1.8) [1.16, 2.54]
|
9 (1.7) [0.79, 3.26]
|
14 (2.4) [1.32, 3.99]
|
|
CV-related deatha, n (%) [95% CI]
|
11 (0.4) [0.21, 0.74]
|
5 (0.3) [0.11, 0.76]
|
1 (0.2) [0.00, 1.07]
|
5 (0.9) [0.28, 1.99]
|
|
VTE-related deatha, n (%) [95% CI]
|
1 (0.0) [0.00, 0.21]
|
0 (0.0) [0.00, 0.24]
|
0 (0.0) [0.00, 0.71]
|
1 (0.2) [0.00, 0.95]
|
Abbreviations: CI, confidence interval; CV, cardiovascular; DVT, deep venous thrombosis;
PE, pulmonary embolism; VTE, venous thromboembolism.
Note: aAdjudicated events.
Overall, the 12-month rate of all-cause death was higher in the ETNA-VTE population
than in the PREFER-VTE patient groups (2.8% in ETNA-VTE vs. 1.5% in PREFER-VTE VKA
and 1.6% PREFER-VTE NOAC), which was also observed during the edoxaban treatment (1.6%
vs. 1.0% and 0.6%, respectively) ([Supplementary Table S13], available in the online version).
Compared with the HOKUSAI-VTE population, the rate of all-cause death within the first
12 months in ETNA-VTE patients on edoxaban was higher (1.6% in ETNA-VTE vs. 0.7% HOKUSAI-VTE
on edoxaban and 0.5% in HOKUSAI-VTE on warfarin; [Supplementary Table S14], available in the online version).
Discussion
In a large, real-world dataset collected from consecutive patients with an acute VTE
who were treated with edoxaban, it was shown that: (1) recurrent VTE persisted beyond
the initial 6 months, with less occurrence among patient on edoxaban treatment compared
with the overall population (annualized rate 1.6%/year vs. 2.7%/year); (2) the rate
of major bleeding on edoxaban was low (annualized rate of 1.5%/year), and (3) the
risk of death appeared to be low (annualized rate 2.1%/year) in patients who received
edoxaban treatment (up to 18 months) with low annual rate of CV (0.5%/year) or VTE-related
death (and 0.0%/year). However, edoxaban treatment based on 30/60 mg dose from baseline
to 18 months showed a higher incidence of any bleeding events (annualized rate 9.3%/year
vs. 7.8%/year) and major bleeding events (annualized rate 2.2%/year vs. 0.8%/year)
in patients receiving 30 mg compared with patients with 60 mg. Assessing the rates
of VTE recurrence based on dose showed a higher incidence of recurrent VTE (annualized
rate 1.5%/year vs. 0.9%/year) in patients receiving 30 mg than patients with 60 mg.
Also, during the first 6 months, the VTE recurrence rate was slightly higher (annualized
rate: 1.1%/year vs. 0.5%/year) in patients with 30 mg. This is likely due to patient
characteristics and the fact that patients perceived to have a higher risk of bleeding
had their anticoagulation dose reduced. Overall, the results indicated that edoxaban
treatment for up to 18 months was associated with a low rate of VTE recurrence, major
bleeding, and death. These results are aligned with those of the randomized clinical
trial, reassuring the use of edoxaban in the treatment of VTE in routine clinical
practice.
Patient Characteristics and Study Outcomes
Compared with the HOKUSAI-VTE clinical study (considering those patients from Europe),
patients included in the ETNA-VTE Europe were older, more often female, and had more
comorbid diseases. Furthermore, more patients in ETNA-VTE Europe had a low body weight
(≤60 kg) and reduced creatinine clearance (CrCl ≤50 mL/min). Dose adjustment was already
required in the HOKUSAI-VTE study for low body weight and those with low creatinine
clearance. However, fewer subjects with these parameters were enrolled in the study
than in ETNA-VTE. Less well-preserved renal function in ETNA-VTE may indicate a specific
selection of subjects in the randomized controlled clinical study (HOKUSAI-VTE), as
reflected by stringent inclusion and exclusion criteria.
Previous studies have shown that the risk of VTE recurrence is higher in patients
≥70 years than in younger subjects.[10] Furthermore, age has been found to increase the risk of bleeding related to anticoagulant
treatment. Recent studies indicated that the increase in risk of VTE recurrence and
bleeding events in patients ≥70 years is less pronounced with DOACs than with VKA.[11]
[12]
[13] Reduced CrCl is also a risk factor for bleeding and seems to affect VKA more than
DOACs.[11] It is reassuring to note that edoxaban treatment in ETNA-VTE Europe was effective
and demonstrated a good safety profile in a real-world population, which included
a high proportion of older patients (50.4% of patients ≥65 years) and patients with
reduced renal function (37.2%).
Effectiveness of Extended Use of Edoxaban
The proportion of patients with recurrent VTE steadily increased from baseline until
the end of the 18-month follow-up, with no particular risk increase during the first
3 months. About two-thirds of recurrent VTE events occurred when subjects had discontinued
edoxaban treatment. The overall rate of recurrent VTE was nominally lower in ETNA-VTE
than in the PREFER-VTE observational study at a comparable 12-month time window (2.4%
vs. 3.3% on VKA and 3.4% on other NOACs) and also lower than in the HOKUSAI-VTE study
(2.4% vs. 2.8% on edoxaban and 3.3% on warfarin). The proportion of patients who died
(all-cause) steadily increased from baseline until the end of the 18-month follow-up,
with no particular risk increase during the first 3 months. One patient died due to
VTE. Although all-cause death was higher in ETNA-VTE on edoxaban than in the HOKUSAI-VTE
population, the difference might be explained by the fact that the subjects of ETNA-VTE
were older, were less often male, and had a higher prevalence of venous insufficiency
than subjects in the HOKUSAI-VTE clinical study.
Safety of Extended Use of Edoxaban
The proportion of subjects with at least one major bleeding event on edoxaban in ETNA-VTE
at 12 months was nominally higher than in either subgroup from PREFER-VTE (1.3% vs.
0.8% on VKA and 0.4% on NOAC). On the other hand, the major bleeding rate in ETNA-VTE
(1.3%) was comparable with the results of HOKUSAI-VTE (1.2% for edoxaban and 1.3%
for warfarin; all rates are on treatment). A meta-analysis showed no apparent difference
in preventing major bleeding events between secondary thromboprophylaxis (warfarin
or aspirin) and placebo (OR 1.84, 95% CI 0.87 to 3.85; low-quality evidence) and also
between rivaroxaban and aspirin (OR 3.06, 95% CI 0.37 to 25.51; p = 0.30; moderate-quality evidence).[14] There was, however, a reduced rate of CRNMB observed in ETNA-VTE (2.3%) compared
with the rates documented in HOKUSAI-VTE on edoxaban (6.3%). A potential reason for
this difference may be a common underreporting of clinically less severe events in
observational studies.
Clinical Implications
International guidelines state that anticoagulation after a VTE event should be continued
for not less than 3 months[5]
[15] or, more explicitly, for 3 to 6 months.[6] Some physicians believe that treatment should be given for a minimum of 6 months,[16] especially as VTE recurrence risk is not resolved within the first 6 months of anticoagulation.[17]
This analysis has provided reasonable evidence of the effectiveness of extended edoxaban
treatment (up to 18 months) in preventing recurrent VTE. A low rate of recurrent VTE,
low death rate, as well as bleeding, indicate a favorable benefit–risk profile of
edoxaban treatment in clinical practice. Furthermore, the prolonged use of edoxaban
led to improved survival, but three-quarters of deaths in this study population were
not related to either CV or VTE events.
Patient safety is paramount and needs to be considered when deciding on treatment,
particularly for the long term. We have shown that extended edoxaban treatment (up
to 18 months) is effective in patients with recurrent VTEs.
