Keywords ACTS - anticoagulation - patient experience - rivaroxaban - venous thromboembolism
Introduction
Anticoagulant treatment for venous thromboembolism (VTE), comprising deep-vein thrombosis
(DVT) and pulmonary embolism (PE), is associated with various benefits and burdens,
some of which are therapy specific. Traditional therapies have several well-known
limitations; agents such as low-molecular-weight heparin (LMWH) are administered via
injection, which makes self-administration problematic for some patients.[1 ] Many patients also have a dislike or fear of needles,[2 ] meaning that repeated injections may not be suitable for them. Vitamin K antagonists
(VKAs) require regular anticoagulation monitoring, which contributes to making their
use in the outpatient setting suboptimal. A meta-analysis of randomized trials and
cohort studies showed that in the first month of treatment, patients were within the
target international normalized ratio (INR) range of 2.0 to 3.0 only 56% of the time.[3 ] Even when the first 3 months of treatment were excluded from the meta-analysis,
the time in therapeutic range was only around 75%. The interaction of VKAs with foods
containing high levels of vitamin K, other medications, and alcohol contributes to
the need for monitoring and management of adverse effects, making their use more burdensome.[4 ] VKAs have slow and variable onsets of action, and there is considerable variation
between individuals in terms of their response to therapy. Other concerns associated
with anticoagulation treatment may adversely affect patients' attitudes toward therapy,
for example, the perceived risk of bleeding events.
In addition to treatment efficacy, the issues described earlier can impact the patient's
subjective impression of the treatment process. Employing measures to assess and quantify
parameters such as treatment satisfaction or enhancement of the patient experience
is becoming common practice as part of clinical studies, especially as treatment satisfaction
has a positive association with adherence and persistence.[5 ]
[6 ] The non-VKA oral anticoagulants (apixaban, dabigatran, edoxaban, and rivaroxaban)
can potentially reduce many of the burdens associated with standard therapies. For
example, their oral route of administration, combined with a lack of requirement for
routine anticoagulation monitoring, would be expected to reduce the burdens associated
with repeated injections and INR testing.
The phase III EINSTEIN DVT and EINSTEIN PE studies demonstrated a higher degree of
patient treatment satisfaction with rivaroxaban versus standard anticoagulation.[7 ]
[8 ]
[9 ]
[10 ] The noninterventional XALIA study of rivaroxaban versus standard anticoagulation
for the treatment of DVT in routine clinical practice was conducted subsequently.
One of the aims of XALIA was to determine whether improved patient experience with
rivaroxaban was replicated outside the clinical trial setting;[11 ] this analysis presents the results of these assessments.
Materials and Methods
The XALIA study methods are described in the XALIA primary article.[11 ] The methodology for the Anti-Clot Treatment Scale (ACTS) substudy is presented below.
Patients
All patients enrolled in XALIA who received rivaroxaban or standard anticoagulation
treatment were eligible for inclusion in this substudy. Patients were aged ≥18 years,
with objectively confirmed DVT (after approval of rivaroxaban in the PE indication,
patients with DVT and concomitant PE were also eligible) and an indication for at
least 3 months of anticoagulation treatment. The type, dose, and duration of anticoagulant
drug therapy administered to each patient were at the discretion of the attending
physician. Patients who received rivaroxaban alone, or who initially received heparin/fondaparinux
for a maximum of 48 hours before enrolment, were included in the rivaroxaban cohort,
consistent with the approach in the EINSTEIN DVT phase III trial.[7 ] Patients who received initial heparin or fondaparinux for >2 to 14 days with or
without a VKA for 1 to 14 days before switching to rivaroxaban were designated as
“early switchers” and were not included in the analysis presented here.
Anti-Clot Treatment Scale
Patients completed the anticoagulation-specific ACTS during follow-up visits; country-specific
and linguistically validated versions of the questionnaire were used. The ACTS comprises
17 items representing negative and positive aspects of anticoagulation treatment:
ACTS Burdens (12 scale items plus 1 global item regarding burdens); ACTS Benefits
(3 scale items plus 1 global item regarding benefits). Item scores were summed across
domains to give an ACTS Burdens score ranging from 12 to 60 and an ACTS Benefits score
ranging from 3 to 15, with higher scores in both indicating a more positive patient
experience with anticoagulation treatment; scores for global items were not included
in the calculations. The ACTS was determined at days 1 to 44 (Visit 1), days 45 to
134 (Visit 2), days 135 to 224 (Visit 3), days 225 to 314 (Visit 4), and any time
from day 315 onward (Visit 5). The questions comprising the ACTS questionnaire are
shown in the [Supplementary Material ] ([Table S1 ]).
Propensity Score-Adjusted Analysis
Propensity score adjustment was used to address the imbalances in baseline characteristics
of the two treatment groups in the safety analysis set. Briefly, analyses were conducted
in two ways, in a set of matched pairs as well as stratified by eight homogeneous
subclasses, both sets derived with propensity scores. Propensity scores, matched pairs,
and subclasses were developed by an independent statistician without knowledge of
outcome events and addressed allocation bias via adjustment of the potential effect
of unbalanced covariates.[12 ]
[13 ] Matched pairs based on the propensity scores were created using the greedy algorithm.[14 ] The propensity score derivation and stratified analysis methodologies are included
in the [Supplementary Material ] of the XALIA primary article.[11 ] The main analysis of the ACTS results in this study was conducted on the set of
patients with matched propensity scores (the propensity score-matched set [PMS]) who
had at least one nonmissing ACTS assessment after implementation of the imputation
method (termed the PMS-ACTS set). As a consequence, only a subset of patients consisted
of matched pairs. Alternatively, patients with propensity scores that were too high/low
were removed from the safety set and eight homogenous subclasses were created for
the adjusted safety analysis (ASAF) set. A sensitivity analysis was conducted in the
adjusted safety set of patients who had at least one nonmissing ACTS assessment after
implementation of the imputation method (ASAF-ACTS).
Subgroups
Several subgroups were also assessed for differences in ACTS Burdens and Benefits
scores between treatments. These subgroups were as follows: age (<60, ≥60 years);
body mass index (≤25, >25 to ≤35, >35 kg/m2 ); sex (male, female); hospitalization for index events (yes, no); immobilization
(yes, no); language (Dutch, German, French, English, Italian, Swedish, Spanish, Canadian
English); provoked VTE (yes, no); reason for treatment (patient's age, patient's living
conditions, comorbidities, distance to treating physician, medical or hospital guidelines,
availability of drug, price of drug, type of health insurance, other); renal disease
(yes, no); active cancer at baseline (yes, no); active cancer at baseline (excluding
patients on VKA; yes, no); race (white, black, Asian, not reported); region (western
Europe, Canada and Israel, eastern Europe); first available weight (≤70, >70 to <90,
≥90 kg); country (France, Germany, Spain, rest); chronic heart failure (yes, no);
diabetes (yes, no); first available creatinine clearance (<30, ≥30 to <50, ≥50 to
<80, ≥80 mL/min); patient health insurance (public, fully private, partially private,
other [including missing]); cardiovascular disease (yes, no); stroke (yes, no); previous
major bleeding event (yes, no); previous VTE (yes, no); thrombophilia (yes, no); venous
insufficiency (yes, no); history of hypertension (yes, no); and chronic obstructive
pulmonary disease (yes, no).
Statistical Analysis
The ACTS was completed in accordance with the developer's guidelines.[15 ] In instances for which more than 50% of questions were missing responses, the scale
was considered to be incomplete; partial completion where missing responses were below
50% was addressed by scale-specific mean imputation.
A mixed-model repeated-measures analysis using an unstructured covariance matrix was
used to analyze the questionnaire data because the questionnaire responses were multiple
measurements on patient experience with treatment over a period of time. The mixed
model included ACTS as outcome, and treatment, cancer at baseline, and treatment by
visit interaction as covariates. From this model, least square (LS) means for each
treatment group and treatment differences and corresponding standard errors (SEs)
were presented per visit. The same analysis without treatment by visit interaction
for the overall effect with LS means and SEs is shown. For ASAF-ACTS, the LS means
were calculated individually per strata and treatment and then combined with a stratified-based
combining rule. All statistical significance testing was performed at a two-sided
0.05α level, and all data analysis was done using SAS statistical software package
version 9.4 (SAS Institute, Cary, North Carolina, United States).
Results
Patients and ACTS Completion Rates
Of the 5,136 patients who received study medication in XALIA, 4,768 (92.8%) were included
in the safety analysis set (368 [7.2%] early switchers were excluded from the primary
analysis). Subsequently, 253 patients were excluded from the ASAF because they had
propensity scores that were either too high or too low. This left an ASAF set of 4,515
patients, 2,505 (55.5%) of who received rivaroxaban and 2,010 (44.5%) who received
standard anticoagulation. A total of 1,124 matched pairs (2,248 patients) were included
in the PMS set.
Because several patients in XALIA had missing ACTS data, the PMS-ACTS set comprised
892 patients in total, of which 458 (51.3%) received rivaroxaban and 434 (48.7%) received
standard anticoagulation. The ASAF-ACTS set included 1,726 patients (1,007 [58.3%]
for rivaroxaban, 719 [41.7%] for standard anticoagulation).
Baseline Demographics and Clinical Characteristics
The baseline demographics and clinical characteristics of the PMS-ACTS set are shown
in [Table 1 ]. Patients in the rivaroxaban cohort were generally younger, had higher rates of
unprovoked VTE, and had lower rates of renal impairment, concomitant PE, and active
cancer than those in the standard anticoagulation group. The baseline characteristics
of both analysis sets were similar to those of the full safety analysis set.[11 ]
Table 1
Baseline demographics and clinical characteristics of patients in the XALIA treatment
satisfaction substudy (PMS-ACTS set)
Characteristic[a ]
Rivaroxaban
(N = 458)
Standard anticoagulation[b ]
(N = 434)
Mean age, y (SD)
60.1 (15.7)
61.7 (16.6)
Age category
< 60 y
198 (43.2)
191 (44.0)
≥ 60 y
260 (56.8)
243 (56.0)
Male sex
196 (42.8)
206 (47.5)
Weight
< 50 kg
4 (0.9)
6 (1.4)
≥ 50 to 70 kg
97 (21.2)
95 (21.9)
> 70 to <90 kg
157 (34.3)
150 (34.6)
≥ 90 kg
119 (26.0)
111 (25.6)
Missing
81 (17.7)
72 (16.6)
First available creatinine clearance
< 30 mL/min
1 (0.2)
8 (1.8)
≥ 30 to <50 mL/min
20 (4.4)
21 (4.8)
≥ 50 to <80 mL/min
95 (20.7)
77 (17.7)
≥ 80 mL/min
201 (43.9)
160 (36.9)
Missing
141 (30.8)
168 (38.7)
Index diagnosis
DVT only
409 (89.3)
382 (88.0)
DVT with PE
49 (10.7)
52 (12.0)
Type of VTE[c ]
Provoked
143 (31.2)
155 (35.7)
Unprovoked
315 (68.8)
279 (64.3)
Previous VTE
124 (27.1)
109 (25.1)
Active cancer at baseline
40 (8.7)
46 (10.6)
Known thrombophilic condition
35 (7.6)
35 (8.1)
Previous major bleeding episode
11 (2.4)
8 (1.8)
Abbreviations: ACTS, Anti-Clot Treatment Scale; DVT, deep-vein thrombosis; PE, pulmonary
embolism; PMS, propensity score matched set; SD, standard deviation; VTE, venous thromboembolism.
a
n (%) unless otherwise stated.
b Standard anticoagulation consisted of initial treatment with unfractionated heparin,
low-molecular-weight heparin, or fondaparinux, which could overlap with and be followed
by an oral vitamin K antagonist.
c Cancer was not considered when defining DVT as provoked or unprovoked.
For the PMS-ACTS set, the countries with the highest rates of representation were
France (30.6%) and Germany (27.2%). A similar pattern was seen with the ASAF-ACTS
set, with France and Germany again forming the largest geographical cohorts (30.9
and 29.5%, respectively).
ACTS Burdens
PMS-ACTS Analysis Set
The overall ACTS Burdens score (LS mean difference) was 2.4 ± 0.4 points higher (higher
scores indicating a reduced burden) in the PMS-ACTS set in the rivaroxaban group versus
the standard anticoagulation group (LS mean scores of 56.1 and 53.7, respectively;
p < 0.0001), indicative of a significantly reduced burden with rivaroxaban. The higher
scores with rivaroxaban were consistent over time, with the LS mean difference in
scores ranging from 2.2 (Visits 2 and 4) to 2.9 (Visit 3); the differences were significant
at each visit ([Table 2 ]).
Table 2
ACTS Burdens scores by visit (PMS-ACTS set)
Visit
Rivaroxaban/standard anticoagulation patients (n /n )[a ]
Rivaroxaban/standard anticoagulation scores (LS mean/LS mean)
LS mean difference (SE)
p -Value
Overall
458/434
56.1/53.7
2.4 (0.4)
< 0.0001
1
333/295
55.2/52.5
2.7 (0.5)
< 0.0001
2
307/279
56.0/53.8
2.2 (0.5)
< 0.0001
3
179/161
56.1/53.2
2.9 (0.6)
< 0.0001
4
68/75
56.7/54.5
2.2 (0.7)
0.002
5
69/77
56.5/54.1
2.4 (0.6)
0.0002
Abbreviations: ACTS, Anti-Clot Treatment Scale; LS, least squares; PMS, propensity
score matched set; SE, standard error.
a ACTS Burdens scores had no effect on treatment discontinuation in the PMS-ACTS analysis
set. Patients treated with standard anticoagulation treatment discontinued 1.6 times
more than patients treated with rivaroxaban.
ASAF-ACTS Analysis Set
The overall ACTS Burdens score (LS mean difference) was 2.4 ± 0.4 points higher in
the ASAF-ACTS set in the rivaroxaban group versus the standard anticoagulation group
(LS mean scores of 55.2 and 52.9, respectively; p = 0.0001), indicative of a significantly reduced burden with rivaroxaban. The higher
scores with rivaroxaban were consistent over time, with LS mean difference in scores
ranging from 2.0 (Visit 2) to 2.8 (Visit 3); the differences were significantly different
at each visit ([Table 3 ]).
Table 3
ACTS Burdens scores by visit (ASAF-ACTS set)
Visit
Rivaroxaban/standard anticoagulation patients (n /n )
Rivaroxaban/standard anticoagulation scores (LS mean/LS mean)
LS mean difference (SE)
p -Value
Overall
1,007/718
55.2/52.9
2.4 (0.4)
0.0001
1
743/475
54.6/52.3
2.5 (0.5)
0.0003
2
656/450
55.1/53.2
2.0 (0.5)
0.001
3
371/246
55.2/52.9
2.8 (0.5)
0.0004
4
166/120
55.6/53.5
2.7 (0.8)
0.009
5
155/115
55.4/52.8
2.7 (0.8)
0.002
Abbreviations: ACTS, Anti-Clot Treatment Scale; ASAF, adjusted safety analysis; LS,
least squares; SE, standard error.
Subgroups
Almost all patient subgroups in the PMS-ACTS and ASAF-ACTS sets showed statistically
significant differences in ACTS Burdens scores between treatment groups in favor of
rivaroxaban, both for overall scores and by treatment visit. Treatment comparisons
for selected subgroups are shown in [Fig. 1 ].
Fig. 1 ACTS Burdens score least squares mean differences by subgroup (PMS-ACTS analysis
set). Subgroups with missing values had too few patients to enable LS mean difference
to be calculated. ACTS, Anti-Clot Treatment Scale; CI, confidence interval; LS, least
squares; PMS, propensity score-matched set; SE, standard error; VTE, venous thromboembolism.
ACTS Benefits
PMS-ACTS Analysis Set
The overall ACTS Benefits score (LS mean difference) was 0.2 ± 0.2 points higher in
the PMS-ACTS set in the rivaroxaban group versus the standard anticoagulation group,
although this difference was not significant (LS mean scores of 12.1 and 11.9, respectively;
p = 0.2). The similar ACTS Benefits scores between the two treatment groups were consistent
over time, with the LS mean difference in scores ranging from –0.1 (Visit 1) to 0.5
(Visit 2; [Table 4 ]).
Table 4
ACTS Benefits scores by visit (PMS-ACTS set)
Visit
Rivaroxaban/standard anticoagulation patients (n /n )[a ]
Rivaroxaban/standard anticoagulation scores (LS mean/LS mean)
LS mean difference (SE)
p -Value
Overall
450/430
12.1/11.9
0.2 (0.2)
0.2
1
326/286
11.6/11.6
–0.1 (0.2)
0.8
2
298/277
12.0/11.4
0.5 (0.2)
0.01
3
177/159
12.1/11.9
0.2 (0.3)
0.4
4
68/75
12.2/12.1
0.1 (0.4)
0.8
5
67/77
12.3/12.2
0.1 (0.3)
0.8
Abbreviations: ACTS, Anti-Clot Treatment Scale; LS, least squares; PMS, propensity
score matched set; SE, standard error.
a ACTS Benefits scores showed an odds ratio of 0.91 (95% CI: 0.84–0.98) for treatment
discontinuation in the PMS-ACTS analysis set; therefore, the lower the ACTS Benefits
score was, the more patients tended to discontinue from the study.
ASAF-ACTS Analysis Set
The overall ACTS Benefits score (LS mean difference) was 0.2 ± 0.1 points higher in
the ASAF-ACTS set in the rivaroxaban group versus the standard anticoagulation group,
although this difference was not statistically significant (LS mean scores of 11.9
and 11.8, respectively; p = 0.4). The similar ACTS Benefits scores between the two treatment groups were consistent
over time, with the LS mean difference in scores ranging from <0.1 (Visit 5) to 0.2
(Visits 1, 3, and 4; [Table 5 ]).
Table 5
ACTS Benefits scores by visit (ASAF-ACTS set)
Visit
Rivaroxaban/standard anticoagulation patients (n /n )
Rivaroxaban/standard anticoagulation scores (LS mean/LS mean)
LS mean difference (SE)
p -Value
Overall
989/712
11.9/11.8
0.2 (0.1)
0.4
1
724/465
11.8/11.6
0.2 (0.2)
0.5
2
639/447
11.8/11.6
0.1 (0.2)
0.3
3
365/244
11.9/11.8
0.2 (0.3)
0.6
4
165/119
12.1/11.9
0.2 (0.4)
0.7
5
152/114
12.2/11.9
<0.1 (0.4)
0.5
Abbreviations: ACTS, Anti-Clot Treatment Scale; ASAF, adjusted safety analysis; LS,
least squares; SE, standard error.
Subgroups
As with the total PMS-ACTS and ASAF-ACTS results, the majority of subgroups showed
no statistically significant differences between treatment groups for ACTS Benefits
scores, either for the overall scores or by treatment visit. For those subgroups that
did show differences, almost all were at Visit 2. Treatment comparisons for selected
subgroups are shown in [Fig. 2 ].
Fig. 2 ACTS Benefits score least squares mean differences by subgroup (PMS-ACTS analysis
set). Subgroups with missing values had too few patients to enable LS mean difference
to be calculated. ACTS, Anti-Clot Treatment Scale; CI, confidence interval; LS, least
squares; PMS, propensity score-matched set; SE, standard error; VTE, venous thromboembolism.
Discussion
The XALIA ACTS substudy compared treatment experience with DVT (also DVT plus concomitant
PE after the approval of rivaroxaban in the PE indication) in patients who were treated
with rivaroxaban or standard anticoagulation. The results demonstrated that patients
given rivaroxaban had an improved treatment experience, both overall and across study
time points, in terms of finding the treatment less burdensome than standard anticoagulation.
Similarly, the ACTS Benefits scores showed a numerical trend in favor of rivaroxaban.
These findings for the ACTS Burdens scores are consistent with results from the ACTS
substudy of the EINSTEIN DVT trial.[9 ] The PMS-ACTS and ASAF-ACTS analysis sets were similar to the overall XALIA safety
analysis set in respect of baseline demographic and clinical characteristics. The
exceptions to this were that patients in the PMS-ACTS set were on average ∼1 year
older (60.9 vs. 59.8 years) and had higher rates of previous VTE (26.1 vs. 23.3%)
than patients in the safety analysis set.
The completion rates of the ACTS for the rivaroxaban and standard anticoagulation
groups in the ACTS substudy in XALIA were consistent with that observed for EINSTEIN
DVT, where around 40% of all patients were analyzed in the treatment satisfaction
substudy. Similar to EINSTEIN DVT, the open-label nature of treatment administration
in XALIA meant that patients were able to evaluate the impact of the therapy directly
because they were not given dummy controls or subject to sham INR monitoring. In addition,
the real-world nature of the XALIA study also means that the findings are reflective
of actual clinical practice. Subgroup analysis by baseline patient demographics and
clinical characteristics revealed a similar pattern to the main analyses for ACTS
Burdens and Benefits scores. For the ACTS Burdens scores, most subgroups showed significant
between-treatment differences favoring rivaroxaban for overall scores and scores by
visit. Of note, the effect size was smaller or showed a marginal trend toward standard
anticoagulation in patients with provoked VTE; a more robust follow-up and health
service contact may explain the almost equivalent response to rivaroxaban. In addition,
the ACTS Burdens scores did not show a statistically significant between-treatment
difference in patients with thrombophilia; again, a more intensive follow-up may explain
this finding. For ACTS Benefits scores, there was a numerical trend favoring rivaroxaban
across visits and subgroups, although only a limited number of subgroups showed statistically
significant differences, and these were almost all at Visit 2.
A limitation of this study is that it assessed only patient experience over a relatively
short period of time (the median treatment duration in XALIA was 181 days with rivaroxaban
and 190 days with standard anticoagulation); therefore, it would also be of interest
to determine whether the patient experience findings observed in XALIA persist in
the longer term. It is important to note that the ACTS questionnaire is a specific
measure that focuses on all medical aspects of the treatment experience that are important
to patients and gathers information on all aspects of anticoagulation therapy, both
positive and negative.[9 ] Therefore, a widely used, generic measure (e.g., the Treatment Satisfaction Questionnaire
for Medication version II [TSQM II]) focusing on the medication per se could complement
the ACTS results.[9 ]
[16 ] It is often highlighted that the use of two measures offers several advantages:
although generic measures allow comparisons of outcomes across different study populations,
thus enhancing the generalizability of findings, specific measures may be more sensitive
for the detection and quantification of small changes that may be important to physicians
or patients.[17 ]
Another limitation of the study was that the patient population was constrained by
the availability of ACTS values; the ASAF-ACTS contained only 38% of the ASAF patients
resulting in 1,726 patients from eight countries (Canada, Denmark, England, France,
Germany, Italy, Spain, and Sweden); therefore, caution is advised when extrapolating
these results to the wider patient population.
In conclusion, despite the limitations described earlier, this substudy confirms the
findings from the EINSTEIN phase III studies that the patient experience is less burdensome
with rivaroxaban treatment than with standard anticoagulation. This finding was expected
based on the characteristics of the two therapies. The differences in ACTS Benefits
scores were numerically in favor of rivaroxaban versus standard anticoagulation treatment
in XALIA, although this did not reach statistical significance. These findings may
potentially impact on factors such as treatment adherence and persistence and could,
therefore, positively impact patients who have a long-term requirement for anticoagulation
treatment.
