Impact of Prolonged Anticoagulation with Rivaroxaban on Provoked Venous Thromboembolism Recurrence: IMPROVE-VTE

Accepted Manuscript

IMpact of Prolonged anticoagulation with Rivaroxaban On proVokEd Venous ThromboEmbolism recurrence: The IMPROVE-VTE Study Craig I. Coleman PharmD , Alexander G.G. Turpie MD , Thomas J. Bunz PharmD, PhD , Jan Beyer-Westendorf MD , William L. Baker PharmD PII: DOI: Reference:

S0002-9343(18)31170-7 https://doi.org/10.1016/j.amjmed.2018.11.040 AJM 14935

To appear in:

The American Journal of Medicine

Please cite this article as: Craig I. Coleman PharmD , Alexander G.G. Turpie MD , Thomas J. Bunz PharmD, PhD , Jan Beyer-Westendorf MD , William L. Baker PharmD , IMpact of Prolonged anticoagulation with Rivaroxaban On proVokEd Venous ThromboEmbolism recurrence: The IMPROVE-VTE Study, The American Journal of Medicine (2018), doi: https://doi.org/10.1016/j.amjmed.2018.11.040

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

IMpact of Prolonged anticoagulation with Rivaroxaban On proVokEd Venous ThromboEmbolism recurrence: The IMPROVE-VTE Study Craig I. Coleman, PharmD; Alexander G. G. Turpie, MD; Thomas J. Bunz, PharmD, PhD; Jan BeyerWestendorf, MD; William L. Baker, PharmD

CR IP T

Institutions:

a. Department of Pharmacy Practice, University of Connecticut School of Pharmacy, Storrs, CT, USA b. Evidence-Based Practice Center, Hartford Hospital, Hartford, CT, USA

c. Department of Medicine, McMaster University, Hamilton, Ontario, Canada

d. Department of Pharmacoepidemiology, New England Health Analytics, LLC, Granby, CT, USA

AN US

e. Thrombosis Research Unit, Department of Medicine I, Division Hematology, University Hospital "Carl Gustav Carus" Dresden, Dresden, Germany f.

Kings Thrombosis Service, Department of Hematology, Kings College London, London, UK

* All authors had access to the data and participated in the writing, critical revision and decision to

M

submit the manuscript.

Corresponding Author:

ED

Craig I. Coleman, PharmD Professor of Pharmacy Practice

University of Connecticut School of Pharmacy

PT

69 North Eagleville Road, Unit 3092 Storrs, CT 06269, USA

CE

860-972-2096

[email protected] Funding:

AC

This study was supported by Bayer AG, Berlin, Germany. The funder had no role in study design; in the collection, analysis and interpretation of data; in the writing of the report; or in the decision to submit the article for publication. Conflicts-of-Interest: Dr. Coleman has received research grants from Bayer AG and Janssen Pharmaceuticals. He has served as a consultant for Boehringer Ingelheim, Janssen Pharmaceuticals, Bayer AG, Portola Pharmaceuticals 1

ACCEPTED MANUSCRIPT

and Medscape. Dr. Beyer-Westendorf has received research grants from Bayer AG, Boehringer Ingelheim, Daiichi Sankyo, Pfizer, Portola Pharmaceuticals, and has served as a consultant to and has received honoraria from Bayer AG, Boehringer Ingelheim, Daiichi Sankyo, Pfizer, Portola Pharmaceuticals, CSL Behring and Medscape. Drs. Bunz and Baker have no relevant disclosures. Key Words: Venous thromboembolism; provoked; extended duration; anticoagulants; rivaroxaban

CR IP T

Running Head: The IMPROVE-VTE Study

HIGHLIGHTS



M



Patients with provoking risk factors (not related to a major surgery or trauma) have an elevated risk of recurrent venous thromboembolism after anticoagulation discontinuation. In patients experiencing a provoked venous thromboembolism and treated with rivaroxaban for the first 3-months, continued rivaroxaban use was associated with a 44% lower risk of recurrent venous thromboembolism versus discontinuation. Continued rivaroxaban use was not associated with an increased risk of major bleeding versus a control group receiving no anticoagulation or non-aspirin antiplatelet agents, with or without aspirin.

AN US



ED

ABSTRACT

PT

Objective: We sought to evaluate the real-world effectiveness and safety of prolonged anticoagulation

CE

with rivaroxaban following a provoked venous thromboembolism. Methods: Using US MarketScan claims from 11/2012-3/2017, we identified adults with ≥1 primary

AC

hospitalization/emergency department diagnosis code for venous thromboembolism, a provoking (major or minor, persistent or transient) risk factor, at least 3-months of continuous rivaroxaban treatment and ≥12-months of continuous insurance benefits prior to their qualifying venous thromboembolism. Patients were categorized as either continuing rivaroxaban or discontinuing anticoagulation (no anticoagulation or non-aspirin antiplatelet agents, but may have received aspirin) after the initial 3-months of rivaroxaban treatment (index date). Differences in baseline covariates 2

ACCEPTED MANUSCRIPT

between cohorts were adjusted for using inverse probability-of-treatment weights based on propensity-scores (absolute standardized differences <0.1 achieved for all covariates after adjustment). Twelve-month incidences of recurrent venous thromboembolism or major bleeding were compared

as hazard ratios (HRs) with 95% confidence intervals (CIs).

CR IP T

between cohorts using Cox regression (according to an intention-to-treat methodology) and reported

Results: Among patients experiencing a provoked venous thromboembolism and treated with

rivaroxaban for the first 3-months (N=4,990), continued rivaroxaban use beyond 3-months [median

AN US

(25%, 75% range) duration of additional rivaroxaban use=3 (2, 5) months] was associated with a 44% (95%CI=9-66%) lower hazard of recurrent venous thromboembolism without altering major bleeding risk (HR=0.87, 95%CI=0.51-1.49) versus anticoagulation discontinuation (with/without aspirin use). Conclusions: Our study suggests continuing rivaroxaban after the initial 3-month period was associated

M

with a decreased risk of recurrent venous thromboembolism. The observed reduction in recurrent

in major bleeding risk.

PT

INTRODUCTION

ED

venous thromboembolism with prolonged rivaroxaban use was not associated with a significant change

CE

Venous thromboembolism, defined as deep-vein thrombosis or pulmonary embolism, is a major health concern in the United States (US), with over 900,000 thrombotic events reported each year, resulting in

AC

at least 600,000 hospitalizations and ~100,000 deaths [1-3]. Total treatment costs during the initial year after an index venous thromboembolism have been estimated to surpass $20,000 per patient [4]. Guidelines recommend a minimum of 3-months of anticoagulation for patients experiencing a venous thromboembolism, with the decision to use longer durations of anticoagulation based upon the balance between the competing risks of recurrent venous thromboembolism and clinically relevant

3

ACCEPTED MANUSCRIPT

bleeding [2-3]. Among patients experiencing an unprovoked venous thromboembolism, the risk of recurrent thrombosis is approximately 10% in the first year and ~30% in the first 5-years after anticoagulation therapy is stopped. Patients with provoking risk factors not related to a major surgery or trauma also have an elevated risk of recurrent venous thromboembolism after anticoagulation

CR IP T

discontinuation, albeit at lower rates of recurrence (5% in year 1 and 15% at year 5).

Two randomized trials have shown extended duration rivaroxaban use in patients experiencing a

venous thromboembolism can reduce patients’ risk of recurrent thrombosis compared to placebo or

AN US

aspirin without a significant increased risk of major bleeding [5-6]. Despite the abovementioned data, fear of bleeding complications often leads to an unwillingness to continue anticoagulation beyond the initial acute treatment phase, particularly in provoked venous thromboembolism patients [7-8]. As a paucity of real-world data evaluating prolonged anticoagulation with rivaroxaban following a

M

provoked venous thromboembolism exists, we sought to assess the effectiveness and safety of prolonged anticoagulation with rivaroxaban in patients experiencing a venous thromboembolism

ED

whom also had a provoking (major or minor, persistent or transient) risk factor.

PT

METHODS

CE

This report of the IMpact of Prolonged anticoagulation with Rivaroxaban On proVokEd Venous ThromboEmbolism recurrence (IMPROVE-VTE) study was written in compliance with the RECORD

AC

(REporting of studies Conducted using Observational Routinely-collected health Data) statement [9]. We performed a retrospective claims database analysis using US Truven MarketScan data from January 1, 2012 through March 31, 2017. Truven MarketScan [10] combines two separate databases, a commercial and a Medicare supplemental database, to cover all age groups; and as of March 2017, contains claims from 260 contributing employers, 40 health plans and government and public

4

ACCEPTED MANUSCRIPT

organizations representing ~240-million lives. The databases capture enrollment records, demographics, International Classification of Diseases, Ninth- and Tenth-Revision (ICD-9 and ICD-10) diagnosis codes, procedure codes, admission and discharge dates, outpatient medical services data and prescription dispensing records. All Truven MarketScan data are de-identified and are in compliance

CR IP T

with the Health Insurance Portability and Accountability Act of 1996 to preserve participant anonymity and confidentiality. This study was determined to not constitute research involving human subjects according to 45 CFR 46.102(f), and therefore, deemed exempt from institutional review board

AN US

oversight.

We included patients who had ≥1 primary discharge diagnosis code for deep vein thrombosis or pulmonary embolism in the primary coding position for a claim associated with a hospital or emergency department (ED) stay (this approach has been shown to have positive predictive value for identifying

M

venous thromboembolism of 95%, 95% confidence interval [CI]=93-97) [1, 11], a provoking risk factor (major persisting, minor persisting, minor transient and major transient), at least 3-months of

ED

continuous rivaroxaban treatment after the qualifying venous thromboembolism and ≥12-months of continuous medical and prescription insurance benefits prior to their venous thromboembolism.

PT

Patients were categorized as either continuing rivaroxaban or discontinuing anticoagulation (no

CE

anticoagulation or non-aspirin antiplatelet agents, but may have received aspirin) after the initial 3months of rivaroxaban treatment (index date). Patients with a venous thromboembolism or atrial

AC

fibrillation disease claim or a prescription claim for any anticoagulant during the baseline period were excluded from this study. The primary endpoints for this study were recurrent venous thromboembolism and major bleeding [1213]. Recurrent venous thromboembolism was defined as a subsequent hospitalization with a primary ICD-9 or -10 discharge diagnosis code for deep vein thrombosis and/or pulmonary embolism [11]. Major

5

ACCEPTED MANUSCRIPT

bleeding was defined as a subsequent hospitalization for a bleeding event and was identified using the validated claims coding algorithm published by Cunningham and colleagues in 2011 [14]. Patients were followed until occurrence of an endpoint, insurance disenrollment or up to 12-months post-index date (intention-to-treat analysis). Subgroup analysis was performed after stratifying patients based upon

CR IP T

the presence of specific provoking risk factor types. We grouped patients into provoking risk factor cohorts using the hierarchical classification scheme previously employed by EINSTEIN Extension and CHOICE trial investigators [15]: 1) major persisting (i.e. cancer); 2) minor persisting (i.e. inflammatory bowel disease, lower extremity paralysis, heart failure, stage III or worse chronic kidney disease,

AN US

hereditary or acquired thrombophilia); 3) minor transient (i.e. admitted to hospital for ≥3-consecutive days in past 3-months [surrogate for immobilization], hormonal therapy, pregnancy or puerperium, leg injury with impaired mobility); and 4) major transient risk factor (i.e. major surgery or trauma).

M

To adjust for potential confounding resulting from imbalances in baseline patient covariates, we utilized inverse probability-of-treatment weighting (IPTW) based on the propensity-score [16]. IPTW

ED

uses propensity-scores (predicted probability of discontinuing treatment given covariates identified using claims data in the 12-months prior to the index event [baseline period]) to generate patient-

PT

specific weights to adjust for any baseline imbalances. Propensity-scores were calculated using

CE

multivariable logistic regression incorporating patient demographics, comorbidities and known venous thromboembolism risk factors as well as concomitant non-oral anticoagulant medications identified

AC

during the 12-month baseline period and reported in Table 1. After propensity-scores were generated, the 2.5% of patients with the lowest and 2.5% with the highest propensity-scores were excluded (propensity-score trimming). Residual differences in baseline covariates between weighted rivaroxaban and warfarin cohorts were assessed by calculating absolute standardized differences; with differences <0.1 considered well-balanced [16].

6

ACCEPTED MANUSCRIPT

Baseline patient characteristics were analyzed using descriptive statistics. Categorical data are reported as proportions; while continuous data were reported as medians with 25%, 75% ranges. Cox proportional hazards regression was performed and results reported as hazard ratios (HRs) with accompanying 95% CIs. As all baseline covariates were sufficiently balanced after IPTW adjustment,

CR IP T

Cox regression only included the dichotomous variable indicating whether rivaroxaban was continued or discontinued as an independent variable. The proportional hazards assumption was checked using statistical tests and graphical diagnostics based on Schoenfeld residuals and was found to be met for all analyses. Statistical analyses were performed using PROC PHREG in SAS version 9.4 (SAS Inc., Cary,

AN US

NC) and a p-value <0.05 was considered significant in all cases. RESULTS

In total, 3806 patients with provoked venous thromboembolism received prolonged rivaroxaban

M

treatment and 1184 discontinuing treatment at 3-months were included in our analysis. Median age was 58 (48, 66) years, 49% were men and 47% had pulmonary embolism±deep vein thrombosis.

ED

Following IPTW, patients were found to be well-balanced on all covariates entered into the propensity-

PT

score logistic regression model as demonstrated by absolute standardized differences between the continued rivaroxaban and discontinuation cohorts of ≤0.06 in all cases.

CE

Among patients experiencing a provoked venous thromboembolism and treated with rivaroxaban for

AC

the first 3-months, continued rivaroxaban use beyond 3-months [median duration of additional rivaroxaban use of 3 (2, 5) months] was associated with a significant 44% lower hazard of recurrent venous thromboembolism without significantly altering major bleeding risk when compared to anticoagulation discontinuation with or without aspirin use (Table 2). The highest risk of recurrent thrombosis following provoked venous thromboembolism (3.1% in the discontinued anticoagulation

7

ACCEPTED MANUSCRIPT

cohort) appeared to occur in patients with a minor persisting risk factor; with continued rivaroxaban use associated with a significant 73% reduction in recurrent venous thromboembolism in this subgroup. DISCUSSION

CR IP T

This study in the US Truven MarketScan administrative claims databases evaluated the real-world effectiveness and safety of prolonged anticoagulation with rivaroxaban following a provoked venous thromboembolism. Despite the low absolute incidence of recurrent thrombosis observed in our study, we found that compared to anticoagulation discontinuation (no anticoagulation or non-aspirin

AN US

antiplatelet agents, but may have received aspirin), prolonged rivaroxaban use was associated with a 44% reduced relative hazard of recurrent venous thromboembolism without increasing major bleeding. Two randomized trials have shown extended duration rivaroxaban use in patients experiencing a venous thromboembolism can reduce patients’ risk of recurrent thrombosis compared to placebo or

M

aspirin without a significant increased risk of major bleeding [5-6]. In the EINSTEIN Extension trial,

ED

rivaroxaban 20 mg once daily was shown to reduce recurrent venous thromboembolism versus placebo (HR=0.18; 95%CI=0.09-0.39) [5]. Only 4 (0.7%) patients in the rivaroxaban group experienced a

PT

nonfatal major bleeding compared to zero patients in the placebo group (p=0.11). In the EINSTEIN CHOICE trial, rivaroxaban at a dose of 10 mg once daily was also shown to significantly reduce recurrent

CE

venous thromboembolism compared to aspirin 100 mg (HR=0.26; 95%CI=0.14-0.47) with similar rates

AC

of major bleeding (0.4% with rivaroxaban versus 0.3% with aspirin; p=0.50)[6]. The American College of Chest Physician (ACCP) guidelines [2] recommend patients with a provoked proximal deep vein thrombosis or pulmonary embolism receive treatment with anticoagulation for at least 3-months with the decision to use longer duration anticoagulation therapy based upon the balance between the competing risks of recurrent venous thromboembolism and clinically important bleeding. A pooled analysis [15] of the EINSTEIN Extension and CHOICE trials separately evaluating 8

ACCEPTED MANUSCRIPT

venous thromboembolism patients with unprovoked venous thromboembolism or a major persisting, minor persisting, minor transient and major transient provoking factors, suggests that not all patients have the same risk/incidence of recurrent venous thromboembolism. Not surprisingly, researchers demonstrated that unprovoked venous thromboembolism patients had the highest risk of venous

CR IP T

thromboembolism recurrence, but among provoked patients, those with a minor persisting risk factor had the highest risk of recurrent thrombosis (and were among the most prevalent type of provoked patients), while those with a major transient risk factor (major surgery or trauma) were had a negligible (0% in the pooled population) risk of recurrent venous thromboembolism. Thus, in order to assess the

AN US

benefit-to-risk balance for prolonged anticoagulation use among patients experiencing a venous

thromboembolism, it is important to determine whether a major or minor, persistent or transient risk factor is present [2, 12].

M

Of note, a previous claims database analysis evaluating prolonged duration rivaroxaban has been published [17]. In this analysis by Berger and colleagues which evaluated patients with unprovoked

ED

venous thromboembolism, prolonged rivaroxaban use was associated with a decreased risk of recurrent venous thromboembolism compared to those discontinuing therapy at 3- (0.57% versus

PT

1.19%; p=0.04), 6- (1.07% versus 2.10%; p=0.02) and 12-months of follow-up (1.45% versus 2.60%,

CE

p=0.02) [17]. No significant differences in the rate of major bleeding were observed between cohorts at any time point. Our findings are similar in overall conclusion to the study by Berger et al. (reduction in

AC

recurrent thrombosis without an increased risk of major bleeding), but is unique in that it is the first real-world study to specifically evaluate provoked venous thromboembolism patients and perform subanalysis based on the type of provoking factor present. Our real-world study’s findings were also generally consistent with the pooled EINSTEIN Extension/CHOICE analysis of patients with provoked venous thromboembolism (N=2669), each

9

ACCEPTED MANUSCRIPT

demonstrating prolonged rivaroxaban use to be associated with a reduction in recurrent venous thromboembolism; with the highest risk of recurrent thrombosis (3.1% in the discontinuation arm) and the most profound reduction (83% relative hazard reduction) observed in those with a minor, persisting risk factor. Agreement between these studies was likely facilitated by our defining of provoked venous

CR IP T

thromboembolism using a majority of the same risk factors as the EINSTEIN Investigators [15] (absent ‘family history of venous thromboembolism’ and ‘travel >8 hours’) and our implementation of the same hierarchal ranking of provoking factors (major persisting, minor persisting, minor transient and major transient risk). There were some differences between our findings and that of the pooled EINSTEIN

AN US

Extension/CHOICE analysis worth further discussion. First, the proportion of patients experiencing major or minor, persisting or transient risk factors varied somewhat between our study and the pooled EINSTEIN analysis (albeit, the presence of patients with minor persisting factors was common in both studies). The difference in provoked venous thromboembolism subtypes observed between studies

M

may in part be the due to differences in patient selection for prolonged duration anticoagulation by

ED

clinical trialists compared to clinicians in routine practice. Next, the incidence of recurrent venous thromboembolism in our anticoagulation discontinuation cohort was lower than that observed in the

PT

pooled EINSTEIN Extension/CHOICE analysis (the incidence of recurrent venous thromboembolism in the rivaroxaban groups were similar [ranging between 0.4% and 1.6%] across studies). The differing

CE

recurrent rates in the anticoagulation discontinuation group could be the result of a higher proportions of patients receiving aspirin versus no antithrombotic therapy in real-world practice (versus the

AC

proportion randomized in the clinical trials). Finally, our study found that continued rivaroxaban anticoagulation in patients with a major persisting risk factor (mostly cancer patients) was associated with a 27% higher venous thromboembolism recurrence rate and a 40% lower rate of major bleeding compared to patients in whom anticoagulation was stopped. Given the 95% CIs for both endpoints were wide and crossed the line of unity, it is possible the direction of effect seen was a chance finding.

10

ACCEPTED MANUSCRIPT

Additionally, as both thrombotic and bleeding event rates in identified major, persisting risk factor (mostly cancer) patients were lower than would by anticipated for a cancer-associated thrombosis population (based upon recent trials such as SELECT-D [18]), it is likely that included patients classified as having cancer did not have “active” cancer (experiencing cancer-associated thrombosis), but rather a

CR IP T

history of cancer (prompting a billing code for cancer in the past 12 months).Being as our study was based in claims data, we cannot rule out such misclassification bias as a potential alternative or supplemental explanation for the above variances [19].

AN US

As a retrospective analysis of claims data, this study has limitations worth discussion. First, both selection bias (selection of patients in a nonrandomized fashion) and misclassification bias (as

suggested above) are always important limitations in claims database studies and may impact a study’s internal validity [19]. Second, while propensity-score adjustment [16] can generate cohorts that are

M

comparable in numerous key characteristics, only those covariates estimable in the US Truven MarketScan databases could be used in the generation of propensity-scores. Regardless of the

ED

sophistication of the methodology and the number of covariates used in developing propensity-scores, residual confounding can never be fully excluded without adequate randomization of patients. Next,

PT

Truven MarketScan does not contain data allowing ascertainment of certain provoking risk factors such

CE

as family history of venous thromboembolism or recent prolonged travel nor does it have data on overthe-counter medication use. As a result, some provoked venous thromboembolisms may have been

AC

missed in our analysis. Moreover, as this is a US claims database analysis and aspirin is used primary over-the-counter in the US (and thus does not generate a prescription claim trackable in MarketScan), we are unable to determine whether control patients did or did not receive aspirin. While the inability to assess whether patients were prescribed aspirin is certainly a limitation of our analysis, it is important to note that aspirin was inferior to rivaroxaban for secondary prevention of venous thromboembolism when administered at either the 10 mg prophylactic or 20 mg treatment dose [5-6], and that the 11

ACCEPTED MANUSCRIPT

comparison made in the present real-world analysis (rivaroxaban 20/10 mg versus no antithrombotic or aspirin) mirrors that of the EINSTEIN investigators’ pooled analysis [15]. Lastly, as the results of EINSTEIN CHOICE [6] were published after the end date of claims data used in our study (3/2017), patients were not able to receive the prophylactic 10 mg dose of rivaroxaban currently approved by the

CR IP T

US Food and Drug Administration for reduction in the risk of recurrent venous thromboembolism in patients after at least 6-months of standard anticoagulant treatment [20]. CONCLUSIONS

AN US

Although the absolute incidence of recurrence was low, our study suggests continuing rivaroxaban after the initial 3-month period was associated with a decreased risk of recurrent venous thromboembolism, particularly in those with a minor persisting risk factor. The observed reduction in recurrent venous thromboembolism with prolonged rivaroxaban use was not associated with a significantly increased

M

risk of major bleeding.

ED

ACKNOWLEDGEMENTS

FUNDING

PT

None.

CE

This study was supported by Bayer AG, Berlin, Germany. ROLE OF THE FUNDER/SPONSOR

AC

The sponsor had no role in study design; in the collection, analysis and interpretation of data; in the writing of the report; or in the decision to submit the article for publication.

12

ACCEPTED MANUSCRIPT

REFERENCES 1. Heit JA, Cohen AT, Anderson Frederick AJ; on Behalf of the VTE Impact Assessment Group. Estimated annual number of incident and recurrent, non-fatal and fatal venous thromboembolism (VTE) events in the US. ASH Annu Meet Abstr 2005; 106:910.

and Expert Panel Report. Chest 2016; 149: 315–352.

CR IP T

2. Kearon C, Akl EA, Ornelas J, et al. Antithrombotic Therapy for VTE Disease: CHEST Guideline

3. Snow V, Qaseem A, Barry P, Hornbake ER, Rodnick JE, Tobolic T, Ireland B, Segal JB, Bass EB, Weiss KB, Green L, Owens DK; American College of Physicians; American Academy of Family

AN US

Physicians Panel on Deep Venous Thrombosis/Pulmonary Embolism. Management of venous thromboembolism: a clinical practice guideline from the American College of Physicians and the American Academy of Family Physicians. Ann Intern Med. 2007;146:204-10. 4. Coleman CI, Baugh C, Crivera C, Milentijevic D, Wang SW, Lu L, Nelson WW. Healthcare costs

ED

Med Econ. 2017;20:200-203.

M

associated with rivaroxaban or warfarin use for the treatment of venous thromboembolism. J

5. EINSTEIN Investigators, Bauersachs R, Berkowitz SD, Brenner B, Buller HR, Decousus H, Gallus

PT

AS, Lensing AW, Misselwitz F, Prins MH, Raskob GE, Segers A, Verhamme P, Wells P, Agnelli G, Bounameaux H, Cohen A, Davidson BL, Piovella F, Schellong S. Oral rivaroxaban for

CE

symptomatic venous thromboembolism. N Engl J Med. 2010;363:2499-510. 6. Weitz JI, Lensing AWA, Prins MH, Bauersachs R, Beyer-Westendorf J, Bounameaux H, Brighton

AC

TA, Cohen AT, Davidson BL, Decousus H, Freitas MCS, Holberg G, Kakkar AK, Haskell L, van Bellen B, Pap AF, Berkowitz SD, Verhamme P, Wells PS, Prandoni P; EINSTEIN CHOICE Investigators. Rivaroxaban or Aspirin for Extended Treatment of Venous Thromboembolism. N Engl J Med. 2017;376:1211-1222.

13

ACCEPTED MANUSCRIPT

7. Moodley O, Goubran H. Should lifelong anticoagulation for unprovoked venous thromboembolism be revisited? Thromb J. 2015;13:33. 8. Prandoni P, Lensing AW, Cogo A, Cuppini S, Villalta S, Carta M, Cattelan AM, Polistena P, Bernardi E, Prins MH. The long-term clinical course of acute deep venous thrombosis. Ann

CR IP T

Intern Med. 1996;125:1-7.

9. Benchimol EI, Smeeth L, Guttmann A, Harron K, Moher D, Petersen I, Sorensen HT, von Elm E, Langan SM, RECORD Working Committee. The Reporting of studies Conducted using Observational Routinely-collected health Data (RECORD) Statement. PLoS Med

AN US

2015;12:e1001885.

10. Hansen L. The Truven Health MarketScan Databases for life sciences researchers. Available at: https://truvenhealth.com/Portals/0/Assets/2017-MarketScan-Databases-Life-SciencesResearchers-WP.pdf [Last accessed December 25, 2017].

M

11. White RH, Garcia M, Sadeghi B, Tancredi DJ, Zrelak P, Cuny J, Sama P, Gammon H, Schmaltz S,

ED

Romano PS. Evaluation of the predictive value of ICD-9-CM codedadministrative data for venous thromboembolism in the United States. Thromb Res.2010;126:61-7.

PT

12. Kearon C, Ageno W, Cannegieter SC, Cosmi B, Geersing GJ, Kyrle PA; Subcommittees on Control of Anticoagulation, and Predictive and Diagnostic Variables in Thrombotic Disease.

CE

Categorization of patients as having provoked or unprovoked venous thromboembolism: guidance from the SSC of ISTH. J Thromb Haemost. 2016; 14: 1480-3.

AC

13. Baglin T, Luddington R, Brown K, Baglin C. Incidence of recurrent venous thromboembolism in relation to clinical and thrombophilic risk factors: prospective cohort study. Lancet. 2003;362:523-6.

14

ACCEPTED MANUSCRIPT

14. Cunningham A, Stein CM, Shung CP, Daugherty JR, Smalley WE, Ray WA. An automated database case definition for serious bleeding related to oral anticoagulant use. Pharmacoepidemiol Drug Saf 2011;20:560-6. 15. Prins MH, Lensing AWA, Prandoni P, Wells PS, Verhamme P, Beyer-Westendorf J, Bauersachs

CR IP T

R, Bounameaux H, Brighton TA, Cohen AT, Davidson BL, Decousus H, Kakkar AK, van Bellen B, Pap AF, Homering M, Tamm M, Weitz JI. Risk of recurrent venous thromboembolism according to baseline risk factor profiles. Blood Adv.2018;2:788-796.

16. Austin PC. An introduction to propensity score methods for reducing the effects of confounding

AN US

in observational studies. Multivariate Behav Res 2011;46:399–424.

17. Berger JS, Seheult, Laliberte MA, Crivera C, Lejeune D, Xiao Y, Schein J, Lefebvre P, Kaatz S. Clinical outcomes of prolonged anticoagulation with rivaroxaban after unprovoked venous thromboembolism. Res Pract Thromb Haemost 2018;2:58-68.

M

18. Gandhi SK, Salmon W, Kong SX, Zhao SZ. Administrative databases and outcomes

ED

assessment: an overview of issues and potential utility. J Manag Care Spec Pharm 1999;5:215– 222.

PT

19. Young AM, Marshall A, Thirlwall J, Chapman O, Lokare A, Hill C, Hale D, Dunn JA, Lyman GH, Hutchinson C, MacCallum P, Kakkar A, Hobbs FDR, Petrou S, Dale J, Poole CJ, Maraveyas A,

CE

Levine M. Comparison of an Oral Factor Xa Inhibitor With Low Molecular Weight Heparin in Patients With Cancer With Venous Thromboembolism: Results of a Randomized Trial (SELECT-

AC

D). J Clin Oncol. 2018;36:2017-2023.

20. XARELTO® (rivaroxaban) tablets, for oral use. Prescribing information. Janssen Pharmaceuticals, Inc., Titusville, NJ, October 2017. Available at: http://www.janssenlabels.com/package-insert/product-monograph/prescribinginformation/XARELTO-pi.pdf (Last accessed on June 23, 2018).

15

ACCEPTED MANUSCRIPT

Variable

PT

CE

AC

Absolute standardized difference

58 (48, 66) 44.7 48.7 46.6

58 (48, 67) 44.2 49.0 46.8

0.03 0.01 0.01 0.04

24.4

21.9

0.06

6.8 14.9 17.3 16.4 42.5 1.6 15.6

AN US

7.2 14.4 17.5 15.9 43.0 1.5 14.6 6.9 7.1 <0.01 4.9 0.1 16.8 2.5 27.8 4.8 0.7 4.2 17.2 4.7 3.1 25.6 15.3 46.2 59.1 20.4 9.8 15.9 6.8 14.2 15.7 4.5 10.1 15.5 24.9 21.9 2.1 3.3 2.1 5.7 13.3 3.2 1.5

M

ED

Demographics Age, years, median (25%, 75% range) Age >60 years Male gender Diagnosis of pulmonary embolism Hospitalization for ≥3-consecutive days in prior 3-months Comorbidities and risk factors 1 2 3 4 ≥5 History of major bleeding Cancer Chronic kidney disease Stage 3 Stage 4 Stage 5 Unstaged Renal dialysis Coagulopathy Central venous catheter Obesity Weight loss Lowe extremity paralysis Varicose veins Major surgery Trauma Fracture Chronic lung disease Pneumonia Hyperlipidemia Hypertension Diabetes mellitus Diabetes mellitus with complications Smoker Rheumatoid arthritis/collagen vascular disease Heart failure Depression Psychosis Liver dysfunction Hypothyroidism Anemia Alcohol abuse Drug abuse Ischemic stroke Transient ischemic attack Acute coronary syndrome Peripheral vascular disease Inflammatory bowel disease Pregnancy or puerperium

N=3806 %

Anticoagulation Discontinuation N=1184 %

CR IP T

Continued Rivaroxaban

6.5 7.2 <0.01 4.8 <0.1 16.9 2.5 27.0 4.7 0.9 4.3 16.9 4.1 3.4 25.4 15.1 45.6 58.9 19.8 9.5 16.0 6.8 14.5 16.0 4.5 10.0 15.5 25.2 21.8 2.0 3.4 2.0 5.6 13.7 3.3 1.8

0.02 0.01 0.00 0.01 0.01 0.01 0.03 0.01 0.00 0.03 0.00 0.00 0.00 0.00 0.03 0.00 0.02 0.00 0.01 0.03 0.02 0.00 0.00 0.01 0.00 0.02 0.01 0.00 0.00 0.01 0.01 0.00 0.00 0.00 0.01 0.00 0.00 0.01 0.00 0.01 0.01 0.01 0.02

16

ACCEPTED MANUSCRIPT

8.2 37.8 28.6 18.1 27.3 1.5 33.0 5.0 13.9 29.5 29.9

8.5 37.2 28.6 17.8 27.2 1.6 32.6 4.5 13.5 30.0 29.5

0.01 0.01 0.00 0.01 0.00 0.00 0.01 0.02 0.01 0.01 0.01

CR IP T

Medication Use Estrogens ACE inhibitors or ARBs Beta-blockers Calcium channel blockers Diuretics Digoxin Statin Insulin Non-insulin diabetes drugs Antidepressants Antiulcer drugs

AC

CE

PT

ED

M

AN US

Table 1. Baseline Characteristics of the Continued Rivaroxaban and Anticoagulation Discontinuation Cohorts

17

Anticoagulation Discontinuation

N

3806 778 1114 515 1399

1.0 1.0 0.9 0.7 1.3

1184 253 329 176 426

3806 778 1114 515 1399

1.0 1.1 1.5 0.5 0.7

N

1184 253 329 176 426

Anticoagulation Discontinuation Event Incidence, %

HR (95%CI)

1.7 0.8 3.1 1.1 1.7

0.56 (0.34-0.91) 1.27 (0.34-4.76) 0.27 (0.11-0.63) 0.59 (0.12-2.94) 0.79 (0.37-1.69)

1.1 1.8 2.1 0 0.4

0.87 (0.51-1.49) 0.60 (0.21-1.67) 0.64 (0.30-1.39) Not calculable 1.64 (0.45-5.88)

CR IP T

Continued Rivaroxaban

Recurrent Venous Thromboembolism Any provoking factor Major persisting Minor persisting Minor transient Major transient Major Bleeding Any provoking factor Major persisting Minor persisting Minor transient Major transient

Continued Rivaroxaban Event Incidence, %

AN US

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

M

Table 2. Incidence of Recurrent Venous Thromboembolism and Major Bleeding According to Risk Factors Associated with the Index Provoked Event CI=confidence interval; HR=hazard ratio; N=number;

18