Prediction of major bleeding in patients receiving DOACs for venous thromboembolism: A prospective cohort study

Journal Pre-proof Prediction of major bleeding in patients receiving DOACs for venous thromboembolism: A prospective cohort study

Maria Cristina Vedovati, Alessandra Mancuso, Lucia Pierpaoli, Ugo Paliani, Serenella Conti, Alessandra Ascani, Giulia Galeotti, Francesco Di Filippo, Carla Caponi, Giancarlo Agnelli, Cecilia Becattini PII:

S0167-5273(19)32687-7

DOI:

https://doi.org/10.1016/j.ijcard.2019.11.105

Reference:

IJCA 28138

To appear in:

International Journal of Cardiology

Received date:

24 May 2019

Revised date:

4 October 2019

Accepted date:

13 November 2019

Please cite this article as: M.C. Vedovati, A. Mancuso, L. Pierpaoli, et al., Prediction of major bleeding in patients receiving DOACs for venous thromboembolism: A prospective cohort study, International Journal of Cardiology(2019), https://doi.org/10.1016/ j.ijcard.2019.11.105

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. 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.

© 2019 Published by Elsevier.

Journal Pre-proof Word count text 3011 Word count abstract 249

Prediction of major bleeding in patients receiving DOACs for venous thromboembolism: a prospective cohort study.

ro

of

Bleeding risk in venous thromboembolism on DOACs

Maria Cristina Vedovati*, MD, Alessandra Mancuso*, MD, Lucia Pierpaoli†, MD, Ugo Paliani‡,

-p

MD, Serenella Conti§, MD, Alessandra Ascani¶, MD, Giulia Galeotti‡, MD, Francesco Di Filippo†,

lP

re

MD, Carla Caponi‡, MD, Giancarlo Agnelli*, MD, Cecilia Becattini*, MD.

Jo ur

na

* Internal, Vascular and Emergency Medicine - Stroke Unit, University of Perugia, Perugia, Italy. This author takes responsibility for all aspects of the reliability and freedom from bias of the data presented and their discussed interpretation. † Emergency Medicine, S. Maria Delle Croci Hospital, Ravenna, Italy. This author takes responsibility for all aspects of the reliability and freedom from bias of the data presented and their discussed interpretation. ‡ Department of Medicine, Hospital of Città di Castello, Città di Castello, Italy. This author takes responsibility for all aspects of the reliability and freedom from bias of the data presented and their discussed interpretation. § Division of Cardiology, S. Matteo degli Infermi Hospital, Spoleto, Italy. This author takes responsibility for all aspects of the reliability and freedom from bias of the data presented and their discussed interpretation. ¶ Emergency Medicine, S. Maria Hospital, Terni, Italy. This author takes responsibility for all aspects of the reliability and freedom from bias of the data presented and their discussed interpretation

Corresponding author: Maria Cristina Vedovati, MD Internal and Cardiovascular Medicine – Stroke Unit University of Perugia, Perugia, Italy Phone +39 075 5786424 [email protected] [email protected]

1

Journal Pre-proof M. C. Vedovati, A. Mancuso, L. Pierpaoli, U. Paliani, S. Conti, A. Ascani, G. Galeotti, F. Di Filippo and C. Caponi have no conflicts to declare. G. Agnelli reports consulting fees from Bayer and Daiichi Sankyo, lecture fees from Bristol-Myers Squibb and Sanofi-Aventis. C. Becattini reports lecture fees from Bayer and Boehringer Ingelheim.

No financial support was received for this study.

Jo ur

na

lP

re

-p

ro

of

Key words: Anticoagulants; Hemorrhage; Prospective Studies; Venous Thromboembolism.

2

Journal Pre-proof Abstract Background In the direct oral anticoagulants (DOACs) era, extended anticoagulation is an attractive strategy after venous thromboembolism (VTE). The role of currently available bleeding risk scores for VTE patients treated with DOACs in clinical practice is undefined. Methods Consecutive patients with VTE were included in a prospective multicenter cohort at the initiation of treatment with DOACs. The role of ATRIA, HAS-BLED, Kuijer, ORBIT, RIETE and VTE-BLEED scores in predicting major bleeding (ISTH definition) while on DOAC treatment was

of

assessed.

ro

Results Overall, 1034 patients were included and followed for one year or until the end of treatment

-p

or the occurrence of major bleeding. During study period, 26 major bleedings occurred in 25

significant predictors of major bleedings.

re

patients (2.8% patient-year). Anemia, bleeding history and creatinine clearance <60 ml/min were

lP

The predictive value of bleeding risk scores was modest. In the 12-month study period, ORBIT (HR

na

intermediate-high vs. low risk patients 3.62, 95% CI 1.65-7.94 and c-statistics 0.645, 95% CI 0.5230.767) and VTE-BLEED (HR high vs. low 16.11, 95% CI 2.18-119.09 and c-statistics 0.674, 95%

Jo ur

CI 0.593-0.755) score significantly predicted major bleeding. The lowest incidence of major bleeding (0.3%) was observed in the low-risk category of VTE-BLEED, while the highest (7.1%) in the high-risk category of ORBIT.

Conclusions In a real-life cohort of patients with VTE treated with DOACs, the predictive value of currently available bleeding risk scores was modest and not statistically different. Whether these scores can be used for decision making on anticoagulation should be assessed in management studies.

3

Journal Pre-proof Introduction Anticoagulants are the mainstay for the treatment of VTE as these agents reduce recurrences by about 90%3. Direct oral anticoagulants (DOACs) have become the agents of choice for the treatment of VTE in the initial, long-term and extended phase4 thanks to improved safety and ease of use. Bleeding risk is the main trade-off of anticoagulant treatment and prediction of bleeding risk in an individual patient basis is a major everyday clinical issue4,5. Prediction rules have been developed to

of

estimate the bleeding risk during anticoagulant therapy, initially in patients with atrial fibrillation

ro

(AF) and then in patients with VTE6-12. Some controversial issues should be mentioned concerning

-p

the derivation and validation studies of these scores in VTE patients. Among these are the inclusion

re

of patients with both provoked or unprovoked VTE 6,7, the different duration of the follow-up periods 6,7 and the combination of major and clinically relevant non-major bleeding as clinical

lP

outcome 6. Moreover, the clinical value of these scores is debated due to limited accuracy and lack

na

of external validation. As an additional limit, most of these studies were carried out in patients receiving vitamin K oral anticoagulants (VKAs) and limited data are available on the predictive

Jo ur

value of bleeding risk scores in patients with VTE treated with DOACs. The VTE-BLEED score was derived and initially validated in the setting of randomized clinical trials.13 14 The predictive value of the VTE-BLEED score was recently confirmed in a prospective cohort of patients receiving rivaroxaban or conventional anticoagulation. 15 The aim of this study was to evaluate the role of six bleeding risk scores (Kuijer, RIETE VTEBLEED and ATRIA HAS-BLED and ORBIT) in predicting major bleeding in a prospective cohort of patients with VTE treated with DOACs in clinical practice.

4

Journal Pre-proof Materials and methods Patients and study design In this study, we analyze data from a multicenter prospective cohort of patient treated with DOACs for objectively confirmed VTE. Patients were enrolled at 5 Italian hospitals (Città di Castello, Perugia, Ravenna, Spoleto and Terni) between March 2014 and November 2017. Consecutive patients aged 18 years or older identified at the in- and outpatient services of the participating study sites could be included if treatment with DOACs was started for an objectively

of

confirmed diagnosis of PE or DVT. These patients could be either anticoagulation naïve or switched

ro

to DOACS from VKAs, heparin or fondaparinux. Exclusion criteria were contraindication to

-p

DOACs, unavailability for study follow-up or refusal of informed consent.

re

The study period started at the time of DOAC prescription. The choice of the individual DOAC and the duration of anticoagulant treatment was left at discretion of the attending physician. Patients

na

whichever came first.

lP

were followed for one year or until the end of treatment or the occurrence of major bleeding,

The study protocol conforms to the ethical guidelines of the 1975 Declaration of Helsinki and was

Study outcome

Jo ur

approved by the Ethical Committee and/or Institutional Review Boards of the participating centres.

The study outcome was major bleeding occurring during treatment with DOACs. Major bleeding was defined according to the ISTH criteria 16: symptomatic bleeding in a critical organ (intracranial, intraspinal, intraocular, retroperitoneal, intraarticular, intramuscular with compartment syndrome, pericardial); bleeding causing a fall in the hemoglobin level of at least 2 g/dl or leading to transfusion of at least two units of whole blood or red cells; fatal bleeding.

Data collection For all included patients the following data were collected: age, gender, weight, height, 5

Journal Pre-proof comorbidities (hypertension, previous stroke, cancer, alcohol abuse, previous bleeding), labile INR in patients previously receiving VKAs, features of the index VTE (initial manifestation as PE or DVT, association with temporary or persistent risk factors or unprovoked), type and dose of DOACs, date of DOAC prescription, previous treatment with heparin or VKA, concomitant medications (non-steroidal anti-inflammatory and antiplatelet agents) and results of blood tests (full blood count, renal and liver function). Creatinine clearance was estimated by Cockcroft–Gault formula 17.

of

The ATRIA, HAS-BLED, Kuijer, ORBIT, RIETE and VTE-BLEED bleeding risk scores were

ro

calculated for each patient 6-10,13,14. Definitions of individual items of the bleeding risk scores are

-p

reported in the Appendix 1. All patients entered a follow-up program by office visit or by phone call

re

every 6 months or whenever clinical issues occurred. At each follow-up visit, data on bleeding events were collected.

lP

For the purpose of this analysis, predictors were defined consistently with their definition in each

Jo ur

Statistical analysis

na

individual bleeding score (see Appendix 1).

Main baseline characteristics of patients were reported as frequencies for categorical data and as mean ± standard deviation (SD) for continuous data. Bleeding risk scores are reported as continuous value and/or as risk category (low, moderate or high) (e-Table 1). The risk for major bleeding was calculated using Cox proportional hazard model for each score. The discriminative power of different scores to correctly classify subjects into a risk category was assessed by the c-statistic in the overall study period (12 months) through Cox regression analysis. The c-statistic represents the concordance between predicted and observed bleeding events, with c=0.5 for prediction no better than chance and c=1.0 for perfect discriminative ability 18. A landmark analysis was performed for each score during the first 6 months of treatment and between month 6 to 12 of treatment. 6

Journal Pre-proof The bleeding risk assessment tool with the best performance (identified by the risk estimate and by the c-statistic value) was further tested with the following sub-analyses: patients receiving DOACs within or after 30 days from the index VTE; patients with PE ± DVT or with DVT only as index event; patients with unprovoked or provoked VTE.

Jo ur

na

lP

re

-p

ro

of

Statistical analysis was performed with SPSS software (version 25).

7

Journal Pre-proof Results Overall, 1034 patients were included in the study. Flow diagram for study inclusion is reported in Figure 1. Patients received a DOAC for the treatment of isolated DVT in 50.8% (525 patients), of isolated PE in 13.5% (140 patients), and of both PE and DVT in 35.7% (369 patients). Venous thromboembolism was unprovoked in 597 patients (57.7%) and associated with active cancer in 164 patients (15.9%). The majority of patients were on treatment with rivaroxaban (82%). Further details are reported in Table 1.

of

The mean follow-up was 312 days (median 264.5 days). Overall, 26 major bleedings occurred in 25

ro

patients (2.8% patient-year): 14 major bleedings occurred in the first 6 months of treatment and 12

-p

from month 6 to 12 in the 654 patients who remained on treatment beyond 6 months.

re

Sites of major bleedings were: gastrointestinal in 11 patients (42%), genitourinary in 4 (15%), traumatic intracranial in 3 (11%), spontaneous intracranial in 2 (8%), pericardial in 2 (8%), epistaxis

lP

in 2 (8%) and muscular or spinal cord hematoma in 1 patient each (4%).

na

Overall, 5 major bleedings occurred in the 164 patients with active cancer (5.2% patient-year): 3 in the first 6 months of treatment and 2 in the 78 patients who remained on treatment beyond 6

Jo ur

months. Basal characteristics of cancer patients are reported in e-Table 2. No significant increase in bleeding risk was observed in patients with compared to those without cancer (HR 1.93, 95% CI 0.72 to 5.17, p=0.191). Mortality in cancer patients was 31% and was mainly due to end stage disease.

Among bleeding risk factors, only anemia, bleeding history and creatinine clearance < 60 ml/min were significant predictors of major bleedings while increasing age, hypertension and cancer were not (e-Table 3).

Performance of the bleeding risk assessment scores The prevalence of patients categorized as low, intermediate or high risk according to bleeding risk score is reported in Table 2. The lowest incidence of major bleeding (0.3%) was observed in the low 8

Journal Pre-proof risk category of VTE-BLEED which includes 38% of patients. The highest incidence of major bleeding (7.1%) was observed in the high-risk category of ORBIT which includes 10.9% of patients. Concerning the performance of the individual scores, no significant differences were observed. A significantly higher risk of major bleeding was observed in patients in the high or intermediate risk categories compared to the low risk category for the ATRIA score (hazard ratio [HR] 3.05, 95% confidence interval [CI] 1.38 to 6.71, p=0.006) and for the ORBIT score (HR 3.62, 95% CI 1.65 to

of

7.94, p=0.001); in patients in the high risk category compared to the low risk category for the VTE-

ro

BLEED score (HR 16.11, 95% CI 2.18 to 119.09, p=0.006). No significant differences were

-p

observed among risk categories of the HAS-BLED, Kuijer and RIETE scores (Table 2). The high

re

vs. low risk category of the 3-risk level score were also compared and reported in Table 2. The risk

Figure 1.

lP

for bleeding associated to specific risk category of the individual scores over time is reported in e-

na

The discriminative power was not significantly different among evaluated scores (Figure 2). In the 12-month study period, an acceptable discrimination was found for the VTE-BLEED (c-statistics

Jo ur

0.674, 95% CI 0.593-0.755) and modest for ORBIT (c-statistics 0.645, 95% CI 0.523-0.767) scores and RIETE (c-statistics 0.604, 95% CI 0.510-0.697) (e-Table 4). No advantage was observed in the accuracy of the HAS-BLED score when it was used as a 3-level (high-, intermediate or low risk) respect to a 2-level (high- or low-risk) tool (e-Table 5). The discriminative value of each bleeding risk score was also calculated separately during the first 6 months of treatment and between the sixth and twelfth month of treatment (e-Table 4). The highest discriminative value, although no significantly different among scores, was observed for the ORBIT in the first 6 months, and for the VTE-BLEED in the sixth to twelfth month of treatment.

Subgroup analyses The VTE-BLEED score showed the best predictive value for major bleeding complications and it 9

Journal Pre-proof was selected for the pre-specified subgroup analyses. Patients with a VTE-BLEED score ≥2 showed a higher rate of major bleedings compared to those with a VTE-BLEED score <2 in all the prespecified subgroups. This difference was significantly different in the subgroup of patients with DVT only as index event (3.6% vs. 0.5% in the high and low risk group, respectively: HR 8.71, 95% CI 1.12 to 67.45) and in those with unprovoked VTE (3.6% vs. 0.4%, HR 9.86, 95% CI 1.28 to 75.80). The discriminative value of the VTE-BLEED score was good for almost all the analyzed

Jo ur

na

lP

re

-p

ro

of

subgroups. See Appendix 2 and e-Table 6 for further details.

10

Journal Pre-proof Discussion In this prospective cohort of patients with VTE treated with DOACs, the predictive value of the currently available bleeding risk scores was modest. No significant difference was observed among evaluated bleeding risk scores. Among all scores, the VTE-BLEED seemed to better identify patients at low risk while ORBIT those at high risk for major bleeding. The VTE-BLEED showed an acceptable discriminative value for major bleedings. This ability was confirmed either in the first 6 months of treatment or in the 6 to 12-month period of treatment.

of

Estimation of bleeding risk is an everyday clinical issue for physicians dealing with VTE. Several

ro

clinical scores have been proposed to stratify patients receiving oral anticoagulant treatment in

-p

categories at different bleeding risk. Bleeding risk scores for patients on anticoagulant treatment for

re

VTE were mainly derived and validated in cohorts of patients receiving VKAs 12,19-22 with the exception of the VTE-BLEED score 13,14. DOACs have become the treatment of choice for VTE

lP

according to current guidelines 4. A limited amount of data is available on the predictive value of

na

bleeding risk scores in cohorts of patients with VTE treated with DOACs outside the setting of clinical trials. This evaluation is essential because of different pharmacokinetics,

Jo ur

pharmacodynamics and risk profile (overall risk as well as prevalent bleeding sites) of DOACs in comparison to VKAs. These differences could translate in a major role of some predictors for bleeding during VKAs (labile INR) and of others for bleeding during DOACs (renal failure). Thus, the performance of bleeding risk scores could differ between patients treated with DOACs or VKAs. In this view, our findings of the apparent good performance of the VTE-BLEED for prediction of major bleeding in VTE patients receiving DOACs is conceivable as this is the only score that was derived and validated from VTE patients receiving DOACs. Recently, the prognostic value of VTE-BLEED in predicting major bleeding was explored in the population of the Xalia study 15. Authors found that VTE-BLEED is able to differentiate VTE patients with higher from those with lower risk of major bleeding during long-term anticoagulation treatment. The main implications of our results concern the management of anticoagulant treatment beyond 11

Journal Pre-proof the initial three months. Current guidelines recommend continuing the anticoagulant therapy until the benefit in the prevention of recurrent thrombosis out-weights the increased risk of bleeding events 4. Uncertainty on which bleeding risk score should be used persists despite the findings of our study. If VTE-BLEED score is being applied, physicians could confidently continue anticoagulation in patients at low risk of bleeding (about 38% of the overall patients and 45% of unprovoked). In our study cohort, the high rate (62%) of patients categorized as high risk according to the VTE-

of

BLEED score (vs. 27% of the score validation cohorts) may be explained by the high prevalence of

ro

active cancer, age of 60 years or over and creatinine clearance lower than 60 ml/min (higher than

-p

that observed in the derivation and validation cohorts of the VTE-BLEED score) 13-15. As a

re

consequence, the rate of major bleeding was higher in our cohort (2.8% patient-year) compared to those of the derivation (1.4% and 2.0% in the dabigatran and warfarin groups, respectively) and

lP

validation cohorts (1.4% and 1.6% in the edoxaban and warfarin groups, respectively). The different

compared to derivation.

na

predictive value of individual risk factors is a common finding in the external validation process as

Jo ur

For patients categorized as high-risk, actually the majority according to the VTE-BLEED score, the balance between the risk of bleeding (in this case 3.7%, 5.5% according to the upper limit of the confidence interval) and the risk of recurrent VTE requires evaluation on an individual patient basis, thus limiting the usefulness of this score. Whether the 3.7% risk is high enough to discontinue treatment probably depends on the risk of recurrent VTE if anticoagulants are discontinued. As an alternative, the ORBIT score showed the highest incidence of major bleeding (7.1%) in patients categorized as high risk (11%). In this view different scores offer different categorization performance. Further evidence, potentially from management study is required. Bleeding rates in our study were similar to those reported in real life cohorts of anticoagulated patients with VTE 12,19, 23. These differences reflect the selection of patients at low risk of bleeding for phase 3 trials. As in previous studies, we also evaluated the performance of scores derived from 12

Journal Pre-proof patients with atrial fibrillation (ATRIA, HAS-BLED and ORBIT). Consistently with these studies, we observed a poor performance by applying bleeding risk scores created for patients with atrial fibrillation to patients with VTE.12,20 It should be taken into account that VTE patients are usually about 20 years younger than atrial fibrillation patients and have a different comorbidity profile (less renal failure and more cancer). In our study, the VTE-BLEED score was able to predict major bleeding events, although with a wide confidence interval and a modest c-statistics. Moreover, the difference with the other bleeding

of

scores was not statistically significant. The stratification in bleeding risk categories differed across

ro

individual bleeding risk scores (Table 2). Patients categorized as low risk ranged from 15.6% by the

-p

Kuijer score up to 96.4% by the HAS-BLED score. As well, the proportion of high-risk patients

re

ranged from 3.6% with the HAS-BLED score to 62.1% with the VTE-BLEED score. These differences can be explained by the inclusion of different predictors in the individual risk scores. In

lP

fact, among all the items included in the scores, only three were confirmed to be predictors of major

na

bleeding in our population. Moreover, the low prevalence of some predictors in our cohort (e.g. abnormal liver function) can have affected the results of our study.

Jo ur

Major bleeding was the study outcome. VTE-BLEED and Kuijer scores have been derived for the prediction of clinically relevant bleedings. However, if these scores should be used to tailor the duration of anticoagulant treatment in VTE patients, the bleeding risk assessment should be probably based on hard clinical outcome, mainly in those patients with an estimated high risk for recurrent VTE. Our study has some limitations. Patients were included in the study at the time of the prescription of DOACs, which differed among patients. Moreover, prescription of DOAC not always corresponded to time of diagnosis of VTE and some early bleeding events may have been lost. To deal with these points, a subgroup analysis was performed in patients receiving DOACs within 30 days from the index VTE. The low number of outcome events should be taken into account when interpreting study results especially of the subgroup analyses as reflected by the presence of wide confidence 13

Journal Pre-proof intervals. Furthermore, number of patients with VTE not receiving DOACs was not collected, thus our study results only apply to patients treated with DOACs. Major bleedings were locally adjudicated by the attending physician. Although central adjudication is essential for clinically relevant non major bleeding, it may be less relevant for major bleedings which are defined by precise criteria16. We did not evaluate the performance of the ACCP score 4, EINSTEIN and Hokusai scores 25,26. Despite these limitations, the prospective data collection and the consecutiveness of the population

of

provides a real-life picture of the accuracy of predicting major bleeding of the currently available

ro

bleeding risk scores. Our results confirm some previous findings and add original information on

-p

the predictive value of bleeding risk scores in one of the largest cohort of patients with VTE treated

re

with DOACs. Conclusions

lP

This study provides original data on currently available bleeding risk scores in predicting major

na

bleeding in a contemporary cohort of patients with VTE treated with DOACs. The predictive value of evaluated bleeding risk scores was modest and not statistically different across scores. The VTE-

Jo ur

BLEED seemed to better identify patients at low risk while ORBIT those at high risk for major bleeding. Whether these scores can be used for decision making on anticoagulation should be assessed in a management study.

Acknowledgments M.C. Vedovati is the guarantor. M.C. Vedovati, A. Mancuso, G. Agnelli and C. Becattini made a substantial contribution to concept and design of the work, to the acquisition, analysis, and interpretation of data, to draft the work and revising it critically for important intellectual content and final approval of the version to be published. L. Pierpaoli, U. Paliani, S. Conti, A. Ascani, G. Galeotti F. Di Filippo and C. Caponi made a 14

Journal Pre-proof substantial contributions to concept of the work, drafting of the work and revising it critically for important intellectual content and final approval of the version to be published.

Support/funding

Jo ur

na

lP

re

-p

ro

of

No financial support was received for this study.

15

Journal Pre-proof References 1. Raskob GE, Angchaisuksiri P, Blanco AN, et al.; ISTH Steering Committee for World Thrombosis Day. Thrombosis: a major contributor to global disease burden. Arterioscler Thromb Vasc Biol. 2014;34:2363-71. 2. Di Nisio M, van Es N, Büller HR. Deep vein thrombosis and pulmonary embolism. Lancet. 2016; 388:3060-3073. 3. Boutitie F, Pinede L, Schulman S, et al. Influence of preceding length of anticoagulant treatment

of

and initial presentation of venous thromboembolism on risk of recurrence after stopping treatment:

ro

analysis of individual participants' data from seven trials. BMJ. 2011; 342:d3036.

-p

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

re

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

5. Konstantinides SV, Meyer G, Becattini C, Bueno H, Geersing GJ, Harjola VP, Huisman MV,

lP

Humbert M, Jennings CS, Jiménez D, Kucher N, Lang IM, Lankeit M, Lorusso R, Mazzolai L,

na

Meneveau N, Ní Áinle F, Prandoni P, Pruszczyk P, Righini M, Torbicki A, Van Belle E, Zamorano JL; ESC Scientific Document Group. 2019 ESC Guidelines for the diagnosis and management of

Jo ur

acute pulmonary embolism developed in collaboration with the European Respiratory Society (ERS). Eur Heart J. 2019 Aug 31.

6. Kuijer PM, Hutten BA, Prins MH, Büller HR. Prediction of the risk of bleeding during anticoagulant treatment for venous thromboembolism. Arch Intern Med. 1999;159:457-460. 7. Ruíz-Giménez N, Suárez C, González R, et al.; RIETE Investigators. Predictive variables for major bleeding events in patients presenting with documented acute venous thromboembolism. Findings from the RIETE Registry. Thromb Haemost. 2008; 100:26-3. 8. O'Brien EC, Simon DN, Thomas LE, et al. The ORBIT bleeding score: a simple bedside score to assess bleeding risk in atrial fibrillation. Eur Heart J. 2015;36:3258-64. 9. Pisters R, Lane DA, Nieuwlaat R, de Vos CB, Crijns HJ, Lip GY. A novel user-friendly score (HAS-BLED) to assess 1-year risk of major bleeding in patients with atrial fibrillation: the Euro 16

Journal Pre-proof Heart Survey. Chest. 2010;138:1093-1100. 10. Fang MC, Go AS, Chang Y, et al. A new risk scheme to predict warfarin-associated hemorrhage: The ATRIA (Anticoagulation and Risk Factors in Atrial Fibrillation) Study. J Am Coll Cardiol. 2011;58:395-401. 11. Gage BF, Yan Y, Milligan PE, et al. Clinical classification schemes for predicting hemorrhage: results from the National Registry of Atrial Fibrillation (NRAF). Am Heart J. 2006;151:713-719. 12. Riva N, Bellesini M, Di Minno MN, et al. Poor predictive value of contemporary bleeding risk

of

scores during long-term treatment of venous thromboembolism. A multicentre retrospective cohort

ro

study. Thromb Haemost. 2014;112:511-21.

-p

13. Klok FA, Hösel V, Clemens A, et al. Prediction of bleeding events in patients with venous

re

thromboembolism on stable anticoagulation treatment. Eur Respir J. 2016;48:1369–1376. 14. Klok FA, Barco S, Konstantinides SV. External validation of the VTE-BLEED score for

na

Haemost. 2017;117:1164-1170.

lP

predicting major bleeding in stable anticoagulated patients with venous thromboembolism. Thromb

15. Klok FA, Barco S, Turpie AGG, Haas S, Kreutz R, Mantovani LG, Gebel M, Herpers M, Bugge

Jo ur

JP, Kostantinides SV, Ageno W. Predictive value of venous thromboembolism (VTE)-BLEED to predict major bleeding and other adverse events in a practice-based cohort of patients with VTE: results of the XALIA study. Br J Haematol. 2018;183(3):457-465. 16. Schulman S, Kearon C; Subcommittee on Control of Anticoagulation of the Scientific and Standardization Committee of the International Society on Thrombosis and Haemostasis. Definition of major bleeding in clinical investigations of antihemostatic medicinal products in non-surgical patients. J Thromb Haemost. 2005;3:692-694. 17. Levey AS, Stevens LA, Schmid CH, et al.; CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration). A new equation to estimate glomerular filtration rate. Ann Intern Med. 2009;150:602-12. 18. Hanley JA, McNeil BJ. The meaning and use of the area under a receiver operating 17

Journal Pre-proof characteristic (ROC) curve. Radiology. 1982;143:29-36. 19. Poli D, Antonucci E, Testa S, Cosmi B, Palareti G, Ageno W; FCSA Italian Federation of Anticoagulation Clinics. The predictive ability of bleeding risk stratification models in very old patients on vitamin K antagonist treatment for venous thromboembolism: results of the prospective collaborative EPICA study. J Thromb Haemost. 2013;11:1053-8. 20. Scherz N, Méan M, Limacher A, et al. Prospective, multicenter validation of prediction scores for major bleeding in elderly patients with venous thromboembolism. J Thromb Haemost.

of

2013;11:435-43.

ro

21. Piovella C, Dalla Valle F, Trujillo-Santos J, et al.; RIETE Investigators. Comparison of four

-p

scores to predict major bleeding in patients receiving anticoagulation for venous thromboembolism:

re

findings from the RIETE registry. Intern Emerg Med. 2014;9:847-52. 22. Nieto JA, Solano R, Trapero Iglesias N, et al.; RIETE Investigators. Validation of a score for

lP

predicting fatal bleeding in patients receiving anticoagulation for venous thromboembolism.

na

Thromb Res. 2013;132:175-9.

23. Kresoja KP, Ebner M, Rogge NIJ, Sentler C, Keller K, Hobohm L, Hasenfuß G, Konstantinides

Jo ur

SV, Pieske B, Lankeit M. Prediction and prognostic importance of in-hospital major bleeding in a real-world cohort of patients with pulmonary embolism. Int J Cardiol. 2019 Mar 16. pii: S01675273(18)34801-0. doi: 10.1016/j.ijcard.2019.03.017. [Epub ahead of print] 24. Palareti G, Antonucci E, Mastroiacovo D, et al. The American College of Chest Physician score to assess the risk of bleeding during anticoagulation in patients with venous thromboembolism. J Thromb Haemost. 2018. doi: 10.1111/jth.14253. 25. Di Nisio M, Ageno W, Rutjes AW, Pap AF, Büller HR. Risk of major bleeding in patients with venous thromboembolism treated with rivaroxaban or with heparin and vitamin K antagonists. Thromb Haemost. 2016;115:424-32. 26. Di Nisio M, Raskob G, Büller HR, et al. Prediction of major and clinically relevant bleeding in patients with VTE treated with edoxaban or vitamin K antagonists. Thromb Haemost. 18

Journal Pre-proof 2017;117:784-793.

Tables and Figures Table 1. Baseline features of the population. Table 2. Distribution of the population and predictive value according to bleeding risk scores

of

categories.

ro

Figure 1. Flow diagram for study inclusion.

Jo ur

na

lP

re

-p

Figure 2. Receiver operating characteristic curves of the bleeding risk scores.

19

Journal Pre-proof Table 1. Baseline features of the population.

Patients (1034) Age (years), mean±SD (range)

66.7±16.2 (19-99)

Age ≥75 years, n (%)

402 (38.9)

Female gender, n (%)

499 (48.3)

PE±DVT, n (%)

509 (49.2) 76.6±15.4 (35-150)

BMI (kg/m2), mean±SD (range)

26.7±5.0 (13-54)

of

Weight (kg), mean±SD (range)

238 (23.0)

Active cancer, n (%)

164 (15.9)

-p

ro

Hypertension, n (%)

History of bleeding, n (%)

48 (4.6)

383 (37.0)

0.88±0.27 (0.2-2.4)

Apixaban, n (%) Dabigatran, n (%) Edoxaban, n (%)

Jo ur

Unprovoked VTE, n (%)

na

lP

Creatinine value (mg/dl), mean±SD (range) Creatinine clearance (ml/min), mean±SD (range) Creatinine clearance <50 ml/min, n (%)

re

Anemia, n (%)

Rivaroxaban, n (%)

84.4±28.9 (20-300) 120 (14.4) 597 (57.7) 141 (13.6) 21 (2.0) 23 (2.2) 849 (82.1)

BMI= body mass index; CrCl= creatinine clearance; SD= standard deviation; VTE= venous thromboembolism.

20

Journal Pre-proof Table 2. Distribution of the population and predictive value according to bleeding risk scores categories.

Major Bleeding (n=25)

(0-3) (4) (≥5)

807 (78.0%) 37 (3.6%) 190 (18.4%)

14 (1.7%) 1 (2.7%) 10 (5.3%)

(<3) (≥3)

997 (96.4%) 37 (3.6%)

23 (2.3%) 2 (5.4%)

161 (15.6%) 716 (69.2%) 157 (15.2%)

1 (0.6%) 20 (2.8%) 4 (2.5%)

HR high vs low risk

(95% CI, p)

(95% CI, p)

of

3.05 (1.38-6.71, 0.006)

3.28 (1.46-7.38, 0.004)

4.84 (0.65-35.74, 0.123)

5.66 (0.63-51.02, 0.122)

13 (1.6%) 4 (3.7%) 8 (7.1%)

3.62 (1.65-7.94, 0.001)

4.64 (1.92-11.19, 0.001)

207 (20.0%) 785 (75.9%) 42 (4.1%)

1 (0.5%) 24 (3.1%) 0

6.33 (0.86-46.76, 0.071)

n.e.+

392 (37.9%) 642 (62.1%)

1 (0.3%) 24 (3.7%)

-

16.11 (2.18-119.09, 0.006)

-p

re

lP

814 (78.7%) 107 (10.3%) 113 (10.9%)

ro

2.81 (0.66-11.92, 0.162)

Jo ur

Kuijer 6 Low (0) Intermediate (1.3-2.9) High (≥3.5) 8 ORBIT Low (0-2) Intermediate (3) High (≥4) 7 RIETE Low (0) Intermediate (1-4) High (≥4.5) 13,14 VTE-BLEED Low (<2) High (≥2)

HR intermediate-high vs low risk

-

na

ATRIA 10 Low Intermediate High HASBLED 9 Low High

Overall population (n=1034)

+not estimable as no event occurred in the high risk group

21

Journal Pre-proof Highlights

Jo ur

na

lP

re

-p

ro

of

- In a real-life cohort of patients with VTE receiving DOACs major bleeding was 2.8% patient-year - Anemia, bleeding history and creatinine clearance <60 ml/min were predictors of major bleedings - In patients with VTE, the predictive value of currently available bleeding risk scores was modest - The VTE-BLEED score <2 points identified patients at low risk for major bleeding

22

Figure 1

Figure 2