1. Introduction
Epistaxis requiring emergency department (ED) treatment is a common bleeding complication in patients taking anticoagulants, accounting for 8% to 16% of treatment-emergent adverse events [1,2]. Limited existing evidence comparing epistaxis events between warfarin and direct oral anticoagulants (DOACs) has been mixed. The Rivaroxaban Once Daily Oral Direct Factor Xa Inhibition compared with Vitamin K Antagonism for Prevention of Stroke and Embolism Trial (ROCKET AF) showed rivaroxaban to have more clinically relevant epistaxis events compared with warfarin, with similar findings in the Japanese ROCKET AF (J-ROCKET AF) trial, although the results were not assessed for statistical significance [1,2]. Conversely, a recent nationwide Icelandic cohort study reported higher rates of epistaxis in patients taking warfarin when compared with DOACs [3].
The purpose of this study was to compare rates of ED visits for epistaxis between patients taking warfarin vs DOACs for atrial fibrillation (AF) or venous thromboembolism (VTE) enrolled in the Michigan Anticoagulation Quality Improvement Initiative (MAQI2) registry.
2. Methods
2.1. Data source
MAQI2 is a Blue Cross Blue Shield of Michigan-funded quality improvement registry of patients taking oral anticoagulation medications from 6 health systems across Michigan, which began patient enrollment in 2009 [4].
Data are retrospectively collected and analyzed from cohorts of patients on warfarin or DOAC therapy, starting from the time of enrollment through the earliest of the last follow-up before May 31, 2022, treatment discontinuation, or death. Data collection is performed by trained abstractors and verified by periodic audits, ensuring agreement with predefined data element definitions.
2.2. Patient population
Patients on warfarin were included from October 2009 to May 2022. Patients on DOACs were included from November 2015 to May 2022. Exclusion criteria included long-term warfarin therapy prior to MAQI2 enrollment or a primary indication for anticoagulation other than AF or VTE. Because patients enrolled in the MAQI2 DOAC registry could not be on any oral anticoagulant for at least 1 year prior to enrollment, no exclusion criteria for long-term oral anticoagulant use were necessary for this group.
2.3. Data collection and outcomes
Baseline data were collected at the time of enrollment. Outcome data were abstracted at each anticoagulation services interaction for warfarin patients or every 6 months via chart reviews for those on DOACs.
The primary outcome was the rate of clinically relevant epistaxis events, defined as epistaxis requiring ED evaluation and treatment. This incorporated major and nonmajor bleeding events. Secondary outcomes included median days from enrollment to the first epistaxis event, epistaxis events meeting criteria for major bleeding as defined by the International Society on Thrombosis and Haemostasis [5], hospital admissions for epistaxis, and recurrent epistaxis. The MAQI2 DOAC registry did not record the number of hospital days per admission and was not compared between groups.
2.4. Statistical analysis
Baseline information was compared using Student’s t-tests for continuous variables and chi-squared and Fisher’s exact tests for categorical variables. Inverse probability weighting (IPW) was used to mitigate differences between groups. Variables included in the IPW model included age, biological sex assigned at birth, non-Hispanic White, modified hypertension, abnormal renal/liver function, stroke, bleeding history of disposition, labile international normalized ratio, elderly (>65 years), drugs/alcohol concomitantly (HAS-BLED) score, comorbid conditions, concomitant antiplatelet or nonsteroidal anti-inflammatory drug (NSAID) use, and treatment indication. A modified HAS-BLED score was calculated for each patient at the time of enrollment, removing the labile international normalized ratio (INR) variable.
Poisson regression was utilized to compare outcomes after IPW. A 2-sided P <.05 was considered statistically significant for all analyses. Analyses were performed using SAS version 9.4 and R-64 4.1.3 software, developed by SAS Institute and the R Development Core Team, respectively.
3. Results
3.1. Participant data
After exclusions, a total of 10,557 and 3889 patients met inclusion for the warfarin and DOAC groups, respectively. In the DOAC group, 2750 (70.9%) patients received apixaban, 1105 (28.5%) received rivaroxaban, 23 (0.59%) received dabigatran, 3 (0.08%) received edoxaban, and 7 had no data on which DOAC was used. Several baseline characteristics were significantly different between cohorts. However, the IPW model demonstrated all standardized differences to be <0.1 (Table 1).
Table 1.
Patient characteristics.
| Warfarin (N = 10,557) | DOACs (N = 3889) | P value | |
|---|---|---|---|
| Demographics | - | - | - |
| Male sex | 5564 (52.7) | 1913 (49.2) | .0002 |
| Age (y), mean (SD) | 65.3 (15.4) | 68 (14.4) | <.001 |
| Non-Hispanic White | 8058 (76.3) | 3069 (78.9) | .001 |
| Black | 1702 (16.1) | 608 (15.6) | <.001 |
| Asian | 104 (1) | 55 (1.4) | <.001 |
| Native American | 29 (0.3) | 7 (0.2) | <.001 |
| Hispanic | 78 (0.7) | 16 (0.4) | <.001 |
| Other | 586 (5.6) | 134 (3.5) | <.001 |
| Comorbidities | - | - | - |
| Hypertension | 6941 (65.8) | 2784 (71.6) | <.0001 |
| Diabetes mellitus | 2717 (25.7) | 1029 (26.5) | .38 |
| Congestive heart failure | 1992 (18.9) | 624 (16.1) | <.0001 |
| Peripheral vascular disease | 576 (5.5) | 188 (4.8) | .14 |
| Alcohol use | 539 (5.1) | 236 (6.1) | .02 |
| Recent bleeding history | 291 (2.8) | 141 (3.6) | .007 |
| Remote bleeding history | 347 (3.3) | 224 (5.8) | <.0001 |
| Chronic liver failure | 267 (2.5) | 170 (4.4) | <.0001 |
| Bleeding diathesis | 53 (0.5) | 4 (0.1) | .001 |
| Concomitant medications | - | - | - |
| NSAIDs | 408 (3.9) | 128 (3.3) | .11 |
| Antiplatelets | 4430 (42) | 1323 (34) | <.0001 |
| Bleeding risk | - | - | - |
| Modified HAS-BLED score, mean (SD) | 2.5 (1.4) | 2.5 (1.4) | .12 |
| Treatment indication | - | - | - |
| AF | 5735 (54.3) | 2337 (60.1) | <.0001 |
| VTE | 4948 (46.9) | 1587 (40.8) | <.0001 |
| Anticoagulant | - | - | - |
| Dabigatran | N/A | 23 (0.59) | N/A |
| Rivaroxaban | N/A | 1105 (28.5) | N/A |
| Apixaban | N/A | 2750 (70.9) | N/A |
| Edoxaban | N/A | 3 (0.08) | N/A |
| Warfarin | 10,557 | N/A | N/A |
Data are presented as n (%) unless otherwise specified.
AF, atrial fibrillation; DOAC, direct oral anticoagulant; HAS-BLED, hypertension, abnormal renal/liver function, stroke, bleeding history of disposition, labile international normalized ratio, elderly (>65 years), drugs/alcohol concomitantly; N/a, not applicable; NSAID, nonsteroidal anti-inflammatory drug; VTE, venous thromboembolism.
3.2. Outcome data
Patients receiving warfarin had significantly higher rates of clinically relevant epistaxis than those on DOACs, with an observed rate of 2.06 events per 100 patient-years (events/100-py) vs 0.75 events/100-py and an estimated rate of 2.02 events/100-py (95% CI, 1.8, 2.2) vs 0.73 events/100-py (95% CI, 0.6, 1.0); P < .0001.
Of 378 ED visits in the warfarin group, 18 (4.8%) met the criteria for major bleeding, and 93 (24.6%) resulted in hospital admission. Of 57 ED visits in the DOAC group, 3 (5.3%) met the criteria for major bleeding, and 10 (17.5%) resulted in hospital admission. Patients on warfarin had higher observed and estimated rates of hospital admission for epistaxis than patients on DOACs (observed rate 0.50 events/100-py vs 0.13 events/100-py; estimated rate 0.49 events/100-py [95% CI, 0.41, 0.58] vs 0.14 events/100-py [95% CI, 0.098, 0.019]; P < .0001), as well as higher observed and estimated rates of epistaxis meeting major bleed criteria (observed rate 0.098 events/100-py vs 0.04 events/100-py; estimated rate 0.096 events/100-py [95% CI, 0.065, 0.143] vs 0.034 events/100-py [95% CI, 0.018, 0.065]; P < .007). Primary and secondary outcome data are shown in Table 2.
Table 2.
Outcomes.
| Warfarin N = 10,557 |
DOACs N = 3889 |
P value | |
|---|---|---|---|
| Median years of follow-up (IQR) | 0.59 (2.1) | 1.5 (2.5) | - |
| Median days from enrollment to first epistaxis event (IQR) | 483 (864) | 369 (719) | .28 |
| ED visits for epistaxis, n | 378 | 57 | - |
| Major epistaxis events | 18 (4.8) | 3 (5.3) | - |
| Hospital admissions | 93 (24.6) | 10 (17.5) | - |
| Outcomes in the no. of events per 100 patient-years of follow-up (95% CI) | |||
| ED visits for epistaxis – observed | 2.06 | 0.75 | - |
| ED visits for epistaxis – estimated | 2.02 (1.86-2.21) | 0.73 (0.63-0.083) | <.0001 |
| Major epistaxis – observed | 0.098 | 0.04 | - |
| Major epistaxis – estimated | 0.096 (0.065-0.143) | 0.034 (0.018-0.065) | <.007 |
| Hospital admissions – observed | 0.50 | 0.13 | - |
| Hospital admissions – estimated | 0.49 (0.41-0.58) | 0.14 (0.098-0.019) | <.0001 |
Data are presented as n (%) unless otherwise specified.
DOAC, direct oral anticoagulant; ED, emergency department.
4. Discussion
Our observational cohort study showed that patients receiving warfarin for AF or VTE had significantly higher rates of clinically relevant epistaxis compared with patients taking DOACs, aligning with a recent study in Icelandic patients [3] but contrasting with others [1,2].
It is unclear why event rates differ between studies. One reason could be differences in epistaxis rates between DOACs. More than 70% of our study population was taking apixaban, which is associated with lower bleeding rates than other DOACs when compared with warfarin [2,6]. The limited existing data have also shown that rivaroxaban has higher epistaxis rates than apixaban [3]. Another reason could be the intensity of anticoagulation at the time of the event. Since INR values could not be compared between groups, they were not included in this study. It is unclear if patients on warfarin had supratherapeutic INRs at the time of ED presentation. Studies including INR values in patients taking vitamin K antagonists have shown average INR values to be mostly therapeutic upon ED presentation [3,7,8], while another found most patients to be supratherapeutic [9]. The differences in epistaxis event rates may be multifactorial: a combination of differences in event rates between DOACs and the intensity of anticoagulation at the time of the epistaxis event.
This study also shows that patients on warfarin have higher rates of epistaxis meeting criteria for major bleeding and hospital admissions compared with DOACs. This adds to previous literature, which has shown mixed results [3,8]. Again, this difference is likely multifactorial and may be impacted by our DOAC group being primarily on apixaban.
There are notable limitations to our study. Our study population was restricted to the state of Michigan within anticoagulation clinics participating in the MAQI2 initiative. However, our study population was 21% to 24% non-White, whereas similar studies had primarily homogenous, non-Hispanic, White study populations [3,8].
Given the retrospective and observational nature of this registry, there is a risk of confounding bias as patients were not randomized. Reasonable measures were taken to correct for confounding, illustrated by the IPW model demonstrating all weighted standardized differences to be <0.1.
This study highlights the importance of efforts to reduce epistaxis and health care utilization in patients on oral anticoagulants. In patients with an increased risk of epistaxis, switching from warfarin to a DOAC may be beneficial if no contraindications exist. Efforts should aim to reduce preventable ED visits for epistaxis in DOAC patients, such as through educational programs similar to those for warfarin [10]. Additional research to further characterize differences in epistaxis rates between DOACs and identify strategies to reduce epistaxis-related ED visits is needed.
Acknowledgments
We would like to recognize the abstractors at each of the participating MAQI2 sites for the collection of data that made this project possible.
Funding
This project was funded by The Blue Cross Blue Shield of Michigan (BCBSM). BCBSM had no role in the conception, design, or conduction of the study, no role in the collection, management, analysis, or interpretation of the data, including drafting of the manuscript, review for important intellectual content, or approval of the manuscript.
Author contributions
Study conception and design: K.S.B., X.K., B.H., G.B., and S.K. Data acquisition, statistical analysis, and interpretation: K.S.B., X.K., B.H., S.K., and D.D. Drafting of the manuscript: K.S.B., X.K., and B.H. Revision of manuscript for important intellectual content: K.S.B., X.K., B.H., M.A., D.D., G.B., and S.K. Final approval of the manuscript for submission: S.K. Obtaining funding: G.B.
Relationship Disclosure
K.S.B., B.H., X.K., M.A., and D.D. report no conflicts of interest. G.B. reports receiving personal fees from Pfizer/Bristol Myers Squibb, Janssen, Bayer, Abbot Vascular, Sanofi, Anthos, and Boston Scientific and grant funding from National Heart, Lung and Blood Institute, Boston Scientific, and Blue Cross Blue Shield of Michigan. He also reports being on the Board of Directors for the Anticoagulation Forum. S.K. reports receiving personal fees from AstraZeneca, Bristol Myers Squibb, Boston Scientific, Gilead, Inari, Janssen, Pfizer, Phase Bio, and Antos. He also reports receiving grant funding from Osmosis Research, Janssen, Bristol Myers Squibb, Bayer, and the National Institutes of Health and is on the Board of Directors for the Anticoagulation Forum, the Pulmonary Embolism Response Team Consortium, and the Scientific Advisory Board for the National Blood Clot Alliance.
Footnotes
Handling Editor:
Appendices
References
- 1.Hori M., Matsumoto M., Tanahashi N., Momomura S., Uchiyama S., Goto S., et al. Rivaroxaban vs. warfarin in Japanese patients with atrial fibrillation – the J-ROCKET AF study. Circ J. 2012;76:2104–2111. doi: 10.1253/circj.cj-12-0454. [DOI] [PubMed] [Google Scholar]
- 2.Patel M.R., Mahaffey K.W., Garg J., Pan G., Singer D.E., Hacke W., et al. Rivaroxaban versus warfarin in nonvalvular atrial fibrillation. N Engl J Med. 2011;365:883–891. doi: 10.1056/NEJMoa1009638. [DOI] [PubMed] [Google Scholar]
- 3.Ingason A.B., Rumba E., Hreinsson J.P., Agustsson A.S., Lund S.H., Palsson D.A., et al. Warfarin is associated with higher rates of epistaxis compared to direct oral anticoagulants: a nationwide propensity score-weighted study. J Intern Med. 2022;292:501–511. doi: 10.1111/joim.13498. [DOI] [PubMed] [Google Scholar]
- 4.Barnes G.D., Kaatz S., Golgotiu V., Gu X., Leidal A., Kobeissy A., et al. Use of warfarin for venous thromboembolism prophylaxis following knee and hip arthroplasty: results of the Michigan Anticoagulation Quality Improvement Initiative (MAQI2) J Thromb Thrombolysis. 2013;35:10–14. doi: 10.1007/s11239-012-0766-y. [DOI] [PubMed] [Google Scholar]
- 5.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. doi: 10.1111/j.1538-7836.2005.01204.x. [DOI] [PubMed] [Google Scholar]
- 6.Granger C.B., Alexander J.H., McMurray J.J., Lopes R.D., Hylek E.M., Hanna M., et al. Apixaban versus warfarin in patients with atrial fibrillation. N Engl J Med. 2011;365:981–992. doi: 10.1056/NEJMoa1107039. [DOI] [PubMed] [Google Scholar]
- 7.Glikson E., Chavkin U., Madgar O., Sagiv D., Nakache G., Yakirevitch A., et al. Epistaxis in the setting of antithrombotic therapy: a comparison between factor Xa inhibitors, warfarin, and antiplatelet agents. Laryngoscope. 2019;129:119–123. doi: 10.1002/lary.27400. [DOI] [PubMed] [Google Scholar]
- 8.Sauter T.C., Hegazy K., Hautz W.E., Krummrey G., Ricklin M.E., Nagler M., et al. Epistaxis in anticoagulated patients: fewer hospital admissions and shorter hospital stays on rivaroxaban compared to phenprocoumon. Clin Otolaryngol. 2018;43:103–108. doi: 10.1111/coa.12904. [DOI] [PubMed] [Google Scholar]
- 9.Smith J., Siddiq S., Dyer C., Rainsbury J., Kim D. Epistaxis in patients taking oral anticoagulant and antiplatelet medication: prospective cohort study. J Laryngol Otol. 2011;125:38–42. doi: 10.1017/S0022215110001921. [DOI] [PubMed] [Google Scholar]
- 10.Neshewat J., Wasserman A., Alexandris-Souphis C., Haymart B., Feldeisen D., Kong X., et al. Reduction in epistaxis and emergency department visits in patients taking warfarin after implementation of an education program. Thromb Res. 2021;199:119–122. doi: 10.1016/j.thromres.2021.01.007. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.