Kidney Res Clin Pract > Epub ahead of print
Mo, Kim, Yang, Kwon, and on behalf of the Vascular Access Working Group of the Korean Society of Dialysis Access: Safety of direct oral anticoagulants in hemodialysis patients with venous thromboembolism: an analysis using the Korean National Health Insurance Service database

Abstract

Background

Limited data exist regarding the safety of direct oral anticoagulants in hemodialysis patients with venous thromboembolic disease. This study aims to investigate the safety of direct oral anticoagulants in hemodialysis patients using national data.

Methods

The National Health Insurance Service database was retrospectively queried to identify chronic kidney disease patients who took direct oral anticoagulants for venous thromboembolism from 2008 to 2019. Bleeding complications and all-cause mortality were compared between 118 hemodialysis patients (HD group) and 227 matched chronic kidney disease patients not undergoing hemodialysis (CKD group).

Results

The use of direct oral anticoagulants among chronic kidney disease patients, with or without dialysis, increased over time. The incidence rate of all-cause mortality per 100 person-years was 38.1 in the HD group and 10.5 in the CKD group (adjusted hazard ratio [HR], 3.28; 95% confidence interval [CI], 2.27–4.75; p < 0.001). When considering death as a competing risk, there was no significant difference in gastrointestinal bleeding (adjusted HR, 1.61; 95% CI, 0.91–2.88; p = 0.115) and intracranial bleeding (adjusted HR, 1.86; 95% CI, 0.73–4.74; p = 0.193) between the HD and CKD groups.

Conclusion

In comparison to chronic kidney disease patients not on hemodialysis, the major bleeding risk, including gastrointestinal and intracranial bleeding, was comparable in hemodialysis patients using direct oral anticoagulants for venous thromboembolism.

Introduction

Venous thromboembolism (VTE), comprising pulmonary embolism (PE) and deep vein thrombosis (DVT), is a prevalent vascular disease with a lifetime risk of around 8% among adults in the United States, gaining increasing awareness [1]. The mainstay for the treatment of VTE is anticoagulation therapy, incorporating medications such as heparin, vitamin K antagonists (VKA), and direct oral anticoagulants (DOACs) [2]. DOACs are increasingly replacing VKAs due to comparable efficacy and a lower risk of complications such as bleeding, as evidenced by several studies [36]. Additionally, DOACs offer several advantages over VKAs, including predictable pharmacokinetics that reduce the need for routine monitoring and allow for fixed-dose administration, coupled with fewer dietary interactions [7].
However, VKAs have not been entirely replaced, especially in cases involving end-stage kidney disease (ESKD), where the safety of DOACs is not fully established since they are excreted through the kidneys [8]. Considering the annual increase in ESKD patients and the significantly higher VTE risk in this population compared to the general population, VTE in ESKD patients emerges as a major health issue [9]. Nevertheless, there is a dearth of research focusing on the safety of using DOACs in ESKD patients with VTE.
Therefore, in this study, our aim is to investigate the differences in major complications arising from the use of DOACs in patients undergoing hemodialysis compared to those with chronic kidney disease (CKD) not undergoing hemodialysis using national claim data.

Methods

Data sources

This nationwide cohort study utilized National Health Insurance Service (NHIS) data covering 98% of the Korean population. The claim data from January 2008 to December 2019 were used. Approval for data collection and publication was granted by the Institutional Review Board of Asan Medical Center (No. 2020-0576), which waived the requirement for written informed consent due to the study’s retrospective nature and deidentification of the study subjects. All methods were performed in accordance with the relevant guidelines and regulations.

Study population and study outcome

The flow diagram is depicted in Fig. 1. First, we identified patients using the N18 or V001 codes and defined hemodialysis patients as those who had been receiving regular dialysis for more than 3 months, as indicated by the hemodialysis procedure codes (O7020/7021). Subsequently, patients older than 18 years were selected, while those who acquired the V001 code or underwent dialysis for over 3 months during the look-back period were excluded as the temporal relationship with the other diagnosis could not be determined. Among these patients, those with VTE, including DVT or PE (International Classification of Disease, 10th Revision [ICD-10] codes; I80.1-3, I80.8-9, I82.9, and I26) were identified. We included patients who initiated DOAC following their VTE diagnosis and excluded cases where there were more than 30 days between the VTE diagnosis and the initiation of DOAC therapy. To select the control group, we used the diagnosis code for advanced CKD (N183-5), and among them, patients who also had a VTE code and were treated with DOACs were included as we did with the study group. The control group was matched to the study group as closely as possible based on age, sex, and other variables, including body mass index (BMI) and comorbidities, with a 1:2 ratio.
Demographic information, ICD-10 diagnosis codes, procedure codes, prescriptions, and survival data from both inpatient and outpatient services, along with health screening data (e.g., health behaviors such as smoking or alcohol, laboratory tests), and comorbidities were collected for each patient. The ICD-10 codes for comorbidities were considered within 1 year of entry.

Outcome measures

The index date of each group was defined as the first event of DVT or PE during the study period. The main safety outcome was the annualized rate of major gastrointestinal bleeding or intracranial hemorrhage, and it was compared between the groups. The secondary outcome was all-cause mortality. Information on death and death dates was retrieved from the registration records of the NHIS. Supplementary Table 1 (available online) summarizes the ICD-10 codes utilized to define bleeding events.

Statistical analysis

Demographic data were presented as a frequency for categorical variables. The statistical differences in patient characteristics between the matched groups were computed using the standardized mean difference (SMD). An SMD less than 0.1 was considered a sign of balance between the groups. The control group was matched to the dialysis group in a 1:2 ratio by closely aligning age, sex, BMI, and comorbidities. Safety outcomes were assessed by Kaplan-Meier curve analysis. The Cox proportional hazard regression model was used to determine the hazard ratio (HR) and corresponding 95% confidence intervals (CI) and adjusted for age and sex. All p-values of <0.05 were considered significant. The statistical analysis was performed using SAS Enterprise Guide version 7.1 (SAS Institute) and R software version 4.0.3 (R Foundation for Statistical Computing).

Results

Baseline characteristics

During the study period, 186 dialysis patients with VTE were newly initiated on DOACs. After 1:2 matching, 118 dialysis patients (HD group) were successfully matched to 227 control group patients (CKD group).
Baseline characteristics of the included patients before and after matching were summarized in Table 1. After matching, the CKD group was older and had more comorbidities such as hypertension, diabetes mellitus, ischemic heart disease, chronic obstructive lung disease, and peripheral artery occlusive disease than the HD group. Additionally, there were more social history factors, such as smoking and drinking, in the CKD group compared with the HD group.

Number of venous thromboembolism and direct oral anticoagulant prescriptions

A total of 186 dialysis patients and 1,195 CKD patients who were not on dialysis were prescribed DOACs for VTE during the study period. As shown in Table 1, the number of DOAC prescriptions per year in patients with or without dialysis has been increasing in recent years. Annual DOAC prescriptions between the HD and CKD groups were similar before matching.

Safety between the groups

The cumulative incidence curves for three outcomes (all-cause mortality, gastrointestinal bleeding, and intracranial hemorrhage) are depicted in Fig. 2.
Table 2 shows that after matching, the incidence rate for all-cause mortality was 10.54 and 38.13 per 100 patient-years for the CKD and HD groups, respectively, with significant differences between the groups (p < 0.001).
The gastrointestinal bleeding rates were 4.54 and 11.52 per 100 patient-years for the CKD and HD groups, respectively (adjusted HR, 2.28; 95% CI, 1.25–4.17; p = 0.008). However, in Cox regression analyses treating death as a competing risk, the HR for gastrointestinal bleeding in the HD group versus the CKD group was 1.61 (95% CI, 0.91–2.88; p = 0.11). The intracranial hemorrhage rates were 1.33 and 4.52 per 100 patient-years for the CKD and HD groups, respectively, and there was also no significant difference between the CKD and HD groups (adjusted HR, 1.86; 95% CI, 0.73–4.74; p = 0.19).

Discussion

VTE, encompassing both PE and DVT, stands as a prevalent vascular disease with a documented increase in prevalence over time [10]. Historically, VKAs were the cornerstone of treatment. However, the past decade has witnessed a significant shift towards DOACs, characterized by their advantageous features such as the absence of the need for laboratory monitoring and dose adjustment [11]. Notably, due to the substantial renal elimination of DOACs, patients with a creatinine clearance (CrCl) of less than 30 mL/min, were advised to use VKAs [12]. The four DOACs—dabigatran, rivaroxaban, edoxaban, and apixaban—demonstrate renal excretion rates of 80%, 66%, 50%, and 27%, respectively [7]. Nevertheless, in practice, DOACs are being employed even in ESKD patients with a CrCl of less than 15 mL/min [13]. Our study also revealed an increasing trend in the prescription of DOACs among CKD patients irrespective of whether they are receiving dialysis. However, there is a noticeable gap in the existing literature concerning the efficacy and safety of DOACs in ESKD patients undergoing dialysis [8]. Consequently, our primary emphasis was on examining the safety aspect of DOACs within this population.
While our study indeed concentrates on safety outcomes, it is crucial to acknowledge that efficacy, particularly measured by recurrent VTE, plays a vital role in assessing the benefits of DOAC use in CKD patients. Bauersachs et al. [14] demonstrated that rivaroxaban effectively prevents recurrent VTE across all renal function categories when compared to enoxaparin/VKA (Pinteraction = 0.72), with recurrence rates varying from 1.8% to 4.8% in patients with different levels of renal impairment. Similarly, Goldhaber et al. [15] found that dabigatran demonstrated better efficacy, indicated by the rates of VTE and VTE-related death, compared to warfarin in patients with renal insufficiency (CrCl 50 to <80 mL/min: 1.9% with dabigatran vs. 1.6% with VKA; CrCl 30 to <50 mL/min: 0% with dabigatran vs. 4.1% with VKA). However, research on the efficacy of DOACs for VTE in ESKD patients undergoing dialysis is limited, with only retrospective studies available. Reed et al. [16] reported no significant difference in efficacy between apixaban and warfarin in ESKD patients undergoing dialysis (4.4% vs. 28.6%, p = 0.99) [16]. While existing research suggests that DOACs appear to be non-inferior to VKAs, the limited number of studies poses challenges in drawing definitive conclusions.
The primary focus of our study was the safety outcome, typically defined as major bleeding or clinically relevant nonmajor bleeding. This is a crucial consideration in the use of DOACs in CKD patients, given their higher bleeding risk compared to the general population [17]. Although DOACs are excreted renally to some extent, leading to reduced clearance and increased exposure in CKD patients [7], most studies have consistently demonstrated that DOACs are similar to or even safer than VKAs in terms of safety, regardless of the stage of CKD. In a study by Bauersachs et al. [14], HRs for major bleeding with rivaroxaban-enoxaparin/VKA were 0.79 (95% CI, 0.46–1.36) for normal renal function, 0.44 (95% CI, 0.24–0.84) for mild renal impairment, and 0.23 (95% CI, 0.06–0.81) for moderate renal impairment (Pinteraction = 0.034). Similarly, in a study by Goldhaber et al. [15], there was no significant difference in safety outcomes between dabigatran and VKA in patients with moderate renal impairment (in subgroup aged <75 years: dabigatran 10.0% vs. VKA 12.9%; in subgroup aged ≥75 years: dabigatran 11.8% vs. VKA 9.6%). Examining safety outcomes in ESKD patients on dialysis, the apixaban group exhibited fewer overall bleeding events (18.9% vs. 42.0%, p = 0.01) and major bleeding events (5.4% vs. 22%, p = 0.01) compared to the warfarin group [16]. In our study, we did not specifically focus on differences between medications but rather observed bleeding complications associated with DOAC use in CKD patients comparing those receiving dialysis with those who were not. Importantly, there was no significant difference in complication rates between the two groups. Fig. 2 suggests that both gastrointestinal and intracranial bleeding may appear higher in the HD group, but this difference is primarily driven by higher mortality in patients on dialysis, and it is not statistically significant. Therefore, there is no compelling reason to completely abandon the use of DOACs in CKD patients on dialysis.
Despite these findings, our study has several limitations. Firstly, it was retrospective, and the number of patients was limited. Secondly, the NHIS data, designed for reimbursement rather than research, lacks detailed clinical information. For instance, we were unable to differentiate the stages of CKD, making it impossible to gauge the severity of CKD in the control group, necessitating caution in interpreting the study results. We were also unable to differentiate between the types of DOACs due to NHIS data policy, preventing us from analyzing the risks associated with each specific medication.
In addition, our study was limited in evaluating the efficacy of VTE treatment in this population. Thus, our focus was on presenting safety outcomes related to DOACs rather than assessing efficacy. The nature of our dataset also restricts our findings to CKD patients, making a direct safety comparison with the general population unfeasible [18]. Consequently, our results do not imply that the safety of using DOACs in dialysis patients is comparable to that in the general population. Nevertheless, these comparisons can offer more targeted and specific information for these distinct patient populations, allowing for a nuanced understanding of the safety profile of DOACs within the CKD spectrum. Additionally, we believe our study possesses strengths owing to the comprehensive coverage of medical practices in Korea provided by the NHIS database. This minimizes the likelihood of missing claims and patient loss, enhancing the robustness of our findings.

Supplementary Materials

Supplementary data are available at Kidney Research and Clinical Practice online (https://doi.org/10.23876/j.krcp.24.093).

Notes

Conflicts of interest

All authors have no conflicts of interest to declare.

Funding

This research was supported by a research grant from the Korean Society of Dialysis Access.

Data sharing statement

The data that support the findings of this study are available from National Health Insurance Service of Korea but restrictions apply to the availability of these data, which were used under license for the current study, and so are not publicly available. Data are however available with permission of National Health Insurance Service of Korea.

Authors’ contributions

Conceptualization: All authors

Data curation, Supervision: HK, SBY, YK

Formal analysis: HM, HK, SBY

Funding acquisition: HK

Methodology: All authors

Writing–original draft: HM

Writing–review & editing: HM, HK

All authors read and approved the final manuscript.

Figure 1.

Flow diagram for patient selection.

CKD, chronic kidney disease; DOAC, direct oral anticoagulant; ESKD, end-stage kidney disease; HD, hemodialysis; VTE, venous thromboembolism.
j-krcp-24-093f1.jpg
Figure 2.

Cumulative incidence of major bleeding events and all-cause mortality in CKD patients with or without dialysis.

Cumulative incidence of gastrointestinal bleeding (A) and intracranial hemorrhage (B), excluding death as a competing event. (C) Cumulative risk of all-cause mortality among CKD patients with or without dialysis.
CKD, chronic kidney disease; HD, hemodialysis.
j-krcp-24-093f2.jpg
Table 1.
Baseline characteristics of venous thromboembolism patients prescribed direct oral anticoagulants
Characteristic Before matching
After matching
HD group CKD group SMD HD group CKD group SMD
No. of patients 186 1,195 118 227
Year of DOAC initiation 0.088 0.234
 2010–2013 7 (3.8) 55 (4.6) 5 (4.2) 17 (7.5)
 2014 14 (7.5) 91 (7.6) 8 (6.8) 21 (9.3)
 2015 25 (13.4) 155 (13.0) 17 (14.4) 37 (16.3)
 2016 37 (19.9) 226 (18.9) 21 (17.8) 47 (20.7)
 2017 53 (28.5) 311 (26.0) 34 (28.8) 54 (23.8)
 2018 50 (26.9) 357 (29.9) 33 (28.0) 51 (22.5)
Age (yr) 70.68 ± 12.7 83.17 ± 4.4 1.312 78.57 ± 5.5 80.11 ± 4.0 0.322
Female sex 106 (57.0) 703 (58.8) 0.037 74 (62.7) 144 (63.4) 0.015
Comorbidities
 Hypertension 148 (79.6) 1,013 (84.8) 0.136 97 (82.2) 195 (85.9) 0.101
 Diabetes 122 (65.6) 557 (46.6) 0.390 77 (65.3) 113 (49.8) 0.317
 Ischemic heart disease 10 (5.4) 38 (3.2) 0.109 7 (5.9) 2 (0.9) 0.281
 COPD 56 (30.1) 319 (26.7) 0.076 40 (33.9) 59 (26.0) 0.173
 Cerebrovascular disease 35 (18.8) 303 (25.4) 0.676 26 (22.0) 56 (24.7) 0.062
 PAOD 41 (22.0) 318 (26.6) 0.107 27 (22.9) 72 (31.7) 0.199
 Malignancies 49 (26.3) 358 (30.0) 0.08 32 (27.1) 60 (26.4) 0.016
Medication
 Antiplatelet 118 (63.4) 697 (58.3) 0.105 76 (64.4) 143 (63.0) 0.029
 Lipid-lowering agent 100 (53.8) 648 (54.2) 0.009 64 (54.2) 140 (61.7) 0.151
Body mass index (kg/m2)
 <18.5 3 (1.6) 23 (1.9) 0.330 2 (1.7) 2 (0.9) 0.568
 ≥18.5 and <23 33 (17.7) 178 (14.9) 21 (17.8) 40 (17.6)
 ≥23 and <25 25 (13.4) 156 (13.1) 15 (12.7) 34 (15.0)
 ≥25 and <30 19 (10.2) 262 (21.9) 11 (9.3) 60 (26.4)
 ≥30 7 (3.8) 40 (3.3) 4 (3.4) 15 (6.6)
 Not available 99 (53.2) 536 (44.9) 65 (55.1) 76 (33.5)
Smoking 0.229 0.449
 Never smoker 64 (34.4) 499 (41.8) 40 (33.9) 114 (50.2)
 Previous smoker 13 (7.0) 122 (10.2) 10 (8.5) 26 (11.5)
 Current smoker 10 (5.4) 38 (3.2) 3 (2.5) 11 (4.8)
 Not available 99 (53.2) 536 (44.9) 65 (55.1) 76 (33.5)
Alcohol 0.190 0.517
 None 79 (42.5) 576 (48.2) 50 (42.4) 127 (55.9)
 Mild-to-moderate 6 (3.2) 61 (5.1) 2 (1.7) 20 (8.8)
 Heavy 2 (1.1) 20 (1.7) 1 (0.8) 3 (1.3)
 Not available 99 (53.2) 538 (45.0) 65 (55.1) 77 (33.9)

Data are expressed as number only, number (%), or mean ± standard deviation.

CKD, chronic kidney disease; COPD, chronic obstructive pulmonary disease; DOAC, direct oral anticoagulant; HD, hemodialysis; PAOD, peripheral arterial occlusive disease; SMD, standardized mean difference.

Table 2.
Safety outcomes of DOACs among CKD patients with or without dialysis
CKD group (n = 227) HD group (n = 118) p-value
All-cause mortality
 No. of event 65 72
 Sum of person-years 616.5 188.8
 Incidence rate for 100 person-years (95% CI) 10.54 (8.14–13.44) 38.13 (29.84–48.02)
 Adjusted HR (95% CI)a 1 (reference) 3.281 (2.269–4.746) <0.001
Gastrointestinal bleeding
 No. of event 27 20
 Sum of person-years 594.8 173.6
 Incidence rate for 100 person-years (95% CI) 4.54 (2.99–6.60) 11.52 (7.04–17.79)
 Adjusted HR (95% CI)b 1 (reference) 2.28 (1.25–4.17) 0.008
 Adjusted HR (95% CI)c 1 (reference) 1.61 (0.91–2.88) 0.11
Intracranial hemorrhage
 No. of events 8 8
 Sum of person-years 599.4 177.4
 Incidence rate for 100 person-years (95% CI) 1.33 (0.58–2.63) 4.51 (1.95–8.88)
 Adjusted HR (95% CI)b 1 (reference) 2.13 (0.83–5.48) 0.12
 Adjusted HR (95% CI)c 1 (reference) 1.86 (0.73–4.74) 0.19

CI, confidence interval; CKD, chronic kidney disease; DOAC, direct oral anticoagulant; HD, hemodialysis; HR, hazard ratio.

aAdjusted for age, sex.

bWithout considering death as a competing event.

cWith death considered as a competing event.

References

1. Bell EJ, Lutsey PL, Basu S, et al. Lifetime risk of venous thromboembolism in two cohort studies. Am J Med 2016;129:339.e19–e26.
crossref pmid
2. Yamashita Y, Morimoto T, Kimura T. Venous thromboembolism: recent advancement and future perspective. J Cardiol 2022;79:79–89.
crossref pmid
3. Hokusai-VTE Investigators, Büller HR, Décousus H, et al. Edoxaban versus warfarin for the treatment of symptomatic venous thromboembolism. N Engl J Med 2013;369:1406–1415.
crossref pmid
4. Agnelli G, Buller HR, Cohen A, et al. Oral apixaban for the treatment of acute venous thromboembolism. N Engl J Med 2013;369:799–808.
crossref pmid
5. EINSTEIN Investigators, Bauersachs R, Berkowitz SD, et al. Oral rivaroxaban for symptomatic venous thromboembolism. N Engl J Med 2010;363:2499–2510.
crossref pmid
6. Schulman S, Kakkar AK, Goldhaber SZ, et al. Treatment of acute venous thromboembolism with dabigatran or warfarin and pooled analysis. Circulation 2014;129:764–772.
crossref pmid
7. Hindley B, Lip GY, McCloskey AP, Penson PE. Pharmacokinetics and pharmacodynamics of direct oral anticoagulants. Expert Opin Drug Metab Toxicol 2023;19:911–923.
crossref pmid
8. Starr JA, Pinner NA, Mannis M, Stuart MK. A review of direct oral anticoagulants in patients with stage 5 or end-stage kidney disease. Ann Pharmacother 2022;56:691–703.
crossref pmid pdf
9. Wattanakit K, Cushman M. Chronic kidney disease and venous thromboembolism: epidemiology and mechanisms. Curr Opin Pulm Med 2009;15:408–412.
crossref pmid pmc
10. Smith SB, Geske JB, Kathuria P, et al. Analysis of national trends in admissions for pulmonary embolism. Chest 2016;150:35–45.
crossref pmid pmc
11. Stevens SM, Woller SC, Kreuziger LB, et al. Antithrombotic therapy for VTE disease: second update of the CHEST guideline and expert panel report. Chest 2021;160:e545–e608.
crossref pmid
12. Kearon C, Akl EA, Ornelas J, et al. Antithrombotic therapy for VTE disease: CHEST guideline and expert panel report. Chest 2016;149:315–352.
crossref pmid
13. United States Renal Data System (USRDS). 2023 USRDS Annual Data Report: epidemiology of kidney disease in the United States. National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases; 2023.
14. Bauersachs RM, Lensing AW, Prins MH, et al. Rivaroxaban versus enoxaparin/vitamin K antagonist therapy in patients with venous thromboembolism and renal impairment. Thromb J 2014;12:25.
crossref pmid pmc
15. Goldhaber SZ, Schulman S, Eriksson H, et al. Dabigatran versus warfarin for acute venous thromboembolism in elderly or impaired renal function patients: pooled analysis of RE-COVER and RE-COVER II. Thromb Haemost 2017;117:2045–2052.
crossref pmid
16. Reed D, Palkimas S, Hockman R, Abraham S, Le T, Maitland H. Safety and effectiveness of apixaban compared to warfarin in dialysis patients. Res Pract Thromb Haemost 2018;2:291–298.
crossref pmid pmc pdf
17. Decousus H, Tapson VF, Bergmann JF, et al. Factors at admission associated with bleeding risk in medical patients: findings from the IMPROVE investigators. Chest 2011;139:69–79.
crossref pmid
18. Mo H, Kwon S, Kim D, et al. Patency of arteriovenous fistulas and grafts for dialysis access: an analysis using the Korean National Health Insurance Service database from 2008 to 2019. J Vasc Access 2024;25:1544–1552.
crossref pmid pdf


ABOUT
BROWSE ARTICLES
EDITORIAL POLICY
FOR CONTRIBUTORS
Editorial Office
#301, (Miseung Bldg.) 23, Apgujenog-ro 30-gil, Gangnam-gu, Seoul 06022, Korea
Tel: +82-2-3486-8736    Fax: +82-2-3486-8737    E-mail: registry@ksn.or.kr                

Copyright © 2025 by The Korean Society of Nephrology.

Developed in M2PI

Close layer