JKDA 2024; 7(1): 19-25
Published online May 25, 2024
https://doi.org/10.56774/jkda24005
© Korean Society of Dialysis Access
Correspondence to : Soo Jeong Choi
Department of Internal Medicine, Soonchunhyang University Bucheon Hospital, 170 Jomaru-ro, Bucheon 14584, Korea
Tel: 82-32-621-5169, Fax: 82-32-621-5016, E-mail: crystal@schmc.ac.kr
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background: Patients with end stage kidney disease (ESKD) are at an increased risk of both thromboembolic events and bleeding. Direct oral anticoagulant (DOAC) agents has been developed as alternatives to warfarin. These medications have advantages such as a predictable anticoagulant effect, fewer drug interactions, and no need for routine monitoring of blood levels. However, their use in patients with ESKD requires consideration.
Material and Methods: This study was a retrospective, observational cohort study who was treated with DOAC and warfarin for anticoagulation. Based on the common data model database, DOAC exposure included apixaban, edoxaban, rivaroxaban, and dabigatran. Outcomes included occurrence of malignancy, gastrointestinal (GI) bleeding, brain hemorrhage, and major adverse cardiovascular event..
Results: Total 680 patients with ESKD on dialysis were enrolled. Apixaban (88.7%) was most frequently used and followed with edoxaban, rivaroxaban, and dabigatran. Drug pathway was different among hospitals. DOAC group was older compared with warfarin group. Outcomes in DOAC group were less compared with those in warfarin group except brain hemorrhage. After propensity score (PS) matching, the estimation of outcomes was not different with DOAC and warfarin.
Conclusion: Outcomes of DOAC in patients with ESKD on dialysis are not inferior with those of warfarin.
Keywords End stage kidney disease, Direct oral anticoagulant, Bleeding, Thromboembolism
Direct oral anticoagulants (DOAC) are used as medications for anticoagulation, particularly in conditions like atrial fibrillation (AF) and venous thromboembolism (VTE). Incidences of AF and VTE in hemodialysis (HD) patients are high, such as 26.5% and 8.4%, respectively [1,2]. However, the use of anticoagulation in patients with HD should consider safety against bleeding risk as well as the effect of thrombosis prevention. Patients in ESKD have a five times higher risk of gastrointestinal (GI) bleeding and mortality than the general population [3,4]. International consensus recommendations concerning DOAC in HD patients mostly suggest individualized decision-making and demand further controlled randomized data [5-7]. When it comes to patients undergoing HD, dose adjustment and choice of agent are important considerations. Until now, apixaban has been considered in HD patients based on small pharmacokinetics and pharmacodynamics studies [8,9]. Rivaroxaban and edoxaban are used with dose reduction in patients with moderate renal impairment. Dabigatran also requires dose reduction in patients with renal impairment. Although avoidance of DOACs is recommended in patients on dialysis, DOAC has been used as off the label in HD patients. Therefore, the efficacy and safety of DOAC in HD patients should be evaluated.
This study was a multicenter retrospective, observational cohort study including five hospital databases investigating the outcomes between target and comparative cohorts. Target cohort was HD patients who were treated with DOAC for anticoagulation. Comparative cohort was HD patients with warfarin. The five hospitals are as follows: Soonchunhyang University Bucheon Hospital (SCHBC), Soonchunhyang University Seoul Hospital (SCHSU), Soonchunhyang University Cheonan Hospital (SCHCA), Soonchunhyang University Gumi Hospital (SCHGM), and Ewha Womans University Medical Center (EWMC). Electronic Health Record data of each hospital were converted to the Common Data Model (CDM) using open-source software (Copyright © 2024 Observational Health Data Sciences and Informatics; https://ohdsi.org/analytic-tools/(accessed on 28 March 2024) and the Observational Medical Outcomes Partnership CDM version 5.3 database. Data of CDM were provided by feedernet.com (evidnet, Gyeonggi-do, Republic of Korea). The characteristics of cohort including total number and sex were evaluated. To reduce the influence of confounders from observational studies, we applied analysis methods, such as propensity score (PS) adjustment and 1:1 matching, and negative control to quantify and adjust for residual unmeasured bias. HD patient cohort was entered when DOAC exposure were over 7 days from January 2013 to May 2023. DOAC included apixaban, edoxaban, rivaroxaban, and dabigatran. Outcomes were defined as occurrence of malignancy, GI bleeding, brain hemorrhage, and major adverse cardiovascular event (acute myocardial infarction, stroke, and VTE). Central Vocabulary Service Athena (http://athena.ohdsi.org) was used to search for outcome concept (Supplementary Table 1).
In each data source, the incidence of malignancy, GI bleeding, brain hemorrhage, and major adverse cardiovascular event between the target and comparator cohorts were compared. We defined the time-at-risk to start on the day of the cohort start date, and stopped 3,650 days from the cohort start date. Incidence was compared using proportions (events/1,000 patients) and rates (events/1,000 person-years). In each hospital data analysis, the treatment pathways as the ordered sequence of medications in DOAC groups could be compared through a sunburst plot. The Cox proportional hazards model was used to estimate the relative risk (RR) of hazard ratio using the R package for Windows, with the level of significance set at p<0.05.
Propensity score matching with a one-to-one was performed in only SCHBC using the Match It package. A 1:1 PS matching with a caliper of 0.1 was used to reduce the differences in baseline characteristics, including age and sex, between the two groups.
This study was approved by the Institutional Review Board (IRB) of Soonchunhyang University Bucheon Hospital (IRB no. SCHBC 2020-08-022-005). The informed consent was waived by the IRB of Soonchunhyang University Bucheon Hospital.
A total of 283 ESKD patients on dialysis treated with DOACs were enrolled (Table 1). The number of patients treated with DOACs increased after 2019 (Fig. 1). There was no difference in sex among patients treated at SCHBC, SCHSU, and SCHCA. However, the DOAC group was older compared to the warfarin group in SCHBC and EWMC.
Table 1 . Characteristics of cohort
DOAC | Warfarin | Std. diff of mean | |
---|---|---|---|
Patients, n | 283 | 382 | |
SCHBC | 62 | 224 | |
SCHSU | 86 | 124 | |
SCHCA | 26 | 42 | |
SCHGM | 5 | 7 | |
EWMC | 66 | 23 | |
Male, n (%) | 127 (44.9) | 202 (52.9) | |
SCHBC | 34 (55) | 115 (51) | –0.0702 |
SCHSU | 42 (49) | 55 (44) | 0.0899 |
SCHCA | 12 (46) | 23 (50) | 0.077 |
SCHGM | 3 (60) | 1 (14) | 1.0738 |
EWMC | 36 (55) | 10 (43) | –0.2228 |
Age, mean±SD | |||
SCHBC | 68.4±11.8 | 65.0±12.8 | –0.2721 |
SCHSU | 64.8±12.8 | 67.7±11.4 | –0.2455 |
SCHCA | 67.1±11.5 | 68.9±12.0 | 0.148 |
SCHGM | 66.0±9.4 | 64.0±11.4 | 0.1922 |
EWMC | 72.41±11.3 | 67.0±11.2 | –0.4759 |
Hypertension | |||
SCHBC | 47 | 136 | –0.3286 |
SCHSU | 58 | 49 | 0.2053 |
SCHCA | 17 | 20 | –0.3642 |
SCHGM | 3 | 1 | 1.0738 |
EWMC | 19 | 5 | –0.1628 |
Among the 283 patients in the DOAC group, apixaban (252, 89%) was the most frequently used, followed by edoxaban, rivaroxaban, and dabigatran (Fig. 2, Supplementary Table 2). Dabigatran was detected in one patient each at SCHBC and SCHCA. The treatment pathways as the ordered sequence of DOAC varied among hospitals. Two, one and one of 252 patients with apixaban at first switched to edoxaban, rivaroxaban, and warfarin, respectively. But, three of 19 who initiated with endoxaban changed to apixaban. One of 17 who initiated with rivaroxaban switched apixaban.
Fig. 3 presents the incidences of death, any malignancy, GI bleeding, brain hemorrhage, and major adverse cardiovascular events according to centers using the CDM. Mortality was 21.3%. Mortality of DOAC group was insignificantly better compared warfarin (18.0% vs. 23.8%, p=0.074). In total cohort, the incidence of GI bleeding was 5 times higher than that of brain hemorrhage (61.5 vs. 9.6/1,000 ptyears). The incidence of stroke was 2.5 times more than that of brain hemorrhage (22.5 vs. 9.6/1,000 ptyears). Nonetheless, the rates of acute myocardial infarction, stroke, and venous thromboembolism were higher than those of warfarin, unlike the proportions (Table 2).
Table 2 . Total outcome comparison of anticoagulants
DOAC | Warfarin | Total | |||||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
n | Cases | Timeatrisk (years) | Proportion | Rate | n | Cases | Timeatrisk (years) | Proportion | Rate | n | Cases | Timeatrisk (years) | Proportion | Rate | |||
Any death | 283 | 51 | 430 | 180.2 | 118.6 | 382 | 91 | 1213 | 238.2 | 75.0 | 665 | 142 | 1643 | 213.5 | 86.4 | ||
All cancer | 237 | 11 | 334 | 46.4 | 32.9 | 344 | 21 | 1054 | 61.0 | 19.9 | 581 | 32 | 1388 | 55.1 | 23.1 | ||
Brain hemorrhage | 267 | 3 | 405 | 11.2 | 7.4 | 372 | 12 | 1151 | 32.3 | 10.4 | 639 | 15 | 1556 | 23.5 | 9.6 | ||
GI bleeding | 235 | 26 | 330 | 110.6 | 78.8 | 314 | 49 | 889 | 156.1 | 55.1 | 549 | 75 | 1219 | 136.6 | 61.5 | ||
Myocardial infarction | 261 | 7 | 391 | 26.8 | 17.9 | 358 | 12 | 1139 | 33.5 | 10.5 | 619 | 19 | 1530 | 30.7 | 12.4 | ||
Stroke | 223 | 12 | 340 | 53.8 | 35.3 | 315 | 18 | 993 | 57.1 | 18.1 | 538 | 30 | 1333 | 55.8 | 22.5 | ||
Thromboembolism | 224 | 9 | 335 | 40.2 | 26.9 | 335 | 12 | 1065 | 35.8 | 11.3 | 559 | 21 | 1400 | 37.6 | 15.0 |
After PS matching, bleeding incidence was not different between DOAC and warfarin. But. There was a trend of increased relative risk about myocardial infarction, stroke and thromboembolism (Table 3).
Table 3 . Total outcome comparison between DOAC and warfarin group in SCHBC after PS matching
DOAC | Warfarin | RR | p-value | |||||
---|---|---|---|---|---|---|---|---|
Total | Cases | Total | Cases | |||||
Any death | 43 | 5 | 43 | 4 | 1.333 | 0.603 | ||
All cancer | 44 | 0 | 44 | 0 | 0.400 | 0.302 | ||
Brain hemorrhage | 41 | 2 | 41 | 2 | 1.000 | 1.000 | ||
GI bleeding | 45 | 10 | 45 | 9 | 1.111 | 0.796 | ||
Myocardial infarction | 45 | 2 | 45 | 0 | 5.000 | 0.294 | ||
Stroke | 45 | 2 | 45 | 6 | 2.000 | 0.616 | ||
Thromboembolism | 45 | 3 | 45 | 1 | 3.000 | 0.333 |
The prevalence of AF and VTE are high among HD patients [1,2], and is associated with age, sex, and distinct comorbidities. Practice patterns of antithrombotic treatment indicate a lack of consensus for stroke prevention in HD patients with AF [1]. Thromboprophylaxis and treatment of VTE carry a high risk of major bleeding and all-cause mortality [2]. KDIGO work group consider apixaban as an alternative to vitamin K antagonist, because apixaban has a substantially better efficacy to toxicity ratio [7]. Because there was no controlled randomized data, international consensus suggests individualized decision-making about weighing bleeding risk and prevention [5,10].
DOACs have varying degrees of renal clearance [8,10-12]. Even though rivaroxaban and apixaban are minimally removed by HD [10,13], dabigatran demonstrates highest renal elimination by HD [10]. Apixaban appears to be the safest DOAC in HD patients, showing equivalent major bleeding outcomes to warfarin [14]. Dabigatran and rivaroxaban showed the increased major bleeding rates [11]. A recent systematic review recommended clinicians should continue to weigh the risk of stroke versus bleeding before prescribing DOACs in the CKD and dialysis population [15]. The 2018 Korean Heart Rhythm Society (KHRS) Practical Guidelines did not recommend the DOAC for AF in HD. KHRS recommends tailoring DOACs selection according to clinical situations in the general population. For instance, in cases of recurrent ischemic stroke despite appropriate anticoagulant therapy, dabigatran administration is suggested, while in high-risk groups for GI bleeding, apixaban or dabigatran administration is preferred [16]. However, many HD patients are prescribed DOAC in real world. As expected, apixaban (89%) is the most used and followed with edoxaban (6.7%), rivaroxaban (6.0%), and dabigatran (0.7%). The choice of agent varied over time and among hospitals (Fig. 1, 2). Since the introduction of DOACs in 2013, their prescription has increased steadily since 2019. Patients who initiated apixaban did not switch DOACs compared to those who initiated edoxaban and rivaroxaban. However, the dominance of apixaban was not compared with other agents due to small number.
Because many previous studies didn’t assess long term outcomes in HD patients, we compared the proportion and rate in outcomes (Supplementary Table 3). In 2016, Tsuruya and Fujisaki reported that infarction is recently increasing compared with brain hemorrhage in ESKD patients. Expansion of the indications for HD among older patients and use of erythropoiesis-stimulating agents to treat anemia, and low heparin dose during dialysis according to improvement in dialysis membranes were considered as possible reasons by them [17]. DOAC user in SCHBC and EWMC was older than warfarin user (Table 1). The proportion and rate of brain hemorrhage in DOAC users are less than those of warfarin users. The incidence rate of GI bleeding and major adverse cardiovascular events in DOAC users are higher than those in warfarin users. While Jung and Park [18] did not comment on renal failure, they suggested that among DOACs, apixaban has the best gastrointestinal safety profile, whereas rivaroxaban has the least gastrointestinal safety.
The rates of major adverse cardiovascular event with DOACs compared to warfarin are higher than the proportions. Because HD patients had individual differences such as age, distinct comorbidities, other medications and dialysis vintage, the rate may be more reasonable measure than proportion. However, these differences in outcomes were not observed after propensity score (PS) matching. PS matching did not overcome selection bias. Considering the short follow up period, DOACs do not appear to offer better safety and efficacy for thromboembolism prevention (Table 2). But, DOAC will show better outcome of brain and GI hemorrhage if it would be followed in long period.
This study has strengths. Firstly, we evaluated total of 283 ESKD patients who recently initiated DOAC therapy and compared the outcomes with those on warfarin. Despite the 2018 Korean Heart Rhythm Society (KHRS) Practical Guidelines not recommending DOAC use in hemodialysis (HD) patients [16], this study provides real-world evidence regarding DOAC use in ESKD patients. Anti-coagulation in ESKD patients is very difficult. There is no consensus according to cardiologist, nephrologist, gastroenterologist, neurologist and other specialists. Secondly, we evaluated DOAC outcomes in 5 hospitals by CDM.
However, there are important limitations to our study. Firstly, it was a small observational study. Matching is unlikely to reduce confounding due to measured or unmeasured factors. In particular, we did not match laboratories results and HD vintage. As previously mentioned, many dialysis factors would influence the outcomes. Secondly, we excluded patients who had experienced DOAC and warfarin therapy. Thirdly, we lacked detailed dialysis data regarding dialysis frequency, duration, dose, and other medications in the cohort. Fourthly, we did not analyze the dose of DOACs. Fifthly, DOAC use, especially apixaban, increased since 2019, and had shorter follow up period compared to warfarin use. Sixthly, CDM method was less evaluated in clinical practice. Missing or misclassified information may happen during data extraction, transformation and loading process. CDM has a noise about diagnosis accuracy and accurate case identification. Each cohort enrolled different AF or/and VTE proportion. In addition, it had limitations including generalization, data Integration, interoperability, and vendor lock-in. But, CDM has benefits such as overcoming data privacy and security [19].
In conclusion, apixaban (89%) was most frequently used, followed by edoxaban, rivaroxaban, and dabigatran. Outcomes of DOAC in patients with ESKD on dialysis are not different with those of warfarin. A prospective study of DOAC use in ESKD population is needed to weight the risk (safety) and benefit (prevention of thromboembolism).
Supplementary Tables can be found via https://doi.org/10.56774/jkda24005.
The authors declare no potential conflict of interest.
This work was supported by the Soonchunhyang University Research Fund.
JKDA 2024; 7(1): 19-25
Published online May 25, 2024 https://doi.org/10.56774/jkda24005
Copyright © Korean Society of Dialysis Access.
MinSung Lee1, YoonWon Choi2, SeonMin Kim2, Geo Neul Park2, Soo Jeong Choi2
1Division of Nephrology, Department of Internal Medicine, Ewha Womans University College of Medicine, Seoul, Korea
2Division of Nephrology, Department of Internal Medicine, Soonchunhyang University College of Medicine, Bucheon, Korea
Correspondence to:Soo Jeong Choi
Department of Internal Medicine, Soonchunhyang University Bucheon Hospital, 170 Jomaru-ro, Bucheon 14584, Korea
Tel: 82-32-621-5169, Fax: 82-32-621-5016, E-mail: crystal@schmc.ac.kr
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background: Patients with end stage kidney disease (ESKD) are at an increased risk of both thromboembolic events and bleeding. Direct oral anticoagulant (DOAC) agents has been developed as alternatives to warfarin. These medications have advantages such as a predictable anticoagulant effect, fewer drug interactions, and no need for routine monitoring of blood levels. However, their use in patients with ESKD requires consideration.
Material and Methods: This study was a retrospective, observational cohort study who was treated with DOAC and warfarin for anticoagulation. Based on the common data model database, DOAC exposure included apixaban, edoxaban, rivaroxaban, and dabigatran. Outcomes included occurrence of malignancy, gastrointestinal (GI) bleeding, brain hemorrhage, and major adverse cardiovascular event..
Results: Total 680 patients with ESKD on dialysis were enrolled. Apixaban (88.7%) was most frequently used and followed with edoxaban, rivaroxaban, and dabigatran. Drug pathway was different among hospitals. DOAC group was older compared with warfarin group. Outcomes in DOAC group were less compared with those in warfarin group except brain hemorrhage. After propensity score (PS) matching, the estimation of outcomes was not different with DOAC and warfarin.
Conclusion: Outcomes of DOAC in patients with ESKD on dialysis are not inferior with those of warfarin.
Keywords: End stage kidney disease, Direct oral anticoagulant, Bleeding, Thromboembolism
Direct oral anticoagulants (DOAC) are used as medications for anticoagulation, particularly in conditions like atrial fibrillation (AF) and venous thromboembolism (VTE). Incidences of AF and VTE in hemodialysis (HD) patients are high, such as 26.5% and 8.4%, respectively [1,2]. However, the use of anticoagulation in patients with HD should consider safety against bleeding risk as well as the effect of thrombosis prevention. Patients in ESKD have a five times higher risk of gastrointestinal (GI) bleeding and mortality than the general population [3,4]. International consensus recommendations concerning DOAC in HD patients mostly suggest individualized decision-making and demand further controlled randomized data [5-7]. When it comes to patients undergoing HD, dose adjustment and choice of agent are important considerations. Until now, apixaban has been considered in HD patients based on small pharmacokinetics and pharmacodynamics studies [8,9]. Rivaroxaban and edoxaban are used with dose reduction in patients with moderate renal impairment. Dabigatran also requires dose reduction in patients with renal impairment. Although avoidance of DOACs is recommended in patients on dialysis, DOAC has been used as off the label in HD patients. Therefore, the efficacy and safety of DOAC in HD patients should be evaluated.
This study was a multicenter retrospective, observational cohort study including five hospital databases investigating the outcomes between target and comparative cohorts. Target cohort was HD patients who were treated with DOAC for anticoagulation. Comparative cohort was HD patients with warfarin. The five hospitals are as follows: Soonchunhyang University Bucheon Hospital (SCHBC), Soonchunhyang University Seoul Hospital (SCHSU), Soonchunhyang University Cheonan Hospital (SCHCA), Soonchunhyang University Gumi Hospital (SCHGM), and Ewha Womans University Medical Center (EWMC). Electronic Health Record data of each hospital were converted to the Common Data Model (CDM) using open-source software (Copyright © 2024 Observational Health Data Sciences and Informatics; https://ohdsi.org/analytic-tools/(accessed on 28 March 2024) and the Observational Medical Outcomes Partnership CDM version 5.3 database. Data of CDM were provided by feedernet.com (evidnet, Gyeonggi-do, Republic of Korea). The characteristics of cohort including total number and sex were evaluated. To reduce the influence of confounders from observational studies, we applied analysis methods, such as propensity score (PS) adjustment and 1:1 matching, and negative control to quantify and adjust for residual unmeasured bias. HD patient cohort was entered when DOAC exposure were over 7 days from January 2013 to May 2023. DOAC included apixaban, edoxaban, rivaroxaban, and dabigatran. Outcomes were defined as occurrence of malignancy, GI bleeding, brain hemorrhage, and major adverse cardiovascular event (acute myocardial infarction, stroke, and VTE). Central Vocabulary Service Athena (http://athena.ohdsi.org) was used to search for outcome concept (Supplementary Table 1).
In each data source, the incidence of malignancy, GI bleeding, brain hemorrhage, and major adverse cardiovascular event between the target and comparator cohorts were compared. We defined the time-at-risk to start on the day of the cohort start date, and stopped 3,650 days from the cohort start date. Incidence was compared using proportions (events/1,000 patients) and rates (events/1,000 person-years). In each hospital data analysis, the treatment pathways as the ordered sequence of medications in DOAC groups could be compared through a sunburst plot. The Cox proportional hazards model was used to estimate the relative risk (RR) of hazard ratio using the R package for Windows, with the level of significance set at p<0.05.
Propensity score matching with a one-to-one was performed in only SCHBC using the Match It package. A 1:1 PS matching with a caliper of 0.1 was used to reduce the differences in baseline characteristics, including age and sex, between the two groups.
This study was approved by the Institutional Review Board (IRB) of Soonchunhyang University Bucheon Hospital (IRB no. SCHBC 2020-08-022-005). The informed consent was waived by the IRB of Soonchunhyang University Bucheon Hospital.
A total of 283 ESKD patients on dialysis treated with DOACs were enrolled (Table 1). The number of patients treated with DOACs increased after 2019 (Fig. 1). There was no difference in sex among patients treated at SCHBC, SCHSU, and SCHCA. However, the DOAC group was older compared to the warfarin group in SCHBC and EWMC.
Table 1 . Characteristics of cohort.
DOAC | Warfarin | Std. diff of mean | |
---|---|---|---|
Patients, n | 283 | 382 | |
SCHBC | 62 | 224 | |
SCHSU | 86 | 124 | |
SCHCA | 26 | 42 | |
SCHGM | 5 | 7 | |
EWMC | 66 | 23 | |
Male, n (%) | 127 (44.9) | 202 (52.9) | |
SCHBC | 34 (55) | 115 (51) | –0.0702 |
SCHSU | 42 (49) | 55 (44) | 0.0899 |
SCHCA | 12 (46) | 23 (50) | 0.077 |
SCHGM | 3 (60) | 1 (14) | 1.0738 |
EWMC | 36 (55) | 10 (43) | –0.2228 |
Age, mean±SD | |||
SCHBC | 68.4±11.8 | 65.0±12.8 | –0.2721 |
SCHSU | 64.8±12.8 | 67.7±11.4 | –0.2455 |
SCHCA | 67.1±11.5 | 68.9±12.0 | 0.148 |
SCHGM | 66.0±9.4 | 64.0±11.4 | 0.1922 |
EWMC | 72.41±11.3 | 67.0±11.2 | –0.4759 |
Hypertension | |||
SCHBC | 47 | 136 | –0.3286 |
SCHSU | 58 | 49 | 0.2053 |
SCHCA | 17 | 20 | –0.3642 |
SCHGM | 3 | 1 | 1.0738 |
EWMC | 19 | 5 | –0.1628 |
Among the 283 patients in the DOAC group, apixaban (252, 89%) was the most frequently used, followed by edoxaban, rivaroxaban, and dabigatran (Fig. 2, Supplementary Table 2). Dabigatran was detected in one patient each at SCHBC and SCHCA. The treatment pathways as the ordered sequence of DOAC varied among hospitals. Two, one and one of 252 patients with apixaban at first switched to edoxaban, rivaroxaban, and warfarin, respectively. But, three of 19 who initiated with endoxaban changed to apixaban. One of 17 who initiated with rivaroxaban switched apixaban.
Fig. 3 presents the incidences of death, any malignancy, GI bleeding, brain hemorrhage, and major adverse cardiovascular events according to centers using the CDM. Mortality was 21.3%. Mortality of DOAC group was insignificantly better compared warfarin (18.0% vs. 23.8%, p=0.074). In total cohort, the incidence of GI bleeding was 5 times higher than that of brain hemorrhage (61.5 vs. 9.6/1,000 ptyears). The incidence of stroke was 2.5 times more than that of brain hemorrhage (22.5 vs. 9.6/1,000 ptyears). Nonetheless, the rates of acute myocardial infarction, stroke, and venous thromboembolism were higher than those of warfarin, unlike the proportions (Table 2).
Table 2 . Total outcome comparison of anticoagulants.
DOAC | Warfarin | Total | |||||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
n | Cases | Timeatrisk (years) | Proportion | Rate | n | Cases | Timeatrisk (years) | Proportion | Rate | n | Cases | Timeatrisk (years) | Proportion | Rate | |||
Any death | 283 | 51 | 430 | 180.2 | 118.6 | 382 | 91 | 1213 | 238.2 | 75.0 | 665 | 142 | 1643 | 213.5 | 86.4 | ||
All cancer | 237 | 11 | 334 | 46.4 | 32.9 | 344 | 21 | 1054 | 61.0 | 19.9 | 581 | 32 | 1388 | 55.1 | 23.1 | ||
Brain hemorrhage | 267 | 3 | 405 | 11.2 | 7.4 | 372 | 12 | 1151 | 32.3 | 10.4 | 639 | 15 | 1556 | 23.5 | 9.6 | ||
GI bleeding | 235 | 26 | 330 | 110.6 | 78.8 | 314 | 49 | 889 | 156.1 | 55.1 | 549 | 75 | 1219 | 136.6 | 61.5 | ||
Myocardial infarction | 261 | 7 | 391 | 26.8 | 17.9 | 358 | 12 | 1139 | 33.5 | 10.5 | 619 | 19 | 1530 | 30.7 | 12.4 | ||
Stroke | 223 | 12 | 340 | 53.8 | 35.3 | 315 | 18 | 993 | 57.1 | 18.1 | 538 | 30 | 1333 | 55.8 | 22.5 | ||
Thromboembolism | 224 | 9 | 335 | 40.2 | 26.9 | 335 | 12 | 1065 | 35.8 | 11.3 | 559 | 21 | 1400 | 37.6 | 15.0 |
After PS matching, bleeding incidence was not different between DOAC and warfarin. But. There was a trend of increased relative risk about myocardial infarction, stroke and thromboembolism (Table 3).
Table 3 . Total outcome comparison between DOAC and warfarin group in SCHBC after PS matching.
DOAC | Warfarin | RR | p-value | |||||
---|---|---|---|---|---|---|---|---|
Total | Cases | Total | Cases | |||||
Any death | 43 | 5 | 43 | 4 | 1.333 | 0.603 | ||
All cancer | 44 | 0 | 44 | 0 | 0.400 | 0.302 | ||
Brain hemorrhage | 41 | 2 | 41 | 2 | 1.000 | 1.000 | ||
GI bleeding | 45 | 10 | 45 | 9 | 1.111 | 0.796 | ||
Myocardial infarction | 45 | 2 | 45 | 0 | 5.000 | 0.294 | ||
Stroke | 45 | 2 | 45 | 6 | 2.000 | 0.616 | ||
Thromboembolism | 45 | 3 | 45 | 1 | 3.000 | 0.333 |
The prevalence of AF and VTE are high among HD patients [1,2], and is associated with age, sex, and distinct comorbidities. Practice patterns of antithrombotic treatment indicate a lack of consensus for stroke prevention in HD patients with AF [1]. Thromboprophylaxis and treatment of VTE carry a high risk of major bleeding and all-cause mortality [2]. KDIGO work group consider apixaban as an alternative to vitamin K antagonist, because apixaban has a substantially better efficacy to toxicity ratio [7]. Because there was no controlled randomized data, international consensus suggests individualized decision-making about weighing bleeding risk and prevention [5,10].
DOACs have varying degrees of renal clearance [8,10-12]. Even though rivaroxaban and apixaban are minimally removed by HD [10,13], dabigatran demonstrates highest renal elimination by HD [10]. Apixaban appears to be the safest DOAC in HD patients, showing equivalent major bleeding outcomes to warfarin [14]. Dabigatran and rivaroxaban showed the increased major bleeding rates [11]. A recent systematic review recommended clinicians should continue to weigh the risk of stroke versus bleeding before prescribing DOACs in the CKD and dialysis population [15]. The 2018 Korean Heart Rhythm Society (KHRS) Practical Guidelines did not recommend the DOAC for AF in HD. KHRS recommends tailoring DOACs selection according to clinical situations in the general population. For instance, in cases of recurrent ischemic stroke despite appropriate anticoagulant therapy, dabigatran administration is suggested, while in high-risk groups for GI bleeding, apixaban or dabigatran administration is preferred [16]. However, many HD patients are prescribed DOAC in real world. As expected, apixaban (89%) is the most used and followed with edoxaban (6.7%), rivaroxaban (6.0%), and dabigatran (0.7%). The choice of agent varied over time and among hospitals (Fig. 1, 2). Since the introduction of DOACs in 2013, their prescription has increased steadily since 2019. Patients who initiated apixaban did not switch DOACs compared to those who initiated edoxaban and rivaroxaban. However, the dominance of apixaban was not compared with other agents due to small number.
Because many previous studies didn’t assess long term outcomes in HD patients, we compared the proportion and rate in outcomes (Supplementary Table 3). In 2016, Tsuruya and Fujisaki reported that infarction is recently increasing compared with brain hemorrhage in ESKD patients. Expansion of the indications for HD among older patients and use of erythropoiesis-stimulating agents to treat anemia, and low heparin dose during dialysis according to improvement in dialysis membranes were considered as possible reasons by them [17]. DOAC user in SCHBC and EWMC was older than warfarin user (Table 1). The proportion and rate of brain hemorrhage in DOAC users are less than those of warfarin users. The incidence rate of GI bleeding and major adverse cardiovascular events in DOAC users are higher than those in warfarin users. While Jung and Park [18] did not comment on renal failure, they suggested that among DOACs, apixaban has the best gastrointestinal safety profile, whereas rivaroxaban has the least gastrointestinal safety.
The rates of major adverse cardiovascular event with DOACs compared to warfarin are higher than the proportions. Because HD patients had individual differences such as age, distinct comorbidities, other medications and dialysis vintage, the rate may be more reasonable measure than proportion. However, these differences in outcomes were not observed after propensity score (PS) matching. PS matching did not overcome selection bias. Considering the short follow up period, DOACs do not appear to offer better safety and efficacy for thromboembolism prevention (Table 2). But, DOAC will show better outcome of brain and GI hemorrhage if it would be followed in long period.
This study has strengths. Firstly, we evaluated total of 283 ESKD patients who recently initiated DOAC therapy and compared the outcomes with those on warfarin. Despite the 2018 Korean Heart Rhythm Society (KHRS) Practical Guidelines not recommending DOAC use in hemodialysis (HD) patients [16], this study provides real-world evidence regarding DOAC use in ESKD patients. Anti-coagulation in ESKD patients is very difficult. There is no consensus according to cardiologist, nephrologist, gastroenterologist, neurologist and other specialists. Secondly, we evaluated DOAC outcomes in 5 hospitals by CDM.
However, there are important limitations to our study. Firstly, it was a small observational study. Matching is unlikely to reduce confounding due to measured or unmeasured factors. In particular, we did not match laboratories results and HD vintage. As previously mentioned, many dialysis factors would influence the outcomes. Secondly, we excluded patients who had experienced DOAC and warfarin therapy. Thirdly, we lacked detailed dialysis data regarding dialysis frequency, duration, dose, and other medications in the cohort. Fourthly, we did not analyze the dose of DOACs. Fifthly, DOAC use, especially apixaban, increased since 2019, and had shorter follow up period compared to warfarin use. Sixthly, CDM method was less evaluated in clinical practice. Missing or misclassified information may happen during data extraction, transformation and loading process. CDM has a noise about diagnosis accuracy and accurate case identification. Each cohort enrolled different AF or/and VTE proportion. In addition, it had limitations including generalization, data Integration, interoperability, and vendor lock-in. But, CDM has benefits such as overcoming data privacy and security [19].
In conclusion, apixaban (89%) was most frequently used, followed by edoxaban, rivaroxaban, and dabigatran. Outcomes of DOAC in patients with ESKD on dialysis are not different with those of warfarin. A prospective study of DOAC use in ESKD population is needed to weight the risk (safety) and benefit (prevention of thromboembolism).
Supplementary Tables can be found via https://doi.org/10.56774/jkda24005.
The authors declare no potential conflict of interest.
This work was supported by the Soonchunhyang University Research Fund.
Table 1 . Characteristics of cohort.
DOAC | Warfarin | Std. diff of mean | |
---|---|---|---|
Patients, n | 283 | 382 | |
SCHBC | 62 | 224 | |
SCHSU | 86 | 124 | |
SCHCA | 26 | 42 | |
SCHGM | 5 | 7 | |
EWMC | 66 | 23 | |
Male, n (%) | 127 (44.9) | 202 (52.9) | |
SCHBC | 34 (55) | 115 (51) | –0.0702 |
SCHSU | 42 (49) | 55 (44) | 0.0899 |
SCHCA | 12 (46) | 23 (50) | 0.077 |
SCHGM | 3 (60) | 1 (14) | 1.0738 |
EWMC | 36 (55) | 10 (43) | –0.2228 |
Age, mean±SD | |||
SCHBC | 68.4±11.8 | 65.0±12.8 | –0.2721 |
SCHSU | 64.8±12.8 | 67.7±11.4 | –0.2455 |
SCHCA | 67.1±11.5 | 68.9±12.0 | 0.148 |
SCHGM | 66.0±9.4 | 64.0±11.4 | 0.1922 |
EWMC | 72.41±11.3 | 67.0±11.2 | –0.4759 |
Hypertension | |||
SCHBC | 47 | 136 | –0.3286 |
SCHSU | 58 | 49 | 0.2053 |
SCHCA | 17 | 20 | –0.3642 |
SCHGM | 3 | 1 | 1.0738 |
EWMC | 19 | 5 | –0.1628 |
Table 2 . Total outcome comparison of anticoagulants.
DOAC | Warfarin | Total | |||||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
n | Cases | Timeatrisk (years) | Proportion | Rate | n | Cases | Timeatrisk (years) | Proportion | Rate | n | Cases | Timeatrisk (years) | Proportion | Rate | |||
Any death | 283 | 51 | 430 | 180.2 | 118.6 | 382 | 91 | 1213 | 238.2 | 75.0 | 665 | 142 | 1643 | 213.5 | 86.4 | ||
All cancer | 237 | 11 | 334 | 46.4 | 32.9 | 344 | 21 | 1054 | 61.0 | 19.9 | 581 | 32 | 1388 | 55.1 | 23.1 | ||
Brain hemorrhage | 267 | 3 | 405 | 11.2 | 7.4 | 372 | 12 | 1151 | 32.3 | 10.4 | 639 | 15 | 1556 | 23.5 | 9.6 | ||
GI bleeding | 235 | 26 | 330 | 110.6 | 78.8 | 314 | 49 | 889 | 156.1 | 55.1 | 549 | 75 | 1219 | 136.6 | 61.5 | ||
Myocardial infarction | 261 | 7 | 391 | 26.8 | 17.9 | 358 | 12 | 1139 | 33.5 | 10.5 | 619 | 19 | 1530 | 30.7 | 12.4 | ||
Stroke | 223 | 12 | 340 | 53.8 | 35.3 | 315 | 18 | 993 | 57.1 | 18.1 | 538 | 30 | 1333 | 55.8 | 22.5 | ||
Thromboembolism | 224 | 9 | 335 | 40.2 | 26.9 | 335 | 12 | 1065 | 35.8 | 11.3 | 559 | 21 | 1400 | 37.6 | 15.0 |
Table 3 . Total outcome comparison between DOAC and warfarin group in SCHBC after PS matching.
DOAC | Warfarin | RR | p-value | |||||
---|---|---|---|---|---|---|---|---|
Total | Cases | Total | Cases | |||||
Any death | 43 | 5 | 43 | 4 | 1.333 | 0.603 | ||
All cancer | 44 | 0 | 44 | 0 | 0.400 | 0.302 | ||
Brain hemorrhage | 41 | 2 | 41 | 2 | 1.000 | 1.000 | ||
GI bleeding | 45 | 10 | 45 | 9 | 1.111 | 0.796 | ||
Myocardial infarction | 45 | 2 | 45 | 0 | 5.000 | 0.294 | ||
Stroke | 45 | 2 | 45 | 6 | 2.000 | 0.616 | ||
Thromboembolism | 45 | 3 | 45 | 1 | 3.000 | 0.333 |
Won-yeol Choi, So-dam Choi, Hoyen Chu, Ming-yeong Choi, Joo-hyeok Choi, Gil-won Choi, Dong-hwa Choi, Dong-hoon Han, Jung-woo Choi, Yeon-hee Choi, Hyangkyoung Kim
JKDA 2018; 1(1): 28-32