Direct Oral Anticoagulants in Chronic Kidney Disease and Dialysis: Dosing, Bleeding Risk, and Comparison with Warfarin

Key Takeaways

• The fundamental clinical dilemma is that chronic kidney disease concurrently amplifies both thrombotic and bleeding risks. Landmark DOAC (direct oral anticoagulant) trials excluded patients with a creatinine clearance below 25 to 30 mL/min. In mild-to-moderate CKD, DOACs are superior to warfarin. In dialysis patients, however, a net clinical benefit for any oral anticoagulant remains unproven; moreover, the incidence of clinically significant bleeding exceeds that of stroke by approximately tenfold.
• Renal elimination rates differ markedly across molecules. Dabigatran, a prodrug-derived direct thrombin inhibitor, is the most dependent on renal function and is dialyzable. Edoxaban occupies an intermediate position. Rivaroxaban and apixaban, which are the least dependent on renal elimination, are highly protein-bound and cannot be meaningfully removed by hemodialysis. Warfarin additionally inhibits vitamin K-dependent matrix Gla protein, which may accelerate vascular calcification.
• Dose reduction cutoffs rely on Cockcroft-Gault creatinine clearance calculated using actual body weight. Rivaroxaban should be reduced to 15 mg daily and edoxaban to 30 mg daily when clearance is between 15 and 50 mL/min. The apixaban dose is reduced to 2.5 mg twice daily when at least two of the following criteria are met: age 80 years or older, body weight 60 kg or less, or serum creatinine 1.5 mg/dL or greater. Dabigatran should be avoided below 30 mL/min according to European regulatory guidelines.
• Randomized evidence in the dialysis population is derived from three small trials. Reduced-dose rivaroxaban decreased cardiovascular events and major bleeding compared with a vitamin K antagonist. Two trials conducted with apixaban were underpowered and inconclusive. Large-scale observational data favor apixaban over warfarin with respect to major bleeding in dialysis patients, while in the broader CKD population, gastrointestinal bleeding risk is lower compared with rivaroxaban.
• At the bedside, creatinine clearance should be reassessed at least every three to six months when it falls below 60 mL/min. Off-label, inappropriate underdosing must be avoided. P-glycoprotein and CYP3A4 interactions should be screened. Left atrial appendage occlusion should be considered in dialysis patients with a prior history of major bleeding.

Introduction and Clinical Background

As renal function declines, the incidence of atrial fibrillation and venous thromboembolism rises. Consequently, patients with chronic kidney disease (CKD) account for a disproportionately large share of anticoagulation decisions across internal medicine, nephrology, and cardiology clinics.

The clinical challenge extends beyond pharmacokinetics. Advanced-stage CKD and end-stage renal disease induce a paradoxical hemostatic milieu: platelet dysfunction, alterations in coagulation factors, and vascular endothelial injury coexist. In concert, these factors heighten both the risk of deep vein thrombosis and pulmonary embolism and the propensity to bleed^[1]. Hemodialysis further compounds this picture with repetitive vascular access cannulations, intradialytic heparin exposure, and fluctuating volume status, each shifting the delicate equilibrium between thrombosis and hemorrhage from session to session.

In patients with non-dialysis-dependent CKD and atrial fibrillation, the broad consensus favors anticoagulating those with elevated thromboembolic risk scores who have no absolute contraindications to bleeding. In dialysis patients, however, the net benefit of anticoagulation has long been contentious owing to the heightened risk of bleeding^[2]. This divergence in approach mirrors the historical evolution of the available evidence.

The pivotal phase 3 trial programs for dabigatran, rivaroxaban, apixaban, and edoxaban enrolled patients with a creatinine clearance roughly above 25 to 30 mL/min. As a result, the CKD data generated by these trials reflect subgroups with moderate renal impairment. Patients with a clearance below 15 mL/min and those receiving maintenance dialysis were systematically excluded. In the United States, the dialysis indication for apixaban was extrapolated from single-dose pharmacokinetic studies rather than clinical outcome data. Clinicians subsequently adopted the agent in end-stage renal disease in the absence of randomized evidence.

Missed clinical opportunities cut both ways. Many patients with stage 3 CKD remain on warfarin despite superior data favoring DOACs, or receive reduced DOAC doses without a label indication. Underutilization and inappropriate dosing remain pervasive, particularly among populations historically underrepresented in clinical trials^[3]. Conversely, some dialysis patients with low thromboembolic risk and recurrent gastrointestinal bleeding are reflexively anticoagulated without a clear appraisal of the questionable net benefit in this clinical setting.

Molecular and Pathophysiological Mechanisms

Renal Elimination by Agent

The four direct oral anticoagulants (DOACs) differ fundamentally in their degree of dependence on renal elimination, and this hierarchical ranking directly governs all subsequent dosing decisions.

• Dabigatran etexilate is a prodrug hydrolyzed by esterases to dabigatran, a direct and competitive thrombin inhibitor. It is excreted predominantly unchanged in the urine; therefore, its half-life prolongs markedly as glomerular filtration declines. Because of its low plasma protein binding, it is effectively removed by hemodialysis.
• Edoxaban, a direct factor Xa inhibitor, has an intermediate renal elimination fraction. It is also a P-glycoprotein substrate; consequently, concomitant potent P-glycoprotein inhibition necessitates dose reduction.
• Rivaroxaban combines renal excretion of unchanged drug with substantial CYP3A4- and CYP2J2-mediated hepatic metabolism.
• Apixaban is the agent with the lowest renal dependence within this class, owing to marked fecal and biliary elimination alongside CYP3A4-mediated metabolism.

Both rivaroxaban and apixaban are highly bound to plasma proteins. Consequently, their clearance cannot be meaningfully achieved with conventional hemodialysis; this constitutes the fundamental clinical distinction between these agents and dabigatran in overdose and bleeding management.

Uremic Hemostasis

Uremia impairs platelet adhesion and aggregation through several mechanisms:

• defective glycoprotein IIb/IIIa activation;
• impaired von Willebrand factor interaction;
• accumulation of uremic toxins and increased nitric oxide production;
• anemia, which diminishes the radial displacement of platelets toward the vessel wall.

Concurrently, endothelial injury, chronic inflammation, elevations in fibrinogen and factor VIII, and repeated vascular access manipulations heighten the thrombotic tendency. Superimposed anticoagulants further exacerbate an already pronounced bleeding diathesis against this pathological backdrop. Due to the high prevalence of angiodysplasias in end-stage kidney disease, the gastrointestinal tract is the primary site of bleeding.

Warfarin, Vitamin K, and Vascular Calcification

By inhibiting vitamin K epoxide reductase, warfarin not only impairs the gamma-carboxylation of clotting factors II, VII, IX, and X, but also blocks the carboxylation of extrahepatic vitamin K-dependent proteins, most notably matrix Gla protein, a local inhibitor of arterial calcification.

In chronic kidney disease (CKD), where the calcification process is already accelerated due to disturbed mineral metabolism, it is biologically plausible that this off-target effect predisposes to medial arterial calcification and calciphylaxis. DOACs act directly on factor Xa or thrombin without interfering with the vitamin K-dependent carboxylation process. This provides a mechanistic rationale for favoring these agents beyond their bleeding profile in advanced kidney disease.

Warfarin control is inherently poorer in the setting of CKD. Fluctuating dietary vitamin K intake, polypharmacy, and altered hepatic CYP2C9 activity in uremia substantially reduce time in therapeutic range (TTR).

Clinical Evidence and Guideline Comparison

Moderate CKD: Subgroups of Landmark Trials

For patients with a creatinine clearance between 30 and 50 mL/min, the evidence base derives from post-hoc subgroup analyses of phase 3 atrial fibrillation trials. These analyses consistently support the use of DOACs over warfarin, with drug-specific signals:

• dabigatran 150 mg and apixaban were associated with lower rates of stroke and systemic embolism;
• apixaban and edoxaban were associated with lower rates of bleeding and mortality^[4].

In patients with mild or moderate renal impairment, each DOAC demonstrated a reduction in both stroke or systemic embolism and major bleeding compared with warfarin. DOACs were also observed to maintain a superior risk-benefit profile compared with warfarin in patients whose renal function deteriorated over time^[5].

For stage 3 CKD, the core message for clinical practice is that DOACs should be preferred, provided dose adjustments are made when required by labeling. Moderate CKD is no longer a valid rationale for continuing warfarin therapy.

Non-Dialysis-Dependent Advanced CKD

In the stage where creatinine clearance falls below 30 mL/min (with the exception of limited randomized data for apixaban in the 25 to 30 mL/min range), the evidence base relies predominantly on observational studies. Retrospective cohorts in patients with advanced CKD and atrial fibrillation show that DOACs exhibit efficacy comparable to warfarin with numerically lower bleeding rates. The level of evidence here is weaker than in moderate renal impairment, and most available data encompass apixaban and rivaroxaban.

Randomized Trials in Hemodialysis Patients

Rivaroxaban (Valkyrie). In this trial, 132 hemodialysis patients with atrial fibrillation were randomized to one of three arms:

• a vitamin K antagonist with an INR target of 2 to 3;
• rivaroxaban 10 mg daily;
• rivaroxaban plus vitamin K2.

Median follow-up was 1.88 years. The composite of fatal and non-fatal cardiovascular events was 63.8 per 100 patient-years with the vitamin K antagonist versus 26.2 per 100 patient-years with rivaroxaban, representing an absolute difference of 37.6 events per 100 patient-years. The hazard ratio (HR) adjusted for competing risks was 0.41 (95% CI 0.25 to 0.68; p = 0.0006). Life-threatening and major bleeding were also significantly reduced with rivaroxaban (hazard ratio (HR) 0.39; 95% CI 0.17 to 0.90; p = 0.03). Symptomatic limb ischemia was markedly less frequent, whereas no differences were detected in all-cause death, cardiac death, or stroke rates^[6].

This finding regarding limb ischemia aligns with the vascular calcification hypothesis. However, the trial had a modest sample size, was open-label, had a permanent treatment discontinuation rate reaching 25%, and the tested 10 mg daily regimen is not an approved dosage for the hemodialysis population in Western guidelines and labels.

Apixaban (RENAL-AF). In this trial, 154 hemodialysis patients were randomized to apixaban or warfarin arms, but the trial was terminated early due to enrollment challenges. Time in therapeutic range (TTR) in the warfarin arm was only 44%. One-year rates of major or clinically relevant non-major bleeding were 32% with apixaban and 26% with warfarin (hazard ratio (HR) 1.20; 95% CI 0.63 to 2.30). Stroke or systemic embolism occurred in 3.0% and 3.3%, respectively.

Death was the most frequent major event, affecting 26% of the apixaban arm and 18% of the warfarin arm. A pharmacokinetic substudy demonstrated that steady-state median 12-hour area under the curve was 2475 ng/mL×h for 5 mg twice daily and 1269 ng/mL×h for 2.5 mg twice daily. Drug exposure overlapped substantially between patients who did and did not develop bleeding. The investigators concluded that the trial was underpowered for bleeding comparisons and that clinically relevant bleeding occurred approximately ten times more frequently than stroke^[7].

Apixaban versus phenprocoumon (AXADIA-AFNET 8). In this European trial, apixaban was compared with phenprocoumon in 97 patients with a mean age of 75 years and a mean CHA₂DS₂-VASc score of 4.5. The composite primary safety endpoint occurred in 45.8% of the apixaban group and 51.0% of the vitamin K antagonist group (hazard ratio (HR) 0.93; 95% CI 0.53 to 1.65). Non-inferiority was not proven (p = 0.157).

The composite efficacy endpoint was 20.8% versus 30.6%, respectively (p = 0.51). All-cause mortality was 18.8% versus 24.5%, major bleeding was 10.4% versus 12.2%, and myocardial infarction was 4.2% versus 6.1%; none of these differences reached statistical significance. Median time in therapeutic range in the vitamin K antagonist group was 50.7%^[8].

The consistent takeaway from these three trials is that dialysis patients remain at exceedingly high risk for cardiovascular events and mortality even when anticoagulated. The clinical picture is dominated by bleeding and fatal events rather than stroke.

Observational Data and Network Meta-Analyses in Dialysis Patients

The largest single observational dataset available is an analysis of 25,523 dialysis patients with atrial fibrillation within the United States Renal Data System. In matched cohorts, apixaban did not differ from warfarin with respect to stroke or systemic embolism (hazard ratio (HR) 0.88; 95% CI 0.69 to 1.12; p = 0.29). In contrast, it was associated with a significantly lower risk of major bleeding (hazard ratio (HR) 0.72; 95% CI 0.59 to 0.87; p < 0.001).

In sensitivity analyses, the standard 5 mg twice-daily dose was associated with lower rates of stroke or systemic embolism and lower mortality compared with the reduced 2.5 mg twice-daily dose. The corresponding hazard ratios (HR) were 0.61 (95% CI 0.37 to 0.98) and 0.64 (95% CI 0.45 to 0.92), respectively^[9]. This dose-related signal is hypothesis-generating, yet vulnerable to confounding by frailty, as reduced doses are preferentially prescribed to patients in poorer clinical condition.

A 2020 network meta-analysis of 16 observational studies (71,877 patients) found that neither apixaban nor warfarin significantly reduced stroke or systemic thromboembolism compared with no anticoagulation. Hazard ratios (HR) were 0.59 (95% CI 0.30 to 1.17) for apixaban 5 mg and 0.91 (95% CI 0.72 to 1.16) for warfarin. Warfarin carried a higher risk of major bleeding compared with both apixaban 5 mg (hazard ratio (HR) 1.41; 95% CI 1.07 to 1.88) and no anticoagulation (hazard ratio (HR) 1.31; 95% CI 1.15 to 1.50). Dabigatran and rivaroxaban were also associated with more major bleeding compared with apixaban^[10].

A contemporary 2024 analysis pooling 42 studies (three of which were randomized) and 185,864 participants demonstrated that vitamin K antagonists increased major bleeding compared with no anticoagulation (hazard ratio (HR) 1.47; 95% CI 1.34 to 1.61), whereas DOACs reduced this risk compared with vitamin K antagonists (hazard ratio (HR) 0.74; 95% CI 0.64 to 0.84). No statistically significant difference was observed across vitamin K antagonists, DOACs, and non-anticoagulated arms in thromboembolic event rates, and all-cause mortality was similar. Although the authors suggested that withholding anticoagulation could represent a clinically reasonable alternative, the predominantly observational nature of the analyzed data precludes definitive inferences of clinical equivalence or non-inferiority^[11].

The favorable signals observed for dabigatran in this analysis should be interpreted with caution. The drug is contraindicated or not recommended in dialysis patients according to current package inserts; thus, users likely represent a highly selected patient cohort.

Venous Thromboembolism in Dialysis Patients

The highest quality data available for VTE are likewise observational. Among 12,206 Medicare dialysis patients with incident VTE, apixaban was associated with lower rates of recurrent VTE over 6 months (hazard ratio (HR) 0.58; 95% CI 0.43 to 0.77) and less major bleeding (hazard ratio (HR) 0.78; 95% CI 0.62 to 0.98) compared with warfarin. No difference in mortality was observed (hazard ratio (HR) 1.04; 95% CI 0.94 to 1.16)^[12].

Unlike atrial fibrillation in dialysis patients, the indication for anticoagulation in acute VTE is undisputed. Consequently, the clinical decision reduces directly to agent selection, and available evidence favors apixaban.

Left Atrial Appendage Occlusion

In a national cohort of 14,849 patients with end-stage renal disease and atrial fibrillation, propensity score matching identified 1,947 patients per group who were followed for a median of 0.9 years. The composite endpoint of stroke or systemic embolism, major bleeding, or death was higher in the warfarin group compared with left atrial appendage occlusion (adjusted hazard ratio (HR) 1.26; 95% CI 1.17 to 1.39). Similarly, this rate was also higher in the apixaban group compared with appendage occlusion (adjusted hazard ratio (HR) 1.27; 95% CI 1.16 to 1.39). Major bleeding, gastrointestinal bleeding, and stroke or systemic embolism rates were higher in both anticoagulant arms, whereas risks of mortality and intracranial hemorrhage were similar^[13]. In such comparisons, selection bias favoring patients who are better candidates and in superior overall condition for surgical or percutaneous intervention represents an inherent limitation.

Guidelines and Consensus Statements

Discrepancies exist among major guidelines. North American guidelines (AHA/ACC/HRS), reflecting US label approvals, consider the use of warfarin or apixaban a reasonable approach in end-stage CKD or dialysis patients. European guidelines (ESC and the EHRA practical guide) take a more restrictive stance: outside of individualized decisions, they do not recommend DOAC use in patients with a creatinine clearance below 15 mL/min or on dialysis, nor dabigatran below 30 mL/min. KDIGO explicitly highlights the lack of evidence and clinical equipoise in dialysis patients.

Recent reviews converge on a pragmatic synthesis: appropriately dose-adjusted DOACs should be preferred over warfarin in moderate to advanced CKD. In dialysis-dependent atrial fibrillation patients, however, the absolute necessity of any oral anticoagulation remains fundamentally questioned.

Practical Implications and Safety Profile

Assessment of Renal Function

Dose thresholds in DOAC package inserts are based on Cockcroft-Gault creatinine clearance calculated using actual body weight, the method employed in the pivotal trials. Estimated GFR calculated with the CKD-EPI equation frequently overestimates clearance in elderly and low-weight patients, potentially masking the need for dose reduction. Therefore, even when eGFR is reported for CKD staging, calculating the Cockcroft-Gault equation directly remains standard clinical practice for anticoagulant dosing.

Renal function is a dynamic and variable parameter. Clearance drops abruptly during acute kidney injury, increasing drug exposure. During recovery, clearance rises again, rendering the dose reduced in the acute phase subtherapeutic. Rather than maintaining acute-phase dose reductions indefinitely, dosing decisions must be systematically reassessed once creatinine levels stabilize.

Phenotypic Tiers and Agent Selection

Tier 1: Creatinine clearance greater than 50 mL/min. Standard dosing is applied, subject to non-renal dose reduction criteria (age and weight for apixaban; weight and P-gp interactions for edoxaban):

• apixaban 5 mg twice daily;
• rivaroxaban 20 mg once daily with the evening meal;
• edoxaban 60 mg once daily;
• dabigatran 150 mg twice daily, or 110 mg twice daily in patients aged 80 years and older or those at high risk of bleeding according to European labeling.

Edoxaban efficacy is diminished at higher clearance levels; thus, US prescribing information does not recommend its use at values greater than 95 mL/min.

Tier 2: Creatinine clearance 30 to 50 mL/min. All DOACs may be used with dose adjustments according to package insert recommendations:

• rivaroxaban 15 mg once daily;
• edoxaban 30 mg once daily;
• apixaban 5 mg twice daily, unless two of three criteria are met (age 80 years or older, body weight 60 kg or less, serum creatinine 1.5 mg/dL or greater);
• dabigatran 110 mg twice daily in selected patients according to European guidelines.

Given subgroup bleeding and mortality data, apixaban or edoxaban are reasonable first-line choices in this cohort.

Tier 3: Creatinine clearance 15 to 29 mL/min. The preferred agent is apixaban:

• apixaban 2.5 mg twice daily according to European labeling; the two-of-three rule remains in effect under the US label;
• alternatively, rivaroxaban 15 mg once daily;
• alternatively, edoxaban 30 mg once daily;
• while European guidelines recommend avoiding dabigatran, the US label permits 75 mg twice daily on pharmacokinetic grounds.

Supporting the edoxaban option, an exploratory series of 46 patients with an eGFR of 15 to 29 mL/min treated with edoxaban 30 mg once daily demonstrated no major bleeding, stroke, systemic embolism, or cardiovascular death over a mean follow-up of 9.13 months, with only five minor bleeds recorded^[14].

Tier 4: Creatinine clearance less than 15 mL/min or patients on dialysis. First, a shared decision-making process balancing the CHA₂DS₂-VASc score against prior bleeding, gastrointestinal angiodysplasia, falls, vascular access complications, and life expectancy must determine whether anticoagulation should be initiated at all. If the decision is made to anticoagulate, apixaban is the agent supported by the strongest observational evidence:

• the US package insert recommends 5 mg twice daily, with a dose reduction to 2.5 mg twice daily in patients aged 80 years or older or weighing 60 kg or less;
• many nephrologists empirically prefer 2.5 mg twice daily, although this practice has yet to be definitively validated;
• warfarin with meticulous INR monitoring is an alternative when DOACs are unavailable;
• dabigatran must not be used under any circumstances.

The Apixaban Dose Reduction Dilemma

Dose-reduction criteria for apixaban depart from conventional paradigms and are frequently misapplied in clinical practice. Pharmacokinetic studies demonstrate increased exposure with advanced age, extreme body weight, and end-stage renal disease; however, post-hemodialysis dosing can offset this increased exposure. Clinical data indicate no loss of safety or efficacy in patients older than 75 years, weighing less than 60 kg, or with an eGFR below 50 mL/min, including dialysis patients^[15].

Inappropriately reducing the dose to 2.5 mg twice daily in a patient with a creatinine clearance above 30 mL/min who meets only one standard dose-reduction criterion (age, body weight, serum creatinine) is a common error; this practice may increase stroke risk without conferring a proven bleeding advantage.

Venous Thromboembolism Dosing

Standard regimens for acute VTE treatment include:

• apixaban 10 mg twice daily for seven days, followed by 5 mg twice daily (the US package insert does not specify routine dose reduction based on renal function, whereas European product information recommends caution in patients with clearance of 15 to 29 mL/min and advises against use below 15 mL/min or on dialysis);
• rivaroxaban 15 mg twice daily for 21 days, followed by 20 mg once daily; European labeling states that a maintenance dose of 15 mg once daily may be considered in patients with clearance between 15 and 49 mL/min whose bleeding risk outweighs the risk of recurrence;
• edoxaban and dabigatran, both requiring a 5- to 10-day parenteral lead-in.

Although apixaban is the most commonly utilized oral agent in clinical practice for dialysis-dependent VTE patients, use in this population is off-label and grounded in observational data. Because low-molecular-weight heparins accumulate in stage 4 to 5 CKD, unfractionated heparin remains the preferred parenteral agent in clinical scenarios requiring rapid titration or rapid reversibility.

Bleeding Risk and Agent-Specific Safety Profile

Across all indications, DOACs reduce intracranial hemorrhage compared with vitamin K antagonists. However, the risk of major and gastrointestinal bleeding varies by specific agent^[16].

Gastrointestinal bleeding represents the most frequent hemorrhagic complication in the CKD population. In a multinational cohort of 527,226 new DOAC users, apixaban was associated with lower rates of gastrointestinal bleeding than rivaroxaban. This association remained consistent in the chronic kidney disease subset (hazard ratio [HR] 0.68, 95% CI 0.59 to 0.77)^[17]. These data support prioritizing apixaban as the first choice in CKD patients with prior gastrointestinal bleeding or known angiodysplasia.

Modifiable bleeding risk factors must be managed proactively. These include:

• concomitant antiplatelet therapy without an active indication;
• nonsteroidal anti-inflammatory drugs;
• uncontrolled hypertension;
• anemia.

Co-prescribing a proton pump inhibitor is a rational strategy in patients at elevated risk for gastrointestinal bleeding.

Drug Interactions

Because a primary elimination route is already impaired, renal impairment accentuates the clinical impact of drug interactions:

• strong dual CYP3A4 and P-glycoprotein inhibitors—such as ketoconazole, itraconazole, ritonavir-boosted regimens, and clarithromycin—increase apixaban and rivaroxaban exposure;
• P-glycoprotein inhibitors, including dronedarone, verapamil, amiodarone, cyclosporine, and quinidine, increase dabigatran and edoxaban exposure;
• strong inducers, such as rifampin, carbamazepine, phenytoin, and St. John’s wort, reduce exposure to all agents across this class and should generally be avoided.

When coadministered with specific P-glycoprotein inhibitors, edoxaban is dose-reduced to 30 mg under certain labeling guidelines. Combining dabigatran with dronedarone is contraindicated in Europe.

Monitoring Intervals

Expert clinical guidelines recommend assessing renal function at baseline and at least annually thereafter. When clearance declines to 30 to 60 mL/min, routine monitoring should occur at least every six months, and at least every three months when clearance is between 15 and 30 mL/min. In centers adopting a more dynamic, precautionary approach, the monitoring interval (in months) can be calculated by dividing creatinine clearance by 10 (e.g., every 3 months for 30 mL/min).

Intercurrent infections and acute illnesses, dehydration, initiation of nephrotoxic or interacting medications, and episodes of acute kidney injury warrant prompt unscheduled checks. Hemoglobin should be measured at least annually, and more frequently in advanced CKD. Routine therapeutic drug monitoring of plasma levels is not standard practice. However, DOAC-calibrated anti-Xa or dilute thrombin time assays can guide clinical management in the setting of extreme renal dysfunction or body weight extremes, unexplained bleeding, or before emergent surgery. Laboratory-level monitoring may be required in frail, obese, or moderately to severely impaired CKD patients^[18].

Bleeding Management Stratified by Clinical Acuity

Tier A: Minor bleeding in stable outpatients. Epistaxis, minor vascular access oozing, and ecchymosis fall into this category. These should be managed with local measures, omitting a single dose if necessary, and reviewing dose appropriateness and drug interactions. Anticoagulation should not be reflexively discontinued altogether.

Tier B: Non-life-threatening major bleeding requiring hospitalization. This category includes gastrointestinal hemorrhage presenting with hemoglobin drop or transfusion requirements. Management comprises:

• discontinuation of the anticoagulant;
• supportive care, including transfusion when indicated;
• endoscopic source control;
• determining the timing of drug resumption based on estimated drug clearance, which is markedly prolonged in severe CKD for dabigatran and edoxaban.

Tier C: Life-threatening or critical-site bleeding. Managed in the intensive care unit, this clinical presentation encompasses intracranial hemorrhage.

• For dabigatran, idarucizumab is the specific reversal agent, and hemodialysis can remove the drug from circulation.
• For apixaban and rivaroxaban, andexanet alfa or four-factor prothrombin complex concentrate is utilized where available. Hemodialysis is ineffective due to high protein binding.
• For warfarin, four-factor prothrombin complex concentrate administered alongside intravenous vitamin K is the standard approach.

Following major bleeding in dialysis patients, left atrial appendage occlusion should be rigorously considered rather than indefinitely re-exposing the patient to high-risk anticoagulation.

Left Atrial Appendage Occlusion: Patient Selection

Optimal candidates are dialysis or stage 5 CKD patients who meet the following criteria:

• high thromboembolic stroke risk associated with atrial fibrillation (substantially elevated CHA₂DS₂-VASc score);
• a documented contraindication to long-term anticoagulation, such as prior intracranial hemorrhage, recurrent angiodysplastic bleeding, or transfusion-dependent gastrointestinal loss;
• sufficient life expectancy to derive clinical benefit;
• favorable cardiac anatomy;
• capacity to tolerate short-term post-procedural antithrombotic therapy.

Patients unlikely to benefit should not be referred. This cohort includes:

• those with another indication for anticoagulation, such as recurrent VTE or mechanical prosthetic valves;
• those with left atrial thrombus;
• those with severely limited life expectancy;
• those unable to tolerate any antiplatelet therapy following implantation.

In clinical practice, clinicians should anticipate a reduction in bleeding burden rather than a mortality benefit from this intervention. Procedural complications and device-related thrombosis rates may run higher in this patient population than in general registry cohorts.

Future Perspectives and Open Questions

The most critical unanswered question is whether any anticoagulation strategy improves net clinical outcomes in dialysis-dependent atrial fibrillation patients. Three completed randomized controlled trials enrolled fewer than 400 patients in total. Two were terminated prior to achieving target recruitment, and none had sufficient statistical power to detect stroke endpoints. The sole trial demonstrating superiority evaluated an off-label dose of rivaroxaban within an open-label design characterized by high discontinuation rates. Placebo- or untreated-control trials in the dialysis population therefore remain an urgent priority. Given network meta-analysis data indicating that no anticoagulation yields comparable performance with respect to thromboembolism, clinical equipoise between apixaban and withholding anticoagulation is now a defensible position.

The optimal apixaban dose in dialysis patients remains uncertain. While observational signals favor 5 mg twice daily, pharmacokinetic data demonstrate that exposure approximately doubles compared with 2.5 mg twice daily. The inability of plasma drug exposure to clearly discriminate bleeders from non-bleeders raises the question of whether plasma level-guided dosing provides clinical value. Validated, widely accessible DOAC-calibrated assays and target therapeutic ranges linked to clinical outcomes remain unavailable.

The vascular calcification hypothesis warrants direct testing. If avoiding vitamin K antagonism mitigates calcific arteriolopathy, limb ischemia, and calciphylaxis, the argument for DOAC use in advanced CKD extends far beyond avoiding bleeding risk alone. Given that coadministration with rivaroxaban in the Valkyrie trial produced no discernible difference compared with rivaroxaban alone, whether vitamin K2 supplementation confers added benefit remains unclear.

Factor XI and XIa inhibitors represent a mechanistically compelling therapeutic pipeline for this population. This class encompasses antisense oligonucleotides, monoclonal antibodies such as abelacimab and osocimab, and small-molecule agents including asundexian and milvexian. The rationale hinges on factor XI’s contribution to pathological thrombosis rather than hemostasis, thereby offering the potential to uncouple antithrombotic efficacy from bleeding risk. Several of these agents exhibit limited renal clearance and are being investigated in end-stage renal disease, including dialysis circuit and vascular access thrombosis settings. Clinical outcome trials in atrial fibrillation patients with renal impairment are underway, and their results will determine whether the bleeding-dominated risk profile of anticoagulation in dialysis can finally be transformed.

Finally, the question of left atrial appendage occlusion must be resolved through randomized trials specifically in the renal impairment population. Propensity-matched data favor occlusion; however, these observational analyses cannot entirely exclude the healthy-candidate bias inherent in interventional procedural selection. Until such trials conclude, individualized decision-making will remain the standard of care. This process must be anchored in agent-specific pharmacology, Cockcroft-Gault-based dosing, meticulous review of interactions, and candid communication regarding the uncertain benefit profile.

References

1. Parul F, Ratnani T, Subramani S, Bhatia H, Ashmawy RE, Nair N, et al. Anticoagulation in Patients with End-Stage Renal Disease: A Critical Review. Healthcare (Basel). 2025;13(12):1373. doi:10.3390/healthcare13121373. PubMed
2. Leung KC, MacRae JM. Anticoagulation in CKD and ESRD. J Nephrol. 2019;32(5):719-731. doi:10.1007/s40620-019-00592-4. PubMed
3. Kumbhani DJ, Armbruster AL, Cheng RK, Chrispin J, Fischer U, Holbrook AM, et al. Direct Oral Anticoagulants in Primary and Secondary Prevention of Thrombotic Events: 2026 ACC Scientific Statement: A Report of the American College of Cardiology. J Am Coll Cardiol. 2026;88(11):1255-1283. doi:10.1016/j.jacc.2026.05.033. PubMed
4. Elenjickal EJ, Travlos CK, Marques P, Mavrakanas TA. Anticoagulation in Patients with Chronic Kidney Disease. Am J Nephrol. 2024;55(2):146-164. doi:10.1159/000535546. PubMed
5. Fanikos J, Burnett AE, Mahan CE, Dobesh PP. Renal Function Considerations for Stroke Prevention in Atrial Fibrillation. Am J Med. 2017;130(9):1015-1023. doi:10.1016/j.amjmed.2017.04.015. PubMed
6. De Vriese AS, Caluwé R, Van Der Meersch H, De Boeck K, De Bacquer D. Safety and Efficacy of Vitamin K Antagonists versus Rivaroxaban in Hemodialysis Patients with Atrial Fibrillation: A Multicenter Randomized Controlled Trial. J Am Soc Nephrol. 2021;32(6):1474-1483. doi:10.1681/ASN.2020111566. PubMed
7. Pokorney SD, Chertow GM, Al-Khalidi HR, Gallup D, Dignacco P, Mussina K, et al. Apixaban for Patients With Atrial Fibrillation on Hemodialysis: A Multicenter Randomized Controlled Trial. Circulation. 2022;146(23):1735-1745. doi:10.1161/CIRCULATIONAHA.121.054990. PubMed
8. Reinecke H, Engelbertz C, Bauersachs R, Breithardt G, Echterhoff HH, Gerß J, et al. A Randomized Controlled Trial Comparing Apixaban With the Vitamin K Antagonist Phenprocoumon in Patients on Chronic Hemodialysis: The AXADIA-AFNET 8 Study. Circulation. 2023;147(4):296-309. doi:10.1161/CIRCULATIONAHA.122.062779. PubMed
9. Siontis KC, Zhang X, Eckard A, Bhave N, Schaubel DE, He K, et al. Outcomes Associated With Apixaban Use in Patients With End-Stage Kidney Disease and Atrial Fibrillation in the United States. Circulation. 2018;138(15):1519-1529. doi:10.1161/CIRCULATIONAHA.118.035418. PubMed
10. Kuno T, Takagi H, Ando T, Sugiyama T, Miyashita S, Valentin N, et al. Oral Anticoagulation for Patients With Atrial Fibrillation on Long-Term Hemodialysis. J Am Coll Cardiol. 2020;75(3):273-285. doi:10.1016/j.jacc.2019.10.059. PubMed
11. Kao TW, Chen ZW, Lin YH. Anticoagulation for Patients With Concomitant Atrial Fibrillation and End-Stage Renal Disease: A Systematic Review and Network Meta-Analysis. J Am Heart Assoc. 2024;13(8):e034176. doi:10.1161/JAHA.123.034176. PubMed
12. Wetmore JB, Herzog CA, Yan H, Reyes JL, Weinhandl ED, Roetker NS. Apixaban versus Warfarin for Treatment of Venous Thromboembolism in Patients Receiving Long-Term Dialysis. Clin J Am Soc Nephrol. 2022;17(5):693-702. doi:10.2215/CJN.14021021. PubMed
13. Al-Abcha A, Saleh G, Sledge H, Abbasi M, Ismayl M, Killu AM, et al. Outcomes of Stroke Prevention Strategies in Patients With Atrial Fibrillation and End-Stage Renal Disease. JACC Cardiovasc Interv. 2025;18(21):2614-2624. doi:10.1016/j.jcin.2025.07.038. PubMed
14. Fazio G, Dentamaro I, Gambacurta R, Alcamo P, Colonna P. Safety of Edoxaban 30 mg in Elderly Patients with Severe Renal Impairment. Clin Drug Investig. 2018;38(11):1023-1030. doi:10.1007/s40261-018-0693-6. PubMed
15. Vu A, Qu TT, Ryu R, Nandkeolyar S, Jacobson A, Hong LT. Critical Analysis of Apixaban Dose Adjustment Criteria. Clin Appl Thromb Hemost. 2021;27:10760296211021158. doi:10.1177/10760296211021158. PubMed
16. Ballestri S, Romagnoli E, Arioli D, Coluccio V, Marrazzo A, Athanasiou A, et al. Risk and Management of Bleeding Complications with Direct Oral Anticoagulants in Patients with Atrial Fibrillation and Venous Thromboembolism: a Narrative Review. Adv Ther. 2023;40(1):41-66. doi:10.1007/s12325-022-02333-9. PubMed
17. Lau WCY, Torre CO, Man KKC, Stewart HM, Seager S, Van Zandt M, et al. Comparative Effectiveness and Safety Between Apixaban, Dabigatran, Edoxaban, and Rivaroxaban Among Patients With Atrial Fibrillation : A Multinational Population-Based Cohort Study. Ann Intern Med. 2022;175(11):1515-1524. doi:10.7326/M22-0511. PubMed
18. Olie RH, Winckers K, Rocca B, Ten Cate H. Oral Anticoagulants Beyond Warfarin. Annu Rev Pharmacol Toxicol. 2024;64:551-575. doi:10.1146/annurev-pharmtox-032823-122811. PubMed

Share this article

LinkedIn X Facebook WhatsApp Copy Link