AF in end-stage renal disease: anticoagulation and burden

Atrial fibrillation in end-stage renal disease is more than a rhythm disturbance: it is a marker of high thromboembolic and bleeding risk. Kidney failure, its treatment and the underlying disease can promote atrial structural remodelling, while hypertension, inflammation, anaemia and repeated fluid shifts alter electrical stability. The result is a population in which stroke prevention, drug handling and arrhythmia control must be planned together.

Many people with advanced kidney failure receive haemodialysis in outpatient units in Melbourne, Sydney, Brisbane, Perth and Adelaide; others live at home or travel long distances to regional services. A dialysis schedule adds constraints to medication adherence, blood tests and procedures. Decisions made in a metropolitan cardiology clinic must also fit the routines of a local renal unit and what a patient can obtain through a community pharmacy.

AF may be persistent, paroxysmal or asymptomatic. A normal screening ECG does not exclude it, and intermittent episodes can be mistaken for dizziness, breathlessness or fatigue associated with uraemia or heart failure. At the same time, the duration of AF is not a substitute for a stroke-risk assessment. Reduced kidney function changes the pharmacokinetics of anticoagulants, and dialysis-dependent patients often have several medicines and comorbidities that narrow the safe options.

For Australian clinicians, the practical issue is not whether dialysis automatically dictates one anticoagulant. It is how to balance expected benefit, bleeding risk, renal clearance, drug interactions, adherence and patient preference while following current product information and guideline advice. A structured assessment should combine Australian prescribing information, renal dosing guidance, local anticoagulation pathways and input from cardiology, nephrology and pharmacy.

Why kidney failure alters thromboembolic risk

Kidney failure raises AF risk through atrial enlargement, fibrosis and conduction abnormalities. Uraemic toxins, chronic inflammation, sympathetic activation, hypertension and left ventricular dysfunction create an atrial substrate vulnerable to re-entry. Haemodialysis adds acute electrolyte and volume changes; missed sessions, intercurrent illness and differing dialysate prescriptions may supply further triggers. These mechanisms can also increase ventricular arrhythmia risk, so symptoms should prompt a broader rhythm assessment rather than an assumption of simple AF.

CHA2DS2-VASc remains a useful starting point, but it does not fully capture the risk attached to kidney failure. Dialysis, prior stroke, heart failure, vascular disease and diabetes may cluster in the same patient, while uraemic endothelial dysfunction and left atrial appendage changes add factors absent from the score. A low numerical score should not be the sole reason to discount anticoagulation in advanced kidney disease, and AF duration or a single episode of sinus rhythm does not erase the need for risk assessment.

Countervailing bleeding risks are common: anaemia, frequent phlebotomy, invasive access, gastrointestinal lesions and antiplatelet or antithrombotic medicines used for other indications. Yet bleeding risk is not a reason to withhold anticoagulation automatically, because thromboembolism can be catastrophic and is often more preventable. The assessment should record bleeding sites, transfusion needs, recent falls, planned procedures and access complications, then address modifiable hazards while preserving stroke protection. The journal’s clinical resource hub brings together evidence and educational material on anticoagulation, monitoring and device-based care.

Comparing anticoagulants for dialysis patients

Warfarin has the longest dialysis experience, and KDIGO suggests a vitamin K antagonist in preference to a direct oral anticoagulant for CKD G5, including dialysis, because direct oral anticoagulant evidence in this group is limited. Randomised outcome data for warfarin in dialysis also remain limited. Its advantages include low cost, familiar protocols and available reversal strategies. Disadvantages include a narrow therapeutic window, frequent international normalised ratio testing, variable dose response and interactions with antibiotics, antifungals, antiarrhythmics and other common drugs.

Stronger randomised evidence exists for apixaban in haemodialysis. The open-label CARAVAGGIO trial found noninferiority to warfarin for a composite of major bleeding and thromboembolic events, but major bleeding was frequent in both groups and the study does not establish that every patient with kidney failure can safely receive apixaban. Other direct oral anticoagulants have less dialysis-specific evidence. Renal clearance changes exposure, and Australian product information for several agents excludes or does not recommend use below defined creatinine-clearance thresholds.

Market availability does not equal guideline endorsement. Apixaban is Pharmaceutical Benefits Scheme-listed for eligible patients with non-valvular AF, and warfarin is also widely available, so affordability can shape the conversation; current PBS criteria should be checked for each prescription. Hospital formularies may differ from community dispensing, and a patient moving between states or between a renal service and a general practice can encounter changes in brand, pack size and brand-generic substitution.

Aspirin and clopidogrel should not be used as substitutes for anticoagulation in AF unless there is a separate indication. Combining them increases bleeding without providing equivalent stroke prevention. Where oral medication is difficult, adherence can be supported through pharmacy repeat collections, dose aids, blister packs, community nursing and chronic disease management plans, but no packaging solution replaces a deliberate choice of agent and dose.

Building an individualised anticoagulation plan

The initial conversation should establish residual urine output, dialysis vintage and access, AF pattern, prior thromboembolism, heart failure and other relevant conditions. It should also include falls, gastrointestinal bleeding, transfusion history, cognitive impairment, medication adherence and the patient’s goals. Age alone should not determine treatment. A person may value continued home or work activities enough to accept monitoring and procedural risk, while another may prioritise fewer blood tests and simpler routines.

For warfarin, Australian renal units commonly coordinate regular INR checks with pathology services, antihypertensive clinics and dialysis appointments. Stable INR results can support use, but dialysis, acute illness and antibiotic courses can shift the dose abruptly. Home INR testing may suit selected patients with good device access and training. Reviewing medicines at every dialysis review is important: an interacting prescription from a general practitioner, hospital or emergency department may be missed if reconciliation is confined to the dialysis ward.

If a direct oral anticoagulant is selected under a specialist-led plan, the team should confirm the indication, renal estimate, product information, dose, interactions and monitoring plan before dispensing. Cockcroft–Gault creatinine clearance is often used for dosing, although its limitations with changing muscle mass and acute illness should be recognised. Reduced dosing must follow the specified criteria, not dialysis alone. Apixaban criteria may involve age and weight as well as serum creatinine. Multidisciplinary teams can turn these structured decisions into comparative evidence; the journal’s author submission guidance sets out pathways for relevant research.

Defining arrhythmia burden beyond the ECG

Arrhythmia burden should be measured with a method matched to the question. Twelve-lead ECG and telemetry may capture persistent or inpatient episodes; ambulatory monitors can detect paroxysmal AF; and an implantable loop recorder or pacemaker/defibrillator interrogation may be needed when episodes remain unexplained. In kidney failure, breathlessness, fatigue and palpitations overlap with fluid overload and anaemia, so timing symptoms against haemodialysis, medication changes and intercurrent illness can make a provisional diagnosis unreliable.

Burden describes more than the number of episodes. Total AF time, longest episode, ventricular rate and symptom correlation influence treatment and may predict heart failure or recurrence. However, no practical monitoring threshold converts directly into a thromboembolic decision. Stroke risk remains linked to the patient’s risk profile and AF pattern, and paroxysmal AF can cause embolism. A low-burden result should therefore narrow the rhythm plan without being used to stop anticoagulation.

Continuous monitoring also changes the value of a negative result. Repeated short recordings may miss clinically important AF, while an implantable device can expose recurrence before it becomes symptomatic. Device interrogation may integrate duration, rate and symptom timing more closely than isolated ECG snapshots. Clinicians should also ask whether symptoms reflect AF, fluid shifts, angina, infection or another arrhythmia before escalating rhythm treatment.

Rate control, rhythm control and clinical action

Rate control must be adapted to renal clearance and cardiac function. Beta-blockers may suit many patients, while verapamil or diltiazem may help selected patients with preserved ejection fraction but can cause bradycardia or worsen systolic dysfunction. Digoxin has a narrow safety margin in kidney failure; reduced dosing, concentration checks and attention to interacting medicines are essential. Amiodarone can be effective, yet cumulative toxicity, thyroid dysfunction, pulmonary complications and warfarin interactions make long-term use less attractive for some patients.

Rhythm control remains valuable when AF contributes to heart failure, haemodynamic instability or substantial symptoms despite rate control. Electrical cardioversion may be appropriate in urgent settings, and catheter ablation with pulmonary vein isolation can reduce recurrence in selected patients. Advanced kidney disease creates technical and prognostic considerations, including atrial fibrosis, access issues, procedural anticoagulation and a higher likelihood of recurrent arrhythmia. Anticoagulation should not be stopped automatically when an ECG looks normal; guidance requires a defined post-ablation period and subsequent reassessment of stroke risk.

Interventional arrhythmia practice extends beyond AF. The journal’s report on surgical denervation outcomes in refractory ventricular arrhythmias illustrates the wider electrophysiology context, but those findings do not justify routine sympathetic denervation for dialysis-dependent AF. The distinction matters when promising procedural options are discussed with patients.

Clinicians and trainees can contribute through original research, case-based reviews or multidisciplinary studies; prospective cohorts linking anticoagulation exposure with bleeding, thromboembolism and device-measured AF burden would be particularly valuable. Start the next dialysis review with a combined stroke-risk, bleeding-risk, renal-function and AF-burden assessment, document the anticoagulation rationale, and arrange timely cardiology or nephrology input whenever the safe choice remains uncertain.