It was with interest that we read “Non-valvular atrial fibrillation in patients undergoing chronic haemodialysis. Should we anti-coagulate?”,1 and we found it engaging and up-to-date, but we want to make some observations.

Thrombotic risk assessment to indicate anti-coagulation in patients with atrial fibrillation (AFib) in the general population has been based on the CHA2DS2-VASc. However, data from the recent SCREAM study call the validity of this score into question across the spectrum of chronic kidney disease (CKD), proposing the use of the modified CHADS.2 In dialysis patients, de Vriese et al. alternatively propose the "dialysis risk score", which includes prior stroke or transient ischaemic attack, diabetes, age >75 years, and gastrointestinal bleeding,3 which would drastically reduce the indication for anti-coagulation in these patients, although its validation is required. The authors comment on the poor predictive value of the bleeding risk of the scores used on the general population of hemodialysis, and mention the CHA2DS2-VASc.1 This does not measure the risk of bleeding, nor is it analysed in the referenced study, which may lead to confusion.

Regarding treatment with coumarins or vitamin K antagonists (VKA), the authors mention the risk of vascular calcification and arterial stiffness,1 which is correct, but it should be noted that they also increase the risk of peripheral vascular disease, valve calcification or calciphylaxis (already prevalent in this population), which calls into question their safety profile in dialysis.4,5

We agree with the authors that the time in therapeutic range (TTR) is low in patients with CKD,1 and especially those on haemodialysis, as demonstrated by clinical trials (CT) (Renal-AF or Valkyrie), in which better control of this is assumed, as well as in numerous observational studies. A worse TTR is associated with a higher risk of stroke or death,6 so nephrologists should be more involved and be aware of the TTR of our patients being treated with VKA so we can proactively suggest other therapeutic alternatives if this is unsatisfactory. Furthermore, an Italian group proposes that the variability of the international normalised ratio (INR) (high in this population) would be a better marker of mortality and risk of bleeding.7 In fact, de Vriese et al. propose that anti-coagulation in patients with AFib on dialysis should be done with apixaban or rivaroxaban for this reason, among others.3

Regarding the CTs with VKA vs no anti-coagulation in dialysis patients with AFib, in addition to the aforementioned AVKDIAL, the Danish Warfarin-Dialysis Study Safety and Efficacy of Warfarin in Patients with Atrial Fibrillation on Dialysis (DANWARD) (NCT03862859) is underway. Both studies will provide valuable information on the risk-benefit of VKAs in dialysis patients, which has been questioned by observational studies and meta-analyses.8,9

Regarding the comments made about warfarin-related nephropathy, this has a clear definition; acute kidney injury after an INR>3. Anticoagulant-related nephropathy has had its definition extended to include direct oral anti-coagulants (DOACs), with which it has also been described (although less frequently than with VKAs) in patients with underlying kidney disease.10 This should be differentiated and not confused with the more rapid deterioration of kidney function associated with VKA and recognised even in clinical guidelines11 and whose cause is probably multifactorial, according to the review by Gómez-Fernández et al.10

Regarding DOACs, apixaban and rivaroxaban seem to be alternatives to classical oral anti-coagulation, although some observational studies question the benefit of DOACs vs no anti-coagulation in this population, which needs to be demonstrated.12 The Valkyrie study, mentioned in the article, has demonstrated a reduction in cardiovascular events with rivaroxaban 10mg/day vs VKA (including fewer peripheral vascular events) with a better safety profile. Therefore, to date it is the only clinical trial that has demonstrated a better risk-benefit profile in this population, pending the results of the two studies with apixaban (AXADIA and SAFE-D) mentioned by the authors.1

Finally, although percutaneous left atrial appendage closure (LAAC) is an attractive alternative to oral anti-coagulation to prevent thromboembolic events in dialysis patients, as demonstrated by the study by Genovesi et al.13 mentioned in the article; a recent analysis of a LAAC registry with a much larger sample found higher mortality during hospitalisation and a trend towards increased mortality in the sub-group of dialysis patients.14

Therefore, anti-coagulant treatment in dialysis patients with AFib will remain a controversial issue due to the scant evidence from CT, starting with the selection of patients who may benefit from it. This requires a greater involvement of the nephrologist in the indication and follow-up of oral anticoagulation and, consideration of the alternative of DOACs (limited by approval for dialysis by the EMA), to avoid/reduce the risks associated with VKAs, and consideration of LAAC as an option, but carefully assessing the risk-benefit profile.