Abstract
Objective
Patients with chronic kidney disease are at high risk of developing atrial fibrillation. However, it remains unclear whether systemic anticoagulants should be utilized, particularly in the later stages of chronic kidney disease.
Methods
This retrospective observational cohort study included patients with chronic kidney disease of various etiologies. This study aimed to record the prevalence of atrial fibrillation, use of systemic anticoagulation, and incidence of thromboembolic events and bleeding events among these patients.
Results
A total of 146 patients with chronic kidney disease were included in this study, with 43.8% of them experiencing atrial fibrillation. There was no significant difference in the prevalence of atrial fibrillation across various stages of chronic kidney disease. Thromboembolic events were not common in patients with atrial fibrillation but were prevalent in those with late-stage chronic kidney disease. Furthermore, patients with atrial fibrillation experienced a higher frequency of bleeding events, with the highest incidence observed in chronic kidney disease stage IV.
Conclusions
As systemic anticoagulation did not lower the frequency of thromboembolic events in patients with chronic kidney disease and atrial fibrillation in our study and considering that these patients experienced a higher frequency of bleeding complications, the use of systemic anticoagulants should be approached with caution, particularly in the later stages of chronic kidney disease.
Introduction
Since the seminal research of Go et al. in 2004, 1 the high risk of cardiovascular complications and cardiovascular death in patients with chronic kidney disease (CKD) has been definitively established. In the most recent iteration of the Kidney Disease: Improving Global Outcomes (KDIGO) guideline on CKD in 2024, 2 the correlation between a decline in estimated glomerular filtration rate (eGFR) or an increase in albuminuria and substantial clinical outcomes was demonstrated with a high degree of clarity. As CKD progresses, individuals below and above the age of 65 years frequently attain notable clinical outcomes (e.g. all-cause mortality and dialysis requirement) in disproportionate numbers. A crucial clinical endpoint is atrial fibrillation (AF). Individuals with CKD have a higher average prevalence of cardiac arrhythmia than those with healthy kidneys. The increase in prevalence is partly due to a number of factors, including arterial hypertension, coronary heart disease, renal anemia, latent systemic inflammation, disorders of bone homeostasis, and electrolyte imbalances. 3 The literature has also documented an increase in the risk of two potentially life-threatening events among those affected. First, AF in CKD is apparently associated with an even higher risk of thromboembolism than that reported in the general population. 4 Second, patients with CKD suffer from an increased tendency to bleed, even without the use of anticoagulant or antithrombogenic substances. 5 This may be due to the anticoagulant effects of uremic mediators and increased blood fluidity due to renal anemia, in addition to platelet dysfunction and increased vascular fragility. 6 To date, the question of whether and, if so, how systemic anticoagulation should be performed in patients with CKD and AF remains unresolved. This is particularly relevant to the late stages of the syndrome (stages IV and V in patients who had no dialysis requirement (ND) and those with dialysis requirement (D)). In patients requiring dialysis, the complexity of the decision-making process is further compounded by the limited available data. The present retrospective monocentric study analyzed patients with CKD of differing stages with or without a diagnosis of AF. The study’s three research questions were as follows: (a) was systemic anticoagulation initiated, and if so, which preparation was used?; (b) what rates of thromboembolism and bleeding can be determined in the sample?; (c) do the findings have any consequences for practical management?
Methods
Design
This single-center, retrospective observational cohort study was conducted at the Department of Internal Medicine 1 (Cardiology, Angiology, Pneumology, Nephrology) of the Brandenburg University Hospital in the Brandenburg Medical School. Due to the retrospective nature of the study and the extraction of all data from the central medical database of the hospital, it was not necessary to obtain written consent for participation. Our study was conducted in accordance with the Helsinki Declaration of 1975 as revised in 2024. 7 The reporting of this study conforms to Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines. 8
Patients
This study included adult patients (aged ≥18 years) who were treated as inpatients at the Brandenburg University Hospital. The identification of patients with CKD was conducted using anonymized case number lists from the Department of Medical Controlling. The recruitment period for the present study ranged from 2021 to 2022.
The inclusion criteria were individuals aged ≥18 years who had a diagnosis of CKD of severity grade 2, irrespective of the etiology of the syndrome. The objective was to obtain a sample that was as representative as possible. The basis was a sustained reduction in eGFR (CKD–Epidemiology Collaboration (EPI)) to <90 mL/min for a minimum of 3 months that could be traced. Albuminuria grading was not recorded separately due to the absence of corresponding determinations in a significant number of cases. The exclusion criteria were individuals with verifiable acute renal dysfunction according to KDIGO 2012 9 and patients who were pregnant or terminally ill, including those receiving palliative therapy.
CKD was categorized into stages II (89–60 mL/min), III (59–30 mL/min), IV (29–15 mL/min), and V (<15 mL/min, either with D or ND) based on the eGFR according to CKD–EPI. 2 Anamnestic, clinical, laboratory chemical, and other examination findings were extracted and documented from the clinic’s central medical database (MEDICO CGM).
Endpoints
The primary endpoint was the prevalence of AF. The endpoint was considered met if patients with CKD suffered from paroxysmal, persistent, or permanent AF at the time of inclusion. Secondary endpoints included the initiation of therapeutic anticoagulation and the prevalence of thromboembolism (thromboembolic events (TE)) and bleeding events (BE). The secondary endpoint—initiation of therapeutic anticoagulation—was met if, at the time of study inclusion, treatment with a preparation of the following active substance groups in therapeutic dosage was ongoing: direct oral anticoagulants (DOACs; such as apixaban, edoxaban, and rivaroxaban), vitamin K antagonists (VKAs; such as phenprocoumon and warfarin), low molecular weight heparin (LMWH), or others (e.g. fondaparinux). The secondary endpoints—TE and BE—were met if any one of the following events was documented within 3 months before study inclusion: BE (gastrointestinal, urogenital, pulmonary, or other bleeding) or TE (cerebrovascular or other peripheral thromboembolism of left atrial origin).
Statistical analysis
The normality of numerical data sets was established using the Shapiro–Wilk test. Normally distributed data were compared using the t-test (two groups) or analysis of variance (more than two groups), whereas non-normally distributed data were compared using the Mann–Whitney U test (two groups) or Kruskal–Wallis test (more than two groups). Finally, the chi-squared test was used to compare categorical data. The chi-squared test was also used to compare categorical data from multiple groups; the groups were formed in such a way that there was a comparable group size. The results were presented as mean values ± standard deviation, median ± interquartile range, or percentage. A p-value <0.05 was considered to indicate statistical significance. All analyses were performed using the following programs: Wizard for MacOS (version 2.0.16) or Graphpad Prism (version 10.4.1). Graphpad Prism and Datagraph for MacOS (version 5.4) were used to create graphs.
Results
Patients
A total of 146 patients with CKD were included in the study. Of these, 43.2% (n = 63) were female and 56.8% (n = 83) were male. The mean age of the sample was 77.4 ± 11.5 years. The distribution of CKD stages is outlined below: II (7.5%), III (52.1%), III (17.1%), and V (23.3%). The average length of hospital stay was 12 ± 10.9 days. The following types of drugs were used for anticoagulation: LMWH, DOACs, VKAs, and others in 7.5%, 52.1%, 17.2%, and 23.3%, respectively. In the DOAC group, the dosage of the medication was adapted to renal function in 88.1% of cases, in accordance with the manufacturer’s instructions. Table 1 summarizes the key demographic, clinical, and laboratory data of the participants at the time of study inclusion.
Patient characteristics.
SD: standard deviation; CKD: chronic kidney disease; LMWH: low molecular weight heparin; DOACs: direct oral anticoagulants; VKA: vitamin K antagonists; TE: thromboembolic event; BE: bleeding event; COPD: chronic obstructive pulmonary disease; CAD: coronary artery disease; MI: myocardial infarction; HF: heart failure; PAD: peripheral artery disease; CVD: cerebrovascular disease.
Endpoints
Primary endpoints
AF was diagnosed in 43.8% (n = 64) of patients in the total sample. No significant difference was observed between the CKD stages with regard to AF diagnosis: II (45.5%), III (43.4%), IV (44%), and V (44.1%) (p = 0.99) (Figure 1).
Primary endpoint. (a) Atrial fibrillation was diagnosed in 43.8% of patients in the total sample and (b) no significant difference was observed between the stages of CKD (p = 0.99).
Secondary endpoints
Overall, 47.3% of all patients with CKD received systemic anticoagulation. As this percentage was higher than the prevalence of AF, some cases had alternative indications for the use of anticoagulants. Overall, 87.5% of patients diagnosed with AF were prescribed anticoagulation therapy, whereas 15.9% of patients without AF received anticoagulants for other indications. The distribution of anticoagulants used in the CKD cohort with AF was as follows: LMWH (5.8%), non-VKAs (66.1%), VKAs (20.3%), and others (6.8%). A TE was detected in 10.4% of all patients, with a higher percentage of 10.9% being recorded in patients with AF. A BE was documented in 12.5% of the total sample and 18.8% of all patients with CKD and AF (Figure 2).
Secondary endpoints. Systemic anticoagulation was used in 47.3% of all patients with chronic kidney disease (a) and in 87.5% of those with atrial fibrillation (b). Thromboembolic events did not occur significantly more often in patients with atrial fibrillation (c), although the risk of bleeding was higher in the latter (d). Direct anticoagulants were by far the most commonly used for systemic anticoagulation (e).
Risk factor analysis
In the risk factor analysis, TE and BE were set as dependent variables.
TE
The prevalence of TE did not differ between six equally sized age categories (44–65, 66–73, 74–80, 81–83, 84–87, and 88–98 years; p = 0.21). The rate of TE was comparable between women and men (9.7% vs. 11%, p = 0.8). Regarding CKD stages, the highest prevalence of TE was detected in stage V (26.5%), followed by stage III (8.1%) and stages IV and II (0% each) (p = 0.003). There was no significant difference in the TE rate between DOAC and VKA groups (6.7 vs. 18.8%, p = 0.16). The rate of thromboembolism did not differ between patients with AF and those without AF (10.9 vs. 10%, p = 0.88) or between patients with systemic anticoagulation and those without systemic anticoagulation (11.6 vs. 9.3%, p = 0.65). In particular, in stage V, the TE rate was not significantly different between patients with AF and those without AF (26.7 vs. 26.3%, p = 0.98). The following comorbidities were also not associated with increased TE rates: arterial hypertension (p = 0.36), diabetes mellitus (p = 0.8), obesity (p = 0.13), chronic obstructive pulmonary disease (COPD) (p = 0.41), hyperlipidemia (p = 0.44), liver cirrhosis (p = 0.43), coronary heart disease (p = 0.08), history of myocardial infarction (p = 0.51), peripheral artery disease (p = 0.06), heart failure (p = 0.56), or depression (p = 0.73). Patients with known cerebrovascular disease, however, had significantly more TEs than those without (47.1 vs. 5.5%, p < 0.001). The TE rate varied among the six categories of hemoglobin (Hb) concentration (equal group size), with the highest rate in the 111–119 g/L category (28%). The rates in other categories were as follows: 50–90 g/L (12.5%), 91–98 g/L (4.3%), 99–109 g/L (4.3%), 120–128 g/L (8.3%), and 129–177 g/L (4%); p = 0.04). Concomitant medication other than possible oral anticoagulation was not associated with different TE rates: ASS (p = 0.33), NSAIDs (p = 0.9), amiodarone (p = 0.9), glycosides (p = 0.89), beta-blockers (p = 0.9), loop diuretics or thiazides (p = 0.44 and p = 0.68), statins (p = 0.65), insulin (p = 0.91), oral antidiabetics (p = 0.68), and proton pump inhibitors (p = 0.57). Table 2 presents all significant findings from the TE (and BE) risk factor analysis.
Risk factor analysis. The dependent variable is named in the second column (thromboembolic events or bleeding events), whereas independent variables are listed in column 1. Regarding sex, for example, the presentation would be as follows: 9.7% of all women developed a TE, compared with 11% of all men (p = 0.8).
CKD: chronic kidney disease; COPD: chronic obstructive pulmonary disease; CAD: coronary artery disease; MI: myocardial infarction; HF: heart failure; PAD: peripheral artery disease; CVD: cerebrovascular disease; ASS: acetylsalicylic acid; NSAIDs: nonsteroidal anti-inflammatory drugs.
BE
In contrast to TE rates, the BE rate differed between six age categories (44–65, 66–73, 74–80, 81–83, 84–87, 88–98 years). The highest rate was observed in the 84–87 years category (29.2%, p = 0.036). There was no statistically significant difference in the BE rate between males and females (12.9% vs. 12.2%, p = 0.89). There were also no significant differences in BE rates between the various CKD stages (II (9.1%), III (12.2%), IV (20%), V (8.8%); p = 0.6). Under DOACs, the BE rate was 24.4% vs. 12.5% under VKA (p = 0.31). AF was associated with a higher BE rate (18.8 vs. 7.5%, p = 0.04), as was the use of anticoagulants (20.3 vs. 5.3%, p = 0.007). None of the comorbidities were associated with different bleeding rates, with one exception (negative neoplasia history (15.1% vs. 0%; p = 0.03)): arterial hypertension (p = 0.41), diabetes mellitus (p = 0.52), obesity (p = 0.06), COPD (p = 0.07), hyperlipidemia (p = 0.89), liver cirrhosis (p = 0.6), coronary heart disease (p = 0.48), history of myocardial infarction (p = 0.92), peripheral artery disease (p = 1), heart failure (p = 0.89), cerebrovascular disease (p = 0.49), or depression (p = 0.3). The BE rate differed significantly between different Hb categories (equal group size per category): 50–90 g/L (29.2%), 90–98 g/L (21.7%), 99–109 g/L (4.3%), 110–119 g/L (12%), 120–128 g/L (4.2%), and 129–177 (4%) (p = 0.029). With regard to concomitant medication, a significantly higher risk of bleeding was identified in patients not receiving ASS therapy compared with those receiving it (15.8% vs. 2.3%, p = 0.04). Finally, patients with AF and BE were analyzed with regard to the influence of dual anticoagulant/antithrombotic therapy on the risk of bleeding. Patients were categorized based on their receipt of dual therapy, comprising systemic anticoagulant and ASS, and compared with those who received only anticoagulant therapy. The findings revealed that 8.6% of patients in the AF group without BE received dual therapy, while no patients in the BE group received dual therapy (p = 0.29). No significant differences in BE rates were found for the other medications: nonsteroidal anti-inflammatory drugs (NSAIDs) (p = 0.43), amiodarone (p = 0.43), glycosides (p = 0.30), beta-blockers (p = 0.38), loop diuretics or thiazides (p = 0.42 and p = 0.12), statins (p = 0.38), insulin (p = 0.45), oral antidiabetics (p = 0.18), and proton pump inhibitors (p = 0.23) (Table 2).
Discussion
First, it should be noted that the study sample is inherently limited. Nevertheless, some findings are immediately striking: on the one hand, the comparable TE rates in patients with compared to those without AF and the comparable rates in individuals with vs. without systemic anticoagulation. TEs were identified most frequently in the highest CKD stage. Second, the elevated bleeding risk observed in patients with AF was unexpected; however, the bleeding rate was also found to be significantly higher in individuals undergoing systemic anticoagulation. Notably, no clear correlation with the CKD stage was observed with respect to the risk of bleeding. Equally surprising was the fact that the other independent variables analyzed, with a few exceptions (Hb, cardiovascular disease, and history of neoplasia), showed no associations with either TE or BE. In theory, obesity, for example, could have been associated with an increased tendency to thrombosis. The question now arises as to what practical implications arise from the data, irrespective of the limited number of test subjects included and the retrospective design. Whether and, if so, how CKD patients with AF should be systemically anticoagulated has been the subject of intense debate for years.
It should be noted that systemic anticoagulants most likely reduce the risk of stroke in patients with CKD and AF. The meta-analysis published by Hart et al. 10 in 2007 indicated a 60% reduction in relative risk with warfarin. Conversely, VKAs increase the risk of bleeding in CKD, as shown in the study by Jun et al. 11 However, the authors explicitly did not consider individuals with CKD stage 5D. In recent years, newer oral anticoagulants (DOACs) have been systematically compared with VKAs, mainly with an aim to reduce the risk of TE and the risk of bleeding to a comparable extent. In 2024, Elenjickal et al. 12 published a review article that addressed this issue in patients with CKD of various severities. It was initially emphasized that up to an eGFR of 30 mL/min, various DOACs are variably superior to VKAs in terms of TE and BE. The use of dabigatran or apixaban is associated with lower TE rates, and apixaban and edoxaban are also associated with a lower risk of bleeding. The use of DOACs in patients on dialysis was evaluated in the prospective, randomized Valkyrie study. 13 A total of 132 patients with CKD-5D were enrolled and treated with either VKA or rivaroxaban alone or in combination with vitamin K. Follow-up was conducted 18 months later, and the primary endpoint was a composite of fatal and nonfatal cardiovascular events. The safety endpoint was bleeding. The primary endpoint was met with a highly significant reduction in incidence in both DOAC groups (p = 0.0006 and p = 0.0003). In particular, there was a clear benefit for the use of DOACs in the symptomatic limb ischemia category. BEs were also significantly less frequent with DOAC therapy when both DOAC groups were pooled (p = 0.02). Finally, the advantage of new therapeutic agents over VKA was highlighted, with a need for dose reduction. The AXADIA-AFNET 8 trial 14 published in 2022 also focused on CKD-5D patients. Apixaban and VKA (phenprocoumon) were compared over a period of >1 year. No significant difference was found between the treatment groups with regard to efficacy (prevention of ischemic stroke, heart attack) or safety (BEs). The study objective of demonstrating that apixaban is not inferior to VKA was achieved to a certain extent. In the same year, the Renal-AF study 15 was published, which involved patients with CKD who were on dialysis. Apixaban was used in this study as well; however, it was compared with warfarin. Ultimately, 154 patients were enrolled, of whom 82 were treated with apixaban and 72 with warfarin. The trial was stopped early due to recruitment problems. However, there was no significant difference between the treatment groups in terms of efficacy and safety. The authors noted that the relatively small number of patients included in the study was a possible limitation, indicating that only limited conclusions could be drawn. The data did not allow a clear recommendation for the preferential use of DOACs in terminal-stage CKD. In addition, it remains unclear whether anticoagulants should be used at all in this situation. However, DOACs are the preferred option for CKD down to an eGFR of 30 mL/min. 4 A 2024 review by Wing et al. 4 discusses the dilemma of decision making for anticoagulation in CKD-5D with AF. The final recommendation or advice is that the responsibility for the use of anticoagulants in this population lies with the physician. The weight given to the two components TE and BE must be considered on an individual basis. At this point, our own data should be recapitulated in an attempt to draw a conclusion for clinical practice. Since our cohort showed no significant advantage of anticoagulation with regard to the risk of TE, but on the other hand the therapy was associated with an increased risk of BE, we would suggest a cautious use of systemic anticoagulants in patients with CKD-5D. This conclusion also seems justifiable because there was no difference in the TE rate between patients with and without AF in CKD stage 5. Perhaps an intervention should only be considered at all if thromboembolism has already manifested itself once. Nevertheless, given the retrospective design of our study, it is not possible to assume that this statement is universally valid.
Limitations
The most important limitation of this study is its retrospective, observational design. It is well known that retrospective studies are associated with a number of disadvantages, such as incomplete data, limited possibility of recognizing causal relationships, and possible overestimation of treatment effects. Another limitation of our study is the relatively small number of patients included. Taken together, there is an urgent need for a prospective follow-up analysis to verify the obtained data. The conclusion of this study should also be interpreted with caution. Our study is an attempt to increase patient safety. Another limitation is the heterogeneity of the systemic anticoagulants used.
Footnotes
Acknowledgements
None.
Author contributions
H. Hansen-Nootbaar identified eligible patients and collected all data. F. Stasche prepared figures and tables. O. Ritter analyzed data. D. Patschan designed the study, analyzed data, and wrote the article. All three authors approved the final version of the article.
Data availability statement
All data will be provided by the corresponding author (
Declaration of conflicting interests
The authors declare that they have no conflicts of interest.
Funding
No funding was provided for the study.
