We included 250 patients with a diagnosis of CKD, all with an eGFR <30 ml/min/1.73 m², corresponding to stages G4 and G5. The mean age was 78.6 (±14.3 years) The cohort was mostly composed of males (64 %), and presented a high burden of comorbidity, highlighting: Diabetes mellitus in 32 %, Hypertension in 82 %, cardiovascular history (previous AMI or peripheral vascular disease) in 21 %, History of stroke or TIA in 14 %.The mean follow-up was 25.5 months, with at least two complete revisions per patient.
Of the 250 patients, 70 (29.2 %) were actively anticoagulated at the time of analysis. The only clinical indication recorded for anticoagulation was the presence of AF. The distribution according to type of anticoagulant was: Dicoumarinics (44.3 %), predominantly with acenocoumarol. Direct-acting oral anticoagulants (44.3 %), mainly apixaban (mostly at a reduced dose of 2.5 mg/12 h) and rivaroxaban (10-15 mg/day). Low molecular weight heparin (8.6%), usually in the context of transition, temporary contraindication of anticoagulate or digestive comorbidity.
Of the 70 anticoagulated patients, only 19 (27%) had been assessed by the nephrology department as appropriate candidates for anticoagulation according to individualized clinical criteria. In the rest of the cases, the initial indication came from primary care or cardiology services, which reflects a frequent referral pattern without first passing through nephrology.
They were analyzed according to the three risk stratification scales most commonly used in clinical practice: CHA2DS2-VASc 98.6% presented a score ≥2, considered high risk of stroke. Only 1 patient (1.4%) had a score <2 (3.1 ± 1.2) reflecting a significant cardiovascular burden. The HAS-BLED 95.7% had a score ≥3 (3.5 ± 0.8) indicating high bleeding risk.
The Mayo Clinic scale, showed markedly different results 51 patients (72.9%) presented a score ≤1, which, would not justify anticoagulation. 21.4% scored =2, which would warrant anticoagulation at a reduced dose. Only 5.7% scored ≥3 points, which would indicate full-dose anticoagulation.
This reflects a significant discordance with the traditional scales. Most of the patients anticoagulated according to CHA2DS2-VASc would not be anticoagulated according to the Mayo Clinic scale, underscoring its capacity to avoid overtreatment in CKD.
Bleeding events
During follow-up, 31 bleeding events (44.3%) were documented among the 70 anticoagulated patients: severe bleeding in 19 cases (61.3%) required hospitalization, transfusion or medical intervention, gastrointestinal (n=11), genitourinary (n=5), intracranial hemorrhage (n=3). Mild bleeding in 12 cases (38.7 %), self-limited or managed on an outpatient basis.
In patients with Mayo score ≤1, 24 of 31 bleeds (77 %) were observed, reinforcing the hypothesis of overanticoagulation in this subgroup.
In comparison, the 6 patients with Mayo score ≥2 and anticoagulated presented only 2 mild bleeding events, with no major complications. Although multivariate regression was not performed, this distribution suggests a strong association between low Mayo score and higher incidence of bleeding on anticoagulation.
DISCUSSION
The results reveal an important discrepancy between the traditional scales and the new tool. More than 98% of anticoagulated patients met criteria for high thromboembolic risk according to the CHA2DS2-VASc score, whereas 96% had a HAS-BLED score ≥3, evidencing a clinical profile that simultaneously suggests the need for anticoagulation and a very high bleeding risk.
This paradox is well known in the CKD population, where the coexistence of procoagulant states and hemostatic dysfunction creates a difficult-to-manage clinical setting.
However, application of the Mayo Clinic scale revealed that 73 % of anticoagulated patients would have been excluded from the indication for anticoagulation, presenting a score ≤1. This finding is clinically significant, since the group with a low score (Mayo Clinic) accumulated the highest number of bleeding events (77%), including 19 severe bleeds, which reinforces the usefulness of this tool as a discriminatory instrument to avoid unnecessary anticoagulation.
The Mayo Clinic scale integrates specific variables in CKD, such as albuminuria, low body weight, use of antiplatelet agents and advanced age, factors that are not included in the classic scales. This approach allows adjustment for the real risk of bleeding complications in the renal population, which represents a significant advantage over traditional models developed in populations without CKD.
Our data coincide with previous studies that warn of the overestimation of the benefit of anticoagulation in patients with CKD when generalist scales are applied. Several studies have shown that the use of vitamin K antagonists in this population not only does not reduce the risk of stroke, but may increase the risk of major bleeding, calciphylaxis and anticoagulant-associated nephropathy. Even direct-acting anticoagulants, although with a better safety profile, have limitations in this population, and their efficacy has not been clearly demonstrated in filtration rates <30 ml/min or in patients on renal replacement therapy.
The high percentage of anticoagulation prescriptions initiated by other specialties, mainly primary care and cardiology, is also a relevant finding. This suggests that nephrology is not yet fully integrated in anticoagulation decision making, despite the complex clinical profile of these patients. Incorporating tools such as the Mayo scale in interdisciplinary shared assessment could improve the selection of real candidates for anticoagulation, optimize therapeutic risk-benefit and avoid serious complications.
Our results support the idea that anticoagulation in patients with CKD should be individualized and reevaluated periodically, beyond the scoring of isolated scales. In cases with intermediate scores, reduced-dose anticoagulation could be considered, always under close control. This attitude is aligned with the current recommendations of the AHA/ACC/HRS 2023 guidelines, which promote a more nuanced and patient-centered approach.
The classic scales can lead to overestimation of anticoagulation in patients with CKD. The Mayo Clinic scale allows restricting anticoagulation to patients with greater potential benefit, avoiding adverse events.
A more individualized and multidisciplinary approach is needed, with nephrology playing a key role in the decision to anticoagulate.
Limitations of the study
Retrospective and single-center design, which limits the generalizability of the results.
The rate of thromboembolic events was not analyzed, so we cannot evaluate the balance between bleeding and stroke prevention.
Nevertheless, the sample is clinically representative and allows us to reflect critically on current clinical practice in patients with CKD and AF.
