Comment on: Differential effect of anticoagulation according to cognitive function and frailty in older patients with atrial fibrillation.

Abstract

We read with interest the recent paper by Dr. Wang et al. highlighting outcomes differences according to physical and cognitive frailty of older patients with atrial fibrillation (AF) receiving or not oral anticoagulant therapy (OAT).1 The authors found cognitive impairment as an independent risk factor for the composite outcome of major bleeding and death, thus, indicating dementia as a deciding component for the anticoagulation prescription. Although current guidelines suggest avoiding OAT prescription on very frail older patients with AF,2 based on Dr. Wang and colleagues' findings, the presence of frailty syndrome does not appear to be associated with an increased risk of major bleeding or death at the 2-year follow-up. This seems to contradict a recent study published in JAGS by Candeloro et al.,3 showing an increased bleeding risk with advancing frailty degree in older outpatients with AF. In our opinion, this finding could be explained by the author's decision to include pre-frail and no-frail patients in a unique entity even though pre-frail patients and fit patients express several clinical and phenotypical intrinsic differences. Conversely, as cognitive impairment represents a major driver of both physical and cognitive frailty, it might have better represented frail patients, thus explaining its significance as a marker of increased bleeding and death risk in Wang et al. paper. In other words, the patients with cognitive impairment were frailer than those incorrectly defined as “frail” in the paper. Furthermore, the composite outcome chosen as the primary outcome (major bleeding or death) seems to be shared almost equally between major bleeding and death in cognitively impaired patients (12.5% and 13.6%, respectively). In a previous study from our lab carried out on more than 1800 oldest old patients with AF,4 a history of previous bleeding emerged as an independent predictor of major bleeding in patients receiving OAT at the 1-year follow-up. To assess the risk of OAT on the composite outcome, the authors correctly decided to perform a multivariable Cox Regression adjusting for a propensity score of several variables, including HASBLED. However, HASBLED has been proven not to be a good predictor in older patients.5 In contrast, a previous history of bleeding has been demonstrated to be strictly associated with bleeding risk following OAT prescription.4 To exclude the absence of an important residual confounding such as previous bleedings, we suggest the authors verify whether cognitively impaired patients receiving OAT showed a more frequent history of previous bleeding than their counterparts. If so, it would be appropriate to perform the multivariable Cox Regression analysis further, using a history of bleeding as covariate rather than HASBLED. Despite these two concerns, we believe that Dr. Wang and colleagues' study provides guidance for anticoagulation treatment in older patients with cognitive impairment and frailty syndrome. Chukwuma Okoye, Cristina Cargiolli, and Giulia Pescatore composed this letter jointly. All authors meet the criteria for authorship stated in the Uniform Requirements for Manuscripts Submitted to Bio-medical Journals. This letter has neither been published nor is it under consideration for publication elsewhere. The authors thank all the Cardio-geriatrics Interest Group of the Geriatrics Unit of Pisa University Hospital. The authors declare no conflicts of interest. No sponsorship was received for this letter.