Our study including elderly patients with AF demonstrates a high risk of worse clinical outcomes in patients hospitalized and with a reduced functional mobility.

In particular, from the Cox multivariate regression model, it emerged that having an SPPB < 8 points is associated with a 2.78-fold higher risk of CVEs compared to an SPPB ≥ 8 points. The analysis of the processed ROC curve and measurement of relative AUC also demonstrated the accuracy of SPPB pathological value as predictor of CVEs occurrence. These results are clinically relevant because functional limitation is frequent in elderly patients. Of particular interest, our study shows that treatment with statins, male gender, and RAASi treatment were associated with a reduced risk of CVEs incidence by 41%, 34%, and 33%, respectively; while pathological GDS and treatment with VKAs prior to DOACs therapy were associated with an increased risk of major cardiovascular events by 54% and 35%, respectively. Functional limitation, therefore, associated with decreased exercise tolerance, seems to increase the risk of nonfatal stroke, nonfatal acute coronary events, and both cardiovascular and non-cardiovascular mortality. This is relevant because AF is often associated with other risk factors and clinical conditions that can contribute to functional limitation and worsen prognosis, such as advanced age, comorbidities, frailty, and lifestyle; these factors increase the risk of adverse events in patients and negatively influence prognosis. To the best of our knowledge, there are no studies that have evaluated SPPB as a prognostic factor associated with the risk of CVEs in hospitalized elderly patients with NVAF receiving DOACs therapy. However, several studies have been conducted on the elderly population and have correlated pathological SPPB with several endpoints. A prospective cohort study evaluated the association between SPPB and CoI in elderly patients and found that decreased physical performance, assessed as SPPB < 9 points, was independently associated with the risk of CoI (HR 2.222, 95% CI 1.047–4.716, p = 0.038) [19]. Similar results were found in our study, as the group with SPPB < 8 points showed significantly lower MMSE and MoCA values compared to the group with SPPB ≥ 8 points [MoCA ≤ 26 points (23.6 ± 2.7 vs. 26.2 ± 2.5, p < 0.0001) and MMSE ≤ 24 points (19.8 ± 5.9 vs. 25.9 ± 4.5, p < 0.0001)]. This can be explained by the presence of common risk factors for both conditions, such as advanced age, comorbidities, and inflammatory status. Literature has revealed data on the association between AF and functional limitation in the elderly, independent of comorbidities and frailty markers. Indeed, a prospective cohort study evaluated the association between AF and the risk of reduced functional ability, assessed with ADL, in elderly patients, independent of the presence of stroke, heart failure, and other comorbidities. The same study also discussed how AF, independently of stroke, can exert degenerative effects on the Central nervous system through hypoperfusion and transient ischemic attacks (TIAs) [20]. Silent strokes and white matter disease are associated with gait and balance abnormalities [21], disability, and a higher risk of falls. Several cross-sectional studies have observed associations between AF and the presence of white matter abnormalities or TIAs [22]. In another study, physical function was assessed with SPPB and its potential association with a composite endpoint consisting of ischemic heart disease, stroke, and heart failure, as well as with individual cardiovascular endpoints, was assessed. It was found that reduced physical performance was an independent predictor of the evaluated cardiovascular outcomes, both as a composite endpoint and individually. This study included 5,570 participants with a mean age of 75 ± 5 years, of whom 58% were female, regardless of prior CVD history. In the Atherosclerosis Risk in Communities (ARIC) study, the population was divided into three groups according to SPPB value: low (0–6 points), intermediate (7–9 points), and high (10–12 points). During a median follow-up of 7 years (IQR range 5.3–7.8 years), 930 composite cardiovascular events occurred. Multivariate Cox regression model showed that the risk of achieving a composite endpoint was 47% higher in the low SPPB group compared to the high SPPB group (HR 1.47, 95% CI 1.20–1.79), and 25% in the intermediate group compared to the low SPPB group, a result that persisted even after adjusting for potential confounders. This study confirms the important prognostic role of functional limitation, assessed with SPPB, as an independent risk factor for composite cardiovascular endpoints in elderly population, regardless of other traditional risk factors [23]. However, the average age of the population in this study was lower than in our study (75 vs. 84 years), and they were outpatients, not all suffering from AF and with a lower burden of comorbidities than in our study.

Of interest, another study showed that after a diagnosis of AF, regardless the presence of stroke, there is a loss of functional autonomy with an increased likelihood of needing assistance, increasing the risk of disability [24].

Our work is also in agreement with previous studies that have demonstrated how reduced functional ability, assessed with ADL < 3, in elderly patients diagnosed with AF and treated with DOACs, is pathologically associated with the risk of CVEs. Therefore, in the hospitalized elderly population with AF, it seems important to evaluate not only the classic risk factors but also the possible presence of functional limitation assessed using SPPB. Indeed, in our work, it appears that a pathological SPPB increases the risk of CVEs by 2.78 times during a follow-up of 3.2 years. Additionally, an increased risk is associated with a treatment with VKAs before DOACs treatment, and the presence of a pathological GDS; while statin treatment, male gender, and RAASi treatment were found to be protective [25]. Furthermore, our study shows that patients with an SPPB score ≥ 8pt exhibited better scores on geriatric scales than the SPPB group < 8 pt, such as GDS 2.9 ± 3. 0 vs 7.3 ± 4.3 pt, p < 0.0001; ADL 5.3 ± 1.1 vs 2.8 ± 1.4 pt, p < 0.0001; and IADL 6.8 ± 1.6 vs 3.4 ± 1.8 pt, p < 0.0001; demonstrating the importance of this test in the elderly population. This study has several limitations; it is an observational study and not a randomized controlled clinical trial. Furthermore, the ratio between the two groups is not 1:1, which represents an intrinsic limitation of the study. Finally, we cannot exclude residual confounding factors not addressed in this study. However, this study has several strengths: it is a study with a large sample size consisting of elderly patients with several comorbidities often excluded from randomized clinical trials. Finally, the long follow-up represents another strength of the study.