Publish in this journal
Journal Information
Visits
Not available
Original article
Full text access
Available online 2 April 2024
Dynamic assessment of CHA2DS2-VASc and HAS-BLED scores for predicting ischemic stroke and major bleeding in atrial fibrillation patients
Evaluación dinámica de las escalas CHA2DS2-VASc y HAS-BLED para predecir ictus isquémico y hemorragia mayor en pacientes con fibrilación auricular
Visits
857
María José Sernaa,1, José Miguel Rivera-Caravacab,c,d,e,1,
Corresponding author
josemiguel.rivera@um.es

Corresponding author.
, Raquel López-Gálvezc,d, Eva Soler-Espejoa, Gregory Y.H. Lipe,f,2, Francisco Marínc,d,2, Vanessa Roldána,2
a Servicio de Hematología y Oncología Médica, Hospital General Universitario Morales Meseguer, Universidad de Murcia, Instituto Murciano de Investigación Biosanitaria (IMIB-Arrixaca), Murcia, Spain
b Facultad de Enfermería, Universidad de Murcia, Murcia, Spain
c Servicio de Cardiología, Hospital Clínico Universitario Virgen de la Arrixaca, Universidad de Murcia, Instituto Murciano de Investigación Biosanitaria (IMIB-Arrixaca), El Palmar, Murcia, Spain
d Centro de Investigación Biomédica en Red de Enfermedades Cardiovasculares (CIBERCV), Spain
e Liverpool Centre for Cardiovascular Science, University of Liverpool and Liverpool Heart and Chest Hospital, Liverpool, United Kingdom
f Department of Clinical Medicine, Aalborg Thrombosis Research Unit, Aalborg University, Denmark
This item has received
Article information
Abstract
Full Text
Bibliography
Download PDF
Statistics
Figures (4)
Show moreShow less
Tables (4)
Table 1. Baseline clinical characteristics
Table 2. C-indexes, c-indexes comparison, IDI and NRI of the dynamic CHA2DS2-VASc compared with the original score
Table 3. C-indexes, c-indexes comparison, IDI and NRI of the dynamic HAS-BLED compared with the original score
Table 4. Net benefits for baseline, 2-year, and 4-year CHA2DS2-VASc/HAS-BLED at different threshold probabilities
Show moreShow less
Additional material (1)
Abstract
Introduction and objectives

Stroke and bleeding risks in atrial fibrillation (AF) are often assessed at baseline to predict outcomes years later. We investigated whether dynamic changes in CHA2DS2-VASc and HAS-BLED scores over time modify risk prediction.

Methods

We included patients with AF who were stable while taking vitamin K antagonists. During a 6-year follow-up, all ischemic strokes/transient ischemic attacks (TIAs) and major bleeding events were recorded. CHA2DS2-VASc and HAS-BLED were recalculated every 2-years and tested for clinical outcomes at 2-year periods.

Results

We included 1361 patients (mean CHA2DS2-VASc and HAS-BLED 4.0±1.7 and 2.9±1.2). During the follow-up, 156 (11.5%) patients had an ischemic stroke/TIA and 269 (19.8%) had a major bleeding event. Compared with the baseline CHA2DS2-VASc, the CHA2DS2-VASc recalculated at 2 years had higher predictive ability for ischemic stroke/TIA during the period from 2 to 4 years. Integrated discrimination improvement (IDI) and net reclassification improvement (NRI) showed improvements in sensitivity and better reclassification. The CHA2DS2-VASc recalculated at 4 years had better predictive performance than the baseline CHA2DS2-VASc during the period from 4 to 6 years, with an improvement in IDI and an enhancement of the reclassification. The recalculated HAS-BLED at 2-years had higher predictive ability than the baseline score for major bleeding during the period from 2 to 4 years, with significant improvements in sensitivity and reclassification. A slight enhancement in sensitivity was observed with the HAS-BLED score recalculated at 4 years compared with the baseline score.

Conclusions

In AF patients, stroke and bleeding risks are dynamic and change over time. The CHA2DS2-VASc and HAS-BLED scores should be regularly reassessed, particularly for accurate stroke risk prediction.

Keywords:
Atrial fibrillation
Stroke
Hemorrhage
CHA2DS2-VASc
HAS-BLED
Risk scores
Risk stratification
Abbreviations:
AF
IDI
NRI
OAC
ROC
TIA
Resumen
Introducción y objetivos

A menudo la evaluación de los riesgos de ictus y hemorragia en la fibrilación auricular (FA) es basal para predecir los resultados años después. Sin embargo, estos riesgos no son estáticos. Se investiga si los cambios dinámicos en CHA2DS2-VASc y HAS-BLED a lo largo del tiempo modifican la predicción del riesgo.

Métodos

Se incluyó a pacientes con FA estables en tratamiento con antagonistas de la vitamina K. Durante 6 años de seguimiento, se registraron todos los ictus isquémicos/accidentes isquémicos transitorios (AIT) y hemorragias mayores. El CHA2DS2-VASc y HAS-BLED se revaluaron cada 2 años y se investigaron los resultados clínicos en periodos de 2 años.

Resultados

Se incluyó a 1.361 pacientes (medias de CHA2DS2-VASc y HAS-BLED, 4,0±1,7 y 2,9±1,2). Durante el seguimiento, 156 pacientes (11,5%) sufrieron un ictus isquémico/AIT y 269 (19,8%), una hemorragia mayor. En comparación con el valor basal, el CHA2DS2-VASc recalculado a los 2 años presentó mayor capacidad predictiva de ictus isquémico/AIT durante el periodo de 2-4 años. El índice de mejoría de la discriminación (IDI) y el índice de reclasificación neta (NRI) mostraron mejoras en la sensibilidad y mejor reclasificación. El CHA2DS2-VASc recalculado a los 4 años arrojó un mejor rendimiento predictivo que el basal durante el periodo de 4-6 años, con una mejora en el IDI y una mejora de la reclasificación. El HAS-BLED recalculado a los 2 años presentó mayor capacidad predictiva de hemorragia mayor que el basal durante el período de 2-4 años, con mejoras significativas en la sensibilidad y la reclasificación. Se observó un ligero aumento en la sensibilidad del HAS-BLED recalculado a los 4 años respecto al basal.

Conclusiones

En pacientes con FA, los riesgos de ictus y hemorragia son dinámicos y cambian con el tiempo. Las escalas CHA2DS2-VASc y HAS-BLED deben revaluarse con regularidad, especialmente para una precisa predicción del riesgo de ictus.

Palabras clave:
Fibrilación auricular
Ictus
Hemorragia
CHA2DS2-VASc
HAS-BLED
Escalas de riesgo
Estratificación del riesgo
Full Text
INTRODUCTION

Atrial fibrillation (AF) is the most common arrhythmia and carries a significant risk of stroke and thromboembolism; hence, oral anticoagulation (OAC) is recommended in guidelines as a fundamental pillar for its appropriate management.1,2

Nonetheless, OAC use requires stroke risk stratification, which should also be balanced with bleeding risk assessment, using clinical risk scores such as the CHA2DS2-VASc and HAS-BLED scores,3,4 as recommended by most clinical practice guidelines. Unfortunately, such risk assessment is usually performed at baseline, as a one-off evaluation that views risk as static process, and these values are usually applied to predict clinical outcomes that occur many years later.

However, risk factors for stroke and bleeding are not static over time, but are rather dynamic in nature, and most AF patients will develop at least 1 new risk factor before presenting with a thromboembolic event.5 These dynamic changes may increase the scores of the CHA2DS2-VASc and HAS-BLED assessed initially, thus modifying the absolute risk (and rate) of stroke and bleeding. Hence, the baseline estimated risk of these outcomes may worsen, due to aging and other incident comorbidities.6,7

The aim of the present study was to investigate whether dynamic evaluation of CHA2DS2-VASc and HAS-BLED scores over time would improve risk prediction in a cohort of AF patients taking OAC therapy and who were prospectively enrolled in the Murcia AF Project.

METHODS

From May 1, 2007, to December 1, 2007, we consecutively included adult outpatients with permanent or paroxysmal AF who were stable on OAC therapy with vitamin K antagonists in the preceding 6 months (ie, INR from 2-3, so the time in therapeutic range at enrolment was 100%) attending our anticoagulation clinic. Patients with prosthetic valves or rheumatic AF were excluded, as well as those who had had an acute coronary syndrome, stroke, surgical interventions, hospitalizations, or any hemodynamic instability in the previous 6 months.

The study protocol was approved by the Ethics Committee of Hospital General Universitario Morales Meseguer and was performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki. All patients gave informed consent for participation.

CHA2DS2-VASc and HAS-BLED scores

At inclusion, a complete clinical history was recorded. The baseline CHA2DS2-VASc and HAS-BLED scores were calculated for each patient at study entry according to their original definitions (appendix A of the supplementary data).3,4 Based on the characteristics of the cohort at entry (see inclusion criteria), the baseline labile INR criterion was quantified as 0 in all patients.

During the follow-up, all risk variables of the scores were reassessed again and the CHA2DS2-VASc and HAS-BLED were recalculated at the end of each 2-year period. Their predictive abilities for clinical outcomes were tested against the baseline scores in 2-year intervals (from baseline to year 2, from year 2 to 4, and from year 4 to 6). Patients were censored if they died between periods of comparisons. For example, if a patient died at 3 years, data were used to compare the baseline vs recalculated scores at 2 years, but not for the following comparisons since their comorbidities could not be obtained again at 4 years. There were no missing data on clinical variables used to calculate scores.

Follow-up and clinical outcomes

Follow-up was performed according to the standard of care at each routine visit to the outpatient anticoagulation clinic or visits for anticoagulation control. If the patient never attended these visits, medical records and telephone calls were used to obtain the information needed and vital status, with no specific interventions and no specific visits for study purposes. Follow-up was extended for 6 years. During this period, all adverse events were recorded. No patient was lost to follow-up.

For the present study, the primary endpoints were ischemic strokes/transient ischemic attacks (TIAs) and major bleeding. Definitions of the primary endpoints are detailed in appendix B of the supplementary data. The investigators identified, confirmed, and recorded all clinical outcomes.

Statistical analysis

Quantitative variables are presented as mean±standard deviation or median and interquartile range [IQR], as appropriate after testing for normality by the Kolmogorov-Smirnov test. Categorical variables are expressed as frequencies and percentages. The Pearson chi-square test was used to compare proportions. The Student t or Mann-Whitney U tests were used to compare continuous and categorical variables, as appropriate.

The proportion of patients moving among different risk categories was visualized in an alluvial plot. The predictive ability (expressed as c-indexes) of baseline and dynamic CHA2DS2-VASc and HAS-BLED scores was assessed by receiver operating characteristics (ROC) curves, using the quantitative version of CHA2DS2-VASc and HAS-BLED in all cases. To contrast prognostic accuracy, we compared ROC curves using the method of DeLong et al.8 The net reclassification improvement (NRI) and integrated discriminatory improvement (IDI) for the baseline against the dynamic scores were also calculated as described by Pencina et al.9

We estimated the clinical usefulness and the net benefit of the original (baseline) scores in comparison with the dynamic scores by using decision curve analysis (DCA), as proposed by Vickers et al.10 The DCA shows the clinical usefulness of each new model based on a continuum of potential thresholds for adverse events (x-axis) and the net benefit of using the model to stratify patients at risk (y-axis) relative to assuming that no patient will have an adverse event. In this study, the prediction models are represented by color lines. The farther the prediction models from the dashed black line (ie, assume all adverse events) and the horizontal black line (ie, assume no adverse events), the higher the net clinical benefit.

A P value <.05 was accepted as statistically significant. Statistical analyses were performed using SPSS v. 25.0 (SPSS, United States), Origin v. 2022, (OriginLab Corporation, United States), MedCalc v. 16.4.3 (MedCalc Software bvba, Belgium), STATA v. 16.0 (Stata Corp, College Station, United States), and PredictABEL package for R v. 4.1.2 for Windows.

RESULTS

This study included 1361 patients, of which 693 (50.9%) were women, with a median age of 76 [IQR 71-81] years. A summary of other baseline characteristics is shown in table 1.

Table 1.

Baseline clinical characteristics

  N=1361 
Demographic data
Male sex  663 (48.7) 
Age, y  76 [71-81] 
Comorbidities
Hypertension  1116 (82.0) 
Diabetes mellitus  363 (26.7) 
Heart failure  429 (31.5) 
History of stroke/TIA/thromboembolism  267 (19.6) 
Renal impairment  144 (10.6) 
Coronary artery disease  255 (18.7) 
Hypercholesterolemia  443 (32.5) 
Current smoking habit  210 (15.4) 
Current alcohol consumption  50 (3.7) 
History of previous bleeding  113 (8.3) 
Concomitant malignant disease  105 (7.7) 
Concomitant treatment
Amiodarone  77 (5.7) 
Digoxin  272 (20.0) 
Calcium antagonist  339 (25.0) 
Beta-blockers  470 (34.5) 
Statins  331 (24.3) 
Diuretics  614 (45.1) 
Antiplatelet therapy  243 (18.0) 
ACE inhibitors/ARBs  717 (52.7) 

ACE inhibitors, angiotensin-converting-enzyme inhibitors; ARBs, angiotensin II receptor blockers; TIA, transient ischemic attack.

Data are expressed as No. (%), or median [interquartile range].

The mean baseline CHA2DS2-VASc was 4.0±1.7, whereas the mean CHA2DS2-VASc scores at 2 years and at 4 years were 4.3±1.6 and 4.5±1.6, respectively. Compared with baseline CHA2DS2-VASc, CHA2DS2-VASc at 2 years and at 4 years were significantly higher (both P <.001). CHA2DS2-VASc at 4 years was also significantly higher than CHA2DS2-VASc at 2 years (P <.001). Accordingly, the proportions of participants categorized as low, intermediate or high risk by CHA2DS2-VASc score at baseline were 2.1% (28), 4.8% (65) and 93.2% (1268), respectively. Of the 1062 patients who were alive at 4 years, 8 (0.8%) were categorized as low risk, 24 (2.3%) as intermediate risk and 1030 (97.0%) as high risk by recalculated CHA2DS2-VASc score (figure 1).

Figure 1.

Alluvial plots showing baseline stroke and bleeding risk stratification (baseline CHA2DS2-VASc and HAS-BLED), and reclassification into different risk categories during follow-up. Green, low risk of stroke or bleeding; orange, moderate risk of stroke or bleeding; red, high risk of stroke or bleeding.

(0.26MB).

The mean HAS-BLED score was 2.7±1.2 at baseline, 2.9±1.3 at 2 years and 3.1±1.2 at 4 years. Compared with baseline HAS-BLED, HAS-BLED at 2 years and at 4 years were significantly higher (both P <.001). HAS-BLED at 4 years was also significantly higher than HAS-BLED at 2 years (P <.001). At baseline, the proportions of participants categorized as low, moderate or high risk by HAS-BLED score were 10.4% (n=141), 28.4% (n=387) and 61.2% (n=833), respectively. Of the 1062 patients who were alive at 4 years, 68 (6.4%) were categorized as low risk, 318 (29.9%) as moderate risk, and 676 (63.7%) as high risk by the recalculated HAS-BLED score (figure 1).

Primary outcomes and predictive abilities

During 6 years of follow-up, 156 (11.5%) patients had an ischemic stroke/TIA and 269 (19.8%) had a major bleeding event. In addition, 472 (34.68%) patients died. The period with the highest incidence rates for ischemic stroke/TIA and major bleeding was from 2 to 4 years, which showed a significantly higher incidence rate ratio compared with the others for major bleeding outcomes. Complementary information on clinical outcomes and incidences among the different periods is shown in appendix C and table 1 of the supplementary data.

The probability of experiencing an ischemic stroke/TIA increased with the use of the recalculated CHA2DS2-VASc, with the 4-year assessment showing the highest higher hazard ratio even after adjustment (table 2 of the supplementary data). Regarding the predictive performance for ischemic stroke/TIA, the CHA2DS2-VASc recalculated at 2 years had significantly higher predictive ability during the period from 2 to 4 years compared with the baseline CHA2DS2-VASc, whereas IDI and NRI showed improvements in sensitivity and reclassification (table 2). Similarly, the CHA2DS2-VASc recalculated at 4 years yielded significantly better predictive performance for ischemic stroke/TIA during the period from 4 to 6 years compared with the baseline CHA2DS2-VASc (figure 2). Again, IDI reported improvement in sensitivity and there was an enhancement of the reclassification ability based on NRI (table 2).

Table 2.

C-indexes, c-indexes comparison, IDI and NRI of the dynamic CHA2DS2-VASc compared with the original score

Risk score assessment  C-index  95%CI  Z score*  P*  IDI  95%CI  P  NRI  95%CI  P 
Ischemic stroke/TIA at 4 y
CHA2DS2-VASc at baseline  0.604  0.576-0.631 
CHA2DS2-VASc at 2 y  0.701  0.675-0.727  3.628  <.001  0.014  0.007/0.020  <.001  0.677  0.427/0.926  <.001 
Ischemic stroke/TIA at 6 y
CHA2DS2-VASc at baseline  0.682  0.653-0.710 
CHA2DS2-VASc at 2 y  0.670  0.640-0.697  0.889  .374  0.002  -0.001/0.004  .211  0.209  -0.092/0.511  .173 
CHA2DS2-VASc at 4 y  0.761  0.734-0.786  2.234  .026  0.030  0.016/0.044  <.001  0.757  0.496/1.018  <.001 

95% CI, 95% confidence interval; IDI, integrated discrimination improvement; NRI, net reclassification improvement.

*

For c-index comparison.

Figure 2.

Receiver operating characteristic curves of the baseline and dynamic (at 2 and 4 years) CHA2DS2-VASc/HAS-BLED scores for the prediction of ischemic stroke/TIA or major bleeding. Blue line, baseline scores; green line, scores recalculated at 2-years; red line, scores recalculated at 4-years.

(0.13MB).

As for stroke, the probability of major bleeding was higher with the use of the recalculated HAS-BLED score (table 3 of the supplementary data). At 2 years, the recalculated HAS-BLED score also showed a significantly higher predictive ability than the baseline HAS-BLED score for major bleeding events during the period from 2 to 4 years. IDI and NRI demonstrated significant improvements compared with the baseline HAS-BLED score (table 3). For major bleeds from 4 to 6 years, the c-index of the HAS-BLED score recalculated at 4 years was numerically higher but not statistically significant compared with baseline. There was a slight enhancement in sensitivity as assessed by the IDI but was not significantly better when reclassified by NRI (table 3).

Table 3.

C-indexes, c-indexes comparison, IDI and NRI of the dynamic HAS-BLED compared with the original score

Risk score assessment  C-index  95%CI  Z score*  P*  IDI  95%CI  P  NRI  95%CI  P 
Major bleeding at 4 y
HAS-BLED at baseline  0.663  0.632-0.693 
vs HAS-BLED at 2 y  0.709  0.680-0.738  2.987  .003  0.016  0.006/0.026  .001  0.444  0.287/0.600  <.001 
Major bleeding at 6 y
HAS-BLED at baseline  0.623  0.593-0.652 
vs HAS-BLED at 2 y  0.613  0.582-0.642  0.467  .640  -0.001  -0.003/0.001  .419  -0.063  -0.246/0.119  .498 
vs HAS-BLED at 4-years  0.631  0.601-0.660  0.318  .751  0.009  0.001/0.016  .018  -0.002  -0.174/0.170  .977 

95% CI, 95% confidence interval; IDI, integrated discrimination improvement; NRI, net reclassification improvement.

*

For c-index comparison.

DCAs would help in the estimation of patients who will experience any of the primary endpoints, based on the predictions of the baseline risk scores in comparison with those recalculated at 2 and 4 years of follow-up. In figure 3, the prediction models that are the farthest away from the dashed black line (ie, assume all events) and the horizontal black line (ie, assume none event) had the highest net benefit. DCAs demonstrated that using the dynamic CHA2DS2-VASc and HAS-BLED scores was clinically useful and provided an overall improvement in the net benefit for the prediction of ischemic stroke/TIA and major bleeding, respectively. A detailed estimate of the net benefit at each threshold of probabilities from 0% to 20% is shown in table 4. A summary of the study findings is shown in figure 4.

Figure 3.

Decision curve analysis of the baseline and dynamic (at 2 and 4 years) CHA2DS2-VASc/HAS-BLED scores for ischemic stroke/TIA or major bleeding. Solid black line, assumes all patients will suffer an adverse event; dashed black line, assumes no patient will suffer an adverse event; blue line, baseline scores; green line, scores recalculated at 2-years; red line, scores recalculated at 4-years.

(0.18MB).
Table 4.

Net benefits for baseline, 2-year, and 4-year CHA2DS2-VASc/HAS-BLED at different threshold probabilities

Threshold probability  Ischemic stroke/TIAMajor bleeding
  Net Benefit for CHA2DS2-VASc  Net Benefit for HAS-BLED 
  Atbaseline  At 2 y  At 4 y  Difference in net benefit(baseline vs recalculated at 4 y)  Atbaseline  At 2 y  At 4 y  Difference in net benefit(baseline vs recalculated at 4 y) 
5%  0.13%  0.12%  1.32%  1.20%  2.00%  1.88%  2.20%  0.20% 
10%  0.00%  0.07%  0.23%  0.23%  0.36%  0.04%  0.15%  0.20% 
15%  0.00%  0.00%  0.07%  0.07%  0.01%  0.00%  0.01%  0.02% 
20%  0.00%  0.00%  0.05%  0.05%  0.00%  0.00%  0.24%  0.24% 
Overall improvement        1.41%        0.24% 
Figure 4.

Central illustration. TIA, transient ischemic attack.

(0.37MB).
DISCUSSION

In this real-world cohort study, our principal findings are as follows: a) consecutive reassessment of stroke and bleeding risks through CHA2DS2-VASc and HAS-BLED scores demonstrated a significantly higher predictive ability and net benefit compared with the baseline scores; and b) the risks of stroke and bleeding in AF patients are dynamic and change during follow-up.

The incidence of new risk factors and the temporal trends in the CHA2DS2-VASc score have been previously investigated. Chao et al.5 included 14 606 AF patients who did not receive antiplatelet agents or OACs, with a baseline CHA2DS2-VASc score of 0 (men) or 1 (women) and with incident risk factors, and observed a dynamic increase in the CHA2DS2-VASc score.5 Indeed, this often translated into a dynamic change in the risk category.6,11,12 A recent study demonstrated that changes in CHA2DS2-VASc score over time were associated with the incidence of stroke, thus indicating that stroke risk is not static.13 In fact, the CHA2DS2-VASc score of AF patients would increase as patients become older and they accumulate more comorbidities, and we observed a significant progressive increase in the CHA2DS2-VASc score in accordance with previously published literature,5,6,11 again emphasizing the dynamic nature of thromboembolic risk.

Although the concept or risk reassessment is well accepted.5,14,15 prior clinical guidelines did not provide clear and concise recommendations on this issue. The landscape has changed with the publication of the latest international guidelines, which emphasize the need for dynamic assessment, at least annually, of thromboembolic risk.1,2,16–19 However, the time interval in which thromboembolic risk should be reassessed is controversial. In a study published by Chao et al.,5 the CHA2DS2-VASc score in patients initially classified as low risk increased by approximately 12.1%/y. The authors therefore suggest that, among low-risk patients, stroke risk should be reassessed every 4 months, with the aim of prescribing OAC therapy in those with an increased CHA2DS2-VASc score.

Regarding bleeding risk, guidelines recommend the use of HAS-BLED for the evaluation of bleeding risk, suggesting its frequent reassessment with particular attention to modifiable bleeding risk factors in patients with a high bleeding risk (HAS-BLED ≥3).1,2,16,18–21 Although the dynamic nature of the variables associated with bleeding is well accepted, few studies have specifically investigated this. As with stroke risk, bleeding risk assessment is often done at baseline only, at the beginning of OAC therapy, while bleeding events can be observed many years later. This may reflect that bleeding risk assessment has been subject to misuse and misinterpretation, and modifiable bleeding risk factors should be addressed as part of a holistic approach to AF patient assessment and management.22 Bleeding risk stratification should be used to flag patients with a high risk of severe bleeding for a more careful and closer follow-up to manage modifiable factors and reduce the potential risk of a major bleeding event. This approach has been prospectively tested in the mAFA-II trial, where the intervention arm using HAS-BLED had a lower risk of major bleeding at 1 year and an increase in OAC use compared with usual care, which showed higher bleeding and a decline in OAC use.23

One study that focused on the dynamic nature of bleeding risk found that dynamic assessments of HAS-BLED had a better risk predictive ability than baseline assessment alone for the prediction of major bleeding.24 These results are similar to that observed in our study, where the predictive ability of the HAS-BLED score calculated at 2 years was significantly better than the baseline assessment. Although there were no significant differences between the dynamic HAS-BLED at 4 years and the baseline score, the dynamic HAS-BLED showed a slight improvement in sensitivity, and the overall clinical usefulness and net benefit of the dynamic HAS-BLED scores were still higher.

Overall, our results reinforce those of previous studies since the use of the dynamic CHA2DS2-VASc and HAS-BLED scores was associated with a higher net benefit and therefore with increased clinical usefulness than the baseline scores. Even though dynamic evaluation and reassessment of stroke and bleeding risks are currently widely accepted, many international clinical practice guidelines for AF patients do not as yet include clear and concise recommendations on how to perform and address this dynamic risk monitoring. The 2021 Asia Pacific Heart Rhythm Society AF guidelines provide recommendations on the dynamic nature of risk in patients with AF, whereby frequent reassessment of patients with AF is suggested, using the CHA2DS2-VASc and HAS-BLED scores.2,25

In our study, we show that changes in the overall score of CHA2DS2-VASc also corresponded to variations in the risk category of AF patients, showing an evolutionary increase in the proportion of patients classified within the high-risk group, to the detriment of a decrease in the proportion of low- and moderate-risk patients. For this reason, periodic reassessment of stroke risk is of particular interest in patients classified as low risk by a baseline CHA2DS2-VASc score (ie, 0 in men; 1 in women) given that OAC is not required in these patients but the progression of aging and new risk factors/comorbidities would change the overall CHA2DS2-VASc score, and therefore OAC may become indicated.15,26,27 Although there is special clinical interest in reassessment of the CHA2DS2-VASc score to reconsider the decision to initiate OAC in these patients initially classified as low-risk, it is also important to reevaluate the risk of stroke in high-risk AF patients already receiving OAC. Indeed, stroke rates can vary significantly between patients with a CHA2DS2-VASc from 3 to 9, even though all of them are in the same high-risk category, and not all stroke risk factors have the same impact.28 Additionally, several risk factors for stroke are also risk factors for bleeding, hence modifiable risk factors could be identified to be managed and reduced appropriately in high-risk patients.

Limitations

This study has some limitations. First, it is limited by its Caucasian-based population and single-center design. Second, although patients with prosthetic valves were excluded, no data were available on other valvular diseases that may have an impact on adverse events in these patients with AF. Third, during the previous 6 months after entry, all patients were stable with vitamin K antagonists (INR 2.0-3.0) and had no adverse events or hemodynamic instability to ensure homogeneity, since these factors may have an impact on baseline risk and subsequent clinical outcomes. As a result, our cohort may have a lower baseline thromboembolic and hemorrhagic risk. This may have interfered with the baseline CHA2DS2-VASc and HAS-BLED scores, underestimating their predictive ability in comparison with the dynamic estimation of risk. These strict selection criteria may not reflect typical clinical practice, but we believe that this initial homogenization of the population limits the possibility that certain variables that generate instability acted as confounding factors, and the long follow-up under standard care make this cohort suitable. Some ischemic strokes that occurred during follow-up may be caused by noncardioembolic reasons, and these have not been investigated in detail. However, participants were carefully followed up and all events (even very early events) were recorded.

We acknowledge that all together, these factors limit the generalizability of the findings to broader and more diverse populations, even those patients under direct-acting OACs, and that the cohort might not be representative of the broader population of AF patients, especially those with higher risk.

Although our dataset was collected prospectively, the baseline assessment and reassessment of risk scores were performed post hoc, which might introduce potential biases. Thus, our results should be interpreted with caution and as hypothesis-generating only. Finally, we were not able to explore the change (Delta’) CHA2DS2-VASc or the Delta HAS-BLED, given that the sample size is limited, and the various (relative short) periods of observation did not allow an adequate evaluation of this metric.

CONCLUSIONS

In AF patients, stroke and bleeding risks are dynamic and change over time with aging and incident comorbidities. The CHA2DS2-VASc and HAS-BLED scores (and clinical risk profiles) should be regularly reassessed, which is particularly necessary for appropriate stroke and bleeding risk prediction.

FUNDING

This work was supported by the Spanish Ministry of Economy, Industry, and Competitiveness, through the Instituto de Salud Carlos III after independent peer review (research grant: PI17/01375 cofinanced by the European Regional Development Fund) and group CB16/11/00385 from CIBERCV.

ETHICAL CONSIDERATIONS

The study protocol was approved by the Ethics Committee of Hospital General Universitario Morales Meseguer and was performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki. All patients gave informed consent for participation. The possible variables of sex and gender have been taken into account in this work in accordance with the SAGER guidelines. In fact, the study is balanced between male and female participants (48.7% men, 51.3% women).

STATEMENT ON THE USE OF ARTIFICIAL INTELLIGENCE

No artificial intelligence was used in the preparation of this work.

AUTHORS’ CONTRIBUTIONS

M.J. Serna, J.M. Rivera-Caravaca, and V. Roldán contributed to data collection, performed statistical analyses, and drafted the manuscript. E. Soler-Espejo and R. López-Gálvez critically revised the manuscript. G.Y.H. Lip and F. Marín conceived and supervised the study and critically revised the manuscript. All authors read and approved the final version of the manuscript.

CONFLICTS OF INTEREST

G.Y.H. Lip is a consultant and speaker for BMS/Pfizer, Boehringer Ingelheim and Daiichi-Sankyo. No fees are received personally.

J.M. Rivera-Caravaca is a consultant for Idorsia Pharmaceuticals LTD.

The remaining authors have nothing to disclose.

WHAT IS KNOWN ABOUT THE TOPIC?

  • -

    Atrial fibrillation (AF) is associated with a high risk of stroke and thromboembolism.

  • -

    Risk assessment in AF is usually performed at baseline, as a one-off evaluation considering risk as a static process.

  • -

    Stroke and bleeding risk are dynamic, which may increase the initial CHA2DS2-VASc and HAS-BLED scores.

WHAT DOES THIS STUDY ADD?

  • -

    Both the CHA2DS2-VASc and HAS-BLED scores were significantly higher at 2 and 4 years.

  • -

    The CHA2DS2-VASc score recalculated at 2 and 4 years had significantly higher predictive ability than the baseline score for ischemic stroke/TIA during the periods from 2 to 4 years and from 4 to 6 years.

  • -

    The HAS-BLED recalculated at 2-years showed significantly higher predictive ability than the baseline score for major bleeding during the period from 2 to 4-years.

  • -

    The dynamic CHA2DS2-VASc and HAS-BLED scores were clinically useful and provided an overall improvement in the net benefit for the prediction of ischemic stroke/TIA and major bleeding.

  • -

    The CHA2DS2-VASc and HAS-BLED scores should be regularly reassessed.

APPENDIX
SUPPLEMENTARY DATA

Supplementary data associated with this article can be found in the online version, at https://doi.org/10.1016/j.rec.2024.02.011

References
[1]
G. Hindricks, T. Potpara, N. Dagres, 2020 ESC Guidelines for the diagnosis and management of atrial fibrillation developed in collaboration with the European Association of Cardio-Thoracic Surgery (EACTS), et al.
Eur Heart J., 42 (2021), pp. 373-498
[2]
T.F. Chao, B. Joung, Y. Takahashi, et al.
2021 Focused Update Consensus Guidelines of the Asia Pacific Heart Rhythm Society on Stroke Prevention in Atrial Fibrillation: Executive Summary.
Thromb Haemost., 122 (2022), pp. 20-47
[3]
G.Y. Lip, R. Nieuwlaat, R. Pisters, D.A. Lane, H.J. Crijns.
Refining clinical risk stratification for predicting stroke and thromboembolism in atrial fibrillation using a novel risk factor-based approach: the euro heart survey on atrial fibrillation.
Chest., 137 (2010), pp. 263-272
[4]
R. Pisters, D.A. Lane, R. Nieuwlaat, C.B. de Vos, H.J. Crijns, G.Y. Lip.
A novel user-friendly score (HAS-BLED) to assess 1-year risk of major bleeding in patients with atrial fibrillation: the Euro Heart Survey.
Chest., 138 (2010), pp. 1093-1100
[5]
T.F. Chao, J.N. Liao, T.C. Tuan, et al.
Incident Co-Morbidities in Patients with Atrial Fibrillation Initially with a CHA2DS2-VASc Score of 0 (Males) or 1 (Females): Implications for Reassessment of Stroke Risk in Initially ‘Low-Risk’ Patients.
Thromb Haemost., 119 (2019), pp. 1162-1170
[6]
T.F. Chao, G.Y.H. Lip, C.J. Liu, et al.
Relationship of Aging and Incident Comorbidities to Stroke Risk in Patients With Atrial Fibrillation.
J Am Coll Cardiol., 71 (2018), pp. 122-132
[7]
L. Fauchier, A. Bodin, A. Bisson, et al.
Incident Comorbidities Aging and the Risk of Stroke in 608,108 Patients with Atrial Fibrillation: A Nationwide Analysis.
J Clin Med., 9 (2020), pp. 1234
[8]
E.R. DeLong, D.M. DeLong, D.L. Clarke-Pearson.
Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach.
Biometrics., 44 (1988), pp. 837-845
[9]
M.J. Pencina, R.B. D’Agostino Sr., R.B. D’Agostino Jr., R.S. Vasan.
Evaluating the added predictive ability of a new marker: from area under the ROC curve to reclassification and beyond.
Stat Med., 27 (2008), pp. 157-1572
[10]
A.J. Vickers, E.B. Elkin.
Decision curve analysis: a novel method for evaluating prediction models.
Med Decis Making., 26 (2006), pp. 565-574
[11]
M. Yoon, P.S. Yang, E. Jang, et al.
Dynamic Changes of CHA2DS2-VASc Score and the Risk of Ischaemic Stroke in Asian Patients with Atrial Fibrillation: A Nationwide Cohort Study.
Thromb Haemost., 118 (2018), pp. 1296-1304
[12]
T.F. Chao, C.E. Chiang, T.J. Chen, G.Y.H. Lip, S.A. Chen.
Reassessment of Risk for Stroke During Follow-up of Patients With Atrial Fibrillation.
Ann Intern Med., 170 (2019), pp. 663-664
[13]
E. Tsiartas, A. Samaras, A.S. Papazoglou, et al.
Changes in CHA(2)DS(2)-VASc score and risk of ischemic stroke among patients with atrial fibrillation.
Heart Vessels., 38 (2023), pp. 1267-1276
[14]
M. Proietti, N. Mujovic, T.S. Potpara.
Optimizing Stroke and Bleeding Risk Assessment in Patients with Atrial Fibrillation: A Balance of Evidence, Practicality and Precision.
Thromb Haemost., 118 (2018), pp. 2014-2017
[15]
M. Domek, J. Gumprecht, M. Mazurek, T.F. Chao, G.Y.H. Lip.
Should We Judge Stroke Risk by Static or Dynamic Risk Scores?. A Focus on the Dynamic Nature of Stroke and Bleeding Risks in Patients With Atrial Fibrillation.
J Cardiovasc Pharmacol., 74 (2019), pp. 491-498
[16]
J.G. Andrade, M. Aguilar, C. Atzema, et al.
The 2020 Canadian Cardiovascular Society/Canadian Heart Rhythm Society Comprehensive Guidelines for the Management of Atrial Fibrillation.
Can J Cardiol., 36 (2020), pp. 1847-1948
[17]
C.T. January, L.S. Wann, H. Calkins, et al.
2019 AHA/ACC/HRS Focused Update of the 2014 AHA/ACC/HRS Guideline for the Management of Patients With Atrial Fibrillation: A Report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines and the Heart Rhythm Society.
J Am Coll Cardiol., 74 (2019), pp. 104-132
[18]
B. Joung, J.M. Lee, K.H. Lee, et al.
2018 Korean Guideline of Atrial Fibrillation Management.
Korean Circ J., 48 (2018), pp. 1033-1080
[19]
G.Y.H. Lip, A. Banerjee, G. Boriani, et al.
Antithrombotic Therapy for Atrial Fibrillation: CHEST Guideline and Expert Panel Report.
Chest., 154 (2018), pp. 1121-1201
[20]
C.E. Chiang, T.J. Wu, K.C. Ueng, et al.
2016 Guidelines of the Taiwan Heart Rhythm Society and the Taiwan Society of Cardiology for the management of atrial fibrillation.
J Formos Med Assoc., 115 (2016), pp. 893-952
[21]
D.A. Gorog, Y.X. Gue, T.F. Chao, et al.
Assessment and Mitigation of Bleeding Risk in Atrial Fibrillation and Venous Thromboembolism: Executive Summary of a European and Asia-Pacific Expert Consensus Paper.
Thromb Haemost., 122 (2022), pp. 1625-1652
[22]
D.A. Lane, G.Y.H. Lip.
Stroke and bleeding risk stratification in atrial fibrillation: a critical appraisal.
Eur Heart J Suppl., 22 (2020), pp. O14-O27
[23]
Y. Guo, D.A. Lane, Y. Chen, G.Y.H. Lip.
Regular Bleeding Risk Assessment Associated with Reduction in Bleeding Outcomes: The mAFA-II Randomized Trial.
Am J Med., 133 (2020), pp. 1195-1202
[24]
T.F. Chao, G.Y.H. Lip, Y.J. Lin, et al.
Incident Risk Factors and Major Bleeding in Patients with Atrial Fibrillation Treated with Oral Anticoagulants: A Comparison of Baseline, Follow-up and Delta HAS-BLED Scores with an Approach Focused on Modifiable Bleeding Risk Factors.
Thromb Haemost., 118 (2018), pp. 768-777
[25]
T.F. Chao, B. Joung, Y. Takahashi, et al.
2021 Focused update of the 2017 consensus guidelines of the Asia Pacific Heart Rhythm Society (APHRS) on stroke prevention in atrial fibrillation.
J Arrhythm., 37 (2021), pp. 1389-1426
[26]
T.Y. Chang, G.Y.H. Lip, S.A. Chen, T.F. Chao.
Importance of Risk Reassessment in Patients With Atrial Fibrillation in Guidelines: Assessing Risk as a Dynamic Process.
Can J Cardiol., 35 (2019), pp. 611-618
[27]
T.F. Chao, S.A. Chen.
Using the scoring schemes in the right way: the dynamic assessment of stroke and bleeding risk in patients with atrial fibrillation.
J Thorac Dis., 10 (2018), pp. S2089-S2091
[28]
S. Zweibel, R. Passman, S. Sarkar, J. Koehler, P. Ziegler.
Are all CHA2DS2-VASc risk factors created equal? An assessment of stroke risk among 34,470 patients with CHA2DS2-VASc scores of 1 or 2.
Eur Heart J., 41 (2020),

Joint first authors.

Joint senior authors.

Copyright © 2024. Sociedad Española de Cardiología
Idiomas
Revista Española de Cardiología (English Edition)

Subscribe to our newsletter

View newsletter history
Article options
Tools
Supplemental materials
es en

¿Es usted profesional sanitario apto para prescribir o dispensar medicamentos?

Are you a health professional able to prescribe or dispense drugs?