Atrial Fibrillation Risk Prediction Research Articles

Association between NT-proBNP levels and risk of atrial fibrillation: a systematic review and meta-analysis of cohort studies

Background and aimsN-terminal pro-B-type natriuretic peptide (NT-proBNP) is a well-established biomarker in clinical practice, particularly for heart failure, but its role in predicting atrial fibrillation (AF) risk is not fully...

Chinese Guidelines for the Diagnosis and Management of Atrial Fibrillation

Atrial fibrillation (AF) is the most common sustained cardiac arrhythmia, significantly impacting patients’ quality of life and increasing the risk of death, stroke, heart failure, and dementia. Over the past 2 decades, there have been significant breakthroughs in AF risk prediction and screening, stroke prevention, rhythm control, catheter ablation, and integrated management. During this period, the scale, quality, and experience of AF management in China have greatly improved, providing a solid foundation for the development of guidelines for the diagnosis and management of AF. To further promote standardized AF management, and apply new technologies and concepts to clinical practice in a timely and comprehensive manner, the Chinese Society of Cardiology of the Chinese Medical Association and the Heart Rhythm Committee of the Chinese Society of Biomedical Engineering have jointly developed the Chinese Guidelines for the Diagnosis and Management of Atrial Fibrillation. The guidelines have comprehensively elaborated on various aspects of AF management and proposed the CHA2DS2-VASc-60 stroke risk score based on the characteristics of AF in the Asian population. The guidelines have also reevaluated the clinical application of AF screening, emphasized the significance of early rhythm control, and highlighted the central role of catheter ablation in rhythm control.

Predicting atrial fibrillation risk following embolic stroke of unknown source using cardiac magnetic resonance imaging

Abstract Background Approximately 25% of embolic strokes of unknown source (ESUS) are attributed to undiagnosed atrial fibrillation (AF).1 The European Stroke Organisation (ESO) recommends implantable loop recorder (ILR) monitoring to identify ESUS patients with AF, permit the initiation of anticoagulation and reduce risk of recurrent stroke.2 However, this approach delays treatment and carries significant cost and resource implications. An alternative approach, where early anticoagulation is initiated in the absence of AF has failed to reduce recurrent stroke in four randomised controlled trials (NAVIGATE ESUS, RE-SPECT ESUS, ATTICUS and ARCADIA).3-6 An urgent need exists for new strategies to identify ESUS patients with undiagnosed AF that are at high risk of future thromboembolism. Purpose The primary objective of this study was to produce a substrate-based predictive model for AF in ESUS patients using atrial cardiac magnetic resonance imaging (CMRi). Method The Atrial CARdiac Magnetic resonance imaging in patients with embolic stroke of unknown source without documented AF (CARM-AF) Study is a prospective, multi-centre, observational study. All patients underwent CMRi and ILR insertion for AF detection within 3 months of ESUS. The main inclusion criteria were expected survival >12 months, CHA2DS2VASc≥3 and eGFR >30ml/min. Patients were allocated to the study group if AF >30 seconds was detected by ILR during the first year of follow-up and control group if no AF detected (Figure 1). Univariable analysis was used to identify clinical, echocardiographic and CMR parameters associated with AF diagnosis. Parameters with p<0.2 and no co-linearity were used to develop multivariable logistic regression models for AF prediction. Results From September 2020 to September 2022, 102 patients were enrolled and 91 were included in the final analysis. AF was detected in 17 patients during the first year of follow-up (18.6%). In univariable analysis, patients with AF were more likely to be female (p=0.02), but there were no significant differences in age, race, body mass index (BMI) or CHA2DS2VASc score (p=0.06, 0.272, 0.195, 0.211) between groups. Increased left atrial (LA) volume, surface area and reduced LA ejection fraction were associated with increased risk of AF detection at 1 year (p=0.006, 0.044, 0.008). LA fibrosis and sphericity were not associated with AF (p=0.84, 0.98). Logistic regression models for AF risk prediction were developed using patient, echocardiographic and CMR parameters. A combined model of patient and CMR parameters outperformed assessment of patient characteristics alone, achieving an AUC of 0.85 for predicting AF occurrence (p<0.005). Pseudo R2 of 0.30 (McFadden) indicated excellent model fit (Figure 2). Conclusion CARM-AF is the first study to demonstrate feasibility and utility of CMR imaging to predict AF risk following ESUS. A randomised controlled trial initiating anticoagulation using the developed model is recommended.Study DesignModel Comparison (ROC Curves)

The cardiac blood transcriptome predicts de novo onset of atrial fibrillation in heart failure

Heart failure (HF) increases the risk of developing atrial fibrillation (AF), leading to increased morbidity and mortality. Therefore, better prediction of this risk may improve treatment strategies. Although several predictors based on clinical data have been developed, the establishment of a transcriptome-based predictor of AF incidence in HF has proven to be more problematic. We hypothesized that the transcriptome profile of coronary sinus blood samples of HF patients is associated with AF incidence. We therefore enrolled 192 HF patients who were selected for biventricular cardioverter defibrillator implantation. Both coronary sinus and peripheral blood samples were obtained during the procedure. Patients were followed-up during two years and AF occurrence was based on interrogation of the defibrillator. A total of 96 patients stayed in sinus rhythm (SR) during follow-up, 13 patients developed AF within 1 year and 10 patients developed AF during the second year of follow up. Gene expression profiling of coronary sinus samples led to the identification of 321 AF predictor genes based on their differential expression between patients developing AF within 1 year of blood sampling and patients remaining in SR. The expression levels of these genes were combined to obtain a molecular atrial fibrillation prediction score for each patient which was significantly different between both patient groups (Mann-Whitney, p = 0.00018). We conclude that the cardiac blood transcriptome of HF patients should be further investigated as a potential AF risk prediction tool.

Chinese guidelines for the diagnosis and management of atrial fibrillation.

Atrial fibrillation (AF) is the most common sustained cardiac arrhythmia, significantly impacting patients' quality of life and increasing the risk of death, stroke, heart failure, and dementia. Over the past two decades, there have been significant breakthroughs in AF risk prediction and screening, stroke prevention, rhythm control, catheter ablation, and integrated management. During this period, the scale, quality, and experience of AF management in China have greatly improved, providing a solid foundation for the development of the guidelines for the diagnosis and management of AF. To further promote standardized AF management, and apply new technologies and concepts to clinical practice timely and fully, the Chinese Society of Cardiology of Chinese Medical Association and the Heart Rhythm Committee of Chinese Society of Biomedical Engineering jointly developed the Chinese Guidelines for the Diagnosis and Management of Atrial Fibrillation. The guidelines comprehensively elaborated on various aspects of AF management and proposed the CHA2DS2‑VASc‑60 stroke risk score based on the characteristics of the Asian AF population. The guidelines also reevaluated the clinical application of AF screening, emphasized the significance of early rhythm control, and highlighted the central role of catheter ablation in rhythm control.

Artificial-intelligence-based risk prediction and mechanism discovery for atrial fibrillation using heart beat-to-beat intervals

Early diagnosis of atrial fibrillation (AF) is important for preventing stroke and other complications. Predicting AF risk in advance can improve early diagnostic efficiency. Deep learning hasbeen used for disease risk prediction; however, it lacks adherenceto evidence-based medicine standards. Identifying the underlyingmechanisms behind disease risk prediction is important and required. We developed an explainable deep learning model called HBBI-AI to predict AF risk using only heart beat-to-beat intervals (HBBIs) during sinus rhythm. We proposed a possible AF mechanism based on the model's explainability and verified this conjecture using confirmed AF risk factors while also examining new AF risk factors. Finally, we investigated the changes in clinicians' ability to predict AF risk using only HBBIs before and after learning the model's explainability. HBBI-AI consistently performed well across large in-house and external public datasets. HBBIs with large changes or extreme stability were critical predictors for increased AF risk, and the underlying cause was autonomic imbalance. We verified various AF risk factors and discovered that autonomic imbalance was associated with all these factors. Finally, cardiologists effectively understood and learned from these findings to improve their abilities in AF risk prediction. HBBI-AI effectively predicted AF risk using only HBBI information through evaluating autonomic imbalance. Autonomic imbalance may play an important role in many risk factors of AF rather than in a limited number of risk factors. This study was supported in part by the National Key R&D Program and the National Natural Science Foundation of China.

Chinese Guidelines for the Diagnosis and Management of Atrial Fibrillation.

Atrial fibrillation (AF) is the most common sustained cardiac arrhythmia, significantly impacting patients' quality of life and increasing the risk of death, stroke, heart failure, and dementia. Over the past two decades, there have been significant breakthroughs in AF risk prediction and screening, stroke prevention, rhythm control, catheter ablation, and integrated management. During this period, the scale, quality, and experience of AF management in China have greatly improved, providing a solid foundation for the development of guidelines for the diagnosis and management of AF. To further promote standardized AF management, and apply new technologies and concepts to clinical practice in a timely and comprehensive manner, the Chinese Society of Cardiology of the Chinese Medical Association and the Heart Rhythm Committee of the Chinese Society of Biomedical Engineering have jointly developed the Chinese Guidelines for the Diagnosis and Management of Atrial Fibrillation. The guidelines have comprehensively elaborated on various aspects of AF management and proposed the CHA2DS2-VASc-60 stroke risk score based on the characteristics of AF in the Asian population. The guidelines have also reevaluated the clinical application of AF screening, emphasized the significance of early rhythm control, and highlighted the central role of catheter ablation in rhythm control.

Abstract 16426: A Risk Model to Predict Atrial Fibrillation in Diabetes Using Machine Learning: The ACCORD Study

Introduction: Atrial fibrillation (AF) is the most common cardiac arrhythmia seen in clinical practice, which necessitates identifying factors that can precipitate this dysrhythmia. We utilized the Action to Control Cardiovascular Risk in Diabetes (ACCORD) trial to build an AF risk prediction model specifically in the type 2 diabetes (T2DM) population, using machine learning (ML) approaches, and comparing to the established CHARGE-AF risk model. Methods: We utilized a patient cohort from the ACCORD clinical trial without prior AF. Baseline characteristics were collected as potential risk factors and incident AF was defined from ECG at follow-up. A random forest (RF) binary classifier was trained with 27 variables, including age, race, sex, smoking, waist circumference, BMI, blood pressure (BP), history of heart disease, medications, cholesterol, renal function, and left ventricular hypertrophy by ECG. All ACCORD subjects had T2DM, therefore this variable was not included. The CHARGE-AF model consisted of a Cox proportional hazard model trained on the following variables: age, race, height, weight, systolic/diastolic BP, current smoking, heart failure, prior MI, and BP medication. Area under ROC curve (AUC) was measured with randomized, stratified, 5-fold cross validation with 80/20 splits. Imputation and resampling methods were used for missing data and class imbalance. Results: Out of 9307 patients in the ACCORD cohort, 175 developed AF during the study period of 6.26 years, with a mean time to AF event of 3.68 years. The RF model, with an AUC of 0.912, showed improved prediction of AF over the Cox model, which had an AUC of 0.784. Permutation feature importance revealed age, race, waist circumference, weekly alcohol consumption, sex, and GFR as the most predictive features of the RF model (Fig 1). Conclusions: The CHARGE-AF risk model performed modestly on the ACCORD cohort, while a RF model showed superior performance.

End-to-end risk prediction of atrial fibrillation from the 12-Lead ECG by deep neural networks

BackgroundAtrial fibrillation (AF) is one of the most common cardiac arrhythmias that affects millions of people each year worldwide and it is closely linked to increased risk of cardiovas- cular diseases such as stroke and heart failure. Machine learning methods have shown promising results in evaluating the risk of developing atrial fibrillation from the electrocardiogram. We aim to develop and evaluate one such algorithm on a large CODE dataset collected in Brazil. MethodsWe used the CODE cohort to develop and test a model for AF risk prediction for individual patients from the raw ECG recordings without the use of additional digital biomarkers. The cohort is a collection of ECG recordings and annotations by the Telehealth Network of Minas Gerais, in Brazil. A convolutional neural network based on a residual network architecture was implemented to produce class probabilities for the classification of AF. The probabilities were used to develop a Cox proportional hazards model and a Kaplan-Meier model to carry out survival analysis. Hence, our model is able to perform risk prediction for the development of AF in patients without the condition. ResultsThe deep neural network model identified patients without indication of AF in the presented ECG but who will develop AF in the future with an AUC score of 0.845. From our survival model, we obtain that patients in the high-risk group (i.e. with the probability of a future AF case being >0.7) are 50% more likely to develop AF within 40 weeks, while patients belonging to the minimal-risk group (i.e. with the probability of a future AF case being less than or equal to 0.1) have >85% chance of remaining AF free up until after seven years. ConclusionWe developed and validated a model for AF risk prediction. If applied in clinical practice, the model possesses the potential of providing valuable and useful information in decision- making and patient management processes.

On Merging Feature Engineering and Deep Learning for Diagnosis, Risk Prediction and Age Estimation Based on the 12-Lead ECG.

Over the past few years, deep learning (DL) has been used extensively in research for 12-lead electrocardiogram (ECG) analysis. However, it is unclear whether the explicit or implicit claims made on DL superiority to the more classical feature engineering (FE) approaches, based on domain knowledge, hold. In addition, it remains unclear whether combining DL with FE may improve performance over a single modality. To address these research gaps and in-line with recent major experiments, we revisited three tasks: cardiac arrhythmia diagnosis (multiclass-multilabel classification), atrial fibrillation risk prediction (binary classification), and age estimation (regression). We used an overall dataset of 2.3M 12-lead ECG recordings to train the following models for each task: i) a random forest taking FE as input; ii) an end-to-end DL model; and iii) a merged model of FE+DL. FE yielded comparable results to DL while necessitating significantly less data for the two classification tasks. DL outperformed FE for the regression task. For all tasks, merging FE with DL did not improve performance over DL alone. These findings were confirmed on the additional PTB-XL dataset. We found that for traditional 12-lead ECG based diagnosis tasks, DL did not yield a meaningful improvement over FE, while it improved significantly the nontraditional regression task. We also found that combining FE with DL did not improve over DL alone, which suggests that the FE were redundant with the features learned by DL. Our findings provides important recommendations on 12-lead ECG based machine learning strategy and data regime to choose for a given task. When looking at maximizing performance as the end goal, if the task is nontraditional and a large dataset is available then DL is preferable. If the task is a classical one and/or a small dataset is available then a FE approach may be the better choice.

Establishment of Risk Model and Analysis of Immunoinfiltration Based on Mitophagy-Related Associated Genes in Atrial Fibrillation.

Atrial fibrillation (AF) is a common tachyarrhythmia whose pathogenesis remains elusive. In the present study, we aimed to investigate the pathological mechanism of mitophagy and immunoinfiltration in AF. First, we identified differentially expressed mitophagy-related genes (DEMRGs) based on the GSE79768 and GSE115574 datasets, subjecting them to functional enrichment analysis. STRING, TRRUST, miRNet, miRwalk, and Cytoscape were used to explore the potential regulatory roles of downstream signaling pathways. Subsequently, the random forest method was used to construct the AF risk model, and the DEMRGs most correlated with AF risk were determined by combining the Gini index. ssGSEA algorithm, NMF algorithm, and unsupervised clustering were used to subdivide AF molecular types. We then studied the characteristics of mitophagy- and immune infiltration-related genes in AF. Ultimately, we detected the expression of key genes in canine atrial tissues and HL-1 cells by immunofluorescence and Western blot. Mitophagy and immune infiltration were significantly enriched and activated in AF samples. Thirty-seven DEMRGs were screened, of which MAPK1, VDAC1, MAPK14, and MTERF3 were most associated with AF risk. The risk model based on these could identify patients at a high risk of AF. The infiltration of immunocells such as mast cells and neutrophils was significantly different among AF types. Finally, expression verification indicated that the expression trend of four key genes in canine atrial muscle tissue and HL-1 cells was consistent. We found that mitophagy may participate in AF progression through immune activation. In addition, the AF risk prediction model composed of VDAC1, MAPK1, MAPK14, and MTERF3 has a good AF prediction performance, which provides new ideas for the study of AF pathogenesis and potential therapeutic targets.

Machine Learning-Based Prediction of Atrial Fibrillation Risk Using Electronic Medical Records in Older Aged Patients

Atrial fibrillation (AF) is an independent risk factor that increases the risk of stroke 5-fold. The purpose of our study was to develop a 1-year new-onset AF predictive model by machine learning based on 3-year medical information without electrocardiograms in our database to identify AF risk in older aged patients. We developed the predictive model according to the Taipei Medical University clinical research database electronic medical records, including diagnostic codes, medications, and laboratory data. Decision tree, support vector machine, logistic regression, and random forest algorithms were chosen for the analysis. A total of 2,138 participants (1,028 women [48.1%]; mean [standard deviation] age 78.8 [6.8]years) with AF and 8,552 random controls (after the matching process) without AF (4,112 women [48.1%]; mean [standard deviation] age 78.8 [6.8]years) were included in the model. The 1-year new-onset AF risk prediction model based on the random forest algorithm using medication and diagnostic information, along with specific laboratory data, attained an area under the receiver operating characteristic of 0.74, whereas the specificity was 98.7%. Machine learning-based model focusing on the older aged patients could offer acceptable discrimination in differentiating the risk of incident AF in the next year. In conclusion, a targeted screening approach using multidimensional informatics in the electronic medical records could result in a clinical choice with efficacy for prediction of the incident AF risk in older aged patients.

Gut microbiome and atrial fibrillation—results from a large population-based study

Atrial fibrillation (AF) is an important heart rhythm disorder in aging populations. The gut microbiome composition has been previously related to cardiovascular disease risk factors. Whether the gut microbial profile is also associated with the risk of AF remains unknown. We examined the associations of prevalent and incident AF with gut microbiota in the FINRISK 2002 study, a random population sample of 6763 individuals. We replicated our findings in an independent case-control cohort of 138 individuals in Hamburg, Germany. Multivariable-adjusted regression models revealed that prevalent AF (N=116) was associated with nine microbial genera. Incident AF (N=539) over a median follow-up of 15 years was associated with eight microbial genera with false discovery rate (FDR)-corrected P<0.05. Both prevalent and incident AF were associated with the genera Enorma and Bifidobacterium (FDR-corrected P<0.001). AF was not significantly associated with bacterial diversity measures. Seventy-five percent of top genera (Enorma, Paraprevotella, Odoribacter, Collinsella, Barnesiella, Alistipes) in Cox regression analyses showed a consistent direction of shifted abundance in an independent AF case-control cohort that was used for replication. Our findings establish the basis for the use of microbiome profiles in AF risk prediction. However, extensive research is still warranted before microbiome sequencing can be used for prevention and targeted treatment of AF. This study was funded by European Research Council, German Ministry of Research and Education, Academy of Finland, Finnish Medical Foundation, and the Finnish Foundation for Cardiovascular Research, the Emil Aaltonen Foundation, and the Paavo Nurmi Foundation.

Generalizable calibrated machine learning models for real-time atrial fibrillation risk prediction in ICU patients

BackgroundAtrial Fibrillation (AF) is the most common arrhythmia in the intensive care unit (ICU) and is associated with increased morbidity and mortality. Identification of patients at risk for AF is not routinely performed as AF prediction models are almost solely developed for the general population or for particular ICU populations. However, early AF risk identification could help to take targeted preemptive actions and possibly reduce morbidity and mortality. Predictive models need to be validated across hospitals with different standards of care and convey their predictions in a clinically useful manner. Therefore, we designed AF risk models for ICU patients using uncertainty quantification to provide a risk score and evaluated them on multiple ICU datasets. MethodsThree CatBoost models, utilizing feature windows comprising data 1.5-13.5, 6-18, or 12-24 hours before AF occurrence, were built using 2-repeat-10-fold cross-validation on AmsterdamUMCdb, the first freely available European ICU database. Furthermore, AF Patients were matched with no-AF patients for training. Transferability was validated using a direct and a recalibration evaluation on two independent external datasets, MIMIC-IV and GUH. The calibration of the predicted probability, used as an AF risk score, was measured using the Expected Calibration Error (ECE) and the presented Expected Signed Calibration Error (ESCE). Additionally, all models were evaluated across time during the ICU stay. ResultsThe model performance reached Areas Under the Curve (AUCs) of 0.81 at internal validation. Direct external validation showed partial generalizability with AUCs reaching 0.77. However, recalibration resulted in performances matching or exceeding that of the internal validation. All models furthermore showed calibration capabilities demonstrating adequate risk prediction competence. ConclusionUltimately, recalibrating models reduces the challenge of generalization to unseen datasets. Moreover, utilizing the patient-matching methodology together with the assessment of uncertainty calibration can serve as a step toward the development of clinical AF prediction models.

Value of baseline characteristics in the risk prediction of atrial fibrillation.

Atrial fibrillation (AF) is prone to heart failure and stroke. Early management can effectively reduce the stroke rate and mortality. Current clinical guidelines screen high-risk individuals based solely on age, while this study aims to explore the possibility of other AF risk predictors. A total of 18,738 elderly people (aged over 60 years old) in Chinese communities were enrolled in this study. The baseline characteristics were mainly based on the diagnosis results of electrocardiogram (ECG) machine during follow up, accompanied by some auxiliary physical examination basic data. After the analysis of both independent and combined baseline characteristics, AF risk predictors were obtained and prioritized according to the results. Independent characteristics were studied from three aspects: Chi-square test, Mann-Whitney U test and Cox univariate regression analysis. Combined characteristics were studied from two aspects: machine learning models and Cox multivariate regression analysis, and the former was combined with recursive feature elimination method and voting decision. The resulted optimal combination of risk predictors included age, atrial premature beats, atrial flutter, left ventricular hypertrophy, hypertension and heart disease. Patients diagnosed by short-time ECG machines with the occurrence of the above events had a higher probability of AF episodes, who are suggested to be included in the focus of long-term ECG monitoring or increased screening density. The incidence of risk predictors in different age ranges of AF patients suggests differences in age-specific patient management. This can help improve the detection rate of AF, standardize the management of patients, and slow down the progression of AF.

Risk prediction of atrial fibrillation and its complications in the community using hs troponin I.

Atrial fibrillation (AF) is becoming increasingly common. Traditional cardiovascular risk factors (CVRF) do not explain all AF cases. Blood-based biomarkers reflecting cardiac injury such as high-sensitivity troponin I (hsTnI) may help close this gap. We investigated the predictive ability of hsTnI for incident AF in 45,298 participants (median age 51.4 years, 45.0% men) across European community cohorts in comparison to CVRF and established biomarkers (C-reactive protein, N-terminal pro B-type natriuretic peptide). During a median follow-up of 7.7 years, 1734 (3.8%) participants developed AF. Those in the highest hsTnI quarter (≥4.2ng/L) had a 3.91-fold (95% confidence interval (CI) 3.30, 4.63; p < .01) risk for developing AF compared to the lowest quarter (<1.4ng/L). In multivariable-adjusted Cox proportional hazards models a statistically significant association was seen between hsTnI and AF (hazard ratio (HR) per 1 standard deviation (SD) increase in log10(hsTnI) 1.08; 95% CI 1.01, 1.16; p=.03). Inclusion of hsTnI did improve model discrimination (C-index CVRF 0.811 vs. C-index CVRF and hsTnI 0.813; p < .01). Higher hsTnI concentrations were associated with heart failure (HR per SD 1.37; 95% CI 1.12, 1.68; p < .01) and overall mortality (HR per SD 1.24; 95% CI 1.09, 1.41; p < .01). hsTnI as a biomarker of myocardial injury does not improve prediction of AF incidence beyond classical CVRF and NT-proBNP. However, it is associated with the AF-related disease heart failure and mortality likely reflecting underlying subclinical cardiovascular impairment.

S-20-3: PLASMA ATRIAL AND BRAIN NATRIURETIC PEPTIDES LEVELS AND THE RISK OF INCIDENT ATRIAL FIBRILLATION: THE SUITA STUDY

Background: Atrial fibrillation (AF), a common cardiac rhythm abnormality, is associated with significantly increased morbidity and mortality risk. We developed the risk prediction of incident AF using traditional risk factors. It is essential to consider the residual risk to increase risk predictions. Heart failure is not a factor that appeared in the Suita AF score but others. There is no prospective study on the relationship between plasma atrial and brain natriuretic peptide (ANP and BNP) levels and incident AF in non-Western residents. Objectives: We hypothesized the association between ANP, BNP, and incident AF in the Suita Study. Methods: A total of 2,828 participants (mean age 66.7 ± 10.4 years without AF at the baseline) have been prospectively followed-up for incident AF since November 2005. AF was diagnosed when AF or atrial flutter was present on ECG at a biannual health examination or when AF was indicated as a current illness or was in the medical records during follow-up. ANP and BNP were measured by the CLEIA methods at the baseline survey examination. Cox proportional hazard ratios (HRs) and 95% confidence intervals (CIs) were analysed after adjusting for the risk components of the recently published Suita AF risk score. Statistical analyses were performed using Cox proportional hazards regression methods and evaluated using C statistics and net reclassification improvement. Results: The baseline prevalence of hypertension according to the BNP levels were 34.5%, 50.1%, and 78.8% in BNP &lt; 18.5 pg/dL, 18.5–99 pg/dL, and P ≧ 4100 pg/dL, respectively. During the 16,584 person-years of follow-up, we observed 65 incidents of AF. Compared to the subjects with ANP&lt; 42 pg/dL and BNP&lt; 18.4 pg/dL, the adjusted HRs (95% confidence intervals) of incident AF were 2.70 (1.53–4.78) for the subjects with ANP ≧ 42 pg/dL and 2.09 (1.09–4.01) and 9.75 (3.82–24.87) for the subjects with BNP = 18.5–99 pg/dL and BNPP ≧ 4100 pg/dL, respectively. The addition of ANP and BNP to variables recently combined in the Suita AF risk score showed an integrated discrimination improvement of 0.061 (0.035–0.087) and 0.068 (0.036–0.100), with 13.3% and 19.1% relative improvement in reclassification analysis, respectively. Furthermore, the addition of ANP to the Suita AF score with BNP showed a significant but slight integrated discrimination improvement of 0.007 (0.001–0.012) (P = 0.01). Conclusions: Each elevated ANP and BNP contributed significantly to the predictive for incident AF to calculate the Suita Risk Score, but adding ANP to the Suita AF score with BNP improved the risk prediction slightly.

Machine learning for atrial fibrillation risk prediction in patients with sleep apnea and coronary artery disease.

Patients with sleep apnea (SA) and coronary artery disease (CAD) are at higher risk of atrial fibrillation (AF) than the general population. Our objectives were: to evaluate the role of CAD and SA in determining AF risk through cluster and survival analysis, and to develop a risk model for predicting AF. Electronic medical record (EMR) database from 22,302 individuals including 10,202 individuals with AF, CAD, and SA, and 12,100 individuals without these diseases were analyzed using K-means clustering technique; k-nearest neighbor (kNN) algorithm and survival analysis. Age, sex, and diseases developed for each individual during 9 years were used for cluster and survival analysis. The risk models for AF, CAD, and SA were identified with high accuracy and sensitivity (0.98). Cluster analysis showed that CAD and high blood pressure (HBP) are the most prevalent diseases in the AF group, HBP is the most prevalent disease in CAD; and HBP and CAD are the most prevalent diseases in the SA group. Survival analysis demonstrated that individuals with HBP, CAD, and SA had a 1.5-fold increased risk of developing AF [hazard ratio (HR): 1.49, 95% CI: 1.18-1.87, p = 0.0041; HR: 1.46, 95% CI: 1.09-1.96, p = 0.01; HR: 1.54, 95% CI: 1.22-1.94, p = 0.0039, respectively] and individuals with chronic kidney disease (CKD) developed AF approximately 50% earlier than patients without these comorbidities in a period of 7 years (HR: 3.36, 95% CI: 1.46-7.73, p = 0.0023). Comorbidities that contributed to develop AF earlier in females compared to males in the group of 50-64 years were HBP (HR: 3.75 95% CI: 1.08-13, p = 0.04) CAD and SA in the group of 60-75 years were (HR: 2.4 95% CI: 1.18-4.86, p = 0.02; HR: 2.51, 95% CI: 1.14-5.52, p = 0.02, respectively). Machine learning based algorithms demonstrated that CAD, SA, HBP, and CKD are significant risk factors for developing AF in a Latin-American population.

CO139 Augmenting Atrial Fibrillation Risk Prediction Tools – How Does Risk Differ by Prior Stroke Type?

CO139 Augmenting Atrial Fibrillation Risk Prediction Tools – How Does Risk Differ by Prior Stroke Type?

Abstract 12814: Atrial Fibrillation Risk Prediction Through Integration of Imaging-Based Biomarker: The Multi-Ethnic Study of Atherosclerosis

Introduction: Despite the best risk prediction tools, there seems to be a significant amount of residual risk of atrial fibrillation (AF) that justifies new risk prediction strategies. The CHARGE- AF simple model outperforms other existing models. This study aims to evaluate the predictive ability of cardiac MRI-based imaging biomarkers in addition to the CHARGE AF simple model. Methods: MESA is an observational cohort study of 6,814 men and women from six U.S. communities between the ages of 45 and 84 years, who were initially free of clinical CVD. We included 4185 participants after excluding baseline AF (n=70) and missing values of imaging biomarkers (n=2559) at the baseline exam. Left atrial (LA) volume and left ventricular (LV) volumes were obtained from cardiac MRI measurements. Multivariable Cox regression was used to assess the independent 5/10-year AF predictive ability of imaging biomarkers in addition to the baseline model while adjusting for confounders. Results: 4,185 participants were included in the study. The average age (SD) was 61(±10) years, 47% male, 13% current smokers, and 39% White with a median follow-up of 17 years. Of this population, 19% developed AF. In multivariable regression models, a 1 ml increase in LA volume was significantly associated with a 3 to 10% higher risk of developing AF. Compared to the baseline model (M1, with no imaging biomarker), the 5 /10-year risk prediction of atrial fibrillation was improved after adding the imaging biomarkers for models 2, 3, and 6 (Table 1). The 5 /10 years concordance index for assessing risk prediction increased marginally from 80 / 79% in M1 to 84 / 81%, 84 / 81%, and 84 / 81% in M2, M3 and M6 respectively (all p values &amp;lt 0.0001). Conclusions: The incorporation of MRI-based LA volume index into the CHARGE AF simple model demonstrated a modest improvement in the 5/10-year risk prediction of AF. The clinical benefit should be interpreted with caution due to marginal improvement in risk prediction.