SETAP: epidemiology and prevention of stroke and transient ischaemic attack in Czech patients with atrial fibrillation

An effective prediction model based on XGBoost for the 12-month recurrence of AF patients after RFA

BACKGROUND

Atrial fibrillation (AF) is a common heart rhythm disorder that can lead to complications such as stroke and heart failure. Radiofrequency ablation (RFA) is a procedure used to treat AF, but it is not always successful in maintaining a normal heart rhythm. This study aimed to construct a clinical prediction model based on extreme gradient boosting (XGBoost) for AF recurrence 12 months after ablation.

METHODS

The 27-dimensional data of 359 patients with AF undergoing RFA in the First Affiliated Hospital of Soochow University from October 2018 to November 2021 were retrospectively analysed. We adopted the logistic regression, support vector machine (SVM), random forest (RF) and XGBoost methods to conduct the experiment. To evaluate the performance of the prediction, we used the area under the receiver operating characteristic curve (AUC), the area under the precision-recall curve (AP), and calibration curves of both the training and testing sets. Finally, Shapley additive explanations (SHAP) were utilized to explain the significance of the variables.

RESULTS

Of the 27-dimensional variables, ejection fraction (EF) of the left atrial appendage (LAA), N-terminal probrain natriuretic peptide (NT-proBNP), global peak longitudinal strain of the LAA (LAAGPLS), left atrial diameter (LAD), diabetes mellitus (DM) history, and female sex had a significant role in the predictive model. The experimental results demonstrated that XGBoost exhibited the best performance among these methods, and the accuracy, specificity, sensitivity, precision and F1 score (a measure of test accuracy) of XGBoost were 86.1%, 89.7%, 71.4%, 62.5% and 0.67, respectively. In addition, SHAP analysis also proved that the 6 parameters were decisive for the effect of the XGBoost-based prediction model.

CONCLUSIONS

We proposed an effective model based on XGBoost that can be used to predict the recurrence of AF patients after RFA. This prediction result can guide treatment decisions and help to optimize the management of AF.

A nomogram to predict left atrial appendage thrombus and spontaneous echo contrast in non-valvular atrial fibrillation patients

Background

Non-valvular atrial fibrillation (NVAF) significantly increases the risk of stroke. Although there is availability of prediction models, their ability to predict the risk of stroke in NVAF patients remains suboptimal. Therefore, there is need to improve prediction of high-risk individuals, which is critical for efficient management of patients with NVAF.

Objective

The objective of our paper is to develop a nomogram for predicting the risk of left atrial appendage thrombus (LAAT) and spontaneous echo contrast (SEC), thereby replacing the risk of stroke in NVAF patients.

Design

This was a retrospective cohort study that analyzed clinical data and echocardiographic indices of 387 patients with NVAF from October 2018 to June 2021. Multivariable logistic regression was used to evaluate independent factors that were used to construct the prediction nomogram.

Analysis

The discriminative ability and calibration of the nomogram to predict LAAT/SEC were tested using C-statistic and calibration plot. The performance of the nomogram was assessed against the CHA2DS2 score, CHA2DS2-VASc score and ATRIA score using the receiver operating characteristic curve (ROC), decision curve analysis (DCA), integrated discrimination index (IDI) and net reclassification index (NRI).

Result

Out of the total 387 patients enrolled in this study, 232 had LAAT/SEC. Multivariable analyses demonstrated that N-terminal pro-B-type natriuretic peptide (NT-proBNP), albumin (ALB), LAA ejection fraction (LAAEF) and LAA global peak longitudinal strain (LAA GPLS) were independent predictors of LAAT/SEC. The constructed nomogram had good discriminative (C = 0.886) and calibration (0.876) abilities after bias correction by the C-index. Compared with other models, the decision curve analyses demonstrated that the nomogram had greater net benefits. Besides, the nomogram had significant improvement in predictive performance, sensitivity and reclassification for LAAT/SEC compared with the CHA2DS2 [(c-index: 0.886 vs. 0.576, p < 0.05), (NRI: 0.539, p < 0.05), (IDI: 0.432, p < 0.05)], CHA2DS2-VASc [(c-index: 0.886 vs0.579, p < 0.05), (NRI: 0.513, p < 0.05), (IDI: 0.432, p < 0.05)] or ATRIA [(c-index: 0.886 vs0.583, p < 0.05), (NRI: 0.546, p < 0.05), (IDI: 0.432, p < 0.05)].

Conclusion

Taken together, our data demonstrated that the developed nomogram was effective and had potential clinical application in the prediction of LAAT/SEC in patients with NVAF.

High Prevalence of Atrial Fibrillation in a Lithuanian Stroke Patient Cohort

Background and Objectives: Atrial fibrillation (AF) is the most common cardiac arrhythmia and is associated with a five-fold increased risk for acute ischemic stroke (AIS). We aimed to estimate the prevalence of AF in a Lithuanian cohort of stroke patients, and its impact on patients regarding case fatality, functional outcome, and health-related quality of life (HRQoL) at 90 days. Materials and Methods: A single-center prospective study was carried out for four non-consecutive months between December 2018 and July 2019 in one of the two comprehensive stroke centers in Eastern Lithuania. A telephone-based follow-up was conveyed at 90 days using the modified Rankin Scale (mRS) and EuroQoL five-dimensional three-level descriptive system (EQ-5D-3L) with a self-rated visual analog scale (EQ-VAS). One-year case fatality was investigated. Results: We included 238 AIS patients with a mean age of 71.4 ± 11.9 years of whom 45.0% were female. A striking 97 (40.8%) AIS patients had a concomitant AF, in 68 (70.1%) of whom the AF was pre-existing. The AIS patients with AF were at a significantly higher risk for a large vessel occlusion (LVO; odds ratio 2.72 [95% CI 1.38−5.49], p = 0.004), and had a more severe neurological impairment at presentation (median NIHSS score (interquartile range): 9 (6−16) vs. 6 (3−9), p < 0.001). The LVO status was only detected in those who had received computed tomography angiography. Fifty-five (80.9%) patients with pre-existing AF received insufficient anticoagulation at stroke onset. All patients received a 12-lead ECG, however, in-hospital 24-h Holter monitoring was only performed in 3.4% of AIS patients without pre-existing AF. Although multivariate analyses found no statistically significant difference in one-year stroke patient survival and favorable functional status (mRS 0−2) at 90 days, when adjusted for age, gender, reperfusion treatment, baseline functional status, and baseline NIHSS, stroke patients with AF had a significantly poorer self-perceived HRQoL, indicated by a lower EQ-VAS score (regression coefficient ± standard error: β = −11.776 ± 4.850, p = 0.017). Conclusions: In our single-center prospective observational study in Lithuania, we found that 40.8% of AIS patients had a concomitant AF, were at a higher risk for an LVO, and had a significantly poorer self-perceived HRQoL at 90 days. Despite the high AF prevalence, diagnostic tools for subclinical AF were greatly underutilized.