Atrial fibrillation and smoking: cardiovascular markers predicting brain natriuretic peptide in patients treated with radiofrequency ablation

Introduction: brain natriuretic peptide is secreted due to stretching due to increased ven-tricular volume. Previous smoking would influence the correlation between this hormone and cardiovascular markers in post-radiofrequency ablation atrial fibrillation. The objective was to analyze the relationship and predictive factors of brain natriuretic peptide with biochemical and echocardiographic markers in patients with and without smoking habit with atrial fibrillation treated with radiofrequency ablation. Material and methods: cross-sectional and analytical study of secondary data from the Dryad digital repository, including 192 participants with post-radiofrequency ablation atrial fibrillation. Student's t-test, Spearman correlation, neural networks (multilayer perceptron), kappa and Cramer's V coefficients, Chisquare, determination of sensitivity and specificity were used. Results: with history of smoking, brain natriuretic peptide was poorly and negatively correlated with LDLcholesterol, moderately and negatively correlated with ejection fraction; low and positively with left atrial volume. In patients without history, brain natriuretic peptide correlated low and positively with diastolic pressure, low and negatively with alanine aminotransferase and ejection fraction. Multilayer perceptron indicated an accuracy percentage for predicting normal or elevated BNP of more than 80% based on smoking history. Conclusions: smoking history influences the correlation between brain natriuretic peptide and cardio-vascular markers in patients post-radiofrequency ablation for atrial fibrillation. It is feasible to create a neural network algorithm for diagnosing elevated levels of brain natriuretic peptide based on the type of previous exposure to tobacco.