Background: As AF-associated morbidity and mortality are increasing, there is an acute need for improved surveillance and prevention strategies to reduce the impact of AF and related strokes. Specific echocardiographic parameters that can best predict future onset of AF within 3 months are lacking.
Methods: Twenty patients with AF, as identified by presence of ICD-9 diagnosis code, were compared with a control group of twenty age- and sex-matched patients selected from the same clinic population but without a diagnosis of AF. Transthoracic echocardiograms (TTE) obtained within 90 days prior to first documented AF episode (study group) or obtained closest to first clinic visit (control) were selected for review.
Results: Baseline characteristics, including age, BMI, presence of hypertension, hyperlipidemia, diabetes, and heart failure were comparable. Increased left atrial (LA) size (end systolic major axis in 2-chamber view: AF 4.62±0.03 vs control 3.79±0.21, P =0.03), increased mitral inflow (E/A ratio: AF 1.35±0.15 vs control 1.06±0.07, P =0.04), and reduced LA global longitudinal strain (AF -2.69±0.26 vs control - 3.59±0.31, P =0.04) were most closely associated with AF compared with the control group. Multivariate logistic regression was used to develop predictive models for AF onset. A combination of imaging and traditional clinical risk factors was the best AF prediction model with AUC of 0.94, which greatly exceeds the current best predictors published. From these parameters, we developed the SMASH2 scoring system for 90- day AF risk estimation.
Conclusions: Risk factors for AF and early features of atrial cardiomyopathy including male sex, hypertension, LA enlargement, reduced mitral inflow, and reduced LA strain are powerful predictors of AF onset within 90 days, and may be used to prognosticate future AF risk.
Keywords: Atrial cardiomyopathy; Atrial fibrillation; Left atrial strain; Prediction.
© 2022. This is a U.S. government work and not under copyright protection in the U.S.; foreign copyright protection may apply.