Paroxysmal or chronic atrial fibrillation (AF) is frequent in hypertrophic cardiomyopathy (HCM), occurring in more than 20%-25% of patients, and is often considered as an important turning point for the quality of life of these patients. Defining the risk of atrial fibrillation in hypertrophic cardiomyopathy patients is an important clinical and prognostic challenge.

The aim of this study is to determine HCM-phenogroups with different risk of AF-occurrence at 5-year.

We applied retrospectively the Bayesian method, which can analyze a large number of variables, to differentiate phenogroups of patients with different risks of AF and prognoses across a French prospective on-going hospital-based registry of adult HCM patients (REMY). Clinical and imaging data were prospectively recorded, and patients were followed for 5 years. A total of 1431 HCM patients were recruited, including 1275 analyzed in the present study after exclusion criteria.

The population included 412 women, and 369 patients with obstructive HCM. AF-occurred in 167 (11.6 %) patients during the 5-year follow-up. Three phenogroups were defined according to their common clinical and echocardiographic-characteristics. Patients at the highest risk were oldest, more often female, with more frequent comorbidities, anteroposterior diameter of the left atrium was significantly greater, with diastolic dysfunction, outflow-tract obstruction and mitral valve abnormality, and presented higher pulmonary artery pressure and/or right ventricular dysfunction. These also had a higher risk of all-cause hospitalizations and death.

Based on a clustering analysis, three phenogroups of HCM according to the risk of AF occurrence can be identified. It can indicate which patients should be more monitored and/or treated, particular to prevent the risk of stroke.