Mitral annulus disjunction: a detailed magnetic resonance study of phenotype and clinical significance

Abstract

Abstract Background The clinical significance and arrhythmogenic propensity of mitral annulus disjunction (MAD) are not well established. Our aim was to characterize MAD in all-comers, analyzing the mitral valve apparatus in detail, scar burden and clinical outcomes. Methods Single-center retrospective analysis of all consecutive patients with MAD described by cardiovascular magnetic resonance (CMR) between 2010-2022. Medical history, imaging and arrhythmia data were analysed. Cardiovascular mortality was assessed. MAD was defined as a separation ≥2mm between the left ventricular myocardium and the mitral annulus. Presence and pattern of left ventricle (LV) late gadolinium enhancement (LGE) was assessed. Scar in the papillary muscles and areas adjacent to the MAD was identified as MAD-related. Ventricular arrhythmias (VA) were grouped into non-sustained (ectopies>10%/24h or ventricular tachycardia) or sustained. Results Out of 341 patients who were initially identified as possibly having MAD, MAD was confirmed in 222. Mean age was 49.2±19.3 years, 126 were male (56.8%). Mean longitudinal MAD extent was 5±2.6mm, while circumferential involvement of a single segment (60º) was most commonly seen (n=99, 44.6%). The inferior wall was the affected segment (n=116, 52.2%) with more extensive MAD, followed by the inferolateral wall (n=71, 32%). Mitral valve (MV) annulus expanded during systole in patients with MAD ≥6mm, while annular contraction occurred in those with MAD <6mm. Annular expansion was associated with MV prolapse (p=0.003), billowing (p=0.018) and curling (p<0.001). LV LGE was present in 82 patients (36.9%), with 23 patients (10.4%) showing MAD-related LGE, but no association with disease extent (MAD length or circumferential involvement). Median follow-up time was 1392 (872-1922) days. No episode of sustained VA was observed. In univariable analysis non-sustained VA was significantly more prevalent in patients with MAD ≥6mm (p=0.002), although this tendency faded after multivariate analysis (p=0.053). The presence of non-sustained VA was associated with global LV LGE (p=0.005), but not MAD-related LGE in isolation - Table 1. Three patients died of cardiovascular causes (1.4%), but none had extensive MAD or MAD-related LGE. None died of sudden cardiac arrest. To the best of our knowledge, this is the biggest dataset of MAD patients that simultaneously analysis phenotype, including LGE, and clinical outcomes. Conclusion Mitral valve dysfunction (prolapse/billowing) is linked to MAD severity. Scar is common, not related to the extent of MAD, but a predictor of VAs independently of its location. The risk of sustained ventricular arrhythmias and cardiovascular death appears to be low in this all-comers population.Table 1