Prognostic impact of atrio-ventricular conduction abnormalities in non-ischemic cardiomyopathy

There is an urgent need to improve the risk-stratification for ventricular arrhythmias (VA) and sudden death (SD) in non-ischaemic cardiomyopathy (NICM), moving beyond left ventricular ejection fraction (LVEF). Late gadolinium enhancement (LGE) at cardiac magnetic resonance (CMR) is a strong and independent predictor of VA and SD in NICM. However, it is necessary to find additional predictors on top of LGE and it is essential to further refine the risk stratification within higher-risk LGE+ cases.

Atrioventricular block (AVB) is a recognised risk factor for VA and SD in specific sub-groups of patients such as laminopathies and cardiac sarcoidosis. However, the prognostic role of AVB in the overall spectrum of NICM has been little explored, with divergent results.

To reassess the prognostic value of AVB in a contemporary cohort of patients with NICM studied with CMR, with a specific focus on the potential improvement of risk stratification in LGE+ cases.

Retrospective cohort study conducted at two tertiary care hospitals, including patients with NICM, evaluation of AV conduction and a CMR with LGE analysis available. The primary endpoint was a combined arrhythmic endpoint including any appropriate ICD therapies, sustained monomorphic ventricular tachycardia, resuscitated cardiac arrest and sudden death.

The secondary endpoint was a combined heart failure endpoint including heart failure hospitalization, heart transplant, or left ventricular assist device implant and death due to end stage heart failure.

We included 616 patients. AVB was observed in 9% of patients, being in most cases (84%) first degree AVB. Patients with AVB were older (64 vs 58, p=0.003) presented with more severe cardiac dysfunction (LVEF 29% vs 37%, p<0.001) and had higher prevalence of LGE (63% vs 46%, p=0.02).

AVB independently predicted the combined arrhythmic endpoint on top of LGE and LVEF (adjusted HR 2.7, p=0.007) and was particularly impactful when present alongside LGE: among LGE+ patients, AVB increased the risk of the primary endpoint by more than three times (adjuster HR 3.4, p<0.001). Three groups of patients were identified: 1) The low-risk group encompassed LGE- patients (incidence rate of 0.4 per 100 people/year). 2) The intermediate-risk group included LGE+ patients without AVB (incidence rate of 3.8 per 100 people/year). 3) The high-risk group was that of LGE+ patients with AVB (incidence rate of 13 per 100 people/year).

By contrast, AV block was not a significant predictor of the heart failure endpoint at multivariate analysis (HR 1.7, p=0.11).

Atrio-ventricular conduction abnormalities are a strong and independent predictor of ventricular arrhythmias and sudden death in NICM, on top of LGE and LVEF. In addition, AVB has a significant impact on the risk-stratification of LGE+ cases: the coexistence of AVB and LGE identifies a small sub-group of patients with extremely high arrhythmic risk.