Introduction Left ventricular non compaction (LVNC) is a heterogeneous entity characterized by prominent LV trabeculae. Fractal analysis, a semi-automatic method to quantify the degree of hypertrabeculation, has been described as a diagnostic criterion for LVNC. However, the prognostic implications of fractal analysis have not been investigated. Purpose Therefore, the aim of our study was to identify the relationship between the trabecular complexity measured by fractal analysis and the occurrence of cardiovascular events in LVNC patients. Methods We conducted a retrospective longitudinal multicenter cohort study of consecutive patients fulfilling LVNC criteria by cardiovascular magnetic resonance (CMR). The endpoints were heart failure (HF), ventricular arrythmias (VA), systemic embolism (SE) and all-cause mortality. Fractal analysis was performed on short-axis CMR cine images using CVI 42 software. The cohort was divided in two groups according to the maximal apical fractal dimension (FD). Results A total of 291 patients were included: age was 44±19 years and 56% were female. LV ejection fraction (LVEF) was 47% ± 14% and 26 patients (9%) presented late gadolinium enhancement (LGE). Mean global FD was 1.29±0.06 and maximal apical FD was 1.39±0.09. Baseline characteristics of patients with maximal apical FD 1.39 and 1.39, including functional LV parameters, were comparable (Table 1). After 3.8±2.5 years of follow-up, 37 patients (12.8%) presented HF, 30 (10.4%) had VA, SE occurred in 3 (1%) and 3 patients (1%) died. Maximal apical FD was not associated with the risk of HF or VA (both p ns), nor was mean global FD. There were no significant differences in the survival curves of both endpoints between the two groups: HR were 2.34 (95% CI 0.74–7.4, p 0.13) and 1.15 (95% CI 0.45–2.41, p 0.71) respectively for HF and VA (Image 1). LVEF and LGE remained consistent predictors of events. Conclusions In a large multicenter retrospective LVNC study, the degree of hypertrabeculation measured by fractal analysis was not associated with functional LV parameters and, most importantly, with the incidence of cardiovascular events. Therefore, our results suggest that diagnostic LVNC criteria should be refined to include parameters with actual prognostic implications. Funding Acknowledgement Type of funding sources: None.

Introduction Left ventricular non compaction (LVNC) is a heterogeneous entity with uncertain prognosis. Cardiac magnetic resonance (CMR) is widely used in the diagnosis of LVNC. However, its role in risk stratification has not been well established. Purpose Therefore, the aim of our study was to identify prognostic CMR variables in LVNC. Methods We conducted a retrospective longitudinal multicentre cohort study of consecutive patients fulfilling CMR LVNC criteria. The endpoints were heart failure (HF), ventricular arrhythmias (VA), systemic embolisms (SE) and all-cause mortality. Biventricular volumes, ejection fraction (LVEF and RVEF) as well as late gadolinium enhancement (LGE) were analysed. Results A total of 310 patients were included: age was 44.4±19 and 43% female. LVEF was 47% ± 15%, RVEF was 48±12 and 28 patients (9%) presented LGE. After a median follow-up of 3.8 2.5 years, 40 patients (13%) presented HF, 31 (10%) had VA, SE occurred in 6 (2%) and 3 patients (1%) died. Baseline characteristics of patients with and without HF and VA are described in Table 1. In univariate analysis, LVEF, LV volumes, LGE, and RVEF were associated with both HF and VA risk. In multivariate analysis, LVEF was the only variable independently associated with HF (HR 0.932, CI 95% 0.88–0.97, p 0.003). Patients with an LVEF >35% were at very low risk of HF (Figure 1A). With regards to VA, LGE was the only independent predictor (HR 2.64, IC 95% 1.059–6.61, p 0.003) (Figure 1B). In LGE negative patients, the arrhythmic risk was higher among those with an LVEF <35% (HR 2.81, 95% CI 1.02–8.12, p 0.047). Conclusions In a large multicentre retrospective LVNC study, left ventricular ejection fraction and late gadolinium enhancement were the main predictors of cardiovascular events. Patients with an LVEF <35% and with LGE were at markedly increased risk. Therefore, we suggest that these variables should be combined to enhance risk stratification in LVNC. Funding Acknowledgement Type of funding sources: None.