Multimodality assessment of risk in dilated cardiomyopathy- the importance of CMR

Background Dilated cardiomyopathy (DCM) has a 20% 5 year mortality. Cardiac MR (CMR) is an established outcome predictor. We evaluate the additive role of novel genetic and circulating biomarkers.Purpose Perform an integrated assessment to evaluate the prognostic importance of CMR parameters in DCM in the context of clinical, genetic and biomarker data.Methods Prospectively recruited DCM patients underwent comprehensive clinical evaluation, CMR with late-gadolinium enhancement (LGE), sequencing for rare variants in major DCM genes (titin-TTNtv, myosin heavy chain-MYH7, troponin T2-TNNT2, lamin-LMNA), biomarker assessment of BNP, troponin I (hsTnI), Galectin 3 and ST2, and follow up for clinical events. Cox proportional hazard modelling evaluated the primary composite endpoint of cardiovascular death, major arrhythmic events (aborted sudden cardiac death, appropriate ICD activation, sustained ventricular tachycardia, ventricular fibrillation) and major heart failure events (heart transplant, left ventricular assist device, unplanned heart failure hospitalisation).Results In total, 423 patients with DCM (mean age 53.6±14.1 years, 67% male) were followed up for a median of 4.0 years (IQR 2.1–5.8). Mean left ventricular ejection fraction (LVEF) was 40%±12.5%. One third of patients had mid-wall fibrosis (MWF-LGE) (n=137, 32%). Mean indexed left atrial volume (LAVi) was 61±27.6 mls.On genetic analysis, 53 patients (12.5%) had TTNtv, 14 patients (3.3%) had non-truncating variants in MYH7, 5 patients (1.2%) had non-truncating variants in TNNT2 and 5 patients (1.2%) had LMNA variants.In total, 44 patients (10.4%) met the primary endpoint. On multivariable analysis, an optimal model predicting the primary endpoint was built without genetic or biomarker data (Table 1). This consisted of LVEF, LAVi, and MWF-LGE (Table 2).On univariable analysis, BNP, Galectin 3 and hsTnI were associated with the primary endpoint (Table 2). When the novel variables were added to the optimised model, Galectin 3 and LMNA variants were predictive of the primary endpoint (Table 3). In the final adjusted model, CMR parameters remained predictive of the primary endpoint. Of these, MWF-LGE was the strongest predictor (adjusted HR 2.12 (1.06–4.02), p=0.02).View this table:Abstract 005 Table 1 Variables evaluated in building baseline Cox proportional hazard model predicting primary endpointView this table:Abstract 005 Table 2 Results of baseline Cox proportional hazard modelling predicting primary endpointView this table:Abstract 005 Table 3 Results of evaluation of biomarkers and genetic variants using Cox proportional hazard modelling predicting primary endpointView this table:Abstract 005 Table 4 Results of final adjusted Cox proportional hazard model predicting primary endpointConclusion On integrated multi-modality risk stratification in DCM, while circulating and genetic biomarkers improve risk stratification, CMR parameters remain strong independent predictors of outcome.Acknowledgements UT is supported by a Medical Research Council (UK) Clinical Research Training Fellowship. The study has also been supported by the Rosetrees Foundation, the Alexander Jansons Foundation, the Wellcome Trust and the NIHR Cardiovascular Biomedical Research Unit of Royal Brompton and Harefield NHS Foundation Trust and Imperial College London.%U http://heart.bmj.com/content/heartjnl/103/Suppl_1/A4.2.full.pdf