Population
Two participating centers, the Centro Cardiologico Monzino, Milan, Italy (center A) and the Humanitas Research Hospital, Milan, Italy (center B) prospectively included patients with known or suspected HCM between January 2011 and April 2017 (center A) and between January 2011 and November 2017 (center B). The diagnosis of HCM was defined according to the European Society of Cardiology guidelines (maximal LV wall thickness of ≥15 mm on echocardiography, in the absence of any other cardiac or systemic disease that would be capable of producing myocardial hypertrophy, such as afterload abnormalities like aortic valve stenosis or arterial hypertension [2]. Clinical and CMR data were collected at each center by two experienced physicians. Each institution’s ethical committee approved the protocol and all patients gave written informed consent.
CMR PROTOCOL
CMR studies were performed with a 1.5-T Discovery MR450 scanner (GE Healthcare, Milwaukee, Wisconsin) in center A. In center B, a 1.5-T CVi scanner (GE Healthcare, Milwaukee, Wisconsin), a 1.5-T Achieva (Philips Medical Systems, Best, Netherlands) or a 1.5-T Aera scanner (Siemens, Erlangen, Germany) were used. Dedicated cardiac software, phased array surface receiver coils, and electrocardiogram triggering were used. Breath-hold steady-state free precession cine imaging was performed in vertical and horizontal long-axis orientations and in short-axis orientations. A stack of short-axis slices encompassing the right and left ventricle from base to apex was used for biventricular volumes, mass, and systolic function assessment. For detecting myocardial fibrosis, we used a contrast-enhanced, breath-hold, segmented T1-weighted inversion-recovery gradient-echo sequence with the LGE technique (i.e., LGE imaging). LGE imaging was performed 10 to 20 min after an intravenous bolus of 0.1 mmol/kg of gadolinium-BOPTA (Multihance, Bracco, Milan, Italy), 0.2 mmol/kg of gadolinium-DOTA (Dotarem, Guerbet, Roissy, France), or 0.2 mmol/kg of Gadobutrol (Gadovist, Bayer AG, Basel, Switzerland). Inversion time was individually adapted to null the signal of remote myocardium (usual range between 220 and 300 ms).
CMR ANALYSIS
All CMR studies were analyzed by expert readers with considerable experience in CMR studies (12 years of experience, Level 3 certified expert), blinded to patient clinical history and data. The short- axis stack was used for calculating RV and LV size and function. Endocardial and epicardial boundaries were delineated in end-diastole and end-systole with a dedicated software (Circle Cardiovascular Imaging, cvi42 version 5.3.2, Calgary, Canada). Based on these data, the following parameters were calculated: RV and LV end-diastolic volumes (RVEDV and LVEDV, respectively), RV and LV end-systolic volumes (RVESV and LVESV, respectively) as well as RV and LV ejection fraction (RVEF and LVEF, respectively), and LV mass. Each volume or mass parameter was indexed to body surface area and expressed as mL/m2 or g/m2, respectively. Additionally, each image was inspected for detecting RV and LV hypertrophy, and the maximal wall thickness of the left and right ventricle was recorded. The presence and amount of scar was assessed using 2D LGE sequences. LGE was defined as hyperintense myocardium with a signal intensity >5 SDs above the mean signal intensity of normal myocardium. For quantitation of LGE, inner and outer myocardial edges were manually delineated. LGE was determined semi-automatically as a percentage of total myocardium.
FOLLOW-UP
All patients underwent clinical visits or telephone interviews performed at each center by experienced cardiologists blinded to CMR data. Outcome measures were a composite of ventricular arrhythmias (non-sustained ventricular tachycardia, i.e. occurrence of 3 or more consecutive ventricular beats for less than 30 second, or sustained ventricular tachycardia), hospitalization for HF and cardiac death due to ventricular arrhythmias or refractory HF.
STATISTICAL ANALYSIS
Unpaired t-test or Kruskal-Wallis test were performed to compare continuous variables between patient without events and those with combined events, while chi-square test or Fisher’s exact test were performed for analysis that involved categorical variables. Univariate and multivariate Cox regression was employed to investigate potential predictors for combined event. The covariates included in the multivariate analysis were selected through epidemiological approach. Continuous variables are presented as mean±SD, skewed variables as median with interquartile range (IQR), while categorical variables are shown as absolute numbers and percentages.