Methods

One hundred twenty-six consecutive healthy subjects aged 18 to 35 years were prospectively enrolled through e-mail and/or word-of-mouth and provided signed informed consent approved by the institution ethics board. Subjects were excluded, based on a standardized questionnaire and a physical examination, if any of the following conditions were present: any congenital or acquired cardiovascular disease, cardiovascular risk factors (hypertension, dyslipidemia, or diabetes mellitus), or presence of renal, hepatic, or blood disorders. Finally, subjects were not included if they possessed standard contraindications to CMR. Pregnant women or those within 1 year of childbirth were excluded. In addition, patients were evaluated with a 12-lead electrocardiogram at rest and N-terminal pro-brain natriuretic peptide (Nt-pro-BNP; analytical sensitivity 5 fmol/ml; proBNP 8-29 ELISA; Biomedica Gruppe and Alpco Diagnostics, Salem, NH). Finally, subjects with cardiac congenital anomalies seen at CMR were excluded. The population, therefore, consisted of 100 young adults without known cardiovascular or systemic disease, without cardiovascular risk factors, with a normal electrocardiogram at rest, with Nt-pro-BNP within normal limits, and without congenital cardiac disease on CMR.

Imaging was performed with a 1.5 T Philips Achieva scanner operating, release 2.6, level 3 (Philips Healthcare, Best, The Netherlands). Cine imaging of cardiac morphology and function was performed by steady-state free precession technique at 30 phases per cardiac cycle in apnea. Short-axis (8 mm thickness and 0 mm gap) and 3 radial long-axis planes were performed covering the entire cardiac silhouette (time repetition/time echo 3.17/1.58 ms, flip angle 60°, number of excitations = 1, in-plane spatial resolution 1.6 × 2 mm). Gadolinium contrast was not used.

Image analysis was performed off-line in an experienced core laboratory using a standardized approach by trained technicians supervised by an experienced cardiologist (EL) following the 16-segment model (CMR Mass, version 7.1; Medis, Leiden, The Netherlands). Cardiac volumes and function measurements were performed as previously described. In summary, for LV volume analysis, the endocardial and epicardial borders were manually determined for all 30 phases of the cardiac cycle, and cardiac phases that had the larger and smallest ventricular cavity volumes were defined as end-diastole (ED) and end-systole (ES), respectively. Papillary muscles were included in the initial LV wall measurements (equivalent to weighting the LV) and excluded from LV cavity measurements (equivalent to blood pool techniques). For cardiac volumes and functional measurements, the endocardial border was defined as the border between trabeculations and the ventricular blood pool, excluding papillary muscles (trabeculations were included in the LV wall and excluded from the blood pool). The LV end-diastolic volume, LV end-systolic volume (LVESV), LV stroke volume, LV ejection fraction (LVEF), and LV mass were computed using Simpson’s rule adjusted to body surface area. Segmental wall thickness was measured at ED by the centerline method (average of 20 to 30 chords/segment) and was compared with the average chord thickening at ES in each segment to determine segmental wall function. Decreased LV segmental wall function was considered present if systolic wall thickening was <30%. Segment 17 was excluded. Left atrial (LA) endocardial borders were also determined for all 30 phases, and the ED (largest) and ES (smallest) volumes and the EF were calculated from Simpson’s method. LA appendix volumes were included in the total LA volumes.

Following standard measurements, LV trabeculated versus nontrabeculated myocardium were specifically analyzed by 2 experienced readers (HT and EL) blinded to each other and to all other variables, followed by resolution of any differences by consensus. Slices from short-axis (SA), 4-chamber (4CH), and 2-chamber (2CH) planes were analyzed. Analysis was performed at ED and ES in 3 planes of SA (basal, mid, and apical LV) and in a single plane of 2CH and a single plane of 4CH views (the planes in which the papillary muscles were easier to distinguish from trabeculations) to mirror the approach used for clinical interpretation by most CMR readers. Segmentation followed the American Heart Association recommendations. We first measured the combined sum of trabeculated (T) and dense (D) myocardium by delineating the epicardial border and the trabeculations border. Delineation of the trabeculations border for measurement of T + D was performed by connecting the inner tips (toward the center of the LV cavity) of all the trabeculations. The full thickness of T + D was then measured by the centerline method along 20 to 30 chords per segment, providing (1) maximum and (2) mean thickness and (3) area objectively and avoiding investigator bias associated with individual selection of the measurement site. Afterward, without altering the delineation of the trabeculations border, the inner border of dense myocardium was delineated. Trabeculations were defined per protocol as any muscular structure that moved synchronously with the inner myocardial border (endocardium) over the cardiac cycle and that was not attached to a papillary muscle at ES. These 2 conditions had to be present in at least 2 orthogonal planes for a muscular structure to be considered trabeculated myocardium. Other structures were excluded ( Supplementary Table 2 ). We then calculated the thickness and area of D myocardium for each segment using the formula: D = (T + D) – T. We determined the trabecular burden by each of the 3 following measurements: (1) thickness of trabeculated over dense myocardium (T/D) ratio, (2) percent trabeculations relative to total myocardium thickness (%T), and (3) percent trabeculations relative to total myocardium area (%TA). All measurements were performed in ED and ES for all segments in SA, 2CH, and 4CH. Taking into account the limits of resolution of the steady-state free precession sequence, we set the analytical sensitivity of our measurements at 1.5 mm ( Supplementary Figure 1 ).

All variables were evaluated by the Shapiro-Wilk test for normality. Categorical variables were expressed as percentages and continuous variables as means ± SD. Gender differences were verified using chi-square tests for categorical variables and Student t tests for continuous variables. When required, logarithmic transformations were performed and linear relations evaluated by Pearson’s correlations between global per-subject T/D ratios in ED and ES and (1) LVEF, (2) Nt-pro-BNP levels, and (3) mean segmental function. Analysis of variance was performed to test associations between categorical and continuous variables. When segmental analysis was performed, repeated observations because of the analysis of 16 myocardial segments per subject were taken into consideration using generalized estimating equation regression modeling. Univariable and multivariable regression analysis were performed to test associations of T/D ratio with LVEF by including age, gender, body mass index (BMI), systolic blood pressure, LA end-diastolic volume (LAEDV), LVESV, Nt-pro-BNP, and mean T/D ratio at ES in the multivariable model. Interobserver agreement for measurements of trabeculations was evaluated using multiple regression taking into account repeated measures (multiple segments per subject) in a random sample of 10% of the cohort. Statistical analysis was performed with Stata 11.0 (StataCorp LP, College Station, Texas).

Results

The general characteristics of study population are listed in Table 1 . No anomalies were present on screening tests (electrocardiograms or plasma levels) on participants. Mean LV function, mass, and volumes and LA function and volumes were within normal limits for age and gender ( Table 1 ). On average, women had lower LV and LA volumes and LV mass indexed to body surface area. Loading conditions, as assessed by blood pressure, heart rate, and total body water percentage (by bioelectrical impedance analysis), did not differ between subjects with greater trabecular burden and those without at the time of CMR examination ( Supplementary Table 3 ).

Trabeculated myocardium was visible by CMR in all subjects. Trabecular burden is reported in Table 2 for the 3 imaging planes in ES and ED as (1) T/D ratio, (2) %T, and (3) %TA. Trabecular burden was not associated to general body morphometrics, defined by BMI (r = 0.07, p = 1.0). Whether they were quantified by T/D ratio, %T, or %TA, trabeculations were consistently more abundant in SA compared with 2CH, and in 2CH compared with 4CH, and by all 3 methods of quantification, more abundant in ED compared with ES. Women also consistently presented greater proportions of trabeculated myocardium compared with men (mean T/D ratio 0.85 ± 0.24 vs 0.72 ± 0.19, p = 0.006). On average, T/D ratios were greater for anterior wall segments compared with nonanterior wall segments, whether at ED (2.01 ± 1.00 vs 1.22 ± 0.56, respectively, p <0.0001) or ES (0.81 ± 0.44 vs 0.52 ± 0.25, p <0.0001). In addition, T/D ratios were greater for apical segments compared with nonapical (basal and midventricular) segments, either at ED (1.87 ± 0.90 vs 1.20 ± 0.62, respectively, p <0.0001) or ES (0.73 ± 0.36 vs 0.53 ± 0.27, p <0.0001; Figure 1 ). T/D ratio medians and upper limits of normal (90th percentile) are reported in Supplementary Table 4 .

Stay updated, free articles. Join our Telegram channel

Join