Differences in Cardiac Mechanics among Genetically At-Risk First-Degree Relatives: The DCM Precision Medicine Study

Abstract

Aims: Among genetically at-risk first-degree relatives (FDRs) of probands with dilated cardiomyopathy (DCM), the ability to detect changes in left ventricular (LV) mechanics with normal LV size and ejection fraction (LVEF) remains incompletely explored. We sought to define a pre-DCM phenotype among at-risk FDRs, including those with variants of uncertain significance (VUSs), using echocardiographic measures of cardiac mechanics.

Methods and results: LV structure and function, including speckle-tracking analysis for LV global longitudinal strain (GLS), were evaluated in 124 FDRs (65% female; median age 44.9 [IQR: 30.6-60.3] years) of 66 DCM probands of European ancestry sequenced for rare variants in 35 DCM genes. FDRs had normal LV size and LVEF. Negative FDRs of probands with pathogenic or likely pathogenic (P/LP) variants (n=28) were a reference group to which negative FDRs of probands without P/LP variants (n=30), FDRs with only VUSs (n=27), and FDRs with P/LP variants (n=39) were compared. In an analysis accounting for age-dependent penetrance, FDRs below the median age showed minimal differences in LV GLS across groups while those above it with P/LP variants or VUSs had lower absolute values than the reference group (-3.9 [95% CI: -5.7, -2.1] or -3.1 [-4.8, -1.4] %-units) and negative FDRs of probands without P/LP variants (-2.6 [-4.0, -1.2] or -1.8 [-3.1, -0.6]).

Conclusions: Older FDRs with normal LV size and LVEF who harbored P/LP variants or VUSs had lower absolute LV GLS values, indicating that some DCM-related VUSs are clinically relevant. LV GLS may have utility for defining a pre-DCM phenotype.

Clinical trial registration: clinicaltrials.gov, NCT03037632.

Keywords: cardiomyopathy; echocardiography; family-based study; genetics.

Figure 1.. A conceptual framework for a pre-DCM phenotype.

Identifying pre-DCM in at-risk individuals using LV GLS provides an opportunity to test interventions to prevent DCM at an earlier stage. Pre-DCM^a is defined as a reduction in the absolute value of LV GLS in a genetically at-risk first-degree relative who still has normal LV size and ejection fraction. Partial DCM^b has been defined as an LVEF <50% or LV enlargement (but not both),^, and has been shown to be a harbinger of DCM. The transition from genetic risk of DCM with normal cardiac structure and function to pre-DCM, partial DCM, clinically identifiable though still asymptomatic DCM (Phase 1), and finally symptomatic DCM (Phase 2) takes months or years. ACC/AHA stages of heart failure aligned with the progression of DCM are shown at the bottom. HF = heart failure.

Figure 2.. Selection of study participants and definition of genetic risk groups.

FDRs of probands who had completed genetic analysis were selected from a sampling frame including FDRs meeting all specified criteria as of October 20, 2021. To ensure adequate representation in each genetic risk group, approximately 50 FDRs were selected from each of three strata defined by the most deleterious variant harbored by the FDR (none, VUS, or P/LP), with preference given to unrelated FDRs. To expand the sample, all remaining FDRs in the sampling frame belonging to families of FDRs selected in the initial round were added without regard to variants harbored. For analysis, FDRs <18 years of age were excluded because DCM is an adult-onset disease with age-dependent penetrance, as were FDRs of probands with less than 87.5% European ancestry, who were not represented in sufficient numbers in the reference group described below to permit statistical adjustment of group comparisons by ancestry in the full sample. FDRs were analyzed in groups with varying levels of genetic risk, with negative FDRs of probands with P/LP variants serving as the reference group with DCM risk no greater than the general population.

Figure 3.. Echocardiographic measurements by genetic risk group in FDRs above the median age in the sample.

A violin plot with a superimposed box-and-whisker plot shows the distribution of the measurements among FDRs in each genetic risk group, with a black diamond at the mean. Next to this, an interval plot shows the estimated marginal mean from the linear mixed model analysis (point) as well as its 95% confidence interval (interval) obtained using Morel-Bokossa-Neerchal bias-corrected empirical standard errors and the standard normal distribution. For each genetic risk group, the estimated marginal mean is a covariate-adjusted estimate of the mean in a population of FDRs above the median age in the sample (44.9 years) that is half female. For LV global longitudinal strain, these populations also have the same mean height and weight and half image quality >2. Table 3 presents covariate-adjusted estimated mean differences between each genetic risk group and the reference group for FDRs above the median age from the same model; Supplemental Table 3 presents these differences comparing the P/LP and VUS groups to negative FDRs of probands without P/LP variants. In both tables, reported differences are the differences between the estimated marginal means in this figure.