Shared Genetic and Environmental Architecture of Cardiac Phenotypes Assessed via Echocardiography

Echocardiographic measures of cardiac structure and function represent important predictors of heart failure.¹ Evidence suggests that both genetic and environmental factors contribute to phenotypic variation of echocardiographic traits.² Identifying these factors could facilitate targeted interventions for optimizing individual risk profiles. However, the degree to which key cardiac traits are influenced by genetic versus environmental exposures, and the interaction between these exposures, remains unknown. Our objective was to investigate the heritability of multiple echocardiographic traits and their environmental correlations.

We studied Framingham Heart Study attendants of Offspring Cohort exam 8 (2005–2008) or Third Generation Cohort exam 1 (2002–2005) with standardized echocardiography performed per guidelines.³ The Third Generation Cohort included children of the Offspring Cohort. Individuals with heart failure or ≥1 missing echocardiographic measure were excluded, providing N=5674 participants (mean age 49 years, 54% women). We selected 5 cardiac structural and 6 functional traits for heritability and paired analyses based on their representing key measures that have been previously related to outcomes and their being reliably and reproducibly measured in our and other core laboratories. The structural traits included: left ventricle (LV) mass index, diastolic dimension, LV wall thickness (LVWT), left atrial systolic dimension, and aortic root diameter. The functional traits included: LV ejection fraction (LVEF), global circumferential and longitudinal strain (GCS, global longitudinal strain), longitudinal synchrony (LSS), MAPSE, and E/e′. All the echocardiographic measures have been deposited in the dbGaP database (https://www.ncbi.nlm.nih.gov/projects/gap/cgi-bin/study.cgi?study_id=phs000007.v32.p13). The Framingham Heart Study was approved by the Institutional Review Boards of Boston University Medical Center, and all participants provided written informed consent.

For all analyses, LV mass index, LSS, and E/e′ were natural log-transformed. For each echocardiographic trait, the associations with age, sex, and height were adjusted for in linear regression models; residuals of regression models were then transformed using a rank-based inverse normal transformation. The normalized values (adjusted for age, sex, height) were used for heritability analyses. Singular and bivariate quantitative trait linkage analysis was performed using Sequential Oligogenic Linkage Analysis Routines,⁴ which estimates additive genetic heritability h² using pedigree information (spousal correlation was considered 0 in the genetic correlation). No shared versus unshared environmental conditions were estimated. Phenotypic correlations among all pairs of traits were calculated and partitioned into genetic and environmental correlations using the variance components method.^4,5

We observed single trait heritabilities ranging from 0.27 (LVWT) to 0.42 (aortic root diameter) for cardiac structural traits and 0.19 (MAPSE) to 0.41 (log[LSS]) for functional traits (P<0.0001 for all; Figure). Heritability of structural traits tended to be higher than for functional traits. Results were unchanged after adjusting for systolic blood pressure and body mass index for traits associated with these characteristics (LV mass index, LV diastolic dimension, and LVWT). Phenotypic correlations between traits ranged from −0.44 (LVEF and GCS) to 0.68 (LV mass index and LVWT). We used bivariate quantitative analysis to examine genetic and environmental correlation between trait pairs. We observed moderate-to-strong genetic correlations among several traits: 0.61 (LVWT with left atrial systolic dimension), and −0.44 (global longitudinal strain with LVEF) to −0.79 (GCS with LVEF). Shared genetic correlation due to pleiotropy ranged from 0.0008 (Log [LSS] and MAPSE) to 0.632 (LVEF and GCS). Modest environmental correlations emerged for other traits: 0.26 (LV diastolic dimension with left atrial systolic dimension), and −0.32 (GCS with LVEF). There were also unmatched pairs with significant genetic but no environmental correlation, or vice versa.

Figure.

Heritability estimates and genetic and environmental correlations among echocardiographic traits. Upper diagonal values are genetic correlations; lower diagonal values are environmental correlations. AoD indicates aortic root diameter; GCS, global circumferential strain; GLS, global longitudinal strain; LAD, left atrial systolic dimension; LSS, longitudinal synchrony; LV, left ventricle/left ventricular; LVDD, LV diastolic dimension; LVEF, LV ejection fraction; LVMi, LV mass index; LVWT, LV wall thickness; and MAPSE, mitral annular plane systolic excursion. All models adjusted for age, sex, and height. Values in bold indicate significance: *P<0.05; †P≤ 0.01; ‡P≤0.001; §P<0.0001.

Overall, we observed that cardiac structural traits tended to demonstrate greater heritability than cardiac functional traits. Prior cardiac phenotyping studies have reported a wide range of heritability estimates, including ranges for LV mass from 0.24 in cohort-based studies to 0.84 in twin studies. We observed particularly strong genetic and weak environmental correlations between LVWT and left atrial systolic dimension, suggesting that genetic factors play a more important role than environmental factors in parallel left ventricular-atrial remodeling—a complex endophenotype that may be especially susceptible to progression towards heart failure in the setting of certain risk exposures. Extending from prior work, we analyzed advanced measures of cardiac function that have been relatively understudied previously. We observed moderate heritability for several of these traits (eg, LSS, GCS), suggesting that novel genetic loci may be identified for these traits in future investigations. Interestingly, for GCS, we found both strong genetic and strong environmental correlation with LVEF, suggesting that these cardiac functional traits have mechanistic features derived from shared genetic origins in addition to similar patterns of pathophysiologic response to risk exposures.

Our analyses were restricted to two related cohorts comprising predominantly individuals of European ancestry and without independent replication. Notwithstanding these limitations, our study involved careful and standardized phenotyping performed in a core laboratory with established quality control and included more than twice as many participants than prior studies, providing statistical power to identify cardiac traits with even modest heritability.

In our large community-based sample, we observed moderate heritability for several key cardiac traits including multiple structural and functional phenotypes. Shared genetic influences among cardiac traits suggest common genetic programming, while shared environmental correlates may reflect conjoint cardiac responses to common risk exposures as well as chronic hemodynamic loading conditions.

Sources of Funding

The study was supported by National Institutes of Health (NIH) contracts 75N92019D00031, HHSN268201500001I, and N01-HC 25195 to the Framingham Heart Study. The study was also supported by NIH grants R01HL093328, R01HL107385, R01HL126136, R01HL143227, and R01HL142983. R.S. Vasan is also supported in part by the Evans Medical Foundation and the Jay and Louis Coffman Endowment from the Department of Medicine, Boston University School of Medicine. C. Castro-Diehl was supported by the Multidisciplinary Training Program (T32) in Cardiovascular Epidemiology (5T32HL125232). A.C. Kwan is supported by the Doris Duke Charitable Foundation Grant 2020059. We acknowledge the dedication of the Framingham Heart Study participants without whom this research would not be possible.

Disclosures

None.