Heritability of Left Ventricular Mass
The Framingham Heart Study
Abstract Left ventricular hypertrophy is associated with an increased risk for cardiovascular disease. The known determinants of left ventricular hypertrophy only partially explain its variability. The purpose of this study was to estimate heritability of left ventricular mass. The study sample included adults in the original Framingham Heart Study and the Framingham Offspring Study who were not receiving antihypertensive medications and who were free of coronary heart disease, congestive heart failure, diabetes mellitus, renal insufficiency, valvular heart disease, and severe left ventricular hypertrophy. Intraclass correlations for left ventricular mass among first-degree relatives, second-degree relatives, and unrelated spouse pairs were calculated to determine the contribution of heredity to the variability in left ventricular mass. After adjustments for age, height, weight, and systolic blood pressure, the intraclass correlations between first-degree relatives were .15 (parent-child, P<.001) to .16 (siblings, P<.001), between second-degree relatives the correlation was .06 (P=NS), and between spouses it was .05 (P=NS). The estimated heritability of adjusted left ventricular mass was between .24 and .32. The proportion of the variance in sex-specific left ventricular mass explained by age, height, weight, and systolic blood pressure was .26 in men and .34 in women. On the basis of intraclass correlations for left ventricular mass, incorporation of adjusted left ventricular mass of a parent or sibling would increase the explained variance by an additional .02 to .03. Heredity explains a small, but discernible proportion of the variance in left ventricular mass. Studies are currently under way to identify genetic markers that predict an individual’s predisposition to left ventricular hypertrophy. This knowledge may lead to advances in the prevention of left ventricular hypertrophy, which is strongly associated with cardiovascular morbidity and mortality.
Left ventricular hypertrophy is associated with an increased risk for cardiovascular disease morbidity and mortality.1 2 3 Elevated systolic blood pressure, obesity, and advanced age are predictors of left ventricular hypertrophy.4 5 6 The known determinants of left ventricular mass only partially explain its variability in the population.
Epidemiological studies in twins suggest that left ventricular hypertrophy may be influenced by genetic factors, in addition to biological variables that are known to influence left ventricular mass.7 8 Efforts are under way to elucidate possible genetic mechanisms responsible for heritability of left ventricular mass. One candidate gene has been the angiotensin-converting enzyme (ACE) gene. Although an association between the deletion-insertion polymorphism for the ACE gene and the prevalence of left ventricular hypertrophy on the electrocardiogram9 has been reported, we recently found no evidence of an association between the ACE gene polymorphism and echocardiographic left ventricular mass in the Framingham Heart Study cohort.10
The purpose of this investigation was to estimate heritability of left ventricular mass among subjects in the Framingham Heart Study. Recognition of the genetic determinants of left ventricular mass may provide insight into the pathogenesis of left ventricular hypertrophy and may offer clues to its prevention.
The Framingham Heart Study is a prospective epidemiological study that was established to evaluate potential risk factors for coronary heart disease. The original cohort included 5209 residents of Framingham, Massachusetts, who were between the ages of 28 and 62 years. The study design and selection criteria have been reported.11 12 13 In 1971, another 5124 subjects, offspring of original participants and spouses of offspring, were entered into the Framingham Offspring Study.14
The present sample was derived (see Table 1⇓) from original Framingham Heart Study subjects who attended biennial examination sixteen (1979-1981) and Offspring Study subjects who attended their second examination (1979-1983). Subjects were excluded if they met any of the following criteria: (1) current use of antihypertensive medications (n=1269), (2) coronary heart disease or congestive heart failure (n=315), (3) diabetes mellitus (n=184), (4) renal insufficiency (n=7), (5) valvular heart disease (n=70), (6) age <20 years (n=12), (7) suboptimal echocardiogram for assessment of left ventricular mass (n=561), and (8) extreme value for left ventricular mass (n=4) (since as extreme outliers, they would have undue greater influence in the statistical analyses).
To minimize misclassification, original cohort participants were asked about the biological status of their offspring, and many adoptees were recognized at that time and excluded from genetic analyses. Body height and weight measurements, medical history, physical examination, serum creatinine, blood glucose level, and echocardiography were obtained at the routine examination. Resting systolic blood pressure was measured twice by a physician during the clinic examination. The mean of the two measurements was used to determine systolic blood pressure. The diagnoses of coronary heart disease and congestive heart failure were established in accordance with previously published criteria15 after review by a committee of three physicians who evaluated records from the Framingham Heart Study clinic, interim hospitalizations, and outside physicians. Diabetes was defined as a fasting blood glucose level >7.77 mmol/L (140 mg/dL), a random nonfasting blood glucose level >11.11 mmol/L (200 mg/dL), or the use of insulin or an oral hypoglycemic agent. Renal insufficiency was defined as a serum creatinine concentration >2.0 mg/dL. The diagnosis of valvular heart disease was based on evidence from physical examination of a systolic murmur of grade 3 or higher intensity on a six-point scale or any diastolic murmur.
Standard M-mode echocardiographic techniques were used in accordance with methods outlined by the American Society of Echocardiography.16 Measurements of left ventricular wall thickness and chamber diameter were made in diastole. Left ventricular mass was estimated by the modified cubed formula using measurements obtained in accordance with the “Penn” convention17 : Left Ventricular Mass (g)=1.04 [(LVID+ VST+PWT)3–(LVID)3]–13.6, where LVID is left ventricular internal diameter, VST is ventricular septal thickness, and PWT is posterior left ventricular wall thickness.
Preliminary analyses used all subjects who met entry criteria, regardless of whether they had relatives in the sample. Several transformations of left ventricular mass were studied to normalize the distributions and equalize the variances in male and female subjects. The natural logarithmic transformation proved best for these tasks, so all further analyses were done on log-transformed data for left ventricular mass, along with log-transformed values of continuous predictor variables. Several linear regression models18 19 were fitted for log left ventricular mass, separately for men and women, to account for sex differences, age, height, weight, and systolic blood pressure, adding variables in the order listed. Residuals from each fitted model were used in subsequent analyses.
To analyze genetic contributions to left ventricular mass, we performed separate analyses on first-degree relatives (parent-child pairs and sibship members), on second-degree relatives (aunt/uncle and niece/nephew pairs), and also on unrelated subjects (spouse pairs). For each defined relationship, similarity of left ventricular mass residuals among the related members was analyzed by the intraclass correlation method18 to accommodate different numbers of subjects in different families. Estimation and testing were carried out with SAS procedures, NESTED and GLM.19 Because we performed multiple testing, we used P<.01 as the criterion for statistical significance. We included spouse pairs, rather than unrelated individuals living apart, as our comparison group. The purpose was to highlight changes in intraclass correlation as one accounts for variables such as age, height, and weight, which tend to be similar both in spouse pairs and sibling pairs.
Estimates of heritability were derived from the intraclass correlation coefficient, r, as h2=2×r (sibling, sibling), h2=2×r (parent, child), and h2=4×r (aunt/uncle, niece/nephew) to account for the average proportion of genes shared by pair members.20
There were 2805 men and 3413 women who attended the index examination. The eligible sample used for the regression analyses included 1669 men and 2127 women, who did not meet any exclusion criteria (Table 1⇑). The ineligible subjects were older, were heavier (women), had higher blood pressure, and had greater body mass index and left ventricular mass. Among eligible subjects, those who had eligible first- or second-degree relatives or a spouse were used for heritability analyses (n=2624). The subjects included in the genetic analyses were similar to the total eligible cohort with regard to the clinical variables measured.
The means and unadjusted correlations of age, height, weight, and systolic blood pressures of the subjects grouped into sibling pairs, parent-child pairs, aunt/uncle-niece/nephew pairs, and spouses are presented in Table 2⇓. The sibling pairs had similar values for these variables and moderate correlations. The parent-child pairs and aunt/uncle-niece/nephew pairs represented different generations, and differed with regard to age, height, weight, and systolic blood pressure. Their correlations were smaller than those for siblings. The spouse pairs were of similar ages and had higher unadjusted correlations than parent-child pairs or aunt/uncle-niece/nephew pairs for age, height, weight, and systolic blood pressure.
Intraclass correlations for left ventricular mass among first-degree relatives, second-degree relatives, and unrelated spouse pairs were calculated to assess the contribution of heredity to the variability in left ventricular mass. These are presented in Table 3⇓. The intraclass correlation is a measure of the degree of similarity in left ventricular mass between subjects. The intraclass correlations for sex-specific left ventricular mass (model 1) were highest for siblings (r=.24), intermediate for spouses (r=.11), and lowest for second-degree relatives (r<.01). After adjustments for age, height, weight, and systolic blood pressure (model 5), the intraclass correlations between first-degree relatives were .15 (parent-child, P<.001) to .16 (siblings, P<.001), between second-degree relatives the correlation was .06 (P=NS), and between spouses it was .05 (P=NS).
For siblings and for spouses, intraclass correlations decreased after adjustments for known determinants of left ventricular mass. Likeness of age and clinical variables among siblings and spouses inflated their similarity with regard to left ventricular mass; when these variables were accounted for, the intraclass correlations declined. In contrast, for parent-child pairs and for second-degree relatives, intraclass correlations increased after adjustments; differences in age and clinical variables likely masked similarities with regard to left ventricular mass.
The proportion of the variance in sex-specific left ventricular mass explained by age, height, weight, and systolic blood pressure was .26 in men and .34 in women (Table 4⇓). Weight was the most important single predictor of left ventricular mass. Including systolic blood pressure in the regression equation only increased the r2 by .01, after age and weight had been added to the model. On the basis of intraclass correlations for left ventricular mass, incorporation of adjusted left ventricular mass of a parent or sibling would increase the explained variance by an additional .02 to .03 (based on model 5 intraclass correlations squared).
On the basis of results from model 5, we estimate heritability of adjusted left ventricular mass as being between .24 (estimate from aunt/uncle-niece/nephew correlations) and .32 (sibling-sibling correlations), with an intermediate estimate of .30 from parent-child data.
A supplemental analysis was completed in which subjects receiving antihypertensive therapy (217 men, 320 women) were included. There were no appreciable differences in the intraclass correlations for left ventricular mass, compared with the original analysis (Table 3⇑), except in model 5 (including adjustment for antihypertensive therapy) the intraclass correlation for parent-child subjects decreased to .12. In addition, the proportion of the variance in left ventricular mass that was explained by the predictors (Table 4⇑) increased to .27 in men and to .37 in women when antihypertensive therapy was included in the model.
Heredity accounted for a small, but discernible proportion of the overall variance in left ventricular mass. The intraclass correlations for left ventricular mass between first-degree relatives were .15 to .16 after accounting for sex, age, height, weight, and systolic blood pressure. We estimate heritability of adjusted left ventricular mass as being between .24 and .32. No significant intraclass correlation was found for second-degree relatives or for unrelated spouse pairs after adjusting for these variables, which accounted for 26% (men) to 34% (women) of left ventricular mass variability. Incorporation of a first-degree relative’s left ventricular mass accounted for an additional 2% to 3% of the variability.
The results of this study are consistent with the findings of Harshfield et al7 who found a greater left ventricular mass index intraclass correlation between black monozygotic twins than between black dizygotic twins, after adjusting for sex, systolic blood pressure, and age (.90 versus .33). The higher intraclass correlation between dizygotic twins in that study than between first-degree relatives in our study may be due to a greater degree of environmental similarity influencing left ventricular mass in dizygotic twins than in non-twin siblings or parent-child pairs, possible racial differences in the heritability of left ventricular mass, and differences in sample size.
The heritable contribution to left ventricular mass variance of 2% to 3% is small; however, in some subjects the genetic contribution may be large, whereas in others it may be negligible. In addition, the genetic contribution was greater than that of systolic blood pressure. The systolic blood pressure only increased the r2 by 1% after age, height, and weight were included in the model.
A limitation of this study is that participants in the Framingham Heart Study are predominantly Caucasian. Since left ventricular hypertrophy may be more prevalent among African-Americans,21 it would be important to determine heritability of left ventricular mass in that population. Another limitation is the possibility that the heritability of left ventricular mass is overestimated due to the inability to control for unknown environmental factors that may influence left ventricular mass. In addition, unrelated spouse pairs who share recent environment may not be the optimal comparison group for related siblings who shared early environment. A strength of this study is the large number of subjects included, which allows for greater control for confounding variables and greater precision in the results.
Heredity explains a small, but discernible proportion of the variance in left ventricular mass. Studies are currently under way to identify genetic markers that predict an individual’s predisposition to left ventricular hypertrophy. This knowledge may lead to advances in the prevention of left ventricular hypertrophy, which is strongly associated with cardiovascular morbidity and mortality.
Reprint requests to Daniel Levy, MD, Framingham Heart Study, National Heart, Lung, and Blood Institute, 5 Thurber St, Framingham, MA 01701.
- Received July 11, 1996.
- Revision received September 11, 1996.
- Accepted June 5, 1997.
Koren MJ, Devereux RB, Casale PN, Savage DD, Laragh JH. Relation of left ventricular mass and geometry to morbidity and mortality in uncomplicated essential hypertension. Ann Intern Med. 1991;114:345-352.
Ghali JK, Liao Y, Simmons B, Castaner A, Cao G, Cooper RS. The prognostic role of left ventricular hypertrophy in patients with or without coronary artery disease. Ann Intern Med. 1992;117:831-836.
Levy D, Anderson KM, Savage DD, Kannel WB, Christiansen JC, Castelli WP. Echocardiographically detected left ventricular hypertrophy: prevalence and risk factors: the Framingham Heart Study. Ann Intern Med. 1988;108:7-13.
Harshfield GA, Grim CE, Hwang C, Savage DD, Anderson SJ. Genetic and environmental influences on echocardiographically determined left ventricular mass in black twins. Am J Hypertens. 1990;3:538-543.
Lindpaintner K, Lee M, Larson MG, Rao VS, Pfeffer MA, Ordovas JM, Schaefer EJ, Wilson AF, Wilson PWF, Vasan RS, Myers RH, Levy D. Absence of association or genetic linkage between the angiotensin-converting-enzyme gene and left ventricular mass. N Engl J Med. 1996;334:1023-1028.
Dawber TR, Meadors GF, Moore FE. Epidemiologic approaches to heart disease: the Framingham study. Am J Public Health. 1951;41:279-286.
Dawber TR, Kannel WB, Lyell LP. An approach to longitudinal studies in a community: the Framingham study. Ann N Y Acad Sci. 1963;107:539-556.
Kannel WB, Feinleib M, McNamara PM, Garrison RJ, Castelli WP. An investigation of coronary heart disease in families: the Framingham offspring study. Am J Epidemiol. 1979;110:281-290.
Sorlie P. Cardiovascular disease and death following myocardial infarction and angina pectoris: Framingham Study, 20 year follow-up. Section 32. DHEW publication No. (NIH) 77-1247. Washington, DC: Government Printing Office, 1977.
Sahn DJ, DeMaria A, Kisslo J, Weyman A. Recommendations regarding quantitation in M-mode echocardiography: results of a survey of echocardiographic measurements. Circulation. 1978;58:1072-1083.
Zar JH. Biostatistical Analysis. 2nd edition. Englewood Cliffs, NJ: Prentice-Hall; 1984:323-325, 328-360.
SAS/STAT User’s Guide. Version 6, 4th edition, Vol 2. Cary, NC: SAS Institute Inc; 1989:846, 891-996, 1127-1134, 1351-1456.
Cavalli-Sforza LL, Bodmer WF. The Genetics of Human Populations. San Francisco, Calif: Freeman; 1971:523-537.