Sibling Resemblance for Left Ventricular Structure, Contractility, and Diastolic Filling
Although there is evidence that left ventricular (LV) function is genetically controlled, the contribution of familial factors to variation and covariation of LV diastolic filling, contractility, and structure is unknown. Single- and cross-trait sibling correlations were estimated using bivariate familial correlation models in 200 white (400 pairs) and 374 black (539 pairs) hypertensive sibships. LV transmitral early and late peak filling velocities, isovolumic relaxation time, atrial filling fraction, stress-corrected midwall shortening, and LV mass and structure were measured and adjusted for important covariates in race-specific linear regression models. Single-trait sibling correlation was strongest for early peak filling velocity. Significant cross-trait sibling correlation was detected between early and late peak filling velocities. In whites, early peak filling velocity and atrial filling fraction, and isovolumic relaxation time and end-diastolic posterior wall thickness, were also significantly correlated. Familial factors common to early and late peak filling velocities contributed to 64% and 54% of sibling resemblance in early peak filling velocity and to 76% and 77% in late peak filling velocity in blacks and whites, respectively. In whites, 100% of sibling resemblance in isovolumic relaxation time was shared by posterior wall thickness, whereas 75% of sibling influence in posterior wall thickness was common to isovolumic relaxation time. In conclusion, significant cross-trait sibling resemblance was detected between (1) early and late filling parameters and (2) isovolumic relaxation time and posterior wall thickness, suggesting pleiotropy and/or common environment on these traits. These data have potential importance in understanding heritability of LV diastolic function in hypertension.
Impaired diastolic left ventricular (LV) function occurs prominently in a variety of common cardiovascular disorders, including hypertension, LV hypertrophy, and coronary artery disease,1 and it is the presumed mechanism of ≈50% of congestive heart failure cases.2–5 Doppler echocardiography provides a noninvasive assessment of diastolic function by measuring diastolic filling.1 LV transmitral early peak filling velocity (MVE) and isovolumic relaxation time (IVRT) reflect early diastolic filling, and LV transmitral late peak filling velocity (MVA), and atrial filling fraction (AFF) reflect late filling. These LV filling parameters correlate with each other and also with LV structure and systolic functional measurements, such as midwall shortening,6,7 suggesting common physiological processes, shared environment, and/or genetic influences.
Evidence from studies in animals and humans suggest variation in LV mass and diastolic function is under genetic control.8–10 In a study of twins aged 18 to 31 years, genetic factors accounted for 43% and 26% of phenotypic variation in MVE and MVA, respectively.11 We previously reported age- and age2-adjusted broad-sense heritabilities12 of diastolic filling of 0.50 for MVE and 0.52 for MVA in black and white hypertensive sibships in the Hypertension Genetic Epidemiology Network Study (HyperGEN).13 In the present study, we extend our prior analysis by investigating single-trait and cross-trait sibling correlations among diastolic filling parameters, including MVE, MVA, IVRT, and AFF, and between diastolic filling parameters and stress-corrected midwall shortening (MWS), LV mass, LV diastolic internal dimension (LVID), posterior wall thickness (PWT), and interventricular septal thickness (IVS), taking into account the influences of demographic, hemodynamic, and environmental factors in the same population. Significant cross-trait sibling correlation suggests pleiotropic genetic and/or familial environmental factors influence the correlation between 2 traits.14 Using this approach, we identified pairs of cardiac structural and functional variables sharing a common familial influence.
The study was approved by the institutional review board at each field center, and all study participants gave informed written consent before participation. The recruitment methods and inclusion criteria in the HyperGEN study, which aims to identify the genetic contributions to hypertension, have been described.15 Hypertension was defined as average systolic blood pressure (SBP) ≥140 mm Hg or average diastolic blood pressure (DBP) ≥90 mm Hg on at least 2 different evaluations, or treatment for hypertension.15 The average SBP and DBP, respectively, were calculated based on the second and third measurements of 3 readings from a selected arm according to a standard blood pressure measurement protocol and using an oscillometric blood pressure instrument (Dinamap 1846 SX/P).15 The Genetics of Left Ventricular Hypertrophy Study, an ancillary study to HyperGEN, performed echocardiography in 4 centers in HyperGEN (Minneapolis, Minnesota; Salt Lake City, Utah; Forsyth County, North Carolina; and Birmingham, Alabama). Sixty-seven subjects whose diastolic filling pattern was classified as restrictive filling (mitral early deceleration time ≤140 ms) were excluded. Included in this analysis are 574 full sibships ranging in size from 2 to 7 members (374 black and 200 white sibships). The full-sibling relation was confirmed by analysis of 387 anonymous markers.
2D-guided M-mode, 2D, and Doppler echocardiograms were performed following a standardized protocol as previously described.7,16,17 LV diastolic filling parameters were measured by pulsed Doppler, with the sample volume placed at the mitral valve leaflet tips and mitral valve annulus.17
Measurements were made at the centralized echo-core laboratory as previously described.17 Because 60% of the subjects had diastolic filling measurements available at the leaflet tips, and the rest had measurements available at the annulus, previously published formulas6 were used to calculate diastolic filling parameters (MVE, MVA, and AFF) at the leaflet tips from the annulus values in the 40% of participants without leaflet tip measurements.
Systolic function was assessed by MWS, and LV mass and stroke volume (Doppler method) were calculated as previously described.18,19 Reproducibility of LV measurements by methods similar to those used in HyperGEN has been reported by the reading center (eg, intraclass correlation coefficients: 0.93 for LV mass, 0.71 for MWS, 0.58 for MVE, 0.57 for MVA).20
Each of the 9 traits was adjusted for the effects of age, age2, gender, heart rate, weight, height, stroke volume, SBP, presence of coronary artery disease and/or diabetes, field center, and number and type of antihypertensive medication classes in race-specific linear regression models. In addition, LV mass was added to the regression models for the functional variables (MVE, MVA, IVRT, AFF, and MWS), and sibling correlations based on this adjustment were compared with previous models to test the influence of LV mass on sibling resemblance of LV function.
The familial correlation model was constructed as follows. For each pair of traits, we calculated 12 correlation parameters, including 6 single-trait sibling correlations (b1b1, b1s1, and s1s1 for the first trait; b2b2, b2s2, and s2s2 for the second trait), 4 cross-trait sibling correlations (b1b2, b1s2, s1s2, b2s1), and 2 intraindividual cross-trait correlations (b12 and s12), where b denotes brother, s denotes sister, and the numbers 1 or 2 refer to the trait number. The calculation was performed using SEGPATH, a maximum likelihood-based computer program.21 Significance of a particular correlation parameter and equality of several correlation coefficients were tested by a likelihood ratio test.
Hypothesis testing started with a general model, in which all parameters were estimated, followed by tests for significance of univariate and cross-trait sibling correlations. If the correlations were at least marginally significant, tests of gender difference and significance of intraindividual sibling correlations proceeded in a series of submodels. The most parsimonious model, which contained the least number of parameters and fit the data as well as the full model, was derived from combining all nonrejected submodels (P>0.05). Because of the large number of pairs analyzed, P<0.01 was considered significant and P<0.05 marginally significant for a given cross-trait correlation. A summary of the hypothesis tests has been filed in an online data supplement available at http://www.hypertensionaha.org.
Table 1 shows the descriptive statistics for echocardiographic measures and covariates by race and gender groups. In this hypertensive sample, SBP, LV mass indexed to body surface area (LVMI), MVE, and MVA were significantly higher in blacks than in whites of the same gender (P<0.05). The ratio of MVE to MVA, a commonly used index of LV relaxation, was higher in blacks than in whites, so that there were more whites classified as having impaired relaxation compared with blacks in each gender group (P<0.05). However, after adjustment for age, the ethnicity-related difference in MVE/MVA or percentage of impaired relaxation was no longer significant, whereas the association between race, SBP, and LVMI persisted.
Table 2 gives the results for the hypothesis testing for single-trait sibling correlations of LV mass, structure, and functional parameters. The single-trait sibling correlation was significant for MVE, MVA, IVRT, AFF, LV mass, LVID, PWT, and IVS (P<0.05). There was no gender difference in the correlations in either race group (except for AFF, which was higher in the sister-sister group than the brother-brother or brother-sister groups in blacks) (Table 3). For MWS, the hypothesis of no sibling correlation was rejected in blacks but not in whites. The univariate sibling correlations (±SE) from both the general model and the most parsimonious model are presented in Table 3. In both race groups, the sibling resemblance of functional parameters was highest for MVE and MVA, intermediate for IVRT and AFF, and least for MWS.
Hypothesis tests for cross-trait correlations that reached at least marginal significance (P<0.05) are summarized in Table 4. In whites, there were significant cross-trait sibling correlations between MVE and MVA, MVE and AFF, and IVRT and PWT (P<0.01). In blacks, the only cross-trait pair showing significant evidence of common familial resemblance was MVE/MVA (P<0.001). All the significant cross-trait sibling correlations were equal across gender groups. For MVE and MVA, the most parsimonious model was no gender difference in interindividual cross-trait correlations in whites, and no gender difference in both interindividual and intraindividual cross-trait correlations in blacks. For MVE/AFF, IVRT/MWS, and IVRT/PWT in whites, the model of “no gender difference in both interindividual and intraindividual cross-trait correlations” fitted the data best. Cross-trait sibling correlations for other trait pairs were not significant (data not shown).
Parameter estimates (±SE) for cross-trait correlations of at least marginal significance are presented in Table 5. In both race groups, MVE exhibits moderate cross-trait sibling resemblance with MVA. According to single-trait and cross-trait sibling correlations, 54.8% and 64.0% of sibling resemblance in MVE was shared by MVA in whites (0.17/0.31) and blacks (0.16/0.25), respectively, and 77.3% and 76.2% of the sibling resemblance in MVA was shared by MVE in whites (0.17/0.22) and blacks (0.16/0.21), respectively. In whites, IVRT and PWT also exhibited modest cross-trait sibling correlation, accounting for 100% of sibling resemblance in IVRT (0.18/0.18) and 75% of sibling resemblance in PWT (0.18/0.24).
After LV mass was added to the regression adjustment models for functional measures (MVE, MVA, IVRT, AFF, and MWS), the most parsimonious models did not change for most of the single- and cross-trait correlations, except that correlations for MWS (blacks) and IVRT/MWS (whites) decreased to below significance level (P>0.05). There were no appreciable changes in single- and cross-trait correlations for other traits (data not shown).
The present study is the first to our knowledge to examine cross-trait familial relationships in LV diastolic filling indices, contractility, LVID, PWT, IVS, and LV mass in humans. After adjusting for a variety of factors known to influence LV structure and function, evidence of significant cross-trait familial resemblance was detected among early LV filling, MVE, and late filling indices (MVA and AFF) and between IVRT and PWT. In addition, modest familial resemblance was found for IVRT/MWS, MVE/LV mass, and MVE with other structure phenotypes. The significant sibling correlation between MVE and MVA was consistent across race groups, explaining >50% of familial resemblance in MVE and MVA. If it is assumed that all the sibling resemblance is caused by genetic factors, single- and cross-trait broad-sense heritability can be calculated by doubling the sibling correlations.12 Based on this calculation, as much as 28% of phenotypic variation and 54% to 77% of genetic variation in each trait (MVE and MVA) were owing to shared genes.
The significant and positive cross-trait sibling correlation between MVE and MVA suggests alterations in early and late diastolic filling are determined by a common familial component. The primary physiological determinant of early filling is myocardial relaxation, whereas it is effective chamber compliance that influences late filling. Impairment of early diastolic active relaxation or compliance causes reciprocal changes in MVE and MVA, with diminished early filling and augmented late filling in uncomplicated hypertension.1 However, our modeling indicates that familial factors common to MVE and MVA are independent of the loading force modulating this LV filling phenotype, and are not mediated by the influence of LV hypertrophy. Although one might expect stroke volume could account for the positive intra- and interindividual correlations of MVE with MVA, our adjustment for the effects of anthropometric, prevalent disease, and hemodynamic covariates, including stroke volume, made these factors unlikely as a source of the sibling covariation between these 2 traits. Alternatively, MVE and MVA could vary reciprocally with mitral annular/valve size, which does not seem to explain the positive association here. Therefore, it is plausible to posit that the familial factors may act through impaired myocardial relaxation and chamber compliance to simultaneously influence MVE and MVA in hypertension, although variation owing to uncontrolled factors is also an interpretation.
The cross-trait sibling correlations of IVRT with PWT and MWS are of clinical relevance because IVRT is an index of impaired diastolic relaxation. Bella et al7 reported that prolonged IVRT was associated with lower MWS and with higher LV mass and PWT in the same population. The present study extends their finding by showing that there are shared familial environments and/or pleiotropic genes that contribute to the association between impaired diastolic relaxation, subnormal myocardial function, and PWT. The correlation between IVRT and MWS became insignificant after LV mass was adjusted for, suggesting LV hypertrophy is the major factor mediating the familial resemblance between LV diastolic relaxation and contractility.
The lack of comparable evidence for some cross-trait sibling resemblance (MVE/AFF and IVRT/PWT) may be explained by heterogeneity in the underlying pathophysiology of altered diastolic filling between the 2 race groups. Although blacks had higher SBP and greater LVMI relative to those of whites, the percentage of impaired relaxation did not differ between blacks and whites after age effect was taken into account, indicating greater influence of protecting factors or pseudonormalization for LV diastolic filling in blacks. An alternative explanation is lack of statistical power to reject the null hypothesis for the 2 cross-trait pairs in blacks, given the smaller magnitude of single-trait sibling correlations for MVE, IVRT, and PWT in blacks.
Many subjects in the study population who were defined as hypertensive on the basis of history and current antihypertensive therapy were normotensive at the time of the investigation. To ensure that potential misclassification of hypertensive subjects had not contaminated the sibling correlations, we repeated the analyses on the subgroup (28.9% whites and 37.5% blacks) verified to be hypertensive based on blood pressure measurements at the time of the investigation. There were no significant changes in the results, except for strengthening of the single-trait correlation for LV mass in blacks (0.51±0.09 compared with 0.22±0.04, P<0.01).
The findings of significant cross-trait sibling resemblance in LV filling and relaxation have important implications. Understanding the heritability and genetics of LV diastolic function could lead to new insights into the pathophysiology and treatment not only of hypertension but also of closely related disorders involving diastolic dysfunction, including congestive heart failure.1–5 Linkage analyses are underway to identify chromosomal regions containing individual and pleiotropic genes for LV structure and function.
The HyperGEN network is funded by National Heart, Lung, and Blood Institute (NHLBI) R01 HL55673 and cooperative agreements (U10) with NHLBI: HL54471, HL54515 (UT), HL54472, HL54496 (MN), HL54473 (MO), HL54495 (AL), and HL54509 (NC). Dr Tang is supported partly by NHLBI training grant 1T32-HL07972-01.
- Received January 25, 2002.
- Revision received March 11, 2002.
- Accepted June 24, 2002.
Senni M, Tribouilloy CM, Rodeheffer RJ, Jacobsen SJ, Evans JM, Bailey KR, Redfield MM. Congestive heart failure in the community: a study of all incident cases in Olmsted County, Minnesota, in 1991. Circulation. 1998; 98: 2282–2289.
Kitzman DW, Gardin JM, Gottdiener JS, Arnold A, Boineau R, Aurigemma G, Marino EK, Lyles M, Cushman M, Enright PL. Cardiovascular Health Study Research Group: importance of heart failure with preserved systolic function in patients >65 years of age. Am J Cardiol. 2001; 87: 413–419.
Palmieri V, Bella JN, DeQuattro V, Roman MJ, Hahn RT, Dahlof B, Sharpe N, Lau CP, Chen WC, Pavan E, de Simone G, Devereux RB. Relations of diastolic left ventricular filling to systolic chamber and myocardial contractility in hypertensive patients with left ventricular hypertrophy: the PRESERVE study. Am J Cardiol. 1999; 84: 558–562.
Bella JN, Palmieri V, Liu JE, Kitzman DW, Oberman A, Hunt SC, Hopkins PN, Rao DC, Arnett DK, Devereux RB. Relationship between left ventricular diastolic relaxation and systolic function in hypertension: the Hypertension Genetic Epidemiology Network (HyperGEN) Study. Hypertension. 2001; 38: 424–428.
Arai M, Alpert NR, MacLennan DH, Barton P, Periasamy M. Alteration in sarcoplasmic gene expression in human heart failure: a possible mechanism for alterations in systolic and diastolic properties of the failing myocardium. Circ Res. 1993; 72: 463–469.
Hajjar RJ, Schmidt U, Matsui T, Guerrero JL, Lee KH, Gwathmey JK, Dec GW, Semigran MJ, Rosenzweig A. Modulation of ventricular function through gene transfer in vivo. Proc Natl Acad Sci U S A. 1998; 95: 5251–5256.
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.
Falconer DS. Introduction to Quantitative Genetics, 3rd ed. New York: Wiley; 1989.
Tang W, Arnett DK, Devereux RB, Hopkins PN, Oberman A, Province MA, Kitzman DW. Heritability of left ventricular filling in hypertensive African Americans and whites: the HyperGEN Echocardiography Study: proceedings of the 40th annual conference on cardiovascular disease epidemiology and prevention, San Diego, March 2000. Circulation. 2000; 101: 725.Abstract.
Williams RR, Rao DC, Ellison RC, Arnett DK, Heiss G, Oberman A, Eckfeldt JH, Leppert MF, Province MA, Mockrin SC, Hunt SC. NHLBI family blood pressure program: methodology and recruitment in the HyperGEN network: Hypertension Genetic Epidemiology Network. Ann Epidemiol. 2000; 10: 389–400.
Palmieri V, Bella JN, Arnett DK, Liu JE, Oberman A, Schuck MY, Kitzman DW, Hopkins PN, Morgan D, Rao DC, Devereux RB. Effect of type 2 diabetes mellitus on left ventricular geometry and systolic function in hypertensive subjects: Hypertension Genetic Epidemiology Network (HyperGEN) Study. Circulation. 2001; 103: 102–107.
Devereux RB, Roman MJ. Evaluation of cardiac function and vascular structure and function by echocardiography and other noninvasive techniques.In: Laragh JH, Brenner BM, eds. Hypertension: Pathophysiology, Diagnosis and Management, 2nd ed. New York: Raven Press Ltd; 1995: 1969–1985.