OUP user menu

The genetic determination of left ventricular mass in healthy adults

L. Swan, D.H. Birnie, S. Padmanabhan, G. Inglis, J.M.C. Connell, W.S. Hillis
DOI: http://dx.doi.org/10.1016/S0195-668X(02)00524-9 577-582 First published online: 2 March 2003


Aims The extent to which left ventricular (LV) mass, an independent cardiovascular risk factor, is determined by genetic factors is unclear. The aim of this study was to assess the heritability of LV mass and its association with three potential candidate genes.

Methods A population-based adult twin study model was utilized. Echocardiographic assessment of LV mass was performed in 110 twin pairs (mean age 55.9±10.9 years). An estimate of genetic determination, heritability, was calculated for the main echocardiographic parameters. The cohort were genotyped for the G-protein beta-3, aldosterone synthase, and beta-1 adrenoceptor genes.

Results The intra-class correlation coefficients for LV mass were 0.69 for monozygotic (r-MZ) twins and 0.32 for dizygotic (r-DZ) twins, P=0.008 (heritability estimate of 0.69). This pattern persisted following correction for known confounding factors. Within-pair differences in the monozygotic, discordant and concordant dizygotic twins showed no differences for the three genes with respect to left ventricular wall thickness or mass. There was a non-significant trend towards a relationship between LV mass and the beta-1 adrenoceptor genotype.

Conclusion Within a normal population left ventricular mass has a significant genetic determination. Further investigation of potential candidate genes is required.

  • Hypertrophy
  • Genetics
  • Ventricle


Left ventricular hypertrophy, defined either by electrocardiography or echocardiography is a well-established risk factor for cardiovascular morbidity and mortality.1,2 Despite the prognostic significance of left ventricular (LV) mass understanding of its determinants is limited. Population studies have attempted to delineate the major determinants of ventricular hypertrophy which have included age, race, sex, body mass index, blood pressure and plasma epinephrine/norepinephrine levels.3,4

Little is known of the genetics of LV mass outside the narrow confines of inherited cardiomyopathies.5 These are, however, rare conditions with a different aetiology and geometry to the ventricular hypertrophy prevalent in the general population and as such may be an inappropriate starting point for the investigation of potential candidate genes for more common forms of hypertrophy.

In the 1980s a number of small studies of children and young adults investigated LV mass and its response to training.6–9 The conclusions from these studies were contradictory with some suggesting a specific genetic component and others simply a familial one.10 The largest twin study to date, performed on 254 pairs of school children, demonstrated that LV mass was genetically determined in childhood.11

This study aimed to determine the degree to which left ventricular hypertrophy is heritable in adults, using a population-based twin methodology. The relationship between heritability and three potential candidate genes (polymorphisms of the G protein beta-3 subunit, aldosterone synthase, and beta-1 adrenoceptor genes) was also studied.



The population studied consisted of a group of monozygotic and dizygotic twins aged between 30 and 85 years of age. Twins were recruited from the general population of the West of Scotland via a media campaign. The nature of the disease process being investigated was not disclosed to minimize volunteer bias. Zygosity was determined by validated questionnaire and verified by analysing tandem repeat polymorphisms. Four markers were used—AFM238×d10, AFM288vb9, AFM273yfl and AFM199zb6 (Centre d'Etudes du Polymorphisme Humain, Paris, France). The probability of inaccurate assessment of zygosity using this technique was less than 10−5.

Details of study

All twins attended the study centre in the fasting state. Blood was withdrawn for genetic analysis from a forearm vein via an intravenous cannula. Resting blood pressures (manually mercury sphygmomanometer) and anthropometric parameters (height and weight) were measured at baseline. Ambulatory blood pressure monitoring was performed in a non-hospital environment using a Spacelabs monitor (Spacelabs 90207, Spacelabs Medical, Inc.) fitted to the non-dominant arm. Recordings were made over a period of at least 24 h. Tracings were disregarded if less than 85% of the readings were unacceptable as defined by the monitor's pre-set upper and lower limits.

Transthoracic echocardiography was performed and analysed in accordance with the American Society of Echocardiography guidelines (Acuson 128XP). M-mode measurements were made over a least three cardiac cycles by a single observer blinded to the twins' other results and zygosity determination. Left ventricular mass was calculated from 2-D guided M-mode using the Penn convention.12 Doppler assessment of the mitral valve was performed with E wave and A wave filling measurements being made over five cardiac cycles.


Common polymorphisms of three candidate genes were analysed. Genomic DNA was isolated by phenol–chloroform extraction of EDTA blood samples.13 Genotyping was performed by PCR by the method previously described for C825T G-protein beta-3 subunit polymorphism,14 SF1 RFLP15 and beta-1 adrenoceptor polymorphism.16

Data analysis

Descriptive data for the cohort are quoted as mean values and standard deviations. Differences between twin groups were compared using a paired t-test. The twin pairs were assessed by differences in within-pair variances between monozygotic and dizygotic twins.17 Intra-class correlation coefficients were calculated for the pairs using a statistics software programme (MINITAB 11.2). This test included a test of homogeneity. Heritability, an estimate of the degree to which a variable is genetically determined, was calculated for the major phenotypes of left ventricular structure. Heritability (h2) was measured by the equation h2=2 (rMZ-rDZ) where rMZ is the correlation coefficient for the monozygotic group and rDZ that of the dizygotic group. The influence of confounders was assessed by estimating heritability of the residuals following multiple regression. Intra-pair differences in left ventricular echocardiographic parameters for monozygous, concordant dizygous and discordant dizygous twins were calculated for each of the potential candidate genes. A P value of less than 0.05 was taken to be statisticallysignificant.


Analysable M-mode data

Fully analysable echocardiograms were obtained in 220 subjects (72.8%). Those without acceptable M-mode measurements were excluded from further analysis. This group was older (58.7±12.5 vs 51.2±11.7, P<0.0001) and had higher systolic blood pressures (136.8±16.2 vs 129.1±15.0, P<0.001) than those with satisfactory data. There was no statistical significance in weight 65.9±16.5 vs 66.0±14.8, P=ns) or height (163.4±9.35 vs 161.5±7.94, P=ns) between the two groups.

Descriptive data

There was no significant difference between the monozygotic (MZ) and dizygotic (DZ) twins with regards to baseline echocardiographic parameters (Table 1). The most important determinants of LV mass in the cohort were weight (r=0.43, P<0.01) and sex (r=−0.41, P<0.01). Daytime SBP was the most important blood pressure parameter (r=0.26, P<0.01) (Table 2). The relationship between LV mass and ambulatory blood pressures was primarily via wall thickness rather than ventricular diameter. The strongest correlates with diastolic wall thickness were daytime SBP for septal measurements (r=0.32, P<0.01) and night-time mean BP for the posterior wall (r=0.27, P<0.01). In systole the same pattern was reproduced (daytime SBP–septal thickness (r=0.34, P<0.01) and night-time mean BP–posterior wall (r=0.40, P<0.01).

View this table:
Table 1

Descriptive data—monozygotic vs dizygotic (means±SD)

MZDZP value
Age (years)57.8±10.554.1±11.50.14
SBP (mmHg)133.8±16.0131.6±17.60.94
Weight (kg)68.2±20.064.1±13.11.0
IVSd (cm)11±211±210.7
LVDd (cm)47±547±50.51
PWd (cm)8±148±1.30.33
LV mass (Penn)178.4±57.9180.4±59.60.43
E wave (ms−1)0.8±0.20.8±0.20.48
A wave (ms−1)0.7±0.10.7±0.20.48
FT (IVS)33.1±18.533.2±17.70.94
  • IVSd=interventricular septal thickness in diastole (in mm); LVDd=left ventricular internal diameter in diastole (in mm); PWd=posterior wall thickness in diastole (in mm); FT=fractional thickening; FS=fractional shortening; E/A=ratio of the mitral valve E wave to the A wave.

View this table:
Table 2

Correlations with LV mass and E/A ratio (as r value)

LV massE/A ratio
Body surface area0.14*−0.04
Resting systolic BP0.14*−0.15*
Resting diastolic BP0.08−0.62**
Smoking (packs)0.010.04
24 h systolic BP0.18**−0.27**
24 h diastolic BP0.17*0.29**
24 h mean BP0.19**−0.33**
Day systolic BP0.26**−0.3**.
Day diastolic BP0.16*−0.27**
Day mean BP0.18**−0.3**
Night systolic BP0.17*−0.26**
Night diastolic BP0.22**−0.23**
Night mean BP0.21**-0.29**
  • *P<0.05, **P<0.01.

Age and resting DBP were the strongest associations with E/A ratio (r=−0.53, P<0.001 and r=−0.62, P<0.01, respectively). There was a significant difference in systolic blood pressure (123.6 vs 112.1, P=0.04), but not LV mass, between the lowest and highest quartile of E/A ratio.

The heritability of left ventricular structure

The intra-class correlation coefficients for the monozygotic and dizygotic groups and their comparison (z) are displayed in Table 3. Left ventricular internal dimensions in diastole and LV mass were both subject to a significant genetic influence (heritability estimates of 0.61 (P=0.02) and 0.69 (P=0.008), respectively). Two-thirds of the variation in LV mass could be explained by a genetic influence. Following correction for age, sex, blood pressure and weight (Corr LV mass) there was still a strong genetic component to left ventricular mass (heritability estimate of 0.53 (P=0.006)).

View this table:
Table 3

Intra-class correlation coefficients and heritability estimates for echocardiographic parameters

LV mass500.690.85410.320.332.390.0080.69
Corr LVM480.530.59390.030.032.520.0060.53
E wave450.490.54400.130.131.840.030.49
A wave440.360.37390.
E/A ratio440.240.25400.500.55−1.360.91−0.52
  • For abbreviations, see.Table 1

    LVDs=left ventricular internal diameter in systole (in mm); IVSs=interventricular septal thickness in systole (in mm); PWs=posterior wall thickness in systole (in mm).


Limited genotyping was performed on the cohort. Demographic data for the subjects is displayed in terms of genotype in Table 4. Only one member of each twin pair was included in association testing. Selecting either twin 1 or twin 2 did not influence the results. Within-pair difference in MZ twins, discordant and concordant DZ twins showed no significant difference for the three genes forposterior wall thickness in diastole (in mm) (PWd). Left ventricular mass and the adrenergic receptor (beta-1 AR) showed a trend towards greater within-pair difference between MZ and DZ discordant twins but this did not reach statistical significance in this small cohort. Left ventricular internal diameter in diastole (in mm) (LVDd) and SF1 polymorphism also showed a similar difference between MZ and DZ discordant twins. Polymorphisms in the G-protein beta-3, aldosterone synthase and beta-1 adrenoceptor genes did not play a significant role in the determination of left ventricular mass in this cohort (Table 5).

View this table:
Table 4

Demographics of subjects according to genotype

Genotype frequency51418
LV mass174.0 (56.8)180.5 (76.1)191.5 (65.9)0.90
LVDd4.7 (0.60)4.6 (0.52)4.8 (0.63)0.83
PWd0.8 (0.11)0.8 (0.14)0.8 (0.22)0.98
Beta- 1ARGenotypeCCCGGG
Genotype frequency47494
LV mass184.8 (63.3)171.9 (49.1)146.7 (22.5)0.43
LVDd4.8 (0.52)4.7 (0.61)4.5 (0.34)0.49
PWd0.8 (0.15)0.8 (0.12)0.8 (0.11)0.84
Genotype frequency245521
LV mass162.4 (44.5)187.3 (75.0)178.1 (48.0)0.29
LVDd4.6 (0.37)4.8 (0.55)4.6 (0.60)0.33
PWd0.8 (0.10)0.8 (0.14)0.8 (0.13)0.12
  • For explanation of abbreviations, see earlier tables.

View this table:
Table 5

Within pair differences according to genotype

GNB3LV mass35.8 (28.3)44.1 (36.1)42.0 (39.2)0.52
LVDd0.3 (0.22)0.4 (0.26)0.4 (0.31)0.26
PWd0.1 (0.09)0.1 (0.11)0.1 (0.10)0.72
Beta 1ARLV mass37.7 (28.5)40.2 (30.7)49.9 (44.1)0.25
LVDd0.3 (0.22)0.4 (0.26)0.4 (0.30)0.51
PWd0.1 (0.09)0.1 (0.10)0.1 (0.11)0.83
SF1LV mass37.9 (28.8)44.0 (37.5)44.8 (33.6)0.38
LVDd0.3 (0.21)0.3 (0.28)0.5 (0.28)0.04
PWd0.1 (0.09)0.1 (0.12)0.1 (0.08)0.46


Although twin studies in children have given conflicting results regarding the heritability of left ventricular structure the largest to date reported that ventricular mass was genetically determined. However the genetic control of a phenotype varies with age and may be ‘watered-down’ by cumulative exposure to environmental factors. Paediatric studies cannot, therefore, be easily extrapolated to adults.18 The current study does, however, demonstrate the genetic determination of various aspects of left ventricular structure in middle-aged adults. Left ventricular mass has a significant heritability (approximately 50%) that is not explained by known confounders.

Functional assessments of the left ventricle such as fractional shortening and fractional thickening did not reveal a significant genetic component, environmental factors being more important. The exception to this was mitral valve E wave forward flow, reflecting the early diastolic filling of the left ventricle, a finding consistent with previous paediatric studies.19

Twin estimates of heritability are often higher than those of sibling and family cohorts. Indeed the estimates for LV mass in this study were higher than those of the Framingham family study (0.24–0.32).20 Although heritability estimates from twin studies attempt to fully dissect out genetic from environmental influences there are some limitations of such models. One basic assumption of the classic twin model is that MZ and DZ environments are the same—however, this is not always the case. One example is that MZ twins, living apart, spend more time with their twin than a corresponding pair of dizygotes. Such differences need to be considered when comparing heritability estimates between twin and family studies. In addition, heritability is population- and situation-specific. It would be of interest to use a similar technique to quantify heritability in groups with a lower prevalence of cardiovascular disease than the current West of Scotland cohort. With these provisos in mind heritability estimates should not be seen as an end in themselves but as a pointer towards those phenotypes that may justify candidate gene investigation. This study identifies LV mass in adults as such a phenotype.

Limited genotyping was performed on this cohort using candidates proposed because of their involvement in cardiovascular regulation. However, polymorphic variation of these genes did not account for the genetic determination of LV mass in this cohort. The first candidate was the polymorphism in the G protein beta subunit (825T), which results in the generation of a splice variant linked with enhanced intracellular sodium/hydrogen exchange. Associations with hypertension,21impaired left ventricular diastolic filling22 and ventricular hypertrophy have been suggested 23 but detailed corroboration of these findings was lacking. The association between the 825T and ventricular hypertrophy was previously described in hypertensive subjects—a finding not confirmed in this study of principally normotensive subjects.

The beta-1 adrenergic receptor (beta-1 AR), a cell surface signalling protein involved in mediating the cardiac influences of the sympathetic nervous system, was also studied. It has been proposed that a polymorphism in the intracellular cytoplasmic tail16 may be a risk factor for cardiac events. The final association examined was with the aldosterone synthase gene. Two polymorphisms of the aldosterone synthase gene—C344T polymorphism at the SF1 transcription factor binding site and a gene conversion in intron2—have been investigated but the data regarding left ventricular hypertrophy has been conflicting. Kupari et al. published a positive study in a small cohort of Finnish subjects aged 36 to 37 years,15 but this was superseded by the larger MONICA study reported by Schunkert.24 Schunkert's study, and the present study, were both negative.

The history of the ACE genotype25,26 demonstrates the potential pitfall in attempting to identify the genetic architecture of non-Mendelian phenotypes. Associations are likely to be small, and population-specific. For example the ACE genotype may only be of importance in specific circumstances such as training-associated LV hypertrophy.27 A cohort of this size is capable of detecting only a sizeable contribution from an individual locus and as such should be viewed as a pilot study aiding the direction of future large (andexpensive) association studies.27 This study demonstrates the lack of any sizeable impact of these candidates on LV mass in this population. The trend towards an association with the beta-1 adrenoceptor genotype may be of potential interest and requires reexamination in a larger cohort.

In conclusion, the genetic determination of LV mass is not limited to children and can be demonstrated in an adult cohort. This association is not explained by confounding variables. The field is open for the investigation of the influence of other candidate genes. More information is needed before comparisons can be made between adult populations to assess whether this degree ofheritability in a high-risk population such as the West of Scotland is reflective of other racialgroups.


Ms Maria Rowan and Mrs A. McKenzie for providing nursing and technical support for this study. This study was supported by a Project Grant from the British Heart Foundation.


View Abstract