You are here

Article Text

Article menu

Download PDFPDF

Cardiac risk factors and prevention
Original article
Left ventricular shape predicts different types of cardiovascular events in the general population
  1. Bharath Ambale-Venkatesh1,
  2. Kihei Yoneyama1,
  3. Ravi K Sharma1,
  4. Yoshiaki Ohyama1,
  5. Colin O Wu2,
  6. Gregory L Burke3,
  7. Steven Shea4,
  8. Antoinette S Gomes5,
  9. Alistair A Young6,
  10. David A Bluemke2,
  11. João AC Lima1
  1. 1Johns Hopkins University, Baltimore, Maryland, USA
  2. 2National Institutes of Health, Bethesda, Maryland, USA
  3. 3Wake Forest University Health Sciences, Winston-Salem, North Carolina, USA
  4. 4Columbia University, New York, New York, USA
  5. 5University of California at Los Angeles, Los Angeles, California, USA
  6. 6University of Auckland, Auckland, New Zealand
  1. Correspondence to Dr Bharath Ambale-Venkatesh, MR 110 Radiology, Johns Hopkins University, Nelson Basement, 600 N Wolfe Street, Baltimore MD 21287, USA; bambale1{at}jhmi.edu

Abstract

Objective To investigate whether sphericity volume index (SVI), an indicator of left ventricular (LV) remodelling, predicts incident cardiovascular events (coronary heart disease, CHD; all cardiovascular disease, CVD; heart failure, HF; atrial fibrillation, AF) over 10 years of follow-up in a multiethnic population (Multi-Ethnic Study of Atherosclerosis).

Methods 5004 participants free of known CVD had magnetic resonance imaging (MRI) in 2000–2002. Cine images were analysed to compute, Embedded Image equivalent to LV volume/volume of sphere with length of LV as the diameter. The highest (greatest sphericity) and lowest (lowest sphericity) quintiles of SVI were compared against the reference group (2–4 quintiles combined). Risk-factor adjusted hazard's ratio (HR) from Cox regression assessed the predictive performance of SVI at end-diastole (ED) and end-systole (ES) to predict incident outcomes over 10 years in retrospective interpretation of prospective data.

Results At baseline, participants were aged 61±10 years; 52% men and 39%/13%/26%/22% Cauc/Chinese/Afr-Amer/Hispanic. Low sphericity was associated with higher Framingham CVD risk, greater coronary calcium score and higher N-terminal pro-brain natriuretic peptide (NT-proBNP); while increased sphericity was associated with higher NT-proBNP and lower ejection fraction. Low sphericity predicted incident CHD (HR: 1.48, 1.55–2.59 at ED) and CVD (HR: 1.82, 1.47–2.27 at ED). However, both low (HR: 1.81, 1.20–2.73 at ES) and high (HR: 2.21, 1.41–3.46 at ES) sphericity predicted incident HF. High sphericity also predicted AF.

Conclusions In a multiethnic population free of CVD at baseline, lowest sphericity was a predictor of incident CHD, CVD and HF over a 10-year follow-up period. Extreme sphericity was a strong predictor of incident HF and AF. SVI improved risk prediction models beyond established risk factors only for HF, but not for all CVD or CHD.

https://doi.org/10.1136/heartjnl-2016-310052

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Introduction

    Left ventricular (LV) remodelling can be defined as the process of structural changes in the LV in response to changes in intrinsic myocardial tissue characteristics and architecture, or to extraneous stimuli caused by increased pressure or volume overload. LV remodelling is considered an important indicator of subclinical cardiovascular (CV) disease and its regression frequently used to index therapeutic efficacy. There is extensive evidence pointing to maladaptive remodelling being a strong predictor of CV disease.1 ,2

    Prior studies have shown hypertrophy characterised using the LV mass index, LV mass:volume ratio and relative wall thickness are predictors of heart failure (HF) and CV disease.3–5 LV end-diastolic and end-systolic volume indices as well as chamber diameters are also markers of HF and LVremodelling.6 ,7 In addition, the evaluation of LV chamber shape has been identified as an important marker of remodelling processes in patients with mitral regurgitation8 non-ischaemic cardiomyopathy9 and myocardial infarction (MI).10 ,11 However, the ability of LV chamber shape to predict different types of CV events in an asymptomatic population free of CV disease is incompletely understood.

    We evaluated the ability of the sphericity volume and dimension indices as surrogates of LV remodelling, derived from cardiac magnetic resonance imaging (MRI), to predict coronary heart disease (CHD) events, all cardiovascular disease (CVD) events, HF and atrial fibrillation (AF) over a 10-year follow-up period in a large multiethnic population free of CV disease at baseline.

    Materials and methods

    Study design and population

    The Multi-Ethnic Study of Atherosclerosis (MESA) has been described in the literature.12 In brief, between 2000 and 2002 (baseline) 6814 men and women (age range: 44–84 years) who were free of clinically apparent CV disease were recruited from six US communities. Of these, 5004 (73%) participants underwent MRI during the baseline exam. These study population has been previously reported.5 ,13–17 While these prior articles dealt with the prognostic utility of LV mass and volumes, their association with risk factors, and advanced statistical modelling techniques to assess ventricular structure information; we report the prognostic utility of LV shape as characterised by the sphericity index over and above these LV structure measures. Sphericity index and traditional CV risk factors together were available in 4884 (97.8%, 4879 for indices at end-systole (ES)) participants5 in this retrospective analysis of prospectively acquired data. Traditional risk factor measures5 serum concentration of N-terminal pro-brain natriuretic peptide (NT-proBNP, available in 4090/4884),18 and coronary artery calcium score (CACS)16 ,19 were obtained as explained previously. The institutional review boards of all MESA field centres approved this HIPAA-compliant study and all participants gave informed consent.

    Cine MRI

    At baseline, cardiac MRI was performed with 1.5-T scanners with determination of LV mass and volumes as previously described.17 Briefly, a stack of short-axis images covering the entire LV was acquired using a fast gradient recalled echo sequence. LV shape indices were calculated both at end-diastole (ED) and ES as follows:

    • Sphericity volume index (SVI)=LV volume/(L3×π/6)8

    • Sphericity dimension index (SDI)=D/L20

    L=length of the LV from the most apical slice to the most basal slice where the endocardium was detectable in the short-axis views and D=the maximal diameter of the LV from the short-axis slices (see figure 1A). Effectively, SVI is equivalent to LV volume divided by the volume of a sphere with length of LV as the diameter.

    Figure 1
    Figure 1

    (A) Illustration of the calculation of the sphericity indices. (B) Scatter plot and associated linear fit between the sphericity volume and dimension indices at end-diastole. (C) Scatter plot and associated linear fit between the sphericity volume and dimension indices at end-systole. LV, left ventricular.

    Endpoint adjudication

    Events adjudicated as HF, AF, CHD and all CVD as part of the MESA study were used as endpoints. In addition to MESA follow-up examinations every 2 years, a telephone interviewer contacted each participant (or representative) every 6–9 months to inquire about interim hospital admissions, CV outpatient diagnoses and deaths. Two physicians reviewed all records for independent endpoint classification and assignment of event dates. Criteria for CHD outcomes included any of—MI, resuscitated cardiac arrest, definite angina, probable angina (if followed by revascularisation) and CHD death. CVD outcomes represented a composite of stroke, stroke death, CHD, CHD death, atherosclerotic death and CVD death. Criteria for HF as an endpoint included symptomatic HF diagnosed by a physician for a patient receiving medical treatment for HF and (1) pulmonary oedema/congestion by chest X-ray and/or (2) dilated ventricle or poor LV function by echocardiography or ventriculography or evidence of LV diastolic dysfunction. Criteria for AF as an endpoint required in-hospital AF diagnosis according to ICD9 codes. A detailed description of the criterion used for each of the endpoints is provided in the online supplementary 1.

    Statistical analysis

    LV shape was divided into quintiles of the SVI with the first quintile representing low sphericity (or high conicity) and the last quintile representing high sphericity. Quintiles 2–4 were combined to form the reference group. In all, there were three categories, with the first (low sphericity, high conicity) and fifth (high sphericity) quintiles representing the extremes of LV shape compared with the combined second to fourth quintiles (reference). Similar categories were formed using SDI. See online supplementary figure S1 for non-linearity in the association and the cut-offs. Table and figure numbers starting with ‘S’ are part of the online supplementary 2.

    Hazard ratios (HRs) and their 95% confidence intervals from multivariable Cox regression models assessed the power of LV shape parameters at baseline to predict endpoints over the follow-up period after adjustments for demographics and traditional CV risk factors. The discrimination improvement of adding LV shape to traditional risk factors was assessed using the increase in the Harrell's C-statistic. Models adjusted for conventional CV risk factors (see table for list). For CHD and CVD—an additional model adjusted for proven risk factors mass:volume ratio and the presence of coronary calcium (CACS>0); for HF and AF—an additional model adjusted for LV mass indexed to body surface area, NT-proBNP and ejection fraction. Calibration of the models was confirmed using the Gronnesby–Borgan tests (see online supplementary table S1 and figure S2). The ‘stcoxgof’ command was used to assess goodness-of-fits, the score statistic was used for computing the added variable version of the Gronnesby and Borgan test.21–23 Kaplan-Meier curves were used to illustrate the abilities of SVI and SDI divided by the three categories to predict the endpoints. p Values of <0.05 were used for significance testing. NT-proBNP was logarithm transformed because of a skewed distribution, before regression analysis. All statistical analyses were done using STATA V.11.0 (StataCorp LP, College Station, TX, USA).

    Results

    Population characteristics

    Table 1 shows population characteristics of all participants included in this study in addition to those with CHD (6.18%), CVD (8.61%), HF (2.91%) and AF (4.16%). The median follow-up time was 10.2 (interquartile range: 9.7–10.7) years. In general, participants who experienced events were older, had a higher BMI, systolic blood pressure and Framingham CVD risk score compared with the overall population. A significantly higher proportion of African Americans had HF and a higher proportion of Caucasians had HF and CVD compared with Chinese Americans at the end of follow-up. The proportion of Chinese Americans and Hispanics with outcomes was lower than in the overall population. Similarly, the proportion of those on antihypertensive medication/s, β-blockers or angiotensin-converting-enzyme inhibitor (ACE) inhibitors was higher in those who experienced events. The proportion of diabetics and current/former smokers was also higher in the participants with future HF, CHD and CVD than the overall population.

    View this table:
    Table 1

    Population characteristics of participants at baseline (before occurrence of events) by incident event categories

    Sphericity index, LV shape and risk factors

    The distribution of risk factors across low sphericity (most conical hearts), reference (intermediate levels of sphericity) and high sphericity groups by SVI at ED and ES, respectively, are shown in table 2 (and see online supplementary table S2). The first and fifth quintiles of SVI were categorised as low sphericity (high conicity, ED: <0.22, ES: <0.11) and highest sphericity (ED: >0.34, ES: 0>0.19), respectively. Similar grouping was also performed for SDI. At both ED and ES, the proportion of Caucasians with high sphericity was higher. Low sphericity was associated with higher Framingham CVD risk, while high sphericity was associated with lower risk as compared with the intermediate sphericity reference group—as seen by the Framingham CVD risk score and associated risk factors (systolic blood pressure, heart rate and calcium score).

    View this table:
    Table 2

    Population characteristics by LV chamber shape characterised by quintiles of the sphericity index at end-diastole

    Concentric hypertrophy, as quantified by the LV mass:volume ratio decreased progressively from the low sphericity≥reference≥high sphericity groups (figure 2E, F); with corresponding increase in LV end-diastolic volume (figure 2C, D). Conversely, absolute hypertrophy assessed as the LV mass index (figure 2A) and the LV ejection fraction decreased from low sphericity≥reference≥high sphericity groups (figure 2H) at ED, while the trend of decreasing ejection fraction was consistent for ES in particular. These are further illustrated in categorical comparisons performed in table 2 and online supplementary table S2.

    Figure 2
    Figure 2

    Scatter plot and associated curve fit (locally weighted scatterplot smoothing) between the sphericity index at end-diastole and end-systole, and—left ventricular (LV) mass indexed to body surface area (A and B); LV end-diastolic volume indexed to body surface area (C and D); LV mass:volume ratio (E and F) and LV ejection fraction (G and H).

    In all, 77% at ES and 85% at ED of the 4879 of the study participants were classified similarly by both the SVI and SDI indices, and there were only eight patients that were categorised as low sphericity/high sphericity by SVI and vice versa by SDI. The SVI and SDI were highly correlated (figure 1B, C).

    Prediction of CHD and CVD

    Low sphericity at ES was a predictor of CHD (figure 3A, C, see online supplementary figure S3a,c) and this predictive ability was independent of CACS and LV mass:volume ratio. For CVD as the outcome of interest, high conicity at both ED and ES were again significant predictors (figure 3B, D, see online supplementary figure S3b,d) after adjustment for conventional risk factors, but this relationship was attenuated after adjustment for subclinical disease markers—CACS and LV mass:volume ratio (model 2, table 3). Table 3 (and see online supplementary table S3) shows the HRs for the prediction of CHD and CVD using SVI (and SDI). Additional adjustment for LV chamber size (LV ED volume) did not change any of the relationships.

    View this table:
    Table 3

    Association, presented as hazard ratios (95% confidence intervals), of sphericity indices at end-diastole and end-systole, for coronary heart disease and all cardiovascular disease as endpoints using Cox regression models

    Figure 3
    Figure 3

    Kaplan-Meier survival curves for coronary heart disease (A) and cardiovascular disease (B) as endpoints across the low sphericity or high conicity (blue), high sphericity (green) and the reference (red) groups of the sphericity volume index at end-diastole. Corresponding survival curves for categories of the sphericity volume index at end-systole are shown (C and D). Individuals were free of all cardiovascular disease at baseline. p<0.05 for low sphericity versus reference groups.

    Prediction of HF and AF

    Table 4 (and see online supplementary table S4) shows the HRs for the prediction of HF and AF using SVI (and SDI). High sphericity at ED was a predictor of both HF and AF (figure 4A, B, see online supplementary figure S4a, b) and this predictive ability was independent of conventional risk factors. This relationship remained significant after adjustment for LV chamber size (LV end-diastolic volume) and for NT-proBNP, LV mass index and LV ejection fraction. The predictive ability of LV shape indices remained unattenuated after adjustment for these parameters and biomarkers. While neither low nor high sphericity at ES predicted AF using SVI, participants having increased sphericity as calculated using SDI experienced higher rates of AF events (see online supplementary table S4). Importantly, extremes of sphericity at ES (using either index) were predictors of HF. High sphericity at ES, in particular, was strongly associated with HF even after adjustment for conventional risk factors, LV mass, ejection fraction and NT-proBNP (figure 4C, see online supplementary figure S4c).

    View this table:
    Table 4

    Association, presented as hazard ratios (95% confidence intervals), of sphericity indices at end-diastole and end-systole, for heart failure and atrial fibrillation as endpoints using Cox regression models

    Figure 4
    Figure 4

    Kaplan-Meier survival curves for heart failure (A) and atrial fibrillation (B) as endpoints across the low sphericity or high conicity (blue), high sphericity (green) and the reference (red) groups of the sphericity volume index at end-diastole. Corresponding survival curves for categories of the sphericity volume index at end-systole are shown (C and D). Individuals were free of atrial fibrillation and heart failure at baseline. p<0.05 for low sphericity versus reference groups (C). p<0.05 for high sphericity versus reference groups (A–C).

    Association of LV shape to NT-proBNP and improvement in prediction models

    Greater NT-proBNP levels were associated with high sphericity at ED and lower sphericity at ES (table 2 and see online supplementary table S2) and these associations paralleled to some extent the relations of LV shape indices with incident HF as shown above.

    There was an improvement in the C-index after the addition of either SVI or SDI to conventional CV risk factors (increased from 0.74 to 0.75 for SVI at both ED and ES) for the prediction of all CVD events, however, this increment was not statistically significant. This was also the case for the prediction of CHD events (0.74to 0.75) and AF (0.79 to 0.80). On the other hand, there was a statistically significant improvement in the C-index after the addition of SVI (0.80 to 0.82, p=0.03) or SDI (0.80 to 0.81, p=0.03) at ES to conventional risk factors for HF.

    Discussion

    In a large population free of known CV disease, LV chamber shape as characterised by sphericity indices was an independent predictor of CV events over a 10-year follow-up period. Low sphericity (or greater conicity) was a predictor of incident CVD and CHD, and this was independent of CV risk factors and known biomarkers. High sphericity was an independent predictor of incident AF. Both extremely low (high conicity) and extremely high sphericity were predictors of incident HF.

    Adverse LV remodelling is considered a valuable phenotyping tool to differentiate CV disease processes and is associated with increased CVD risk.1 Prior work has largely been restricted to studying LV mass, LV volume or a combination of the two to derive indices (or groups) of concentric and eccentric remodelling and hypertrophy.4 ,24 ,25 Here, we explored the ability of low and high sphericity, as opposed to a more or less conical chamber shape, to predict CV events. The sphericity indices, used here as surrogates of LV shape, have found use in a number of studies evaluating adaptations to the LV chamber shape—after acute MI,10 ,11 ,25 and in patients with mitral regurgitation and HF,26 as well as a marker of reverse remodelling after cardiac resynchronisation therapy for HF.

    The low sphericity or increased conicity is, perhaps, indicative of a stiffer and more fibrotic ventricle with decreased myocardial deformation and a stiffer arterial system.27 In our study the group classified to have low sphericity had greater concentric remodelling and ventricular hypertrophy although the relationships were not entirely linear, leading to increased cardiomyocyte stress as seen by higher levels of NT-proBNP. The ability of LV concentric hypertrophy to predict CVD, CHD and cerebrovascular diseases in large populational studies have been demonstrated.2 ,3 ,5 ,18 LV concentric remodelling is associated with decreased global and regional function,28 hypertension and insulin resistance syndrome. The low sphericity group also had higher CVD risk (Framingham score, higher blood pressure, more diabetics) as compared with the reference group consistent with what has been observed in previous studies.2 Importantly, over the 10-year follow-up period, low sphericity was an independent (of known risk factors) predictor of CHD and CVD. The increased ejection fraction in this group is also consistent with previous reports of ‘supra-ejection’ fraction as a result of increased torsion associated with concentricremodelling.29 ,30

    Raghava et al in the Framingham Heart Study have shown that concentric and eccentric hypertrophy are related to future incidence of preserved and reduced ejection fraction, respectively, in a large free-living community.4 Similarly, increased LV mass or wall dimension5 ,31 has been shown to predict HF. In our study, extremely low and high sphericity, particularly at ES, were strong predictors of HF. High sphericity was a strong predictor of AF in our population, suggestive of the relationship of volume overload and AF.32 The group classified as having a more spherical LV chamber, interestingly, had a lower CVD risk (Framingham score, lower blood pressure, fewer diabetics and coronary calcium), and a slower heart rate as compared with both the high conicity and reference groups. The group with high sphericity had slightly increased LV volumes indicative of LV dilation. This could be conceived as both a cause and an effect of myocyte stress and dysfunction as the myocytes pump against a greater preload.33 This is also thought to be a compensatory mechanism to maintain the stroke volume in the event of decreasing myofibre shortening.31 Importantly, the high sphericity group had higher NT-proBNP levels than the reference group reflecting greater risk for HF and AF, and greater cardiomyocyte stress. This group also had a lower ejection fraction, although the mean ejection fraction was within the normal range at 64%.

    Using sphericity indices to phenotype LV chamber shape, we have shown that LV shape represents the adaptation of myocardial architecture to changes in tissue characteristics and external CV risk factors. While these indices were independent predictors of events, they do not provide additional discrimination ability to conventional CV risk factors and imaging and biomarkers such as CACS and NT-proBNP. Advanced modelling techniques that incorporate different aspects of remodelling—LV hypertrophy, LV shape and LV size may help in identifying phenotypes with superior prediction and discriminatory abilities.34 ,35 Perhaps as importantly, additional work relating LV shape to regional myocardial function and aortic function may also yield further insight into the mechanisms behind alterations in LV shape change that are associated with CV events. A limitation of this study is that it was performed in participants who were free of CV disease at baseline, and therefore extension of results presented here to populations with prevalent disease would require additional investigation. This study is of an observational nature, which sought to provide insights into the relationship of progressive remodelling and incident events. Application of the same as a screening tool in large populations using MRI might not be cost-effective and echocardiography, particularly, with the advent of three-dimensional echocardiography may be more practical.

    In a large multiethnic population free of CV disease at baseline, LV shape is a predictor of CV events over a 10-year follow-up period. Extreme conicity of the LV was an independent predictor of CHD, CVD and HF, while extreme sphericity, despite lower exposure to traditional risk factors, is an independent predictor of HF and AF.

    Key messages

    What is already known on this subject?

    • Adverse left ventricular (LV) remodelling can be used to study cardiovascular disease processes and is associated with increased cardiovascular disease risk. LV mass, volume and relative wall thickness have all been studied extensively and have found use in event prediction, risk stratification and as endpoints in clinical trials. The ability of LV shape, as a surrogate of LV remodelling, derived from cardiac magnetic resonance imaging, to predict incident cardiovascular disease and its' relationship to risk factors in the general population is less understood.

    What might this study add?

    • Lowest sphericity is associated with a higher risk for coronary heart disease, cardiovascular disease (CVD) and heart failure (HF) among asymptomatic individuals. By contrast, highest degrees of sphericity are associated with higher incidence of HF and atrial fibrillation. Different types of adverse left ventricular remodelling may be used to differentiate CVD processes and may be associated with different types of disease risk.

    How might this impact on clinical practice?

    • Left ventricular (LV) shape is a strong predictor of cardiovascular events independent of LV size, hypertrophy and other components of LV remodelling and risk factors in an asymptomatic population. The use of this index under different disease conditions may yield a better understanding of disease progression and complex remodelling adaptations.

    Acknowledgments

    This research was supported by contracts N01-HC-95159, N01-HC-95160, N01-HC-95161, N01-HC-95162, N01-HC-95163, N01-HC-95164, N01-HC-95165, N01-HC-95166, N01-HC-95167, N01-HC-95168 and N01-HC-95169 from the National Heart, Lung, and Blood Institute and by grants UL1-TR-000040 and UL1-TR-001079 from NCRR. The authors thank the other investigators, the staff and the participants of the MESA study for their valuable contributions. A full list of participating MESA investigators and institutions can be found at http://www.mesa-nhlbi.org.

    References

    1. ↵
      1. Gjesdal O,
      2. Bluemke DA,
      3. Lima JA
      . Cardiac remodeling at the population level—risk factors, screening, and outcomes. Nat Rev Cardiol 2011;8:673–85. doi:10.1038/nrcardio.2011.154
      OpenUrlCrossRefPubMed
    2. ↵
      1. Lavie CJ,
      2. Patel DA,
      3. Milani RV, et al
      . Impact of echocardiographic left ventricular geometry on clinical prognosis. Prog Cardiovasc Dis 2014;57:3–9. doi:10.1016/j.pcad.2014.05.003
      OpenUrlCrossRefPubMed
    3. ↵
      1. Bang CN,
      2. Gerdts E,
      3. Aurigemma GP, et al
      . Four-group classification of left ventricular hypertrophy based on ventricular concentricity and dilatation identifies a low-risk subset of eccentric hypertrophy in hypertensive patients. Circ Cardiovasc Imaging 2014;7:422–9. doi:10.1161/CIRCIMAGING.113.001275
      OpenUrlAbstract/FREE Full Text
    4. ↵
      1. Velagaleti RS,
      2. Gona P,
      3. Pencina MJ, et al
      . Left ventricular hypertrophy patterns and incidence of heart failure with preserved versus reduced ejection fraction. Am J Cardiol 2014;113:117–22. doi:10.1016/j.amjcard.2013.09.028
      OpenUrlCrossRefPubMed
    5. ↵
      1. Bluemke DA,
      2. Kronmal RA,
      3. Lima JAC, et al
      . The relationship of left ventricular mass and geometry to incident cardiovascular events: the MESA (Multi-Ethnic Study of Atherosclerosis) study. J Am Coll Cardiol 2008;52:2148–55. doi:10.1016/j.jacc.2008.09.014
      OpenUrlFREE Full Text
    6. ↵
      1. Lieb W,
      2. Gona P,
      3. Larson MG, et al
      . The Natural history of left ventricular geometry in the community: clinical correlates and prognostic significance of change in LV geometric pattern. JACC Cardiovasc Imaging 2014;7:870–8. doi:10.1016/j.jcmg.2014.05.008
      OpenUrlAbstract/FREE Full Text
    7. ↵
      1. Yu C-M,
      2. Fung W-H,
      3. Lin H, et al
      . Predictors of left ventricular reverse remodeling after cardiac resynchronization therapy for heart failure secondary to idiopathic dilated or ischemic cardiomyopathy. Am J Cardiol 2003;91:684–8. doi:10.1016/S0002-9149(02)03404-5
      OpenUrlCrossRefPubMedWeb of Science
    8. ↵
      1. Sabbah HN,
      2. Kono T,
      3. Rosman H, et al
      . Left ventricular shape: a factor in the etiology of functional mitral regurgitation in heart failure. Am Heart J 1992;123:961–6. doi:10.1016/0002-8703(92)90703-X
      OpenUrlCrossRefPubMedWeb of Science
    9. ↵
      1. Kim J,
      2. Kochav JD,
      3. Gurevich S, et al
      . Left ventricular geometric remodeling in relation to non-ischemic scar pattern on cardiac magnetic resonance imaging. Int J Cardiovasc Imaging 201430:1559–67. doi:10.1007/s10554-014-0487-8
      OpenUrlCrossRefPubMed
    10. ↵
      1. Di Donato M,
      2. Dabic P,
      3. Castelvecchio S, et al
      . Left ventricular geometry in normal and post-anterior myocardial infarction patients: sphericity index and ‘new’ conicity index comparisons. Eur J Cardiothorac Surg 2006;29(Suppl 1):S225–30. doi:10.1016/j.ejcts.2006.03.002
      OpenUrlCrossRefPubMed
    11. ↵
      1. Lamas GA,
      2. Vaughan DE,
      3. Parisi AF, et al
      . Effects of left ventricular shape and captopril therapy on exercise capacity after anterior wall acute myocardial infarction. Am J Cardiol 1989;63:1167–73. doi:10.1016/0002-9149(89)90173-2
      OpenUrlCrossRefPubMedWeb of Science
    12. ↵
      1. Bild DE,
      2. Bluemke DA,
      3. Burke GL, et al
      . Multi-ethnic study of atherosclerosis: objectives and design. Am J Epidemiol 2002;156:871–81. doi:10.1093/aje/kwf113
      OpenUrlCrossRefPubMedWeb of Science
    13. ↵
      1. Yeboah J,
      2. Bluemke DA,
      3. Hundley WG, et al
      . Left ventricular dilation and incident congestive heart failure in asymptomatic adults without cardiovascular disease: Multi-Ethnic Study of Atherosclerosis (MESA). J Card Fail 2014;20:905–11. doi:10.1016/j.cardfail.2014.09.002
      OpenUrlCrossRefPubMed
    14. ↵
      1. Zhang X,
      2. Ambale-Venkatesh B,
      3. Bluemke D, et al
      . Orthogonal shape modes describing clinical indices of remodeling. In: van Assen H, Bovendeerd P, Delhaas T, eds. Functional imaging and modeling of the heart. Springer International Publishing, 2015:273–81.
    15. ↵
      1. Medrano-Gracia P,
      2. Cowan BR,
      3. Bluemke DA, et al
      . Atlas-based analysis of cardiac shape and function: correction of regional shape bias due to imaging protocol for population studies. J Cardiovasc Magn Reson 2013;15:80. doi:10.1186/1532-429X-15-80
      OpenUrlCrossRefPubMed
    16. ↵
      1. Lakoski SG,
      2. Greenland P,
      3. Wong ND, et al
      . Coronary artery calcium scores and risk for cardiovascular events in women classified as “low risk” based on Framingham risk score: the multi-ethnic study of atherosclerosis (MESA). Arch Intern Med 2007;167:2437. doi:10.1001/archinte.167.22.2437
      OpenUrlCrossRefPubMedWeb of Science
    17. ↵
      1. Natori S,
      2. Lai S,
      3. Finn JP, et al
      . Cardiovascular function in multi-ethnic study of atherosclerosis: normal values by age, sex, and ethnicity. AJR Am J Roentgenol 2006;186:S357–65. doi:10.2214/AJR.04.1868
      OpenUrlCrossRefPubMedWeb of Science
    18. ↵
      1. Choi EY,
      2. Rosen BD,
      3. Fernandes VR, et al
      . Prognostic value of myocardial circumferential strain for incident heart failure and cardiovascular events in asymptomatic individuals: the Multi-Ethnic Study of Atherosclerosis. Eur Heart J 2013;34:2354–61. doi:10.1093/eurheartj/eht133
      OpenUrlCrossRefPubMedWeb of Science
    19. ↵
      1. Carr JJ,
      2. Nelson JC,
      3. Wong ND, et al
      . Calcified coronary artery plaque measurement with cardiac CT in population-based studies: standardized protocol of Multi-Ethnic Study of Atherosclerosis (MESA) and Coronary Artery Risk Development in Young Adults (CARDIA) Study. Radiology 2005;234:35–43. doi:10.1148/radiol.2341040439
      OpenUrlCrossRefPubMedWeb of Science
    20. ↵
      1. Wong SP,
      2. French JK,
      3. Lydon A-M, et al
      . Relation of left ventricular sphericity to 10-year survival after acute myocardial infarction. Am J Cardiol 2004;94:1270–5. doi:10.1016/j.amjcard.2004.07.110
      OpenUrlCrossRefPubMedWeb of Science
    21. ↵
      1. Coviello E
      . STCOXGOF: Stata module to produce goodness-of-fit test and plot after a Cox model. 2008. http://EconPapers.repec.org/RePEc:boc:bocode:s456894
    22. ↵
      1. Hosmer Jr.,
      2. David W,
      3. Lemeshow S
      . Applied survival analysis: regression modelling of time to event data (1999). Eur Orthodontic Soc 1999:561–2. doi:10.1016/j.amjcard.2004.07.110
    23. ↵
      1. May S,
      2. Hosmer DW
      . Hosmer and Lemeshow type goodness-of-fit statistics for the Cox proportional hazards model. Handb Stat 2003;23:383–94. doi:10.1016/S0169-7161(03)23021-2
      OpenUrlCrossRef
    24. ↵
      1. Zile MR,
      2. Gaasch WH,
      3. Patel K, et al
      . Adverse left ventricular remodeling in community-dwelling older adults predicts incident heart failure and mortality. JACC: Heart Fail 2014;2:512–22. doi:10.1016/j.jchf.2014.03.016
      OpenUrlAbstract/FREE Full Text
    25. ↵
      1. Li F,
      2. Chen YG,
      3. Yao GH, et al
      . Usefulness of left ventricular conic index measured by real-time three-dimensional echocardiography to predict left ventricular remodeling after acute myocardial infarction. Am J Cardiol 2008;102:1433–7. doi:10.1016/j.amjcard.2008.07.034
      OpenUrlCrossRefPubMed
    26. ↵
      1. Kono T,
      2. Sabbah HN,
      3. Stein PD, et al
      . Left ventricular shape as a determinant of functional mitral regurgitation in patients with severe heart failure secondary to either coronary artery disease or idiopathic dilated cardiomyopathy. Am J Cardiol 1991;68:355–9. doi:10.1016/0002-9149(91)90831-5
      OpenUrlCrossRefPubMedWeb of Science
    27. ↵
      1. Ohyama Y,
      2. Ambale-Venkatesh B,
      3. Noda C, et al
      . Association of aortic stiffness with left ventricular remodeling and reduced left ventricular function measured by magnetic resonance imaging: the multi-ethnic study of atherosclerosis. Circ Cardiovasc Imaging 2016;9:pii: e004426. doi:10.1161/CIRCIMAGING.115.004426
      OpenUrlCrossRef
    28. ↵
      1. Rosen BD,
      2. Edvardsen T,
      3. Lai S, et al
      . Left ventricular concentric remodeling is associated with decreased global and regional systolic function the multi-ethnic study of atherosclerosis. Circulation 2005;112:984–91. doi:10.1161/CIRCULATIONAHA104.500488
      OpenUrlAbstract/FREE Full Text
    29. ↵
      1. Ahmed MI,
      2. Desai RV,
      3. Gaddam KK, et al
      . Relation of torsion and myocardial strains to LV ejection fraction in hypertension. JACC Cardiovasc Imaging 2012;5:273–81. doi:10.1016/j.jcmg.2011.11.013
      OpenUrlAbstract/FREE Full Text
    30. ↵
      1. Yoneyama K,
      2. Gjesdal O,
      3. Choi EY, et al
      . Age, sex, and hypertension-related remodeling influences left ventricular torsion assessed by tagged cardiac magnetic resonance in asymptomatic individuals: the multi-ethnic study of atherosclerosis. Circulation 2012;126:2481–90. doi:10.1161/CIRCULATIONAHA.112.093146
      OpenUrlAbstract/FREE Full Text
    31. ↵
      1. Vasan RS,
      2. Larson MG,
      3. Benjamin EJ, et al
      . Left ventricular dilatation and the risk of congestive heart failure in people without myocardial infarction. N Engl J Med 1997;336:1350–5. doi:10.1056/NEJM199705083361903
      OpenUrlCrossRefPubMedWeb of Science
    32. ↵
      1. Giri S,
      2. White CM,
      3. Dunn AB, et al
      . Oral amiodarone for prevention of atrial fibrillation after open heart surgery, the Atrial Fibrillation Suppression Trial (AFIST): a randomised placebo-controlled trial. Lancet 2001;357:830–6. doi:10.1016/S0140-6736(00)04196-9
      OpenUrlCrossRefPubMedWeb of Science
    33. ↵
      1. Katz AM
      . Cardiomyopathy of overload. A major determinant of prognosis in congestive heart failure. N Engl J Med 1990;322:100–10. doi:10.1056/NEJM199001113220206
      OpenUrlCrossRefPubMedWeb of Science
    34. ↵
      1. Medrano-Gracia P,
      2. Cowan BR,
      3. Ambale-Venkatesh B, et al
      . Left ventricular shape variation in asymptomatic populations: the multi-ethnic study of atherosclerosis. J Cardiovasc Magn Reson 2014;16:56. doi:10.1186/s12968-014-0056-2
      OpenUrlCrossRefPubMed
    35. ↵
      1. Young AA,
      2. Frangi AF
      . Computational cardiac atlases: from patient to population and back. Exp Physiol 2009;94:578–96. doi:10.1113/expphysiol.2008.044081
      OpenUrlCrossRefPubMedWeb of Science

    Footnotes

    • Contributors BA-V: study design, data analysis and interpretation, manuscript drafting. KY, RKS, YO, GLB, SS, ASG, AAY and COW: study design, data interpretation, critical manuscript revision. DAB and JACL: study design, data acquisition and interpretation, critical manuscript revision. All authors read and approved the final manuscript.

    • Funding This research was supported by contracts N01-HC-95159 through N01-HC-95168 from the National Heart, Lung, and Blood Institute.

    • Competing interests None declared.

    • Ethics approval Institutional Review Boards, USA.

    • Provenance and peer review Not commissioned; externally peer reviewed.

    Linked Articles