Author + information
- Michael R. Zile, MD∗,†∗ (, )
- William H. Gaasch, MD‡,
- Kanan Patel, MBBS, MPH§,
- Inmaculada B. Aban, MS, PhD§ and
- Ali Ahmed, MD, MPH§,‖
- ∗Medical University of South Carolina, Charleston, South Carolina
- †Ralph H. Johnson Department of Veterans Affairs Medical Center, Charleston, South Carolina
- ‡Lahey Clinic Medical Center, Burlington, Massachusetts
- §University of Alabama at Birmingham, Birmingham, Alabama
- ‖Department of Veterans Affairs Medical Center, Birmingham, Alabama
- ↵∗Reprint requests and correspondence:
Dr. Michael R. Zile, Medical University of South Carolina, Department of Medicine, Division of Cardiology, Ashley River Towers, Room 7067, 25 Courtenay Drive, Charleston, South Carolina 29425.
Objectives This study sought to determine whether specific patterns of adverse left ventricular (LV) structural remodeling are associated with differential rates of cardiovascular (CV) outcomes.
Background It is not known whether a stepwise combinatorial assessment of LV volume, mass, and geometry done to define specific remodeling patterns provides incremental prognostic information.
Methods A total of 3,181 Cardiovascular Health Study participants (mean age, 73 years of age; 60% women, 5% African American) were categorized by LV remodeling patterns and related to a multivariate-adjusted (age, sex, race, ejection fraction, hypertension, myocardial infarction, diabetes mellitus, chronic kidney disease) analysis of CV outcomes (incident heart failure [HF], all-cause mortality, and a combined endpoint of HF and mortality) over a 13-year follow-up period.
Results Examined independently, either left ventricular enlargement (LVE) or left ventricular hypertrophy (LVH) was associated with a higher risk of HF (32%, 34%, respectively) than with normal geometry (17%; p < 0.001). When LV volume and mass were used in combination, important incremental prognostic information was achieved. In the absence of LVE, HF was more common in those with LVH than in those with normal mass (32% vs. 16%, respectively; p < 0.001). In the presence of LVE, HF was more common in those with LVH than in those with normal mass (37% vs. 29%, respectively; p = 0.021). The subgroup with normal volume and mass but relative wall thickness (RWT) >0.42 carried a higher risk of HF (21%) than those with normal geometry (15%; p = 0.011). Once LVH or LVE was present, the addition of RWT to this analysis did not affect HF rate. Similar results were obtained for the other CV outcomes.
Conclusions Stepwise combinatorial assessment of LV volume, mass, and geometry provides incremental prognostic information regarding CV outcomes.
Adverse structural remodeling of the heart has long been recognized as a pivotal process in the progression of cardiovascular (CV) disease, especially the development of heart failure (HF) (1). This process is not a single event, nor is it homogeneous. It is time-dependent and related to the nature of the injury or load imposed on the ventricle as well as the rapidity of application and the duration of the injury or load. Despite such complexity, the pattern of left ventricular (LV) remodeling can be characterized by a combinatorial analysis of the volume, wall mass, and geometry of the ventricle (2,3). One objective of the current study was to test the applicability of such pattern characterization in a large population of older adults (4). Thus, we determined whether all subjects in this population could be categorized according to specific pattern definitions, and we report the prevalence of each remodeling pattern.
Recognizing that the presence of left ventricular hypertrophy (LVH) or left ventricular chamber enlargement (LVE) have been shown to be risk factors for CV morbidity and mortality (5,6), we considered the possibility that the assessment of LV volume, mass, and geometry in a stepwise and combinatorial fashion might provide more specific risk assessment than that provided by the chamber size or wall mass alone. Accordingly, a second objective of this study was to determine whether the combinatorial assessment of the 3 characteristics of structural remodeling would provide incremental prognostic information.
CHS (Cardiovascular Health Study) is an ongoing, prospective, population-based epidemiologic study of CV disease risk factors in older adults, sponsored by the National Heart, Lung, and Blood Institute. CHS participants include 5,888 community-dwelling adults 65 years of age and older and were recruited from 4 U.S. counties. The rationale and design of the CHS have been previously reported (4). Participants were recruited from a random sample of Medicare-eligible older adults in 2 phases: 1) an initial cohort of 5,201 participants (1989 to 1990); and 2) another 687 African American participants (1992 to 1993). The current analysis is based on a public-use copy of CHS data. Of the 5,888 CHS participants, 93 did not consent to be included in the de-identified public-use data; 274 with prevalent HF were excluded; and 2,340 did not have adequate baseline echocardiographic data. Thus, the remaining 3,181 participants were the subjects of this report. Demographic data from this population are shown in Table 1.
Echocardiographic measurements and calculations
The echocardiographic methods have been described in previous publications (7,8). Measurements included LV end-diastolic and end-systolic dimensions, and end-diastolic posterior wall and septal wall thicknesses. LV end-diastolic volume (using the Teicholz method), wall mass, and ejection fraction (EF) were calculated using previously published formulas (7,8). Relative wall thickness (RWT) was defined as the ratio of the septal plus posterior wall thicknesses to end-diastolic dimension. LVH was defined as the sex-neutral value for wall mass (normalized to height2.7) exceeding 51 g/m2.7 (8). A second or alternate index of LVH was defined as sex-dependent value LV mass (normalized to body surface area) exceeding 115 g/m2 for men and 95 g/m2 for women (8). LVE was defined as a sex-neutral value for LV end-diastolic dimension exceeding 56 mm (7,8). A second or alternate index of LV chamber enlargement was defined as sex-dependent value for LV end-diastolic diameter exceeding 59 mm for men and 53 mm for women (8). The normal range of RWT was defined as 0.32 to 0.42 (3). The normal range for the LV wall mass-to-end-diastolic volume (M/V) ratio was defined as 1.0 to 1.5 (3).
LV structural remodeling categories
The scheme used to categorize LV structural remodeling in the current study was an extension of that proposed by Linzbach (2) and other investigators (3). These definitions of remodeling patterns used in this study differ from those proposed in the American Society of Echocardiography (ASE) guidelines. The rationale, implications, and importance of these differences are presented in the discussion section. In the current study, normal LV architecture was defined as normal chamber size, wall mass, and RWT. Concentric hypertrophy was defined as a normal LV end-diastolic dimension and increased wall mass. Concentric remodeling was defined as having a normal LV end-diastolic dimension, normal wall mass, and a high RWT. Eccentric hypertrophy was defined as an increased end-diastolic dimension and an increased wall mass. Eccentric remodeling was defined as having an increased end-diastolic dimension and a normal wall mass.
New-onset (incident) HF is emphasized in this study, but CV mortality, all-cause mortality, and a combined endpoint of incident HF and all-cause mortality are also reported. The methods used to define HF and CV mortality were described in previous publications (9), and both were centrally adjudicated.
Clinical outcomes were related to structural remodeling in a stepwise and combinatorial fashion. First, outcome data were related to LV chamber size (end-diastolic dimension) or LV wall mass as single structural characteristics; this analysis is shown in Table 2. Second, outcome data were related to LV geometry by separating the subjects into 4 groups: those with normal chamber size (with and without increased wall mass) and those with LV enlargement (with and without increased wall mass); data from these 4 groups are presented in Table 3. Third, the incremental value of adding RWT to the analysis was assessed. This was done in a combinatorial fashion by relating chamber size, wall mass, and RWT to incident of HF. These data are presented in Table 4 and are summarized in Figure 1 and Online Tables 1 and 2. The associations between various remodeling categories and outcomes were examined in an unadjusted fashion, after adjustment for age, sex, and race, and after multivariate adjustment for age, sex, race, ejection fraction, and the presence of hypertension, previous myocardial infarction, diabetes mellitus, and chronic kidney disease. For descriptive analyses, the Pearson chi-square test was used for categorical variables, and ANOVA analysis was used for continuous variables.
Frequencies of HF incidence, all-cause mortality, CV mortality, and a combined endpoint of incident HF and all-cause mortality were compared between various LV remodeling categories by using the Pearson chi-square test. Kaplan-Meier analysis (Figure 2, Online Table 1) and Cox proportional hazard models were used to examine the associations of outcomes with various LV remodeling categories. Outcomes were examined over a 13-year follow-up period. All statistical tests were 2-sided, and results with p values <0.05 were considered significant, except for data presented in Online Tables 1 and 2, where a p value <0.01 was considered significant in order to correct for the large number of multiple pairwise comparisons made in these additional analyses. SPSS software (version 20, release 2011, IBM Corp., Armonk, New York) for Windows (Microsoft, Redmond, Washington) was used for data analysis.
The baseline demographic characteristics in the 4 LV size and mass groups are shown in Table 1, and the echocardiographic data in these 4 groups are shown in Table 2. We examined the association of incident HF (and other outcome data) with the various patterns of LV structural remodeling in a stepwise fashion. First, chamber size (normal vs. LVE) and wall mass (normal vs. LVH) data were tabulated and incident HF was examined in each of these 4 groups (Table 2). Second, we examined the impact of wall mass in subjects with and without LVE (Table 3). Third, the assessment was expanded in a combinatorial fashion by adding RWT to the chamber size and wall mass analysis (Table 4, Online Tables 1 and 2).
Baseline demographic characteristics
Demographic information is presented for groups with and without LVE and for groups with and without LVH (Table 1). Chamber enlargement was seen in 16% and increased wall mass was present in 15%. There were significantly more comorbidities and more medication use in the group with LVE than in those with a normal chamber size. A similar pattern of more comorbidities and medication use was seen in the group with LVH compared with the group with normal LV wall mass.
Chamber size, wall mass, and incident HF
As shown in Table 2, baseline echocardiographic data from groups with LVE exhibit an eccentric geometry distinctly different from that seen in those without LVE. There was significantly more incident HF in the group with LVE than in those with a normal LV chamber size (32% vs. 17%, respectively; unadjusted hazard ratio [HR]: 2.24; 95% confidence interval [CI]: 1.86 to 2.69; p < 0.001). Likewise, there was significantly more incident HF in the group with LVH than was seen in those with a normal LV mass (34% vs. 17%, unadjusted HR: 2.62; 95% CI: 2.18 to 3.16; p < 0.001). Thus, the mere presence of LVE or increased wall mass was associated with a 2-fold increase in incident HF. These results were virtually the same when LVE and LVH were defined using sex-specific and body surface area specific partition values (second/alternate analysis, see the Methods section). For example, analysis in which LVH was defined using sex-neutral versus sex-specific partition values of LV mass are presented in Online Table 3.
Most subjects in this study exhibited a normal or nearly normal LV EF (Table 2). In the absence of LVE and/or LVH, the EF was normal or nearly normal (>45%) in 99% of the subjects. When LVE and/or LVH was present, 90 to 92% had an EF >45%. It appears, therefore, that the incident HF seen in subjects with adverse structural remodeling was not closely related to LV systolic dysfunction at the time of the baseline study.
When the associations between LVE and LVH and incident HF described above were adjusted for baseline demographic variables and EF, the differences remained statistically significant. After adjustment for age, sex, race, and EF and the presence of hypertension, myocardial infarction, diabetes mellitus, and chronic kidney disease, incident HF was more common in patients with either LVE (HR: 1.95; 95% CI: 1.60 to 2.39; p < 0.001) or LVH (HR: 2.24; 95% CI: 1.85 to 2.70; p < 0.001).
Fifty-three of the 617 patients who developed incident HF during follow-up had an incident myocardial infarction before developing HF; this included patients with incident HF and a myocardial infarction at the same time. When incident myocardial infarction was added to the multivariate adjustment model, none of the data shown in Tables 3 and 4 were altered either numerically or statistically.
Impact of combining wall mass and chamber size data
The stepwise analysis shown in Table 3 emphasized the association between LVH with incident HF (and other outcome data) in groups with and without LVE. In the group with normal chamber size, those with LVH had significantly more incident HF than was seen in those without LVH (32% vs. 16%, respectively, HR: 2.60; 95% CI: 2.03 to 3.32; p < 0.001). In the group with LVE, incident HF was more common in those with LVH than in those without LVH (37% vs. 29%, respectively, HR: 1.57; 95% CI: 1.14 to 2.16; p = 0.006).
When these associations were adjusted for baseline demographic variables and EF, the differences remained statistically significant (Table 3). For example, after adjustment for age, sex, race, EF, and the presence of hypertension, myocardial infarction, diabetes mellitus, and chronic kidney disease, incident HF was higher in the group with normal LV end-diastolic dimension and LVH than in those with a normal chamber size and no LVH (HR: 1.71; 95% CI: 1.33 to 2.21; p < 0.001). Differences among the 4 remodeling groups in all-cause mortality, CV mortality, and the combined endpoint of incident HF and all-cause mortality showed similar trends (Table 3, Online Table 1). For example, all-cause mortality was higher in the group with normal LV end-diastolic dimension and LVH than in those without LVH (55% vs. 38%, respectively, unadjusted HR: 1.78; 95% CI: 1.48 to 2.14; p < 0.001; HR multivariate adjusted: 1.29; 95% CI 1.07 to 1.55, p = 0.009). The Kaplan-Meier analysis shown in Figure 2 likewise indicates that the highest likelihood of developing incident HF over 13 years of follow-up would occur in patients with LVE and/or LVH.
Impact of adding relative wall thickness to the analysis
In Table 4 and Figure 1, each of the 4 groups analyzed in Table 3 was divided into subgroups with high and low values for RWT. This allowed an assessment of the impact of all 3 parameters (in combination) on incident HF (and other outcome data). In those with a normal LV chamber size and normal wall mass, there is significantly more incident HF in the subgroup with RWT >0.42 (concentric remodeling) than in the subgroup with RWT ≤0.42 (normal geometry; 21% vs. 15%, respectively, HR: 1.58; 95% CI: 1.24 to 2.00; p < 0.001). In those with a normal LV chamber size and increased wall mass (concentric hypertrophy), there is a nonsignificant tendency for more incident HF in the subgroup with RWT >0.42 than in the subgroup with RWT <0.42 (34% vs. 30%, HR: 1.41; 95% CI: 0.90 to 2.22; p = 0.137). All-cause mortality and the combined endpoint of incident HF and all-cause mortality were significantly higher in the subgroup with RWT >0.42 (Table 4, Online Table 1).
In those with LVE and no hypertrophy (eccentric remodeling), incident HF was 29%; the subgroup with RWT <0.32 had a nonsignificant tendency for more incident HF than those with RWT >0.32 (30% vs. 20%, respectively, HR: 1.42; 95% CI: 0.58 to 3.52; p = 0.446). In the group with LVE and LVH (eccentric hypertrophy), incident HF was 37% and was similar in the 2 RWT subgroups (36% vs. 37%, respectively, HR: 0.94; 95% CI: 0.58 to 1.52; p = 0.785).
When these associations using the addition of RWT were adjusted for baseline demographic variables using a multivariate adjustment, the results were similar. In addition, differences in all-cause mortality and the combined HF endpoint of incident HF and all-cause mortality showed similar associations (Table 4, Online Tables 1 and 2). All pairwise comparisons within the normal LV Left ventricular end-diastolic dimension (EDD) and increased LV EDD groups were made and presented in Online Tables 1 and 2. These analyses allowed examination of the interactive effects of volume and mass with RWT to be examined.
The results of the new classification scheme presented in this study were directly compared with results using the American Society of Echocardiography (ASE) classification scheme with respect to 2 parameters: 1) the number of subjects reclassified; and 2) the effects of this reclassification on risk prediction for incident HF. Twenty-six percent of the subjects were reclassified. The ASE classification did not include 3 classifications that were included in the new classification scheme: mixed hypertrophy, physiologic hypertrophy, and eccentric remodeling; 11% of the population was reclassified into these 3 new classification categories. In addition, the ASE scheme overestimated the number of subjects classified as normal geometry (73% prevalence using ASE vs. 64% using the new scheme) and overestimated the number of patients classified with eccentric hypertrophy (10% prevalence using ASE vs. 3.6% using the new scheme).
These differences in classification were associated with differences in rates of incident HF. For example, some subjects who were classified as having normal geometry using the ASE criteria with a predicted incident HF of 17% were reclassified as having eccentric remodeling with an incident HF of 30%. This reclassification is responsible, in part, for the fact that the incident HF in the normal geometry group was overestimated using ASE criteria. In addition, some subjects who were classified as having eccentric hypertrophy by using ASE criteria were reclassified as having physiologic hypertrophy; this reclassification is responsible, in part, for the fact that the incident HF in the eccentric hypertrophy group was underestimated by 10% using ASE criteria.
Our analysis of echocardiographic and outcome data from the CHS allows several conclusions. First, we confirmed that the presence of LVE alone or LVH alone is associated with almost twice as much incident HF as was seen in the absence of these abnormalities. Such associations, first recognized by the Framingham investigators, have since been confirmed by others (5,6,9–15). These reports, however, did not use measurements of chamber size and wall mass in the combinatorial approach that was used herein. In the presence of a normal LV chamber size, incident HF was twice as frequent in those with hypertrophy (concentric hypertrophy) as that seen in those with normal LV mass. Likewise with LVE, the presence of hypertrophy (eccentric hypertrophy) was associated with more incident HF than was seen in those without hypertrophy. We also found significantly more incident HF in the group with eccentric hypertrophy than in the group with concentric hypertrophy. This relatively simple method of categorizing LV structure allowed us to classify 100% of the population and to separate all subjects into 4 groups with highly significant differences in the incidence of HF. These differences in incident HF (all-cause mortality, CV mortality, and the combined endpoint of incident HF and all-cause mortality) remained significant even after adjustment for age, sex, race, EF, and the presence of hypertension, myocardial infarction, diabetes mellitus, and chronic kidney disease.
Second, when RWT was added to the analysis in a stepwise combinatorial manner, 96% of the population could be categorized according to the specific definitions of the 7 subgroups (Figure 1). In the subjects with normal chamber size and normal wall mass, the subgroup with RWT exceeding 0.42 (concentric remodeling) had significantly more incident HF than those with a normal architecture. This association of concentric remodeling with adverse outcomes, first reported by Ganau et al. (16), has been confirmed in other reports (3,17–20). Incident HF was even higher in those with concentric hypertrophy. Subdividing these subjects with concentric hypertrophy into subgroups with high and normal RWT, suggests a trend toward more incident HF and a significantly higher mortality when RWT exceeds 0.42 (Table 4). However, in the subjects with LVE, with or without hypertrophy, subgroup analysis according to RWT did not indicate significant differences in incident HF. We conclude, therefore, that combining chamber size, wall mass, and RWT in a stepwise and combinatorial manner provides incremental value when the LV chamber size is normal, but not when LVE is present. Moreover, this incremental value remained significant even after adjustment for age, sex, race, EF and the presence of hypertension, myocardial infarction, diabetes mellitus, and chronic kidney disease. The most important finding was that the risk of incident HF is closely related to the pattern of adverse structural remodeling of the ventricle, not merely to the presence or absence of LVE or LVH.
Definitions of remodeling patterns
The definitions of the remodeling patterns used in this study differ from those proposed in the American Society of Echocardiography (ASE) guidelines. The ASE guidelines define populations on the basis of the presence or absence of LVH and then use a RWT partition value of 0.42 to subdivide the population into those with normal geometry, concentric remodeling, concentric hypertrophy, and eccentric hypertrophy. A RWT of 0.42 was proposed as the upper limit of normal with a RWT ≥0.42, indicative of concentric geometry. In addition, RWT <0.42 was proposed to indicate the presence of eccentric geometry. However, virtually all normal hearts have RWT in the range of 0.32 to 0.42, and most dilated/failing hearts with eccentric geometry have RWT in this same range or lower. Thus, the definitions proposed by the ASE guidelines do not provide a classification that differentiates physiologic hypertrophy from eccentric remodeling from eccentric hypertrophy. In an attempt to rectify this limitation, we applied the ideas and methods described by Linzbach (2) to develop a classification scheme and examined the utility of this scheme by applying it to the CHS database of community dwelling older adults. The classification scheme used in the current study relies primarily on measurements of LV chamber size and wall mass, and we limited the use of the term “eccentric” to ventricles with an LV chamber size that exceeded normal. In addition, this scheme uses an RWT of 0.32 as the lower limit of normal.
Our study, which includes a wide spectrum of remodeling patterns, indicates that the stepwise combinatorial analysis of LV chamber size and wall mass carries important prognostic value in a general population of older adults and that subgroup analysis by inclusion of RWT provides added utility when the LV chamber is not enlarged. The limited utility of separating the subjects with LVE into subgroups on the basis of RWT was unexpected, but there may be several reasons why this is the case. First, relatively small measurement errors can lead to significant differences in RWT, which can result in subgroup misclassification. For example, a typical dilated ventricle with an end-diastolic dimension of 62 mm and a wall thickness of 9 mm would have an abnormal/low RWT of 0.29. If the chamber dimension was measured at 60 mm and the thickness increased by a corresponding 2 mm, the result would be a normal RWT of 0.37. Second, specific information about the diagnosis/etiology (e.g., hypertension, mitral regurgitation, coronary heart disease, and so on) and medication use as well as sex and physical activity were not available for inclusion in our analysis. It is also important to consider the possibility that changes in LV architecture (e.g., development of a new disorder or progression of established disease) could have occurred after the baseline data were collected. Third, it is possible that the limits used to define the subgroups are imperfect, or even that the classification scheme itself is flawed. It appears that the upper limit of normal for RWT (0.42) is an appropriate normative value, but the lower limit of RWT (0.32) is less certain (3). Fourth, the power of this analysis may have been limited by small sample sizes and the number of events in the various subgroups. Despite this, our data confirm the prognostic utility of a classification on the basis of a sequential analysis of chamber size, wall mass, and RWT and have potential clinical implications. For example, the approach to the management of patients with LV remodeling, but without overt HF (Stage B), may be tailored by the prognosis attendant to the degree and classification of remodeling.
Concentric remodeling and hypertrophy
The systolic pressure load imposed on the left ventricle by systemic arterial pressure or aortic valve stenosis generally results in an increment in ventricular mass, often with a high RWT (concentric hypertrophy). The earliest change appears to be an increase in RWT before there is a measurable increase in LV mass (concentric remodeling). These architectural changes provide a mechanism for maintenance of normal LV systolic wall stress in the presence of a high systolic pressure, and they are said to indicate that hypertrophy “develops in a pattern that is unique to the inciting overload” (21). Such preservation of systolic wall stress allows maintenance of normal or nearly normal LV systolic function and performance (22,23). Thus, concentric remodeling/hypertrophy is generally thought to be a compensatory adaptation to LV systolic pressure overload, but unfortunately, it yields a substrate for LV diastolic dysfunction and diastolic HF.
It should be recognized, however, that “compensatory” pressure overload hypertrophy may not always be essential to the maintenance of normal LV systolic function and performance. For example, experimental studies indicate that it is possible to inhibit or even prevent the development of hypertrophy with little or no depression of systolic function, despite the presence of systolic pressure overload (24,25). Clinicians have also recognized the absence of hypertrophy in some patients with severe aortic stenosis and normal ejection fraction (26). Such observations suggest an augmented contractile state that allows maintenance of systolic function despite a systolic pressure overload, but the underlying mechanisms are unknown. If mechanisms other than hypertrophy can allow maintenance of normal systolic function despite high levels of afterload, we must recognize this as a potentially limiting factor in our attempts to classify subjects on the basis of specific patterns structural remodeling. It should also be acknowledged that our analysis did not specifically dissect out a subgroup with excessive or inappropriate hypertrophy, which is said to be associated with poor clinical outcomes (27,28). These caveats notwithstanding, our analysis of the population with normal LV chamber size confirms a progressive increase in incident HF in the normal (15%), concentric remodeling (21%), and concentric hypertrophy (34%) groups.
The diastolic volume overload imposed on the ventricle in dilated cardiomyopathies, coronary heart disease, regurgitant lesions of the mitral or aortic valves, and in some athletes and pregnant females typically results in LVE (eccentric geometry) with an increase in mass (eccentric hypertrophy). Under these conditions, RWT ranges from normal to low. In athletes, pregnant women, and people with the early/compensated phase of chronic mitral regurgitation, RWT remains normal. Such “physiologic” hypertrophy is thought to provide a compensatory mechanism for maintenance of normal systolic wall stress and normal EF. Under other conditions, the magnitude of chamber enlargement tends to exceed the degree of hypertrophy, and although the RWT ranges widely, it is most often low (eccentric hypertrophy). In acute myocarditis and other apoptotic conditions, LVE is prominent, but the hypertrophic process can be blunted, and there may be little or no increase in mass (eccentric remodeling). In both of these situations, afterload excess contributes to the depressed LV systolic performance and to progressive adverse remodeling. Moreover, in coronary cardiomyopathy, the extreme nonuniformity of the LV wall in combination with marked variability in the extent or degree of adverse remodeling would be expected to contribute to considerable overlap of RWT in our subgroups with eccentric geometry. This may be one reason why the results of the VALIANT study in patients with ischemic heart disease indicate that the risk of incident HF was approximately twice as high in concentric hypertrophy than in eccentric hypertrophy (10), whereas our results indicate the opposite.
Recognizing the disappointing performance of adding RWT to our sequential classification scheme in eccentric geometry, the accuracy and reproducibility of our echocardiographic data might be questioned. Perhaps another population could be studied using cardiac magnetic resonance imaging, which many authorities consider the most accurate method available. Such studies require attention to specific methodological details regarding definition of the endocardium. For example, when the trabecular and papillary myocardium is included in the LV wall mass rather than the chamber volume, the mass increases by 23% and the volume decreases by a corresponding amount of 28% (29,30). Such differences in normal subjects result in a normal M/V ratio of 1.3:1.4 when the trabecular and papillary myocardium is included in the mass and an abnormal to low M/V ratio of 0.8 when the trabecular and papillary myocardium is included in the volume. These M/V ratios correspond to RWT values of 0.38 and 0.28, respectively (normal range is 0.42 to 0.32) (3). Differences of this magnitude obviously have a significant impact on any analysis that incorporates these parameters (3,29,30). If a goal is to develop normative values that accurately distinguish subjects with different LV architecture, future studies should use values for chamber size and wall mass that do not include trabeculae and papillary muscles in the LV chamber volume.
A combinatorial assessment of LV chamber size and wall mass is valuable in the identification of older adults who are at risk of developing HF. This remains true even after adjustment for age, sex, race, and EF. Such an assessment allows classification of virtually all subjects with or without heart disease. The addition of RWT to the analysis of LV remodeling patterns and outcomes appears to have incremental value only when the LV is not enlarged. This latter finding is not what we would have predicted on the basis of our previous review of published information (3) and until other data become available, we would not emphasize RWT in hearts exhibiting eccentric geometry.
For supplemental tables, please see the online version of this article.
Dr. Zile is supported by the Research Service of the Department of Veterans Affairs (5101CX000415-02 and 5101BX000487-04). Dr. Ahmed is supported by National Institutes of Health R01-HL097047 from the National Heart, Lung, and Blood Institute and a generous gift from Ms. Jean B. Morris of Birmingham, Alabama. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- ejection fraction
- heart failure
- hazard ratio
- left ventricular
- left ventricular enlargement
- left ventricular hypertrophy
- relative wall thickness
- Received March 6, 2014.
- Accepted March 21, 2014.
- American College of Cardiology Foundation
- Cohn J.N.,
- Ferarri R.,
- Sharpe N.
- Gaasch W.H.,
- Zile M.R.
- Feigenbaum H.
- Lang R.M.,
- Bierig M.,
- Devereux R.B.,
- et al.
- Gottdiener J.S.,
- Arnold A.M.,
- Aurigemma G.P.,
- et al.
- Verma A.,
- Meris A.,
- Skali H.,
- et al.
- Udelson J.E.,
- Konstam M.A.
- Konstam M.A.,
- Kramer D.G.,
- Patel A.R.,
- Maron M.S.,
- Udelson J.E.
- Kramer D.G.,
- Trikalinos T.A.,
- Kent D.M.,
- Antonopoulos G.V.,
- Konstam M.A.,
- Udelson J.E.
- Zile M.R.,
- Gottdiener J.S.,
- Hetzel S.J.,
- et al.
- Ganau A.,
- Devereux R.B.,
- Roman M.J.,
- et al.
- Verdecchia P.,
- Schillaci G.,
- Borgioni C.,
- et al.
- Krumholz H.M.,
- Larson M.,
- Levy D.
- Quinones M.A.,
- Gaasch W.H.,
- Cole J.S.,
- Alexander J.K.
- Dumesnil J.G.,
- Shoucri R.M.
- Hill J.A.,
- Karami M.,
- Kutschke W.,
- et al.
- Esposito G.,
- Rapacciuolo A.,
- Nagaprasad S.V.,
- et al.
- Kupari M.,
- Turto H.,
- Lommi J.
- Orsinelli D.A.,
- Aurigemma G.P.,
- Battista S.,
- Krendel S.,
- Gaasch W.H.
- Cioffi G.,
- Faggiano P.,
- Vizzardi E.,
- et al.
- Mor-Avi V.,
- Jenkins C.,
- Kuhl H.P.,
- et al.