The Associations of Body Size and Body Composition With Left Ventricular Mass: Impacts for Indexation in Adults

451 LEFT VENTRICULAR HYPERTROPHY The Associations of Body Size and Body Composition With Left Ventricular Mass: Impacts for Indexation in Adults HANS-WERNER HENSE, MD, BIRGIT GNEITING, MSC,* MICHAEL MUSCHOLL, MD, ULRICH BROECKEL, MD, BERNHARD KUCH, MD, ANGELA DOERING, MD,* GÜNTER A. J. RIEGGER, MD, HERIBERT SCHUNKERT, MD Germany Objectives. We investigated the relationship between body size, body composition and left ventricular mass (LVM) in adults, and assessed the impact of different indexations of LVM on its associations with gender, adiposity and blood pressure. Background. The best way to normalize LVM for body size to appropriately distinguish physiologic adaptation from morbid heart morphology was discussed. Methods. We undertook a community survey of 653 men and 718 women, aged 25 to 74 years. Lean body mass (LBM) was determined by bioelectric impedance analyses and LVM was assessed by two-dimensional guided M-mode echocardiography. Results. After traditional indexations to body height, body height 2.7, or body surface area, men had higher LVM than women (p < 0.001). These gender differences disappeared (p > 0.05) when LVM was indexed to LBM. The type of indexation also modified the strength of the association between adiposity and LVM. The estimated impact of body fat on LVM indexed to LBM was less than half that obtained with traditional indexations. In contrast, the magnitude of the associations of blood pressure with LVM was entirely independent of the type of indexation. Conclusions. This study showed the prominent influence of body composition on adult heart size. Indexation for LBM removed gender differences for LVM and reduced the impact of adiposity, but left the effects of blood pressure unchanged. We suggest that this approach be used for clinical and research applications. (J Am Coll Cardiol 1998;32:451 7) 1998 by the American College of Cardiology Increases in left ventricular mass (LVM) and left ventricular hypertrophy (LVH) are independent and strong predictors of cardiovascular morbidity and mortality (1 7). The occurrence of LVH is determined by blood pressure levels (8 10) and overweight, either alone or with hypertension (11 16). A number of reports have concluded that the influence of obesity on LVM may even exceed that of hypertension (15,17,18). However, the complex relations between body size, body composition and physiologic adaptation of the heart make the assessment of abnormal LVM a difficult task (19,20). Comparative studies have suggested that heart size follows body growth to accommodate the greater metabolic demands of larger bodies. Although lean body mass (LBM) largely determines LVM in children irrespective of gender (21), the LVM From the Institute of Epidemiology and Social Medicine, Clinical Epidemiology Unit, University Münster, Münster; *Institute for Epidemiology, GSF National Research Center, Munich-Neuherberg, Department of Internal Medicine II, University Regensburg, Regensburg; Germany. This study was supported by the Bundesministerium für Forschung und Technologie grant no. KBF01-GB9403 and Deutsche Forschungsgemeinschaft (Schu 672/9-1, Schu 672/10-1), Germany. Manuscript received September 26, 1997; revised manuscript received April 1, 1998, accepted April 17, 1998. Address for correspondence: Prof. Dr. Hans-Werner Hense, University Münster, Institute of Epidemiology and Social Medicine - Clinical Epidemiology Unit-, D-48129 Münster, Germany. E-mail: hense@uni-muenster.de. differences of male and female adults seems to be explained by physiologic cardiac hypertrophy in men, which occurs in parallel to the gender divergence in body growth after puberty (22). There is considerable controversy over the optimal method for normalizing LVM to body size. The indexation to body surface area is very common (23,24), but it has been criticized for disregarding the effects of obesity (12,20,25). Alternatively, the indexation of LVM to body height was proposed (18,25). Subsequent refinements introduced allometric signals to better account for the nonlinear association of body size with LVM (20,22,26,27). In theory, the best option is indexation to LBM because the latter represents the fat-free body compartments and their dominating metabolic demands (17,20,21,27,28). Indexing to LBM should enable a better distinction of the physiologic adaptations of LVM from morbid alterations due to adiposity or hypertension. However, the difficulties of validating LBM in most studies have impeded its evaluation in adults to date. We present data from a community-based survey of men and women, aged 25 to 74 years, who underwent echocardiography and bioelectric impedance analysis (BIA) for the estimation of LBM. Various indexations, including that for LBM, were evaluated in terms of their influence on the associations of gender, adiposity and blood pressure with LVM. 1998 by the American College of Cardiology 0735-1097/98/$19.00 Published by Elsevier Science Inc. PII S0735-1097(98)00240-x

452 HENSE ET AL. JACC Vol. 32, No. 2 LVM AND BODY COMPOSITION Abbreviations and Acronyms BIA bioelectrical impedance analysis BMI body mass index LBM lean body mass LVH left ventricular hypertrophy LVM left ventricular mass Methods Study population. The subjects of this study participated in the Monitoring trends and determinants in cardiovascular disease (MONICA) Augsburg survey from October 1994 to June 1995. The MONICA Augsburg project is part of the international World Health Organization (WHO) MONICA Project; its objective is to assess the trends and determinants of morbidity and mortality of cardiovascular diseases in a defined region of Southern Germany by use of multiple, independent population surveys and the establishment of a populationbased myocardial infarction registry. The study design, sampling frame and data collection have been described in detail elsewhere (29,30). Briefly, 6,640 individuals, aged 25 to 74 years, were randomly sampled by two-stage, age-sex stratified cluster sampling from the population registry of the city of Augsburg and two adjacent counties. A total of 4,856 men and women (74.9%) of all eligible patients participated. The present study was organized as a substudy, for which only the 2,376 participants residing in Augsburg were offered an additional echocardiographic examination. A total of 1,678 individuals (substudy response 70.6%) agreed to be examined. Interview and anthropometric measurements. Data were obtained by interview, physical examination, BIAs, and echocardiography. The interview comprised questions on the patient s own and family medical history, life style, behavioral and psychosocial factors, and medications used the week before the examination. Blood pressure at rest was measured in a highly standardized fashion (16) with random zero sphygmomanometers after subjects had been sitting for at least 30 minutes. Blood pressure was measured to the nearest even digit three times on the right arm. The mean of the second and third measurements were used for this study. Body height and weight were measured in light clothing with calibrated stadiometers and balance scales. Height was measured at 0.5 cm and weight at 0.5 kg intervals. Body mass index (BMI) was calculated as weight in kilograms divided by height in meters squared (kg/m 2 ). Bioelectric impedance analyses. Lean body mass was determined by measurement of bioelectric impedance with a Body Composition Analyzer TVI-10 (Danninger Medical Technology, Heidelberg, Germany). Bioelectric impedance analyses were performed under highly standardized conditions (31) with all subjects in a supine position. All measurements were performed with an alternating current with a frequency of 50 khz and an amplitude of 800 ma. Tetrapolar placement of electrodes was used (32). Bioelectric impedance analysis was based on the resistance of body tissue to the flow of an applied alternating current. The LBM of men and women was calculated from equations containing the resistance (in ohms), age, body height and body weight. We employed two formulas obtained from the validation studies of Segal et al. (33) in United States adults, aged 17 to 62 years, using densitometry as the reference method, and in adult Danes aged 35 to 64 years, by Heitmann (34), who used a four-compartment model to estimate LBM in kilograms. Body fat was calculated after subtraction of LBM from total body weight in kilograms. There were hardly any differences between the two methods in terms of estimated mean values of LBM and its sample variation. We selected the Heitmann equations (34) for the present analyses, because in our sample the correlations of LBM, body fat and relative body fat with age and BMI were slightly more consistent for men and women than the formulas of Segal et al. (33). Analyses of the intraand interobserver variability of BIA measurements indicated a high reliability with coefficients of variation consistently below 1% (31). Echocardiographic measurements. Two-dimensional guided M-mode echocardiograms were performed by two expert sonographers (M.M., U.B.) using the Sonos 1500 (Hewlett Packard Inc., Andover, Massachusetts) M-mode tracings were recorded on strip chart paper at 50 mm/s. All M-mode tracings were analyzed by a single cardiologist (M.M.) who was unaware of the clinical data. Measurements were made according to the Penn convention and LVM was calculated as described by Devereux and Reichek (23) as: LVM (in grams) 1.04 [(LVED SWT PWT) 3 LVED 3 ] 13.6, where LVED is the left ventricular end-diastolic diameter, SWT is the septal wall thickness and PWT is the posterior wall thickness (each given in millimeters). The rank correlation for 144 duplicate LVM measurements of the two sonographers was 0.91. Bland-Altman plots (35) revealed a mean difference (systematic bias) between both observers of 0.9 g with a SD of 10.8 g. Statistical analyses. Men and women were compared with respect to the mean values and SDs of their anthropometric characteristics, i.e., body weight, body height, LBM, body fat, body surface area, and BMI, and in addition for age, systolic and diastolic blood pressure and crude and indexed LVM. For indexation, LVM was divided by body height, by body surface area (36), by body height to the power of 2.7 and 2.0, and by LBM. Differences between men and women in crude and indexed LVM were assessed in 10-year age groups and overall by means of t tests. Additionally, multivariate linear regression analyses were computed using the data of male and female study subjects to estimate gender differences in crude and indexed LVM, adjusting for age, body fat and systolic blood pressure. Correlation analyses evaluated the strength of the linear associations of the anthropometric factors and systolic and diastolic blood pressure with crude and indexed LVM. Partial correlation coefficients, controlling for age, are reported for men and women. Furthermore, multivariate linear

HENSE ET AL. LVM AND BODY COMPOSITION 453 Table 1. Means and SDs of Anthropometric Measurements and Blood Pressure (BP), by Gender Men (n 653) Women (n 718) Male Female Difference Mean SD Mean SD (p Value) Age (yr) 49.6 13.9 49.3 13.6 0.80 Body height (m) 1.75 0.62 1.62 0.65 0.001 Body weight (kg) 82.1 10.7 68.5 11.8 0.001 Body surface area (m 2 ) 1.97 0.14 1.72 0.14 0.001 BMI (kg/m 2 ) 26.8 3.3 26.3 4.7 0.01 LBM (kg) 59.9 5.5 43.8 4.4 0.001 Body fat (kg) 22.1 6.8 24.7 9.0 0.001 Systolic BP (mm Hg) 135.9 18.6 129.7 20.3 0.001 Diastolic BP (mm Hg) 82.7 11.5 78.1 11.1 0.001 regression analyses were performed separately in men and women; these included age, body fat and systolic blood pressure as independent and crude and indexed LVM as dependent variables. Estimations for LVM changes associated with increased body fat and systolic blood pressure by 1 SD were obtained from these multivariate analyses. The results provide an impression of the quantitative impact that body fat and systolic blood pressure exert on LVM in men and women based on their variability in this population. Left ventricular mass changes are expressed as absolute values and as a percentage of the respective mean values. Results for the indexation of LVM to height 2.7 and height 2.0 were very similar in our study and only data indexed to height 2.7 are presented. Likewise, because systolic blood pressure showed the stronger and more consistent associations with LVM than diastolic or mean blood pressure, multivariate analyses were run with systolic blood pressure only. All analyses were carried with the SAS System for Windows Release 6.10 (Carey, North Carolina). Results Baseline characteristics. Baseline characteristics of the study sample are given in Table 1. In the absence of age differences, men had higher average values for body height, body weight, body surface area, BMI and LBM, whereas body fat was significantly higher in women. Men had also higher mean blood pressure values. Hypertension was more common in men (systolic 140 mm Hg or diastolic 90 mm Hg; 42.5% of men vs. 30.5% of women, p 0.001), whereas the proportion of men and women taking antihypertensive medications were similar (16% vs. 18%, respectively). Indexation and gender differences. The unindexed mass of male left ventricles was substantially greater than in females (Table 2). Indexations of LVM to height, to height 2.7 or to body surface area resulted in attenuations of this gender difference. Expressed in relative terms, the male female difference for crude LVM was greatest with 28.9%, whereas the most pronounced reduction was achieved with indexation to height 2.7. Nevertheless, gender differences remained quantitatively substantial, highly statistically significant and unaltered by additional adjustment for age, body fat and systolic blood pressure. On the other hand, indexation of LVM to LBM reduced male to female differences to marginal, nonsignificant amounts (0.9 g/10 kg of LBM or 2.7%, p 0.09), which persisted after multivariate adjustments (Table 2). Analyses by 10-year age group revealed that significant differences between men and women occurred at any age with conventional indexations in univariate and multivariate analyses (Fig. 1). In contrast, gender differences for the LBM indexation were significant only in the youngest age group; these were eliminated after control for blood pressure differences. Indexation and the associations with adiposity and hypertension. Partial correlation coefficients, controlled for age, were compared to investigate how the type of indexation affected the associations of LVM with measurements of body size, body composition and blood pressure (Table 3). Crude LVM showed strong and positive correlations with all variables considered. Indexation to height affected correlations with weight, body fat and BMI only moderately. Likewise, indexation to height 2.7, although attenuating the strength of the associations with weight and body fat, was identical with crude LVM in respect to the association with BMI. In contrast, indexation to LBM quite effectively reduced the correlations of LVM with any of the measurements of body size and composition in both men and women. Interestingly, LVM indexation Table 2. Mean Values and SDs of Unindexed and Indexed Left Ventricular Mass (LVM) by Gender, and Absolute and Relative Gender Differences Men (n 653) Women (n 718) Male Female Difference Measures of LVM Mean SD Mean SD Absolute Relative* Adjusted absolute LVM unindexed (g) 201.1 60.7 142.9 46.5 58.2 28.9 62.8 0.0001 LVM indexed to height (g/m) 115.1 35.2 88.6 29.5 26.5 23.0 28.6 0.0001 LVM indexed to height 2.7 (g/m 2.7 ) 44.7 14.4 39.4 14.0 5.3 11.9 5.9 0.0001 LVM indexed to body surface area (g/m 2 ) 101.9 28.8 82.7 24.3 19.2 18.8 19.2 0.0001 LVM indexed to LBM (g/10 kg) 33.6 9.9 32.7 10.3 0.9 2.7 0.9 0.06 *Difference expressed as percentage of male mean ([men women]/men). Adjusted for age, body fat and systolic blood pressure by multivariate linear regression analysis. p Value of gender term in the above multivariate model. p Value

454 HENSE ET AL. JACC Vol. 32, No. 2 LVM AND BODY COMPOSITION Figure 1. Mean LVM indexed to height, height 2.7, body surface area and LBM, by 10-year age groups in men and women. *p 0.001 men versus women. to body surface area resulted in partial correlations that were similar to those obtained with LBM (Table 3). Of note, comparisons of the correlations between indexed LVM and systolic or diastolic blood pressure demonstrated that all coefficients were of equal magnitude and, therefore, apparently independent of the type of indexation. Stronger associations were found in men and women with systolic rather than with diastolic blood pressure (Table 3). The LVM changes in response to increased body fat or systolic blood pressure by 1 SD were estimated separately for men and women from multivariate regressions controlling for age (Table 4). The results are expressed in absolute and relative terms. Although the predicted impact of body fat was most pronounced on crude, height and height 2.7 indexed LVM (between 10.3% and 14.8% of the sample mean values), the indexations to LBM, as well as to body surface area, reduced Table 3. Partial Correlation Coefficients (r), Controlled for Age, of Body Size, Body Composition and Blood Pressure With Unindexed and Indexed Left Ventricular Mass (LVM) LVM Unindexed (g) to Height (g/m) to Height 2.7 (g/m 2.7 ) to Body Surface Area (g/m 2 ) to LBM (g/10 kg) Variables r r r r r Men Body weight (kg) 0.43* 0.38* 0.28* 0.21* 0.16* Body height (m) 0.15 0.03 0.16* 0.02 0.07 LBM (kg) 0.37* 0.29* 0.15* 0.14* 0.06 Body fat (kg) 0.43* 0.41* 0.35* 0.23* 0.22* BMI (kg/m 2 ) 0.41* 0.42* 0.42* 0.25* 0.23* Systolic BP (mm Hg) 0.23* 0.25* 0.27* 0.25* 0.26* Diastolic BP (mm Hg) 0.16* 0.16* 0.16* 0.13* 0.13* Women Body weight (kg) 0.52* 0.48* 0.39* 0.27* 0.24* Body height (m) 0.11 0.01 0.22* 0.06 0.12* LBM (kg) 0.46* 0.38* 0.22* 0.21* 0.13* Body fat (kg) 0.49* 0.48* 0.44* 0.27* 0.26* BMI (kg/m 2 ) 0.48* 0.50* 0.52* 0.30* 0.30* Systolic BP (mm Hg) 0.28* 0.29* 0.29* 0.27* 0.28* Diastolic BP (mm Hg) 0.22* 0.22* 0.20* 0.19* 0.17* *p 0.001; p 0.01. BP blood pressure.

HENSE ET AL. LVM AND BODY COMPOSITION 455 Table 4. Changes in Left Ventricular Mass (LVM) Associated With an Increase of 1 SD in Body Fat (adiposity) or Systolic Blood Pressure (BP) Results Are Expressed as Absolute Changes With 95% Confidence Intervals and as Relative Changes LVM Unindexed to Height to Height 2.7 to Body Surface Area to LBM Predictor variables (g) (g/m) (g/m 2.7 ) (g/m 2 ) (g/10 kg) Men (n 653) Body fat 24.6 12.2 13.1 11.4 4.5 10.3 5.8 5.7 1.7 5.1 (1 SD 6.8 kg) (20.2 28.8) (10.7 15.5) (3.5 5.5) (3.7 7.9) (1.0 2.4) Systolic BP 10.6 5.3 6.7 5.8 3.0 6.7 6.2 6.1 2.2 6.5 (1 SD 18.6 mm Hg) (6.4 14.8) (4.3 9.1) (2.0 4.0) (4.2 8.4) (1.5 2.9) Women (n 718) Body fat 21.1 14.8 12.7 14.3 5.3 13.5 5.4 6.5 2.2 6.7 (1 SD 9.0 kg) (18.2 24.0) (10.9 14.5) (4.4 6.2) (3.7 7.1) (1.6 2.8) Systolic BP 9.9 6.9 6.5 7.3 3.1 7.9 6.0 7.3 2.4 7.3 (1 SD 20.3 mm Hg) (6.7 13.1) (4.5 8.5) (2.2 4.0) (4.2 7.8) (1.7 3.1) *Results are derived from multivariate linear regression models containing age, body fat and systolic blood pressure. Absolute change expressed as percentage of the respective LVM mean. the impact of body fat to less than half (5.1% to 6.7%). Moreover, the multivariate analyses confirmed that the estimated impact of blood pressure on LVM were the same irrespective of the indexation applied, ranging from 5.3% to 6.7% in men and from 6.9% to 7.9% in women. It has to be noted that the relative affect of adiposity and blood pressure on LVM were of similar size after indexation to LBM and body surface area, whereas the estimated effect of adiposity exceeded that of blood pressure substantially with the other indexations. Exclusion of treated hypertensives. To account for potential confounding of the observed associations by use of antihypertensive medication, we excluded all pharmaceutically treated hypertensive patients (104 men and 129 women) from the analyses. The results did not change materially. Similar to what was demonstrated in Table 4 for all study subjects, the impact per 1 SD of body fat, which was as high as 9.7% of the unindexed LVM mean in men and 13.9% in women, was now reduced by the LVM/LBM ratio to 3.1% in men and to 6.5% in women. Of note, the impact of systolic blood pressure again remained unaffected by indexation and ranged from 6.1% to 8.0% in men and from 7.1% to 8.2% in women. Discussion This study shows that cardiac mass is similar in adult men and women once their differences in fat-free body mass are taken into account. Based on normalizations for LBM, previously reported associations of adiposity with cardiac mass appear markedly overestimated. In contrast, the strength of the relation of blood pressure with LVM is unaffected by different types of indexation. The results suggest that body size and composition are basic determinants of heart size in adults, that their effects are similar in men and women and that the mode of its analytic consideration, by means of different options for indexation, has a prominent influence on the observed strengths of associations. In fact, the simple LVM/ LBM ratio was effective in removing residual effects of body size and composition present in conventional indexations of LVM without impairing the relation to hypertension. Measurements of lean body mass. We employed BIA to estimate LBM values in this population sample. The BIA method is a rapid and easily applicable technique based on measurements of electrical resistance (31,32) and was validated in the past against a variety of more laborious techniques (32,33,37 39). Studies in children and adults have shown that BIA can be validly applied to assess body composition in epidemiologic studies if proper consideration is given to population-specific characteristics (40). We selected the BIA equation of Heitmann (34) because it had been derived in a population-based sample from a Danish MONICA community. The ranges of LBM found in this study were very comparable to our study, ranging in men from 46 to 83 kg and in women from 35 to 64 kg. The respective ranges in MONICA Augsburg, which also included the age range 25 to 34 years, were 42 to 78 kg and 32 to 58 kg. We further suggest that genetic background and dietary factors, each likely determinants of LBM, can be validly assumed not to be extremely diverse in these two populations. Second, the range of LBM and the results obtained with this equation were plausible and consistent: differences in body build of men and women were indicated by significantly higher male LBM levels and higher average body fat level as well as proportionate body fat mass (35.0% vs. 26.4% in males, p 0.01) in females. These relations are consistently found in population studies (41 43) and support our assumption that BIA produced valid assessments of body composition in our analyses. Body size, body composition and indexation of left ventricular mass. Studies that monitored normal growth of hearts from childhood into adulthood have revealed that cardiac size follows body growth and its metabolic demands (20,22,44,45). Thus, physiologically adequate increases of heart size in response to body requirements must be distinguished from morbid rises of LVM. Due to the lack of appropriate data on

456 HENSE ET AL. JACC Vol. 32, No. 2 LVM AND BODY COMPOSITION body composition, previous attempts to normalize LVM for body size were inadvertently restricted to indicators of body frame, that is, to body surface area, body height, or exponentials of body height (18,22 25,27). However, there is a considerable variability in LBM at a given body frame that arises as a result of individual differences in, for example, muscle mass. Moreover, disregarding body composition is particularly problematic in middle-aged and elderly individuals where estimations based on body frame are increasingly modified by intraabdominal adiposity and loss of muscle mass (41,43). We employed the simple ratio of LVM over LBM for indexation in this study because this ratio reflects the mathematical first-power relation between heart size and LBM (20,44,45). A residual positive correlation coefficient of 0.13 between LBM and the LVM/LBM ratio in women appears to indicate that this assumption of linearity is strictly fulfilled only in men. On the other hand, these residuals were so weak that exponentiation of LBM in women did not notably change results. Gender and adult heart size. In the general population, men have on average larger hearts than women (by 63 g in this study, Table 2). It has been shown that heart size differs only marginally between boys and girls in early childhood and that heart size begins to diverge only after puberty in response to differentials in body growth (26). The authors suggested that male heart size may represent physiologic hypertrophy in adaptation to body size. Nevertheless, in most studies, indexations of LVM to body weight, height, exponentials of height, or body surface area were unable to remove all of the gender differences for LVM, leaving open the question whether unidentified factors add to male heart sizes. Our results indicate that heart sizes of adult men and women are indeed very similar when an appropriate normalization for body size is applied. Moderate, but significant elevations of the LVM/LBM ratio in younger men (Figure 1), disappeared after adjustment for the higher blood pressure of males in this age range. Thus, use of the LVM/LBM ratio reveals that the hearts of men and women, despite their markedly different absolute weights and masses, have a similar size in relation to the metabolic demands of their bodies. This finding may render new perspectives on previously reported influences of gender on LVM (11,15,17,46,47) and, in particular, on the common use of gender specific partition values for LVH (24). Adiposity and left ventricular mass. The most striking inconsistencies were observed when the relative impact of adiposity, assessed in this study as body fat, on LVM was analyzed. Traditional indexations of LVM resulted in an overestimation of the influence of adiposity. This was particularly true for the indexation to height and height 2.7. Interestingly, the indexation to body surface area, which has been criticized for disregarding the affect of obesity (12,18,25), produced results for the affect of adiposity on LVM that were very close to those found after indexation to LBM. Evidence from observational studies (9,11) and clinical trials (48) seems to support a causal role of adiposity for the development of LVH. Nevertheless, its contribution relative to that of arterial hypertension is still unclear. We present data that allow a quantitative estimation of the impact that adiposity has on LVM in the general population. In contrast to other studies (11,15,47), adiposity seems to influence LVM in both men and women to a similar degree. Moreover, the magnitude of its impact is barely different from that of blood pressure if indexations to LBM or to body surface area are applied. Indexations of LVM for height or height 2.7 tend to inflate these estimates to about twice the size of the blood pressure effect. Blood pressure and left ventricular mass. Interestingly, indexations had no influence on the association of systolic or diastolic blood pressure with LVM in this study. This may indicate that the effects of blood pressure elevation on ventricular morphology are completely dissociated from those occurring as physiologic adaptations to body size or in response to adiposity. The LVM/LBM ratio seems to effectively distinguish the different component causes that have an impact on LVM. Study limitations. This study is cross sectional by design and, therefore, does not allow the evaluation of prognostic implications of the proposed indexation to LBM. However, it may be very appealing to investigate how the LVM/LBM ratio affects previously reported influential survival differences between men and women with conventionally defined LVH (6). Lack of data on physical activity in this study also prohibited investigation of the important question of how exercise affects LBM and its relation with LVM. Furthermore, the BIA equation used in this study cannot be assumed to be universally applicable in other populations. The usefulness of our results for clinical practice may also appear limited because, to date, measurements of LBM are only rarely available. However, commercial devices are being developed in increasing numbers and studies have been conducted to assess their validity (33). Therefore, the introduction of the BIA technique may offer an attractive option to improve the clinical interpretation of echocardiographic measurements of LVM. Conclusions. The results of this study indicate that the heart size of adults in the general population is mostly determined by body size and body composition. Indexation of LVM to LBM appropriately accounts for this. Use of the LVM/LBM ratio removes gender differences, reduces the overestimated impact of adiposity and leaves the effects of blood pressure on LVM unchanged. We suggest that the LVM/LBM ratio be considered for the generation of normative values of LVH and for the clinical evaluation of echocardiographic measurements of LVM. Most importantly, however, its superiority over other indexations has to be ultimately confirmed in prospective studies investigating the prognosis of abnormal LVM. References 1. Casale PN, Devereux RB, Milner M, et al. Value of echocardiographic left ventricular mass in predicting cardiovascular morbid events in hypertensive men. Ann Intern Med 1986;105:173 8. 2. Levy D, Garrison RJ, Savage DD, Kannel WB, Castelli WP. Left ventricular mass and incidence of coronary heart disease in an elderly cohort: The Framingham Study. Arch Intern Med 1989;110:101 7.

HENSE ET AL. LVM AND BODY COMPOSITION 457 3. Levy D, Garrison RJ, Savage DD, Kannel WB, Castelli WP. Prognostic implications of echocardiographically determined left ventricular mass in the Framingham Heart Study. N Engl J Med 1990;322:1561 6. 4. Koren MJ, Devereux RB, Casale PN, Savage DD, Laragh JH. Relation of left ventricular mass and geometry to morbidity and mortality in uncomplicated essential hypertension. Ann Intern Med 1991;114:345 52. 5. Krumholz HM, Larson M, Levy D. Prognosis of left ventricular geometric patterns in the Framingham Heart Study. J Am Coll Cardiol 1995;25:879 44. 6. Liao Y, Cooper RS, Mensah GA, McGee DL. Left ventricular hypertrophy has a greater impact on survival in women than in men. Circulation 1995;92:805 10. 7. Verdecchia P, Schillaci G, Borgioni C, et al. Adverse prognostic significance of concentric remodeling of the left ventricle in hypertensive patients with normal left ventricular mass. J Am Coll Cardiol 1995;25:871 78. 8. Frohlich ED, Apstein C, Chobanan AV, et al. The heart in hypertension. N Engl J Med 1992;327:998 1008. 9. Devereux RB, Alderman MH. Role of preclinical cardiovascular disease in the evolution from risk factor exposure to development of morbid events. Circulation 1993;88:1444 55. 10. Devereux RB, Roman MJ. Inter-relationships between hypertension, left ventricular hypertrophy and coronary heart disease. J Hypertension 1993; 11:S3 S9. 11. de Simone G, Devereux RB, Roman MJ, Alderman MH, Laragh JH. Relation of obesity and gender to left ventricular hypertrophy in normotensive and hypertensive adults. Hypertension 1994;23:600 6. 12. Lauer MS, Anderson KM, Kannel WB, Levy D. The impact of obesity on left ventricular mass and geometry. The Framingham Heart Study. JAMA 1991;266:231 6. 13. Savage DD, Garrison RJ, Kannel WB, et al. The spectrum of left ventricular hypertrophy in a general population sample: the Framingham Study. Circulation 1987;75:1 26. 14. Savage DD, Levy D, Dannenberg AL, Garrison RJ, Castelli WP. Association of echocardiographic left ventricular mass with body size, blood pressure and physical activity (The Framingham Study). Am J Cardiol 1990;65:371 6. 15. Kuch B, Muscholl M, Luchner A, et al. Geschlechtsunterschiede in der Beziehung zwischen Übergewicht und Hypertonie mit linksventrikulärer Masse und Hypertrophie. Zeitschr Kardiol 1996;85:334 42. 16. Hense HW, Koivisto AM, Kuulasmaa K, et al. Assessment of blood pressure measurement quality in the baseline surveys of the WHO MONICA Project. J Hum Hypertens 1995;9:935 46. 17. Hammond IW, Devereux RB, Alderman MH, Laragh JH. Relation of blood pressure and body build to left ventricular mass in normotensive and hypertensive employed adults. J Am Coll Cardiol 1988;12:996 1004. 18. Levy D, Anderson KM, Savage DD, Kannel WB, Christiansen JC, Castelli WP. Echocardiographically detected left ventricular hypertrophy: prevalence and risk factors. Ann Intern Med 1988;108:7 13. 19. Schmidt-Nielsen K. Scaling Why Is Animal Size So Important? New York: Cambridge University Press, 1984. 20. de Simone G, Daniels SR, Devereux RB, et al. Left ventricular mass and body size in normotensive children and adults: assessment of allometric relations and impact of overweight. J Am Coll Cardiol 1992;20:1251 60. 21. Daniels SR, Kimball TR, Morrison JA, Khoury P, Witt S, Meyer RA. Effect of lean body mass, fat mass, blood pressure, and sexual maturation on left ventricular mass in children and adolescents. Circulation 1995;92:3249 54. 22. de Simone G, Devereux RB, Daniels SR, Koren MJ, Meyer RA, Laragh JH. Effect of growth on variability of left ventricular mass: assessment of allometric signals in adults and children and their capacity to predict cardiovascular risk. J Am Coll Cardiol 1995;25:1056 62. 23. Devereux RB, Reichek N. Echocardiographic determination of left ventricular mass in man. Anatomic validation of the method. Circulation 1977;55: 613 8. 24. Abergel E, Tase M, Bohlender J, Menard J, Chatellier G. Which definition for echocardiographic left ventricular hypertrophy? Am J Cardiol 1995;75: 498 502. 25. Levy D, Savage DD, Garrison RJ, Anderson KM, Kannel WB, Castelli WP. Echocardiographic criteria for left ventricular hypertrophy: The Framingham Heart Study. Am J Cardiol 1987;59:956 60. 26. de Simone G, Devereux RB, Daniels SR, Meyer RA. Gender differences in left ventricular growth. Hypertension 1995;26:979 83. 27. Lauer MS, Anderson KM, Larson MG, Levy D. A new method for indexing left ventricular mass for differences in body size. Am J Cardiol 1994;74:487 91. 28. Goble MM, Mosteller M, Moskowitz WB, Schieken RM. Sex differences in the determinants of left ventricular mass in childhood. The Medical College of Virginia Twin Study. Circulation 1992;85:1661 5. 29. Keil U, Stieber J, Döring A, et al. The cardiovacular risk factor profile in the study area Augsburg: results from the first MONICA survey 1984/85. Acta Med Scand 1988;728:119 28. 30. Hense HW, Döring A, Stieber J, Keil U. The association of antihypertensive treatment patterns and adverse lipid effects in population-based studies. J Clin Epidemiol 1992;45:1423 30. 31. Kussmaul B, Döring A, Filipiak B. Bioelektrische Impedanzanalyse (BIA) in einer epidemiologischen Studie. Ernährungs-Umschau 1996;43:46 8. 32. Kushner R. Bioelectrical impedance analysis: a review of principles and applications. J Am Coll Nutr 1992;11:199 209. 33. Segal KR, Van Loan M, Fitzgerald PI, Hodgdon JA, Van Itallie TB. Lean body mass estimation by bioelectrical impedance analysis: a four-site crossvalidation study. Am J Clin Nutr 1988;47:7 14. 34. Heitmann B. Prediction of body water and fat in adult Danes from measurement of electrical impedance. A validation study. Int J Obesity 1990;14:789 802. 35. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1:307 10. 36. Dubois D, Dubois EF. A formula to estimate the approximate surface area if height and weight be known. Arch Intern Med 1961;17:863 71. 37. Fuller NJ. Comparison of abilities of various interpretations of bio-electrical impedance to predict reference method body composition assessment. Clin Nutr 1993;12:236 42. 38. Franssila Kalunki A. Comparison of near-infrared light spectroscopy, bioelectrical impedance and tritiated water techniques for the measurement of fat-free mass in humans. Scand J Clin Lab Invest 1992;52:879 85. 39. Gray DS, Bray GA, Gemayel N, Kaplan N. Effect of obesity on bioelectrical impedance. Am J Clin Nutr 1989;50:255 60. 40. Deurenberg P, Kusters C, Smit H. Is the bioelectrical impedance method suitable for epidemiological field studies? Eur J Clin Nutr 1989;43:647 54. 41. Heitmann B. Body fat in adult Danish population aged 35 65 years: an epidemiological study. Int J Obesity 1991;15:535 45. 42. Durnin JV, Womersley J. Body fat assessed from total body density and its estimation from skinfold thickness: measurements on 481 men and women aged from 16 to 72 years. Brit J Nutr 1974;77 97. 43. Roubenoff R, Dallal GE, Wilson PW. Predicting body fatness: the body mass index vs. estimation of bioelectrical impedance. Am J Publ Health 1995; 726 8. 44. Gutgesell HP, Rembold CM. Growth of the human heart relative to body surface area. Am J Cardiol 1990;65:662 8. 45. MacMahon T. Size and shape in biology. Science 1973;179:1201 4. 46. Gardin JM, Siscovick D, Anton Culver H, et al. Sex, age, and disease affect echocardiographic left ventricular mass and systolic function in the freeliving elderly. The Cardiovascular Health Study. Circulation 1995;91:1739 48. 47. Marcus R, Krause L, Weder AB, Dominguez-Mejia A, Schork NJ, Julius S. Sex-specific determinants of increased left ventricular mass in the Tecumseh Blood Pressure Study. Circulation 1994;90:928 36. 48. MacMahon S, Wilcken D, MacDonald G. The effect of weight reduction on left ventricular mass: a randomized controlled trial in young, overweight hypertensive patients. N Engl J Med 1986;334 9.