Increase in left ventricular mass (LVM) and development

Transcription

1 AJH 2005; 18: Indexation Criteria of Ventricular Mass and Predictive Role of Blood Pressure and Body Composition Liberato Aldo Ferrara, Olga Vaccaro, Ondina Cardoni, Martino Laurenzi, Mario Mancini, and Alberto Zanchetti Background: There is no definite consensus on which indexation of left ventricular mass (LVM) should be used to better identify left ventricular hypertrophy (LVH). Left ventricular mass has been adjusted to height, to height 2.7 (h 2.7 ) and to body surface area (BSA). The aims of the present study were to evaluate the prevalence of LVH according to different indexations and different cutoffs and to identify the most useful indexation of LVM to detect hypertension-related LVH. Methods: Echocardiographic LVH was defined as LVM to h 2.7, LVM to BSA, LVM to height, LVM values in the upper 5th percentile of our gender-related LVM distribution, using different partition values suggested in previous population-based studies. Results: Prevalence of LVH in the general population was 32% using the less restrictive criterion (LVH 49.2/46.7 g/m 2.7 ), 15% with the criterion of LVH 116/104 g/m 2, and Increase in left ventricular mass (LVM) and development of left ventricular hypertrophy (LVH), as measured echocardiographically, are strong predictors of cardiovascular morbidity and mortality. 1 4 There is no doubt that heart mass parallels body size to satisfy the increased metabolic requests of heavier bodies 5 ; therefore, obesity appears to be a very strong predictor of cardiac mass, even more than increased afterload. 6 This is the why several investigators have suggested adjusting LVM to body size, using body surface area (BSA) 7,8 or height 9 as an indexation criterion. Further attempts have been made to separate the influence of different body weight components (fatty body mass [FM] versus fat-free body mass [FFM]) 5,10,11 and different distribution of adiposity (measured as waist-to-hip ratio [w/h]). Different indexations of LVM, along with different cut-offs 3.8% with the most restrictive one (LVH 134/110 g/m 2 ). Prevalence of LVH in hypertensive subjects was almost twice than in normotensive subjects with all criteria. Only 20 subjects out of the 707 evaluated were found to have LVH with all six criteria. In multiple regression analysis SBP was independently associated with nonindexed LVM and was indexed to both BSA and h 2.7. On the other hand, fat-free mass was a poweful predictor of nonindexed LVM or of LVM to BSA, whereas body mass index was the strongest predictor of LVM to h 2.7. Conclusions: The indexation of LVM to BSA, possibly with the cut-off of LVH 116/104, is probably the best criterion for identifying blood pressure related LVH. Am J Hypertens 2005;18: American Journal of Hypertension, Ltd. Key Words: Left ventricular mass, left ventricular hypertrophy, blood pressure, body composition. for normality, strongly influence the prevalence of LVH in the population. 9,12 15 The present analyses, performed in a population sample of the Gubbio Study, had the following aims: 1) to evaluate the prevalence of LVH according to different indexations of the cardiac mass as well as to different cut-offs for the same indexation; and 2) to detect the most reliable and useful indexation of LVM to identify the independent weight of several suggested determinants of cardiac mass. Patients and Methods Study Population The Gubbio Population Study is a clinical survey performed in the small town of Gubbio, in central Italy, not Received January 17, First decision April 29, Accepted May 2, From the Department of Clinical and Experimental Medicine (LAF, OV, MM), Federico 2nd University of Naples, Naples, Italy; Ospedale di Gubbio, Rome, Italy; Merck Sharp and Dohme (ML); Centro di Fisiologia Clinica ed Ipertensione (AZ), Università di Milano, Ospedale Maggiore and Istituto Auxologico Italiano, Milan, Italy. This study was supported in part by a grant from the National Institute of Health and by Merck Sharp and Dohme. Deceased. Address correspondence and reprint requests to Dr. Liberato Aldo Ferrara, Department of Clinical and Experimental Medicine, Federico 2nd University, Via S. Pansini 5, Naples, Italy; ferrara@ unina.it /05/$ by the American Journal of Hypertension, Ltd. doi: /j.amjhyper Published by Elsevier Inc.

2 AJH October 2005 VOL. 18, NO. 10 INDEXATION OF LVM AND PREVALENCE OF LVH 1283 too far from Perugia. The study was initally conducted from 1983 to 1985 when 5376 male and female individuals, 5 to 90 years of age, were initially examined. The entire population living within the area bounded by the medieval walls of the town was invited to participate in the study. Of the 3836 who had been invited, 3536 accepted the invitation and the other 1840, who resided outside the walls, were evaluated at the request of the health administration. Age and sex distribution in Gubbio was similar to that for Italy overall, as well as the distribution of BMI and the prevalence of hypertension, diabetes, and hyperlipidemia. 16 Subsequently, these individuals underwent a follow-up examination during 1989 to ,18 In both surveys the comprehensive examination included measurement of body weight, height, waist-to-hip ratio, subscapular and triceps skinfold thicknesses, blood pressure (BP) and heart rate (HR), 12-lead electrocardiography, standardized blood and urine laboratory tests, individual and family medical histories, and questionnaires on lifestyle and awareness of hypertension and other cardiovascular risk factors. 16 Both FM and FFM were calculated from the logarithm of the sum of the thickness of two skinfolds (subscapular and triceps), according to published equations. 19,20 Blood pressure was measured by doctors trained by the Istituto Superiore di Sanità, Rome. Measurements were taken after the participant had been seated quietly for 5 min; pressure was measured on the right arm, with the use of an appropriate cuff size. Three recordings were made, 1 min apart. The first and the fifth phases of Korotkoff sounds were defined as systolic and diastolic BP, respectively. Subjects who had BP 140/90 mm Hg or who were taking antihypertensive therapy were classified as hypertensive subjects. After the second survey, all persons in the age range 40 to 60 years (n 851) were invited for an echocardiographic examination. The protocol of the study had been approved by the local Ethics Committee, and informed consent had been obtained by all participants in the study. Laboratory Analyses Serum cholesterol, HDL-cholesterol, total triglycerides, and blood glucose were measured according to methods fully described elsewhere. 16 Echocardiography All echocardiograpic procedures were performed during 1990 to 1992 by the same echocardiographer (O.C.). Echocardiograms were obtained with participants lying in the left lateral decubitus position with the head 30 degrees from the horizontal plane. A Vingmed (Oslo, Denmark) 700 echocardiographic machine with a 3.5-MHz transducer was used. Pictures of M-mode left ventriclular images were made by a fiber-optic recorder on light-sensitive paper at 50 mm/sec. All examinations were recorded on VHS videotape. Measurements of interventricular septal thickness (IVST), left ventricular posterior wall thickness (PWT), and left ventricular diastolic internal diameter (LVIDd) were taken at the mitral valve chordal level. Measurements were made from three consecutive cardiac cycles on M-mode pictures, and averaged. The LVIDd and end-systolic (LVIDs) left ventricular internal diameters, PWT, and IVST were measured according to the recommendations of the American Society of Echocardiography. 21 A second set of measurements was also taken according to the Penn convention criteria to calculate left ventricular mass (LVM). 7 Left ventricular mass was indexed to BSA, height, and height raised to the allometric power of 2.7. These indexations have well-defined prognostic impacts on cardiovascular disease, which were found to be roughly equivalent. 22 Left ventricular hypertrophy (LVH) was defined according to the following different cut-offs: 1) the LVM to height to the power of 2.7 using a partition value of 49.2 g/m 2.7 for men and 46.7 g/m 2.7 for women 12 (cut-off 1) or 51 g/m 2.7 for both sexes 13 (cut-off 2); 2) the LVM to BSA, using the partition values of 116 g/m 2 for men and 104 g/m 2 for women 15 (cut-off 3) or values of 134 g/m 2 for men and 110 g/m 2 for women 14 (cutoff 4); 3) the LVM to height, using the partition values of 143 g/m for men and 102 g/m for women 9 (cut-off 5); and 4) the LVM values in the upper 5th percentile of our gender-related LVM distribution ( g for women and g for men) (cutoff 6). In addition we indexed LVM to FFM, which is an indexation not yet prognostically validated, and we calculated the prevalence of LVH according to the partition value of LVM/FFM 4.1, which is the cut-off value in two previous studies. 10,11 All echocardiograms were read by the same operator (O.C.) and in a 20% sample of tracings that were randomly chosen, intrareader variability of the measurement was evaluated. The two measurements for each variable were closely related to each other (P.001), as the percent error was 3.76 for IVST, 4.28 for PWT, and 1.06 for LVM. 23 Statistical Analyses Data are expressed as mean standard deviations. The data were analyzed by using the SPSS statistical package (SPSS Italia, Bologna, Italy). Comparisons of means were performed by unpaired t test and by one-way analysis of variance with Tukey s multiple comparisons. The strength of correlation between different variables was evaluated by Pearson s linear correlation; multiple regression analysis with stepwise forward procedure was used to explore the independent influence of some variables on LVM. The intraclass correlation coefficient was calculated to evaluate the relationship between the different criteria used for the diagnosis of LVH.

3 1284 INDEXATION OF LVM AND PREVALENCE OF LVH AJH October 2005 VOL. 18, NO. 10 Table 1. Prevalence of left ventricular hypertrophy (LVH) according to different indexation and cut-offs LVH cut-off n % n % n % 1 LVM/h (M)/46.7(F) LVM/h LVM/m (M)/104 (F) LVM/m (M)/110 (F) LVM/m 143 (M)/102 (F) LVM (M)/194.2 (F) F female; LVM left ventricular mass; M male. Results All subjects (n 707) Of the 851 invited subjects, 707 had technically adequate echocardiograms for the analysis of LVM. Of these individuals, 457 were normotensive and 250 hypertensive. Of the hypertensive subjects, 55 were not taking any antihypertensive drug, whereas 195 were on antihypertensive treatment, either monotherapy or combination therapy. The prevalence of obesity was 23.8% in men and 26.2% in women.the prevalence of diabetes was 5.2% and 3.0% in men and women, respectively. As shown in Table 1, prevalence of LVH widely changed according to the indexation of LVM and the cut-off criteria used. In particular, in both normotensive and hypertensive subjects, the prevalence was highest when LVM was indexed to height 2.7 and the lower cut-off was used, intermediate with the height 2.7 indexation and the higher cut-off or the BSA indexation with the lower cut-off or the height indexation. Indexation to BSA with the higher cut-off gave the same low prevalence as use of the upper 5th percentile. Only 20 subjects (seven normotensive and 13 hypertensive) had LVH according to all six criteria. By all criteria LVH prevalence was two to three times higher in hypertensive subjects, without a significant difference between treated and untreated subjects. According to the LVM/FFM cut-off of 4.1, 4% of the sample had LVH. The prevalence of LVH in obese subjects was significantly higher than in nonobese subjects when LVM was indexed to h 2.7 (58.1% v 22.7%, P.001) but was roughly similar when LVM was indexed to BSA (18.4% v 13.8%, P NS). Influence of gender on the prevalence of LVH was significant only when it was determined according to the indexation of LVM to BSA (17.8% and 7.1% in men v 13.2% and 1.9% in women, P.01, with the less restrictive and more restrictive criterion, respectively). A significant relationship was found between LVH detected with the different methods (intraclass correlation coefficient 0.510, P.001). Differences in age, BMI, BP, blood glucose, waist-tohip ratio, subscapular skinfold, FM, and FFM between subjects with and without LVH are shown in Table 2. When LVM was indexed to height 2.7 the differences in all Normotensive subjects (n 457) Hypertensive subjects (n 250) variables were statistically significant. After indexation to BSA, on the other hand, differences in BP and blood glucose maintained statistical significance whereas differences in BMI, waist-to-hip ratio, and FM did not always reach statistical significance. Linear correlation analysis showed that BMI, waist-tohip ratio, FFM, fasting blood glucose, SBP, and DBP were strongly related to absolute LVM and LVM indexed to height or to BSA in the entire population and in the subgroups of normotensive and hypertensive subjects. Fatty mass, on the other hand, was significantly correlated only with absolute LVM and LVM indexed to height 2.7 but not with LVM indexed to BSA. Multiple regression analysis investigating the independent relationship between the variables with highest correlation coefficients and LVM was thereafter performed. Variables included in the regression equations were SBP (or DBP), fasting blood glucose (FBG), BMI (or body weight [BW] or waist-to-hip ratio) and FFM (or FM). The best regression equation included BMI, fasting blood glucose, SBP, and FFM. These variables were strongly and independently associated with nonindexed LVM (R ) in the entire population and in both normotensive and hypertensive subjects. The SBP and FFM were significantly included in the equation in all cases. (For SBP, t 5.08, P.001 in all subjects; t 5.40, P.001 in normotensive subjects; and t 2.08, P.05 in hypertensive subjects. For FFM, t in all subjects, t in normotensive subjects, and t 9.98 in hypertensive subjects, P.001). The BMI was significantly included only in the equation for all subjects (t 1.92, P.05). When LVM was indexed to height 2.7, BMI and SBP were strongly associated in all subjects (R ), in normotensive subjects (R ), and in hypertensive subjects (R ). For BMI, t in all subjects, 7.52 in normotensive subjects, and 7.05 in hypertensive subjects, P.001. For SBP, t 5.71, P.001 in all subjects, 2.27 in normotensive subjects, and 2.69 in hypertensive subjects, P.01). When LVM was indexed to BSA, FFM, rather than BMI, entered the best regression

4 Table 2. Selected characteristics in the Gubbio population according to six different cut-offs for left ventricular hypertrophy LVH Age (y) BMI (ww/h 2 ) SBP (mm Hg) DBP (mm Hg) FBG (mg/dl) W/H SS LVH cut-off 1 Yes No * * 2 Yes No * 3 Yes No * 4 Yes No * 5 Yes No Yes No * * BMI body mass index; DBP diastolic blood pressure; FBG fasting blood glucose; FFM fat-free mass; FM fatty mass; SBP systolic blood pressure; SS subscapular skinfold; W/H waist-to-hip ratio. * P.01; P 0.001; P.05. FM (kg) FFM (kg) AJH October 2005 VOL. 18, NO. 10 INDEXATION OF LVM AND PREVALENCE OF LVH 1285

5 1286 INDEXATION OF LVM AND PREVALENCE OF LVH AJH October 2005 VOL. 18, NO. 10 equation (R in all subjects, 0.19 in normotensive subjects, and 0.21 in hypertensive subjects. Both FFM (t 9.31 in all subjects, 8.22 in normotensive subjects, and 4.55 in hypertensive subjects, P.001) and SBP (t 5.50 P.001 in all subjects, 2.44 in normotensive subjects, and 2.55 in hypertensive subjects, P.01) were the independent variables most closely related to LVM/BSA. When waist-to-hip ratio replaced BMI in the equation, the predictive power of the equations remained similar or was only slightly reduced. Discussion In the last 25 years ultrasound techniques for measurement of LVM have greatly contributed to the epidemiology and pathophysiology of LVH in hypertension, to the extent that the detection of LVH is recommended for stratification of total cardiovascular risk in hypertensive patients and for assessing prognosis, 24 and its therapeutic regression is considered as a surrogate of morbid event reduction. 25 However echocardiographically determined LVM is obviously dependent on body size; and various indices of normalization have been used, each of which is likely to correct to a different degree for different components of body size. Therefore the prevalence of LVH is going to be influenced not only by the cohort of investigated subjects (general population, hypertensive patients, obese subjects, and so forth) and by the more or less stringent arbitrary cut-off used to define LVH, but also by the indexation criteria used. Comparisons of different indexation criteria have previously been performed in different populations, indicating that the relative risk of morbidity and mortality is similar with the different indexations. 10,22,26 Only one of these studies, which was performed in communities with a very high prevalence of obesity (nearly three times that in the present investigation) and diabetes ( 15 times that in this study), found that the indexation to the power of 2.7 was more suitable to detect prognostically adverse LVH in that particular population. The present study, performed in a sample of individuals 40 to 60 years of age from the Gubbio Study, is, on the other hand, representative of a middle-aged Italian population and provided a considerable cohort (n 707) of individuals within a relatively narrow age range who underwent a careful echocardiographic examination. This population was suitable for analyzing how much different indexation criteria, as well as different cut-offs, do indeed influence the prevalence of LVH, and how well different components of body size are taken care of by various indexation criteria. In the Gubbio population, aged 40 to 60 years, the prevalence of LVH has been found to range from 31.6%, when the least selective cut-off (LVH 49.2/46.7) was used, to 3.8% with the most selective one (LVH 134/110). Among normotensive subjects the range was from 25.2% to 2.4% and in hypertensive subjects 43.6% to 6.4%. The two most stringent cut-offs (LVH 134/110 and LVH 276.6/194.2) yielded very similar prevalences in the overall population, as well as separately in normotensive and hypertensive subjects, largely identifying the same subjects. This was not the case when these criteria were correlated with the most generous ones (LVH 49.2/46.7 and LVH 51) using height to the power 2.7 as indexation. On the whole, only a small number of subjects (n 20) had LVH according to all six cut-offs. When considering the characteristics of the populations with and without LVH, indexation to height to the power 2.7 appeared to differentiate the two populations strongly, as most of the subjects characteristics markedly and significantly differed between the two groups. On the other hand, indexation to BSA, largely adjusted for body size, with small and often nonsignificant differences in BMI and fat body mass: the main differences, therefore, were confined to blood pressure, fasting blood glucose, and FFM. Accordingly, multiple regression analyses indicate that BMI was independently associated with LVM particularly when it was indexed to height 2.7, whereas the association of the FFM became particularly evident when LVM was indexed to BSA. Blood pressure, and in particular systolic pressure, was largely confirmed to be the independent variable most strongly associated with LVH in all cases. In conclusion, the results of the present study suggest the following. First, the variability of LVM is largely related to systolic BP in both the normotensive and hypertensive components of a general population. However, BMI, or waist-to-hip ratio, also plays an important role, particularly when LVM is indexed to height 2.7. The importance of FFM becomes particularly evident when indexation to BSA is used. Second, prevalence of LVH largely depends on the type of indexation used and on the cut-off chosen. Indexation to height to the 2.7 power, whatever the cut-off used, gives the highest prevalence of LVH and may add obesity-related LVH to that related to BP. On the other hand, indexation to BSA appears to correct for obesity-related increases in LVM and to identify LVH associated with hypertension to a better extent. Therefore the two types of indexation may have different application in epidemiologic studies in which the role of body weight or that of BP is the main object of investigation. Third, the risk of cardiovascular disease with increasing LVM is likely to be a continuous one; and the use of any cut-offs, although clinically practical, is obviously arbitrary. It is apparent from our analyses, however, that the two most widely used cut-offs of LVM indexed to height 2.7 provide large prevalences; on the other hand the stringent cut-off of LVM indexed to BSA ( 134 in men and 110 in women) is too strict and identifies only the upper 5th percentile of LVM values. For practical reasons, until definite data are available on which identification criterion of LVH gives the best prediction of cardiovascular risk in the general population, the indexation to BSA appears to us to be most appropriate to detect the prognostically adverse hypertension-related LVH. Among the different cut-offs, that of 116/104 g/m 2 does not limit its use to the heaviest ventricular masses for the diagnosis of

6 AJH October 2005 VOL. 18, NO. 10 INDEXATION OF LVM AND PREVALENCE OF LVH 1287 cardiac hypertrophy, as does the criterion LVH 134/110, and therefore seems to give a reliable prevalence of LVH in the general population. According do this reasoning, the recent European Society of Hypertension/European Society of Cardiology (ESH/ESC) guidelines 24 suggest using the same indexation of LVM to BSA with an intermediate cut-off (ie, 125/ 110 g/m 2 ) to evaluate LVH in hypertensive subjects. References 1. 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: Cooper RS, Simmons BE, Castaner A, Santhanam V, Ghali JK, Mar M: Left ventricular hypertrophy is associated with worse survival independent of ventricular function and number of coronary arteries severely narrowed. Am J Cardiol 1990;65: 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: 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: 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: Levy D, Anderson KM, Savage DD, Kannel WB, Christiansen JC, Castelli WP: Echocardiographically detected left ventricular hypertrophy: prevalence and risk factors. The Framingham Heart Study. Ann Intern Med 1988;108: Devereux RB, Reichek N: Echocardiographic determination of left ventricular mass in man. Anatomic validation of the method. Circulation 1977;55: Abergel E, Tase M, Bohlender J, Menard J, Chatellier G: Which definition for echocardiographic left ventricular hypertrophy? Am J Cardiol 1995;75: 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: Bella JN, Devereux RB, Roman MJ, O Grady MJ, Welty TK, Lee ET, Fabsitz RR, Howard BV: Relations of left ventricular mass to fat-free and adipose body mass: the strong heart study. The Strong Heart Study Investigators. Circulation 1998;98: Kuch B, Hense HW, Gneiting B, Doring A, Muscholl M, Brockel U, Schunkert H: Body composition and prevalence of left ventricular hypertrophy. Circulation 2000;102: de Simone G, Daniels SR, Devereux RB, Meyer RA, Roman MJ, de Divitiis O, Alderman MH. 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: 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: Devereux RB, Lutas EM, Casale PN, Kligfield P, Eisenberg RR, Hammond IW, Miller DH, Reis G, Alderman MH, Laragh JH: Standardization of M-mode echocardiographic left ventricular anatomic measurements. J Am Coll Cardiol 1984;4: Devereux RB, Dahlof B, Levy D, Pfeffer MA: Comparison of enalapril versus nifedipine to decrease left ventricular hypertrophy in systemic hypertension (the PRESERVE trial). Am J Cardiol 1996;78: Laurezi M, Cirillo M, Angeletti, Buongiorno A, Morisi G, Panarelli W, Panfili M, Stamler J, Terradura O, Trevisan M: Gubbio Population Study: baseline findings. Nutr Met Cardiovasc Dis 1991;1: S1 S Laurenzi M, Trevisan M: Sodium-lithium countertransport and blood pressure: the Gubbio Population Study. Hypertension 1989; 13: Menotti A, Lanti M, Zanchetti A, Puddu PE, Cirillo M, Mancini M, Vignarelli OT: Impact of the Gubbio Population Study on community control of blood pressure and hypertension. Gubbio Study Research Group. J Hypertens 2001;19: 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. Br J Nutr 1974;32: Siri WE: The gross composition of the body. Adv Biol Med Phys 1956;4: Sahn DJ, De Maria A, Kisslo J, Weyman A: The Committee on M-mode Standardization of the American Society of Echocardiography: recommendations regarding quantitation in M-mode echocardiography: results of a survey of echocardiographic measurements. Circulation 1978;58: Liao Y, Cooper RS, Durazo-Arvizu R, Mensah GA, Ghali JK: Prediction of mortality risk by different methods of indexation for left ventricular mass. J Am Coll Cardiol 1997;29: Ferrara LA, Vaccaro O, Cardoni O, Panarelli W, Laurenzi M, Zanchetti A: Is there a relationship between left ventricular mass and plasma glucose and lipids independent of body mass index? Results from the Gubbio Study.Nutr Met Cardiovac Dis 2003;13: Cifkova R, Erdine S, Fagard R, Farsang C, Heagerty AM, Kiowski W, Kjeldsen S, Luscher T, Mallion JM, Mancia G, Poulter N, Rahn KH, Rodicio JL, Ruilope LM, van Zwieten P, Waeber B, Williams B, Zanchetti A, ESH/ESC Hypertension Guidelines Committee: Practice guidelines for primary care physicians: 2003 ESH/ESC hypertension guidelines. J Hypertens 2003;21: Devereux RB, Agabiti-Rosei E, Dahlof B, Gosse P, Hahn RT, Okin PM, Roman MJ: Regression of left ventricular hypertrophy as a surrogate end-point for morbid events in hypertension treatment trials. J Hypertens 1996;14(Suppl):S95 S de Simone G, Kizer JR, Chinali M, Roman MJ, Bella JN, Best LG, Lee ET, Devereux RB, Strong Heart Study Investigators: Normalization for body size and population-attributable risk of left ventricular hypertrophy: the Strong Heart Study. Am J Hypertens 2005; 18: