& 2005 Nature Publishing Group All rights reserved /05 $30.00

Transcription

1 (2005) 19, & 2005 Nature Publishing Group All rights reserved /05 $ ORIGINAL ARTICLE Left ventricular filling abnormalities and obesity-associated hypertension: relationship with overproduction of circulating transforming growth factor b1 G Parrinello 1, A Licata 1, D Colomba 1, T Di Chiara 1, C Argano 1, P Bologna 1, S Corrao 2, G Avellone 1, R Scaglione 1 and G Licata 1 1 Department of Internal Medicine, University of Palermo, Palermo, Italy; 2 Unit of Clinical Methodology, Epidemiological and Statistics, National Relevance Hospital Trust, Civico e Benfratelli, Palermo, Italy This study has been designed to evaluate the relationship among transforming growth factor b1 (TGFb1) and some measurements of diastolic function in a population of hypertensive subjects with normal left ventricular ejection fraction. We studied 67 hypertensive outpatients who according to their BMI levels were subdivided into three groups: lean (L), overweight (OW) and obese (OB) hypertensives (HT). Circulating TGFb1 and M- and B-mode echocardiography was determined. All hypertensives were further subgrouped, according to European Society of Cardiology Guidelines, into two subsets of patients with normal diastolic function or with diastolic dysfunction. Prevalence of left ventricular hypertrophy (LVH) was determined in all the groups. TGFb1, left ventricular mass (LVM), LVM/h 2.7, E-wave deceleration time and isovolumic relaxation time (IVRT) were significantly (Po0.005) higher and E/A velocity ratio was significantly (Po0.05) lower in OW-HT and OB- HT than in L-HT. Prevalence of LVH was significantly higher (Po0.03) in group OB-HT than in L-HT. TGFb1 (Po0.004), LVM/h 2.7 (Po0.001) and prevalence of LVH were (Po0.01) significantly higher in hypertensives with diastolic dysfunction than hypertensives with normal diastolic function. TGFb1 levels were positively correlated with BMI (r ¼ 0.60; Po0.0001), LVM/h 2.7 (r ¼ 0.28; Po0.03), IVRT (r ¼ 0.30; Po0.02) and negatively with E/A ratio (r ¼ 0.38; Po0.002) in all HT. Multiple regression analysis indicated that TGFb1, BMI and IVRT were independently related to E/A ratio explaining 71% of its variability (r ¼ 0.84; Po0.0001). This relationship was independent of LVH, age and HR suggesting that TGFb1 overproduction may be considered a pathophysiological mechanism in the development of left ventricular filling abnormalities in obesity-associated hypertension. (2005) 19, doi: /sj.jhh Published online 21 April 2005 Keywords: obesity-associated hypertension; TGFb1; left ventricular hypertrophy; left ventricular diastolic function Introduction Epidemiological and pathophysiological data indicate that obesity has to be actually considered a major risk factor for cardiovascular disease or events. 1,2 Obesity and hypertension are often associated with each other and each one, per se, can induce left ventricular hypertrophy (LVH) and left ventricular diastolic dysfunction. 3,4 In view of this, alterations in systemic haemodynamics and in left ventricular geometry and function have been reported both in normotensive and in hypertensive obese subjects. They include an Correspondence: Professor G Parrinello, Department of Internal Medicine, Istituto di Clinica Medica, Piazza delle Cliniche 2, University od Palermo, Palermo 90127, Italy. gaspare.parrinello@unipa.it Received 6 November 2004; revised 21 February 2005; accepted 22 February 2005; published online 21 April 2005 increased cardiac preload and ventricular mass that contribute to the occurrence of early preclinical left ventricular dysfunction. 3 9 Impairment of left ventricular relaxation, which seems often to be independent of cardiac load and geometry, has been reported both in obese and hypertensive subjects. 3,9,10 Excessive myocardial fibrosis has been implicated in progression of cardiac dysfunction, especially diastolic dysfunction, in hypertensive hearts. 11 Disproportional accumulation of fibrous tissue is able to impair stiffness and pumping capacity accounting for a spectrum of ventricular dysfunction that first appears during diastole and subsequently involves systole. 12 In this context a main role might be attributed to transforming growth factor b1 (TGFb1), a multifunctional cytokine that regulates cell growth, differentiation, matrix production, blocks matrix degradation and induces fibrosis in many tissues,

2 544 including heart, blood vessels, kidney and lung Overproduction of TGFb1 is involved in obesityassociated hypertension and in the long-term sequelae of hypertension, including LVH, vascular remodelling and progressive renal disease Moreover, some recent data suggest that TGFb1 might play a casual role in myocardial fibrosis and diastolic dysfunction both in pressure overload hearts 19 and after heart transplantation. 20 In addition, the blocking of TGFb1 with monoclonal antibodies has been reported to be effective to neutralize myocardial fibrosis. 19 Although these results emphasize a link between TGFb1 overproduction and abnormalities in left ventricular structure and function, so far no data are available on the relationship between TGFb1 and diastolic function in hypertensive subjects. Accordingly, the present study has been designed to evaluate the relationship between TGFb1 and some measurements of diastolic function in a population of hypertensive subjects. For these reasons, circulating TGFb1, ratio between peak early transmitral flow velocity and peak late transmitral flow velocity (E/A ratio), the deceleration time of early transmitral flow velocity (DTE) and isovolumic relaxation time (IVRT) were analysed in lean, overweight and obese hypertensive subjects. The principal goal of this study was to evaluate whether higher levels of circulating TGFb1 were associated with abnormalities in diastolic function in hypertensive obese subjects. To minimize the effects of left ventricular systolic function, only hypertensive patients with normal ejection fraction (X50%) were included in our investigation. Methods Patients We performed a multiple regression sample size calculation based on a beta value of 0.10 (90% power), an alpha value of 0.05, a prior squared regression coefficient of 0.25 and a maximum of six tested variables. 21 An appropriate algorithm was used to obtain exact answers. The sample size consisted of 59 patients and this number was assumed as the minimum one for this study. In all, 77 consecutive outpatients were enrolled (39 males and 28 females, aged years). The patients were attending the antihypertensive center of the Department of Internal Medicine at the University of Palermo (Italy). Subjects taking antihypertensive drugs were included when blood pressure (average of three determinations) was greater than 140/90 mmhg on two occasions when the patients were off medications for at least 2 weeks. 22 Systolic (SBP), diastolic (DBP) and mean blood pressure (MBP) were taken. MBP was calculated from the equation [DBP þ 1 3 (SBP DBP)]. Subjects were defined as overweight and obese on the basis of sex-specific 85th percentile of BMI values as reported in the Italian Consensus Conference of Obesity. 23 Accordingly, the men with BMI X25 and o30 kg/m 2 and women with BMI X24.7 and o27.3 kg/m 2 were considered overweight subjects, whereas the men with BMI X30 kg/m 2 and women with BMI X27.3 kg/m 2 were considered obese. In addition, the men with BMI o25 kg/m 2 and women with BMI o24.7 kg/m 2 were considered lean subjects. Central fat distribution was defined on the basis of sexspecific 85th percentile of waist-to-hip ratio (WHR). The cutoff value of central obesity was considered X0.81 for women and X0.92 for men. 23 Exclusion criteria included secondary or severe hypertension, diabetes mellitus, cardiovascular diseases (defined as myocardial infarction, recent stroke within 6 months, heart failure), systolic dysfunction (EFo50%), endocrinal diseases, renal failure, lung and liver fibrosis. Patients with microalbuminuria and proteinuria were also excluded since a TGFb1 overproduction has been reported in these conditions An accurate diagnostic procedure was performed in all the subjects with suspected coronary heart disease. This includes basal and 24 h ECG, echocardiography and, where necessary, ECG after exercise. According to BMI values, all the hypertensive subjects were classified as follows: Lean hypertensives: comprised 18 subjects (eight females and 10 males BMI kg/m 2 ; WHR cm). Overweight hypertensives: included 26 subjects (10 females and 16 males BMI 2872kg/m 2 ; WHR cm). Obese hypertensives: encompassed 23 subjects (10 females and 13 males BMI 3373 kg/m 2 ; WHR cm). Each patient gave informed consent after receiving a detailed description of the study procedure. The study was approved by the Ethics Commitee of our Institution. Methods Patients underwent a general analytical laboratory parameters profile including BUN, creatinine, glycaemia and electrolytes (serum sodium, potassium, chloride), by routine laboratory methods. Urinary albumin excretion (UAE) was also determined by immunonephelometric assay (Bohering Institute; limit of detection, 0.1 mg/dl; interassay coefficient 3.5%). To eliminate the intraindividual day-to-day variability of UAE, three consecutive 24 h urine collections were used. In addition, to assess the completeness of 24 h urine collection, measurements of urinary rate of clearance of creatinine were evaluated. Microalbuminuria was defined as level of UAE X20 and p300 mg/24 h.

3 TGFb1 Peripheral venous blood was obtained from each patient and the sera were isolated and stored at 701C. TGFb1 levels were determined by using a solid-phase-specific sandwich (R&D Systems Inc., Minneapolis, USA) ELISA technique as previously described. 17,18,24 The interassay and intraassay variations for determining TGFb1 were 8 and 6%, respectively. The sensitivity, hence, minimum level of detection of TGFb1 by sandwich ELISA, was 5 pg/ml. Echocardiographic parameters All patients underwent an echocardiography examination M- and B-mode, by a computerized echocardiography (ESAOTE, Italy) for the determination of following parameters: left ventricular telediastolic internal diameter (LVIDd), interventricular septum diastolic (IVSTd) and posterior wall thickness (PWTd). The Penn convention was used to calculate left ventricular mass (LVM). LVM was normalized for height to the 2.7 power. 25 The relative wall thickness (RWT) by formula [(PWTd/LVIDd) 2] was also calculated. Ejection fraction from left ventricular end-diastolic and end-systolic volumes was measured from the apical four chamber view, using the ellipsoidal single-plane algorithm. Mean ejection fraction was automatically calculated by the echocardiographic processing system. In our laboratory, the ejection fraction calculated over five consecutive beats permitted optimal reproducibility and accuracy also in obese subjects. 7 The presence of LVH was established when LVM/h 2.7 was 450 g/m 2.7 for men and 447 g/m 2.7 for women. 25 LV relaxation and filling were evaluated by pulsed-wave Doppler interrogation of the LV inflow tract from the apical four-chamber view, with the sample volume placed at the tips of the mitral valve. After a stable signal of the transmitral flow velocity was obtained, the Doppler cursor was moved towards the LV outflow tract in the apical fivechamber view for recording both mitral and aortic signals, including the closing click of the aortic valve and the opening click of the mitral valve. Doppler signals were recorded at high speed ( mm/s) with the subjects in held expiration. An average of five beats was used for analysis. IVRT was calculated as the time from the closure click of the aortic valve to the opening click of the mitral valve. When either the closing or opening click was not identified, the time from the end of the aortic flow to the onset of mitral flow from the continuous wave interrogation of the LV inflow outflow tract was used. Peak early transmitral flow velocity (E), peak late transmitral flow velocity (A) and the deceleration time of E velocity (DTE) were measured at the tips of mitral leaflets at the maximum amplitude of E velocity. DTE was measured as the time from peak E velocity to the time when E wave descent intercepted the zero line. According to European Guidelines 26 only subjects with normal systolic function (EF450%) and IVRT492 ms (o30 years), IVRT4100 ms (30 50 years), IVRT4105 ms (450 years) and/or E/A o1 þ DTE4220 ms (o50 years), E/A o0.5 þ DTE4 280 ms (450 years) were considered with diastolic dysfunction. To better evaluate the relationship between diastolic function and circulating TGFb1, all the hypertensives were further subgrouped into two subsets of patients with normal diastolic function or with diastolic dysfunction (Table 3). Statistical analysis Differences among the three groups were analysed by one-way analysis of variance and the post hoc Newman Keuls test. Since an asymmetry of TGFb1 data distribution was evident (high skewness), the difference in TGFb1 values among the groups were analysed by a nonparametric test, that is, the Kruskal Wallis test. Differences among LVH prevalence of three groups were analysed by w 2 test. Differences between hypertensives with normal diastolic function or with diastolic dysfunction were analysed by unpaired t-test. The relationship among TGFb1 and the other parameters was analysed by univariate and multiple regression analysis. Multiple regression analysis was used to investigate the relationship between TGFb1 and diastolic function measurements when corrected for other variables. Firstly, for selection of variables, we analysed the correlation matrix and then TGFb1, BMI, LVM/h 2.7, MBP, HR, IVRT, E/A ratio, age and gender were entered into the model. The best subset selection by examination of all possible regressions was used computing Mallow s Cp statistics. We also controlled for collinearity and this confirmed the selection of variables. Finally, TGFb1, BMI, E/A velocity ratio and IVRT were the best variables to fit the model. On the contrary, age, HR and LVM/h 2.7, were less relevant and were excluded by the model. Accordingly, E/A was considered an independent variable and TGFb1, BMI and IVRT were considered dependent variables. A regression equation and multiple correlation coefficients were computed. Both controlled for autocorrelation using the Dubin Watson test statistic and examined the scatter plot of the residuals against fitted TGFb1 and E/A velocity ratio to investigate distributions of data for deviation from normality. Data are expressed as mean value7s.d. A Po0.05 was considered statistically significant. Results Characteristics of study groups are reported in tables and figures. There were no significant changes among the groups in age, sex distribution, blood pressure values, BUN and creatinine levels (Table 1). 545

4 546 Table 1 Clinical characteristics and TGFb1 values in lean, overweight and obese hypertensive (HT) patients Table 2 Left ventricular geometry and function in lean, overweight and obese hypertensives (HT) Lean HT (no 18) Overweight HT (no 26) Obese HT (no 23) Lean HT (no 18) Overweight HT (no 26) Obese HT (no 23) Sex (F/M) 8/10 10/16 10/13 Age (years) BMI (kg/m 2 ) * *, ** WHR * *, ** HR (b/min) SBP (mmhg) DBP (mmhg) MBP (mmhg) BUN (mg/dl) Creatinine (mg/dl) TGFb1 (ng/ml) * 49718*, ** Minimum Maximum Median BMI: body mass index; WHR: waist hip ratio; HR: heart rate; SBP: systolic blood pressure; DBP: diastolic blood pressure; MBP: mean blood pressure; BUN: blood urea nitrogen; TGFb1: transforming growth factor b1. *Po0.05 vs A; **Po0.05 vs B. Differences of TGFb1 among the groups were analysed by the Kruskal Wallis test. LVIDd (mm) IVSTd (mm) PWTd (mm) RWT [(PWTd/LVIDd) 2] LVM (g) * *, ** LVM/h 2.7 (g/m 2.7 ) * 50712* LVEF (%) %LVH 2/18 (11%) 8/26 (35%) 9/23 (48%)*** E/A velocity ratio * * DTE (ms) * * IVRT (ms) * 98722* LVIDd: left ventricular internal diastolic diameter; IVSTd: interventricular septum thickness diastolic; PWTd: posterior wall thickness; RWT: relative wall thickness; LVEF: left ventricular ejection fraction; LVM: left ventricular mass; LVM/h 2.7 : left ventricular mass normalized to height 2.7 ; %LVH: percentage of subjects with left ventricular hypertrophy; E/A velocity ratio: peak early transmitral flow velocity (E), peak late transmitral flow velocity (A) ratio; DTE: E deceleration time; IVRT: isovolumic relaxation time. *Po0.05 vs A; **Po0.05 vs B; ***z-test: Po0.03 vs A. BMI, WHR and TGFb1 values were significantly (Po0.05) higher in obese and overweight than lean, and in obese than overweight hypertensives (Table 1). Cardiac parameters Total LVM and DTE values were significantly (Po0.05) higher in both overweight and obese than lean, and in obese than in overweight hypertensives. Idexed LVM and IVRT values were significantly (Po0.05) higher in obese and in overweight than lean hypertensives, but not in obese than in overweight hypertensives. E/A velocity ratio was significantly (Po0.05) lower in both overweight and obese than lean hypertensives, but not in obese than overweight hypertensives. A significant (Po0.05) higher prevalence in LVH was found in obese than lean hypertensives (48 vs 11%). An evident but not significant higher prevalence in LVH was found in overweight than lean hypertensives (35 vs 11%). No significant changes in the remaining cardiac parameters were observed among the three groups (Table 2). Table 3 Characteristics in hypertensive (HT) subjects with altered or normal diastolic function HT with normal diastolic function (no 47) HT with diastolic dyfunction (no 20) Sex (F/M) 16/31 12/8 Age (years) * BMI (kg/m 2 ) WHR HR (b/min) SBP (mmhg) DBP (mmhg) MBP (mmhg) TGFb1 (ng/ml) ** Minimum Maximum Median LVM/h 2.7 (g/m 2.7 ) * %LVH 8/47 (17%) 13/20 (65%)*** BMI: body mass index; WHR: waist hip ratio; HR: heart rate; SBP: systolic blood pressure; DBP: diastolic blood pressure; MBP: mean blood pressure; TGFb1: transforming growth factor b1; LVM/h 2.7 :left ventricular mass normalized to height 2.7 ; %LVH: percentage of subjects with LVH. *Po0.001; **Po0.04; ***w 2 test Po Differences of TGFb1 among the groups were analysed by the Kruskal Wallis test. Hypertensives with normal diastolic function vs hypertensives with diastolic dysfunction To better evaluate the impact of TGFb1 on diastolic function, all the hypertensive subjects were also subdivided according to the presence or the absence of diastolic dysfunction (Table 3). The two groups of hypertensives were comparable for sex distribution, BMI, WHR and blood pressure levels. On the contrary, age (Po0.001), LVM/h 2.7 (Po0.001), TGFb1 (Po0.004) levels and prevalence of LVH (Po0.001) were significantly higher in hypertensives with diastolic dysfunction than hypertensives with normal diastolic function (Table 3).

5 Correlation In all hypertensive subjects, circulating TGFb1 levels were positively correlated with BMI (r ¼ 0.60; Po0.0001), LVM/h 2.7 (r ¼ 0.28; Po0.03), IVRT (r ¼ 0.30; Po0.02) and negatively with E/A ratio (r ¼ 0.38; Po0.002) (Figures 1 and 2). Multiple regression analysis indicated that TGFb1, BMI and IVRT were independently related to E/A ratio and they explained 71% of E/A variability (r ¼ 0.84; Po0.0001) (Table 4). TGF-beta r = 0.30 p< Discussion and conclusion In the current study, a strong relationship among circulating TGFb1, BMI, LVM and some measurements of diastolic function have been found in hypertensive subjects with normal left ventricular ejection fraction. In particular, the main interesting finding was the recognition in these subjects of an association between circulating TGFb1 overproduction and impaired left ventricular filling. This association was independent of changes in age, HR and LVM. In our opininon this finding might be TGF-beta IVRT r = p< TGF-beta r = 0.60 p< EIA Figure 2 Correlation between circulating TGFb1, IVRT and E/A. Table 4 Multiple regression analysis Coefficients t-value ¼ P-valueo TGF-beta BMI r = 0.28 p< LVM/H 2.7 Figure 1 Correlation between circulating TGFb1, BMI and LVM/ H 2.7. Intercept TGFb IVRT BMI E/A velocity ratio: peak early transmitral flow velocity (E); peak late transmitral flow velocity (A) ratio; BMI: body mass index; TGFb1: transforming growth factor b1; IVRT: isovolumic relaxation time. E/A ¼ TGFb IVRT BMI. F ¼ ; Po Multiple correlation coefficients: r ¼ 0.84; r 2 ¼ 71.04%; Ra 2 ¼ 69.66%. supported by a common opinion that abnormalities in the left ventricular diastolic properties can be detected also when systolic function is maintained. 3,10,27,28 In addition, the evidence that congestive heart failure may occur in the presence of normal ejection fraction reinforces the hypothesis of a relative independence of LV diastole from systolic mechanics, reported in normotensive and hypertensive obese subjects. 3,9,29,30

6 548 Nevertheless, it is often difficult to separate diastolic and systolic abnormalities from either mechanical or energetic stand points, although a time sequence between these phenomena cannot be excluded. However, some recent data suggest that abnormalities of LV diastolic properties may precede the appearance of LV systolic dysfunction. 10 Accordingly, to minimize the effects of systolic dysfunction, only hypertensive patients with normal ejection fraction were included in our investigation. To the best of our knowledge, this is the first study testing the hypothesis that circulating TGFb1 overproduction can influence the occurrence and development of left ventricular diastolic abnormalities in obese hypertensive subjects, and we think that our results support this hypothesis. In fact, we found higher circulating TGFb1 levels in hypertensive patients with increased BMI as compared with TGFb1 levels in hypertensives with normal BMI. In addition, deceleration time of E wave and IVRT was significantly prolonged and E/A velocity ratio was significantly reduced in obese and overweight than in lean hypertensives. Thus, the pattern emerging from our study is an association between abnormal left ventricular filling and circulating TGFb1 overproduction. Nevertheless, the nature of LV filling abnormalities with increasing BMI is partially unclear and a precise discrimination between pathologic or physiologic changes is difficult. However, a possible relationship between TGFb1 overproduction and left ventricular diastolic dysfunction has been recently reported in experimental studies and in patients after heart transplantation. 19,20 In our opinion the reasons for the early appearance of impaired left ventricular filling and delayed active relaxation (ie prolongation of both E deceleration time and IVRT) in hypertensive obese subjects with normal ejection fraction may be the relationship between BMI, blood pressure and TGFb1. In fact, circulating TGFb1 overproduction has been reported in hypertensive subjects, central obeses and related to the target-organ disease. 17,18,31 A genetic TGFb1 DNA polimorphysm, 32 angiotensin II activity 33 and shear stress 34 have been proposed to explain TGFb1 overproduction in obesity-associated hypertension. Experimental data indicate an elevated expression of TGFb1 in the adipose tissue in obese mice, 35 a direct relationship between BMI and TGFb1 in human adipose tissue 36 and a reduction in TGFb1 levels after weight loss. 16 Finally, an increased expression of TGFb1 associated with a left ventricular diastolic dysfunction has been reported in experimental and clinical studies suggesting a pathophysiological role of this cytokine in the occurrence and development of myocardial fibrosis. 19,20 To better explain the link between TGFb1 overproduction and diastolic function, all hypertensive patients were further subgrouped according to the European Society of Cardiology Guidelines 26 to detect diastolic dysfunction. A significant increase in TGFb1 values and LVM/h 2.7 and a higher prevalence of subjects with left ventricular hypertrophy were found in the subset of hypertensives with diastolic dysfunction in comparison with patients with normal diastolic function. According to literature data and to higher prevalence of LVH found in our hypertensives with altered diastolic function, TGFb1 overproduction might be able to affect diastolic function through an altered left ventricular geometry. Surprisingly, in our study, multivariate analysis indicated an independent role of TGFb1 in the occurence of alteration in left ventricular filling. In fact, our regression model indicated that 71% of E/A ratio variability may be explained by TGFb1, BMI and IVRT changes, independent of LVM values. Theoretically because TGFb1 overproduction stimulates collagen synthesis, chronic TGFb1 overproduction might affect left ventricular filling through an increase in myocardial fibrosis. In this context some data suggest that impairment of diastolic function might be related in obese hypertensives to metabolic or neurohormonal abnormalities (ie hyperinsulinaemia, insulin resistance or renin angiotensin hyperactivity) rather than LVH. 39 Obviously, the present study does not explore the pathophysiology of the possible link between TGFb1 overproduction, hyperinsulinaemia or insulin resistance, RAS hyperactivity and diastolic dysfunction, but we could hypothesize that TGFb1 overproduction may be able to affect left ventricular filling early in obese hypertensives, independent of the effects related to increased LVM. This hypothesis might be supported by the statement that, with the fibrotic remodelling process, an increase in the stiffness of left ventricular myocardium can lead to an early impairment of the diastolic function. 40 In addition, some clinical implications arise from the results of our study. In particular, an early recognition of altered left ventricular filling in the presence of the normal left ventricular systolic function may be beneficial in single subjects as well as in a strategy for general health care. The present study is limited by its exclusive utilization of echocardiographic techniques for the diagnosis of diastolic dysfunction. In view of this, most of noninvasive parameters suitable to assess LV diastolic function by Doppler echocardiography are reported to be heart rate, preload and after load dependent and might change overtime in a given patient. However, alternative techniques are either invasive or require exposure to radioisotopes so that they are not currently feasible. Another limitation might be the use of an ejection fraction X50% as the sole index of normal systolic function. Nevertheless, the present as well as other populations are consistent with a relatively low prevalence of systolic dysfunction so that a misclassification is highly unlikely.

7 A further potential source of error is also the variability of the echocardiographic parameters, but this is due to the wide variation of flow patterns than linear relationship between the E/A ratio and severity of diastolic function. Finally, more invasive techniques may be required in order to directly measure diastolic filling pressure, chamber and muscle stiffness constants as well as systolic function for a complete estimation of diastolic dysfunction in the population. The lack of evaluation of exercise training in subjects studied by us could represent another limit of our study, since some recent data 41 indicate that exercise may induce changes in circulating levels of TGFb1 in humans. However, none of the subjects included in this study was an athlete and their physical activity was very low, so that they may be considered as sedentary subjects. In conclusion, our data seem to indicate an important pathophysiological role of TGFb1 overproduction on altered left ventricular filling in a subset of hypertensives as well as central obese hypertensives. However, further studies are necessary to explain the nature of the association between circulating TGFb1 overproduction and impaired E/A velocity ratio in these subjects. What is already known on this topic Diastolic abnormalities may be detectable early in obese and hypertensive subjects TGFb1 plays a role in the occurrence of both myocardial fibrosis and diastolic dysfunction The block of TGFb1 is able to neutralize myocardial fibrosis What this study adds A first demonstration of an association between TGFb1 overproduction and impaired left ventricular filling in human obese hypertensives This relationship appears independent of LVM values, age and HR The recognition of obese hypertensives with TGFb1 overproduction and early diastolic abnormalities may improve the prevention strategy References 1 Must A et al. The disease burden associated with overweight and obesity. JAMA 1999; 282: Eckel RH, Krauss RM. American Heart Association call to action: obesity as a major risk factor for coronary artery disease. American Heart Association Nutrition Committee. Circulation 1998; 97(21): Licata G, Scaglione R, Parrinello G, Corrao S. Rapid left ventricular filling in untreated hypertensive subjects with or without left ventricular hypertrophy. Chest 1992; 120: Mureddu GF et al. Left ventricular filling pattern in uncomplicated obesity. Am J Cardiol 1996; 77: Licata G et al. Hypertension in obese subjects: distinct hypertensive subgroup. J Hum Hypertens 1990; 4: de Simone G et al. Relation of obesity and gender to left ventricular hypertrophy in normotensive and hypertensive adults. Hypertension 1994; 23: Licata G et al. Heredity and obesity associated hypertension. Impact of hormonal characteristics and left ventricular mass. J Hypertens 1995; 13: Licata G et al. Left ventricular function response to exercise in normotensive obese subjects: influence of degree and duration of obesity. Int J Cardiol 1992; 27: Scaglione R et al. Left ventricular diastolic and systolic function in normotensive obese subjects: influence of degree and duration of obesity. Eur Heart J 1992; 13: de Simone G et al. Relation of left ventricular diastolic properties to systolic function in arterial hypertension. Circulation 2000; 101: Mann DL. Mechanisms and models in heart failure. Circulation 1999; 100: Weber KT, Brilla CG, Janicki JS. Myocardial fibrosis: functional significance and regulatory factors. Cardiovasc Res 1993; 27: Roberts AM, Sporn MB. Transforming growth factor b. Adv Cancer Res 1988; 51: Roberts AB, Mc Cune BK, Sporn MB. TGF-b: regulation of extracellular matrix. Kidney Int 1992; 41: Border WA, Nobel NA. Transforming growth factor-b in tissue fibrosis. N Engl J Med 1994; 331: Porreca E et al. Transforming growth factor-b1 levels in hypertensive patients: association with body mass index and leptin. Am J Hypertens 2002; 15: Scaglione R et al. Relationship between transforming growth b1 and progression of hypertensive renal disease. J Hum Hypertens 2002; 16: Scaglione R et al. Central obesity and Hypertensive renal disease: association between higher levels of BMI, circulating transforming growth factor b1 and urinary albumin excretion. Blood Pressure 2003; 12: Fumitaka K et al. Transforming growth factor-b function blocking prevents myocardial fibrosis and diastolic dysfunction in pressure-overloaded rats. Circulation 2002; 106: Aziz T et al. Transforming growth factor b and diastolic left ventricular dysfunction after heart transplantation: echocardiographic and hystologic evidence. J Heart Transplant 2003; 22: Kohen J. Statistical Power Analysis for the Behavioural Sciences. Lawrence Erlbaum Associates: Hillsdale, NJ, Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure. The Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure. Arch Intern Med 1997; 157: Crepaldi G et al. Special report: Italian Consensus Conference-Overweight, Obesity and Health. Int J Obesity 1991; 15: Khanna A, Li B, Stenzel KH, Suthanthiran M. Regulation of new DNA synthesis in mammalian cells by ciclosporine: demonstration of a transforming growth b factor dependent mechanism of inhibition of cell growth. Transplantation 1994; 57:

8 de Simone G et al. Left ventricular mass and body size in normotensive children and adults: assessment of allometric relations and impact of overweight. JAm Coll Cardiol 1992; 20: European Study Group on Diastolic Heart Failure. How to diagnose diastolic heart failure. Eur Heart J 1998; 19: Inouye I et al. Abnormal left ventricular filling: an early finding in mild to moderate systemic hypertension. Am J Cardiol 1984; 53: Pearson AC et al. Assessment of diastolic function in normal and hypertrophied hearts: comparison of Doppler echocardiography and M-mode echocardiography. Am Heart J 1987; 113: Vasan RS, Benjamin EJ, Levy D. Congestive heart failure with normal left ventricular systolic function: clinical approaches to diagnosis and treatment of diastolic heart failure. Arch Intern Med 1996; 156: Bonow RO, Udelson JE. Left ventricular diastolic dysfunction as a cause of congestive heart failure: mechanisms and management. J Intern Med 1992; 117: Laviades C, Varo N, Diez J. Transforming growth factor b in hypertensives with cardiorenal damage. Hypertension 2000; 36: Grainger DJ et al. Genetic control of the circulating concentration of transforming growth factor type b1. Hum Mol Genet 1999; 8: Kim S et al. Angiotensin II type 1 receptor antagonist inhibits the gene expression of transforming growth factor b1 and extracellular matrix in cardiac and vascular tissues of hypertensive rats. J Pharmacol Exp Ther 1995; 273: Neghishi M et al. Up-regulatory expression of furin and transforming growth factor b by fluid shear stress in vascular endothelial cells. Arterioscl Thromb Vasc 2001; 21: Samad F, Yamamoto K, Pandey M, Loskutoff DJ. Elevated expression transforming growth factor type beta in adipose tissue from obese mice. Mol Med 1997; 3: Alessi MC et al. Plasminogen activator inhibitor, transforming growth factor-b, and BMI are closely associated in human adipose tissue during morbid obesity. Diabetes 2000; 49: Yang CM, Kandaswamy V, Young D, Sen S. Changes in collagen phenotypes during progression and regression of cardiac hypertrophy. Cardiovascular Res 1997; 36: Muiesan ML. Left ventricular hypertrophy: a new approach for fibrosis. J Hypertens 2002; 20: Grandi AM et al. Insulin and diastolic dysfunction in lean and obese hypertensives. Genetic influence. Hypertension 1999; 34: Schannwell CM et al. Left ventricular diastolic function in physiologic and pathologic hypertrophy. Am J Hypertens 2002; 15(6): Heinemeier K, Langberg H, Kjaer M. Exercise-induced changes in circulating levels of transforming growth factor-beta-1 in humans: methodological considerations. Eur J Appl Physiol 2003; 90: