Journal of Cardiac Failure Vol. 20 No. 2 2014

Association of Left Ventricular Longitudinal and Circumferential Systolic Dysfunction With Diastolic Function in Hypertension: A Nonlinear Analysis Focused on the Interplay With Left Ventricular Geometry PIERCARLO BALLO, MD,1 STEFANO NISTRI, MD, PhD,2 MATTEO CAMELI, MD,3 BARBARA PAPESSO, MD,2 FRANK LLOYD DINI, MD,4 MAURIZIO GALDERISI, MD,5 ALFREDO ZUPPIROLI, MD,6 AND SERGIO MONDILLO, MD,3 ON BEHALF OF THE WORKING GROUP NUCLEUS ON ECHOCARDIOGRAPHY OF THE ITALIAN SOCIETY OF CARDIOLOGY Florence, Altavilla Vicentina, Siena, Pisa, and Naples, Italy

ABSTRACT Background: The relationships of left ventricular (LV) longitudinal and circumferential systolic dysfunction with diastolic performance in hypertensive patients have never been compared. Methods and Results: In 532 asymptomatic hypertensive patients, circumferential function was assessed with the use of midwall fractional shortening (mFS) and stress-corrected mFS (SCmFS), whereas longitudinal function was assessed with the use of left atrioventricular plane displacement (AVPD) and systolic mitral annulus velocity (s0 ). Early diastolic annular velocity (e0 ) and the E/e0 ratio were measured. Global longitudinal and circumferential strain were determined in a subset of 210 patients. e0 was linearly related to all systolic indexes (AVPD: R 5 0.40; s0 : R 5 0.39; mFS: R 5 0.16; SCmFS: R 5 0.17; all P ! .0001), but the correlations were stronger with longitudinal indexes than with circumferential ones (P ! .0001). E/e0 was nonlinearly related to AVPD (R 5 0.49; P ! .0001) and s0 (R 5 0.34; P ! .0001) and showed no relationship with mFS and SCmFS. Longitudinal indexes were superior to circumferential ones in predicting e0 !8 cm/s, E/e0 !8, and E/e0 $13. The effect of LV geometry on LV diastolic function was evident among patients with preserved systolic longitudinal function, but was blunted among patients with impaired longitudinal function. In multivariable analyses, only longitudinal indexes remained associated with e0 and E/e0 . Analyses using strains provided similar results. Conclusions: In asymptomatic hypertensive subjects, LV diastolic performance is independently associated with longitudinal systolic dysfunction, but not with circumferential systolic dysfunction. Subtle longitudinal systolic impairment plays a role in mediating the effect of LV geometry on diastolic performance. These findings may support the need of critically revising the concept of isolated diastolic dysfunction in these patients. (J Cardiac Fail 2014;20:110e120) Key Words: Systole, diastolic dysfunction, hypertension, tissue Doppler imaging, speckle tracking.

Systemic hypertension is a major cause of left ventricular (LV) diastolic dysfunction, which has been associated with subtle systolic impairment in both longitudinal and circumferential midwall systolic function.1e10 Although a nonlinear relation between LV longitudinal and circumferential midwall systolic indexes has been demonstrated

in hypertensive patients,11 the relationships between systolic and diastolic indexes have yet to be elucidated. In particular, the hypothesis that the prevalence and severity of diastolic dysfunction might increase nonuniformly over the entire range of LV systolic dysfunction has never been tested. From a clinical point of view, this analysis may be helpful

From the 1Cardiology Unit, S Maria Annunziata Hospital, Florence, Italy; 2Cardiology Service, CMSR Veneto Medica, Altavilla Vicentina, Italy; 3Department of Cardiovascular Disease, University of Siena, Siena, Italy; 4Cardiac, Thoracic and Vascular Department, University of Pisa, Pisa, Italy; 5Department of Translational Medical Sciences, Federico II University Hospital, Naples, Italy and 6Local Health Unit, Department of Cardiology, Florence, Italy. Manuscript received July 2, 2012; revised manuscript received December 2, 2013; revised manuscript accepted December 11, 2013.

Reprint requests: Dr. Piercarlo Ballo, MD, Department of Cardiology, S Maria Annunziata Hospital, Via dell’Antella 58, Florence, Italy. Tel: þ390552496582; Fax: þ390552496553. E-mail: [email protected] See page 118 for disclosure information. 1071-9164/$ - see front matter Ó 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.cardfail.2013.12.009

110

Systolic vs Diastolic Dysfunction in Hypertension

in identifying appropriate cutoff values for longitudinal or circumferential systolic indexes associated with increased risk of advanced diastolic dysfunction. In addition, the analysis could be useful to better clarify the interplay between systolic impairment, diastolic dysfunction, and abnormalities in LV mass and geometry in hypertensive patients. The aim of the present study was to explore and compare the relationships of both longitudinal and circumferential midwall systolic performance with LV diastolic dysfunction in a population of asymptomatic uncomplicated patients with systemic hypertension. Materials and Methods Study Population A total of 532 asymptomatic patients aged $16 years and affected by systemic hypertension, consecutively enrolled at our laboratories, were included in this study. Hypertension was diagnosed as blood pressure $140/90 mm Hg on average measurements obtained in O2 visits, or as antihypertensive therapy with documented history of hypertension.12 Exclusion criteria were: overt coronary artery disease (defined by $ 1 of the following: history of effort angina, acute coronary syndrome, or revascularization procedures; evidence of positive exercise stress test; segmental wall abnormalities at echocardiography); history of atrial fibrillation, atrial flutter or other major arrhythmias; mitral regurgitation of higher degree than trivial; aortic regurgitation; valvular stenosis; severe mitral annulus calcification; previous mitral valve surgery; hypertrophic cardiomyopathy; left bundle branch block; pacemaker implantation; heart transplantation; inadequate acoustic windows; secondary hypertension; diabetes mellitus; chronic kidney disease (estimated glomerular filtration rate ! 60 mL min1 1.73 m2); other significant systemic disease. All procedures were conducted in accordance with the Declaration of Helsinki, and all subjects gave their consent to participate in the study. Echocardiography

General Measurements. Studies were performed with the use of commercially available ultrasound systems (Vivid 7; GE Healthcare, Milwaukee, Wisconsin) equipped with 3S and 3V probes. In accordance with American Society of Echocardiography (ASE) recommendations,13 LV dimensions, mass, relative wall thickness (RWT), and geometric pattern were determined from parasternal views, and sex-specific cutoff values for indexed LV mass (O115 g/m2 for men and O95 g/m2 for women) were used to identify LV hypertrophy. An RWT of O0.42 was considered to define concentric geometry. LV volumes and ejection fraction were measured with the use of the modified biplane Simpson rule. Indexes of LV inflow were measured by pulsed Doppler from the apical 4-chamber view. Left atrial volume was calculated using the biplane method of disks. LV Circumferential Systolic Function. Midwall fractional shortening (mFS) was estimated with the use of an ellipsoidal model of the left ventricle.14 Circumferential end-systolic stress was calculated with the use of the Gaasch formula.15 The logarithmic stress-mFS relationship observed in 133 age- and sex-matched healthy control subjects (age 63.7 6 13.9 years, 51.9% male) was used to calculate predicted mFS. Matching for



Ballo et al

111

age was performed by defining groups of 4 randomly selected hypertensive patients, and by matching 1 control to the average age of each group with the use of a 65 years criterion. Stresscorrected mFS (SCmFS), a load-independent index of circumferential myocardial contractility, was obtained by dividing observed by predicted mFS.16 Sex-specific cutoff values for mFS (!14% in men, !15% in women) and SCmFS (!89.2% for both men and women) were used to identify reduced circumferential systolic function.13,17 Current ASE guidelines recommend the use of midwall indexes to detect subtle LV circumferential systolic dysfunction, because they better reflect intrinsic contractility than LV endocardial fractional shortening or ejection fraction.13 LV Longitudinal Systolic Function. M-Mode and pulsed tissue Doppler imaging of mitral annulus motion were obtained from the apical 4-chamber view. Atrioventricular plane displacement (AVPD), peak systolic annular velocity (s0 ), and peak early diastolic annular velocity (e0 ) were calculated by averaging values measured at the septal and lateral annuluar sites.18 Cutoffs of !12 mm for AVPD and of !8 cm/s for s0 were used to identify reduced longitudinal systolic function.19 LV Diastolic Function. Average e0 was used as a measure of LV relaxation, whereas the E/e0 ratio was used as an estimate of LV filling pressure.20 Abnormal LV relaxation was identified by an e0 of !8 cm/s, whereas E/e0 values of !8 and $13 were used to identify normal and raised LV filling pressure, respectively.21 Normal diastolic function and grade IeIII diastolic dysfunction were defined in accordance with ASE recommendations for the evaluation of LV diastolic function by echocardiography.22 Speckle Tracking. Speckle tracking echocardiography was performed in a subset of 210 patients. Apical 4-, 3-, and 2chamber images and short-axis basal, middle, and apical images of the left ventricle were obtained with the use of 2-dimensional grayscale imaging, during breath-hold with a stable electrocardiogram recording, and with a frame rate of 60e80/s. Images were analyzed offline with the use of a 18-segment LV model and dedicated software (Echopac 7.0; GE Healthcare), as previously described.23 Briefly, the software allows quantifying cardiac motion based on frame-to-frame tracking of ultrasonic speckles, thus providing curves that express myocardial deformation over time. During systole, strain curves are characterized by a negative deflection, reflecting myocardial negative deformation (ie, fiber shortening). Peak global longitudinal strain (GLS) and global circumferential strain (GCS) during systole were calculated by averaging peak strain in all LV segments.24 Segments in which no adequate tracking quality could not be obtained despite manual adjustment were excluded, and patients in whom no adequate tracking quality occurred in O3 segments were excluded from the analysis. Reproducibility. Standardization of image acquisition and data analysis to meet current ASE guidelines for correct measurements of LV systolic and diastolic indexes was defined in a previous meeting of the Echocardiography Study Group of the Italian Society of Cardiology.25 All measurements were obtained as the average of 3 values in consecutive cycles. Intraobserver coefficients of variation and intraclass correlation coefficients, evaluated in a randomly selected subset of 65 subjects, were: AVPD 3.8%, R 5 0.94; s0 4.0%, R 5 0.92; mFS 4.9%, R 5 0.89; SCmFS 5.9%, R 5 0.86; e0 3.8%, R 5 0.95; E/e0 5.1%, R 5 0.91; GLS 6.2%, R 5 0.84; GCS 6.5%, R 5 0.83 (P ! .0001 for all). Corresponding values for interobserver analysis were: AVPD 4.4%, R 5 0.91; s0 4.7%, R 5 0.90; mFS 5.6%, R 5 0.86; SCmFS

112 Journal of Cardiac Failure Vol. 20 No. 2 February 2014 Table 1. General Characteristics in the Overall Study Population and According to Normal or Reduced Longitudinal Left Ventricular (LV) Systolic Function

Age (y) Male (n) Body surface area (m2) Body mass index (kg/m2) Systolic blood pressure (mm Hg) Diastolic blood pressure (mm Hg) Heart rate (beats/min) LV mass index (g/m2) Relative wall thickness Indexed LV end-diastolic volume (mL/m2) Indexed LV end-systolic volume (mL/m2) Ejection fraction (%) E/A ratio Deceleration time (ms) mFS (%) SCmFS (%) AVPD (mm) s0 (cm/s) e0 (cm/s) E/e0 ratio Indexed left atrial volume (mL/m2)

Overall Population

Normal AVPD (n 5 399)

Reduced AVPD (n 5 133)

P Value

Normal s0 (n 5 343)

Reduced s0 (n 5 189)

P Value

64.2 6 9.9 277 (52.1%) 1.85 6 0.21 26.8 6 4.4 139.2 6 18.0 80.6 6 8.9 72.7 6 12.8 105.1 6 25.5 0.43 6 0.07 59.4 6 15.9 18.7 6 7.1 62.5 6 7.7 0.91 6 0.34 234 6 66 16.9 6 2.5 90.7 6 13.3 13.5 6 2.3 9.0 6 2.1 9.3 6 2.6 8.3 6 3.2 38.3 6 12.4

63.2 6 9.9 220 (55.7%) 1.87 6 0.22 26.7 6 4.3 139.1 6 18.4 81.2 6 9.0 72.6 6 12.9 103.5 6 25.7 0.42 6 0.07 59.3 6 15.6 18.4 6 7.0 63.1 6 7.0 0.89 6 0.33 236 6 68 16.9 6 2.5 91.0 6 13.5 14.5 6 1.8 9.4 6 2.1 9.7 6 2.6 7.6 6 2.5 36.6 6 11.3

67.1 6 9.5 57 (43.2%) 1.81 6 0.19 27.0 6 4.6 139.6 6 16.9 78.8 6 8.3 73.2 6 12.8 110.0 6 24.2 0.45 6 0.07 59.7 6 16.6 19.5 6 7.5 61.1 6 8.1 0.95 6 0.39 229 6 61 16.8 6 2.3 90.0 6 12.6 10.7 6 1.0 7.9 6 1.8 8.0 6 2.2 9.7 6 4.0 43.3 6 14.6

!.0001 .014 .0010 .60 .80 .07 .72 .013 .0008 .82 .17 .012 .13 .31 .38 .42 !.0001 !.0001 !.0001 !.0001 !.0001

63.2 6 10.2 202 (58.9%) 1.89 6 0.22 26.8 6 4.2 138.4 6 17.6 80.7 6 9.5 72.1 6 12.4 102.0 6 23.6 0.42 6 0.07 59.7 6 14.8 18.1 6 6.8 63.6 6 7.2 0.92 6 0.33 234 6 66 17.0 6 2.4 91.3 6 13.1 14.2 6 2.3 10.2 6 1.7 10.0 6 2.5 7.6 6 2.6 37.4 6 11.8

66.2 6 9.0 75 (39.7%) 1.80 6 0.19 26.8 6 4.9 140.7 6 18.7 80.4 6 7.6 73.9 6 13.0 111.7 6 27.9 0.45 6 0.07 58.8 6 17.9 19.9 6 7.5 60.7 6 7.2 0.88 6 0.36 236 6 68 16.6 6 2.5 89.5 6 13.2 12.4 6 2.0 6.9 6 0.9 7.9 6 2.4 9.5 6 3.8 41.3 6 14.2

.0011 !.0001 !.0001 .97 .18 .83 .34 .0003 .0004 .60 .017 !.0001 .18 .76 .21 .23 !.0001 !.0001 !.0001 !.0001 .015

A, peak late diastolic mitral flow velocity; AVPD, left atrioventricular plane displacement; E, peak early diastolic mitral flow velocity; e0 , peak early diastolic mitral annulus velocity; mFS, midwall fractional shortening; s0 , peak systolic mitral annulus velocity; SCmFS, stress-corrected mFS.

6.7%, R 5 0.83; e0 4.4%, R 5 0.92; E/e0 5.9%, R 5 0.85; GLS 6.7%, R 5 0.83; GCS 7.1%, R 5 0.82 (P ! .0001 for all). Statistical Analysis Data were expressed as mean 6 SD. Nonlinear regression analysis using cubic splines was performed to explore the relationships between systolic and diastolic indexes. The Akaike information criterion was used to select the best fitting function after considering the number of regression parameters. Correlation

coefficients were compared with the use of the Fisher Z-transform. The performance of systolic indexes for the discrimination of e0 ! 8 cm/s, E/e0 !8, and E/e0 $13 was analyzed by receiver operating characteristic (ROC) analysis. The areas under the ROC curves (AUCs) were compared with the use of the nonparametric method of DeLong et al. The odds ratios (ORs) and 95% confidence intervals (CIs) for the prediction of e0 !8 cm/s, E/e0 !8, and E/e0 O15 were calculated for each systolic index. Differences in the distribution of diastolic dysfunction degrees were explored by the

Table 2. Clinical and Echocardiographic Characteristics After Stratification by Left Ventricular (LV) Diastolic Dysfunction Degree

Age (y) Male gender (n) Body surface area (m2) Body mass index (Kg/m2) Systolic blood pressure (mm Hg) Diastolic blood pressure (mm Hg) Heart rate (bpm) LV mass index (g/m2) Relative wall thickness Indexed LV end-diastolic volume (mL/m2) Indexed LV end-systolic volume (mL/m2) Ejection fraction (%) E/A ratio Deceleration time (ms) mFS (%) SCmFS (%) AVPD (mm) s0 (cm/s) e0 (cm/s) E/e0 ratio Indexed left atrial volume (mL/m2)

Normal (n 5 150)

Grade 1 (n 5 267)

Grade 2e3 (n 5 115)

P Value (ANOVA)

59.5 6 9.9 96 (63.5%) 1.90 6 0.20 26.3 6 3.5 136.6 6 13.3 81.8 6 9.6 72.5 6 13.3 97.2 6 22.3 0.43 6 0.06 56.1 6 12.7

65.3 6 9.3* 129 (48.9%)* 1.87 6 0.22 27.5 6 4.4 140.6 6 18.8 79.9 6 8.5 72.7 6 13.2 103.9 6 23.5 0.43 6 0.07 60.1 6 14.0

67.8 6 9.0* 52 (46.4%)* 1.76 6 0.18*,y 25.4 6 4.9y 139.5 6 21.2 80.6 6 8.8 73.0 6 10.0 118.9 6 28.8 0.45 6 0.08 62.1 6 22.1*

!.0001 .0049 !.0001 .010 .10 .42 .92 !.0001 !.0001 .015

19.0 6 6.2

18.0 6 6.9

62.4 0.96 208 17.0 91.4 14.3 9.7 11.6 6.4 28.4

6 6 6 6 6 6 6 6 6 6

7.0 0.35 65 2.6 13.6 1.9 2.0 1.9 1.8 9.5

Abbreviations as in Table 1. *P ! .05 vs normal (Tukey post hoc pairwise comparison). y P ! .05 vs grade 1 (Tukey post hoc pairwise comparison).

63.1 0.84 249 16.9 91.1 13.6 9.1 9.0 7.9 38.9

6 6 6 6 6 6 6 6 6 6

7.5 0.29* 64* 2.4 13.1 2.4* 2.0* 2.3* 2.5* 11.2*

,y

20.1 6 8.3

.07

6 6 6 6 6 6 6 6 6 6

.33 .0003 !.0001 .30 .21 !.0001 !.0001 !.0001 !.0001 !.0001

61.9 0.98 218 16.5 88.8 12.4 7.8 6.9 11.7 49.4

7.4 0.42y 60y 2.4 13.2 2.3*,y 1.9*,y 1.6*,y 3.5*,y 14.2*,y

Systolic vs Diastolic Dysfunction in Hypertension



Ballo et al

113

Fig. 1. Relationships of left atrioventricular plane displacement (AVPD), peak systolic mitral annulus velocity (s0 ), midwall fractional shortening (mFS), and stress-corrected mFS (SCmFS) with peak early diastolic mitral annulus velocity (e0 ). chi-square test. GLM analysis was performed to identify interactions between LV geometric patterns and systolic indexes. Multivariable linear and nonlinear regressions were used to assess the relations of systolic indexes with e0 and E/e0 . For multivariable nonlinear regression, the Levenberg-Marquardt algorithm was used. The likelihood ratio test was used to compare the fit of different models. The significance level was set at 0.05. All tests were 2 tailed. Analyses were performed with the use of SPSS 15.0 (SPSS, Chicago, Illinois). The sample size was calculated by hypothesizing that a difference in the AUC for the prediction of e0 !8 cm/s would have been found between s0 and SCmFS. Assuming a prevalence of 30% for reduced e0 and an AUC of !0.60 for the systolic index characterized by the worse performance, a sample of 427 subjects would have achieved 80% power to detect a 0.10 difference between the 2 AUCs with the use of a 2-sided z test at P ! .05. Power calculations were performed with the use of Pass 2008 (NCSS, Kaysville, Utah).

Results General Features

Main characteristics of the overall study population and after stratification by longitudinal systolic function and diastolic dysfunction degree are presented in Tables 1 and 2. A total of 187 subjects (35.1%) had normal geometry, whereas concentric remodeling, eccentric hypertrophy, and concentric hypertrophy were present in 187 (35.1%), 133 (25.0%), 89 (16.7%), and 123 (23.1%) patients, respectively. Normal diastolic function and grade I, II, and III diastolic dysfunction were observed in 150 (28.2%), 267 (50.2%), 89 (16.7%), and 26 (4.9%) patients, respectively. A total of 184 subjects (34.6%) showed an e0 !8 cm/s,

whereas the E/e0 ratio was !8 in 301 patients (56.6%) and $13 in 43 (8.1%). Most used antihypertensive medications in the overall population were angiotensin-converting enzyme inhibitors (46.1%), diuretics (36.1%), calcium antagonists (24.8%), beta-blockers (21.8%), angiotensin II receptor antagonists (19.0%), and alpha-blockers (4.3%). Relationships With e0

The relationships between systolic indexes and e0 were all best expressed by linear functions (Fig. 1). The correlations observed for longitudinal indexes were stronger than those observed for circumferential indexes (P ! .0001 for all comparisons). ROC analysis (Fig. 2) showed that longitudinal indexes had larger AUCs for the prediction of e0 !8 cm/s compared with circumferential indexes. An AVPD !12.6 mm was associated with O3-fold higher risk of e0 !8 cm/s (OR 3.6 [95% CI 2.4e5.2]; P ! .0001), whereas s0 !8.3 cm/s was associated with almost 5-fold increased risk (OR 4.8 [95% CI 3.3e7.0]; P ! .0001). Relationships With E/e0

The relationships of longitudinal systolic indexes with E/e0 were best expressed by nonlinear functions, whereas no significant relationships between E/e0 and circumferential midwall indexes were found (Fig. 3). ROC analysis (Fig. 4) showed a better accuracy of longitudinal indexes in predicting E/e0 of !8 and E/e0 $13 compared with circumferential indexes. Values of AVPD !11.5 mm and s0 !8.3 cm/s were associated with almost 6-fold and 4-fold increased risk of having an E/e0 ratio $13,

114 Journal of Cardiac Failure Vol. 20 No. 2 February 2014 echocardiographic confounders (model 1), and in a second model built by adding circumferential indexes to the group of covariates (model 2). In contrast, circumferential indexes showed a borderline association with e0 after adjustment for clinical and echocardiographic confounders (model 1), which was no longer evident after insertion of longitudinal indexes (model 2). When a similar analysis was performed to identify predictors of E/e0 , longitudinal indexesdbut not circumferential indexesdremained independently associated with E/e0 in both model 1 and model 2 (Table 4). In analyses testing the effect of 2-way interaction terms added to each of these multivariable models, the AVPDRWT interaction emerged as an additional independent predictor for the prediction of e0 (b 5 0.178; P ! .0001). A borderline significance of the s0 -RWT interaction for the prediction of e0 was also observed (b 5 0.102; P 5 .074). Circumferential indexes showed no significant interactions. Fig. 2. Receiver operating characteristic curves showing the performance of longitudinal and circumferential systolic indexes to predict peak early diastolic mitral annulus velocity (e0 ) !8 cm/s. Best cutoffs (sensitivity, specificity): AVPD !12.6 mm (58.9%, 72.1%); s0 !8.3 cm/s (64.3%, 73.8%); mFS !16.2% (53.3%, 64.5%); and SCmFS !89.4% (56.0%, 60.8%). Comparisons between areas under the curve: AVPD vs mFS: P 5 .0020; AVPD vs SCmFS: P 5 .0031; s0 vs mFS: P 5 .0004; s0 vs SCmFS: P 5 .0006.

respectively (AVPD: OR 5.8 [95% CI 3.1e11.0; P ! .0001]; s0 : OR 3.8 [95% CI 1.9e7.4; P ! .0001]). Association With Diastolic Dysfunction Degree

Compared with patients with normal AVPD or s0 , those with reduced longitudinal systolic indexes showed a lower prevalence of normal diastolic function and a higher prevalence of grade IIeIII diastolic dysfunction (Supplemental Fig. 1). In contrast, the severity of diastolic dysfunction did not differ between patients with normal and reduced mFS or SCmFS. Interaction With LV Hypertrophy and Geometry

The effect of geometrical patterns on e0 was more evident among patients with normal AVPD or normal s0 than in those with reduced longitudinal indexes, as shown by significant interactions (Supplemental Fig. 2). No interaction effects on the E/e0 ratio were observed. Circumferential midwall indexes showed no significant interactions with LV geometric pattern for either e0 (mFS: P 5 .87; SCmFS: P 5 .99) or E/e0 (mFS: P 5 .60; SCmFS: P 5 .97). Multivariable Analysis

Multivariable analysis for the prediction of e0 is presented in Table 3. Longitudinal indexes remained independently associated with e0 after adjustment for clinical and

Speckle Tracking Subset Analysis

In the subset of patients studied by speckle tracking, both GLS and GCS showed a significant correlation with e0 , but the correlation with GLS (R 5 0.55; P ! .0001) was stronger than that with GCS (R 5 0.19; P 5 .0062; P ! .0001 by comparison of correlation coefficients). GLS also showed a nonlinear negative relation with E/e0 (R 5 0.51, P ! .0001), whereas no significant relationship between GCS and E/e0 was found (R 5 0.11, P 5 .10). Results of multivariable analysis for strain indexes are presented in Table 5. GLS remained independently associated with both e0 and E/e0 after adjustment for clinical and echocardiographic confounders, and in a second model built by adding GCS to the group of covariates. GCS did not show significant associations with either e0 or E/e0 in multivariable models. Discussion Main Findings

A considerable proportion of hypertensive patients with LV diastolic dysfunction show a subtle impairment in systolic function that is not detected by LV ejection fraction. Longitudinal and circumferential midwall indexes are both sensitive markers of LV systolic impairment that can be used to identify this subtle systolic dysfunction, because they have been shown to be depressed in hypertensive patients with diastolic dysfunction and in those with heart failure and normal ejection fraction.26,27 However, little is known about the relationships of longitudinal and circumferential midwall systolic performance with diastolic dysfunction in hypertension. The present study shows that in uncomplicated and asymptomatic hypertensive patients: 1) the association of LV longitudinal systolic dysfunction with diastolic performance is stronger than that observed for circumferential midwall systolic dysfunction, independently from relevant covariates; 2) LV longitudinal systolic

Systolic vs Diastolic Dysfunction in Hypertension



Ballo et al

115

Fig. 3. Relationships of left atrioventricular plane displacement (AVPD), peak systolic mitral annulus velocity (s0 ), midwall fractional shortening (mFS), and stress-corrected mFS (SCmFS) with the ratio of peak early diastolic transmitral flow velocity to peak early diastolic mitral annulus velocity (E/e0 ). Top: The relationships between longitudinal indexes and E/e0 were nonlinear. The best AVPDeE/e0 fitting curve had a single knot at AVPD 5 13.5 mm, whereas the s0 eE/e0 fitting curve had a single knot at 8.8 cm/s. Bottom: No significant relationships were found between circumferential indexes and E/e0 with the use of either linear or cubic spline regression.

dysfunction plays an important role in mediating the effect of LV geometry on LV diastolic function; and 3) LV longitudinal systolic indexes are linearly related to e0 but nonlinearly related to the E/e0 ratio. Relationship Between Systolic and Diastolic Dysfunction in Hypertension

Earlier studies suggested that LV longitudinal and circumferential midwall systolic impairment are both related to LV diastolic dysfunction in patients with hypertension. An independent association between impaired circumferential midwall systolic mechanics and abnormal standard Doppler measures of LV diastolic filling has been reported in these patients,4,5 although other studies showed that mildly abnormal LV relaxation can be present despite normal circumferential midwall function.6 Evidence also showed that, in the presence of normal LV ejection fraction, longitudinal systolic velocities and strains are impaired in hypertensive subjects with diastolic dysfunction compared with those with normal diastolic function.7,8 LV longitudinal systolic dysfunction has also been hypothesized to play a key role in favoring the progression from asymptomatic hypertensive heart disease to heart failure with preserved ejection fraction.9 Our results add to these findings, showing that in asymptomatic subjects with uncomplicated hypertension, the association of LV diastolic performance with longitudinal systolic dysfunction is stronger than that with circumferential systolic dysfunction, and that only longitudinal indexes are independently associated with tissue Doppler measures of LV relaxation

and filling pressure. Interestingly, these results may be in accordance with a recent study in which LV global longitudinal strain, but not global circumferential strain, was an independent contributor of E/e0 in young untreated hypertensive patients and top-level rowers.10 From a pathophysiologic point of view, our findings show that an important link between diastolic and systolic dynamics emerges when the 2 processes are explored by a similar approach focused on longitudinal shortening and relengthening measures, pointing out the need of moving away from the concept of isolated diastolic dysfunction as a mechanism of heart failure. Interplay With LV Geometry

In the present population, the effect of geometric patterns on LV relaxation was relevant in the group with preserved longitudinal function, but blunted among patients in whom longitudinal dysfunction had already developed. This interaction effect was still evident in multivariable analysis. These findings suggest the intriguing hypothesis that longitudinal dynamics could play a major role in mediating the relationship between progression of LV diastolic dysfunction and abnormalities in LV mass and geometry in patients with hypertension. It is known that the development of different LV geometric patterns in hypertension depends on complex interaction between biochemical and hemodynamic factors, which may favor a progressive decline in longitudinal systolic function related to increased collagen synthesis and myocardial fibrosis.28 On the other

116 Journal of Cardiac Failure Vol. 20 No. 2 February 2014

Fig. 4. Receiver operating characteristic curves (ROCs) showing the performance of longitudinal and circumferential systolic indexes to predict a ratio of peak early diastolic transmitral flow velocity to peak early diastolic mitral annulus velocity (E/e0 ) (left) ! 8 and (right) $13, respectively. Best cut-offs (sensitivity, specificity) for predicting E/e0 !8: AVPD O13.6 mm (71.5%, 53.5%); s0 O8.3 cm/s (69.1%, 60.0%); mFS O16.5% (57.0%, 48.3%); and SCmFS O89.3% (57.7%, 46.9%); P ! .0001 for all comparisons of areas under the ROCs (AUCs) between longitudinal and circumferential indexes. Best cut-offs (sensitivity, specificity) for predicting E/e0 $13: AVPD !11.5 mm (51.2%, 84.7%); s0 !8.3 cm/s (69.8%, 63.4%); mFS !16.5% (51.2%, 55.1%); and SCmFS !85.4% (39.9%, 66.3%). AUC comparisons: AVPD vs mFS: P 5 .0005; AVPD vs ScmFS: P 5 .0022; s0 vs mF: P 5 .0008; s0 vs SCmFS: P 5 .0027.

hand, myocardial fibrosis is a major determinant of the deterioration in LV diastolic function associated with increased LV mass and concentric geometry.29 In this regard, recent evidence confirmed that the early stages of hypertensive heart disease are characterized by a strict relationship between impairment in longitudinal systolic function and increase in passive myocardial stiffness, suggesting that subtle abnormalities in long-axis systolic function may play a role in both progression of LV diastolic dysfunction and transition from asymptomatic hypertensive heart disease to heart failure with preserved ejection fraction.30 These considerations further support the need of critically reconsidering the concept of isolated diastolic dysfunction, and are in accordance with the increasing role of longitudinal measures of both systolic and diastolic function as stronger predictors of symptoms and outcome than radial measures and ejection fraction in patients with hypertension.31 Nonlinear Relationships With E/e0

An additional finding in the present study is that the relationships of longitudinal systolic indexes with e0 were best described by linear functions, whereas those with E/e0 were best described by nonlinear functions. Thus, changes in the higher range of longitudinal systolic indexes (ie, among patients with preserved longitudinal function) yielded relatively small changes in E/e0 , whereas similar

changes in the lower range (ie, among patients with advanced longitudinal dysfunction) yielded considerable increases in E/e0 . For example, a 25% decrease in AVPD from 20 mm to 15 mm corresponded to an increase in E/e0 from 6.4 to 7.6 (relative increase þ18.3%), whereas a 25% decrease in AVPD from 10 mm to 7.5 mm corresponded to an increase in E/e0 from 11.0 to 21.3 (relative increase þ94.0%). On the other side, as a result of linear relationship, proportional changes in longitudinal indexes were associated with similar relative changes in e0 across the whole range of either AVPD or s0 . From a clinical point of view, these findings may reflect the evidence that an impairment in LV relaxation is an early event in the history of hypertensive heart disease, whereas a substantial increase in LV filling pressure usually occurs in relatively advanced stages.31e34 Clinical Implications and Study Limitations A major prognostic value of LV diastolic dysfunction was previously demonstrated in several cardiovascular diseases,35e38 including systemic hypertension.39,40 According to the strong link between diastolic and longitudinal systolic mechanics, our findings suggest that the detection of longitudinal systolic dysfunction in asymptomatic subjects with uncomplicated hypertension might be helpful to identify a

Systolic vs Diastolic Dysfunction in Hypertension



Ballo et al

117

Table 3. Multivariable Linear Regression Analysis Showing the Associations of Systolic Indexes With Peak Early Diastolic Mitral Annulus Velocity (e0 ) in Unadjusted Models and After Controlling for Confounders Unadjusted

AVPD 0

b

95% CI

0.450

0.361e0.538 R2 5 0.158 0.383e0.579 R2 5 0.149 0.082e0.262 R2 5 0.026 0.018e0.051 R2 5 0.030

s

0.481

mFS

0.172

SCmFS

0.034

Model 2y

Model 1* P Value

b

95% CI

!.0001

0.412

!.0001

0.394

.0002

0.111

!.0001

0.023

0.324e0.500 R2 5 0.292 0.292e0.495 R2 5 0.260 0.004e0.218 R2 5 0.184 0.001e0.045 R2 5 0.185

P Value

b

95% CI

!.0001

0.417

!.0001

0.385

.042

0.082

.036

0.018

0.328e0.505 R2 5 0.302 0.283e0.488 R2 5 0.263 0.016e0.180 R2 5 0.327 0.002e0.037 R2 5 0.327

P Value

P Value for R2 Changez

!.0001

.12

!.0001

.39

.10

!.0001

.08

!.0001

Abbreviations as in Table 1. *Model 1: adjusted for age, sex, body mass index, heart rate, blood pressure, LV ejection fraction, indexed LV mass, relative wall thickness, E/A ratio, Ewave deceleration time, indexed left atrial volume, history of hypercholesterolemia, smoking, and use of each major pharmacologic class (as listed in Table 1). y Model 2: adjusted for AVPD and s0 as in model 1 þ circumferential indexes; adjusted for mFS and SCmFS as in model 1 þ longitudinal indexes. z P value expressing significance for R2 change from model 1 to model 2, calculated with the use of the likelihood ratio test.

subset of patients at high risk of advanced diastolic impairment. In this regard, it should be pointed out that longitudinal systolic indexes are simple, reproducible, easy to measure, usually feasible even in patients with relatively suboptimal image quality, and can be rapidly obtained even in busy echocardiography laboratories without refined last-generation software and postprocessing time capabilities. AVPD, in particular, can also be measured with the use of low-cost echocardiographic systems without tissue

Doppler imaging (eg, hand-held devices).41 In our population, an AVPD of !11.5 mm discriminated an E/e0 $13 with 85% specificity and was associated with a 6-fold increased risk of raised LV filling pressures. In clinical practice, this information could be particularly useful for an initial assessment of hypertensive patients in settings where tissue Doppler capabilities are not always available (eg, home visitations, bedside evaluation, emergency departments). Additionally, the results of this study suggest that

Table 4. Multivariable Nonlinear Regression Showing the Associations of Systolic Indexes With the Ratio of Early Diastolic Mitral Flow to Mitral Annulus Velocity (E/e0 ) in Unadjusted Models and After Controlling for Confounders Unadjusted

AVPD Intercept b1 b2 b3 b4

b

95% CI

148.5 30.7 2.25 0.055 0.053

84.1e212.9 14.3e47.2 0.96e3.54 0.021e0.089 0.034e0.072 R2 5 0.240

s0 Intercept b1 b2 b3 b4 mFS SCmFS

30.9 6.1 0.546 0.017 0.013 0.003 0.003

13.7e48.1 2.2e10.0 0.048e1.044 0.003e0.031 0.002e0.024 R2 5 0.121 0.107e0.114 R2 5 0.01 0.024e0.018 R2 5 0.01

Model 2y

Model 1* P Value

b

95% CI

!.0001 !.0001 !.0001 !.0001

159.9 33.8 2.52 0.063 0.064

97.1e222.7 18.3e49.4 1.24e3.79 0.028e0.097 0.016e0.113 R2 5 0.297

.0023 .031 .018 .021

47.9 13.7 1.540 0.059 0.076

.96

0.004

.78

0.003

21.7e74.2 3.1e24.2 0.166e2.913 0.002e0.117 0.011e0.141 R2 5 0.190 0.112e0.104 R2 5 0.074 0.023e0.018 R2 5 0.074

P Value

b

95% CI

!.0001 !.0001 .0003 .0097

161.8 34.2 2.55 0.063 0.066

98.1e225.6 18.4e49.7 1.26e3.84 0.029e0.098 0.017e0.115 R2 5 0.301

.011 .028 .046 .020

47.6 13.9 1.576 0.061 0.079

.94

0.048

.81

0.007

21.0e74.2 3.3e24.5 0.191e2.962 0.002e0.120 0.008e0.150 R2 5 0.193 0.055e0.142 R2 5 0.189 0.012e0.026 R2 5 0.188

P Value

P Value for R2 changez

!.0001 !.0001 .0003 .0088 .27 .010 .026 .042 .030 .34 .36 .48

!.0001 !.0001

Abbreviations as in Table 1. The analysis was performed with the use of the same models identified by univariable regression. Thus, cubic spline regression models were used for longitudinal indexes and linear models were used for circumferential indexes. The coefficients b1, b2, and b3 refer to the linear, quadratic, and cubic terms, respectively, of the cubic spline function. The coefficient b4 refers to the (X  k)3 term, where X is the systolic index (AVPD or s0 ) and k is the knot value of the spline curve (13.5 mm for AVPD, 8.8 cm/s for s0 ). Because b1 and b3 were negative in all models, to facilitate readability the coefficients were described as b1 and b3 in the first column, and their absolute values were reported in the table. *Model 1: adjusted for clinical and echocardiographic variables as in model 1 of Table 2. y Model 2: for AVPD and s0 adjusted as in model 1 þ circumferential indexes; for mFS and SCmFS adjusted as in model 1 þ longitudinal indexes. z P value expressing significance for R2 change from model 1 to model 2, calculated with the use of the likelihood ratio test.

118 Journal of Cardiac Failure Vol. 20 No. 2 February 2014 Table 5. Multivariable Nonlinear Regression Analysis Showing the Associations of Global Longitudinal Strain (GLS) and Global Circumferential Strain (GCS) With Peak Early Diastolic Mitral Annulus Velocity (e0 ) and With the Ratio of Early Diastolic Mitral Flow to Mitral Annulus Velocity (E/e0 ) in Unadjusted Models and After Controlling for Confounders Unadjusted

Prediction of e0 GLS GCS

b

95% CI

0.208e0.320 R2 5 0.301 0.021e0.145 R2 5 0.032

!.0001

0.255

.0084

0.068

0.202e0.308 R2 5 0.415 0.005e0.141 R2 5 0.276

15.1e25.5 0.57e2.54 0.010e0.105 0.001e0.003 R2 5 0.272 0.013e0.131 R2 5 0.011

!.0001 !.0001 .0017 .016

22.3 1.52 0.047 0.002

.46

0.51

95% CI

0.264 0.083

Prediction of E/e0 GLS Intercept 20.3 1.56 b1 0.058 b2 0.002 b3 GCS

P Value

b

0.59

Model 2y

Model 1*

16.3e28.2 0.59e2.45 0.008e0.086 0.001e0.003 R2 5 0.345 0.020e0.122 R2 5 0.153

P Value

b

95% CI

P Value

!.0001

0.248

!.0001

.075

0.023

0.195e0.303 R2 5 0.431 0.055e0.101 R2 5 0.431

!.0001 !.0001 .0019 .022

22.5 1.47 0.046 0.002

!.0001 !.0001 .0073 .027

.65

0.42

16.1e28.9 0.49e2.51 0.006e0.086 0.001e0.003 R2 5 0.356 0.043e0.127 R2 5 0.356

.74

.83

The analysis was performed in a subset of 210 patients with the use of the same models identified by univariable regression. Thus, a cubic spline regression model was used for the prediction of E/e0 using GLS, whereas linear models were used in all other cases. The coefficients b1, b2, and b3 refer to the linear, quadratic, and cubic terms, respectively, of the cubic spline function. The curve included no knots, so that the b4 coefficient was not computed. Because b1 and b3 were negative in all models, to facilitate readability the coefficients were described as b1 and b3 in the first column, and their absolute values were reported in the table. *Model 1: adjusted for clinical and echocardiographic variables as in model 1 of Table 2. y Model 2: adjusted as in model 1 þ both GLS and GCS.

the impact of LV geometric pattern on LV diastolic function in asymptomatic hypertensive patients could be particularly evident among those with preserved longitudinal LV function, ie, those with an AVPD O12 mm or a s0 O8 cm/s. Because of the expected strict association of LV pattern with diastolic dysfunction, careful assessment of LV mass and geometry might therefore be considered of particular importance in this subset. The results of this observational cross-sectional study can not allow drawing any conclusion about changes in diastolic function over time. However, considering the increased risk of abnormal LV diastolic function associated with impaired longitudinal systolic indexes, a strict clinical and echocardiographic follow-up might be reasonable in hypertensive patients with longitudinal systolic dysfunction. Moreover, future studies should evaluate whether impaired systolic longitudinal function could be useful in selecting patients with exertional dyspnea for a stress test aimed at assessing diastolic functional reserve.42,43 Some limitations in this study should be taken into account. Our population included asymptomatic and relatively young subjects with uncomplicated hypertension, so that caution is needed in generalizing the findings. We did not assess any component of arterial load, but the potential confounding effects of arterial stiffness, wave reflections, and the augmentation index on LV systolic and diastolic function should be considered.44,45 Natriuretic peptide plasma concentrations were not assessed. Because natriuretic peptide levels are increased in hypertensive patients with diastolic dysfunction,46 an analysis aimed at exploring the interaction among different components of

LV systolic shortening, LV diastolic indexes, and natriuretic peptide levels in patients with hypertension may be of clinical interest. By design, our study population included asymptomatic subjects in New York Heart Association functional class I. As a result, we were not able to investigate the relationships of systolic and diastolic indexes with functional impairment in these patients. Speckle tracking data were available only in a subset of our population. Also, the use of 3-dimensional speckle tracking or magnetic resonance imaging for the analysis of the relationship between systolic and diastolic function could have provided further information. Finally, future studies comparing the association of longitudinal and circumferential indexes with invasively determined diastolic measures would provide more solid data to explore these relationships. Disclosures None.

Supplementary Data Supplementary data related to this article can be found online at http://dx.doi.org/10.1016/j.cardfail.2013.12.009. References 1. Koulouris SN, Kostopoulos KG, Triantafyllou KA, Karabinos I, Bouki TP, Karvounis HI, et al. Impaired systolic dysfunction of left

Systolic vs Diastolic Dysfunction in Hypertension

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

ventricular longitudinal fibers: a sign of early hypertensive cardiomyopathy. Clin Cardiol 2005;28:282e6. Ballo P, Mondillo S, Guerrini F, Barbati R, Picchi A, Focardi M. Midwall mechanics in physiologic and hypertensive concentric hypertrophy. J Am Soc Echocardiogr 2004;17:418e27. Adebayo AK, Oladapo OO, Adebiyi AA, Ogunleye OO, Ogah OS, Ojji DB, et al. Characterisation of left ventricular function by tissue Doppler imaging technique in newly diagnosed, untreated hypertensive subjects. Cardiovasc J Afr 2008;19:259e63. Palmieri V, Bella JN, DeQuattro V, Roman MJ, Hahn RT, Dahlof B, et al. Relations of diastolic left ventricular filling to systolic chamber and myocardial contractility in hypertensive patients with left ventricular hypertrophy (the PRESERVE study). Am J Cardiol 1999;84:558e62. Wachtell K, Papademetriou V, Smith G, Gerdts E, Dahl€of B, Engblom E, et al. Relation of impaired left ventricular filling to systolic midwall mechanics in hypertensive patients with normal left ventricular systolic chamber function: the Losartan Intervention for Endpoint Reduction in Hypertension (LIFE) study. Am Heart J 2004;148:538e44. de Simone G, Greco R, Mureddu G, Romano C, Guida R, Celentano A, et al. Relation of left ventricular diastolic properties to systolic function in arterial hypertension. Circulation 2000;101:152e7. Sciarretta S, Paneni F, Ciavarella GM, de Biase L, Palano F, Baldini R, et al. Evaluation of systolic properties in hypertensive patients with different degrees of diastolic dysfunction and normal ejection fraction. Am J Hypertens 2009;22:437e43. Dini FL, Galderisi M, Nistri S, Buralli S, Ballo P, Mele D, et al. SPHERE Hypertension Collaborators of the Working Group on Echocardiography of the Italian Society of Cardiology. Abnormal left ventricular longitudinal function assessed by echocardiographic and tissue Doppler imaging is a powerful predictor of diastolic dysfunction in hypertensive patients: the SPHERE study. Int J Cardiol 2013;168:3351e8. Kono M, Kisanuki A, Ueya N, Kubota K, Kuwahara E, Takasaki K, et al. Left ventricular global systolic dysfunction has a significant role in the development of diastolic heart failure in patients with systemic hypertension. Hypertens Res 2010;33:1167e73. Galderisi M, Lomoriello VS, Santoro A, Esposito R, Olibet M, Raia R, et al. Differences of myocardial systolic deformation and correlates of diastolic function in competitive rowers and young hypertensives: a speckle-tracking echocardiography study. J Am Soc Echocardiogr 2010;23:1190e8. Ballo P, Quatrini I, Giacomin E, Motto A, Mondillo S. Circumferential versus longitudinal systolic function in patients with hypertension: a nonlinear relation. J Am Soc Echocardiogr 2007;20:298e306. Mancia G, de Backer G, Dominiczak A, Cifkova R, Fagard R, German o G, et al. 2007 guidelines for the management of arterial hypertension: the Task Force for the Management of Arterial Hypertension of the European Society of Hypertension (ESH) and of the European Society of Cardiology (ESC). J Hypertens 2007;25:1105e87. Recommendations for chamber quantification: a report from the American Society of Echocardiography’s Guidelines and Standards Committee and the Chamber Quantification Writing Group, developed in conjunction with the European Association of Echocardiography, a branch of the European Society of Cardiology. J Am Soc Echocardiogr 2005;18:1440e63. Ballo P, Zaca V, Giacomin E, Galderisi M, Mondillo S. Impact of obesity on left ventricular systolic function in hypertensive subjects with normal ejection fraction. Int J Cardiol 2010;141:316e20. Gaasch WH, Zile MR, Hoshino PK, Apstein CS, Blaustein AS. Stressshortening relations and myocardial blood flow in compensated and failing canine hearts with pressure-overload hypertrophy. Circulation 1989;79:872e83. Lisi M, Ballo P, Cameli M, Gandolfo F, Galderisi M, Chiavarelli M, et al. Mitral annular longitudinal function preservation after mitral valve repair: the MARTE study. Int J Cardiol 2012;157:212e5. Palmieri V, Wachtell K, Gerdts E, Bella JN, Papademetriou V, Tuxen C, et al. Left ventricular function and hemodynamic features

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.



Ballo et al

119

of inappropriate left ventricular hypertrophy in patients with systemic hypertension: the LIFE study. Am Heart J 2001;141:784e91. Mondillo S, Galderisi M, Ballo P, Marino PN. Left ventricular systolic longitudinal function: comparison among simple M-mode, pulsed and M-mode color-coded Tissue Doppler of mitral annulus in healthy individuals. J Am Soc Echocardiogr 2006;19:1085e91. Elnoamany MF, Abdelhameed AK. Mitral annular motion as a surrogate for left ventricular function: correlation with brain natriuretic peptide levels. Eur J Echocardiogr 2006;7:187e98. Mondillo S, Ballo P, Galderisi M, Focardi M, Giacomin E, Maffei S, et al. Assessment of left ventricular diastolic events interrelations: an integrated approach. Int J Cardiol 2010;145:426e31. Deswal A, Richardson P, Bozkurt B, Mann DL. Results of the Randomized Aldosterone Antagonism in Heart Failure with Preserved Ejection Fraction trial (RAAM-PEF). J Card Fail 2011;17:634e42. Nagueh SF, Appleton CP, Gillebert TC, Marino PN, Oh JK, Smiseth OA, et al. Recommendations for the evaluation of left ventricular diastolic function by echocardiography. Eur J Echocardiogr 2009;10:165e93. Iwano H, Yamada S, Watanabe M, Mitsuyama H, Nishino H, Yokoyama S, et al. Novel strain rate index of contractility loss caused by mechanical dyssynchronyda predictor of response to cardiac resynchronization therapy. Circ J 2011;75:2167e75. Kalogeropoulos AP, Georgiopoulou VV, Gheorghiade M, Butler J. Echocardiographic evaluation of left ventricular structure and function: new modalities and potential applications in clinical trials. J Card Fail 2012;18:159e72. Galderisi M, Nistri S, Mondillo S, Losi MA, Innelli P, Mele D, et al. Working Group of Echocardiography, Italian Society of Cardiology. Methodological approach for the assessment of ultrasound reproducibility of cardiac structure and function: a proposal of the study group of Echocardiography of the Italian Society of Cardiology (Ultra Cardia SIC) part I. Cardiovasc Ultrasound 2011;9:26. Wang YC, Yu CC, Chiu FC, Klepfer R, Hilpisch K, Splett V, et al. Impact of ventricular dyssynchrony on postexercise accommodation of systolic myocardial motion in hypertensive patients with heart failure and a normal ejection fraction: a tissue-Doppler echocardiography study. J Card Fail 2012;18:134e9. Liu YW, Tsai WC, Su CT, Lin CC, Chen JH. Evidence of left ventricular systolic dysfunction detected by automated function imaging in patients with heart failure and preserved left ventricular ejection fraction. J Card Fail 2009;15:782e9. Plaksej R, Kosmala W, Frantz S, Herrmann S, Niemann M, St€ork S, et al. Relation of circulating markers of fibrosis and progression of left and right ventricular dysfunction in hypertensive patients with heart failure. J Hypertens 2009;27:2483e91. Bernardo BC, Weeks KL, Pretorius L, McMullen JR. Molecular distinction between physiological and pathological cardiac hypertrophy: experimental findings and therapeutic strategies. Pharmacol Ther 2010;128:191e227. Borlaug BA, Lam CS, Roger VL, Rodeheffer RJ, Redfield MM. Contractility and ventricular systolic stiffening in hypertensive heart disease insights into the pathogenesis of heart failure with preserved ejection fraction. J Am Coll Cardiol 2009;54:410e8. Ballo P, Barone D, Bocelli A, Motto A, Mondillo S. Left ventricular longitudinal systolic dysfunction is an independent marker of cardiovascular risk in patients with hypertension. Am J Hypertens 2008; 21:1047e54. Kim SH, Cho GY, Baik I, Lim SY, Choi CU, Lim HE, et al. Early abnormalities of cardiovascular structure and function in middle-aged korean adults with prehypertension: the Korean Genome Epidemiology study. Am J Hypertens 2011;24:218e24. di Bello V, Talini E, dell’Omo G, Giannini C, delle Donne MG, Canale ML, et al. Early left ventricular mechanics abnormalities in prehypertension: a two-dimensional strain echocardiography study. Am J Hypertens 2010;23:405e12. Kosmala W, Plaksej R, Strotmann JM, Weigel C, Herrmann S, Niemann M, et al. Progression of left ventricular functional abnormalities

120 Journal of Cardiac Failure Vol. 20 No. 2 February 2014

35.

36.

37. 38.

39.

40.

in hypertensive patients with heart failure: an ultrasonic two-dimensional speckle tracking study. J Am Soc Echocardiogr 2008;21:1309e17. Sharma GV, Woods PA, Lindsey N, O’Connell C, Connolly L, Joseph J, et al. Noninvasive monitoring of left ventricular enddiastolic pressure reduces rehospitalization rates in patients hospitalized for heart failure: a randomized controlled trial. J Card Fail 2011;17:718e25. Fukuta H, Ohte N, Wakami K, Goto T, Tani T, Kimura G. Prognostic value of left ventricular diastolic dysfunction in patients undergoing cardiac catheterization for coronary artery disease. Cardiol Res Pract 2012;2012:243735. Sherazi S, Zare˛ba W. Diastolic heart failure: predictors of mortality. Cardiol J 2011;18:222e32. Schulmeyer MC, Arriaza N. Good prognostic value of the intraoperative tissue Dopplerederived index E/e0 after non-cardiac surgery. Minerva Anestesiol 2012;78:1013e8. Sharp AS, Tapp RJ, Thom SA, Francis DP, Hughes AD, Stanton AV, et al. ASCOT Investigators. Tissue Doppler E/e0 ratio is a powerful predictor of primary cardiac events in a hypertensive population: an ASCOT substudy. Eur Heart J 2010;31:747e52. Wang M, Yip GW, Wang AY, Zhang Y, Ho PY, Tse MK, et al. Tissue Doppler imaging provides incremental prognostic value in patients

41.

42.

43.

44.

45.

46.

with systemic hypertension and left ventricular hypertrophy. J Hypertens 2005;23:183e91. Giannotti G, Mondillo S, Galderisi M, Barbati R, Zaca V, Ballo P, et al. Hand-held echocardiography: added value in clinical cardiological assessment. Cardiovasc Ultrasound 2005;3:7. Ha JW, Choi D, Park S, Choi EY, Shim CY, Kim JM, et al. Left ventricular diastolic functional reserve during exercise in patients with impaired myocardial relaxation at rest. Heart 2009;95:399e404. Gharacholou SM, Scott CG, Borlaug BA, Kane GC, McCully RB, Oh JK, et al. Relationship between diastolic function and heart rate recovery after symptom-limited exercise. J Card Fail 2012;18:34e40. Fukuta H, Ohte N, Wakami K, Asada K, Goto T, Mukai S, et al. Impact of arterial load on left ventricular diastolic function in patients undergoing cardiac catheterization for coronary artery disease. Circ J 2010;74:1900e5. Weber T, O’Rourke MF, Ammer M, Kvas E, Punzengruber C, Eber B. Arterial stiffness and arterial wave reflections are associated with systolic and diastolic function in patients with normal ejection fraction. Am J Hypertens 2008;21:1194e202. Chatzis D, Tsioufis C, Tsiachris D, Taxiarchou E, Lalos S, Kyriakides Z, et al. Brain natriuretic peptide as an integrator of cardiovascular stiffening in hypertension. Int J Cardiol 2010;141:291e6.