View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector

Journal of the American College of Cardiology Vol. 38, No. 2, 2001 © 2001 by the American College of Cardiology ISSN 0735-1097/01/$20.00 Published by Elsevier Science Inc. PII S0735-1097(01)01378-X Relationship of the Electrocardiographic Strain Pattern to Left Ventricular Structure and Function in Hypertensive Patients: The LIFE Study Peter M. Okin, MD, FACC,* Richard B. Devereux, MD, FACC,* Markku S. Nieminen, MD, PHD, FACC,† Sverker Jern, MD,‡ Lasse Oikarinen, MD, PHD,† Matti Viitasalo, MD, PHD,† Lauri Toivonen, MD, PHD,† Sverre E. Kjeldsen, MD, PHD,§ Stevo Julius, MD, SCD, FACC,࿣ Bjo¨rn Dahlo¨f, MD, PHD,‡ for the LIFE Study Investigators New York, New York; Helsinki, Finland; Go¨teborg, Sweden; Oslo, Norway; and Ann Arbor, Michigan

OBJECTIVES This study was designed to assess the relation of electrocardiographic (ECG) strain to increased left ventricular (LV) mass, independent of its relation to coronary heart disease (CHD). BACKGROUND The classic ECG strain pattern, ST depression and T-wave inversion, is a marker for left (LVH) and adverse prognosis. However, the independence of the relation of strain to increased LV mass from its relation to CHD has not been extensively examined. METHODS Electrocardiograms and echocardiograms were examined at study baseline in 886 hyperten- sive patients with ECG LVH by Cornell voltage-duration product and/or Sokolow-Lyon voltage enrolled in the Losartan Intervention For End point (LIFE) echocardiographic substudy. Strain was defined as a downsloping convex ST segment with inverted asymmetrical T-wave opposite to the QRS axis in leads V5 and/or V6. RESULTS Strain occurred in 15% of patients, more commonly in patients with than without evident CHD (29%, 51/175 vs. 11%, 81/711, p Ͻ 0.001). When differences in gender, race, diabetes, systolic pressure, serum creatinine and high density lipoprotein cholesterol were controlled, strain on baseline ECG was associated with greater indexed LV mass in patients with (152 Ϯ 33 vs. 131 Ϯ 32 g/m2,pϽ 0.001) or without CHD (131 Ϯ 24 vs. 119 Ϯ 22 g/m2,pϽ 0.001). In logistic regression analyses, strain was associated with an increased risk of anatomic LVH in patients with CHD (relative risk 5.14, 95% confidence interval [CI] 1.16 to 22.85, p ϭ 0.0315), without evident CHD (relative risk 2.91, 95% CI 1.50 to 5.65, p ϭ 0.0016), and in the overall population when CHD was taken into account (relative risk 2.98, 95% CI 1.65 to 5.38, p ϭ 0.0003). CONCLUSIONS When clinical evidence of CHD is accounted for, ECG strain is likely to indicate the presence of anatomic LVH. Greater LV mass and higher prevalence of LVH in patients with strain offer insights into the known association of the strain pattern with adverse outcomes. (J Am Coll Cardiol 2001;38:514–20) © 2001 by the American College of Cardiology

The classic strain pattern of ST depression and T-wave ship of ECG strain to LV mass, as opposed to its potential inversion on the rest electrocardiogram (ECG) is a well- relation to CHD, has not been extensively examined. Thus, recognized marker of the presence of anatomic left ventric- the present study examined the relation of the strain pattern ular hypertrophy (LVH) (1–6). This abnormality of repo- on the rest ECG to LV mass and the presence of anatomic larization has been associated with an adverse prognosis in a LVH, taking into account potential effects of coexistent variety of clinical populations (7–11) and implicated as the CHD and controlling for other clinical and demographic strongest marker of untoward outcomes when ECG LVH variables that could potentially affect these relationships. criteria have been utilized for risk stratification (8–10). Moreover, the strain pattern could also reflect underlying METHODS coronary heart disease (CHD), a relationship that could in part explain the clinical consequences of this ECG finding Subjects. The Losartan Intervention For End point beyond those directly attributable to high left ventricular (LIFE) study (12,13) enrolled hypertensive patients with (LV) mass (3,4,8,10). However, the independent relation- ECG LVH by Cornell voltage-duration product (14,15) and/or Sokolow-Lyon voltage (1) on a screening ECG in a From the *Department of Medicine, Division of Cardiology, Cornell University prospective double-blind study to determine whether appre- Medical Center, New York, New York; †Division of Cardiology, Department of ciable reduction in mortality and morbid events is associated Medicine, Helsinki University Central Hospital, Helsinki, Finland; ‡Sahlgrenska University Hospital/O¨ stra, Go¨teborg, Sweden; §Ullevål University Hospital, Oslo, with the use of losartan as opposed to atenolol (12). Eligible Norway; and the ࿣University of Michigan Medical Center, Ann Arbor, Michigan. patients for LIFE were hypertensive men and women aged Supported in part by grant COZ-368 from Merck and Co., Inc., West Point, 55 to 80 years with a seated blood pressure of 160 to 200/95 Pennsylvania. Manuscript received November 10, 2000; revised manuscript received March 1, to 115 mm Hg after one and two weeks on placebo. Study 2001, accepted April 10, 2001. inclusion and exclusion criteria have been previously pub- JACC Vol. 38, No. 2, 2001 Okin et al. 515 August 2001:514–20 Strain and Hypertrophy

LV mass index was Ͼ104 g/m2 in women or Ͼ116 g/m2 Abbreviations and Acronyms in men. Hypertrophy was considered concentric if LV ASE ϭ American Society of Echocardiography relative wall thickness was Ͼ0.430 and eccentric if relative CHD ϭ coronary heart disease wall thickness was normal (16); patients with normal LV ECG ϭ electrocardiogram, electrocardiographic ϭ mass were considered to have normal LV geometry if ESS end-systolic stress Յ HDL ϭ high density lipoprotein relative wall thickness was 0.43 or to have concentric LIFE ϭ Losartan Intervention For End point study remodeling if relative wall thickness was increased. Alter- LV ϭ left ventricular native analyses used LV mass/height2.7 partition values of LVH ϭ left ventricular hypertrophy 46.7 and 49.2 g/m2.7 in women and men, respectively. Myocardial contractile performance was assessed by ex- amining LV systolic shortening in relation to circumferen- lished (12,13). A previously described (16) representative tial end-systolic stress (ESS) (26). The relation of LV sample of the whole LIFE study, totaling 964 patients, midwall shortening to midwall circumferential ESS at the underwent baseline echocardiograms; 21 patients had un- LV minor axis was calculated as previously described (27), measurable echocardiographic LV mass and 57 had incom- and was expressed as a percent of the value predicted from plete baseline ECG measurements. There were 361 eligible circumferential ESS by an equation derived in apparently women and 525 eligible men whose mean age was 66 Ϯ 7 normal adults (28). This variable, stress-corrected midwall years. shortening was considered low if Ͻ89.2%, the 5th percentile . All ECGs were interpreted at the in a separate reference population of 280 normal adults (28). Core Laboratory at Sahlgrenska University Hospital/O¨ stra To estimate myocardial oxygen demand, the ESS-LV in Go¨teborg, Sweden, by investigators blinded to the mass-heart rate product was calculated as previously de- clinical information (12,13). The product of QRS dura- scribed (29). Fractional and midwall shortening could be ϫ ϩ tion Cornell voltage (RaVL SV3, with 8 mm added in determined in 100% of patients, whereas circumferential women [14,15]) was used with a threshold value of ESS, stress-corrected midwall shortening and ESS-LV 2,440 mm ⅐ ms to identify LVH. After design of LIFE, mass-heart rate product could be determined in 97% (857/ studies were published suggesting a smaller gender adjust- 886) of patients in this study. ment (6,17), and feedback from LIFE investigators showed Statistics. Patients were classified as having CHD if they that otherwise eligible patients had ECG LVH by highly had self-reported or , diagnos- specific but insensitive Sokolow-Lyon voltage criteria (1), tic Q-waves by Minnesota code on the ECG or segmental but not by Cornell product. Accordingly, changes were wall motion abnormalities on the two-dimensional echocar- made in ECG entry criteria: the gender adjustment of diogram. Differences in prevalence were compared using ␹2 Cornell voltage was reduced from 8 to 6 mm and Sokolow- analyses and mean values of continuous variables by un- ϩ Ͼ Lyon voltage (SV1 RV5/6) 38 mm was accepted for paired t test. Echocardiographic variables were further ECG eligibility (13). Additional ECG measurements were compared using analysis of covariance to adjust for differ- performed at Helsinki University Central Hospital. Repo- ences in age, gender, race, systolic pressure, cholesterol and larization abnormalities in leads V5 and/or V6 indicated high density lipoprotein (HDL) cholesterol, creatinine and typical strain (11) when there was a downsloping convex ST prevalent diabetes between groups. Independent relations of segment with an inverted asymmetrical T-wave opposite to echocardiographic LVH, abnormal LV geometry and ab- the QRS axis. normal stress-corrected midwall shortening to ECG strain Echocardiography. Studies were performed using funda- were determined using stepwise logistic regression analyses mental imaging with commercially available phased-array and the same covariates. A two-tailed p Ͻ 0.05 was required echocardiographs as previously described (16,18). Left ven- for statistical significance. tricular internal dimension and wall thicknesses were mea- sured at end-diastole by American Society of Echocardiog- RESULTS raphy (ASE) recommendations (19) on up to three cardiac cycles. When M-mode recordings could not be optimally Patient characteristics. Strain was present in 132 patients oriented, correctly oriented linear dimension measurements (15%) and was more common in patients with CHD than in were made using two-dimensional imaging by leading-edge those without (29%, 51/175 vs. 11%, 81/711, p Ͻ 0.001). ASE convention (20). Methods for measurement and evi- Clinical and demographic characteristics of patients with or dence of interchangeability of two-dimensional and without strain on the ECG, stratified by CHD status, are M-mode LV dimensions have been described in detail (18). shown in Table 1. Among patients without CHD, those Relative wall thickness was calculated as end-diastolic pos- with typical strain were more likely to be men and African- terior wall thickness/LV radius; LV mass was calculated American with higher systolic pressures, lower HDL cho- (21) and was indexed for body surface area and alternatively lesterol and higher serum creatinine levels than patients for height2.7 (22). Under criteria known to predict an without strain. Among patients with evident CHD, those adverse prognosis (23–25), LVH was considered present if with strain similarly were more likely to be African- 516 Okin et al. JACC Vol. 38, No. 2, 2001 Strain and Hypertrophy August 2001:514–20

Table 1. Demographic and Clinical Characteristics in Relation to Electrocardiographic Strain in Hypertensive Patients With and Without Evident CHD No CHD CHD Strain؊ Strain؉ Strain؊ Strain؉ p Value (51 ؍ n) (124 ؍ p Value (n (81 ؍ n) (630 ؍ Variable (n Age (yrs) 65.8 Ϯ 7.1 66.6 Ϯ 5.8 0.239 67.2 Ϯ 7.0 68.8 Ϯ 6.7 0.178 Gender (% male) 55.6 71.6 0.009 65.3 70.6 0.597 Race (% African-American) 11.7 30.9 Ͻ 0.001 8.9 27.5 0.003 Diabetes (%) 8.4 8.6 1.000 13.7 31.4 0.010 Systolic blood pressure (mm Hg) 172.4 Ϯ 14.2 178.1 Ϯ 12.9 0.001 174.4 Ϯ 15.6 175.3 Ϯ 13.7 0.721 Diastolic blood pressure (mm Hg) 98.7 Ϯ 8.3 100.4 Ϯ 8.8 0.093 96.8 Ϯ 9.7 94.9 Ϯ 11.7 0.265 Body mass index (kg/m2) 27.3 Ϯ 4.3 26.9 Ϯ 4.5 0.386 27.6 Ϯ 5.8 27.4 Ϯ 6.0 0.820 Total cholesterol (mM) 6.06 Ϯ 1.12 5.88 Ϯ 1.14 0.195 5.65 Ϯ 1.05 5.60 Ϯ 1.12 0.781 HDL cholesterol (mM) 1.56 Ϯ 0.47 1.44 Ϯ 0.46 0.042 1.45 Ϯ 0.37 1.33 Ϯ 0.37 0.062 Creatinine (mM) 88.0 Ϯ 22.2 97.2 Ϯ 24.3 0.002 93.6 Ϯ 20.3 103.3 Ϯ 22.8 0.007 Serum sodium (mM/l) 140.2 Ϯ 2.3 140.4 Ϯ 2.4 0.551 139.9 Ϯ 2.8 139.8 Ϯ 2.7 0.774 Serum potassium (mM/l) 4.16 Ϯ 0.42 4.06 Ϯ 0.37 0.058 4.15 Ϯ 0.39 4.19 Ϯ 0.35 0.511

CHD ϭ coronary heart disease; HDL ϭ high density lipoprotein.

American, have higher serum creatinine, a slightly lower dium, as measured by fractional shortening, and with similar HDL cholesterol, and a greater prevalence of diabetes, but levels of circumferential ESS, but patients with strain had were similar to patients without strain with regard to lower LV midwall performance than predicted for the level gender, systolic pressure and serum sodium and potassium of ESS, as reflected by lower stress-corrected midwall levels. shortening. Electrocardiographic strain was associated with Echocardiographic findings in relation to ECG strain. increased wall mass-stress-heart rate product only among Relations of LV structure and function to the presence or patients without CHD. absence of ECG strain are examined in Table 2. In patients Prevalences of echocardiographic LVH, abnormal LV with or without CHD, strain on the ECG was associated geometry and LV midwall function in patients with or with greater LV wall thicknesses, LV mass and LV mass without ECG strain are shown in Table 3. In this popula- indexed for either body surface area or for height2.7. Among tion selected on the basis of ECG LVH by Cornell product patients without CHD, LV internal dimensions were also and/or Sokolow-Lyon voltage, strain was associated with a greater in patients with strain. In both patient groups, higher prevalence of LVH defined by LV mass/body surface because strain was associated with a greater increase in LV area in patients with and without CHD. Alternative anal- wall thickness than cavity dimension, strain was also asso- yses using gender-specific LV mass/height2.7 partitions ciated with increased relative wall thickness. Strain was confirmed higher prevalences of LVH in patients with associated with similar LV performance at the endocar- ECG strain, both in those with CHD (92% vs. 84%, p ϭ

Table 2. Echocardiographic Findings in Relation to Electrocardiographic Strain in Hypertensive Patients With and Without Evident CHD No CHD CHD Strain؊ Strain؉ Strain؊ Strain؉ p Value (51 ؍ n) (124 ؍ p Value (n (81 ؍ n) (630 ؍ Variables (n LVIDD (cm) 5.20 Ϯ 0.53 5.34 Ϯ 0.59 0.035 5.56 Ϯ 0.65 5.62 Ϯ 0.64 0.533 IVSD (cm) 1.14 Ϯ 0.14 1.24 Ϯ 0.19 Ͻ 0.001 1.14 Ϯ 0.14 1.25 Ϯ 0.19 Ͻ 0.001 PWTD (cm) 1.06 Ϯ 0.11 1.14 Ϯ 0.15 Ͻ 0.001 1.06 Ϯ 0.10 1.15 Ϯ 0.17 Ͻ 0.001 Relative wall thickness 0.41 Ϯ 0.06 0.43 Ϯ 0.09 0.007 0.39 Ϯ 0.06 0.41 Ϯ 0.08 0.017 LV mass (g) 223 Ϯ 48 259 Ϯ 68 Ͻ 0.001 249 Ϯ 59 286 Ϯ 71 Ͻ 0.001 LV mass index* (g/m2) 118 Ϯ 22 134 Ϯ 29 Ͻ 0.001 131 Ϯ 27 151 Ϯ 34 Ͻ 0.001 LV mass index† (g/m2.7) 54 Ϯ 13 59 Ϯ 14 0.002 60 Ϯ 18 69 Ϯ 19 0.005 Fractional shortening (%) 34 Ϯ 534Ϯ 6 0.487 30 Ϯ 730Ϯ 8 0.781 Midwall shortening (%) 15.8 Ϯ 2.0 15.0 Ϯ 1.9 0.001 14.8 Ϯ 2.1 13.9 Ϯ 2.5 0.029 Circumferential end-systolic stress 176 Ϯ 41 183 Ϯ 54 0.255 211 Ϯ 59 207 Ϯ 70 0.750 (kdynes/cm2) Stress-corrected midwall shortening 98 Ϯ 13 95 Ϯ 12 0.046 96 Ϯ 12 90 Ϯ 16 0.006 (% predicted) Stress-mass-HR product 2.72 Ϯ 1.13 3.21 Ϯ 1.48 0.001 3.77 Ϯ 1.95 4.08 Ϯ 1.76 0.340 (kdynes⅐g/min⅐cm2 ϫ 106)

*Indexed to body surface area; †indexed to height2.7. CHD ϭ coronary heart disease; HR ϭ heart rate; IVSD ϭ intraventricular septal thickness in diastole; LV ϭ left ventricular; LVIDD ϭ LV internal dimension in diastole; PWTD ϭ posterior wall thickness in diastole. JACC Vol. 38, No. 2, 2001 Okin et al. 517 August 2001:514–20 Strain and Hypertrophy

Table 3. Prevalence of Echocardiographic LVH, Abnormal LV Geometry and Abnormal LV Midwall Function in Relation to Electrocardiographic Strain in Hypertensive Patients With and Without Evident CHD No CHD CHD Strain؊ Strain؉ Strain؊ Strain؉ p Value (51 ؍ n) (124 ؍ p Value (n (81 ؍ n) (630 ؍ Variables (n LVH (%)* 64.0 81.5 0.003 83.1 96.1 0.025 LV geometry (%) 0.003 0.011 Normal (n ϭ 155) 22.8 13.6 11.3 2.0 Concentric remodeling (n ϭ 88) 13.2 4.9 5.6 2.0 Eccentric hypertrophy (n ϭ 307) 43.0 45.7 62.9 54.9 Concentric hypertrophy (n ϭ 161) 21.0 35.8 20.2 41.2 Stress-corrected midwall shortening Ͻ89.2% (%) 24.0 27.8 0.471 26.1 49.0 0.006

*Left ventricular mass Ͼ104 g/m2 in women and Ͼ116 g/m2 in men. CHD ϭ coronary heart disease; LV ϭ left ventricular; LVH ϭ left ventricular hypertrophy.

0.226) and without CHD (82% vs. 70%, p ϭ 0.039). strain, but midwall shortening remained moderately de- Independent of the presence or absence of evident CHD, pressed in those with strain. Similarly, after taking into patients with strain were less likely than those without strain account differences in systolic pressure and normalizing for to have normal LV geometry or concentric remodeling and differences in LV dimension between groups, ESS was had a similar prevalence of eccentric LVH, but were much similar in patients with and without strain, but stress- more likely to have concentric LVH than patients without corrected midwall shortening remained lower in patients strain on the rest ECG. Depressed LV contractility, as with strain, attaining statistical significance in patients with estimated by stress-corrected midwall shortening Ͻ89.2%, CHD. Wall stress-mass-heart rate product was elevated in was more common in CHD patients with strain, but did not patients without evident CHD after adjusting for baseline differ in prevalence in patients without CHD according to differences. the presence or absence of strain. In parallel fashion, relations of echocardiographic LVH, Because patients with and without strain differed with abnormal LV geometry and depressed LV contractility to respect to demographic and clinical variables that could strain were further examined using separate logistic regres- affect LV structure and function (Table 1), independent sion analyses in patients with and without CHD (Table 5). relations of echocardiographic findings to the presence of After adjusting for baseline differences, ECG strain was strain were examined in patients with and without CHD by associated with a nearly threefold greater risk of LVH in analysis of covariance (Table 4). After adjusting for baseline patients without CHD and a greater than fivefold increased differences, strain remained strongly associated with in- risk of echocardiographic LVH in patients with CHD. In creased LV wall thicknesses, relative wall thickness, LV nominal regression analyses where the risk of abnormal LV mass and LV mass indices in patients with and without geometry compared with normal geometry was determined CHD. Left ventricular internal dimensions, fractional for the presence of strain, ECG strain was most strongly shortening and ESS were similar in patients with or without related to increased risk of concentric LVH, was strongly

Table 4. Echocardiographic Findings in Relation to Electrocardiographic Strain in Hypertensive Patients With and Without Evident CHD Adjusted for Differences in Clinical and Demographic Variables* No CHD CHD Strain؊ Strain؉ Strain؊ Strain؉ p Value (51 ؍ n) (124 ؍ p Value (n (81 ؍ n) (630 ؍ Variables (n LVIDD (cm) 5.21 Ϯ 0.53 5.26 Ϯ 0.57 0.401 5.56 Ϯ 0.67 5.61 Ϯ 0.70 0.672 IVSD (cm) 1.14 Ϯ 0.15 1.22 Ϯ 0.17 Ͻ 0.001 1.15 Ϯ 0.17 1.25 Ϯ 0.17 Ͻ 0.001 PWTD (cm) 1.06 Ϯ 0.13 1.13 Ϯ 0.13 Ͻ 0.001 1.06 Ϯ 0.13 1.16 Ϯ 0.13 Ͻ 0.001 Relative wall thickness 0.41 Ϯ 0.08 0.43 Ϯ 0.07 0.014 0.39 Ϯ 0.07 0.42 Ϯ 0.07 0.009 LV mass (g) 224 Ϯ 46 249 Ϯ 50 Ͻ 0.001 250 Ϯ 65 286 Ϯ 68 0.002 LV mass index* (g/m2) 119 Ϯ 22 131 Ϯ 24 Ͻ 0.001 131 Ϯ 32 152 Ϯ 33 Ͻ 0.001 LV mass index (g/m2.7) 55 Ϯ 13 58 Ϯ 14 0.040 59 Ϯ 16 70 Ϯ 17 Ͻ 0.001 Fractional shortening (%) 34 Ϯ 535Ϯ 5 0.736 30 Ϯ 830Ϯ 7 0.945 Midwall shortening (%) 15.8 Ϯ 2.5 15.3 Ϯ 1.8 0.060 14.6 Ϯ 2.2 13.9 Ϯ 2.2 0.083 Circumferential end-systolic stress 176 Ϯ 43 176 Ϯ 47 0.933 212 Ϯ 66 204 Ϯ 69 0.476 (kdynes/cm2) Stress-corrected midwall shortening 98 Ϯ 13 95 Ϯ 14 0.081 95 Ϯ 13 90 Ϯ 14 0.016 (% predicted) Stress-mass-HR product 2.73 Ϯ 1.11 3.03 Ϯ 1.13 0.040 3.84 Ϯ 1.91 3.99 Ϯ 1.95 0.675 (kdynes⅐g/min⅐cm2 ϫ 106)

*Adjusted for age, gender, race, prevalent diabetes, systolic blood pressure, high density lipoprotein cholesterol and creatinine. Abbreviations as in Table 2. 518 Okin et al. JACC Vol. 38, No. 2, 2001 Strain and Hypertrophy August 2001:514–20

Table 5. Multivariate Logistic Regression Analyses to Assess the Predictive Value of Electrocardiographic Strain for the Presence of LVH, Abnormal LV Geometry and Abnormal LV Midwall Function in Patients With and Without Evident CHD* Variable ␹2 p Value Beta Coefficient Relative Risk 95% CI No Evident CHD LVH 9.91 0.0016 1.067 2.91 1.50–5.65 LV geometry† 14.17 0.003 Concentric remodeled 0.985 0.013 1.01 0.29–3.59 Eccentric hypertrophy 0.030 0.919 2.51 1.09–5.76 Concentric hypertrophy 0.002 1.358 3.89 1.64–9.25 Stress-corrected midwall shortening Ͻ89.2% 0.23 0.630 0.137 1.15 0.66–2.00 CHD LVH 4.63 0.0315 1.637 5.14 1.16–22.85 LV geometry† 10.13 0.016 Concentric remodeled 0.554 0.905 2.47 0.12–49.52 Eccentric hypertrophy 0.124 1.687 5.41 0.63–46.33 Concentric hypertrophy 0.023 2.557 12.89 1.43–116.65 Stress-corrected midwall shortening Ͻ89.2% 6.61 0.010 0.951 2.59 1.25–5.35

*Relative risk of strain compared to absence of strain, adjusted for age, gender, race, prevalent diabetes, systolic blood pressure, high density lipoprotein cholesterol and creatinine; †relative risks calculated for presence of strain and compared with risk of normal geometry. CHD ϭ coronary heart disease; CI ϭ confidence interval; LV ϭ left ventricular; LVH ϭ left ventricular hypertrophy. associated with presence of eccentric LVH and was not patients without clinically evident CHD, but did not exam- related to concentric remodeling in patients with or without ine differences in LV structure and mass in relation to strain. evident CHD. Strain remained associated with decreased In a subset of 40 patients with severe isolated aortic LV contractility, as determined by stress-corrected midwall regurgitation without significant obstructive CHD at an- shortening Ͻ89.2%, in patients with CHD but not in giography, Roman et al. (11) demonstrated that strain was patients without evident CHD. Of note, logistic regression associated with greater LV mass. However, neither study analyses performed in the entire study population using an adjusted for other variables that could affect the relation indicator variable for CHD confirmed these relations, dem- between strain and increased LV mass. The current study onstrating that ECG strain was strongly associated with significantly extends these observations to a large cohort of echocardiographic LVH, especially of concentric pattern, patients with hypertension, demonstrating that strain is and was modestly related to depressed stress-corrected associated with greater LV mass, indexed LV mass and midwall shortening. higher prevalences of echocardiographic LVH in patients with and without CHD, even after taking into account DISCUSSION baseline differences that could contribute to differences in This study demonstrates that typical strain on the ECG in the prevalence of ECG strain. hypertensive patients identifies patients with greater LV Development of typical ST segment and T-wave changes mass, a higher prevalence of echocardiographic LVH that is of the strain pattern in response to hypertrophy is predicted more likely to be concentric, and lower myocardial contrac- by a distributed dipole model of the ECG response to tility and higher estimated myocardial oxygen demand, hypertrophy (30), which demonstrates that T-wave ampli- independent of the presence of clinically evident CHD and tudes are proportional to the square of the cell radius and other demographic and clinical differences between patients that the flattening or inversion of the T-wave observed with with and without strain. These findings offer insights into LVH can be attributed to the contiguity effect of adjacent the known association of the strain pattern with adverse layers having different transmembrane action potential du- outcomes in hypertensive patients and suggest that further rations. These findings suggest that greater degrees of ST evaluation of the diagnostic and prognostic values of quan- depression and T-wave inversion will be associated with titative measures of strain, as reflected by the magnitude of larger increases in LV mass, consistent with the linear ST depression and/or T-wave amplitude, is indicated. relations of these variables to LV mass (31). Relationship of strain to LVH. Previous studies have Relationship of strain to LV geometry. The relationship documented a clear relation between strain and anatomic of strain to LV geometry has been less well studied. In 107 LVH (1–3,5,6) and an increased prevalence of strain with patients not on digitalis (3), strain was more strongly more severe hypertrophy (3). However, most studies did not associated with an LV internal dimension Ͼ55 mm (50% vs. account for potential effects of concomitant CHD, which is 13%, p Ͻ 0.01) than with posterior wall thickness Ͼ12 mm highly covariate with LVH (4,24,25) and could contribute (31% vs. 14%, p ϭ 0.06), suggesting a stronger relationship to the genesis of ST/T-wave changes (3,8,10). Schillaci et between strain and eccentric as opposed to concentric al. (6) demonstrated low sensitivity (16%) and high speci- hypertrophy. Among patients with pure aortic regurgitation ficity (98%) of strain for LVH in 923 white hypertensive (11), typical strain was associated with both increased LV JACC Vol. 38, No. 2, 2001 Okin et al. 519 August 2001:514–20 Strain and Hypertrophy chamber dimension and wall thicknesses, and with an Implications. These findings have important clinical im- increased relative wall thickness that nonetheless remained plications. First, the increased morbidity and mortality within the normal range, such that the overall LV geometric associated with anatomic LVH (24,25), and particularly pattern reflected an eccentric hypertrophic response consis- with concentric hypertrophy (25) and abnormal midwall LV tent with increased LV volume load. The present study mechanics (29), provide new insights into the demonstrated extends these observations to a more heterogeneous popu- prognostic value of strain (8–10). The association of strain lation of hypertensive patients, demonstrating that ECG with male gender and African-American race may in part strain was associated with a greater prevalence of concentric explain the higher event rates in these groups. These as opposed to eccentric LVH due to increased wall thickness findings and the subjective and qualitative nature of strain out of proportion to slightly increased LV chamber size after criteria currently employed provide impetus to develop controlling for other factors (Table 4). quantitative, computerized ST segment and T-wave ampli- Relationship of strain to LV performance. The relation tude measurements to identify LVH and stratify risk. of ECG strain to measures of LV function has not been Further study is needed to determine whether hypertrophy previously examined in hypertensive patients. Devereux and regression is associated with reversal or reduction of repo- Reichek (3) and Roman et al. (11) found the prevalence of larization abnormalities of strain, and whether serial evalu- reduced LV systolic function to be higher in patients with ation of quantitative ST segment and T-wave measure- the strain pattern. Moreover, Roman et al. (11) reported ments will provide additional insight into this process. lower fractional shortening by echocardiography and lower ejection fraction on radionuclide angiography in patients Reprint requests and correspondence: Dr. Peter M. Okin, with strain. However, neither study examined measures of Cornell University Medical Center, 525 East 68th Street, New LV contractility, nor did they adjust for possible confound- York, New York 10021. E-mail: [email protected]. ers by multivariate analyses. The present study extends these observations to hypertensive patients with CHD demon- strating lower myocardial contractility, as estimated by REFERENCES stress-corrected midwall shortening (Table 4) and a more 1. Sokolow M, Lyon TP. The ventricular complex in left ventricular than 2.5-fold greater prevalence of abnormally low stress- hypertrophy as obtained by unipolar precordial and limb leads. Am corrected midwall shortening (Table 5), among CHD Heart J 1949;37:161–86. 2. Carter WA, Estes EH Jr. Electrocardiographic manifestations of patients with strain. However, strain was not associated ventricular hypertrophy: a computer study of ECG-anatomic correla- with depressed midwall shortening in patients without tions in 319 cases. Am Heart J 1964;68:173–82. evident CHD. 3. Devereux RB, Reichek N. abnormalities of left ven- tricular hypertrophy. Clinical, echocardiographic and hemodynamic Mechanisms of abnormal repolarization. The present correlates. J Electrocardiol 1982;14:47–54. study provides additional support for several possible mech- 4. Pringle SD, Macfarlane PW, McKillop JH, Lorimer AR, Dunn FG. anisms for the repolarization abnormalities of the strain Pathophysiologic assessment of left ventricular hypertrophy and strain in asymptomatic patients with essential hypertension. J Am Coll pattern in the absence of CHD. First, the association Cardiol 1989;13:1377–81. between strain and a primarily concentric geometric pattern 5. Levy D, Labib SB, Andersen KM, Christiansen JC, Kannel WB, of LVH, due to greater increases in LV wall thickness than Castelli WP. Determinants of sensitivity and specificity of electrocar- diographic criteria for left ventricular hypertrophy. Circulation 1990; in chamber dimension, suggests a possible primary effect of 81:815–20. myocardial cell hypertrophy, consistent with the model 6. Schillaci G, Verdecchia P, Borgioni C, et al. Improved electrocardio- proposed by Thiry et al. (30). Second, ST depression and graphic diagnosis of echocardiographic left ventricular hypertrophy. Am J Cardiol 1995;74:714–9. T-wave inversion may reflect true subendocardial 7. Sokolow M, Perloff D. The prognosis of essential hypertension treated in the absence of CHD attributable to increased LV mass conservatively. Circulation 1961;33:697–713. and wall thickness (32). This hypothesis is supported by the 8. Verdecchia P, Schillaci G, Borgioni C, et al. Prognostic value of a new electrocardiographic method for diagnosis of left ventricular hypertro- association of strain with increased wall stress-mass-heart phy. J Am Coll Cardiol 1998;31:383–90. rate product among patients without CHD, providing 9. Kannel WB, Gordon T, Offutt D. Left ventricular hypertrophy by evidence of a major demand-side predisposition to myocar- electrocardiogram: prevalence, incidence, and mortality in the Fra- mingham Study. Ann Intern Med 1969;71:89–105. dial ischemia. Indeed, although coronary artery size in- 10. Kannel WB. Prevalence and natural history of electrocardiographic left creases with LV mass and is related to regional myocardial ventricular hypertrophy. Am J Med 1983;75 Suppl 3A:4–11. mass (33), compensatory increases in coronary size, as 11. Roman MJ, Kligfield P, Devereux RB, et al. Geometric and functional correlates of electrocardiographic repolarization and voltage abnormal- measured by lumen area to LV mass ratio, are frequently ities in aortic regurgitation. J Am Coll Cardiol 1987;9:500–8. inadequate to match the increased mass (33,34). The 12. Dahlo¨f B, Devereux R, de Faire U, et al., for the LIFE Study Group. finding that the ratio of coronary lumen area to regional LV The Losartan Intervention For End point (LIFE) Reduction in Hypertension Study: rationale, design, and methods. Am J Hypertens mass can partially normalize with regression of LVH after 1997;10:705–13. replacement for (35) suggests that 13. Dahlo¨f B, Devereux RB, Ibsen H, et al., for the LIFE Study Group. repolarization abnormalities of strain may also diminish or The Losartan Intervention For End point (LIFE) Reduction in Hypertension Study: baseline characteristics of 9,194 patients with left disappear with regression of hypertrophy after antihyper- ventricular hypertrophy. Hypertension 1998;32:989–97. tensive therapy. 14. Molloy TJ, Okin PM, Devereux RB, Kligfield P. Electrocardiographic 520 Okin et al. JACC Vol. 38, No. 2, 2001 Strain and Hypertrophy August 2001:514–20

detection of left ventricular hypertrophy by the simple QRS voltage- prognostic role of left ventricular hypertrophy in patients with or duration product. J Am Coll Cardiol 1992;20:1180–6. without . Ann Intern Med 1992;117:831–6. 15. Okin PM, Roman MJ, Devereux RB, Kligfield P. Electrocardio- 25. Koren MJ, Devereux RB, Casale PN, Savage DD, Laragh JH. graphic identification of increased left ventricular mass by simple Relation of left ventricular mass and geometry to morbidity and voltage-duration products. J Am Coll Cardiol 1995;25:417–23. mortality in uncomplicated essential hypertension. Ann Intern Med 16. Devereux RB, Bella JN, Boman K, et al. Echocardiographic left 1991;114:345–52. ventricular geometry in hypertensive patients with electrocardiographic 26. Gaasch WH, Zile MR, Hosino PK, Apstein CS, Blaustein AS. left ventricular hypertrophy: The LIFE Study. Blood Press 2001. In Stress-shortening relations and myocardial blood flow in compensated Press. and failing canine hearts with pressure-overload hypertrophy. Circu- 17. Norman JE, Levy D. Improved detection of echocardiographic left lation 1989;79:872–83. ventricular hypertrophy: a correlated data base approach. J Am Coll 27. de Simone G, Devereux RB, Koren MJ, Mensah GA, Casale PN, Cardiol 1995;26:1022–9. Laragh JH. Midwall left ventricular mechanics. An independent 18. Ilercil A, O’Grady MJ, Roman MJ, et al. Reference values for predictor of cardiovascular risk in arterial hypertension. Circulation echocardiographic left ventricular measurements in urban and rural 1996;93:259–65. populations of differing ethnicity: The Strong Heart Study. J Am Soc 28. Palmieri V, Bella JN, DeQuattro V, et al. Relation of diastolic left Echocardiogr 2001;14:601–11. ventricular filling to systolic chamber and myocardial contractility in 19. Sahn DJ, De Maria A, Kisslo J, Weyman A. The Committee on hypertensive patients with left ventricular hypertrophy (the PRE- SERVE study). Am J Cardiol 1999;94:558–62. M-mode Standardization of the American Society of Echocardiogra- 29. Devereux RB, Roman MJ, Palmieri V, et al. Left ventricular wall phy: recommendations regarding quantitation in M-mode echocardi- stresses and wall stress-mass products in hypertensive patients: The ography. Results of a survey of echocardiographic measurements. LIFE Study. J Hypertens 2000;18:1129–38. Circulation 1978;58:1072–83. 30. Thiry PS, Rosenberg RM, Abbott JA. A mechanism for the electro- 20. Schiller NB, Shah PM, Crawford M, et al. American Society of cardiogram in response to left ventricular hypertrophy and acute Echocardiography Committee on Standards, Subcommittee on Quan- ischemia. Circ Res 1975;36:92–104. titation of Two-Dimensional Echocardiograms: recommendations for 31. Kornreich F, Montague TJ, van Herpen G, et al. Improved prediction quantitation of the left ventricle by two-dimensional echocardiogra- of left ventricular mass by regression analysis of body surface potential phy. J Am Soc Echocardiogr 1989;2:358–67. maps. Am J Cardiol 1990;66:485–92. 21. Devereux RB, Alonso DR, Lutas EM, et al. Echocardiographic 32. Rembert JC, Kleinman LH, Fedor JM, Wechsler AS, Greenfield JC assessment of left ventricular hypertrophy: comparison to necropsy Jr. Myocardial blood flow distribution in concentric left ventricular findings. Am J Cardiol 1986;57:450–8. hypertrophy. J Clin Invest 1978;62:379–86. 22. de Simone G, Daniels SR, Devereux RB, et al. Left ventricular mass 33. Anderson HV, Stokes MJ, Leon M, et al. Coronary artery flow and body size in normotensive children and adults: assessment of velocity is related to lumen area and regional left ventricular mass. allometric relations and of the impact of overweight. J Am Coll Circulation 2000;102:48–54. Cardiol 1992;20:1251–60. 34. Villari B, Hess OM, Moccetti D, Vassalli G, Krayenbuehl HP. Effect 23. Devereux RB, Dahlo¨f B, Levy D, Pfeffer MA. Comparison of of progression of left ventricular hypertrophy on coronary artery enalapril vs. nifedipine to decrease left ventricular hypertrophy in dimensions in aortic valve disease. J Am Coll Cardiol 1992;20:1073–9. systemic hypertension (the PRESERVE trial). Am J Cardiol 1996;78: 35. Villari B, Hess OM, Meier C, et al. Regression of coronary artery 61–5. dimensions after successful aortic valve replacement. Circulation 1992; 24. Ghali JK, Liao Y, Simmons R, Castaner A, Cao G, Cooper RS. The 85:972–8.