Additional impact of electrocardiographic over echocardiographic diagnosis of left ventricular hypertrophy for predicting the risk of ischemic stroke

Additional impact of electrocardiographic over echocardiographic diagnosis of left ventricular hypertrophy for predicting the risk of ischemic stroke Shun Kohsaka, MD,a Robert R. Sciacca, EngScD,a Kenichi Sugioka, MD,a Ralph L. Sacco, MD,b Shunichi Homma, MD,a and Marco R. Di Tullio, MDa New York, NY

Background Patients with left ventricular hypertrophy (LVH) have an increased risk of ischemic stroke. Although echocardiography is commonly used for the diagnosis of LVH, there is little information about the potential role of electrocardiography in providing additional prognostic information. The purpose of this study is to determine if electrocardiographically derived criteria for LVH provide additional prognostic value over echocardiography for predicting ischemic stroke in a multiethnic population. Methods

A population-based, case-control study was conducted in 177 patients who had had a first ischemic stroke and in 246 control patients matched for age, gender, and race or ethnicity. Left ventricular mass was measured by using 2-dimensional transthoracic echocardiography. Logistic regression analysis was performed to assess the risk of stroke associated with the presence of LVH diagnosed by electrocardiography (defined by 4 established criteria) after adjustment for the presence of other stroke risk factors and for echocardiographically determined LVH.

Results After adjustment for the presence of other established stroke risk factors, ECG-LVH was associated with ischemic stroke, using Sokolow-Lyon (odds ratio [OR] 2.12, 95% CI 1.05-4.30), Cornell voltage (OR 2.06, 95% CI, 1.26-3.35), and Cornell product criteria (OR 2.12, 95% CI, 1.13-3.97). Cornell voltage criterion (men, N 2.8mV; women, N 2.0mV) was associated with ischemic stroke even after adjustment for echocardiographically determined LVH (OR 1.73, 95% CI, 1.04-2.88). The combination of echo-LVH and a positive Cornell voltage criterion was associated with a 3.5-fold increase in stroke risk. Conclusions

Our study indicates that the presence of ECG-LVH is associated with an increased risk of ischemic stroke after adjustment for other stroke risk factors. For Cornell voltage criteria, this relationship persisted even after adjustment for echocardiographic LVH. Electrocardiographic results can provide independent information for left ventricular myocardial changes and should be considered together with echocardiographic results to fully assess the risk of ischemic stroke. (Am Heart J 2005;149:181- 6.)

Stroke ranks as the third leading cause of death in the United States. Although the treatment of acute ischemic stroke has improved, the greatest reductions in stroke mortality and morbidity are likely to be achieved through more effective prevention strategies.1,2 The identification of high-risk or stroke-prone individuals

From the aDepartment of Medicine and the bNeurological Institute, Columbia University College of Physicians and Surgeons, New York, NY. Submitted November 19, 2003; accepted June 12, 2004. Reprint requests: Marco R. Di Tullio, MD, Associate Professor of Clinical Medicine, Associate Director, Adult Echocardiography Laboratory, Columbia University College of Physicians and Surgeons, PH 3-342 622 West 168th St, New York, NY 10032. E-mail: [email protected] 0002-8703/$ - see front matter n 2004, Elsevier Inc. All rights reserved. doi:10.1016/j.ahj.2004.06.006

could lead to the implementation of specific intervention strategies. Left ventricular hypertrophy (LVH) is among the risk factors for ischemic stroke.3,4 An accurate anatomical diagnosis of LVH predominantly depends on echocardiographic measurements. Calculations of left ventricular (LV) mass, determined by 2D-guided M-mode echocardiographic measurements, have been standardized.5 Little is known, however, about whether electrocardiography (ECG) -derived LVH may provide additional prognostic information for stroke to that of echocardiographically-detected LVH.6 Although ECG is a less sensitive tool for the detection of LVH, it is universally available, technically easy to perform, and highly specific. The severity of LVH detected by ECG is also an important prognostic

American Heart Journal January 2005

182 Kohsaka et al

indicator because it is known to correlate with the risk of cardiovascular disease and carotid atherosclerosis.7-10 The purpose of this study is to evaluate the additional prognostic value of ECG-LVH over echo-LVH for predicting ischemic stroke.

Methods The patient population was drawn from the Northern Manhattan Stroke Study (NOMASS), a community-based epidemiological study aimed at assessing stroke incidence, risk factors, and prognosis in the multiethnic population of northern Manhattan. Prospective case surveillance consisted of daily screening of all hospital admission and discharge records and of head computed tomography (CT ) scan logs at Presbyterian Hospital, where approximately 80% of all patients with cerebral infarction in northern Manhattan are hospitalized. Community control subjects were eligible for the study if a diagnosis of stroke had never been made, they were aged N39 years, and they had resided in northern Manhattan for at least 3 months in a household with a telephone. Control subjects were identified by using random-digit dialing (a dual-frame sampling for identifying both published and unpublished telephone numbers). Control subjects were matched to study subjects by age (within 5 years), gender, and race or ethnicity. Data were collected through interviews by trained research assistants, review of the medical records, physical and neurological examination by the study physicians, in-person measurements, and fasting blood specimens for lipid and glucose measurements. Data were obtained directly from stroke patients and control subjects with the standardized data collection instruments. If a stroke patient was unable to answer questions because of death, aphasia, coma, dementia, or other conditions, a proxy who was knowledgeable about the patient’s history was interviewed. Patients were interviewed as soon as possible after their stroke (median time of 4 days from stroke onset). Control subjects were interviewed in person and evaluated in the same manner as the stroke patients. From June 1993-December 1996, as a part of NOMASS, 234 first-time ischemic stroke patients (aged N39 years) and 286 stroke-free control subjects underwent standard 12-lead ECG and 2-dimensional transthoracic echocardiography. Written informed consent was obtained from all subjects. The study was approved by the Institutional Review Board of ColumbiaPresbyterian Medical Center.

Diagnostic evaluation Stroke risk factors were recorded by direct interview or medical record review for all stroke patients and by direct interview for control subjects. Routine laboratory tests included a complete blood count (CBC), coagulation studies, serum electrolyte studies, liver function tests, and glucose and cholesterol testing. Arterial hypertension was defined by a history of hypertension, the participants’ taking an antihypertensive agent, or blood pressure readings of N140 mm Hg (systolic) or N90 mm Hg (diastolic) during the interview. Hypercholesterolemia was defined by a total serum cholesterol N 240 mg/dL or the presence of appropriate drug treatment. Diabetes mellitus was defined by an abnormal fasting blood sugar, a history of diabetes mellitus, or the need for oral or injectable insulin treatment. Criteria for coronary artery disease

Table I. Patient characteristics and stroke risk factors Group Stroke patients Control subjects (n = 177) (n = 246) P value Demographic criteria Age (years F SD) Male (%) Race (%) Black Hispanic White Stroke risk factors (%) Hypertension Diabetes mellitus Cigarette smoking Coronary artery disease Congestive heart failure Atrial fibrillation Echo-LVH

69 F 12 76 (42.9)

67 F 12 105 (42.7)

79 (32.5) 103 (59.2) 21 (12.1)

50 (28.7) 122 (50.2) 42 (17.3)

134 50 35 53

154 34 37 45

(75.7) (28.3) (20.6) (29.9)

(62.6) (13.8) (15.6) (18.3)

b .005 b .001 .19 b .01

16 (9.0)

10 (4.1)

.04

16 (9.0) 80 (45.2)

4 (1.6) 58 (13.7)

b .001 b .001

LVH, Left ventricular hypertrophy.

(CAD) included a history of myocardial infarction (MI ) or typical angina, previous bypass surgery or angioplasty, or drug treatment. The presence of atrial fibrillation was documented based on the results of a current or past ECG. A diagnosis of congestive heart failure was made based on the presence of clinical signs and symptoms. The diagnosis of ischemic stroke was confirmed in all cases by means of a head CT or magnetic resonance imaging (MRI). The cardiac evaluation included 12lead ECG and 2-dimensional color Doppler transthoracic echocardiography. Stroke diagnostic subtypes were determined by a neurologist on the basis of the presenting clinical syndrome and the results of the diagnostic work-up (head computed tomography or magnetic resonance imaging; carotid Doppler ultrasound; transcranial Doppler ultrasound; a cardiac work-up; and cerebral angiography, when indicated).

Electrocardiographic evaluation A standard 12-lead ECG was recorded at 50 mm/s and 10 mm/ mV and analyzed for LV configuration with the aid of a 5-power magnifying lens and calipers. LVH was defined as (1) a left ventricular strain pattern (a downsloping ST-segment depression) N0.1mV, with T-wave flattening or inversion in leads V4 -V5; (2) Sokolow-Lyon voltage amplitude (SV1 + RV5 or RV6) N3.5mV; (3) Cornell voltage (SV3 + RaVL) N 2.8mV for men and N2.0mV for women; or (4) Cornell product [(SV3 + RaVL)QRS duration] N244uVd s. The left ventricular strain pattern, Sokolow-Lyon voltage, and Cornell voltage have previously been validated in prospective studies.7,8,11 The Cornell product has not been validated in a prospective study but has been used as inclusion criterion in the Losartan Intervention for Endpoint Reduction in Hypertension (LIFE) study.12 The ECG examinations and readings were performed by an experienced physician who was unaware of the other data of the subjects. Subjects were excluded from the present study if they had pathological Q-waves indicating an old MI or patterns associated with

American Heart Journal Volume 149, Number 1

Kohsaka et al 183

Table II. Characteristics among stroke patients LVH present (n = 80) Demographic matching criteria (%) Age (years F SD) 68 F 12 Male (%) 31 (41.3) Race (%) Black 18 (22.8) Hispanic 51 (64.6) White 10 (12.7) Stroke risk factors (%) Hypertension Diabetes mellitus Cigarette smoking Coronary artery disease Congestive heart failure Atrial fibrillation

63 24 17 30 8 12

(78.8) (30.0) (21.8) (37.5) (10.0) (15.0)

Table III. Electrocardiographic measurements Group

LVH absent (n = 97)

P value

68 F 11 43 (44.3)

1.0 .76

32 (33.7) 52 (54.7) 11 (11.3)

.28

71 26 18 23 8 4

.48 .74 .85 .05 .79 .02

(73.2) (26.8) (19.6) (23.7) (8.3) (4.1)

ventricular pacing or bundle branch block. There were 57 patients with ischemic stroke and 40 control patients who were excluded because of these electrocardiographic exclusion criteria.

Echocardiographic evaluation Transthoracic echocardiography was performed in all study subjects with Hewlett-Packard Sonos 1000 or 2500 ultrasound equipment (Hewlett-Packard Imaging Systems Division). Studies were performed and measurements taken according to the guidelines of the American Society of Echocardiography.13 Left ventricular end-diastolic dimension (LVEDD) and interventricular septal thickness (IVS) and posterior wall (PW) thickness at end-diastole were measured. The LV mass was calculated according to the corrected American Society of Echocardiography method14: 0.8(1.04[(IVS + LVEDD + PW)3 – LVEDD3] ) + 0.6. Relative LV mass was defined as the ratio of observed LV mass to predicted LV mass based on height and gender.15 Predicted LV mass was derived from the results of a regression of observed LV mass (log-transformed to increase normality) on height and gender in a control population without cardiovascular risk factors. Left ventricular hypertrophy was defined as any value for relative LV mass N90th percentile. The interpretation of the echocardiographic studies was blinded to casecontrol status and other clinical characteristics.

Statistical analysis Data were reported as the mean F SD for continuous variables and as frequency for categorical variables. Differences between proportions were assessed by the m2 test and replaced by the Fisher exact test when the expected cell count was b5. Differences between mean values were assessed by the unpaired Student t test. A 2-tailed P value of b.05 was considered significant. Univariate and multivariate logistic regression analysis (PROC LOGISTIC, SAS statistical package, version 8.2; SAS Inc, Cary, NC) was used to test the association between ECG-LVH (independent variable) and ischemic stroke (dependent variable). Unadjusted odds ratios for the association between ECGLVH and ischemic stroke were calculated.

Stroke patients Control subjects (n = 177) (n = 246) P value Sokolow-Lyon voltage Mean (mV) 25.5 F 8.8 N3.5mV 24 (13.6%)

22.3 F 7.7 16 (6.5%)

Left ventricular strain Mean (mV) N0.1mV

0.074 F 0.044 38 (21.5%)

0.071 F 0.038 38 (15.5%)

.51 .12

2.00 F 0.94

1.74 F 0.67

.001

Cornell voltage Mean (mV) N2.8 mV (men) N2.0 mV (women)

57 (32.2%) Cornell product Mean (AV d s) N244 AV d s

176 F 90 31 (17.5%)

47 (19.1%)

144 F 63 20 (8.1%)

b .001 .01

.002

b .001 .004

Multivariate analysis was used to determine the adjusted odds ratio for ECG-LVH and ischemic stroke after other established stroke risk factors were included as potential confounding factors in the model. Variables significantly associated with ischemic stroke by univariate analysis (arterial hypertension, diabetes mellitus, atrial fibrillation, coronary artery disease, and congestive heart failure) were entered along with age and gender as independent variables in the analysis. The presence of echo-LVH was included in a second multivariate model to determine whether the association of ECGLVH and ischemic stroke was independent of measurements derived from the echocardiogram.

Results The present report is based on data from 177 stroke patients (76 men, 101 women; mean age 68 F 12 years) and 246 control subjects for whom all morphologic, electrocardiographic, and echocardiographic variables required for the study could be obtained. Demographics and stroke risk factors for the study population are shown in Table I. This was a predominantly elderly cohort. Women comprised 57% of the cohort. The raceethnic distribution of the patients was as follows: Hispanic (54%), non-Hispanic black (31%), and nonHispanic white (15%). There was a greater prevalence of hypertension, diabetes, coronary artery disease, atrial fibrillation, and echo-LVH among ischemic stroke patients than among the control subjects. Demographics of the stroke patients with and without LVH are shown in Table II. Stroke patients with LVH had a higher incidence of atrial fibrillation; however, the distribution of the other stroke risk factors was equivalent. With respect to stroke subtype, information was obtained for 176 of the 177 patients: 31 patients (17.6%) had embolic strokes, 71 (40.3%) had cryptogenic strokes,

American Heart Journal January 2005

184 Kohsaka et al

Table IV. Association between various LVH criteria and ischemic stroke Odds ratio (95% CI)

LVH criteria Echocardiography Sokolow-Lyon Strain pattern Cornell voltage Cornell product

Unadjusted 2.67 2.26 1.49 2.01 2.40

Adjusted for stroke risk factors

(1.76-4.06) (1.16-4.38) (0.90-2.45) (1.29-3.15) (1.32-4.37)

2.23 2.12 1.24 2.06 2.12

(1.43-3.46) (1.05-4.30) (0.73-2.10) (1.26-3.35) (1.13-3.97)

Adjusted for stroke risk factors and echo-LVH

1.96 1.02 1.73 1.79

N/A (0.96-4.02) (0.59-1.78) (1.04-2.88) (0.94-3.43)

LVH, Left ventricular hypertrophy.

43 (24.4%) had lacunar strokes, 24 (13.6%) had atherosclerotic strokes, 5 (2.8%) had a conflicting mechanism, and 2 (1.1%) had strokes of another etiology. Electrocardiographic measurements are listed in Table III. Echo-LVH was associated with ischemic stroke after adjustment for other stroke risk factors (adjusted OR 2.23, 95% CI, 1.43-3.46). ECG-LVH was also associated with ischemic stroke, using either Sokolow-Lyon (unadjusted OR 2.26, 95% CI, 1.16-4.38), Cornell voltage (unadjusted OR 2.01, 95% CI, 1.29-3.15), or Cornell product criteria (unadjusted OR 2.40, 95% CI, 1.324.37). An increased risk was still present for these criteria even after adjustment for the presence of other established stroke risk factors [OR 2.12 (95% CI, 1.054.30), OR 2.06 (95% CI, 1.26-3.35), and OR 2.12 (95% CI, 1.13-3.97), respectively]. The Cornell voltage criterion was significantly associated with stroke even after echoLVH was added to the model (OR 1.73, 95% CI, 1.042.88) (Table IV). The combination of echo-LVH and a positive Cornell voltage criterion carried a more than 3-fold increase in stroke risk compared with the absence of either variable (unadjusted OR 3.87, 95% CI, 2.20-6.79, adjusted OR 3.45, 95% CI, 1.88-6.34).

Discussion The identification of patients at high-risk for stroke is important because epidemiological data suggest a substantial leveling-off of prior declines in stroke-related mortality and a possible increase in stroke incidence. In the present study, 3 of the 4 established ECG-LVH criteria were associated with an increased risk of ischemic stroke, even after adjustment for the presence of other established stroke risk factors. Furthermore, there was a significant association between ischemic stroke and ECGLVH–defined Cornell voltage criterion (N2.8mV for men, N2.0 mV for women), even after additional analysis to adjust for the presence of echo-LVH. LVH imparts an increased cardiovascular risk, whether it is diagnosed by ECG 7,8,16 or by echocardiography.17,18 The first insights into the worst survival associated with LVH originally came from the ECG

literature; however, in the early 1990s, the prognosis of echo-LVH was validated in a prospective, populationbased, large-scale study.19 Currently, ECG is considered a less-sensitive method for detecting anatomical abnormalities such as LVH.5 In fact, in the entire Framingham Study population, ECG-LVH was less prevalent than echo-LVH, with overall rates of 2.4% and 17.4%, respectively.20 Although ECG is a less accurate tool for the detection of anatomic LVH, our study validates its role as an adjunct tool to echocardiography in accurately estimating the risk of stroke. In addition, ECG is an easy and inexpensive screening tool for LVH, and aids in the selection of patients for the more quantitative echocardiographic evaluation. Because the association between the Cornell voltage criterion and ischemic stroke persisted after adjusting for echocardiographicallydetected LV mass, it appears that ECG can provide separate and complementary diagnostic information to that derived from echo-LVH. After adjustment for other stroke risk factors, the combination of echo-LVH and a positive Cornell voltage criterion was associated with a 3.5-fold increase in stroke risk, which was substantially higher than that of echo-LVH alone. ECG-LVH may detect electrical abnormalities induced by the hypertrophic process that are not detectable by echocardiography, and this information may aid in stroke risk prediction. Furthermore, the use of ECG may be helpful for detecting patients exhibiting levels of LVH that exceed the need to sustain cardiac workload, since echocardiography alone is insufficient to accurately differentiate this bpathologicalQ LVH from compensatory anatomical LV adaptation that balances chronic overload.21 Therefore, ECG-LVH and echo-LVH appear to measure different entities, and should be considered independent and non-parallel clinical variables. Different LVH criteria measure different cardiac properties, and changes on ECG are not directly comparable to those observed with echocardiography.22 Of the 4 ECG-LVH criteria we analyzed, the Cornell voltage criterion has been validated in prospective studies, and its serial changes are known to

American Heart Journal Volume 149, Number 1

reflect changes in the risk of cardiovascular events.7,11 In our study, the Cornell voltage was the only ECG criterion that was independently predictive of ischemic stroke after adjusting for echo-LVH. Although they remained independent predictors after adjustment for various other stroke risk factors, the Sokolow-Lyon and Cornell product criteria were not significantly associated with stroke after adjustment for echo-LVH. The reason for the different performance of different ECG parameters is not clear. It should be noted that different limitations exist for each criterion. The Sokolow-Lyon criterion has similar limitations to those of echocardiography, including the attenuating effects of aging and obesity.23 There is probably a larger overlap of patients with LVH detected by the SokolowLyon criterion and echo-LVH than with the Cornell voltage criterion and echo-LVH, making the former less likely to provide different information from that of echo. On the other hand, the Cornell product criterion was developed to reflect conduction delay from LVHrelated remodeling. It is possible that LV remodeling does not play a significant role in the risk of ischemic stroke, decreasing the predictive relevance of the product criterion. Finally, racial differences in the performance of ECG criteria for LVH have been reported.25-29 The studies investigating the relationship between LVH and prognosis have been conducted in cohorts drawn from predominantly white populations, or in patients who were hypertensive or had been referred for coronary angiography. The prognosis of LVH in a community-based cohort with a large representation of minorities is largely unexplored. The broad socio-demographic distribution of the NOMASS study allowed us to evaluate the risk of ischemic stroke in a multiethnic population largely composed of Caribbean Hispanics, who represent the fastest growing minority in the United States. Our study does not have the power to address the association between ECG findings and different ischemic stroke subtypes. Hemorrhagic strokes also were not included. The case-control design has some limitations, including the possibility that differences between the patients and controls may have existed in variables that were not measured. Finally, the relatively small sample size precludes separate analyses of ECG-LVH criteria in different race-ethnic subgroups. In summary, the presence of ECG-LVH is associated with an increased risk of ischemic stroke after adjustment for other stroke risk factors. For the Cornell voltage criterion, this relationship persisted after adjustment for echo-LVH. Although the exact mechanism underlying the role of ECG-LVH in the prediction of stroke risk is not clear, our study indicates that it can provide independent prognostic information from that of echo-LVH, and that the data from both tests should be assessed together to fully assess the risk of ischemic stroke.

Kohsaka et al 185

References 1. Bonita R, Beaglehole R. Does treatment of hypertension explain the decline in mortality from stroke? Br Med J (Clin Res Ed) 1986;292:191 - 2. 2. Gorelick PB. Stroke prevention. An opportunity for efficient utilization of health care resources during the coming decade. Stroke 1994;25:220 - 4. 3. Rodriguez CJ, Sacco RL, Sciacca RR, et al. Physical activity attenuates the effect of increased left ventricular mass on the risk of ischemic stroke: The Northern Manhattan Stroke Study. J Am Coll Cardiol 2002;39:1482 - 8. 4. Bots ML, Nikitin Y, Salonen JT, et al. Left ventricular hypertrophy and risk of fatal and non-fatal stroke. EUROSTROKE: a collaborative study among research centres in Europe. J Epidemiol Community Health 2002;56(1 Suppl ):i8-i13. 5. Lorell BH, Carabello BA. Left ventricular hypertrophy: pathogenesis, detection, and prognosis. Circulation 2000;102:470 - 9. 6. Sundstrom J, Lind L, Arnlov J, et al. Echocardiographic and electrocardiographic diagnoses of left ventricular hypertrophy predict mortality independently of each other in a population of elderly men. Circulation 2001;103:2346 - 51. 7. Levy D, Salomon M, D’Agostino RB, et al. Prognostic implications of baseline electrocardiographic features and their serial changes in subjects with left ventricular hypertrophy. Circulation 1994;90:1786 - 93. 8. Kannel WB, Gordon T, Castelli WP, et al. Electrocardiographic left ventricular hypertrophy and risk of coronary heart disease. The Framingham study. Ann Intern Med 1970;72:813 - 22. 9. Kannel WB. Prevalence and natural history of electrocardiographic left ventricular hypertrophy. Am J Med 1983;75:4 - 11. 10. Okin PM, Roman MJ, Devereux RB, et al. Association of carotid atherosclerosis with electrocardiographic myocardial ischemia and left ventricular hypertrophy. Hypertension 1996;28:3 - 7. 11. Verdecchia P, Schillaci G, Borgioni C, et al. Prognostic value of a new electrocardiographic method for diagnosis of left ventricular hypertrophy in essential hypertension. J Am Coll Cardiol 1998;31:383 - 90. 12. Dahlof B, Devereux RB, Julius S, et al. Characteristics of 9194 patients with left ventricular hypertrophy: the LIFE study. Losartan intervention for endpoint reduction in hypertension. Hypertension 1998;32:989 - 97. 13. Sahn DJ, DeMaria A, Kisslo J, et al. Recommendations regarding quantitation in M-mode echocardiography: results of a survey of echocardiographic measurements. Circulation 1978;58:1072 - 83. 14. Devereux RB, Alonso DR, Lutas EM, et al. Echocardiographic assessment of left ventricular hypertrophy: comparison to necropsy findings. Am J Cardiol 1986;57:450 - 8. 15. Lauer MS, Larson MG, Levy D. Gender-specific reference M-mode values in adults: population-derived values with consideration of the impact of height. J Am Coll Cardiol 1995;26:1039 - 46. 16. De Bacquer D, De Backer G, Kornitzer M, et al. Prognostic value of ECG findings for total, cardiovascular disease, and coronary heart disease death in men and women. Heart 1998;80:570 - 7. 17. Levy D, Garrison RJ, Savage DD, et al. Left ventricular mass and incidence of coronary heart disease in an elderly cohort. The Framingham heart study. Ann Intern Med 1989;110:101 - 7. 18. Verdecchia P, Porcellati C, Reboldi G, et al. Left ventricular hypertrophy as an independent predictor of acute cerebrovascular events in essential hypertension. Circulation 2001;104: 2039 - 44.

American Heart Journal January 2005

186 Kohsaka et al

19. Levy D, Garrison RJ, Savage DD, et al. Prognostic implications of echocardiographically determined left ventricular mass in the Framingham heart study. N Engl J Med 1990;322:1561 - 6. 20. Levy D, Savage DD, Garrison RJ, et al. Echocardiographic criteria for left ventricular hypertrophy: the Framingham heart study. Am J Cardiol 1987;59:956 - 60. 21. de Simone G, Verdecchia P, Pede S, et al. Prognosis of inappropriate left ventricular mass in hypertension: the MAVI Study. Hypertension 2002;40:470 - 6. 22. Van Den Hoogen JP, Mol WH, Kowsoleea A, et al. Reproducibility of electrocardiographic criteria for left ventricular hypertrophy in hypertensive patients in general practice. Eur Heart J 1992;13:1606 - 10. 23. Levy D, Labib SB, Anderson KM, et al. Determinants of sensitivity and specificity of electrocardiographic criteria for left ventricular hypertrophy. Circulation 1990;81:815 - 20. 24. Kronmal RA, Smith VE, O’Leary DH, et al. Carotid artery measures are strongly associated with left ventricular mass in older adults

25.

26.

27.

28.

29.

(a report from the Cardiovascular Health Study). Am J Cardiol 1996;77:628 - 33. Lee DK, Marantz PR, Devereux RB, et al. Left ventricular hypertrophy in black and white hypertensives. Standard electrocardiographic criteria overestimate racial differences in prevalence. JAMA 1992;267:3294 - 9. Chapman JN, Mayet J, Chang CL, et al. Ethnic differences in the identification of left ventricular hypertrophy in the hypertensive patient. Am J Hypertens 1999;12:437 - 42. Crow RS, Prineas RJ, Rautaharju P, et al. Relation between electrocardiography and echocardiography for left ventricular mass in mild systemic hypertension (results from Treatment of Mild Hypertension Study). Am J Cardiol 1995;75:1233 - 8. Jaggy C, Perret F, Bovet P, et al. Performance of classic electrocardiographic criteria for left ventricular hypertrophy in an African population. Hypertension 2000;36:54 - 61. Xie X, Liu K, Stamler J, et al. Ethnic differences in electrocardiographic left ventricular hypertrophy in young and middle-aged employed American men. Am J Cardiol 1994;73:564 - 7.