Cornell Product Left Ventricular Hypertrophy in Electrocardiogram and the Risk of Stroke in a General Population
Left ventricular hypertrophy (LVH), assessed by ECG, is associated with an increased risk for cardiovascular events among hypertensive subjects. We evaluated the risks of LVH in a Japanese general population including normotensive and prehypertensive subjects. We measured ECG and blood pressure in 10 755 subjects at baseline. The Cornell product (CP) and Sokolow-Lyon (SL) voltage were calculated as markers of LVH (CP ≥2440 mm×ms and SL voltage ≥38 mm). Follow-up was performed for 10 years, and the incidence of stroke and myocardial infarction was evaluated. The prevalence of CP-LVH was 2.7% for normotensives, 5.2% for prehypertensives, and 11.0% for hypertensives, and the prevalence of SL-LVH was 5.0%, 8.2%, and 15.2%, respectively. In all of the subjects, CP-LVH and SL-LVH were both predictors of stroke (CP-LVH: hazard risk: 1.62, 95% CI: 1.19 to 2.20, P=0.002; SL-LVH: hazard risk: 1.29, 95% CI: 0.98 to 1.71, P=0.07) after adjustment for confounding factors but were not predictors of myocardial infarction. The adjusted hazard ratio of CP-LVH predicting stroke was especially high in the normotensives (hazard risk: 7.53; 95% CI: 3.39 to 16.77). In the normotensives, diabetes mellitus and hyperlipidemia were significant determinants of CP-LVH but not of SL-LVH. In all of the hypertensive subgroups (normotensives, prehypertensives, and hypertensives), the c-statistic for the equation predicting stroke increased when CP-LVH was added to the model but not when SL-LVH was added. In conclusion, both CP-LVH and SL-LVH are risk factors for stroke in the Japanese general population. CP-LVH is related to glucose abnormality, and its predictive value for stroke is seen even in normotensives and prehypertensives.
Left ventricular hypertrophy (LVH), a measure of hypertensive target organ damage in the heart, has been reported to be associated with increased morbidity and mortality.1,2 LVH can be evaluated by echocardiography (Echo) and/or a 12-lead ECG. LVH defined by ECG (ECG-LVH) has been evaluated using standard voltage criteria reported by Sokolow and Lyon (SL)3 and more recently using the Cornell product (CP) criteria.4 LVH detected by CP (CP-LVH) is reported to have a higher sensitivity for the presence of LVH evaluated by echocardiography (Echo-LVH) than the Sokolow-Lyon criteria (SL-LVH).5 CP-LVH and SL-LVH were independently associated with Echo-LVH6 and with stroke events, cardiovascular morbidity, and mortality7 in subjects with essential hypertension. These data show that evaluation of CP-LVH is beneficial in Western hypertensive subjects; however, in Japanese populations, the incidence of stroke is higher than that of ischemic heart disease,8 and there are no data on the cardiovascular risk in Japanese subjects with CP-LVH.
In addition, hypertension is the major cause of LVH, and the cardiovascular risks in normotensive and prehypertensive subjects with CP-LVH and/or SL-LVH remain unknown. It is reported that hypertensive patients with diabetes mellitus have a higher prevalence and greater severity of LVH than those without diabetes mellitus.9,10 Okin et al11 reported that diabetes mellitus, per se, attenuates the regression of hypertensive LVH during antihypertensive treatments. However, there are no data examining whether cardiovascular risk factors such as diabetes mellitus might contribute to an increase of ECG-LVH in normotensive subjects or whether LVH in normotensive and prehypertensive subjects can be a risk factor for cardiovascular events.
The purpose of the present study was to evaluate the cardiovascular risks of CP-LVH and SL-LVH in the Japanese general population and to explore the differences in backgrounds and prognostic factors that predict the presence of CP-LVH and SL-LVH, especially in normotensive and prehypertensive subjects.
The Jichi Medical School Cohort Study was begun in 1992, with the primary aim of clarifying the risk factors for cardiovascular and cerebrovascular diseases in the Japanese general population. The details of the protocol of the Jichi Medical School Cohort Study have been reported previously.12 Baseline data were collected between April 1992 and July 1995 in 12 rural districts using a government-sponsored mass screening system. In each community, a local government office sent personal invitations by mail to all of the subjects in accordance with the health and medical service law for the aged. The subjects for the mass screening examinations were residents aged 40 to 69 years in 8 areas (Iwaizumi, Tako, Kuze, Sakuma, Sakugi, Okawa, Ainoshima, and Akaike). Subjects included those aged ≥30 years in 1 area (Wara), and other age groups were also included in 3 areas (Hokudan, Yamato, and Takasu). The total number of subjects in the Jichi Medical School Cohort Study at baseline was 12 490 (4911 men and 7579 women). The participation rate varied in each community (26.0% to 90.0%), and the overall participation rate of those invited to the mass screening examination program was 65.4%.13
ECG Measurement and Interpretation
ECG was measured at a paper speed of 25 mm/s, at a gain of 10 mm/mV (or 5 mm/mV), using ECG devices that the institutes had (FCP130-A9, FCP145-M4, and FCP270-M5, Fukuda Denshi, etc). A trained person, who did not know the subjects’ backgrounds, measured ECGs at a central laboratory using a ruler with 0.01-mm graduations. Both SL voltage (SV1+RV5) and Cornell voltage (RaVL+SV3, with 6 mm added for women)4,5 were measured. QRS duration was measured manually from lead II (or lead I or III if the measurement of QRS duration was difficult from lead II) on a single heart beat. CP was calculated as the product of Cornell voltage times QRS duration afterward. SL-LVH was defined as ≥38 mm (3.8 mV), and CP-LVH was defined as 2440 mm×ms according to a previous report of the Losartan Intervention for Endpoint Reduction in Hypertension Study.7
Questionnaire and Other Measurements
Information about medical history and lifestyle was obtained with a questionnaire at baseline. Age is the value at baseline. Smoking status was reported as current smoker, ex-smoker, or never smoked. Alcohol drinkers were defined as those who were reported consuming ≥20 g/d. Body mass index (BMI) was calculated as weight (kilograms)/height (meters squared). The systolic blood pressure (SBP) and diastolic blood pressure (DBP) at baseline were measured using a fully automated and validated upper arm cuff-oscillometric device, the BP203RV-II (Nippon Colin).14 Blood pressure was measured once after resting for ≥5 minutes while seated. Hypertension was defined as either a SBP/DBP of ≥140/90 mm Hg or taking antihypertensive medications. Prehypertension was defined as SBP/DBP 120/80 to 139/89 mm Hg. Normotension was defined as SBP/DBP <120/80 mm Hg. Diabetes mellitus was defined by a fasting glucose level ≥7.0 mmol/L (126 mg/dL), a casual glucose level >11.1 mmol/L (200 mg/dL), or the use of an oral hypoglycemic agent or insulin. Impaired fasting glucose was defined as a fasting glucose level of 110 to 125 mg/dL. Hyperlipidemia was defined as a total cholesterol level ≥5.7 mmol/L (220 mg/dL), a triglyceride level ≥1.7 mmol/L (150 mg/dL), or the use of an oral lipid-lowering agent, according to the Japanese Atherosclerosis Society Guidelines for Prevention of Atherosclerotic Cardiovascular Diseases.
The internal review board of the Jichi Medical University School of Medicine approved this study. Written informed consent for the study was obtained individually from all of the subjects during the mass screening examination health checkup.
Follow-Up and Diagnostic Criteria
The mass screening examination system was used to check the subjects every year for 10 years. The details of follow-up are shown in the online supplemental data file (please see http://hyper.ahajournals.org). The diagnosis was determined independently by an end points committee, which included radiologists, neurologists, and cardiologists, in accordance with the World Health Organization Monitoring of Trends and Determinants in Cardiovascular Disease Project.15 The details of diagnostic criteria are also shown in the online supplemental data file.
Among the 12 490 subjects who were initially enrolled in the Jichi Medical School Cohort Study, we analyzed 10 755 subjects who had adequate follow-up, after excluding 1735 subjects with no ECG recording (n=1285), immeasurable ECG findings (n=28), complete left bundle branch block (n=20), complete right bundle branch block (n=189), atrial fibrillation (n=53), or no blood pressure data (n=160).
Data are shown as means±1 SDs for continuous variables and as percentages for dichotomous variables. Differences in characteristics between subjects with and without CP-LVH and SL-LVH were evaluated using Student t test or χ2 test. Because the prevalence of LVH varied among the hypertensive subgroups (normotension, prehypertension, and hypertension), we performed analyses of characteristics of the subjects with CP-LVH in the hypertensive subgroups. Determinants of CP-LVH and SL-LVH in all of the subjects and separately for each hypertensive subgroup were evaluated using multivariate logistic regression analysis including age, gender, BMI, smoking status, alcohol drinking, SBP, antihypertensive medication use (only in the hypertensives), presence of hyperlipidemia, and status of diabetes mellitus. The incidence risks of stroke and myocardial infarction in both unadjusted models and those adjusted for significant covariates were evaluated using Cox regression analysis in the total sample and separately for each hypertension subgroup. The c-statistic was calculated, according to the method of Pencina and D’Agostino,16 for the baseline model that included the unmodifiable cardiovascular risk factors (age and gender) and then for a series of models in which each cardiovascular risk factor was added separately to the baseline model. Computer software SPSS 16.0 (SPSS Inc) and SAS 9.1 (SAS Institute) were used for the analysis, and a P value <0.05 was considered statistically significant.
The mean age was 55.6±11.2 years (men: 37.8%). The average SBP/DBP was 130±21/78±12 mm Hg. The percentages of subjects with normotension, prehypertension, and hypertension were 32.9%, 32.6%, and 34.5%, respectively. The characteristics of the study subjects were as follows: history of stroke, 1.0%; history of myocardial infarction, 0.5%; antihypertensive medication use, 11.1%; former smoker, 12.4%; current smoker, 21.8%; alcohol drinker, 27.6%; hyperlipidemia, 35.3%; impaired fasting glucose, 2.5%; and diabetes mellitus, 3.6%. The prevalences of CP-LVH and SL-LVH in the full sample were 6.4% were 9.5%, respectively.
Characteristics and Determinants of CP-LVH and SL-LVH in All Subjects
Comparisons of subjects with and without CP-LVH and SL-LVH are shown in Table 1. Subjects with CP-LVH and those with SL-LVH were older and had a higher prevalence of hypertension and antihypertensive medication use; however, there were differences in characteristics such as gender, BMI, and hyperlipidemia, and diabetes mellitus depended on whether we grouped subjects by CP-LVH or SL-LVH.
In all of the subjects, significant determinants of both CP-LVH and SL-LVH were SBP (CP-LVH: odds ratio [OR]: 1.20 per 10 mm Hg, 95% CI: 1.15 to 1.25; SL-LVH: OR: 1.26 per 10 mm Hg, 95% CI: 1.21 to 1.30) and antihypertensive medication use (CP-LVH: OR: 1.79, 95% CI: 1.45 to 2.20; SL-LVH: OR: 1.71, 95% CI: 1.41 to 2.08). Additional determinants of only CP-LVH were age (OR: 1.16 per 10 years; 95% CI: 1.07 to 1.27) and higher BMI (OR: 1.03 per 1 kg/m2; 95% CI: 1.00 to 1.06) and of only SL-LVH were male gender (OR: 3.00; 95% CI: 2.47 to 3.65) and lower BMI (OR: 0.92 per 1 kg/m2; 95% CI: 0.89 to 0.93).
Characteristics and Determinants of Subjects With CP-LVH and SL-LVH in the Hypertensive Groups
The characteristics of subjects with and without CP-LVH and those with and without SL-LVH in the 3 hypertensive subgroups (normotensives, prehypertensives, and hypertensives) are shown in the online supplemental data file (Tables S1 and S2).
In the normotensives, the significant determinants of both CP-LVH and SL-LVH were male gender (CP-LVH: OR: 2.15, 95% CI: 1.22 to 3.78; SL-LVH: OR: 3.55, 95% CI: 2.30 to 5.48) and lower BMI (CP-LVH: OR: 0.90 per 1 kg/m2, 95% CI: 0.83 to 0.98; SL-LVH: OR: 0.93 per 1 kg/m2, 95% CI: 0.87 to 0.99). The additional determinants of only CP-LVH were nonsmokers (ex-smokers: OR: 0.45, 95% CI: 0.24 to 0.87; current smokers: OR: 0.24, 95% CI: 0.09 to 0.66), the presence of hyperlipidemia (OR: 1.66; 95% CI: 1.05 to 2.61), and diabetes mellitus (OR: 3.26; 95% CI: 1.24 to 8.53), and that of only SL-LVH was age (OR: 1.15 per 10 years; 95% CI: 1.01 to 1.32).
In the prehypertensives, the significant determinants of both CP-LVH and SL-LVH were SBP (CP-LVH: OR: 1.38 per 10 mm Hg, 95% CI: 1.07 to 1.79; SL-LVH: OR: 1.27 per 10 mm Hg, 95% CI 1.03 to 1.57). The significant determinant of only CP-LVH was age (OR: 1.22 per 10 years; 95% CI: 1.04 to 1.42), and those of only SL-LVH were male gender (OR: 2.88; 95% CI: 2.00 to 4.14) and lower BMI (OR: 0.90 per 1 kg/m2; 95% CI: 0.86 to 0.95).
In the hypertensives, the significant determinants of both CP-LVH and SL-LVH were antihypertensive medication use (CP-LVH: OR: 1.51, 95% CI: 1.20 to 1.89; SL-LVH: OR: 1.59, 95% CI: 1.28 to 1.96) and SBP (CP-LVH: OR: 1.07 per 10 mm Hg, 95% CI: 1.01 to 1.14; SL-LVH: OR: 1.17 per 10 mm Hg; 95% CI: 1.11 to 1.24). The significant determinants of only CP-LVH were female gender (male: OR: 0.68; 95% CI: 0.49 to 0.95) and higher BMI (OR: 1.04 per 1 kg/m2; 95% CI: 1.00 to 1.07), and those of only SL-LVH were male gender (OR: 2.82; 95% CI: 2.14 to 3.71), lower BMI (OR: 0.91 per 1 kg/m2; 95% CI: 0.88 to 0.94), nonsmokers (current smokers: OR: 0.71; 95% CI: 0.51 to 0.99), and absence of impaired fasting glucose or diabetes mellitus (diabetes: OR: 0.53; 95% CI: 0.33 to 0.86).
We followed up the subjects for an average of 127.5±30.3 months (total: 114 270 person-years). There were 391 clinical stroke events (cerebral hemorrhage: n=91; cerebral infarction: n=252; subarachnoid hemorrhage: n=47; and unknown: n=1) and 79 myocardial infarction events. In separate Cox hazard models, both CP-LVH (hazard ratio [HR]: 1.67; 95% CI: 1.23 to 2.26; P=0.001) and SL-LVH (HR: 1.35; 95% CI: 1.02 to 1.78; P=0.035) were significant risk factors for stroke after adjustment for the following potential confounding factors: age, gender, BMI, history of stroke, history of myocardial infarction, smoking status, alcohol drinking, antihypertensive medication use, SBP level, presence of hyperlipidemia, and status of diabetes mellitus. The results of Cox regression analyses that included CP-LVH and SL-LVH together are shown in Table 2. In the total sample, CP-LVH and SL-LVH were both significant predictors of stroke risk in the unadjusted model; CP-LVH, but not SL-LVH, remained a significant predictor after adjustment for the confounding factors. In the unadjusted Cox regression analysis predicting myocardial infarction, SL-LVH was associated with an increased risk, whereas CP-LVH was not. However, the increased risk for subjects with SL-LVH disappeared after adjustment for the confounding factors (Table 2). When we divided the subjects into 4 groups based on presence/absence of CP-LVH and SL-LVH, the incidence of stroke was significantly higher in subjects with CP-LVH alone and those with both CP-LVH and SL-LVH compared to those with neither CP-LVH nor SL-LVH (Figure 1), and SL-LVH increased the risk for stroke in addition to CP-LVH.
Incidence of Stroke in Subjects With CP-LVH and SL-LVH in the Hypertension Subgroups
The incidences of stroke in subjects with CP-LVH and those with SL-LVH in the 3 hypertension subgroups are shown in Figure 2. In the normotensives, the incidence of stroke was higher in subjects with CP-LVH than in those with SL-LVH, although it was similar for subjects with CP-LVH and SL-LVH within prehypertensives and hypertensives. In separate Cox regression analyses for the 3 groups, the HR for stroke events in subjects with CP-LVH was higher in the normotensives (HR: 7.53; 95% CI: 3.39 to 16.77) than in the prehypertensives (HR: 1.49; 95% CI: 0.65 to 3.46) and hypertensives (HR: 1.41; 95% CI: 0.99 to 2.02; Figure 2A), after adjustment for age, gender, BMI, smoking status and alcohol drinking, history of stroke, history of myocardial infarction, diabetes mellitus status, hyperlipidemia, and SBP. However, the HR for stroke events in subjects with SL-LVH was not a significant predictor of stroke in subjects for any of the hypertensive groups (normotensives: HR: 1.15, 95% CI: 0.40 to 3.32; prehypertensives: HR: 1.71, 95% CI: 0.89 to 3.25; hypertensives: HR: 1.26, 95% CI: 0.91 to 1.73; Figure 2B).
C-Statistics of LVH for Predicting Stroke Event in Hypertensive Groups
The c-statistics for the models that separately contain ECG-LVH and each of the other cardiovascular risk factors for the Cox regression models predicting stroke in hypertensive groups are shown in Table 3. The inclusion of CP-LVH increased the c-statistic in all of the hypertensive groups, but the inclusion of SL-LVH did not. Inclusion of CP-LVH resulted in similar c-statistic values to that of status of diabetes mellitus in prehypertensives and resulted in the highest c-statistic value in normotensives.
Both CP-LVH and SL-LVH were risk factors for the incidence of stroke in the Japanese general population. Screening of ECG using CP-LVH, in addition to SL-LVH, is beneficial for detecting individuals who are at high risk for stroke. The predictive value of CP-LVH was seen even in normotensive subjects, where those with CP-LVH were more likely to have hyperlipidemia and diabetes mellitus. In normotensive subjects, the presence of CP-LVH is the strongest predictor of future stroke events compared with the conventional cardiovascular risk factors. In addition, in prehypertensive subjects, for whom antihypertensive medication is controversial, the presence of CP-LVH independently predicted stroke as did the status of diabetes mellitus, history of stroke, history of myocardial infarction, and smoking status.
CP-LVH and SL-LVH were related to different risk factors, although they were both ECG markers of LVH. Okin et al17 reported in hypertensive patients for whom the presence of CP-LVH was predominantly associated with higher BMI, whereas SL-LVH was predominantly related to lower BMI and male gender. The determinants of CP-LVH in the present study of community-dwelling Japanese individuals were similar to the results in the Losartan Intervention for Endpoint Reduction in Hypertension Study (hypertensive subjects).17
CP-LVH was related to metabolic factors (hyperlipidemia and diabetes mellitus), especially in normotensive subjects, but SL-LVH was not. Okin et al6 reported that abnormal left ventricular geometry was more common in patients with only CP-LVH than in patients with only SL-LVH, and Shirai et al18 reported that CP-LVH is a better marker of relative wall thickness (RWT) than SL-LVH. We reported previously that hypertensive patients with type 2 diabetes mellitus had greater RWT (but not increased left ventricular mass index) than those without diabetes mellitus.19 Therefore, CP-LVH might reflect increased RWT related to diabetes mellitus. In diabetic patients, RWT is a better marker of cardiovascular events than left ventricular mass index,20 and the higher predictive value of CP-LVH for stroke might be derived from that of diabetes-associated RWT. It is reported that protein glycation can cause myocardial damage in diabetes mellitus, and receptor binding of advanced glycation products induces proinflammatory cytokines, inflammation, growth factor release, and fibrosis.21 On the other hand, the presence of diabetes mellitus was a negative predictor of SL-LVH in the hypertensive group of the present study. Diabetic subjects are more often overweight, which can reduce the SL product and, thus, SL-LVH on ECG. SL-LVH was more strongly associated with hypertension than with metabolic factors such as diabetes mellitus and hyperlipidemia compared with CP-LVH.
Evaluation of both CP-LVH and SL-LVH may be useful for detecting subjects at an increased risk of stroke events. The subjects with both CP-LVH and SL-LVH had an increased risk for stroke (Figure 1). In the present study, the predictive value of SL-LVH for stroke lost the significance when we adjusted for CP-LVH and conventional risk factors (Table 2); however, the risks of both CP-LVH and SL-LVH were significant when we preformed a parallel analysis excluding the subjects with history of stroke and myocardial infarction (data not shown).
In prehypertensives, the c-statistic increased when CP-LVH, history of stroke, history of myocardial infarction, status of diabetes, and smoking were each added to a baseline model that included only age and gender; however, the c-statistic did not increase when SBP was added. Therefore, a prehypertensive with these risk factors should be monitored cautiously and may be a candidate for lifestyle modification (especially cessation of smoking and glycemic control if needed). Moreover, CP-LVH was the strongest predictor of stroke in normotensives. Evaluation of CP-LVH may be useful for detecting normotensive diabetic subjects at greatest risk for stroke.
The specificity of ECG-LVH for detecting Echo-LVH is higher than its sensitivity,5 and it is difficult to apply the results of the present study to Echo-defined LVH. Okin et al22 reported that the use of ECG-LVH for detecting increased left ventricular mass index varies by race. Sundstrom et al23 reported that Echo-LVH and CP-LVH have independent predictive values for mortality. ECG-LVH (“electronic” LVH) is clearly an imperfect proxy measure of Echo-LVH (structural LVH) for the prediction of cardiovascular events. Finally, there are no data with which to determine whether the adjustment of Cornell voltage (6 mm) for women is appropriate for Japanese people.
Both CP-LVH and SL-LVH predicted future stroke in a Japanese population. The prediction of stroke by the CP-LVH is high even in normotensive subjects, in whom CP-LVH is related to the presence of diabetes mellitus and hyperlipidemia. In prehypertensives, CP-LVH predicts future stroke events, as do history of stroke, history of myocardial infarction, smoking status, and presence of diabetes mellitus.
Sources of Funding
This work was supported by a Grant-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology of Japan and grants from the Foundation for the Development of the Community (Tochigi, Japan).
J.I. is supported in part by a grant of Mitsubishi Pharma Research Foundation. The remaining authors report no conflicts.
- Received June 12, 2008.
- Revision received July 4, 2008.
- Accepted October 23, 2008.
Okin PM, Devereux RB, Jern S, Kjeldsen SE, Julius S, Nieminen MS, Snapinn S, Harris KE, Aurup P, Edelman JM, Wedel H, Lindholm LH, Dahlof B, for the LIFE Study Investigators. Regression of electrocardiographic left ventricular hypertrophy during antihypertensive treatment and the prediction of major cardiovascular events. JAMA. 2004; 292: 2343–2349.
Palmieri V, Bella JN, Arnett DK, Liu JE, Oberman A, Schuck M-Y, Kitzman DW, Hopkins PN, Morgan D, Rao DC, Devereux RB. Effect of type 2 diabetes mellitus on left ventricular geometry and systolic function in hypertensive subjects: Hypertension Genetic Epidemiology Network (HyperGEN) study. Circulation. 2001; 103: 102–107.
Bella JN, Devereux RB, Roman MJ, Palmieri V, Liu JE, Paranicas M, Welty TK, Lee ET, Fabsitz RR, Howard BV. Separate and joint effects of systemic hypertension and diabetes mellitus on left ventricular structure and function in American Indians (the Strong Heart Study). Am J Cardiol. 2001; 87: 1260–1265.
Okin PM, Devereux RB, Gerdts E, Snapinn SM, Harris KE, Jern S, Kjeldsen SE, Julius S, Edelman JM, Lindholm LH, Dahlof B, for the LIFE Study Investigators. Impact of diabetes mellitus on regression of electrocardiographic left ventricular hypertrophy and the prediction of outcome during antihypertensive therapy: the Losartan Intervention for Endpoint (LIFE) reduction in hypertension study. Circulation. 2006; 113: 1588–1596.
Ishikawa S, Kayaba K, Gotoh T, Nago N, Nakamura Y, Tsutsumi A, Kajii E. Incidence of total stroke, stroke subtypes, and myocardial infarction in the Japanese population: the JMS Cohort Study. J Epidemiol. 2008; 18: 144–150.
Sekine M, Gotoh E, Ochiai H, Umezu M, Ishii M. Office blood pressure in patients with essential hypertension. Ther Res. 1997; 18: 122.
Okin PM, Devereux RB, Jern S, Kjeldsen SE, Julius S, Dahlof B. Baseline characteristics in relation to electrocardiographic left ventricular hypertrophy in hypertensive patients: the Losartan Intervention for Endpoint reduction (LIFE) in hypertension study. Hypertension. 2000; 36: 766–773.
Eguchi K, Kario K, Hoshide S, Ishikawa J, Morinari M, Shimada K. Type 2 diabetes is associated with left ventricular concentric remodeling in hypertensive patients. Am J Hypertens. 2005; 18: 23–29.
Eguchi K, Ishikawa J, Hoshide S, Ishikawa S, Pickering TG, Schwartz JE, Homma S, Shimada K, Kario K. Differential impact of left ventricular mass and relative wall thickness on cardiovascular prognosis in diabetic and nondiabetic hypertensive subjects. Am Heart J. 2007; 154: 79.e9–e15.
Marwick TH. Diabetic heart disease. Heart. 2006; 92: 296–300.
Okin P, Wright J, Nieminen MS, Jern S, Taylor A, Phillips R, Papademetriou V, Clark L, Ofili E, Randall O, Oikarinen L, Viitasalo M, Toivonen L, Julius S, Dahlöf B, Devereux R. Ethnic differences in electrocardiographic criteria for left ventricular hypertrophy: the LIFE Study. Am J Hypertens. 2002; 15: 663–671.
Sundstrom J, Lind L, Arnlov J, Zethelius B, Andren B, Lithell HO. Echocardiographic and electrocardiographic diagnoses of left ventricular hypertrophy predict mortality independently of each other in a population of elderly men. Circulation. 2001; 103: 2346–2351.