Usefulness of T-axis deviation as an independent risk indicator for incident cardiac events in older men and women free from coronary heart disease (the Cardiovascular Health Study)∗

Usefulness of T-axis deviation as an independent risk indicator for incident cardiac events in older men and women free from coronary heart disease (the Cardiovascular Health Study)∗

Usefulness of T-Axis Deviation as an Independent Risk Indicator for Incident Cardiac Events in Older Men and Women Free from Coronary Heart Disease (T...

160KB Sizes 7 Downloads 19 Views

Usefulness of T-Axis Deviation as an Independent Risk Indicator for Incident Cardiac Events in Older Men and Women Free from Coronary Heart Disease (The Cardiovascular Health Study)* Pentti M. Rautaharju, MD, PhD, Jennifer Clark Nelson, PhD, Richard A. Kronmal, Zhu-Ming Zhang, MD, John Robbins, MD, MHS, John S. Gottdiener, MD, Curt D. Furberg, MD, PhD, Teri Manolio, PhD, and Linda Fried, MD, PhD

PhD,

T-axis shift has been reported to be an indicator of increased mortality risk. We evaluated the association of spatial T-axis deviation with incident coronary heart disease (CHD) events in older men and women free from clinically overt CHD. Spatial T-axis deviation was measured from the standard 12-lead electrocardiogram of a subgroup of 4,173 subjects considered free of CHD at baseline in the Cardiovascular Health Study, a prospective cohort study of risk factors for CHD and stroke in older men and women. Cox regression analysis was used to evaluate the association of altered repolarization with the risk of incident CHD events. The prevalence of marked T-axis deviation (>45ⴗ) was 12%. During the median follow-up of 7.4 years, there were 161 CHD deaths, 743 deaths from all causes, and 679 incident

CHD events. Adjusting for demographic and clinical risk factors, including other electrocardiographic abnormalities, there was a nearly twofold excess risk of CHD death, and approximately a 50% excess risk of incident CHD and all-cause mortality for those with marked Taxis deviation. From other electrocardiographic abnormalities, only QT prolongation was associated with excess risk for incident CHD comparable to that for abnormal T-axis deviation. These results suggest that T-axis deviation is an easily quantified marker for subclinical disease and an independent indicator for the risk of incident CHD events in older men and women free of CHD. 䊚2001 by Excerpta Medica, Inc. (Am J Cardiol 2001;88:118 –123)

he investigators of a European study in a cohort of older men and women with and without coronary T heart disease (CHD) recently reported that a simple

METHODS

indicator of T-wave abnormality, the T axis, was a stronger predictor of fatal and nonfatal cardiac events than any other electrocardiographic (ECG) abnormality or established cardiovascular risk indicator.1 We investigated the risk of CHD events associated with T-axis alterations in a subgroup of CHD-free participants from the Cardiovascular Health Study (CHS), a prospective cohort study of risk factors for CHD and stroke in older men and women. From the Department of Public Health Sciences, Wake Forest University School of Medicine, Winston-Salem, North Carolina; CHS Coordinating Center, University of Washington, Seattle, Washington; Department of Medicine, University of California Davis, Sacramento, California; Department of Cardiology, St. Francis Hospital, Roslyn, New York; the National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, Maryland; and the Johns Hopkins University, Baltimore, Maryland. This study was supported by Contracts N01-HC-85079 to N01-HC-85086 from the National Heart, Lung, and Blood Institute, Bethesda, Maryland. Manuscript received December 12, 2000; revised manuscript received and accepted February 12, 2001. Address for reprints: Pentti M. Rautaharju, MD, PhD, EPICARE Center, Suite 505, Piedmont Plaza Two, 2000 West First Street, Winston-Salem, North Carolina 27104. E-mail: prautaha@wfubmc. edu. *See Appendix for list of participating institutions and principal staff.

118

©2001 by Excerpta Medica, Inc. All rights reserved. The American Journal of Cardiology Vol. 88 July 15, 2001

Study population: CHS is a prospective cohort study of risk factors for CHD and stroke in older men and women2– 4 selected as random samples from Medicare eligibility lists in 4 field centers. Recruitment of 5,201 participants was completed in 1990. An additional sample of 687 African-Americans was recruited in 1992 to 1993 (year 4). Eligible participants included all noninstitutionalized persons aged ⱖ65 years who were living in the household of each person sampled from the Health Care Financing Administration lists. All participants gave informed consent and were expected to remain in the area for 3 years. Among those contacted and eligible, 57.3% were enrolled. Baseline examination: A home interview and a clinic examination were performed at the baseline visit. Standard questionnaires were used to assess a variety of risk factors including smoking, physical activity, and history of cardiovascular conditions. The self-reported conditions such as myocardial infarction were validated.2 Participants underwent clinic examination after a 12-hour overnight fast. The examination included blood pressure measurement (seated), ECG recording, and venipuncture. Blood samples were analyzed in a standardized fashion at the Central Blood Analysis Laboratory as previously described.2 Diabetes was defined by fasting glucose ⱖ126 mg/dl. Sys0002-9149/01/$–see front matter PII S0002-9149(01)01604-6

tolic and diastolic blood pressure ⱖ140 and 95 mm Hg, respectively, or taking antihypertensive medication defined hypertension. Carotid sonography was performed with the Toshiba SSA-270A sonographic units (Toshiba America Medical Systems, Tustin, California) and the recordings were read by the Ultrasound Reading Center.5 The participant was categorized into the CHD-free subgroup in the absence of prevalent CHD that was defined by the following conditions at baseline of the study: (1) confirmed definite history of past myocardial infarction; (2) confirmed definite past history of angina pectoris; (3) self-reported history of coronary artery bypass surgery; and (4) self-reported history of angioplasty. Electrocardiographic methods: The standard 12-lead resting electrocardiogram was recorded in all participants using MAC PC-DT ECG acquisition units (GE Marquette Medical Division, Milwaukee, Wisconsin), which followed a strictly standardized procedure for ECG acquisition and placement of chest electrodes using a HeartSquare device (Nova Heart Inc., Winston-Salem, North Carolina).6 Ten-second segments of the electrocardiograms, sampled at 250 samples per second per lead, were transmitted daily to the ECG Reading Center for analysis and classification using the Novacode ECG program (Nova Heart Inc.)7,8; this program performs ECG wave measurements and classifies electrocardiograms according to the Novacode8 and Minnesota Code.9 The program also includes algorithms for calculation of QT prolongation index and for estimating left ventricular mass index from ECG measurements and body weight.10,11 There were 5,752 ECG recordings available from baseline examination of the CHS cohort. A total of 210 electrocardiograms (3.7%) were excluded because of inadequate quality of ⱖ1 ECG lead needed for generation of the orthogonal XYZ leads and for calculation of various spatial T and QRS angles and for other ECG variables. In addition, 489 electrocardiograms (8.5%) with complete bundle branch blocks (QRS duration ⱖ120 ms) were excluded because these conditions will cause secondary T-wave changes due to altered ventricular depolarization sequence. This left 5,053 valid recordings for evaluation. In this study, we present data from analyses in a subgroup of 4,173 participants (83%) considered free of CHD at baseline. From this study group, 123 subjects were excluded from some of the multivariate analyses due to missing covariates. Definition of angular variables: Eight standard ECG leads were first transformed into the orthogonal leads using the inverse Dower transform matrix.12 The integrated values of the XYZ lead amplitudes of QRS and T complexes were then used to determine the following spatial angles in the frontal (XY) and horizontal (XZ) plane (Figure 1): 1. Alpha, the mean frontal plane T axis (0° left, 90° down, ⫺90° up, and ⫾ 180 right, referring to the chest of a standing subject). 2. Beta, the T-axis azimuth or the mean horizontal

FIGURE 1. Schematic of reference angles for T-axis deviations: (1) frontal (XY) plane axis ␣ (Alpha) (0ⴗ left, ⴙ90 down, ⴚ90ⴗ up, ⴞ 180ⴗ right); (2) horizontal (XZ) plane axis or azimuth ␤ (Beta) (0ⴗ left, ⴙ90ⴗ front, ⴚ90ⴗ back, ⴞ 180ⴗ right), and spatial angle ␭ (Lambda)— deviation from the reference direction Vref (␣ 45ⴗ and ␤ 45ⴗ). Normal (apicobasal) repolarization sequence is opposite the Vref direction. Anterior shift of T axis (solid arrow) took place in two thirds of marked axis deviations, and the remaining one third were posterior shifts (broken arrow).

plane T axis (0° left, 90° front, ⫺90° back, ⫾ 180° right). 3. Epsilon, the T-axis elevation (0° down along the positive Y axis, 90° horizontal, 180° up). 4. Lambda, the spatial T-axis deviation from the normal reference direction (x ⫽ 1/公3, y ⫽ 1/公3, and z ⫽ ⫺1/公3, where x, y, and z are the unit vector components in X, Y, and –Z directions). Thus, this reference direction is 45° anteriorly, with a 45° angle from the ⫹Y axis in the frontal and sagittal plane projections. 5. Phi, the mean frontal plane QRS axis. 6. Rho, the QRS axis azimuth or the mean horizontal plane QRS axis. The reference directions for ␾ and ␳ are the same as for T-axis ␣ and ␤. Follow-up and classification of events: Participants returned to annual follow-up clinic examinations. They were also contacted between the clinic examinations by telephone interviews and asked about cardiovascular events and all hospitalizations. Discharge summaries and diagnoses were obtained from all hospitalizations. Hospitalization records and additional information were collected for all potential incident cardiovascular events, including cardiac enzymes and serial electrocardiograms. Incident CHD was defined by the occurrence of the following events during the follow-up period: new angina pectoris, incident myocardial infarction, coronary artery bypass surgery or angioplasty, and CHD death. The algorithms for classification of myocardial infarction based on a decision matrix containing elements of chest pain, cardiac enzymes, and serial ECG changes have been published elsewhere.13

CORONARY ARTERY DISEASE/T AXIS, SUBCLINICAL DISEASE, AND CHD RISK

119

TABLE 1 Characteristics of the Study Group by T-Axis Category T-Axis Category

Characteristics Age (yrs) Women Black Body mass index (kg/m2) Systolic blood pressure (mm Hg) Diastolic blood pressure (mm Hg) Hypertension Diabetes mellitus Current smoker Total cholesterol (mg/dl) High-density lipoprotein (HDL) cholesterol Total cholesterol/HDL cholesterol ratio Internal carotid wall (mm) T-axis (XY plane)* T axis (XZ plane)* T-axis elevation* QRS axis (XY plane)* QRS axis (XZ plane)* QT index Left ventricular mass index Silent myocardial infarction† ST depression‡ T-wave inversion§

All (n ⫽ 4,173)

Normal (n ⫽ 2,665) (64%)

72 ⫾ 6 62% 15% 27 ⫾ 5 136 ⫾ 21 71 ⫾ 11 41% 14% 13% 212 ⫾ 39 56 ⫾ 6

72 ⫾ 5 65% 13% 26 ⫾ 5 134 ⫾ 20 70 ⫾ 11 37% 13% 12% 214 ⫾ 39 56 ⫾ 16

4.1 ⫾ 1.23 1.4 24 45 72 15 ⫺29 102 102

⫾ 0.54 ⫾ 39 ⫾ 32 ⫾ 15 ⫾ 28 ⫾ 26 ⫾ 5% ⫾ 10% 4% 5% 8%

Borderline Deviation (n ⫽ 997) (24%) 73 ⫾ 6 54% 16% 27 ⫾ 4 138 ⫾ 21 72 ⫾ 11 43% 14% 12% 210 ⫾ 39 55 ⫾ 16

4.1 ⫾ 0.23 1.3 30 42 67 18 ⫺27 102 100

⫾ 0.52 ⫾ 13 ⫾ 14 ⫾ 10 ⫾ 26 ⫾ 25 ⫾ 5% ⫾ 8% 2% 1% 2%

4.1 ⫾ 1.23 1.4 18 56 79 10 ⫺31 102 104

⫾ 0.55 ⫾ 62 ⫾ 44 ⫾ 16 ⫾ 30 ⫾ 27 ⫾ 5% ⫾ 10% 4% 4% 6%

Marked Deviation (n ⫽ 511) (12%) 74 ⫾ 6 61% 28% 27 ⫾ 5 144 ⫾ 23 73 ⫾ 12 59% 19% 15% 211 ⫾ 40 54 ⫾ 16 4.2 ⫾ 1.24 ⫾ 0.59 ⫾ 101 ⫾ 78 ⫾ 24 ⫾ 34 ⫾ 30 ⫾ 5% ⫾ 15% 9% 25% 46%

1.5 3 59 85 12 ⫺33 102 109

p Value ⬍0.01 ⬍0.01 ⬍0.01 ⬍0.01 ⬍0.01 ⬍0.01 ⬍0.01 ⬍0.01 NS 0.05 0.02 NS ⬍0.01 ⬍0.01 ⬍0.01 ⬍0.01 ⬍0.01 ⬍0.01 NS ⬍0.01 ⬍0.01 ⬍0.01 ⬍0.01

*Frontal plane (XY) axis (0° left, 90° down, ⫺90° up, ⫾180° right), horizontal (XZ) plane axis (0° left, 90° front, ⫺90° back, ⫾180° right), elevation (0° down, 90° horizontal, 180° up). † Minnesota code 1.1, 1.2, or (1.3 and 5.1 or 5.2). ‡ Minnesota code 4.1 or 4.2. § Minnesota code 5.1 or 5.2. Values are expressed as mean ⫾ SD or percentages. For other definitions, see Methods section.

Statistical analyses: All angular variables were statistically examined in their continuous forms for their strength of association with cardiovascular events. Analyses and results presented in this study are for the T-axis deviation angle (␭), which consistently showed the strongest association with increased risk across all events. For ease of interpretation, the study group was categorized into 3 subgroups: (1) normal T axis or the reference group (␭ ⬍30°); (2) borderline deviation (␭ ⱖ30° to 44°); and (3) marked deviation (␭ ⱖ45°). Chi-square tests of independence and analysis of variance methods were used, respectively, to compare categorical and continuous baseline study characteristics between T-axis deviation groups. Kaplan-Meier estimates were computed and displayed by T-axis deviation groups to explore the form of the survivorship functions over time for cardiovascular events. Estimated curves were compared using unweighted log-rank tests. Cox proportional-hazards regression models were constructed to examine the bivariate and multivariate associations between T-axis deviation and cardiovascular events. Diagnostic procedures were performed to check for the model fit and assumptions. In addition, the functional form for T-axis deviation (and 120 THE AMERICAN JOURNAL OF CARDIOLOGY姞

VOL. 88

other continuous variables) was explored by fitting the variable as a smooth function (using splines). Continuous forms were subsequently approximated using categorical terms. Interactions between T-axis deviation and other covariates and between gender and other covariates were tested for significance using nested likelihood ratio tests. These analyses were performed using Stata, release 6.0 (Stata Corp., College Station, Texas),14 S-Plus, version 3.3 (Mathsoft Inc., Cambridge, Massachusetts),15 and SPSS (SPSS Inc., Chicago, Illinois)16 software packages.

RESULTS

Characteristics of study population: The prevalence of marked T-axis deviation was 12% and of borderline T-axis deviation 24% (Table 1). There were several statistically significant differences between the T-axis groups. The proportion of African-Americans, the proportion of hypertensives, and the level of systolic and diastolic blood pressure were higher in the borderline and abnormal T-axis groups than in the normal reference group. Similarly, the proportion of diabetics was significantly higher (19%) in the marked deviation group than in the 2 lower T-axis groups (13% and 14%, respectively), and the internal carotid intimaJULY 15, 2001

TABLE 2 Cardiovascular Event and All-Cause Mortality Rates (per 1,000 person-years) With 95% Confidence Intervals (CI) by TAxis Group Among Men and Women Free from CHD T-Axis Category Normal T Axis (n ⫽ 2,665) (64%)

All (n ⫽ 4,173) Event Category Incident CHD Women Men All CHD death Women Men All All-cause death Women Men All

Borderline Deviation (n ⫽ 997) (24%)

Marked Deviation (n ⫽ 511) (12%)

No. at Risk

No. of Events

Event Rate (CI)

No. of Events

Event Rate (CI)

No. of Events

Event Rate (CI)

No. of Events

Event Rate (CI)

2,576 1,597 4,173

329 350 679

19.9 (17.8, 22.1) 37.3 (33.6, 41.4) 26.2 (24.3, 28.2)

196 181 377

17.2 (15.0, 19.8) 31.4 (27.1, 36.3) 22.0 (19.9, 24.3)

64 107 171

18.8 (14.7, 24.0) 40.7 (33.7, 49.2) 28.3 (24.4, 32.9)

69 62 131

38.9 (30.7, 49.2) 62.4 (48.6, 80.0) 47.3 (39.9, 56.1)

2,576 1,597 4,173

72 89 161

4.1 (3.3, 5.2) 8.6 (7.0, 10.6) 5.8 (5.0, 6.7)

21 21 42

10.9 (7.1, 16.8) 18.5 (12.1, 28.4) 13.7 (10.2, 18.6)

2,576 1,597 4,173

371 372 743

21.3 (19.3, 23.6) 35.9 (32.4, 39.7) 26.7 (24.9, 28.7)

76 70 146

39.6 (31.6, 49.5) 61.7 (48.8, 78.0) 47.8 (40.6, 56.2)

36 40 76 214 192 406

3.0 (2.2, 4.2) 6.4 (4.7, 8.7) 4.2 (3.3, 5.2) 17.9 (15.7, 20.5) 30.6 (26.6, 35.3) 22.3 (20.2, 24.6)

media thickness increased progressively with increasing T-axis deviation. In terms of ECG abnormalities indicating repolarization abnormalities, the prevalence of ST depression and T-wave inversion was low in the normal and borderline T-axis groups (1% and 4% for ST depression, and 2% and 6% for T-wave inversion, respectively). In the marked T-axis group, 25% of the subjects had ST depression ⱖ50 ␮V and nearly 50% had T-wave inversion. The percentage of subjects with possible silent myocardial infarction by electrocardiography increased from 2% in the reference group to 9% in the marked deviation group. The mean left ventricular mass index was significantly higher with increasing severity of T-axis deviation. The differences cited for all above ECG variables were significant at a p level ⬍0.01. The differences between the T-axis groups in the mean QT index were not significant. With increasing spatial T-axis deviation, the spatial T axis shifted progressively closer to the horizontal plane, and with marked T-axis deviation the T-axis elevation angle reached 85°. QRS axis orientation differences between the T-axis groups were minor. The QRS axis remained approximately 30° posteriorly, and its orientation in the frontal plane shifted slightly to the left (more horizontal) with increasing T-axis deviation. Compared with the pronounced spatial T-axis orientation differences, the relatively unchanged QRS axis, even with marked T-axis deviation, supports the assertion that the repolarization changes with T-axis deviation reflect primary repolarization abnormalities. Event rates and T-axis deviation: During the median follow-up of 7.4 years, there were 161 CHD deaths, 679 incident CHD events, and 743 deaths from all causes in the study group. CHD death rates showed increasing rates by T-axis group (Table 2), although the increase was most pronounced in the marked Taxis group. The increase was similar both in men and in women, although the overall rates in men were

15 28 43 81 110 191

4.2 (2.5, 6.9) 9.4 (6.5, 13.6) 6.5 (4.9, 8.8) 22.5 (18.1, 28.0) 37.0 (30.7, 44.6) 29.1 (25.2, 33.5)

considerably higher than in women. The KaplanMeier curves (Figure 2) reflect similar trends as shown by the mortality rates. The pairwise differences between the survivor curves for the normal T-axis and marked T-axis deviation groups were significant for incident CHD and CHD death (p ⬍0.001 for both); pairwise differences between the normal and borderline T-axis groups were also significant for both incident CHD (p ⫽ 0.0053) and CHD (p ⫽ 0.0153) death. The hazard ratios indicate that with adjustment for age, gender, and race, marked T-axis deviation was associated with a twofold excess risk for incident CHD and with nearly a threefold excess risk for CHD death (Table 3, model A). After additional adjustment for demographic and clinical factors, for myocardial infarction by electrocardiogram, QT index, ST depression, left ventricular mass index, and frontal plane QRS axis, the hazard ratio was 1.47 (95% confidence intervals 1.16, 1.87) for incident CHD and 1.88 (95% confidence intervals 1.19, 2.98) for CHD death (Table 3, model B). When tested in additional models (not shown), the hazard ratios for none of these other ECG variables were significant except for QT prolongation (QT index ⱖ108%), which had hazard ratios comparable to those for marked T-axis deviation. The inclusion of T-wave inversion in the multivariate model did not significantly alter the hazard ratios for T-axis deviation. The study group included 147 subjects (3.6%) with possible silent myocardial infarction at baseline by the Minnesota Code criteria (codes 1.1 or 1.2, or code 1.3 with code 5.1 or 5.2). The exclusion of this small subgroup produced only minor alterations in the hazard ratios for T-axis deviation (hazard ratio for marked T-axis deviation in model B changed from 1.49 to 1.46 for total mortality, from 1.47 to 1.46 for incident CHD, and from 1.88 to 1.81 for CHD death). T-axis shifts and standard 12-lead electrocardiogram: Although an abnormal spatial T axis is a readily

derived quantity using computer algorithms, the question remains whether this information can be extracted

CORONARY ARTERY DISEASE/T AXIS, SUBCLINICAL DISEASE, AND CHD RISK

121

FIGURE 2. Kaplan-Meier survival curves for CHD and all-cause mortality by T-axis deviation in the CHD-free group. Proportion indicates the estimated fraction with no CHD death or alive, respectively, for CHD and all-cause death curves. The curves for incident CHD (not shown) were closely similar to those for all-cause death.

TABLE 3 Hazard Ratios With 95% Confidence Intervals (CI) for Categorized TAxis Deviation for Incident CHD, CHD Death, and All-Cause Death in Men and Women Free from CHD Model A*

Incident CHD T-axis normal (⬍30°) Marginal deviation (30° to 44°) Marked deviation (ⱖ45°) CHD death T-axis normal (⬍30°) Marginal deviation (30° to 44°) Marked deviation (ⱖ45°) All-cause death T-axis normal (⬍30°) Marginal deviation (30° to 44°) Marked deviation (ⱖ45°)

Model B*

Hazard Ratio

(CI)

Hazard Ratio

(CI)

1.00 1.20 2.10

— (1.00, 1.44) (1.72, 2.57)

1.00 1.09 1.58

— (0.90, 1.31) (1.25, 1.99)

1.00 1.41 2.88

(0.97, 2.06) (1.96, 4.23)

1.00 1.20 1.98

(0.82, 1.78) (1.28, 3.08)

1.00 1.21 1.93

(1.01, 1.44) (1.59, 2.35)

1.00 1.12 1.49

(0.93, 1.34) (1.19, 1.88)

of a simple set of visual criteria for abnormal T-axis shift, namely T in V1 positive ⱖ1 mm (100 ␮V) or T in V6 negative ⱖ0.5 mm (50 ␮V) (Table 4). The prevalence of abnormal repolarization defined by these criteria was very high (43%), and was associated with a nearly 50% excess risk of CHD and a nearly 40% excess risk of incident CHD, with the adjustment for demographic and clinical factors and other ECG abnormalities.

DISCUSSION

The results from the present study in CHD-free men and women sup*Model A, adjusted for age, gender, and race; Model B, adjusted for demographic and clinical port the main findings in the Rotterfactors, silent myocardial infarction, QT index, ST depression, left ventricular mass, and frontal plane dam study population that included QRS axis. subjects with CHD. T-axis deviation appears to be a subclinical sign of by visual inspection of the standard 12-lead electro- cardiac abnormality and a significant marker for excardiogram. T-axis rotation anterior and to the right cess risk of future CHD events. Primary repolarization abnormalities can be excan be expected to result in gradually increasing T amplitude in lead V1 and a decreasing T amplitude in pected to alter T axis or QT interval, or both. Although lead V6. We tested potential value in risk identification action potential duration prolongation in the ventric122 THE AMERICAN JOURNAL OF CARDIOLOGY姞

VOL. 88

JULY 15, 2001

TABLE 4 Hazard Ratios With 95% Confidence Intervals (CI) for Incident CHD, CHD Death, and All-Cause Death for Abnormal T-Wave Patterns by Simple Visual Criteria* in Men and Women Free from CHD Model A† Hazard Ratio Incident CHD T in V1 and V6 normal T in V1 or V6 abnormal* CHD death T in V1 and V6 normal T in V1 or V6 abnormal All-cause death T in V1 and V6 normal T in V1 or V6 abnormal

Model B†

(CI)

Hazard Ratio

(CI)

1.00 1.61

— (1.37, 1.88)

1.39

(1.18, 1.64)

1.00 1.79

— (1.28, 2.48)

1.00 1.47

— (1.04, 2.07)

1.00 130

— (1.12, 1.51)

1.00 1.15

— (0.98, 1.34)

*T in V1 positive (ⱖ100 ␮V) or T in V6 negative (⬍50 ␮V) (prevalence 43%). † Model A, adjusted for age, gender and race; Model B, adjusted for demographic and clinical factors, silent myocardial infarction, QT index, ST depression, left ventricular mass index, and frontal plane QRS axis.

Elaine Meilahn, Peg Meyer, Roberta Moyer, Anne Newman, Richard Schulz, Vivienne E. Smith, Sidney K. Wolfson; University of California, Irvine—Echocardiography Reading Center (baseline): Hoda AntonCulver, Julius M. Gardin, Margaret Knoll, Tom Kurosaki, Nathan Wong; Georgetown Medical Center— Echocardiography Reading Center (Follow-Up): John Gottdiener, Eva Hausner, Stephen Kraus, Judy Gay, Sue Livengood, Mary Ann Yohe, Retha Webb; New England Medical Center, Boston—Ultrasound Reading Center: Daniel H. O’Leary, Joseph F. Polak, Laurie Funk; University of Vermont—Central Blood Analysis Laboratory: Elaine Cornell, Mary Cushman, Russell P. Tracy; University of Arizona-Tucson Pulmonary Reading Center: Paul Enright; University of Washington, Seattle—Coordinating Center: Alice Arnold, Annette L. Fitzpatrick, Richard A. Kronmal, Bruce M. Psaty, David S. Siscovick, Will Longstreth, Patricia W. Wahl, David Yanez, Paula Diehr, Corrine Dulberg, Bonnie Lind, Thomas Lumley, Ellen O’Meara, Jennifer Nelson, Chuck Spiekerman; NHLBI Project Office: Robin Boineau, Teri A. Manolio, Peter J. Savage, Patricia Smith.

1. Kors JA, de Bruyne MC, Hoes AW, van Herpen G,

ular region repolarizing last (or delayed excitation of that region) will prolong the QT interval, action potential duration shortening or prolongation in any ventricular region can be expected to influence T axis to a variable degree. T-axis deviation can thus be expected to be a sensitive indicator of sustained derailed ionic mechanisms. Very few epidemiologic studies have evaluated the prognostic value of ST and T-wave abnormalities as separate entities. The independent prognostic value of isolated T-wave abnormalities without ST findings has generally been found to be not significant17 or only of borderline significance.18 Importantly, abnormal spatial T-axis shifts do not necessarily result in abnormal (negative) T waves in the standard 12-lead electrocardiogram by conventional criteria. For instance, we observed that a positive T wave in V1 (ⱖ1 mm), not considered abnormal by the conventional clinical criteria, may convey potentially important risk information. Derivation of optimal visual criteria for extraction from the standard 12-lead electrocardiogram of the counterpart of risk information contained in the spatial T-axis shift deserves attention in future studies.

APPENDIX Participating Institutions and Principal Staff: Forsyth County, NC— Wake Forest University School of Medicine: Gregory L. Burke, Sharon Jackson, Alan Elster, Curt D. Furberg, Gerardo Heiss, Dalane Kitzman, Margie Lamb, David S. Lefkowitz, Mary F. Lyles, Cathy Nunn, Ward Riley, John Chen, Beverly Tucker; Winston-Salem, NC—Wake Forest University School of Medicine–ECG Reading Center: DeDe Pemberton, Ron Prineas, Farida Rautaharju, Pentti Rautaharju, Zhu-Ming Zhang; Sacramento County, CA—University of California, Davis: William Bonekat, Charles Bernick, Michael Buonocore, Mary Haan, Calvin Hirsch, Lawrence Laslett, Marshall Lee, John Robbins, William Seavey, Richard White; Washington County, MD—The Johns Hopkins University: M. Jan Busby-Whitehead, Joyce Chabot, George W. Comstock, Adrian Dobs, Linda P. Fried, Joel G. Hill, Steven J. Kittner, Shiriki Kumanyika, David Levine, Joao A. Lima, Neil R. Powe, Thomas R. Price, Jeff Williamson, Moyses Szklo, Melvyn Tockman; Washington County, MD—The Johns Hopkins University –MRI Reading Center: Norman Beauchamp, R. Nick Bryan, Douglas Fellows, Melanie Hawkins, Patrice Holtz, Naiyer Iman, Michael Kraut, Cynthia Quinn, Grace Lee, Carolyn C. Meltzer, Larry Schertz, Earl P. Steinberg, Scott Wells, Linda Wilkins, Nancy C. Yue; Allegheny County, PA—University of Pittsburgh: Diane G. Ives, Charles A. Jungreis, Laurie Knepper, Lewis H. Kuller,

Hofman A, van Bemmel JH, Grobbee DE. T axis as an independent indicator of risk of cardiac events in elderly people. Lancet 1998;352:601– 605. 2. Fried LP, Borhani, NO, Enright P, Furberg CD, Gardin JM, Kronmal RA, Kuller LH, Manolio TA, Mittelmark MB, Newman A, O’Leary DH, Psaty BM, Rautaharju P, Tracy RP, Weiler PG. The Cardiovascular Health Study: design and rationale. Ann Epidemiol 1991;1:263–276. 3. Furberg CD, Manolio TA, Psaty BM, Bild DE, Borhani NO, Newman A, Tabatznik B, Rautaharju PM. Major ECG abnormalities in persons aged 65 years and older (The Cardiovascular Health Study). Am J Cardiol 1992;69:1329 –1335. 4. Psaty BM, Kuller LH, Bild D, Burke GL, Kittner SJ, Mittelmark M, Price TR, Rautaharju PM, Robbins J. Methods of assessing prevalent cardiovascular disease in the Cardiovascular Health Study. Ann Epidemiol 1995;5:270 –277. 5. O’Leary DH, Polak JF, Wolfson SK Jr, Bond MG, Bommer W, Sheth S, Psaty BM, Sharrett AR, Manolio T. The use of sonography to evaluate carotid atherosclerosis in the elderly: The Cardiovascular Health Study. Stroke 1991;22:1155–1163. 6. Rautaharju PM, Park L, Rautaharju FS, Crow R. A standardized procedure for locating and documenting ECG chest electrode positions. Consideration of the effect of breast tissue on ECG amplitudes in women. J Electrocardiol 1998;31:17–29. 7. Rautaharju PM, MacInnes PJ, Warren JW, Wolf HK, Rykers PM, Calhoun HP. Methodology of ECG interpretation in the Dalhousie Program. Novacode ECG classification procedures for clinical trials and population studies. Methods Inf Med 1990;29:362–374. 8. Rautaharju PM, Calhoun HP, Chaitman BR. Novacode serial ECG classification system for clinical trials and epidemiological studies. J Electrocardiol 1992;24:163–172. 9. Blackburn H, Keys A, Simonson E, Rautaharju PM, Punsar S. The electrocardiogram in population studies. A classification system. Circulation 1960;21: 1160 –1175. 10. Rautaharju PM, Warren JW, Calhoun HP. Estimation of QT prolongation: a persistent, avoidable error in computer electrocardiography. J Electrocardiol 1991;23:111–117. 11. Rautaharju PM, Park LP, Siscovick D, Boineau R, Gottdiener JS, Smith V, Powe NR. Race- and gender-specific ECG models for left ventricular mass in older populations. Factors influencing overestimation of left ventricular hypertrophy by ECG prevalence by ECG criteria in African-Americans. J Electrocardiol 2000;33:205–213. 12. Edenbrandt L, Pahlm, O. Vectorcardiogram synthesized from a 12-lead ECG: superiority of the inverse Dower transform matrix. J Electrocardiol 1988;21: 361–367. 13. Ives DG, Fitzpatrick AL, Bild DE, Psaty BM, Kuller LH, Crowley PM, Cruise RG, Theroux S. Surveillance and ascertainment of cardiovascular events: The Cardiovascular Health Study. Ann Epidemiol 1995;5:278 –285. 14. Stata Corporation, Statistical Software. Release 6.0. College Station, TX: Stata Corporation. 1999. 15. Statistical Sciences, S-Plus Guide to Statistical, and Mathematical Analysis, Version 3.3, Seattle: StatSci, a division of MathSoft, Inc., 1995. 16. SPSS Inc., SPSS 6.1 Syntax Reference Guide, Chicago: SPSS Inc., 1994. 17. Tervahauta M, Pekkanen J, Punsar S, Nissinen A. Resting electrocardiographic abnormalities as predictors of coronary events and total mortality among elderly men. Am J Med 1996;100:641– 645. 18. Bacquer D, de Backer G, Kornitzer M, Muny K, Doyen Z, Blackburn H. Prognostic value of ischemic electrocardiographic findings for cardiovascular mortality in men and women. J Am Coll Cardiol 1998;32:680 – 685.

CORONARY ARTERY DISEASE/T AXIS, SUBCLINICAL DISEASE, AND CHD RISK

123