T-loop morphology as a marker of cardiac events in the elderly

Journal of Electrocardiology Vol. 31 Supplement

T - L o o p M o r p h o l o g y as a Marker of Cardiac E v e n t s in t h e Elderly

J a n A. K o r s , P h D , * M a r t i n e

C. d e B r u y n e ,

MD, PhD,*J-

Arno W. Hoes, MD, PhD,t{

Gerard van Herpen,

MD, PhD,*

Albert Hofman,

J a n H. v a n B e m m e l ,

PhD,* and

MD, PhD,t

D i e d e r i c k E. G r o b b e e , M D , P h D t $

Abstract: ST-T wave changes of electrocardiographic (ECG) leads have long been recognized as predictors of future cardiac events, but they only imperfectly characterize T-loop morphology. Using vectorcardiographic (VCG) parameters, we investigated the predictive value of T-loop abnormality for fatal and nonfatal cardiac events in a prospective cohort study among 5,815 elderly. Separately, the predictive value of an easily obtainable T-loop parameter, the T axis, was also assessed. Measurements were determined by a computer program, using VCGs reconstructed from the standard 12-lead ECGs. During the 3 to 6 (mean 4) years of follow-up, 166 fatal and 193 nonfatal cardiac events occurred. Subjects with an abnormal T-loop morphology had increased risks for fatal cardiac events (hazard ratio 4.3; 95% CI 3.0-6.4) and nonfatal cardiac events (3.0; 1.9-4.8). Risks associated with an abnormal T axis alone were only slightly lower. Additional adjustment for established cardiovascular risk indicators resulted in lower, but still highly significant risks. Both T-loop and T-axis abnormalities appear to be strong, independent risk indicators of cardiac events in the elderly. K e y words: electrocardiography, repolarization disturbance, risk factors, coronary heart disease, computerized analysis.

Repolarization abnormalities as manifested by ST depression and T-wave inversion of electrocardiographic (ECG) leads have long been recognized as predictors of future fatal and nonfatal cardiac events (1-8). These parameters, however, only

imperfectly characterize the spatial T loop inscribed by the heart vector, so that possibly important prognostic information m a y be missed. While other ECG parameters might be marshaled to provide a more comprehensive picture of T-loop m o r p h o l o g y ,

From the *Department of Medical Informatics; and the 1-Department of Epidemiology & Biostatistics, Erasmus University Medical School, Rotterdam, The Netherlands; and ~the Julius Center for Patient-Oriented Research, Utrecht University, The Netherlands.

Supported by the Netherlands Institute for Health Sciences. The Rotterdam Study is supported by grants from several institutions, including the Municipality of Rotterdam, the NESTORprogram for research in the elderly (supported by the Netherlands Ministries of Health and Education), the Netherlands Heart Foundation, the Netherlands Prevention Fund, and the Rotterdam Medical Research Foundation (ROMERES). Reprint requests: Jan A. Kors, PhD, Department of Medical Informatics, Erasmus University Medical School, PO Box 1738, 3000 DR Rotterdam, The Netherlands. Copyright © 1998 by Churchill Livingstone ®

0022-0736/98/310S- 10l I$5.00/0 54

T-Loop Morphology as a Marker of Cardiac Events

its description and interpretation are m u c h simpler using vectorcardiographic (VCG) parameters. Two reasons can be given. First, the information in the ECG is contained in 12 leads that exhibit a large degree of redundancy, whereas in the VCG basically the same information is compressed into three orthogonal leads. Second, temporal relationships b e t w e e n leads are m u c h easier visualized in the VCG t h a n in the ECG and facilitate the design of diagnostic criteria. In this study, we investigated w h e t h e r an overall measure of T-loop abnormality, based on VCG parameters, is a m a r k e r for fatal or nonfatal cardiac events in older adults. Because determination of an overall measure m a y present practical difficulties, we also studied the predictive value of one easily obtainable T-loop parameter: the T axis.

Materials and Methods Study Population and Baseline Data Collection The present investigation is part of the Rotterdam Study, a population-based cohort study aimed at assessing the occurrence of and risk indicators for chronic disease in the elderly. Objectives and m e t h ods of the Rotterdam Study have been described in detail elsewhere (9). Briefly, in the Rotterdam Study, all m e n and w o m e n aged 55 years or older, living in the Rotterdam district of Ommoord, were invited to participate (response rate 78 %). Of 7,129 participants, baseline data were collected over the years 1990 to 1993, including established cardiovascular risk indicators, use of medications, history of cardiovascular disease, and an ECG. All participants gave their written informed consent. Of the participants, 85% were enlisted in computerized practices of general practitioners (GPs) working in the study district of O m m o o r d . The remaining 15 % were u n d e r the care of GPs outside the O m m o o r d area. Hypertension was defined as systolic blood pressure above 160 m m Hg or diastolic blood pressure above 95 m m Hg or use of antihypertensive medication for the indication hypertension. Diabetes mellitus was defined as a nonfasting blood glucose level above 11.1 m m o l / L or use of antidiabetic medication. History of myocardial infarction was defined as self-reported myocardial infarction with hospital admission. Presence of angina pectoris was established t h r o u g h the Rose questionnaire (10). Digitally stored ECGs of 6,160 (86%) participants were available. Missing ECGs were mainly

•

Kors et al.

55

due to t e m p o r a r y technical problems with the ECG recording. After exclusion of 345 subjects w i t h o u t follow-up data, mainly due to u n k n o w n addresses, the study population consisted of 2,360 m e n and 3,455 w o m e n .

Follow-Up Procedures The follow-up period, starting at the baseline examination and for the present analysis lasting until April 1996, was 3 to 6 (mean 4) years. Information on fatal and nonfatal endpoints of the participants was obtained from the GPs and from the municipal health authorities in Rotterdam at regular intervals. Complete follow-up information was available for 94% of the population. Classification of fatal and nonfatal events was based on the 10th revision of the International Classification of Diseases (ICD-10) (11). We defined cardiac mortality as death from myocardial infarction or other heart diseases (ICD-10:I2 I-28, 42, 43, 4 6 - 5 0 ) , or sudden cardiac death. Sudden cardiac death was defined as death occurring within 1 h o u r after onset of symptoms, or unwitnessed death w h e r e a cardiac cause could not be excluded (12,13). Nonfatal cardiac events were defined as myocardial infarction or chronic ischemic heart disease (I21-25), coronary artery bypass graft, or p e r c u t a n e o u s transluminal coronary angioplasty (no ICD-10 codes). All events were classified i n d e p e n d e n t l y by two research physicians. If there was disagreement, a consensus was reached in a separate session. Finally, all events were verified by a medical expert in the field of cardiovascular disease. In case the expert disagreed with the research physicians, the expert's j u d g m e n t was considered final. The research physicians and the expert based their decisions on the same source data.

ECG Measurements and Interpretation A 12-lead resting ECG was recorded with an ACTA electrocardiograph (ESAOTE, Florence) at a sampling frequency of 500 Hz, and stored digitally. All ECGs were processed by the Modular ECG Analysis System (MEANS) (14). MEANS computes a representative averaged beat for each of the 12 leads from which ECG m e a s u r e m e n t s and a diagnostic interpretation are derived. The program is also able to perform a classification according to the Minnesota Code (15). MEANS has b e e n extensively evaluated, both by the developers themselves (14,16) and by others (17,18).

56

Journal of Electrocardiology Vol. 31 Supplement

Table l. Baseline Characteristics of All Participants and According to Three Categories of T-Loop Morphology* Characteristic

All (n = 5,815)

Normal (n = 4,967)

Borderline (n = 534)

Abnormal (n = 314)

Age (yr) Female sex (%) Systolic blood pressure (mm Fig) Diastolic blood pressure (mm Hg) Body mass index (kg/m2) TC/HDLt ratio Current smoking (%) Use of cardiovascular medication (%) Hypertension (%) Diabetes mellitus (%) History of angina pectoris (%) History of myocardial infarction (%) T axis (degrees) ST depression (%) T-wave inversion (%)

69.3 + 9.0 59.4 139.4 _+22.4 73.5 + 11.6 26.3 -+ 3.7 5.2 _+ 1.6 21.0 36.1 29.6 12.7 6.8 7.1 38.0 + 39.2 10.4 8.2

68.4 60.8 138.6 73.3 26.3 5.2 21.0 32.8 28.2 11.6 5.9 4.9 37.2 4.9 2.6

73.0 49.3 145.0 75.0 26.5 5.4 20.I 49.6 39.4 17.4 11.6 20.8 42.0 35.4 31.9

76.8 54.0 143.8 73.8 26.1 5.6 21.8 63.9 36.i 21.2 12.5 19.9 43.8 54.4 54.8

Category-specificvalues have been adjusted for age and sex using one-way analysis of covariance. * Plus-minus values are means _+ SD. t TC = total cholesterol (retool/L), HDL = high-density lipoprotein cholesterol (retool/L).

After reconstructing VCG X, Y, and Z leads from the standard ECG leads (19,20), the following Tloop parameters were computed: m a x i m u m spatial amplitude, width and sense of inscription of the T loop in the horizontal plane, direction of the m e a n T axis in the frontal and horizontal planes, and direction and m a g n i t u d e of the J-point displacem e n t in the two planes. Using these parameters, a decision tree (2I) t h e n classified a given T loop as "normal," "borderline," or "abnormal." Based on previously published reports (22-25), the frontal T axis was classified into three groups: "normal" (15 ° to 75°), "borderline" ( - 1 5 ° to 15 ° and 75 ° to 105°), and "abnormal" ( - 1 8 0 ° to - 1 5 ° and 105 ° to 180°). ST depression is taken as Minnesota Code 4.1 or 4.2. T-wave inversion is defined as Minnesota Code 5.1 or 5.2.

Data Analysis Using o n e - w a y analysis of covariance and adjusting for age and sex w h e n appropriate, differences in baseline characteristics b e t w e e n those with and the 345 w i t h o u t follow-up data, and b e t w e e n those with normal, borderline, and abnormal T-loop morp h o l o g y were examined. Cox's proportional hazards analysis (26) was used to determine the risk for cardiac death, sudden cardiac death, nonfatal cardiac events, and fatal or nonfatal cardiac events associated with borderline and abnormal T loops and T axes, taking subjects with n o r m a l T loops and T axes, respectively, as the reference group. We adjusted for two sets of risk indicators: age and sex (model A); and age, sex, and

established cardiovascular risk factors: body mass index, ratio of total to high density lipoprotein cholesterol, current cigarette smoking, diabetes, hypertension, history of angina, and history of m y o cardial infarction (model B). Subgroup analyses were carried out to assess the modifying influence of age, sex, and history of myocardial infarction on the risk for cardiac m o r tality associated with an a b n o r m a l T-loop m o r p h o l ogy or T axis.

Results Table 1 shows baseline characteristics of all participants and of those with normal, borderline, and abnormal T-loop morphologies, after adjustment for age and sex. Most variables have a clear association with T-loop m o r p h o l o g y . Differences in baseline characteristics b e t w e e n those with and w i t h o u t follow-up data s h o w e d that those w i t h o u t follow-up data had a lower prevalence of diabetes (7.3% vs 12.7%, P = .008) and a lower body mass index (25.7 vs 26.3, P = .006). Differences for all other baseline characteristics were nonsignificant (P > .10). During the foll o w - u p period, 166 subjects (2.9%) died from a cardiac cause, 73 (1.3%) of w h o m died suddenly. Nonfatal cardiac events were experienced by 193 participants (3.3 %). Subjects with an abnormal T loop had a m o r e t h a n fourfold age- and sex-adjusted risk for cardiac death (hazard ratio [HR] 4.3; 95% CI 3 . 0 - 6 . 4 ) and sudden cardiac death (HR 4.6; 95% CI 2 . 6 - 8 . 3 )

T-Loop Morphology as a Marker of Cardiac Events

Table 2, Hazard Ratios (95% CI) of Borderline and

Abnormal T-Loop Morphology and Frontal T Axis, Adjusted for Age and Sex (Model A) Endpoint Cardiac death Sudden death Nonfatal cardiac event Fatal or nonfatal cardiac event

Variable

Borderline

Abnormal

T loop T axis T loop T axis T loop T axis T loop T axis

2.7 (1.8-3.9) 2.9 (1.9~L4) 2.9 (1.6-5.2) 3.9 (2.1-7.1) 2.2 (1.5-3.2) 1.9 (1.2-2.9) 2.2 (1.7-3.0) 2.3 (1.7-3.1)

4.3 (3.0-6.4) 3.8 (2.7-5.5) 4.6 (2.6-8.3) 4.4 (2.5-7.7) 3.0 (1.9-4,8) 2.7 (1.8-4,1) 3.7 (2.8-5,0) 3.3 (2.5-4,4)

(Table 2). In addition, a m a r k e d increased risk for nonfatal cardiac events (HR 3.0; 9 5 % CI 1.9-4.8), a n d for fatal a n d nonfatal cardiac events c o m b i n e d (HR 3.7; 95% CI 2.8-5.0) was observed. The risks associated w i t h a borderline T-loop m o r p h o l o g y w e r e also significantly increased for all outcomes. Subjects w i t h a borderline or a b n o r m a l T axis h a d c o m p a r a b l e or slightly l o w e r risks for all endpoints. For c o m p a r i s o n purposes, Figure 1 shows the age- a n d sex-adjusted risk for cardiac d e a t h associated w i t h each of the risk indicators separately. The risk associated w i t h either an a b n o r m a l T loop or an a b n o r m a l T axis p r o v e d to be higher t h a n that of a n y other risk indicator, including ST depression or T-wave inversion. After a d j u s t m e n t for different established cardiovascular risk indicators (model B), the hazard ratios associated w i t h a n a b n o r m a l T-loop m o r p h o l o g y

7-

~ 4-

"I-

•

Kors et al.

57

attenuated, but r e m a i n e d statistically highly significant for all endpoints (Table 3). Again, the risks associated w i t h a n a b n o r m a l frontal T axis w e r e comparable. Subgroup analyses indicated that relative risk estimates for cardiac d e a t h associated w i t h an abn o r m a l T-loop m o r p h o l o g y w e r e only slightly m o d ified b y age (HR 6.0 (95% CI 2.8-12.8) for age --< 75 years a n d 4.8 (95% CI 3.1-7.4) for age > 75), b y sex (HR 4.1; 9 5 % CI 2 . 4 - 6 . 9 ) for males a n d 4.3 (95% CI 2.4-7.8) for females), a n d b y history of myocardial infarction (HR 3,6; 9 5 % CI 2.1-6.0) in subjects w i t h o u t a n d 3.5 (95% CI 1 . 8 - 6 . 6 ) in subjects w i t h a history of myocardial infarction.

Discussion

Both T-loop m o r p h o l o g y and T axis are strong a n d i n d e p e n d e n t risk indicators of cardiac events in o u r study p o p u l a t i o n of older adults. Their prognostic i m p o r t a n c e was higher t h a n that of a n y other established cardiovascular risk indicator. ST-T w a v e changes as m a n i f e s t e d in the ECG are k n o w n to be predictive for c o r o n a r y h e a r t disease (1-8). Also in our study, ST depression a n d T-wave inversion are associated w i t h m a r k e d l y increased risks for cardiac death (cf. Fig. 1). However, abnormality of the T loop proves to be a m u c h stronger predictor of future cardiac events. Apparently, the m o r p h o l o g y of the T loop carries prognostic inform a t i o n w h i c h is not or only incompletely c o n v e y e d b y the usual ST-T w a v e change codes. This m a y not c o m e as a surprise considering t h a t the M i n n e s o t a Code has primarily b e e n designed for a n u n a m b i g uous description of the ECG, n o t as a diagnostic tool (27,28). Our choice for a set of VCG m e a s u r e m e n t s to describe a n d classify T-loop m o r p h o l o g y is based on practical rather t h a n principal considerations. For-

.

2 -

,~,

Table 3. Hazard Ratios (95% CI) of Borderline and Abnormal T-Loop Morphology and Frontal T Axis,

0-

r o

.~

.~

.~

o

o

~ ~

r

~

'~

._

Adjusted for Age, Sex, and Cardiovascular Risk Indicators (Model B)

~ ~

.~

Endpoint .

.

.

.

Cardiac death Sudden death

Fig. 1. Age- and sex-adjusted hazard ratios (95% CI) of cardiac death for T-loop abnormality, frontal T-axis deviation, and other commonly used risk indicators and ECG abnormalities.

Nonfatal cardiac event Fatal or nonfatal cardiac event

Variable

Borderline

Abnormal

T loop T axis T loop T axis T loop T axis T loop T axis

2.0 (1.3-3.0) 2.2 (1.4-3.5) 1.9 (1.0-3,6) 2.7 (1.4-5.3) 1.6 (1.0-2.3) 1.3 (0.8-2.1) 1.6 (1.2-2.2) 1.7 (1.2-2.4)

3.2 (2.I--4.9) 3.0 (2.0-4.5) 3.4 (1.8-6.4) 3.4 (1.8-6.1) 2.0 (1.2-3.2) 1.9 (1.2-2.9) 2.6 (1.9-3.5) 2.4 (1.8-3.2)

58

Journal of Electrocardiology Vol. 31 Supplement

malization of the cardiologist's expertise in a decision algorithm is m u c h simpler w h e n the VCG is used instead of the ECG, mainly because the T loop is readily characterized by features that are easily derived from the planar VCG projections, involving only three orthogonal components. Surprisingly, the predictive value of an abnormal frontal T axis, w h i c h is only one, albeit important, parameter in assessing overall T-loop abnormality, is only slightly less than the risk associated with abnormal T-loop m o r p h o l o g y . Since the m e c h a nisms underlying repolarization disturbances are still largely unclear, a physiological explanation is difficult to give. However, it seems p r u d e n t to say that both parameters are markers of (sub)clinical myocardial damage. An advantage of measuring the T axis over T-loop m o r p h o l o g y is its simplicity. The frontal T axis can readily be estimated similarly to the physician's derivation of the frontal QRS axis: Using the same leads, the procedure is simply applied to the T wave instead of the QRS complex.

Conclusion T-loop m o r p h o l o g y and T axis seem to be important, general markers of a compromised myocardium in the elderly and are strong, i n d e p e n d e n t predictors for fatal and nonfatal cardiac events. If confirmed, their m e a s u r e m e n t should be considered in identifying individuals p r o n e to develop coronary heart disease both in clinical practice and in screening programs.

References 1. Cedres BL, Liu K, Stamler J et ah Independent contribution of electrocardiographic abnormalities to risk of death from coronary heart disease, cardiovascular diseases and all causes. Findings of three Chicago epidemiologic studies. Circulation 65:146, 1982 2. Cullen K, Stenhouse NS, Wearne KL et al: Electrocardiograms and 13 year cardiovascular mortality in Busselton study. Br Heart J 47:209, 1982 3. Rabkin SW, Mathewson FAL, Tate RB: The electrocardiogram in apparently healthy men and the risk of sudden death. Br Heart J 47:546, 1982 4. Kannel WB, Anderson K, McGee DL et al: Nonspecific electrocardiographic abnormality as a predictor of coronary heart disease: the Framingham study. Am Heart J 113:370, 1987 5. Liao Y, Liu K, Dyer A et al: Major and minor electrocardiographic abnormalities and risk of death

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

from coronary heart disease, cardiovascular disease and all causes in men and women. J Am Coll Cardiol 12:I494, 1988 Knutsen R, Knutsen SF, Curb JD et al: The predictive value of resting electrocardiograms for I2-year incidence of coronary heart disease in the Honolulu heart program. J Clin Epidemiol 4I:293, 1988 Sigurdsson E, Sigfusson N, Sigvaldason H et ah Silent ST-T changes in an epidemiological cohort study--a marker of hypertension or coronary heart disease, or both: the Reykjavik study. J Am Coll Cardiol 27: 1140, 1996 Tervahauta M, Pekkanen J, Punsar S et al: Resting electrocardiographic abnormalities as predictors of coronary events and total mortality among elderly men. Am J Med 100:641, 1996 Hofman A, Grobbee DE, De Jong PTVM et ah Determinants of disease and disability in the elderly: the Rotterdam elderly study. Eur J Epidemiol 7:403, 1991 Rose G, McCartney P, Reid DD: Self-administration of a questionnaire on chest pain and intermittent claudication. Br J Prev Soc Med 31:42, 1977 World Health Organization: International Statistical Classification of Diseases and Related Health Problems, 10th rev. World Health Organization, Geneva, 1992 Cupples LA, Cagnon DR, Kannel WB: Long- and short-term risk of sudden coronary death. Circulation 85(suppl I):ll, 1992 Myerburg RJ, Kessler KM, Castellanos A: Sudden cardiac death: Structure, function, and time-dependence of risk. Circulation 85(suppl I):2, 1992 Van Bemmel JH, Kors JA, Van Herpen G: Methodology of the modular ECG analysis system MEANS. Methods Inf Med 29:346, 1990 Prineas RJ, Crow RS, Blackburn H: The Minnesota Code Manual of Electrocardiographic Findings: Standards and Procedures for Measurement and Classification. Wright PSG, Boston, 1982 Kors JA, Van Herpen G, Wu J e t ah Validation of a new computer program for Minnesota coding. J Electrocardiol 29(Suppl):83, 1996 Willems JL, Arnaud P, Van Bemmel JH et al: A reference data base for multilead electrocardiographic computer measurement programs. J Am Coll Cardiol 10:1313, 1987 Willems JL, Abreu-Lima C, Arnaud P et al: The diagnostic performance of computer programs for the interpretation of electrocardiograms. N Engl J Med 325:1767, I991 Edenbrandt L, Pahlm O: Vectorcardiogram synthesized from a 12-lead ECG: superiority of the inverse Dower matrix. J Electrocardiol 21:361, 1988 Kors JA, Van Herpen G, Sittig AC et ah Reconstruction of the Frank vectorcardiogram from standard electrocardiographic leads: diagnostic comparison of different methods. Eur Heart J 11:1083, 1990 Talmon JE, Van Herpen G: Quantitative classification

T-Loop Morphology as a Marker of Cardiac Events

22. 23.

24.

25.

of T-wave abnormalities in the VCG. pp. 233-236. Adv Cardiol, vol. 16. Karger, Basel, 1976 Bristow JD: A study of the normal Frank vectorcardiogram. Am Heart J 61:242, 1961 Draper HW, Peffer CJ, Stallmann FW et al: The corrected orthogonal electrocardiogram in 510 normal m e n (Frank lead system). Circulation 30:853, 1964 Silverberg SM: A quantitative study of the Frank vectorcardiogram. A comparison of younger and older normal populations. A m J Cardiol i8:672, 1966 Cooksey JD, Dunn M, Massie E: Clinical Vectorcar-

•

Kors et al.

59

diography and Electrocardiography (2nd ed), Year Book Medical Publishers, Chicago, 1977 26. Cox DR: Regression models and life tables. J Royal Stat Soc Series B 34:187, 1972 27. Pipberger HV, Simonson E, Lopez AE et al: The electrocardiogram in epidemiologic investigations. A n e w classification system. Circulation 65:1456, 1982 28. Tuinstra CL, Rautaharju PM, Prineas RJ et al: The performance of three visual coding procedures and three computer programs in classification of electrocardiograms according to the Minnesota Code. J Electrocardiol 15:345, 1982