Blood Pressure Measures and Electrocardiogram‐Defined Myocardial Infarction in an Iranian Population: Tehran Lipid and Glucose Study

Abstract

The purpose of this study was to describe blood pressure values in Iranian adults with electrocardiogram (ECG) evidence of a myocardial infarction (MI). High blood pressure is a risk factor, and an ECG can be diagnostic of coronary artery disease. In recent studies the role of pulse pressure in predicting coronary artery disease has been suggested to be more important than that of blood pressure. From among participants of the Tehran Lipid and Glucose study, data for 2479 men and 3060 women aged ≥30 years not currently using any antihypertensive medication were collected. The study used the mean of two separate blood pressure measurements for each individual. ECG findings of all subjects were coded according to Minnesota ECG coding criteria, and they were categorized into probable/possible MI or no MI. ECG evidence of probable or possible MI was found in 1.2% of subjects (1.8% in men vs. 0.8% in women, p<0.001). Prevalence of ECG‐defined MI in hypertensive persons was two‐fold higher than in normotensives. Adjusted for age, sex, and body mass index, mean diastolic blood pressure was significantly lower in cases with ECG‐defined MI than in subjects without MI (p<0.03). There was a strong positive correlation between pulse pressure and systolic blood pressure in both hypertensive/normotensive and MI/no MI groups at the p<0.001 level. There was a weak inverse correlation between diastolic blood pressure and pulse pressure in hypertensive/normotensive/no MI groups (−0.32 and −0.14, both p<0.001). Diastolic blood pressure was not correlated with pulse pressure in cases with MI. Prevalence of ECG‐defined MI in hypertensive cases was higher than in normotensives. Systolic blood pressure is a better predictor for pulse pressure than diastolic blood pressure in both normotensive and hypertensive populations with or without ECG‐defined MI.

Hypertension is one of the important modifiable risk factors of cardiovascular diseases. The usefulness of its treatment in preventing coronary artery disease has been demonstrated in several controlled clinical trials. ¹ , ² For many years there was a tendency to consider diastolic blood pressure (DBP) rather than systolic blood pressure (SBP) as a better predictor of cardiovascular disease. In recent studies, however, the role of SBP and pulse pressure (PP) in predicting coronary artery disease has been emphasized and considered to be more important. ³ , ⁴ , ⁵ , ⁶ PP is defined as SBP less DBP and is related to cardiac output and arterial compliance. ⁵ A recent study ⁷ reported by the Framingham group has emphasized the role of PP in predicting coronary events in males older than 50 years. Studies in older men and women have found that PP remains important even after controlling either SBP or DBP. ⁷ , ⁸ , ⁹

The recent Framingham study showed that the risk of coronary artery disease for any SBP increases with a simultaneous decreasing DBP in middle‐aged men. ⁷ Similar findings have been reported elsewhere. ¹⁰

In our study we investigated the prevalence of SBP, DBP, and PP in hypertensive and normotensive subjects with or without electrocardiogram (ECG) evidence of an acute myocardial infarction (MI) in the Tehran Lipid and Glucose Study (TLGS).

TLGS Design

TLGS is a longitudinal study in which the first phase is a cross‐sectional population study or a baseline examination survey. It has been designed to measure the prevalence of risk factors including dyslipoproteinemia, hyperglycemia, obesity, smoking, and hypertension in a representative sample of an Iranian urban population in district No. 13 of Tehran. Details of TLGS, its rationale, and its design have been published elsewhere. ¹¹ , ¹² , ¹³ Crude response rate in the TLGS participants was approximately 57.5%. The reasons for nonresponse have been investigated to find out if there is any essential difference between respondents and nonrespondents. Data revealed there was no significant difference between responders and nonresponders.

The data were collected by means of participant interview, completion of a questionnaire by a trained interviewer for demographic data; a physical examination for blood pressure, pulse rate, and anthropometrical measurements; and laboratory measurements of lipid profiles, fasting blood sugar, 2 hours post‐glucose challenge, and thyroid profile. For those subjects older than age 30 years, ECGs were also taken. Data collected were directly stored in a computer. ¹⁴

Blood Pressure Measurement

After 15 minutes rest, blood pressure was measured three times by a physician trained in blood pressure measurement using a standard mercury sphygmomanometer, initially once for determining peak inflation level and twice for determining blood pressure value. The shygmomanometers were calibrated by the Iranian Institute of Standards and Industrial Researches. On the basis of the circumference of the participant's arm, a regular adult or large cuff was chosen. The cuff was placed on the participant's right arm at the heart level and inflated until the cuff pressure was 30 mm Hg above the level at which the radial pulse disappeared. The SBP was defined as the appearance of the first sound (Korotkoff phase 1), and the DBP was defined as the disappearance of the sound (Korotkoff phase 5). The participant was expected to refrain from drinking tea or coffee, smoking, participating in any physical activity, and should have emptied his or her bladder during the 30 minutes preceding the measurement.

All physicians were required to participate in a period of specialized training for the use of a standardized protocol for measuring blood pressure, following which they took a qualifying exam to ensure the quality of the blood pressure measurement. A digit preference score was also calculated to ascertain the quality and precision of physicians' performances. ¹² Hypertension was defined as SBP ≥140 mm Hg or DBP ≥90 mm Hg according to the Joint National Committee on Prevention, Prevention Detection, Evaluation, and Treatment of High Blood Pressure's seventh report (JNC 7) criteria. ¹⁵ All individuals participating in this study were required to sign an informed consent form.

ECG and Definition of MI

A 12‐lead rest ECG was recorded for each individual aged ≥30 years in the TLGS population by two trained and qualified technicians, according to a standard recording protocol developed by the University of Minnesota, ¹⁶ School of Public Health, using a PC‐ECG 1200 machine (models S and M for DOS, version 4.07, Cardioscan NV/SA, Brussels, Belgium). Two trained physicians coded the ECGs independently, according to the Minnesota codes, using a measuring loupe specially manufactured by the University of Minnesota. ¹⁶ To ensure quality, a third trained physician recoded 10% of the ECGs and all the data were doubly entered. ¹⁷

Statistical Analysis

Subjects aged ≥30 years, currently using no antihypertensive medication, were initially selected and categorized into hypertensive and normotensive groups according to JNC 7 criteria. ¹⁵ They also were divided into three subgroups based on the ECG evidence ¹⁸ : 1) probable or definite MI; 2) possible or equivocal MI; or 3) no MI. For more detailed analysis and because of the small numbers of people with probable and possible MI among our subjects, we considered the first two groups as a single group named MI.

Analysis of variance and a general multivariate linear model were used to compare mean PP, SBP, and DBP between groups or subgroups. The Pearson correlation test was used to determine the association between SBP and PP, and also between DBP and PP in both hypertensive and normotensive groups. SPSS 10.0 statistical software package (SPSS Inc., Chicago, IL) was used in statistical analysis.

RESULTS

ECG data for 5539 subjects (2479 men and 3060 women) aged 30 years or older were collected and analyzed. Mean age was 48±13 years and 45±11 years in men and women, respectively. Hypertension was seen in 19% of men and 18% of women. ECG‐defined probable or possible MI was found in 1.2% of cases (1.8% for men and 0.8% for women, p<0.001) (Table I). The prevalence of MI was twice as high in hypertensive subjects compared with normotensives (1% vs. 2%, p<0.01). When the subjects were categorized into different sex groups, the significant difference was only seen in women (1.6% vs. 2.7%, p=0.08 in men; 0.6% vs. 1.4%, p<0.05 in women).

Table I.

Distribution of Electrocardiogram (ECG)‐Defined Myocardial Infarction (MI) by Sex and Blood Pressure Status in the Tehran Lipid and Glucose Study Adult Population

			Ecg Definition
		MI		No MI
Sex	Blood Pressure	N	%	N	%
Men	Hypertensive	13	2.7	468	97.3
	Normotensive	32	1.6	1966	98.4
	Total	45	1.8*	2434	98.2
Women	Hypertensive	8	1.4**	544	98.6
	Normotensive	14	0.6	2494	99.4
	Total	22	0.7	3038	99.3
*p<0.001 compared with women; **p=0.03 compared with normotensives of the same sex

Blood Pressure

Mean blood pressure values for the studied population are shown in Table II. Men had significantly higher mean SBP and PP than women. Mean blood pressure values in both genders after adjustment for age, sex, smoking, waist‐to‐hip ratio, and body mass index are presented in Table III. There was no significant difference between blood pressure measures in subjects with or without ECG‐defined MI. However, when the cases were sudivided into normotensive and hypertensive groups, the mean DBP was significantly lower only in the normotensive cases (73 mm Hg vs. 75 mm Hg, p<0.03).

Table II.

Mean Blood Pressure Values by Sex in the Tehran Lipid and Glucose Study

Sex	Blood Pressure	Mean±SD
Men	SBP	121±18.0*
(n=2479)	DBP	78±10.1
	PP	42±13.3*
Women	SBP	119±17.4
(n=3060)	DBP	78±9.9
	PP	40±12.7
SBP=systolic blood pressure; DBP=diastolic blood pressure; PP=pulse pressure;*p<0.001 compared with women

Table III.

Blood Pressure Values by Sex, Blood Pressure Status, and Electrocardiogram (ECG)‐Defined Myocardial Infarction (MI) in the Tehran Lipid and Glucose Study Adult Population With ECG Data

			ECG‐Defined MI
		MI		No MI
Variable	Blood Pressure Status	Mean±SE	95% CI	Mean±SE	95% CI
SBP	Hypertensive	140±3.4	133–144.4	145±0.5	144–146
	Normotensive	112±1.6	109–115	114±0.2	113–114
	Total	118±1.9	114–121	120±0.2	119–120
DBP	Hypertensive	91±1.9	87–94	92±0.3	92–93
	Normotensive	73±1.1*	71–75	75±0.1	75–76
	Total	76±1.2	74–79	78±0.1	78–79
PP	Hypertensive	49±3.3	43–55	53±0.5	52–54
	Normotensive	39±1.3	36–42	38±0.1	38–39
	Total	41±1.4	39–44	41±0.2	41–42
CI=Confidence interval; SBP=systolic blood pressure; DBP=diastolic blood pressure; PP=pulse pressure; *p=0.01 compared with DBP in normotensives without MI

Correlations

There was a positive correlation between PP and SBP in both hypertensive and normotensive groups (r=0.87, p<0.001 and r=0.75, p<0.001, respectively). The corresponding values for cases with MI and without MI were r=0.78 and r=0.82, respectively, both p<0.001. Negative correlation was found between DBP and PP; (r=−0.32 in hypertensives and r=−0.14 in normotensives, both p<0.001). There was no significant correlation or a weak correlation between DBP and PP in cases with and without MI (r=−0.08 and p=0.6, r=0.15 and p<0.001, respectively) (data not shown) (Table IV).

Table IV.

Systolic and Diastolic Blood Pressure Correlation Coefficient With Pulse Pressure by Blood Pressure Status and Electrocardiogram‐Defined Myocardial Infarction (MI) in the Tehran Lipid and Glucose Study Adult Population

Pulse Pressure
		Hypertensive				Normotensive
	MI	p	No MI	p	MI	p	No MI	p
Diastolic blood pressure	−0.37	0.13	−0.21	0.001	−0.38	0.01	−0.17	0.01
Systolic blood pressure	0.84	0.001	0.84	0.001	0.70	0.01	0.72	0.01

DISCUSSION

This study shows that hypertension is twice as prevalent in subjects with ECG‐defined MI as in subjects without ECG‐defined MI. This is consistent with the cardiovascular literature. In spite of a higher prevalence of hypertension in subjects with MI compared with those without MI, mean blood pressure measures are lower in the first group. Except for DBP in normotensive subjects with MI, which is significantly different from DBP in normotensive subjects without MI, there are neither statistical nor clinical differences in SBP or PP in subjects with or without MI in both normotensive and hypertensive groups. However, it seems a small number of, and highly deviated blood pressure measures in, subjects with MI compared with those without MI, are the best explanations for lower blood pressure in former group. Although we eliminated any pharmacologic effect of medical interventions in subjects with MI by excluding cases receiving antihypertensive medication, other possible medical interventions such as dietary or physical activity regimens prescribed by physicians for these subjects should be considered.

Recently, it has been shown that high PP is a risk factor for MI. ⁷ , ⁸ , ⁹ We did not show a statistically different mean PP in subjects with MI and without MI. There was, however, a strong positive correlation between PP and SBP; a correlation between PP and DBP was weakly negative in our study. This is in contrast to some studies such as those by Franklin et al. ⁷ and Millar et al. ¹⁹ and consistent with certain other studies. ²⁰ , ²¹ The first group showed that coronary heart disease risk had a strong (not weak as in our study) inverse relation to DBP in middle‐aged and older individuals; Sesso and colleagues ²⁰ and Antikainen et al. ²¹ implied that DBP was not associated with coronary artery disease risk. Franklin et al. ⁷ showed the risk of coronary artery disease with a given SBP increases with decreasing DBP in older middle‐aged men. In the Medical Research Council (MRC) trial of treatment of mild hypertension, ¹⁹ PP was a stronger predictor of coronary events than was SBP, DBP, or mean blood pressure in males. Fatal and nonfatal coronary event rates increased progressively in ascending quartiles of age‐adjusted PP, but there was also a strong correlation with SBP. Coronary risk was inversely related to DBP. ¹⁹ Sesso et al. ²⁰ showed average SBP, DBP, and mean arterial pressure strongly predict cardiovascular disease among younger men, whereas either average SBP or PP predicts cardiovascular disease among older men. In this study, average DBP was not associated with coronary artery disease risk in older men. Average PP was associated with the risk of cardiovascular accident in both younger and older men. ²⁰ Antikainen and coworkers, ²¹ during a 15‐year follow‐up period, showed that relative risk of coronary heart disease, stroke, cardiovascular disease, and all‐cause mortality increased with increasing PP in individuals aged 45–64 years, independent of DBP level. After adjustment for SBP, the positive association between mortality and increasing PP disappeared. Studies suggested that increasing PP is a predictor for death from coronary heart disease, stroke, cardiovascular disease, and all other causes in men and women aged 45–64 years, but that increase in risk is entirely associated with the increase in SBP. ²¹

There are several possible reasons that the data reported here are different from earlier studies. Our study was cross‐sectional in nature. Our cases comprised a younger population: the mean age in our cases was 46±11.9 years. In young individuals, as vascular resistance increases there is a proportional increase in SBP and DBP. With the onset of middle age, SBP increases more than DBP, resulting in elevation of PP. ⁷ , ²² Thus, DBP rises with increased peripheral artery resistance during middle age and falls with increased central artery stiffness in older age.

Considering the results of our study and results of other studies, it is clear that most of the effects of PP on prediction of coronary artery disease resulted from higher SBP rather than from lower DBP.

Acknowledgments and disclosure:

The authors thank the participants of district No. 13 of Tehran for their enthusiastic support in this study and the social relations unit and steering committee of the Tehran Lipid and Glucose Study for their guidance in the preparation of this article. The participation of the staffs of the Endocrine and Metabolism Research Center, Tehran Lipid and Glucose Study Unit, and the Health Department of Shaheed Beheshti University of Medical Sciences is gratefully acknowledged. The authors also thank Dr. Nasrollah Rezaieghaleh for his assistance in the preparation of this manuscript and the ECG Analyzer Team of Tehran Endocrine Research Center for its assistance in coding ECGs. This research project was supported by national grant No. 121 from NRCI Research Projects and by the National Research Council of the Islamic Republic of Iran.