EPIDEMIOLOGIC STUDIES
Risk Factors for Myocardial Infarction and Death Due to lschemic Heart Disease and Other Causes GijSTA TIBBLIN, MD LARS WILHELMSEN, MD LARS WERKij, MD, FACC Gteborg,
Sweden
From the Section of Preventive Cardiology, Medical Department I, Sahlgren’s Hospital, Goteborg, Sweden. This study was supported by grants from the Swedish Association Against Heart and Chest Diseases, and the Forenade Liv Insurance Company. Manuscript accepted September 12, 1974. Address for reprints: Gosta Tibblin, MD, Medical Department I, Sahlgren’s Hospital, S-413 45, Goteborg, Sweden.
514
April 1975
As part of a study of the male population in an industrial city in Sweden, one third of all male inhabitants of Goteborg born in 1913 were invlted to an examination in 1963. Of those invited, 655 (66 percent) accepted. This report examines the incidence of nonfatal myocardial infarction and death from ischemic heart disease and other causes in this group of men during the ensuing 10 years. There were 61 deaths; autopsy was performed in 56 cases. Nineteen men died of ischemic heart disease and 16 of cancer; 12 men died violently. Thirty-one men survived an acute myocardial infarction. Cigarette smoking and registration with the Temperance Board at the time of the initial examination were more common in men who later had a nonfatal myocardial infarction or died of ischemic heart disease or other causes than in surviving subjects and men who did not have an infarction. Dyspnea was more common in men who died of ischemit heart disease but was less common in those who died of other causes than in the remaining subjects. Values for systolic blood pressure were higher and those for peak expiratory flow lower in men who died of ischemlc heart disease. Serum cholesterol values were higher and those for serum triglycerides tended to be higher in men who died of ischemic heart disease than in other subjects. Heart size tended to be greater in those who had nonfatal or fatal ischemic heart disease. Obesity, the level of physical activity, fasting blood glucose levels, coffee consumption, hematocrit and erythrocyte sedimentation rate as determined at age 50 years had no predictive value for assessing the risk of nonfatal myocardial infarction, fatal ischemic heart disease or death from other causes before age 60. The results indicate that many so-called risk factors have a different relation to fatal than to nonfatal ischemic heart disease.
Investigators are agreed that age has an important bearing on the development of ischemic heart disease although the rather small incidence of this disease, especially in young persons, has made it difficult to study populations within narrowly defined age groupings. In most studies, subjects have been classified within age spans of 10 or more years; frequently the age ranges at entry have been 30 to 49 or 50 to 59 years, and there has been no clear description of the distribution of the population sample within these age ranges. The effect may be to skew data on risk factors and other variables, although most investigators try to compensate mathematically for possible distortion. This possibility constitutes one good reason for eliminating the age factor in prospective studies. Another is that several of the conventional risk factors, such as serum cholesterol and blood pressure, are in themselves age-related. Furthermore, all risk factors carry a different weight at different ages. For example, many studies may underestimate the importance of smoking, which tends to decrease with age, or exaggerate the significance of increased blood pressure or cholesterol, which tends to increase with age.
The American Journal of CARDIOLOGY
Volume 35
RISK FACTORS
The present study is part of a long-term investigation of the development of ischemic heart disease and other diseases in a random population sample of the same age and sex. It is intended to answer the following questions: Are there differences among groups classified as having (1) nonfatal myocardial infarction, (2) fatal ischemic heart disease, (3) death from other causes, or (4) no disease? If so, can such differences be explained by differences between one or more of the first three groups and the group with no disease? Furthermore, are there differences between the group with nonfatal myocardial infarction and the group with fatal ischemic heart disease?
Study Population and Methods All inhabitants of Sweden have a national registration number that includes their date of birth and other vital statistics. Names, addresses and registration numbers are registered with the official county census bureau and were accessible before the sample was drawn for the present study. The study population was recruited from men living in Giiteborg, Sweden, born in 1913 and still alive at age 50 years (1963). All men meeting these criteria who were born on a date divisible by three-that is, the third, sixth, ninth day (and so on) of each month-comprised the study sample. Nine hundred seventy-three men met these criteria. Of these 855 (88 percent) agreed to be examined in 1963 at Sahlgren’s Hospital, Giiteb0rg.l Of the 118 nonparticipants, 7 had died, 4 were hospitalized and 9 did not live in the town at the time of the investigation. The remaining 98 men refused study for various reasons, usually because of a negative attitude toward medical care. Compared with participants, nonparticipants had a lower mean income, were more frequently unmarried and were more often registered with the Temperance Board. Slightly more nonparticipants than participants had obtained sick allowance.2 For most of the participants (847 persons) complete records for the present analysis were available. It was also possible to trace the death certificates of the men in the overall study sample who died in Gijteborg between 1950 and 1963, that is, between the ages of 37 and 50 years. A total of 33 deaths had occurred, only 4 of which had been listed as due to ischemic heart disease. A detailed medical history was obtained from each participant, and all clinical examinations were performed by one of us (G.T.). Examinations were performed in the morning, with the subjects fasting until after blood sampling. A parental death score was constructed, an analysis facilitated by the identical age at entry (50 years) of all participants in the study. The cause of the parents’ death was not taken into account because thk data could not, be verified. Participants received 1 point if a parent had died before age 65 years, 2 points if the parent had died between the ages of 66 and 79 years, 3 points if the parent was living and aged 66 to 79 years and 4 points if the parent was living or had died after age 80 years. Thus, a subject received a minimal score of 2 if both parents had died before age 65 years and a maximal score of 8 if both parents were still alive or had died after age 80 years. Blood groups in the ABO, Rh, Kell, Duffy and MN systems were determined according to methods described earlier.3 For analysis of smoking data the subjects were grouped
FOR DISEASE-TIBBLIN
ET AL.
as follows: never smoked, stopped smoking, smoking 1 to 14 cigarettes/day, smoking 15 to 24 cigarettes/day and smoking 25 or more cigarettes/day. Those who smoked only cigars, “cigarillos” or a pipe (75 subjects) were classified in a single group without consideration of their exact coasumption. In part of the analysis those who had never smoked and former smokers were grouped together, those who smoked 1 to 14 cigarettes a day were referred to a second group and those who smoked 15 cigarettes a day or more were referred to a third group. Coffee consumption was assessed by interview and recorded as the mean number of ordinary cups consumed each day. Reliable data concerning alcohol consumption are difficult to obtain from interviews, but in Sweden records of local temperance boards provide some guide to alcoholic intemperance. Board registration may cover the spectrum from severe alcoholism to a single offense of driving a car with a high level of alcohol in the blood. The records provide minimal data since heavy alcohol consumption that may be socially acceptable or take place only in the home does not lead to board registration. Nevertheless, 21 percent of this representative population sample of 50 year old men had at some time been registered with the temperance board. Physical activity during work was classified as sedentary, moderate or heavy according to the type of occupation held at age 50 years. Dyspnea on exertion was defined according to the criteria of Rose and Blackburn. Subjects experiencing dyspnea when hurrying on level ground or when walking up a small hill or undertaking other less taxing exertion were classified as having dyspnea. Blood pressure was measured by the same observer throughout the study after a 5 to 10 minute interview with the subject seated. Both systolic and diastolic blood pressure were measured to the nearest 5 mm Hg. Men found to have casual blood pressure readings of 175 mm Hg or more systolic and 115 mm Hg or more diastolic were given treatment. This fact has to be considered when high blood pressure is discussed as a risk factor in this study. Measures of overweight-obesity-were obtained by measuring skinfold thickness and calculating relative weight.. The latter was obtained by dividing the actual weight in kilograms by the height in centimeters minus 100. Heart rate at rest was determined from the resting electrocardiogram. Peak expiratory flow, as described by Wright and McKerrow,s was measured with patients standing; the mean of three measurements was used. Heart oolume was determined from full scale roentgenograms of the chest in the frontal and the oblique lateral projections according to the technique of Jonsell. All films were interpreted by the same roentgenologist. The absolute figure in milliliters was related to the body surface area in square meters. The fasting levels of serum cholesterol and serum triglycerides, blood sugar, hematocrit and erythrocyte sedimentation rate during 1 hour were determined at the same laboratory, and regular checks were performed to determine the accuracy of these methods.’ The criterion used for diagnosis of acute ischemic heart disease was (1) hospitalization with the clinical diagnosis of myocardial infarction, or (2) fresh ischemic heart disease at autopsy. The diagnoses were established by the clinicians or pathologists, respectively. Since November 1968, all clinical cases of myocardial infarction have been record-
April 1975
The American Journal of CARDIOLOGY
Volume 35
515
RISK FACTORS
TABLE
FOR DISEASE-TIBBLIN
ET AL.
In the total group of men studied, there were 61 deaths. Autopsy, usually performed in a standardized fashion, was carried out in 56 cases. The data were analyzed for the three end point groups: nonfatal myocardial infarction, fatal ischemic heart disease and death from other causes. If none of these end points occurred during the 10 year follow-up period the subjects were classified as having no disease. For the statistical study chi-square analysis was used to test the equality of distributions of quantitative variables in the four groups. No further statistical analysis was performed if significant differences were found. Analysis of variance was used to test equality of means. If significant
I
Number of Subjects Who Died or Had Nonfatal Myocardial Infarction Number Deaths Fatal ischemic heart disease Fresh myocardial infarction Other acute ischemrc heart disease Pulmonary carcinoma Other carcinoma Other nonviolent death Violent death Nonfatal myocardral infarction * Subjects
who drd not undergo
11 8 (l*) 9 (I*) 9 (2*) 12 12 (l*) 31
differences between means were found, these differences were localized by the Scheffe method. The significance level of P = 0.05 was used.
ed in collaboration with the World Health Organization in a Myocardial Infarction Register. The clinical criteria for myocardial infarction used in this study were those adopted by the Swedish Society of Cardiology: central chest pain, shock or syncope suggesting a myocardial infarction together with a typical transaminase spectrum or appearance of a pathologic Q wave or localized S-T changes in the electrocardiogram. The criterion for ischemic heart disease at autopsy was a fresh myocardial scar or, in the absence of any macroscopically observed scar, total or almost total occlusion of a coronary artery together with a medical history suggesting myocardial infarction or sudden coronary death.
TABLE
II
Clinical Course of Subjects lschemic Heart Disease Case IlO.
Who Died of
Durabon of Acute Symptoms
Hosprtalrzabon
A. Fresh Myocardral Infarctron 1 2 3 4 5 6 7 8 9
Days Days Days Days 1 l/2 hours 10 hours 15 hours Days 1
10 11
Day Days
516
No Yes, 2 hours Yes, 1 day Yes, 9 days No Yes, other disease No Yes, 1 hour Yes, operation for infarction Yes, 1 day Yes, 13 days
B. Acute lschemic Heart Disease ~No Found dead No Seconds No l/2 hour No Seconds Yes, other disease Seconds l/4 hour No No Seconds No Seconds
1 2 3 4 5 6 7 8
April 1975
The group with no disease was separated into quartiles (25 percent of the subjects in each group) according to values for systolic blood pressure, peak expiratory flow and serum cholesterol. The cutoff points obtained were then used to define the distributions in the three end point groups (see Fig. 5 to 7).
autopsy.
The American Journal of CARDIOLOGY
Results Fatal and Nonfatal lschemic
Heart Disease
This analysis concerns findings in the 10 year follow-up period from entry in 1963 to the end of June 1973. The number of events-nonfatal myocardial infarction and death from ischemic heart disease or other causes-is shown in Table I. Of the five subjects who died but did not undergo autopsy, one had a clear-cut sudden cardiac death and was classified in the group with ischemic heart disease. Of the other four, three died of cancer and one had a fatal accident. The clinical course of the men who died of ischemic heart disease is presented in Table II. Thirty-one men with nonfatal myocardial infarction were treated at the hospital. Of these 19 subjects who died with a diagnosis of ischemic heart disease, all except 1 had severe coronary atherosclerosis (Fig. 1). Seven of the 11 who had a fresh myocardial infarction at autopsy also had an old scar indicating an earlier myocardial infarction. The majority of those who died of other causes had few signs of coronary arterial changes or myocardial scars. Two subjects with fresh myocardial infarction at autopsy had died with other clinical diagnoses, extensive bronchial carcinoma in one, and severe aortic stenosis in the other. Most of those with a fresh myocardial infarction at autopsy had a history of chest pain lasting for at least 1 hour. Eight of 11 were hospitalized. Of those who died of acute ischemic heart disease all died within 30 minutes and none were hospitalized because of ischemit heart disease. Risk Variables Distributions for qualitative variables and means and standard deviations for quantitative variables are shown in Tables III and IV, respectively. The parental death score did not differ significantly among the four groups (chi-square = 15.0; df = 9)
Volume 35
RISK FACTORS
Fatal
lschemic
Heart
FOR DISEASE-TIBBLIN
ET AL.
Other Deaths
Disease
n=19
n=L2
Severe coronary at herosclerosis
Severe coronarv rtherosclerosis’
Fresh M I
Fresh M I
0
Nonsmokers
1-IL g/day
215 g/day
Non-fatal Myocardial
Infarction
n=31
u 1
FIGURE 1. Autopsy data on the coronary arteries and myocardium of subjects who died of fatal ischemic heart disease and other causes. Autopsy was performed in 18 and 38 cases, respectively. Fresh Ml = recent myocardial infarction. _
UScar
NonI-U smokers g/day Fatal Heart
El5
Non-
g/day
smokers
lschemic Disease
01 her
1-u g/day Deaths n=L2
?I5 g/day
NonI-14 smokers g/day
Scar
?15 g/day
No Disease n=755
n.19
FIGURE 2. Distribution of smoking habits in the four groups. g = cigarettes.
although the score tended to be lower in the group with fatal ischemic heart disease; that is, their parents died at an earlier age. There was a trend, although it was not significant, toward a higher frequency of blood group A and lower frequency of group 0 among those who had nonfatal or fatal ischemic heart disease. Similar trends were not observed in the other blood group systems. Smoking habits differed significantly among the four groups (chi-square = 20.1; df = 6). This was due to a higher proportion of smokers in all three end point groups than in the group with no disease (Fig. 2). In the group that died of causes other than ischemit heart disease, the overrepresentation of smokers among those who later had pulmonary carcinoma could account for the result. There were significant differences among the four groups with respect to registration by the temperance board (chi-square = 11.4; df = 3). This finding was due to an overrepresentation in all end point groups (Fig. 3).
The type of physical activity during work did not differ significantly among the four groups (chi-square = 6.5; df = 6). Dyspnea on exertion differed significantly among groups (chi-square = 14.2; df = 3). There was a high rate of dyspnea among those who later had fatal ischemit heart disease, but the rate was lower than expected among subjects who died of other causes (Fig. 4). With respect to blood pressure, there was a significant difference among the mean values for systolic (F = 4.65; df = 3) and diastolic (F = 3.05; df = 3) blood pressure (Fig. 5). For systolic but not for diastolic blood pressure the difference was localized between those with fatal ischemic heart disease and those with no disease. The mean values for heart rate at rest did not differ significantly among groups (F = 1.63; df = 3), but the group with fatal ischemic heart disease tended to have a higher heart rate. The mean values for heart volume did not differ significantly (F = 1.25; df = 3), but volume tended to
April 1975
The American Journal of CARDIOLOGY
Volume 35
517
RISK FACTORS
FOR DISEASE-TIBBLIN
ET AL.
be greater in the two groups with ischemic heart disease. The mean values for peak expiratory flow rate differed significantly among the four groups (F = 5.32; df = 3). This difference was localized between the group with fatal ischemic heart disease, which TABLE
III
Qualitative
Variables
in Relation to Different
Parental death score 2-3 4-5 6-7 8 Blood group A B AB 0 Smoking habits (cigarettes) Nonsmoker (never or stopped) Smoking 1-14/day Smoking 15/day Registered by temperance board Yes No Physical activity during work Sedentary Moderate Heavy Dyspnea Yes No The total number
of subjects
Fatal lschemrc Heart Disease (no. = 19)
8 (26%) 18 (58%)
9 5 4 9
4 (13%) 1(3%)
18 (58%)
Other Deaths (no. = 42)
No Drsease (no. = 755)
14 (33%) 29 (48%) 7 (17%) 1(2%)
178 (23%) 337 (45%) 179 (24%) 5103%)
20 (48%)
(50%) (23%) (22%) (9%)
4 (13%) 9 (9%) 9 (29%)
10 (56%) 1(‘3%) 2 (11%) 5 (25%)
2 (5%) 1(2%) 19 (45%)
342 (45%) 77 (10%) 29 (4%) 306 (41%)
4 (13%) 15 (48%) 12 (39%)
5 (28%) 8 (44%) 5 (28%)
13 (31%) 17 (41%) 12 (28%)
350 (46%) 255 (34%) 149 (20%)
11(35%) 20 (65%)
7 (39%) 11(51%)
13 (31%) 29 (69%)
145 (19%) 609 (81%)
9 (29%) 16(52%) 6 (19%)
6 (4’3%) 4 (27%) 5 (33%)
14 (32%) 12 (29%) 16 (39%)
262 (35%) 235 (32%) 243 (33%)
9 (29%)
10 (56%)
9 (14%) 36 (85%)
162W%,t 593(79%)
22 (71%)
in the various subgroups
8 (44%)
does not always equal the total number
of the group because
of missing
IV
Quantitative
Variables
in Relation to Different
Variable Systolic blood pressure (mm Hg) Diastolic blood pressure (mm Hg) Heart rate Heart volume (ml/m2 BSA) (standing) Peak expiratory flow (liters/min) Serum cholesterol (mg/lOO ml) Serum triglycerides (millimoles/liter) Fasting blood sugar (mg/lOO ml) Erythrocyte sedimentation rate (mm in 1 hour) Hematocrit (%) Coffee consumption (cups/day) As in Table III there are occasional BSA = body surface area.
518
End Points
Nonfatal Myocardral Infarctron (no. = 31)
Vanable
TABLE
had a low mean value, and the group with no disease (Fig. 6). The measures of obesity, skinfold thickness and relative weight, showed no relation with the end points. The mean values for both serum cholesterol and
April 1975
missing
End Points (Mean i
standard
deviation)
Nonfatal Myocardial Infarction (no. = 31)
Fatal lschemic Heart Disease (no. = 19)
Other Deaths (no. = 42)
144 f 30 95 i 15 86 * 12 408 =!z 71 487 zt 88 260 i 37 1.36 X!Z0.53 83 i 15
152 z+z33 98 + 15 95 i 16 402 I!= 68 436 & 106 279 z!= 37 1.59 i 0.57 88 i 18
143 i 94 + 92-f 384 +I 490 * 257 & 1.41 i 84Jr
9& 8 46 I!= 2 5.7 f 2.4
lOi 9 45 f 3 5.9 z!z 3.1
data that do not affect the statistical
The American Journal of CARDIOLOGY
Volume 35
23 17 18 64 99 49 0.76 17
81 10 46 & 4 5.4 ZIZ2.8 tests or conclusions.
No Disease (no. = 755) 137 i 91 It 91 It 391 f 506 & 245 + 1.23 & 82 f.
21 13 15 58 80 42 0.71 15
8f7 45 * 4 5.1 f 3.0
data.
RISK FACTORS
FOR DISEASE-TIBBLIN
ET AL.
% IOO-
n
80.60-LO--
-
20-O-
Non-fatal Fatal lschemic Other DeathsNo Disease Myocardral Heart Disease n=42 n=755 Infarction n=19 n:31 FIGURE 3. Registration by the temperance board in the four groups. + = registered: - = not registered.
B
-
l--tlnm +
l
-
Fatal
Non-fatal
+
lschemrc
Other
Deaths
No Drsease
Heart Disease Myocardral n=42 n-755 Infarction n=19 n.31 FIGURE 4. Dyspnea on exertion in the four groups. -I = present; = not present.
% 60-
-
40--
20--
FIGURE 5. Distribution of systolic blood pressure. The gr6up with no disease was classified by quartiles and the cutoff points obtained were then used for all groups.
-
I
O- . > Systolrc brooa 122 132 146 Dressure mmHg No of cases 7 4 11 9
122 132 146
122 132 146 4
2
4
8
Non-fatal
Fatal
lschemic
Myocardial
Heart
Disease
10
6
Other
9
17
Deaths
122 132 146 188 188 I88 188 No
Disease
Infarction
60
40
20
pop FIGURE 6. Distribution of peak expiratory flow rate (litersjmin). The group with no disease was classified by quartiles and the cutoff points obtained were then used for all groups.
0 f;,!‘$atury
No of cases
0l ln -
466 512 557
11
6
5
-
-_
466 512 557
8
Non-fatal
10
Fatal Heart
Myocardial Infarction
April 1975
6
0
2
lschemic Disease
The American
466 512 557
13
Other
9
8
II
Deaths
Journal of CARDIOLOGY
466 512 557 182 182 182 182 No
Disease
Volume 35
519
RISK FACTORS
FOR
Cholesterol mg/lOOml No of cases
215 211 271 .4
DISEASE-TIBBLIN
L
9
Non-fatal Myocardial lnfarctlon
ET AL.
215 2Ll 271
215 241 271 0
f4
Fatal Heart
1
9
8
lschemic Disease
6
10
Other
serum triglycerides differed significantly among groups (F = 6.00 and 2.67, respectively; df = 3). For serum cholesterol, the difference was localized between the group with fatal ischemic heart disease and the group with no disease (Fig. 7), and the same tendency, although not significant, was seen for serum triglycerides. The mean values for fasting blood sugar level did not differ significantly, nor did those for erythrocyte sedimentation rate or hematocrit. However, blood sugar level tended to be higher in the group with fatal ischemic heart disease. Coffee consumption did not differ significantly among groups but tended to be greater in the two groups with ischemic heart disease. Discussion The proportion of nonparticipants in this study is considered small. Nonparticipants were more often registered by the temperance board and, since there is a positive relation between alcohol consumption and smoking, they may very well be heavier smokers than participants. Since both smoking and alcohol abuse were associated with ischemic heart disease and the nonparticipants had a higher death rate, at least during the first part of the follow-up period, it is possible that these factors are more strongly related to ischemic heart disease than our results indicate. In other prospective studies in which the participation rate usually has been lower, the effect of socially related variables, such as excessive use of alcohol, may be more distorted than in the present investigation. Familial history: Several clinical studies7-i3 have suggested that ischemic heart disease tends to occur in families. A shortcoming of several of these studies is that it is necessary to collect the information from subjects who seldom know much about the cause of their parents’ death. Since most of the studies have concerned subjects from different age groups, it has been difficult to standardize for the age of the par-
520
April 1975
The American Journal of CARDIOLOGY
10 16 Deaths
al
215 241 271
188 188 188 188 No Disease
FIGURE 7. Distribution of serum cholesterol. The group with no disease was classified by quartiles and the cutoff points obtained were then used for all groups.
ents. In our study this standardization was possible. A familial predisposition to fatal ischemic heart disease was suggested by our results, although the data were not significant. Blood groups: A person’s blood groups mirror his genetic constitution. Several investigators14-l6 have reported an overrepresentation of blood group A (A and AB) among subjects with ischemic heart disease; this trend, although not significant, was found in our study. Some genetic traits are linked to disturbed lipid metabolism as well as high blood pressure; the familial tendency in ischemic heart disease may be due in part to such a disturbance although other factors are probably also responsible. Smoking: Extensive data from several prospective studies have demonstrated a significant relation between cigarette smoking and ischemic heart disease. It has also repeatedly been found that the risk is lower in subjects who have stopped smoking than in those who continue to smoke. 17The strong relation between the number of cigarettes smoked and the risk of ischemic heart disease in one study indicated a causal relation, not merely a genetic predisposition to ischemic heart disease. Prestonls has shown that much of the changing incidence of ischemic heart disease in different countries, as evident in the increasing male mortality, can be attributed to a preceding increase in tobacco consumption. Alcohol consumption: Even though some investigators have suspected that excessive alcohol consumption might cause cardiac damage, drinking alcohol has not been consistently shown before to increase the risk of ischemic heart disease. As mentioned, we did not study the amount of alcohol consumed during a specific period but an index of alcohol abuse that is related to social or psychological disturbances. This index was also related to tobacco and coffee consumption and, upon multivariate analysis, was shown to increase the risk of ischemic heart disease independently of these two other factors.lg At
Volume 35
RISK FACTORS
least part of the association of alcohol abuse with death from other causes in our study was due to the high rate of violent deaths among subjects who misuse alcohol. It should be pointed out that the per capita alcohol consumption according to official statistics is 5.6 liters of 100 percent alcohol in Sweden compared with 16.7 liters in France and 6.3 liters in the United States. The inhabitants of Gdteborg are representative of Sweden in this respect. Physical activity: We found no relation between physical inactivity during work and ischemic heart disease in our study, but in part this may be an expression of a limited variation in physical strain during ordinary work in an industrialized city. Our results are at variance with those of others.2c-24 A thorough analysis of both occupational and leisure time activity was performed in 1967.25 The follow-up data from that investigation showed a trend toward reduced effort during leisure time activities but not during occupational activity among those who later had ischemic heart disease. Dyspnea on exertion: There was an increased risk of fatal ischemic heart disease among subjects complaining of dyspnea on exertion. That this relation was associated with factors other than smoking has been shown earlier.lg Some subjects may have difficulty in differentiating between chest pain and dyspnea, but chest pain, as determined by a standardized questionnaire, was infrequent at age 50 years in this population. Thus, early dyspnea may reflect early deterioration of myocardial function due to incipient coronary vascular disease. 26 Blood pressure: In this study, as in several other prospective studies, blood pressure, and especially systolic blood pressure, was associated with an increased risk of premature ischemic heart disease. According to our data the risk was almost entirely related to fatal ischemic heart disease, Subjects with markedly increased blood pressure in 1963 were given treatment that might have caused some underestimation of the effect of blood pressure. The slightly higher heart rate and heart size on Xray examination in those who had ischemic heart disease might, at least in part, be related to the increased blood pressure in this group. Positive correlations between blood pressure and heart rate and between blood pressure and heart volume were found in this series. Vital capacity: Only 2 of the 19 men who died of ischemic heart disease had values for peak expiratory flow rate above the median for the entire study population in 1963 (Fig. 6). It is conceivable that low peak flow values indicate only excessive cigarette smoking. This is probably not the explanation since subjects who died of ischemic heart disease were not heavier smokers than those with nonfatal ischemic heart disease. Recently, Keys et a1.27 found that decreased lung function is a predictor of ischemic heart disease. However, in our study, this factor was found to be insignificant when studied with other factors in multivariate analysis. Nonfatal and fatal ischemic heart
FOR DISEASE-TIBBLIN
ET AL.
disease were not studied separately in the investigation of Keys et a1.27 Obesity: Prospective population studies in the United States have shown that obesity increases the risk of ischemic heart disease.24 In our subjects obesity by itself was not a risk factor for ischemic heart disease. This finding is even more remarkable when one considers that obese persons are more likely than lean persons to be hypertensive, hyperlipidemic or hyperglycemic. However, the degree of obesity was less severe in our sample than in the Framingham study. Serum lipids: As in a series of other prospective population studies,28 our study indicated that serum cholesterol was a significant risk factor for ischemic heart disease. The study also indicated that fatal ischemic heart disease is more strongly associated with increased serum cholesterol values than is nonfatal ischemic heart disease. Serum triglyceride values were related to serum cholesterol values (correlation coefficient = 0.35). As revealed in an earlier multivariate analysis with smoking, cholesterol, triglycerides, blood pressure, alcoholic intemperance and physical activity,lg triglyceride values in our study did not add further information of predictive importance for nonfatal and fatal ischemic heart disease when cholesterol values were known. Hyperglycemia: Other studies 2L,2g,30 have shown that disturbances of carbohydrate metabolism are associated with an increased risk of ischemic heart disease. Since fasting blood sugar determinations are fairly insensitive indicators of such disturbances, measurements of glucose and insulin after administration of a standardized dose of glucose have been advocated. Such an investigation was performed in a subsample of our group and the results were compared with those in nonselected patients with nonfatal myocardial infarction. 31 The data indicated that disturbances of this kind play only a small role in the increased risk for ischemic heart disease. Manifest clinical diabetes was present in 0.8 percent in our sample of 50 year old men. This finding is associated with an increased risk according to several other investigations as well as ours.32,33 Multivariate analysis of risk factors: Many factors are implicated in the development of ischemic heart disease, and it is possible that several of these factors are interrelated. The relations among various factors in this study population were indicated earlier in this paper and analyzed in detail. IsGenerally, risk factors that are not associated with each other in bivariate analysis are also shown to be risk factors upon multivariate analysis. Thus, it is probable that the findings indicated here will also be revealed upon multivariate analysis. The latter cannot yet be performed because there are too few end-point cases when the groups are differentiated by the present methods. Conclusions of study: In accordance with results from other studies, and earlier results in our study,
April 1975
The American Journal of CARDIOLOGY
Volume 35
521
RISK FACTORS
FOR DISEASE-TIBBLIN
ET AL.
smoking, alcoholic intemperance, dyspnea on exertion, elevated blood pressure, and serum cholesterol levels and diminished peak expiratory flow rate were found to be related to ischemic heart disease. There was also a trend toward more abnormal findings for all these variables except smoking habits in subjects who later had fatal ischemic heart disease than in those who had nonfatal myocardial infarction. In future analyses of the results of prospective studies, it is thus mandatory to differentiate between these two manifestations of ischemic heart disease that may, at least in part, have different clinical backgrounds. For several of the studied variables, except dyspnea on exertion, heart rate and heart volume, there
was also a trend toward differences between the group with no disease and the groups with nonfatal myocardial infarction or death from causes other than ischemic heart disease. A search for similar trends in other prospective studies is recommended. Such findings may indicate that there is a nonspecific influence of risk factors on nonischemic heart disease or that the risk factor pattern merely mirrors a general liability to greater risk of disease, ischemic as well as nonischemic. Acknowledgment We thank Karin Winqvist, MSc, for valuable discussion of the statistical methods and for assistance with data processing.
References 1. Tibblin G: High blood pressure in men aged 50-population study of men born in 1913, Acta Med Stand [Suppl] 470:1025, 1967 2. Tibblin G: A population study of 50-year-old men. An analysis of the nonparticipation group. Acta Med Stand 176453-459, 1965 3. Nilsson L-A, Ryttinger L, Tibblin G: Distribution of the ABO, MN, Rh, Duffy and Kell blood groups in a random sample of Swedish men aged fifty. Acta Pathol Microbial Stand 68: 117-l 22, 1966 4. Rose GA, Blackburn H: Cardiovascular Survey Methods. Geneva, World Health Organization, 1968, p 176 5. Wright BM, McKerrow CB: Maximum forced expiratory flow rate as a measure of ventilatory capacity. Br Med J 2:10411047, 1959 6. Jonsell SA: A method for the determination of the heart size by teleroentgenography (a heart volume index). Acta Radio1 (Stockh) 20:325-340, 1939 7. Gertler MM, WhHe TD: Coronary heart disease in young adults. Cambridge, Mass, Harvard University Press, 1954, p 23, 54 multia. Forssman 0, Lindegard B: The post coronary patient-a disciplinary investigation of middle-aged Swedish males. J Psychosom Res 369-169, 1958 9. Harvald B: Genetic factors in coronary heart disease. In, Coronary Heart Disease and Physical Fitness (Larsen AO, Malmborg RO, ed). Copenhagen, Munksgaard, 197 1, p 229-234 10. Slack J, Evans KA: The increased risk of death from ischaemic heart disease in first degree relatives of 121 men and 96 women with ischaemic heart disease. J Med Genet 3:239-257, 1966 11. Epstein FH: Risk factors in coronary heart disease-environmental and hereditary influences. Isr J Med Sci 3:594-607, 1967 12. Hammond EC, Garflnkel L, Seidman H: Longevity of parents and grandparents in relation to coronary heart disease and associated variables. Circulation 43:31-44, 197 1 13. Lundman T: Genetic predisposition of ischaemic heart diseasetwin studies. In, Early Phases of Coronary Heart Disease (Waldenstrom J, Larsson T, Ljungstedt N, ed). Stockholm, Nordiska Bokhandeln, Skandia International Symposia, 1973, p 186-195 14. Bronte-Stewart B, Botha MC, Krut LH: ABO blood groups in relation to ischaemic heart disease. Br Med J 2:1646-1650, 1962 15. Vogel F, Kruger J: Statistische Beziehungen zwischen den ABO-Blutgruppen und Krankheiten mit Ausnahme der Infektionskrankheiten. Blut 16:351-376, 1966 16. Medalie J, Levine C, Neufeld H, et al: Blood groups, cholesterol and myocardial infarction. Lancet 2:723, 1970
522
April 1975
The American Journal of CARDIOLOGY
17. The Health Consequences of Smoking. A Report of the Surgeon General. US Department of Health, Education, and Welfare, Washington DC, 1972, p 2 18. Preston SH: Older male mortality and cigarette smoking. A demographic analysis. Population Monograph Series No. 7. Berkeley, Calif.. University of California, Berkeley, 1970 19. Wilhelmsen L, Wedel H, Tlbblfn G: Multivariate analysis of risk factors for coronary heart disease. Circulation 46950-958, 1973 20. Morris J, Heady J, Raffle P, et al: Coronary heart disease and physical activity of work. Lancet 2: 1053-1057, 1953 21. Zukel WJ, Lewis RH, Enterline PE, et al: A short-term community study of the epidemiology of coronary heart disease. A preliminary report on the North Dakota Study. Am J Public Health 49:1630-1639, 1959 22. Brunner D, Manells G: Myocardial infarction among members of communal settlements in Israel. Lancet 2: 1049-1050, 1960 23. Taylor HL, Klepetar E, Keys A, et al: Death rates among physically active and sedentary employees of the railroad industry. Am J Public Health 52: 1697-1707, 1962 24. Kannel WB, Gordon T: Assessment of coronary vulnerability the Framingham study. In Ref 13, p 123-143 25. Wilhelmsen L, Tibblin G: Fysisk aktivitet och risk for hjartinfarkt. Lakartidningen (Stockh) 72: 343-345, 1975 26. Wilhelmsen L: Early diagnosis of coronary heart disease. In, Third International Symposium on Atherosclerosis. Berlin, Heidelberg and New York, Springer Verlag, 1974, p 705-707 27. Keys A, Aravanis C, Blackburn H, et al: Lung function as a risk factor for coronary heart disease. Am J Public Health 62: 1506-1511, 1972 28. Inter-Society Commission for Heart Disease Resources: Primary prevention of the atherosclerotic diseases. Circulation 42: A55-A95, 1970 risk factor in coronary dis29. Epstein FH: “Hyperglycemia’‘-a ease. Circu!ation 36:609-619, 1967 30. Gerber MM, Leetma HE, Saluste E, et al: lschaemic heart disease. Insulin, carbohydrate, and lipid interrelationships. Circulation 46:103-111, 1972 31. Bjorntorp P, Berchtold P, Grimby G, et al: Effects of physical training on glucose tolerance, plasma insulin and lipids and on body composition in men after myocardial infarction. Acta Med Stand 192:439-443, 1972 32. Wilhelmsen L: The Myocardial Infarction Clinic in Goteborgorganization and preliminary results. Pehr Dubb J (Goteborg) 3: 43-54, 1969 33. Bengtsson C, Bjorntorp P. Blohme G, et al: lschaemic heart disease in women. Acta Med Stand Suppl549:65-74, 1973
Volume 35
Risk Factors for Myocardial Infarction and Death Due to lschemic Heart Disease and Other Causes GijSTA TIBBLIN, MD LARS WILHELMSEN, MD LARS WERKij, MD, FACC Gteborg,
Sweden
From the Section of Preventive Cardiology, Medical Department I, Sahlgren’s Hospital, Goteborg, Sweden. This study was supported by grants from the Swedish Association Against Heart and Chest Diseases, and the Forenade Liv Insurance Company. Manuscript accepted September 12, 1974. Address for reprints: Gosta Tibblin, MD, Medical Department I, Sahlgren’s Hospital, S-413 45, Goteborg, Sweden.
514
April 1975
As part of a study of the male population in an industrial city in Sweden, one third of all male inhabitants of Goteborg born in 1913 were invlted to an examination in 1963. Of those invited, 655 (66 percent) accepted. This report examines the incidence of nonfatal myocardial infarction and death from ischemic heart disease and other causes in this group of men during the ensuing 10 years. There were 61 deaths; autopsy was performed in 56 cases. Nineteen men died of ischemic heart disease and 16 of cancer; 12 men died violently. Thirty-one men survived an acute myocardial infarction. Cigarette smoking and registration with the Temperance Board at the time of the initial examination were more common in men who later had a nonfatal myocardial infarction or died of ischemic heart disease or other causes than in surviving subjects and men who did not have an infarction. Dyspnea was more common in men who died of ischemit heart disease but was less common in those who died of other causes than in the remaining subjects. Values for systolic blood pressure were higher and those for peak expiratory flow lower in men who died of ischemlc heart disease. Serum cholesterol values were higher and those for serum triglycerides tended to be higher in men who died of ischemic heart disease than in other subjects. Heart size tended to be greater in those who had nonfatal or fatal ischemic heart disease. Obesity, the level of physical activity, fasting blood glucose levels, coffee consumption, hematocrit and erythrocyte sedimentation rate as determined at age 50 years had no predictive value for assessing the risk of nonfatal myocardial infarction, fatal ischemic heart disease or death from other causes before age 60. The results indicate that many so-called risk factors have a different relation to fatal than to nonfatal ischemic heart disease.
Investigators are agreed that age has an important bearing on the development of ischemic heart disease although the rather small incidence of this disease, especially in young persons, has made it difficult to study populations within narrowly defined age groupings. In most studies, subjects have been classified within age spans of 10 or more years; frequently the age ranges at entry have been 30 to 49 or 50 to 59 years, and there has been no clear description of the distribution of the population sample within these age ranges. The effect may be to skew data on risk factors and other variables, although most investigators try to compensate mathematically for possible distortion. This possibility constitutes one good reason for eliminating the age factor in prospective studies. Another is that several of the conventional risk factors, such as serum cholesterol and blood pressure, are in themselves age-related. Furthermore, all risk factors carry a different weight at different ages. For example, many studies may underestimate the importance of smoking, which tends to decrease with age, or exaggerate the significance of increased blood pressure or cholesterol, which tends to increase with age.
The American Journal of CARDIOLOGY
Volume 35
RISK FACTORS
The present study is part of a long-term investigation of the development of ischemic heart disease and other diseases in a random population sample of the same age and sex. It is intended to answer the following questions: Are there differences among groups classified as having (1) nonfatal myocardial infarction, (2) fatal ischemic heart disease, (3) death from other causes, or (4) no disease? If so, can such differences be explained by differences between one or more of the first three groups and the group with no disease? Furthermore, are there differences between the group with nonfatal myocardial infarction and the group with fatal ischemic heart disease?
Study Population and Methods All inhabitants of Sweden have a national registration number that includes their date of birth and other vital statistics. Names, addresses and registration numbers are registered with the official county census bureau and were accessible before the sample was drawn for the present study. The study population was recruited from men living in Giiteborg, Sweden, born in 1913 and still alive at age 50 years (1963). All men meeting these criteria who were born on a date divisible by three-that is, the third, sixth, ninth day (and so on) of each month-comprised the study sample. Nine hundred seventy-three men met these criteria. Of these 855 (88 percent) agreed to be examined in 1963 at Sahlgren’s Hospital, Giiteb0rg.l Of the 118 nonparticipants, 7 had died, 4 were hospitalized and 9 did not live in the town at the time of the investigation. The remaining 98 men refused study for various reasons, usually because of a negative attitude toward medical care. Compared with participants, nonparticipants had a lower mean income, were more frequently unmarried and were more often registered with the Temperance Board. Slightly more nonparticipants than participants had obtained sick allowance.2 For most of the participants (847 persons) complete records for the present analysis were available. It was also possible to trace the death certificates of the men in the overall study sample who died in Gijteborg between 1950 and 1963, that is, between the ages of 37 and 50 years. A total of 33 deaths had occurred, only 4 of which had been listed as due to ischemic heart disease. A detailed medical history was obtained from each participant, and all clinical examinations were performed by one of us (G.T.). Examinations were performed in the morning, with the subjects fasting until after blood sampling. A parental death score was constructed, an analysis facilitated by the identical age at entry (50 years) of all participants in the study. The cause of the parents’ death was not taken into account because thk data could not, be verified. Participants received 1 point if a parent had died before age 65 years, 2 points if the parent had died between the ages of 66 and 79 years, 3 points if the parent was living and aged 66 to 79 years and 4 points if the parent was living or had died after age 80 years. Thus, a subject received a minimal score of 2 if both parents had died before age 65 years and a maximal score of 8 if both parents were still alive or had died after age 80 years. Blood groups in the ABO, Rh, Kell, Duffy and MN systems were determined according to methods described earlier.3 For analysis of smoking data the subjects were grouped
FOR DISEASE-TIBBLIN
ET AL.
as follows: never smoked, stopped smoking, smoking 1 to 14 cigarettes/day, smoking 15 to 24 cigarettes/day and smoking 25 or more cigarettes/day. Those who smoked only cigars, “cigarillos” or a pipe (75 subjects) were classified in a single group without consideration of their exact coasumption. In part of the analysis those who had never smoked and former smokers were grouped together, those who smoked 1 to 14 cigarettes a day were referred to a second group and those who smoked 15 cigarettes a day or more were referred to a third group. Coffee consumption was assessed by interview and recorded as the mean number of ordinary cups consumed each day. Reliable data concerning alcohol consumption are difficult to obtain from interviews, but in Sweden records of local temperance boards provide some guide to alcoholic intemperance. Board registration may cover the spectrum from severe alcoholism to a single offense of driving a car with a high level of alcohol in the blood. The records provide minimal data since heavy alcohol consumption that may be socially acceptable or take place only in the home does not lead to board registration. Nevertheless, 21 percent of this representative population sample of 50 year old men had at some time been registered with the temperance board. Physical activity during work was classified as sedentary, moderate or heavy according to the type of occupation held at age 50 years. Dyspnea on exertion was defined according to the criteria of Rose and Blackburn. Subjects experiencing dyspnea when hurrying on level ground or when walking up a small hill or undertaking other less taxing exertion were classified as having dyspnea. Blood pressure was measured by the same observer throughout the study after a 5 to 10 minute interview with the subject seated. Both systolic and diastolic blood pressure were measured to the nearest 5 mm Hg. Men found to have casual blood pressure readings of 175 mm Hg or more systolic and 115 mm Hg or more diastolic were given treatment. This fact has to be considered when high blood pressure is discussed as a risk factor in this study. Measures of overweight-obesity-were obtained by measuring skinfold thickness and calculating relative weight.. The latter was obtained by dividing the actual weight in kilograms by the height in centimeters minus 100. Heart rate at rest was determined from the resting electrocardiogram. Peak expiratory flow, as described by Wright and McKerrow,s was measured with patients standing; the mean of three measurements was used. Heart oolume was determined from full scale roentgenograms of the chest in the frontal and the oblique lateral projections according to the technique of Jonsell. All films were interpreted by the same roentgenologist. The absolute figure in milliliters was related to the body surface area in square meters. The fasting levels of serum cholesterol and serum triglycerides, blood sugar, hematocrit and erythrocyte sedimentation rate during 1 hour were determined at the same laboratory, and regular checks were performed to determine the accuracy of these methods.’ The criterion used for diagnosis of acute ischemic heart disease was (1) hospitalization with the clinical diagnosis of myocardial infarction, or (2) fresh ischemic heart disease at autopsy. The diagnoses were established by the clinicians or pathologists, respectively. Since November 1968, all clinical cases of myocardial infarction have been record-
April 1975
The American Journal of CARDIOLOGY
Volume 35
515
RISK FACTORS
TABLE
FOR DISEASE-TIBBLIN
ET AL.
In the total group of men studied, there were 61 deaths. Autopsy, usually performed in a standardized fashion, was carried out in 56 cases. The data were analyzed for the three end point groups: nonfatal myocardial infarction, fatal ischemic heart disease and death from other causes. If none of these end points occurred during the 10 year follow-up period the subjects were classified as having no disease. For the statistical study chi-square analysis was used to test the equality of distributions of quantitative variables in the four groups. No further statistical analysis was performed if significant differences were found. Analysis of variance was used to test equality of means. If significant
I
Number of Subjects Who Died or Had Nonfatal Myocardial Infarction Number Deaths Fatal ischemic heart disease Fresh myocardial infarction Other acute ischemrc heart disease Pulmonary carcinoma Other carcinoma Other nonviolent death Violent death Nonfatal myocardral infarction * Subjects
who drd not undergo
11 8 (l*) 9 (I*) 9 (2*) 12 12 (l*) 31
differences between means were found, these differences were localized by the Scheffe method. The significance level of P = 0.05 was used.
ed in collaboration with the World Health Organization in a Myocardial Infarction Register. The clinical criteria for myocardial infarction used in this study were those adopted by the Swedish Society of Cardiology: central chest pain, shock or syncope suggesting a myocardial infarction together with a typical transaminase spectrum or appearance of a pathologic Q wave or localized S-T changes in the electrocardiogram. The criterion for ischemic heart disease at autopsy was a fresh myocardial scar or, in the absence of any macroscopically observed scar, total or almost total occlusion of a coronary artery together with a medical history suggesting myocardial infarction or sudden coronary death.
TABLE
II
Clinical Course of Subjects lschemic Heart Disease Case IlO.
Who Died of
Durabon of Acute Symptoms
Hosprtalrzabon
A. Fresh Myocardral Infarctron 1 2 3 4 5 6 7 8 9
Days Days Days Days 1 l/2 hours 10 hours 15 hours Days 1
10 11
Day Days
516
No Yes, 2 hours Yes, 1 day Yes, 9 days No Yes, other disease No Yes, 1 hour Yes, operation for infarction Yes, 1 day Yes, 13 days
B. Acute lschemic Heart Disease ~No Found dead No Seconds No l/2 hour No Seconds Yes, other disease Seconds l/4 hour No No Seconds No Seconds
1 2 3 4 5 6 7 8
April 1975
The group with no disease was separated into quartiles (25 percent of the subjects in each group) according to values for systolic blood pressure, peak expiratory flow and serum cholesterol. The cutoff points obtained were then used to define the distributions in the three end point groups (see Fig. 5 to 7).
autopsy.
The American Journal of CARDIOLOGY
Results Fatal and Nonfatal lschemic
Heart Disease
This analysis concerns findings in the 10 year follow-up period from entry in 1963 to the end of June 1973. The number of events-nonfatal myocardial infarction and death from ischemic heart disease or other causes-is shown in Table I. Of the five subjects who died but did not undergo autopsy, one had a clear-cut sudden cardiac death and was classified in the group with ischemic heart disease. Of the other four, three died of cancer and one had a fatal accident. The clinical course of the men who died of ischemic heart disease is presented in Table II. Thirty-one men with nonfatal myocardial infarction were treated at the hospital. Of these 19 subjects who died with a diagnosis of ischemic heart disease, all except 1 had severe coronary atherosclerosis (Fig. 1). Seven of the 11 who had a fresh myocardial infarction at autopsy also had an old scar indicating an earlier myocardial infarction. The majority of those who died of other causes had few signs of coronary arterial changes or myocardial scars. Two subjects with fresh myocardial infarction at autopsy had died with other clinical diagnoses, extensive bronchial carcinoma in one, and severe aortic stenosis in the other. Most of those with a fresh myocardial infarction at autopsy had a history of chest pain lasting for at least 1 hour. Eight of 11 were hospitalized. Of those who died of acute ischemic heart disease all died within 30 minutes and none were hospitalized because of ischemit heart disease. Risk Variables Distributions for qualitative variables and means and standard deviations for quantitative variables are shown in Tables III and IV, respectively. The parental death score did not differ significantly among the four groups (chi-square = 15.0; df = 9)
Volume 35
RISK FACTORS
Fatal
lschemic
Heart
FOR DISEASE-TIBBLIN
ET AL.
Other Deaths
Disease
n=19
n=L2
Severe coronary at herosclerosis
Severe coronarv rtherosclerosis’
Fresh M I
Fresh M I
0
Nonsmokers
1-IL g/day
215 g/day
Non-fatal Myocardial
Infarction
n=31
u 1
FIGURE 1. Autopsy data on the coronary arteries and myocardium of subjects who died of fatal ischemic heart disease and other causes. Autopsy was performed in 18 and 38 cases, respectively. Fresh Ml = recent myocardial infarction. _
UScar
NonI-U smokers g/day Fatal Heart
El5
Non-
g/day
smokers
lschemic Disease
01 her
1-u g/day Deaths n=L2
?I5 g/day
NonI-14 smokers g/day
Scar
?15 g/day
No Disease n=755
n.19
FIGURE 2. Distribution of smoking habits in the four groups. g = cigarettes.
although the score tended to be lower in the group with fatal ischemic heart disease; that is, their parents died at an earlier age. There was a trend, although it was not significant, toward a higher frequency of blood group A and lower frequency of group 0 among those who had nonfatal or fatal ischemic heart disease. Similar trends were not observed in the other blood group systems. Smoking habits differed significantly among the four groups (chi-square = 20.1; df = 6). This was due to a higher proportion of smokers in all three end point groups than in the group with no disease (Fig. 2). In the group that died of causes other than ischemit heart disease, the overrepresentation of smokers among those who later had pulmonary carcinoma could account for the result. There were significant differences among the four groups with respect to registration by the temperance board (chi-square = 11.4; df = 3). This finding was due to an overrepresentation in all end point groups (Fig. 3).
The type of physical activity during work did not differ significantly among the four groups (chi-square = 6.5; df = 6). Dyspnea on exertion differed significantly among groups (chi-square = 14.2; df = 3). There was a high rate of dyspnea among those who later had fatal ischemit heart disease, but the rate was lower than expected among subjects who died of other causes (Fig. 4). With respect to blood pressure, there was a significant difference among the mean values for systolic (F = 4.65; df = 3) and diastolic (F = 3.05; df = 3) blood pressure (Fig. 5). For systolic but not for diastolic blood pressure the difference was localized between those with fatal ischemic heart disease and those with no disease. The mean values for heart rate at rest did not differ significantly among groups (F = 1.63; df = 3), but the group with fatal ischemic heart disease tended to have a higher heart rate. The mean values for heart volume did not differ significantly (F = 1.25; df = 3), but volume tended to
April 1975
The American Journal of CARDIOLOGY
Volume 35
517
RISK FACTORS
FOR DISEASE-TIBBLIN
ET AL.
be greater in the two groups with ischemic heart disease. The mean values for peak expiratory flow rate differed significantly among the four groups (F = 5.32; df = 3). This difference was localized between the group with fatal ischemic heart disease, which TABLE
III
Qualitative
Variables
in Relation to Different
Parental death score 2-3 4-5 6-7 8 Blood group A B AB 0 Smoking habits (cigarettes) Nonsmoker (never or stopped) Smoking 1-14/day Smoking 15/day Registered by temperance board Yes No Physical activity during work Sedentary Moderate Heavy Dyspnea Yes No The total number
of subjects
Fatal lschemrc Heart Disease (no. = 19)
8 (26%) 18 (58%)
9 5 4 9
4 (13%) 1(3%)
18 (58%)
Other Deaths (no. = 42)
No Drsease (no. = 755)
14 (33%) 29 (48%) 7 (17%) 1(2%)
178 (23%) 337 (45%) 179 (24%) 5103%)
20 (48%)
(50%) (23%) (22%) (9%)
4 (13%) 9 (9%) 9 (29%)
10 (56%) 1(‘3%) 2 (11%) 5 (25%)
2 (5%) 1(2%) 19 (45%)
342 (45%) 77 (10%) 29 (4%) 306 (41%)
4 (13%) 15 (48%) 12 (39%)
5 (28%) 8 (44%) 5 (28%)
13 (31%) 17 (41%) 12 (28%)
350 (46%) 255 (34%) 149 (20%)
11(35%) 20 (65%)
7 (39%) 11(51%)
13 (31%) 29 (69%)
145 (19%) 609 (81%)
9 (29%) 16(52%) 6 (19%)
6 (4’3%) 4 (27%) 5 (33%)
14 (32%) 12 (29%) 16 (39%)
262 (35%) 235 (32%) 243 (33%)
9 (29%)
10 (56%)
9 (14%) 36 (85%)
162W%,t 593(79%)
22 (71%)
in the various subgroups
8 (44%)
does not always equal the total number
of the group because
of missing
IV
Quantitative
Variables
in Relation to Different
Variable Systolic blood pressure (mm Hg) Diastolic blood pressure (mm Hg) Heart rate Heart volume (ml/m2 BSA) (standing) Peak expiratory flow (liters/min) Serum cholesterol (mg/lOO ml) Serum triglycerides (millimoles/liter) Fasting blood sugar (mg/lOO ml) Erythrocyte sedimentation rate (mm in 1 hour) Hematocrit (%) Coffee consumption (cups/day) As in Table III there are occasional BSA = body surface area.
518
End Points
Nonfatal Myocardral Infarctron (no. = 31)
Vanable
TABLE
had a low mean value, and the group with no disease (Fig. 6). The measures of obesity, skinfold thickness and relative weight, showed no relation with the end points. The mean values for both serum cholesterol and
April 1975
missing
End Points (Mean i
standard
deviation)
Nonfatal Myocardial Infarction (no. = 31)
Fatal lschemic Heart Disease (no. = 19)
Other Deaths (no. = 42)
144 f 30 95 i 15 86 * 12 408 =!z 71 487 zt 88 260 i 37 1.36 X!Z0.53 83 i 15
152 z+z33 98 + 15 95 i 16 402 I!= 68 436 & 106 279 z!= 37 1.59 i 0.57 88 i 18
143 i 94 + 92-f 384 +I 490 * 257 & 1.41 i 84Jr
9& 8 46 I!= 2 5.7 f 2.4
lOi 9 45 f 3 5.9 z!z 3.1
data that do not affect the statistical
The American Journal of CARDIOLOGY
Volume 35
23 17 18 64 99 49 0.76 17
81 10 46 & 4 5.4 ZIZ2.8 tests or conclusions.
No Disease (no. = 755) 137 i 91 It 91 It 391 f 506 & 245 + 1.23 & 82 f.
21 13 15 58 80 42 0.71 15
8f7 45 * 4 5.1 f 3.0
data.
RISK FACTORS
FOR DISEASE-TIBBLIN
ET AL.
% IOO-
n
80.60-LO--
-
20-O-
Non-fatal Fatal lschemic Other DeathsNo Disease Myocardral Heart Disease n=42 n=755 Infarction n=19 n:31 FIGURE 3. Registration by the temperance board in the four groups. + = registered: - = not registered.
B
-
l--tlnm +
l
-
Fatal
Non-fatal
+
lschemrc
Other
Deaths
No Drsease
Heart Disease Myocardral n=42 n-755 Infarction n=19 n.31 FIGURE 4. Dyspnea on exertion in the four groups. -I = present; = not present.
% 60-
-
40--
20--
FIGURE 5. Distribution of systolic blood pressure. The gr6up with no disease was classified by quartiles and the cutoff points obtained were then used for all groups.
-
I
O- . > Systolrc brooa 122 132 146 Dressure mmHg No of cases 7 4 11 9
122 132 146
122 132 146 4
2
4
8
Non-fatal
Fatal
lschemic
Myocardial
Heart
Disease
10
6
Other
9
17
Deaths
122 132 146 188 188 I88 188 No
Disease
Infarction
60
40
20
pop FIGURE 6. Distribution of peak expiratory flow rate (litersjmin). The group with no disease was classified by quartiles and the cutoff points obtained were then used for all groups.
0 f;,!‘$atury
No of cases
0l ln -
466 512 557
11
6
5
-
-_
466 512 557
8
Non-fatal
10
Fatal Heart
Myocardial Infarction
April 1975
6
0
2
lschemic Disease
The American
466 512 557
13
Other
9
8
II
Deaths
Journal of CARDIOLOGY
466 512 557 182 182 182 182 No
Disease
Volume 35
519
RISK FACTORS
FOR
Cholesterol mg/lOOml No of cases
215 211 271 .4
DISEASE-TIBBLIN
L
9
Non-fatal Myocardial lnfarctlon
ET AL.
215 2Ll 271
215 241 271 0
f4
Fatal Heart
1
9
8
lschemic Disease
6
10
Other
serum triglycerides differed significantly among groups (F = 6.00 and 2.67, respectively; df = 3). For serum cholesterol, the difference was localized between the group with fatal ischemic heart disease and the group with no disease (Fig. 7), and the same tendency, although not significant, was seen for serum triglycerides. The mean values for fasting blood sugar level did not differ significantly, nor did those for erythrocyte sedimentation rate or hematocrit. However, blood sugar level tended to be higher in the group with fatal ischemic heart disease. Coffee consumption did not differ significantly among groups but tended to be greater in the two groups with ischemic heart disease. Discussion The proportion of nonparticipants in this study is considered small. Nonparticipants were more often registered by the temperance board and, since there is a positive relation between alcohol consumption and smoking, they may very well be heavier smokers than participants. Since both smoking and alcohol abuse were associated with ischemic heart disease and the nonparticipants had a higher death rate, at least during the first part of the follow-up period, it is possible that these factors are more strongly related to ischemic heart disease than our results indicate. In other prospective studies in which the participation rate usually has been lower, the effect of socially related variables, such as excessive use of alcohol, may be more distorted than in the present investigation. Familial history: Several clinical studies7-i3 have suggested that ischemic heart disease tends to occur in families. A shortcoming of several of these studies is that it is necessary to collect the information from subjects who seldom know much about the cause of their parents’ death. Since most of the studies have concerned subjects from different age groups, it has been difficult to standardize for the age of the par-
520
April 1975
The American Journal of CARDIOLOGY
10 16 Deaths
al
215 241 271
188 188 188 188 No Disease
FIGURE 7. Distribution of serum cholesterol. The group with no disease was classified by quartiles and the cutoff points obtained were then used for all groups.
ents. In our study this standardization was possible. A familial predisposition to fatal ischemic heart disease was suggested by our results, although the data were not significant. Blood groups: A person’s blood groups mirror his genetic constitution. Several investigators14-l6 have reported an overrepresentation of blood group A (A and AB) among subjects with ischemic heart disease; this trend, although not significant, was found in our study. Some genetic traits are linked to disturbed lipid metabolism as well as high blood pressure; the familial tendency in ischemic heart disease may be due in part to such a disturbance although other factors are probably also responsible. Smoking: Extensive data from several prospective studies have demonstrated a significant relation between cigarette smoking and ischemic heart disease. It has also repeatedly been found that the risk is lower in subjects who have stopped smoking than in those who continue to smoke. 17The strong relation between the number of cigarettes smoked and the risk of ischemic heart disease in one study indicated a causal relation, not merely a genetic predisposition to ischemic heart disease. Prestonls has shown that much of the changing incidence of ischemic heart disease in different countries, as evident in the increasing male mortality, can be attributed to a preceding increase in tobacco consumption. Alcohol consumption: Even though some investigators have suspected that excessive alcohol consumption might cause cardiac damage, drinking alcohol has not been consistently shown before to increase the risk of ischemic heart disease. As mentioned, we did not study the amount of alcohol consumed during a specific period but an index of alcohol abuse that is related to social or psychological disturbances. This index was also related to tobacco and coffee consumption and, upon multivariate analysis, was shown to increase the risk of ischemic heart disease independently of these two other factors.lg At
Volume 35
RISK FACTORS
least part of the association of alcohol abuse with death from other causes in our study was due to the high rate of violent deaths among subjects who misuse alcohol. It should be pointed out that the per capita alcohol consumption according to official statistics is 5.6 liters of 100 percent alcohol in Sweden compared with 16.7 liters in France and 6.3 liters in the United States. The inhabitants of Gdteborg are representative of Sweden in this respect. Physical activity: We found no relation between physical inactivity during work and ischemic heart disease in our study, but in part this may be an expression of a limited variation in physical strain during ordinary work in an industrialized city. Our results are at variance with those of others.2c-24 A thorough analysis of both occupational and leisure time activity was performed in 1967.25 The follow-up data from that investigation showed a trend toward reduced effort during leisure time activities but not during occupational activity among those who later had ischemic heart disease. Dyspnea on exertion: There was an increased risk of fatal ischemic heart disease among subjects complaining of dyspnea on exertion. That this relation was associated with factors other than smoking has been shown earlier.lg Some subjects may have difficulty in differentiating between chest pain and dyspnea, but chest pain, as determined by a standardized questionnaire, was infrequent at age 50 years in this population. Thus, early dyspnea may reflect early deterioration of myocardial function due to incipient coronary vascular disease. 26 Blood pressure: In this study, as in several other prospective studies, blood pressure, and especially systolic blood pressure, was associated with an increased risk of premature ischemic heart disease. According to our data the risk was almost entirely related to fatal ischemic heart disease, Subjects with markedly increased blood pressure in 1963 were given treatment that might have caused some underestimation of the effect of blood pressure. The slightly higher heart rate and heart size on Xray examination in those who had ischemic heart disease might, at least in part, be related to the increased blood pressure in this group. Positive correlations between blood pressure and heart rate and between blood pressure and heart volume were found in this series. Vital capacity: Only 2 of the 19 men who died of ischemic heart disease had values for peak expiratory flow rate above the median for the entire study population in 1963 (Fig. 6). It is conceivable that low peak flow values indicate only excessive cigarette smoking. This is probably not the explanation since subjects who died of ischemic heart disease were not heavier smokers than those with nonfatal ischemic heart disease. Recently, Keys et a1.27 found that decreased lung function is a predictor of ischemic heart disease. However, in our study, this factor was found to be insignificant when studied with other factors in multivariate analysis. Nonfatal and fatal ischemic heart
FOR DISEASE-TIBBLIN
ET AL.
disease were not studied separately in the investigation of Keys et a1.27 Obesity: Prospective population studies in the United States have shown that obesity increases the risk of ischemic heart disease.24 In our subjects obesity by itself was not a risk factor for ischemic heart disease. This finding is even more remarkable when one considers that obese persons are more likely than lean persons to be hypertensive, hyperlipidemic or hyperglycemic. However, the degree of obesity was less severe in our sample than in the Framingham study. Serum lipids: As in a series of other prospective population studies,28 our study indicated that serum cholesterol was a significant risk factor for ischemic heart disease. The study also indicated that fatal ischemic heart disease is more strongly associated with increased serum cholesterol values than is nonfatal ischemic heart disease. Serum triglyceride values were related to serum cholesterol values (correlation coefficient = 0.35). As revealed in an earlier multivariate analysis with smoking, cholesterol, triglycerides, blood pressure, alcoholic intemperance and physical activity,lg triglyceride values in our study did not add further information of predictive importance for nonfatal and fatal ischemic heart disease when cholesterol values were known. Hyperglycemia: Other studies 2L,2g,30 have shown that disturbances of carbohydrate metabolism are associated with an increased risk of ischemic heart disease. Since fasting blood sugar determinations are fairly insensitive indicators of such disturbances, measurements of glucose and insulin after administration of a standardized dose of glucose have been advocated. Such an investigation was performed in a subsample of our group and the results were compared with those in nonselected patients with nonfatal myocardial infarction. 31 The data indicated that disturbances of this kind play only a small role in the increased risk for ischemic heart disease. Manifest clinical diabetes was present in 0.8 percent in our sample of 50 year old men. This finding is associated with an increased risk according to several other investigations as well as ours.32,33 Multivariate analysis of risk factors: Many factors are implicated in the development of ischemic heart disease, and it is possible that several of these factors are interrelated. The relations among various factors in this study population were indicated earlier in this paper and analyzed in detail. IsGenerally, risk factors that are not associated with each other in bivariate analysis are also shown to be risk factors upon multivariate analysis. Thus, it is probable that the findings indicated here will also be revealed upon multivariate analysis. The latter cannot yet be performed because there are too few end-point cases when the groups are differentiated by the present methods. Conclusions of study: In accordance with results from other studies, and earlier results in our study,
April 1975
The American Journal of CARDIOLOGY
Volume 35
521
RISK FACTORS
FOR DISEASE-TIBBLIN
ET AL.
smoking, alcoholic intemperance, dyspnea on exertion, elevated blood pressure, and serum cholesterol levels and diminished peak expiratory flow rate were found to be related to ischemic heart disease. There was also a trend toward more abnormal findings for all these variables except smoking habits in subjects who later had fatal ischemic heart disease than in those who had nonfatal myocardial infarction. In future analyses of the results of prospective studies, it is thus mandatory to differentiate between these two manifestations of ischemic heart disease that may, at least in part, have different clinical backgrounds. For several of the studied variables, except dyspnea on exertion, heart rate and heart volume, there
was also a trend toward differences between the group with no disease and the groups with nonfatal myocardial infarction or death from causes other than ischemic heart disease. A search for similar trends in other prospective studies is recommended. Such findings may indicate that there is a nonspecific influence of risk factors on nonischemic heart disease or that the risk factor pattern merely mirrors a general liability to greater risk of disease, ischemic as well as nonischemic. Acknowledgment We thank Karin Winqvist, MSc, for valuable discussion of the statistical methods and for assistance with data processing.
References 1. Tibblin G: High blood pressure in men aged 50-population study of men born in 1913, Acta Med Stand [Suppl] 470:1025, 1967 2. Tibblin G: A population study of 50-year-old men. An analysis of the nonparticipation group. Acta Med Stand 176453-459, 1965 3. Nilsson L-A, Ryttinger L, Tibblin G: Distribution of the ABO, MN, Rh, Duffy and Kell blood groups in a random sample of Swedish men aged fifty. Acta Pathol Microbial Stand 68: 117-l 22, 1966 4. Rose GA, Blackburn H: Cardiovascular Survey Methods. Geneva, World Health Organization, 1968, p 176 5. Wright BM, McKerrow CB: Maximum forced expiratory flow rate as a measure of ventilatory capacity. Br Med J 2:10411047, 1959 6. Jonsell SA: A method for the determination of the heart size by teleroentgenography (a heart volume index). Acta Radio1 (Stockh) 20:325-340, 1939 7. Gertler MM, WhHe TD: Coronary heart disease in young adults. Cambridge, Mass, Harvard University Press, 1954, p 23, 54 multia. Forssman 0, Lindegard B: The post coronary patient-a disciplinary investigation of middle-aged Swedish males. J Psychosom Res 369-169, 1958 9. Harvald B: Genetic factors in coronary heart disease. In, Coronary Heart Disease and Physical Fitness (Larsen AO, Malmborg RO, ed). Copenhagen, Munksgaard, 197 1, p 229-234 10. Slack J, Evans KA: The increased risk of death from ischaemic heart disease in first degree relatives of 121 men and 96 women with ischaemic heart disease. J Med Genet 3:239-257, 1966 11. Epstein FH: Risk factors in coronary heart disease-environmental and hereditary influences. Isr J Med Sci 3:594-607, 1967 12. Hammond EC, Garflnkel L, Seidman H: Longevity of parents and grandparents in relation to coronary heart disease and associated variables. Circulation 43:31-44, 197 1 13. Lundman T: Genetic predisposition of ischaemic heart diseasetwin studies. In, Early Phases of Coronary Heart Disease (Waldenstrom J, Larsson T, Ljungstedt N, ed). Stockholm, Nordiska Bokhandeln, Skandia International Symposia, 1973, p 186-195 14. Bronte-Stewart B, Botha MC, Krut LH: ABO blood groups in relation to ischaemic heart disease. Br Med J 2:1646-1650, 1962 15. Vogel F, Kruger J: Statistische Beziehungen zwischen den ABO-Blutgruppen und Krankheiten mit Ausnahme der Infektionskrankheiten. Blut 16:351-376, 1966 16. Medalie J, Levine C, Neufeld H, et al: Blood groups, cholesterol and myocardial infarction. Lancet 2:723, 1970
522
April 1975
The American Journal of CARDIOLOGY
17. The Health Consequences of Smoking. A Report of the Surgeon General. US Department of Health, Education, and Welfare, Washington DC, 1972, p 2 18. Preston SH: Older male mortality and cigarette smoking. A demographic analysis. Population Monograph Series No. 7. Berkeley, Calif.. University of California, Berkeley, 1970 19. Wilhelmsen L, Wedel H, Tlbblfn G: Multivariate analysis of risk factors for coronary heart disease. Circulation 46950-958, 1973 20. Morris J, Heady J, Raffle P, et al: Coronary heart disease and physical activity of work. Lancet 2: 1053-1057, 1953 21. Zukel WJ, Lewis RH, Enterline PE, et al: A short-term community study of the epidemiology of coronary heart disease. A preliminary report on the North Dakota Study. Am J Public Health 49:1630-1639, 1959 22. Brunner D, Manells G: Myocardial infarction among members of communal settlements in Israel. Lancet 2: 1049-1050, 1960 23. Taylor HL, Klepetar E, Keys A, et al: Death rates among physically active and sedentary employees of the railroad industry. Am J Public Health 52: 1697-1707, 1962 24. Kannel WB, Gordon T: Assessment of coronary vulnerability the Framingham study. In Ref 13, p 123-143 25. Wilhelmsen L, Tibblin G: Fysisk aktivitet och risk for hjartinfarkt. Lakartidningen (Stockh) 72: 343-345, 1975 26. Wilhelmsen L: Early diagnosis of coronary heart disease. In, Third International Symposium on Atherosclerosis. Berlin, Heidelberg and New York, Springer Verlag, 1974, p 705-707 27. Keys A, Aravanis C, Blackburn H, et al: Lung function as a risk factor for coronary heart disease. Am J Public Health 62: 1506-1511, 1972 28. Inter-Society Commission for Heart Disease Resources: Primary prevention of the atherosclerotic diseases. Circulation 42: A55-A95, 1970 risk factor in coronary dis29. Epstein FH: “Hyperglycemia’‘-a ease. Circu!ation 36:609-619, 1967 30. Gerber MM, Leetma HE, Saluste E, et al: lschaemic heart disease. Insulin, carbohydrate, and lipid interrelationships. Circulation 46:103-111, 1972 31. Bjorntorp P, Berchtold P, Grimby G, et al: Effects of physical training on glucose tolerance, plasma insulin and lipids and on body composition in men after myocardial infarction. Acta Med Stand 192:439-443, 1972 32. Wilhelmsen L: The Myocardial Infarction Clinic in Goteborgorganization and preliminary results. Pehr Dubb J (Goteborg) 3: 43-54, 1969 33. Bengtsson C, Bjorntorp P. Blohme G, et al: lschaemic heart disease in women. Acta Med Stand Suppl549:65-74, 1973
Volume 35
