Adv. Cardiol., vol. IS, pp. 217-230 (Karger, Basel 1976)
Physical Activity, Physical Fitness and Risk of Myocardial Infarction 1 L. WILHELMSEN, G. TIBBLIN, M. AURELL, J. BlURE, B. EKSTROM-JODAL and G. GRIMBY Medical Department I and Department of Clinical Physiology I, Sahlgren's Hospital, Gothenburg
l. Introduction
Supported by grants from the Swedish National Organization against Heart and Chest Diseases, and the Bank of Sweden Tercentenary Fund.
Downloaded by: Université René Descartes Paris 5 193.51.85.197 - 1/28/2018 10:34:37 AM
An essential prerequisite for the prevention of coronary heart disease (CHD) is demonstration of relationships between personal characteristics and risk of manifest CHD. Some of the risk factors detected in this way may only be indicators of an early, subclinical disease, while other risk factors may have causative significance and be amenable to treatment which may prevent or postpone the disease. An example of the first type of risk factor is ECG changes, and examples of the second type are tobacco smoking, hypertension and hyperlipidemia. MORRIS et al. [12] showed that the risk of CHD was lower in London bus conductors than in bus drivers. Later, a long series of studies by Fox and HASKELL [6], KANNEL and McNAMARA [10], KANNEL et al. [11] and CASSEL et al. [3], as well as others, showed that high physical activity was associated with lower risk for future CHD than low physical activity. The earlier studies mainly concerned occupational activity. Later studies have also investigated the importance of leisure-time activity. An important recent study into this question was performed by MORRIS et al. [13], who studied 17,000 civil servants and related their leisure-time exercise habits to subsequent CHD after adjustment for age and smoking habits. These authors did not find any CHD differences until the highest classification (frequent peak activities to a level of 7.5 kcal/min) was reached. In the studies of both KANNEL et al. [10, 11] and MORRIS et al. [12, 13], leisure-time activity was
WILHELMsENjTmBLINjAURELLjBJUREjEKSTROM-JODALjGRlMBY
218
of significant importance even in multivariate analysis when other confounding factors were taken into consideration. There is a relationship between physical activity, especially during leisure time, and physical fitness, but several constitutional factors seem to be of great importance in the degree of physical fitness. No studies have been published which have related physical fitness to subsequent CHD. The prognostic significance of ECG changes during rest or after exercise has, however, been reported [1, 2, 4, 14, 15]. In the present paper, studies concerning physical inactivity as a risk factor in middle-aged men from the general population in G6teborg, Sweden, are reported. In addition, the prognostic signification of an exercise test in a general population sample of men has been analyzed. The relatively large number of persons studied within a very narrow age span facilitated the analysis.
The study involving men born in 1913. The study population was recruited from men living in Gothenburg, Sweden, born in 1913 and still alive at the age of 50 years in 1963. All men meeting these criteria who were born on dates divisible by three, i. e. the 3rd, 6th, 9th (etc.) of each month, comprised the study sample. These criteria were met by 973 men. Of them, 855 (88 %) agreed to be examined in 1963, and 803 were reexamined in 1967. Analyses of the participants and the non-participants have been published previously [17, 18, 21]. In 1967, a comprehensive interview concerning physical activity at work and during leisure time in lO-year intervals from age of 20 to the age of 54 was performed by one interviewer (L. W.). The questionnaire has been described previously [16]. It was later modified for use as a postal questionnaire [19]. The exercise test was performed on an electrically braked bicycle ergometer. The subjects cycled for 6 min at each load and the load was increased stepwise, starting with 49 W (300 kpmjmin) and increasing to 98 W (600 kpmjmin). According to the protocol some subjects were tested at 147 W (900 kpmjmin) and finally at a maximal load. The procedure has been described in detail previously [8, 9]. The work performed in kilopondmeters per minute at a heart rate of 150 beats per minute (W150) was estimated by interpolation. Maximum work load was defined as the subjects' apprehension of exhaustion or a plateau of heart rate or both when increasing the work load in small steps. This level was reached by 84 % of subjects. Submaximal and maximal oxygen uptake and ventilation were determined in a 10 % subsample and a formula was constructed for calculation of maximal oxygen uptake in the total sample [7]. Heart volume determined by X-ray examination was not better correlated with maximal oxygen uptake than any combination of anthropometric variables. The primary preventive trial in Gothenburg. One third of all men born between
Downloaded by: Université René Descartes Paris 5 193.51.85.197 - 1/28/2018 10:34:37 AM
II. Methods and Population Samples
Risk of Myocardial Infarction
219
1915-1922 and 1924-1925 in Gothenburg are included in the intervention group of a primary preventive trial which began in 1970. The postal questionnaire was answered by 83 % of the 10,000 men, and 75 % agreed to examination. The study design and an analysis of the non-participants has been published [20, 22]. The follow-up of end-point cases in both investigations has been done by the collection of death certificates in the city (autopsy rate 90 %) and since 1968 by a special Myocardial Infarction Register [5]. Standard statistical methods were used for x2-analysis, Student's t-test, and regression analysis.
An important point concerning our questionnaire on physical activity is whether it can detect true differences in physical activity. With the aid of the exercise test, we studied how the different types of activity were related to physical working capacity. The percentage of men who could perform the maximal exercise test increased from 63 % in the group who had the lowest physical activity both in their occupation (grade 1) and during leisure time (grade 1) up to 95% among those who had high activity both in their occupation (grade 3 or 4) and during leisure time (grade 3 or 4). There was a tendency towards leisure-time activity being more important than occupational activity. A similar tendency for the importance of leisure-time activity was also shown when the physical working capacity was measured at heart rate 150 (PWC150), (table I) or as calculated maximum oxygen uptake (table II). In this analysis, activity grades 1 and 2 were combined because small differences were found for PWC150, etc. Grades 3 and 4 were combined because of very few subjects in grade 4. During an 8-year follow-up, 30 men suffered non-fatal myocardial infarction and 19 died from CHD. The relationships between physical activity in the subjects' occupation and during leisure time appear in figures 1 and 2. There was no relationship whatsoever with occupational activity, but the leisure-time activity the year before interview was related to later CHD (p < 0.05). There was no difference between non-fatal and fatal infarcts. The estimated leisure-time activity between ages 20 and 29 was not, however, related to CHD during middle-age. When analyzing the relationship between physical activity and CHD, it is important also to know the relationships between other risk factors and CHD. A multivariate logistic analysis in this sample has previously been reported [21] and has been repeated for an 11.5-year follow-up (from 1963). The three risk factors, tobacco smoking, hypercholesterolemia, and hyper-
Downloaded by: Université René Descartes Paris 5 193.51.85.197 - 1/28/2018 10:34:37 AM
III. Results
220
WILHELMSENjTIBBLINjAURELLjBJUREjEKSTROM-JODALjGRIMBY
Table I. Physical working capacity at heart rate 150 (PWC150) in relation to physical activity
Activity during leisure time
Grade
Activity at work
1-2
3-4
total
1-2
Mean SD N
824 185 294
871 208 116
837 192 410
3-4
Mean SD N
846 209 156
909 179 53
862 203 209
Total
Mean SD N
832 194 450
882 199 169
845 196 619
Difference
)
t = 1.51 n.s.
~
Difference
= 2.88;
t
P < 0.005
Table II. Calculated maximum oxygen uptake (mI/kg x min) in relation to physical activity
Activity during leisure time
Grade
Activity at work
1-2
3-4
total
1-2
Mean SD N
29.8 3.5 303
31.1 3.4 118
30.2 3.5 421
3-4
Mean SD N
30.9 3.6 162
31.3 3.8 58
31.0 3.7 220
Total
Mean SD N
30.2 3.6 465
31.2 3.5 176
30.5 3.6 641
Difference
t = 2.90 P < 0.005
t
=
3.25; p < 0.005
Downloaded by: Université René Descartes Paris 5 193.51.85.197 - 1/28/2018 10:34:37 AM
~
Difference
Risk of Myocardial Infarction MI,Ofo 10 Activity, age 20-29
1 8
2
3
AcbvltY,age 54
4
1
2
4 r-
3
r-
r-
6
221
-
-
-
r-
4
r-
4
o
NI number 7 Total number 95
12 237
17 281
10 165
13
164
18 357
8 183
6
73
Fig. 1. Incidence of myocardial infarction (MI, fatal and non-fatal) in relation to physical activity during work. Men born in 1913 followed from age 54 to 62 years.
MI J % 10 Activity, age 20-29
1 8
2
-
3
Acbvlty, age 54
4
1
4
0-
-
-
2
o
3
r-
6 4
2
r-
NI number 1 Total number 25
24 291
16 326
5 133
11
138
30 439
Jl
5 188
0 2
tension were found to be strong risk factors in this study as well as in several others. In addition to these, a genetic factor, 'early parental death', and dyspnea during exertion according to a questionnaire were significant risk factors [18]. The factor 'early parental death' was based upon a scoring system of the age at which the parents died. The cause of death was not taken into account because the data concerning parents could not be verified. An analysis of confounding variables in relation to physical activity showed that there was a relationship between smoking and physical activity. Thus, those who were physically active in their occupation were more often smokers, whereas those who were physically active during leisure time were more often non-smokers (table III). The same relationships were found with
Downloaded by: Université René Descartes Paris 5 193.51.85.197 - 1/28/2018 10:34:37 AM
Fig. 2. Incidence of myocardial infarction (MI, fatal and non-fatal) in relation to physical activity during leisure time. Men born in 1913 followed from age 54 to 62 years.
WILHELMSEN/TIBBLIN/AURELL/BJURE/EKSTROM-JODAL/GRIMBY
222
Table III. Percent smokers in relation to physical activity Grade
Activity at work
1-2 3-4 Total
Activity during leisure time, % 1-2
3-4
total
56 60 57
43 46 44
53 57 54
X2
= 10.9; p < 0.001
}
Difference
X2 = 1.07 n. s.
Table IV. Percent registered for intemperance by the Temperance Board in relation to physical activity Grade
Activity at work
1-2 3-4 Total
Activity during leisure time, % 1-2
3-4
19 29 22
20 12
8
Difference
total 16 27 20
}
X2 = 13.2 p
Physical Activity, Physical Fitness and Risk of Myocardial Infarction 1 L. WILHELMSEN, G. TIBBLIN, M. AURELL, J. BlURE, B. EKSTROM-JODAL and G. GRIMBY Medical Department I and Department of Clinical Physiology I, Sahlgren's Hospital, Gothenburg
l. Introduction
Supported by grants from the Swedish National Organization against Heart and Chest Diseases, and the Bank of Sweden Tercentenary Fund.
Downloaded by: Université René Descartes Paris 5 193.51.85.197 - 1/28/2018 10:34:37 AM
An essential prerequisite for the prevention of coronary heart disease (CHD) is demonstration of relationships between personal characteristics and risk of manifest CHD. Some of the risk factors detected in this way may only be indicators of an early, subclinical disease, while other risk factors may have causative significance and be amenable to treatment which may prevent or postpone the disease. An example of the first type of risk factor is ECG changes, and examples of the second type are tobacco smoking, hypertension and hyperlipidemia. MORRIS et al. [12] showed that the risk of CHD was lower in London bus conductors than in bus drivers. Later, a long series of studies by Fox and HASKELL [6], KANNEL and McNAMARA [10], KANNEL et al. [11] and CASSEL et al. [3], as well as others, showed that high physical activity was associated with lower risk for future CHD than low physical activity. The earlier studies mainly concerned occupational activity. Later studies have also investigated the importance of leisure-time activity. An important recent study into this question was performed by MORRIS et al. [13], who studied 17,000 civil servants and related their leisure-time exercise habits to subsequent CHD after adjustment for age and smoking habits. These authors did not find any CHD differences until the highest classification (frequent peak activities to a level of 7.5 kcal/min) was reached. In the studies of both KANNEL et al. [10, 11] and MORRIS et al. [12, 13], leisure-time activity was
WILHELMsENjTmBLINjAURELLjBJUREjEKSTROM-JODALjGRlMBY
218
of significant importance even in multivariate analysis when other confounding factors were taken into consideration. There is a relationship between physical activity, especially during leisure time, and physical fitness, but several constitutional factors seem to be of great importance in the degree of physical fitness. No studies have been published which have related physical fitness to subsequent CHD. The prognostic significance of ECG changes during rest or after exercise has, however, been reported [1, 2, 4, 14, 15]. In the present paper, studies concerning physical inactivity as a risk factor in middle-aged men from the general population in G6teborg, Sweden, are reported. In addition, the prognostic signification of an exercise test in a general population sample of men has been analyzed. The relatively large number of persons studied within a very narrow age span facilitated the analysis.
The study involving men born in 1913. The study population was recruited from men living in Gothenburg, Sweden, born in 1913 and still alive at the age of 50 years in 1963. All men meeting these criteria who were born on dates divisible by three, i. e. the 3rd, 6th, 9th (etc.) of each month, comprised the study sample. These criteria were met by 973 men. Of them, 855 (88 %) agreed to be examined in 1963, and 803 were reexamined in 1967. Analyses of the participants and the non-participants have been published previously [17, 18, 21]. In 1967, a comprehensive interview concerning physical activity at work and during leisure time in lO-year intervals from age of 20 to the age of 54 was performed by one interviewer (L. W.). The questionnaire has been described previously [16]. It was later modified for use as a postal questionnaire [19]. The exercise test was performed on an electrically braked bicycle ergometer. The subjects cycled for 6 min at each load and the load was increased stepwise, starting with 49 W (300 kpmjmin) and increasing to 98 W (600 kpmjmin). According to the protocol some subjects were tested at 147 W (900 kpmjmin) and finally at a maximal load. The procedure has been described in detail previously [8, 9]. The work performed in kilopondmeters per minute at a heart rate of 150 beats per minute (W150) was estimated by interpolation. Maximum work load was defined as the subjects' apprehension of exhaustion or a plateau of heart rate or both when increasing the work load in small steps. This level was reached by 84 % of subjects. Submaximal and maximal oxygen uptake and ventilation were determined in a 10 % subsample and a formula was constructed for calculation of maximal oxygen uptake in the total sample [7]. Heart volume determined by X-ray examination was not better correlated with maximal oxygen uptake than any combination of anthropometric variables. The primary preventive trial in Gothenburg. One third of all men born between
Downloaded by: Université René Descartes Paris 5 193.51.85.197 - 1/28/2018 10:34:37 AM
II. Methods and Population Samples
Risk of Myocardial Infarction
219
1915-1922 and 1924-1925 in Gothenburg are included in the intervention group of a primary preventive trial which began in 1970. The postal questionnaire was answered by 83 % of the 10,000 men, and 75 % agreed to examination. The study design and an analysis of the non-participants has been published [20, 22]. The follow-up of end-point cases in both investigations has been done by the collection of death certificates in the city (autopsy rate 90 %) and since 1968 by a special Myocardial Infarction Register [5]. Standard statistical methods were used for x2-analysis, Student's t-test, and regression analysis.
An important point concerning our questionnaire on physical activity is whether it can detect true differences in physical activity. With the aid of the exercise test, we studied how the different types of activity were related to physical working capacity. The percentage of men who could perform the maximal exercise test increased from 63 % in the group who had the lowest physical activity both in their occupation (grade 1) and during leisure time (grade 1) up to 95% among those who had high activity both in their occupation (grade 3 or 4) and during leisure time (grade 3 or 4). There was a tendency towards leisure-time activity being more important than occupational activity. A similar tendency for the importance of leisure-time activity was also shown when the physical working capacity was measured at heart rate 150 (PWC150), (table I) or as calculated maximum oxygen uptake (table II). In this analysis, activity grades 1 and 2 were combined because small differences were found for PWC150, etc. Grades 3 and 4 were combined because of very few subjects in grade 4. During an 8-year follow-up, 30 men suffered non-fatal myocardial infarction and 19 died from CHD. The relationships between physical activity in the subjects' occupation and during leisure time appear in figures 1 and 2. There was no relationship whatsoever with occupational activity, but the leisure-time activity the year before interview was related to later CHD (p < 0.05). There was no difference between non-fatal and fatal infarcts. The estimated leisure-time activity between ages 20 and 29 was not, however, related to CHD during middle-age. When analyzing the relationship between physical activity and CHD, it is important also to know the relationships between other risk factors and CHD. A multivariate logistic analysis in this sample has previously been reported [21] and has been repeated for an 11.5-year follow-up (from 1963). The three risk factors, tobacco smoking, hypercholesterolemia, and hyper-
Downloaded by: Université René Descartes Paris 5 193.51.85.197 - 1/28/2018 10:34:37 AM
III. Results
220
WILHELMSENjTIBBLINjAURELLjBJUREjEKSTROM-JODALjGRIMBY
Table I. Physical working capacity at heart rate 150 (PWC150) in relation to physical activity
Activity during leisure time
Grade
Activity at work
1-2
3-4
total
1-2
Mean SD N
824 185 294
871 208 116
837 192 410
3-4
Mean SD N
846 209 156
909 179 53
862 203 209
Total
Mean SD N
832 194 450
882 199 169
845 196 619
Difference
)
t = 1.51 n.s.
~
Difference
= 2.88;
t
P < 0.005
Table II. Calculated maximum oxygen uptake (mI/kg x min) in relation to physical activity
Activity during leisure time
Grade
Activity at work
1-2
3-4
total
1-2
Mean SD N
29.8 3.5 303
31.1 3.4 118
30.2 3.5 421
3-4
Mean SD N
30.9 3.6 162
31.3 3.8 58
31.0 3.7 220
Total
Mean SD N
30.2 3.6 465
31.2 3.5 176
30.5 3.6 641
Difference
t = 2.90 P < 0.005
t
=
3.25; p < 0.005
Downloaded by: Université René Descartes Paris 5 193.51.85.197 - 1/28/2018 10:34:37 AM
~
Difference
Risk of Myocardial Infarction MI,Ofo 10 Activity, age 20-29
1 8
2
3
AcbvltY,age 54
4
1
2
4 r-
3
r-
r-
6
221
-
-
-
r-
4
r-
4
o
NI number 7 Total number 95
12 237
17 281
10 165
13
164
18 357
8 183
6
73
Fig. 1. Incidence of myocardial infarction (MI, fatal and non-fatal) in relation to physical activity during work. Men born in 1913 followed from age 54 to 62 years.
MI J % 10 Activity, age 20-29
1 8
2
-
3
Acbvlty, age 54
4
1
4
0-
-
-
2
o
3
r-
6 4
2
r-
NI number 1 Total number 25
24 291
16 326
5 133
11
138
30 439
Jl
5 188
0 2
tension were found to be strong risk factors in this study as well as in several others. In addition to these, a genetic factor, 'early parental death', and dyspnea during exertion according to a questionnaire were significant risk factors [18]. The factor 'early parental death' was based upon a scoring system of the age at which the parents died. The cause of death was not taken into account because the data concerning parents could not be verified. An analysis of confounding variables in relation to physical activity showed that there was a relationship between smoking and physical activity. Thus, those who were physically active in their occupation were more often smokers, whereas those who were physically active during leisure time were more often non-smokers (table III). The same relationships were found with
Downloaded by: Université René Descartes Paris 5 193.51.85.197 - 1/28/2018 10:34:37 AM
Fig. 2. Incidence of myocardial infarction (MI, fatal and non-fatal) in relation to physical activity during leisure time. Men born in 1913 followed from age 54 to 62 years.
WILHELMSEN/TIBBLIN/AURELL/BJURE/EKSTROM-JODAL/GRIMBY
222
Table III. Percent smokers in relation to physical activity Grade
Activity at work
1-2 3-4 Total
Activity during leisure time, % 1-2
3-4
total
56 60 57
43 46 44
53 57 54
X2
= 10.9; p < 0.001
}
Difference
X2 = 1.07 n. s.
Table IV. Percent registered for intemperance by the Temperance Board in relation to physical activity Grade
Activity at work
1-2 3-4 Total
Activity during leisure time, % 1-2
3-4
19 29 22
20 12
8
Difference
total 16 27 20
}
X2 = 13.2 p
