International Journal of Obesity (2000) 24, 1465±1474
ß 2000 Macmillan Publishers Ltd All rights reserved 0307±0565/00 $15.00
www.nature.com/ijo
Body mass index, physical inactivity and low
level of physical ®tness as determinants of allcause and cardiovascular disease mortality Ð
16 y follow-up of middle-aged and elderly men
and women
N Haapanen-Niemi1,2*, S Miilunpalo1, M Pasanen1, I Vuori1, P Oja1 and J Malmberg1,2
1
UKK Institute for Health Promotion Research, Tampere, Finland; and 2University of Tampere, Tampere School of Public Health,
Tampere, Finland
OBJECTIVE: To investigate the independent associations and the possible interaction of body mass index (BMI),
leisure time physical activity (LTPA) and perceived physical ®tness and functional capability with the risk of mortality.
DESIGN: Prospective 16 y follow-up study.
SUBJECTS: A regionally representative cohort of 35 ± 63 -y-old Finnish men (n ˆ 1090) and women (n ˆ 1122).
MEASUREMENTS: All-cause, cardiovascular disease (CVD) and coronary heart disease (CHD) mortality were derived
from the national census data until the end of September 1996 while the initial levels of BMI, LTPA, physical ®tness
and function were determined from self-administered questionnaires.
RESULTS: After adjustment for age, marital and employment status, perceived health status, smoking and alcohol
consumption, the Cox proportional hazards model showed that BMI was not associated with the risk of death among
the men or the women. Compared with the most active subjects the men and women with no weekly vigorous
activity had relative risks of 1.61 (95% con®dence interval, CI, 0.98 ± 2.64) and 4.68 (95% CI, 1.41 ± 15.57), respectively,
for CVD mortality, and for the men there was a relative risk of 1.66 (95% CI, 0.92 ± 2.99) for CHD mortality. When
compared with the men who perceived their ®tness as better than their age-mates, the men with the `worse'
assessment had a relative risk of 3.29 (95% CI, 1.80 ± 6.02) for all-cause mortality and 4.37 (95% CI, 1.80 ± 10.6) for CVD
mortality. Men with at least some dif®culty in walking a distance of 2 km had a relative risk of 1.62 (95% CI, 1.05 ± 2.50)
for all-cause mortality when compared with those who had no functional dif®culties. In addition, in the comparison
with subjects with no functional dif®culties, the men and women who had some dif®culty climbing several ¯ights of
stairs had relative risks of 1.47 (95% CI, 0.97 ± 2.23) and 2.39 (95% CI, 1.25 ± 4.60) for all-cause mortality, respectively.
For CVD mortality the relative risks were 1.85 (95% CI, 1.04 ± 3.30) and 3.38 (1.22 ± 9.41), respectively.
CONCLUSIONS: Although BMI did not prove to be an independent risk factor for mortality from CVD, CHD or from all
causes combined, perceived physical ®tness and functional capability did. An increase in LTPA seems to have a similar
bene®cial effect on the mortality risk of obese and nonobese men and women, and the effect also seems to be similar
for ®t and un®t subjects.
International Journal of Obesity (2000) 24, 1465±1474
Keywords: body mass index; cardiovascular diseases; ®tness; mortality; obesity; physical activity
Introduction
Obesity is a well-known risk factor for several chronic
disorders, and it is possibly associated with an
increased risk of premature mortality.1 ± 11 The summary of evidence suggests that the relation between
body weight or body mass index (BMI) and all-cause
mortality is usually J- or U-shaped. In other words the
*Correspondence: N Haapanen-Niemi, UKK Institute for Health
Promotion Research, P O Box 30, FIN-33501 Tampere, Finland.
E-mail: nina.haapanen@uta.®
Received 26 November 1999; revised 8 May 2000; accepted
12 June 2000
optimal body weight may be higher than that de®ned
as a desirable average, and leanness may be associated
with an increased risk of premature death.8,10 ± 14 The
results of previous studies concerning disease-speci®c
mortality and moderate body weights are con¯icting
and often vary between men and women. In addition
to the U- and J-shaped associations, ®ndings on
the relationship between body weight or BMI and
mortality have included no, direct or even inverse
associations.10,14
A convincing body of epidemiological evidence
shows that physical inactivity and a low level of
physical ®tness are risk factors for premature allcause and cardiovascular disease (CVD) mortality.15 ± 22 Besides the effects on mortality, low levels
BMI, physical inactivity, low physical ®tness and mortality
N Haapanen-Niemi et al
1466
of physical activity and ®tness have been shown to be
associated with an unfavourable level of body
weight.23 ± 27
In several recent studies the health consequences of
physical activity, ®tness and obesity have been studied
in separate groups of subjects. Little is known, however about the relative importance of these factors and
their interrelationships to the risk of premature allcause and CVD mortality, especially among women.
Some results suggest that an active way of life may
have bene®cial effects on mortality even for those
who remain overweight.28 In recent studies in the
United States it was concluded that moderate and
high levels of physical ®tness were associated with a
lower risk for all-cause mortality among both overweight and normal-weight men.28,29
The main objective of this epidemiological study
was to determine both the independent associations
and the possible interaction of BMI, leisure-time
physical activity (LTPA) and perceived physical
®tness and functional capability with the risk of
all-cause, CVD and coronary heart disease (CHD)
mortality in middle-aged and elderly men and
women during a follow-up of 16 y. It was hypothesized that a moderate or high level of LTPA and
good perceived physical ®tness and functional capability have independent protective effects against
all-cause, CVD and CHD mortality both among the
men and among the women. According to our
earlier ®ndings30 it was also assumed that, when
compared with low level of LTPA, a perceived low
level of physical ®tness and functional capability
are stronger predictors of premature mortality,
while, compared with low level of LTPA and a
perceived low level of ®tness and functional capability, obesity is an independent but weaker risk
factor for all-cause, CVD and CHD mortality.
Methods
Subjects and design
A systematic representative sample (n ˆ 6787) of
residents between the ages of 19 and 63 y was
drawn from the census data of a medium-size industrial town and two rural municipalities in northeastern
Finland. The sample represented 20% of the workingage population of the target area in 1980.
A self-administered questionnaire including structured questions concerning living habits, health behaviour, health status, functional capacity and
sociodemographic background factors was sent to
the subjects at the beginning of 1980. The ®nal
Figure 1 Design of the prospective follow-up study carried out in northeastern Finland, 1980 ± 1996.
International Journal of Obesity
BMI, physical inactivity, low physical ®tness and mortality
N Haapanen-Niemi et al
response rate was 77.5% (n ˆ 5259). Follow-up questionnaires requesting similar information were sent to
the respondents 1, 5, 10 and 16 y later. The response
rates for these surveys were 88, 84, 85 and 85%,
respectively. The study design of the 16 y prospective
follow-up study carried out in northeastern Finland
during 1980 ± 1996 is shown in Figure 1.
Because of the very low mortality rates of the
persons initially less than 35 y, the analyses were
targeted at subjects aged 35 y and older (men,
n ˆ 1340; women, n ˆ 1500), who were 51 ± 79 y of
age at the end of the follow-up. So that any confounding effect of a subclinical disease would be as low as
possible,31 those having an initial BMI of less than
20.0 and those suffering from a disease or symptoms
that totally prevented participation in LTPA were
excluded from the analysis. After all the exclusion
criteria 1090 men and 1122 women were included in
the analysis. The frequencies of the baseline characteristics (age, socioeconomic status, employment
status, marital status, perceived health, alcohol consumption and smoking) of the surviving and deceased
men and women are given in Table 1.
1467
Assessment of body mass index
The initial self-administered questionnaire included
an open question concerning the subject's body
weight and height. The validity of self-reports of
body weight was evaluated for a subsample of 490
men and 615 women aged initially 40 ± 63 y; the body
weight distributions obtained from the questionnaire
data and measured data in 1996 were used for this
purpose. The reported body weight values were highly
correlated with measured body weight both for the
men (Spearman r ˆ 0.96) and for the women (Spearman r ˆ 0.98). The men underestimated their weight
by 0.8 kg and the women by 1.5 kg, however. The
total error, including both systematic and random
error, for self-reported body weight was 3.1 kg for
the men and 2.7 kg for the women. For the women the
total error for self-reported body weight increased
slightly with age. For the analysis the BMI measure,
de®ned as weight in kilogrammes divided by the
square of height in metres, was used. In accordance
with recent statements,10,14 the BMI categories for the
®nal analysis were de®ned as follows: < 25 (normal
weight), 25.0 ± 29.9 (overweight), and 30 (obese).
Table 1 Baseline characteristics of men and women
Men
Stratum
Age (y)
35 ± 44
45 ± 54
55 ± 63
Socioeconomic status
Upper-level employee
Lower-level employee
Manual worker
Farmer or other own-account worker
Other
Housewife
Missing data
Employment status
Participant in work life
Not participant in work life
Housewife
Missing data
Marital status
Married
Single
Widowed
Divorced, separated
Missing data
Perceived health
Good or fairly good
Average
Poor or rather poor
Missing data
Smoking status
Never smoked
Past smoker
Current smoker
Missing data
Alcohol consumption (g=day)
0
0.1 ± 12
> 12
Missing data
Women
Alive (n ˆ 882) %
Deceased (n ˆ 208) %
Alive (n ˆ 1035) %
Deceased (n ˆ 87) %
48.1
37.0
15.0
20.7
35.6
43.8
44.2
35.7
20.2
12.6
32.2
55.2
19.7
13.2
54.2
11.6
0.8
Ð
0.6
12.5
8.2
60.6
16.3
1.9
Ð
0.5
9.3
34.9
30.7
15.7
1.0
8.3
0.2
5.7
27.6
34.5
11.5
5.7
13.8
1.1
83.3
15.6
Ð
1.0
50.5
47.1
Ð
2.4
61.8
18.9
18.6
0.6
34.5
44.8
18.4
2.3
82.2
12.8
1.1
3.6
0.2
80.8
14.4
Ð
4.8
Ð
78.1
8.1
8.5
5.0
0.3
63.2
13.8
17.2
5.7
Ð
44.0
38.9
16.9
0.2
24.0
34.6
41.3
Ð
43.5
37.4
19.0
0.1
23.0
37.9
39.1
Ð
27.9
37.2
32.5
2.4
13.5
33.2
47.1
6.3
75.9
9.1
13.6
1.4
63.2
10.3
23.0
3.4
16.1
54.1
29.4
0.5
15.9
47.1
35.1
1.9
54.9
41.5
2.4
1.2
69.0
24.1
5.7
1.1
International Journal of Obesity
BMI, physical inactivity, low physical ®tness and mortality
N Haapanen-Niemi et al
1468
Assessment of leisure-time physical activity and
physical ®tness and function
LTPA was assessed by two measures indicating the
activity level during the year preceding the baseline
questioning: an index for total energy expenditure in
LTPA and a single-item self-assessment of physical
activity. For an index LTPA was assessed primarily
from the data obtained with 23 individual questions
concerning conditioning exercise, sports, physical
recreation, different leisure time and household
chores and active commuting to and from work. The
physical activity energy expenditure (kcal=week) was
calculated by multiplying the weekly frequency and
the average duration of each type of physical activity
reported and two coef®cients, one estimating the rate
of the energy cost32 and the other the seasonal duration of each activity. The construction of the index has
been described in more detail elsewhere.21
The subjects were divided into high, moderate and
low physical activity groups according to the index of
total physical activity. For the men, the classes were
designated as 0 ± 1000, 1000.1 ± 1900 and > 1900 kcal
per week, and for the women the respective categories
were 0 ± 800, 800.1 ± 1500 and > 1500 kcal per week.
A single-item self-assessment of global physical
activity during the previous 12 months covered the
intensity, frequency and duration of weekly exercise
sessions as follows. `Which of the following categories
best describes your physical activity during the past 12
months? Consider all types of LTPA, including walking and cycling to and from work, if the trip takes at
least 15 min one way'. The response alternatives were:
(1) `vigorous activity twice or more a week'; (2)
`vigorous activity once a week and some light intensity
activity'; (3) `some activity each week'; and (4) `no
regular weekly activity'. In the analysis, the ®rst two
classes were considered as active while the last two
categories were considered as inactive. In our subsample of 1996 data that consisted of the men and
women initially aged 40 ± 63 y, the validity of the selfreported, single-item, global physical activity had been
found to be good. For example the average measured
1 km walking time was found to decrease as the level
of self-reported LTPA increased both among the men
and the women. The men who were vigorously active
at least twice a week had a 20% shorter 1 km walking
time than the physically inactive men. Among the
women the difference was 15%.
Physical ®tness was assessed by three measures
indicating perceived ®tness and functional status.
The rating of perceived physical ®tness compared
with that of age-mates was classi®ed with the
response alternatives `better', `similar' and `worse'.
This ®tness measurement was associated with the risk
of death in our earlier analysis with a shorter followup period as well.30 Functional status was assessed by
a self-reported estimate on ability to walk 2 km and
climb several ¯ights of stairs without rest. The
response alternatives were `no dif®culties', `some
dif®culties', `major dif®culties' and `not at all able
International Journal of Obesity
to walk 2 km or climb several ¯ights of stairs'. In the
®nal analysis the categories were combined as follows: `no dif®culties' and `at least some dif®culties'.
In a subsample of elderly subjects in 1996 the measured average time for climbing several ¯ights of
stairs decreased as the level of self-reported dif®culty
in climbing stairs decreased. The men reporting no
dif®culty in climbing several ¯ights of stairs had an
18% lower measured stair climbing time than the men
with at least some dif®culties, while the difference
among the women was 24%. Similarly, the average
1 km walking time decreased as the level of dif®culty
in the self-reported ability to walk 2 km decreased
among both the men and the women. The men with no
dif®culties had a 15% shorter 1 km walking time than
those with at least some dif®culties. Among the
women the difference was 13%.
Assessment of death events
The all-cause, CVD and CHD mortality of the respondents was monitored in national census data from the
Statistics Finland until the end of September 1996.
This monitoring represented a follow-up period of 16 y
and 7 months. During the follow-up period, 208 men
and 87 women aged 35 y and older died. Altogether
54% of the deaths among the men and 52% of the
deaths among the women were due to CVD (ICD-9:
390 ± 458). The proportion of CHD (ICD: 410 ± 414)
deaths among the CVD deaths of the men was 73%.
Due to the very small number of CHD deaths among
the women the analysis was restricted to only men.
Statistical analyses
The number of subjects and deaths, the number of
person-y and the age-adjusted mortality rates per 1000
person-y of follow-up were computed for the categories of BMI, LTPA and perceived physical ®tness
and functional capability. The mortality rates were
adjusted by the direct method using the total population as the standard. The following age groups were
designated: 35 ± 44, 45 ± 54 and 55 ± 63 y.
A multivariate analysis for estimating the relative
risk of death with respect to person-y was computed
using the Cox proportional hazards model.33 The
models for mortality included adjustment for initially
assessed age as a continuous variable and marital
status, employment status, perceived health status,
smoking and alcohol consumption as classi®ed variables. In the models all the two-way interactions
between LTPA, perceived physical ®tness and BMI
were also tested in relation to mortality risk.
Results
Of those men having highest total energy expenditure
for LTPA, 27% were categorized as having BMI value
BMI, physical inactivity, low physical ®tness and mortality
N Haapanen-Niemi et al
equal to or higher than 27 (results not shown). Among
the least active men the proportion of overweight
(BMI 27) subjects was 38%. The percentage proportions among the most and least active women were
29% and 39%, respectively.
Among the men who perceived that their physical
®tness level was better than that of their age-mates,
47% were classi®ed into the highest third of total
energy expenditure for LTPA while 19% of those men
perceiving their physical ®tness worse than that of
their age-mates were classi®ed into this group (results
not shown). Among the women the percentages were
44% and 17%, respectively.
Tables 2 and 3 present the age-adjusted mortality
rates for BMI, LTPA and perceived ®tness and function. The overweight subjects with a BMI of 25.0 ±
29.9 had a lower level of age-adjusted mortality than
the normal-weight subjects, while obese men and
women (BMI 30) had increased mortality rates for
all causes, CVD and CHD when compared with the
normal-weight subjects. The mortality rates for allcauses, CVD and CHD were highest for the men with
a low total energy expenditure for LTPA and for those
with no weekly vigorous physical activity. In addition
the mortality rates were highest for those who perceived that their physical ®tness level was worse than
that of their age-mates. Among the men the ageadjusted mortality rates for all-causes, CVD and
CHD were highest for those with at least some
dif®culty climbing several ¯ights of stairs or some
dif®culty walking 2 km. Among the women the mortality rates were increased only among those reporting
at least major dif®culty with these functions.
The relative risks for all-cause, CVD and CHD
mortality are described in Tables 4 and 5. During the
16 y follow-up period the BMI was not associated
with the risk of all-cause or cause-speci®c mortality
either among the men or the women after adjustment
for age, marital and employment status, perceived
health status, smoking and alcohol consumption. As
a consequence, no interaction effect between the
LTPA=®tness measurement and BMI was found for
the relative risk of death.
The men with a low level of energy expenditure for
LTPA had a suggestively increased risk for CVD and
CHD mortality when they were compared with the
most active men after adjustment for age, marital
status, employment status, perceived health status,
smoking and alcohol consumption. When compared
with those who were the most active, the men with a
low energy expenditure for LTPA had relative risks of
1.69 (95% con®dence interval, CI, 0.97 ± 2.93) for
CVD mortality and 1.70 (95% CI, 0.90 ± 3.21) for
CHD mortality. Among the women the inactive subjects had increased relative risks for all-cause and
CVD mortality when compared with the women with
1469
Table 2 Age-adjusted all-cause, CVD and CHD mortality rates per 1000 person-years for the men in 1980 ± 1996
CVD a
All causes
Stratum
PersonAge-adjusted
Age-adjusted
Age-adjusted
No. of
years of No. of
rate=1000
No. of
rate=1000
No. of
rate=1000
Subjects follow-up deaths person-years deaths person-years deaths person-years
BMIa
20.0 ± 24.9
440
25.0 ± 26.9
293
27.0 ± 29.9
249
30.0
108
a
Total LTPA energy expenditure index (kcal=week)
High
347
Moderate
369
Low
322
Single-item self-assessment of LTPA
Vigorous activity twice or more a week
213
Vigorous activity once a week and
224
some light activity
Light intensity activity weekly
493
No regular weekly activity
116
Perceived physical ®tness compared with age-mates
Better
197
Similar
635
Worse
236
2 km walking ability
No dif®culties
691
Some dif®culties
84
Substantial dif®culties or not at all
40
able to walk
Climbing several ¯ights of stairs
No dif®culties
623
Some dif®culties
180
Substantial dif®culties or not at all
54
able to climb
a
CHD a
6609.2
4465.5
3838.4
1528.8
92
46
43
27
14.0
10.6
10.6
16.1
45
27
21
20
6.8
6.3
5.1
11.8
34
21
15
13
5.2
4.9
3.6
7.7
5430.6
5728.4
4610.8
50
57
86
10.2
10.3
16.0
23
28
54
4.9
5.0
10.1
16
21
41
3.3
3.8
7.6
3322.8
3564.5
29
27
9.9
8.3
14
12
5.0
3.7
11
7
4.1
2.0
7281.1
1732.4
113
23
14.3
12.9
65
12
8.2
6.7
46
11
5.9
6.1
3164.1
9884.3
3169.7
19
99
82
5.2
10.9
23.6
8
51
49
2.2
5.7
14.0
5
38
37
1.4
4.3
10.6
10153.7
1058.7
456.9
97
31
16
10.2
29.2
33.7
51
15
11
5.5
12.1
20.6
39
10
7
4.2
7.6
6.5
9258.2
2354.2
694.4
73
60
18
8.6
21.3
23.2
32
35
15
3.9
11.5
21.3
24
26
11
2.9
8.6
12.8
CVD indicates cardiovascular disease; CHD, coronary heart disease; BMI, body mass index, LTPA, leisure-time physical activity.
International Journal of Obesity
BMI, physical inactivity, low physical ®tness and mortality
N Haapanen-Niemi et al
1470
Table 3 Age-adjusted all-cause and CVD mortality rates per 1000 person-years for the women, 1980 ± 1996
CVD a
All causes
Stratum
No. of
Subjects
BMIa
20.0 ± 24.9
523
25.0 ± 26.9
226
27.0 ± 29.9
211
30.0
151
a
Total LTPA energy expenditure index (kcal=week)
High
328
Moderate
392
Low
310
Single-item self-assessment of LTPA
Vigorous activity twice or more a week
214
Vigorous activity once a week and
203
some light activity
Light intensity activity weekly
561
No regular weekly activity
99
Perceived physical ®tness compared with age-mates
Better
201
Similar
692
Worse
208
2 km walking ability
No dif®culties
723
Some dif®culties
105
Substantial dif®culties or not
42
at all able to walk
Climbing several ¯ights of stairs
No dif®culties
576
Some dif®culties
242
Substantial dif®culties or not at
94
all able to climb
Personyears of
follow-up
No. of
deaths
Age-adjusted
rate=1000
person-years
No. of
deaths
Age-adjusted
rate=1000
person-years
8434.9
3628.8
3376.3
2368.1
35
15
15
22
5.1
3.9
3.9
7.9
17
8
9
11
2.7
2.1
2.2
3.7
5328.4
6374.9
4851.7
17
17
40
4.3
2.7
6.9
9
7
21
2.5
1.1
3.3
3497.1
3328.0
8
7
2.4
2.3
2
1
0.6
0.7
8886.1
1585.9
56
9
5.7
5.4
34
6
3.4
3.4
3249.7
11191.2
3230.9
15
36
31
4.0
3.7
8.9
8
19
15
2.0
2.0
4.0
10980.0
1572.7
548.2
40
9
13
4.0
4.7
21.1
21
3
7
2.2
1.3
8.7
8835.3
3643.5
1270.7
19
18
24
2.7
4.1
18.5
7
9
13
1.1
1.9
8.6
a
CVD indicates cardiovascular disease; BMI, body mass index; LTPA, leisure-time physical activity.
Table 4 Relative risks for all-cause-, CVD and CHD mortality for the men in 1980 ± 1996
CVD a
All causes
Stratum
RR a,b
95% CI a
P
RR b
95% CI
CHD a
P
RR b
95% CI
P
a
BMI
20.0 ± 24.9
25.0 ± 29.9
30.0
1.00
0.87
0.64 ± 1.19
0.379
1.06
0.67 ± 1.69
0.796
LRa ˆ 1.1, df ˆ 2, P ˆ 0.577
a
Total LTPA energy expenditure index (kcal=week)
High
1.00
Moderate
0.82
0.55 ± 1.24
0.350
Low
1.20
0.82 ± 1.76
0.349
LR ˆ 4.2, df ˆ 2, P ˆ 0.120
Single-item self-assessment of LTPA
Vigorous activity at least once a week
1.00
and some light activity
No or light intensity activity weekly
1.26
0.89 ± 1.77
0.193
LR ˆ 1.7, df ˆ 1, P ˆ 0.187
Perceived physical ®tness compared with age-mates
Better
1.00
Similar
1.93
1.15 ± 3.24
0.013
Worse
3.29
1.80 ± 6.02 < 0.001
LR ˆ 16.9, df ˆ 2, P < 0.001
2 km walking ability
No dif®culties
1.00
At least some dif®culties
1.62
1.05 ± 2.50
0.028
LR ˆ 4.7, df ˆ 1, P ˆ 0.030
Climbing several ¯ights of stairs
No dif®culties
1.00
At least some dif®culties
1.47
0.97 ± 2.23
0.070
LR ˆ 3.3, df ˆ 1, P ˆ 0.070
a
1.00
0.87 0.56 ± 1.35
0.541
1.44 0.80 ± 2.56
0.223
LR ˆ 2.6, df ˆ 2, P ˆ 0.274
1.00
0.94 0.57 ± 1.56
0.813
1.23 0.61 ± 2.50
0.564
LR ˆ 0.5, df ˆ 2, P ˆ 0.767
1.00
0.94 0.52 ± 1.71
0.846
1.69 0.97 ± 2.93
0.063
LR ˆ 6.5, df ˆ 2, P ˆ 0.039
1.00
0.88 0.44 ± 1.76
0.709
1.70 0.90 ± 3.21
0.105
LR ˆ 5.8, df ˆ 2, P ˆ 0.056
1.00
1.00
1.61 0.98 ± 2.64
0.058
LR ˆ 3.8, df ˆ 1, P ˆ 0.050
1.66 0.92 ± 2.99
0.090
LR ˆ 3.1, df ˆ 1, P ˆ 0.079
1.00
2.39 1.09 ± 5.22
0.029
4.37 1.80 ± 10.62 0.001
LR ˆ 12.5, df ˆ 2, P ˆ 0.002
1.00
2.82 1.06 ± 7.46
0.037
4.64 1.56 ± 13.84 0.006
LR ˆ 9.1, df ˆ 2, P ˆ 0.011
1.00
1.00
1.25 0.71 ± 2.22
0.438 1.03 0.51 ± 2.05
0.941
LR ˆ 0.6, df ˆ 1, P ˆ 0.441 LR ˆ 0.005, df ˆ 1, P ˆ 0.941
1.00
1.85 1.04 ± 3.30
0.036
LR ˆ 4.5, df ˆ 1, P ˆ 0.034
1.00
1.61 0.82 ± 3.16
0.167
LR ˆ 1.9, df ˆ 1, P ˆ 0.164
CVD indicates cardiovascular disease; CHD, coronary heart disease; RR, relative risk; CI, con®dence interval; BMI, body mass index;
LTPA, leisure-time physical activity; LR, likelihood ratio.
Adjusted for age, employment status, marital status, perceived health status, smoking status and alcohol consumption.
b
International Journal of Obesity
BMI, physical inactivity, low physical ®tness and mortality
N Haapanen-Niemi et al
1471
Table 5 Relative risks for all-cause- and CVD mortality for the women in 1980 ± 1996
CVD a
All causes
Stratum
RR a,b
BMIa
20.0 ± 24.9
25.0 ± 29.9
30.0
1.00
0.87
1.35
Total LTPAa energy expenditure index (kcal=week)
High
1.00
Moderate
0.59
Low
1.27
Single-item self-assessment of LTPA
Vigorous activity at least once a week
and some light activity
No or light intensity activity weekly
95% CI a
P
0.52 ± 1.46
0.599
0.76 ± 2.41
0.310
LRa ˆ 2.2, df ˆ 2, P ˆ 0.341
0.30 ± 1.18
0.136
0.69 ± 2.34
0.440
LR ˆ 6.8, df ˆ 2, P ˆ 0.033
1.00
1.61
Perceived physical ®tness compared with age-mates
Better
1.00
Similar
0.82
Worse
1.71
2 km walking ability
No dif®culties
At least some dif®culties
1.00
1.45
Climbing several ¯ights of stairs
No dif®culties
At least some dif®culties
1.00
2.39
RR b
1.00
0.85
1.36
1.00
0.43
1.17
95% CI
P
0.42 ± 1.74
0.664
0.60 ± 3.07
0.459
LR ˆ 1.3, df ˆ 2, P ˆ 0.527
0.16 ± 1.16
0.093
0.51 ± 2.68
0.717
LR ˆ 6.2, df ˆ 2, P ˆ 0.046
1.00
0.89 ± 2.92
0.114
LR ˆ 2.7, df ˆ 1, P ˆ 0.101
0.41 ± 1.65
0.582
0.72 ± 4.05
0.221
LR ˆ 5.9, df ˆ 2, P ˆ 0.054
0.78 ± 2.70
0.237
LR ˆ 1.4, df ˆ 1, P ˆ 0.243
1.25 ± 4.60
0.009
LR ˆ 7.2, df ˆ 1, P ˆ 0.007
4.68
1.00
0.82
1.89
1.00
1.25
1.00
3.38
1.41 ± 15.57
0.012
LR ˆ 9.4, df ˆ 1, P ˆ 0.002
0.32 ± 2.16
0.693
0.57 ± 6.27
0.299
LR ˆ 3.7, df ˆ 2, P ˆ 0.154
0.53 ± 2.90
0.611
LR ˆ 0.3, df ˆ 1, P ˆ 0.614
1.22 ± 9.41
0.020
LR ˆ 6.2, df ˆ 1, P ˆ 0.013
a
CVD indicates cardiovascular disease; RR, relative risk; CI, con®dence interval; BMI, body mass index; LTPA, leisure time physical
activity; LR, likelihood ratio.
b
Adjusted for age, employment status, marital status, perceived health status, smoking status and alcohol consumption.
moderate energy expenditure for LTPA, and the
LTPA index was statistically signi®cantly associated
with the risk of death. The men not engaging in
weekly vigorous LTPA had a suggestively increased
risk for CVD and CHD mortality. When compared
with the most active men, the men with no weekly
vigorous activity had relative risks of 1.61 (95% CI,
0.98 ± 2.64) for CVD mortality and 1.66 (95% CI,
0.92 ± 2.99) for CHD mortality. Among the inactive
women the relative risk for CVD mortality was 4.68
(95% CI, 1.41 ± 15.57) in the comparison with the
women who were vigorously active at least once a
week.
The men who perceived their physical ®tness level
to be worse than or similar to that of their age-mates
had an increased risk of death. Compared with the
men perceiving their physical ®tness level as better
than their age-mates, the men with a `worse' assessment had an increased mortality risk that varied
between 3.29 (95% CI, 1.80 ± 6.02) for all causes
and 4.64 (95% CI, 1.56 ± 13.84) for CHD. Among
the women the perceived physical ®tness measure was
suggestively associated with the all-cause mortality
risk but no single perceived ®tness category differed
statistically signi®cantly from the `better' category.
The men, but not the women, with self-reported
dif®culty in walking 2 km had an increased risk of
all-cause mortality. Compared with the subjects with
no functional dif®culties, the men with at least some
dif®culty walking 2 km had a relative risk of 1.62
(95% CI, 1.05 ± 2.50) for all-cause mortality. Dif®culty in climbing several ¯ights of stairs was associated with increased relative risks for all-cause and
CVD mortality among both the men and the women.
Compared with those with no functional dif®culties,
the men who had at least some dif®culty climbing
several ¯ights of stairs had relative risks of 1.47 (95%
CI, 0.97 ± 2.23) for all-cause and 1.85 (95% CI, 1.04 ±
3.30) for CVD mortality. Among the women the
relative risks were 2.39 (95% CI, 1.25 ± 4.60) for
all-cause and 3.38 (95% CI, 1.22 ± 9.41) for CVD
mortality.
There was no consistent interaction between the
LTPA and the perceived ®tness measurements with
respect to the risk of death, except for LTPA and the
perceived ability to climb several ¯ights of stairs,
which showed interaction with respect to the relative
risk of death from CVD among the men (results not
shown). These results indicate that the bene®cial
effects of LTPA on mortality are similar among
physically ®t and un®t subjects.
Discussion
This prospective follow-up study among middle-aged
and elderly men and women indicates that obesity (as
assessed by increased BMI) is not related to an
increased risk of all-cause and CVD mortality, but
International Journal of Obesity
BMI, physical inactivity, low physical ®tness and mortality
N Haapanen-Niemi et al
1472
low-level LTPA and a low level of perceived physical
®tness and functional capability are. The low selfreported ability to climb several ¯ights of stairs was a
very strong predictor of premature all-cause and
cause-speci®c mortality among both the men and the
women. The bene®cial effects of LTPA were similar
among the both obese and nonobese subjects and
among both the ®t and un®t subjects.
Several studies relating obesity and mortality have
included biases that have led to a systematic underestimate of the impact of obesity on premature mortality. According to Manson, et al 12 the discrepancy
between the results exists at least because of the
following three major biases common in previous
studies: (1) failure to control for cigarette smoking;
(2) inappropriate control of the biological effects of
obesity, such as hypertension, hyperglycemia and
diabetes; and (3) failure to control for weight loss
due to subclinical disease.
In this study several potential sources of obesityrelated bias were considered. Smoking, which has
been generally found to be more prevalent among
lean subjects, was controlled for at the beginning of
the study. Furthermore, to eliminate possible clinical
or subclinical illnesses that reduce body weight and
thereby attenuate the weight ± mortality association,
we excluded those men and women from the analysis
who had an initial BMI of less than 20.0 and those
who suffered from a disease or symptoms that totally
prevented participation in LTPA. Furthermore, initially assessed perceived health status was considered
as a confounding factor in multivariate analysis.
The major weaknesses of the current study were
related to the use of self-reported information about
physical ®tness and functional capability, LTPA and
BMI, for example. However, the association of LTPA
and perceived physical ®tness and functional capability with mortality was measured with several questions that were consistently associated with the
measured ®tness. Furthermore, the LTPA measures
that we used were consistently associated with
morbidity and mortality in our earlier analyses.21,26
The validity of the self-reported weight and thereby
also BMI, relative to measured estimates, was
assessed as high for these data.4 The information on
death was obtained from a national census, which is a
valid data source. Therefore, we believe that our
results indicate unbiased associations between living
habits, obesity and mortality among a middle-aged
and elderly Finnish population.
In this 16 y follow-up study the age-adjusted allcause and CVD mortality rates were lowest for the
men and women with a BMI of 25.0 ± 29.9, but in the
multivariate analysis these associations disappeared
for both the men and the women. According to current
evidence the direction between the level of BMI and
mortality can vary, even within the same study, across
the sex, age, race and cause of mortality.34 ± 36 A
recent meta-analysis demonstrated that, for 50 -y-old
white men followed for 30 y the lowest risk for all-
International Journal of Obesity
cause mortality is in the BMI range of 23.1 ± 27.9.8 In
a recent US study carried out among more than 1
million adults the lowest rates of death from all causes
were found at BMI level which varied from 23.5 to
24.9 in men and from 22.0 to 23.4 in women.11 The
risk of death increased with an increasing level of
BMI in white but not in black men and women. Due to
the mild association between moderate level of obesity (BMI < 32) and the risk of death in this exceptionally large study sample, the clinical importance
may be uncertain, however. In our study the number
of obese subjects was relatively small, but if the
relation between obesity and the risk of death were
strong, the association would probably have been
apparent also in our data.
The optimal level of BMI for the risk of death may
vary according to the disease in question, however. In
one male-population based study Shaper et al 9 found
that all-cause mortality was increased for men with a
BMI less than 20 and for men with a BMI equal to or
higher than 30, while the risk of CVD death, heart
attack, and diabetes increased progressively with
increasing BMI from the level of less than 20.
Furthermore, it has been suggested that central obesity
may be a stronger predictor of morbidity and mortality
than body weight or BMI.13
Our study showed that, although a low level of
LTPA was not associated with an increased risk of allcause mortality among men or women, physically
inactive men had about a 70% increased risk for
CVD and CHD mortality when compared with
highly active men. According to the energy-expenditure-based index, with which the weekly energy
expenditure for physical recreation, household
chores, commuting to and from work, ®tness exercise
and sport was estimated, the inactive women had an
increased risk of death when compared with the
women with moderate energy expenditure for
LTPA. The women with no weekly vigorous activity
had nearly a ®ve-fold risk for CVD mortality when
compared with the most active subjects. The current
evidence indicates that, when compared with people
who are highly active, sedentary people usually
experience between a 1.2-fold and a 2-fold increased
risk of dying from any cause.22 In considering CHD
mortality, the recent reviews and meta-analyses suggest that physically inactive subjects have even
higher, about double, the risk of the most active
subjects.22,37 ± 39 This conclusion is close to our estimates in men, but in women the protective effect of
vigorous LTPA was extremely strong.
A low level of perceived physical ®tness was a
strong predictor of premature mortality. Compared
with the men who perceived their ®tness status as
better than that of their age-mates, the men with the
`worse' assessment had a three-to nearly ®ve-fold risk
for premature mortality. One explanation for the
strong association may be that perceived physical
®tness is an integrated summary function of the
functional status and health of many organ systems,
BMI, physical inactivity, low physical ®tness and mortality
N Haapanen-Niemi et al
and the level of physical ®tness is frequently perceived as the degree of effort required to perform
daily activities.32 Among the women no single perceived ®tness category differed from the reference
category, however.
Perceived ability to climb several ¯ights of stairs
was a strong predictor of premature mortality among
both the men and the women. The predictive power of
stair climbing may be due to the high loading caused
by this customary activity, which in turn reveals
effectively insuf®cient cardiorespiratory and musculoskeletal function. A poor perceived ability to walk
2 km predicted an increased risk of all-cause mortality
only among the men. Our ®nding of a strong association between a low level of physical ®tness and an
increased risk of mortality agrees with the results of
several studies, carried out usually among men, in
which physical ®tness has been found to have a
graded, independent, long-term effect on all-cause
and CVD mortality among healthy, middle-aged and
elderly subjects.16,19,40
According to our statistical tests of interaction the
bene®cial role of LTPA on mortality were, with the
exception of the interrelationship between LTPA and
the ability to climb several ¯ights of stairs among the
men, similar for both the ®t and un®t subjects. Thus
far, only a few studies have compared the relative
importance of LTPA and physical ®tness as predictors
of mortality in the same study. Two studies compared
the role of LTPA and physical ®tness in the prevention of fatal or nonfatal CHD.18,41 In their study, Hein
et al 18 concluded that, among sedentary men, being
very ®t does not provide protection against CHD and
all-cause mortality; however un®t and sedentary men
had a higher risk of CHD than un®t and active men.
In our study the possible interaction effect between
the ®tness=LTPA measurements and BMI was estimated in relation to mortality. BMI was not associated
with an increased risk of mortality in our data and no
interaction effect was found after adjustment for
potential confounders. This ®nding indicates that the
bene®cial effects of LTPA and ®tness are similar
among obese and nonobese men and women. Some
earlier studies with male cohorts have concluded that
®t and overweight men have a lower rate of all-cause
mortality than un®t and normal weight men.28,29
Similarly, it has been concluded, even though not
tested, that ®t and overweight men have less risk of
CVD mortality than un®t men of normal weight.27
In conclusion, in contrast with our initial hypothesis, obesity was not found to be an independent
predictor of mortality among middle-aged and elderly
men and women. However, low-level LTPA seemed
to predict and a low level of perceived physical ®tness
and functional capability predicted an increased risk
of all-cause and CVD mortality among both men and
women. Self-reported dif®culty in climbing several
¯ights of stairs seemed to predict very strongly an
increased risk of death from all causes, CVD and
CHD both among the men and among the women. The
inverse association of high level of LTPA with mortality seemed to be similar among the obese and nonobese men and women and also among the ®t and
un®t subjects. Therefore, our results support the recent
recommendations,22,42 ± 44 according to which bene®cial effects of an active way of life are obtained not
only with respect to morbidity and mortality but also
with respect to body weight development.23 ± 27
1473
Acknowledgements
This study was supported by grants from the Juho
Vainio Foundation, The YrjoÈ Jahnsson Foundation,
the Finnish Ministry of Education and partially from
the Emil Aaltonen Foundation.
References
1 Hubert HB, Feinleib M, McNamara PM, Castelli WP. Obesity
as an independent risk factor for cardiovascular disease: a 26-y
follow-up of participants in the Framingham Heart Study.
Circulation 1983; 67: 968 ± 977.
2 Rissanen A, Knekt P, HelioÈvaara M, Aromaa A, Reunamaa A,
Maatela J. Weight and mortality in Finnish women. J Clin
Epidemiol 1991; 44: 787 ± 795.
3 Lee IM, Manson JAE, Hennekens CH, Paffenbarger RS. Body
weight and mortality: a 27-y follow-up of middle-aged men.
JAMA 1993; 270: 2823 ± 2828.
4 Manson JAE, Willett WC, Stampfer MJ, Colditz GA, Hunter
DJ, Hankinson SE, Hennekens CH, Speizer FE. Body weight
and mortality among women. New Engl J Med 1995; 333:
677 ± 685.
5 Rimm EB, Stampfer MJ, Giovannucci E, Ascherio A, Spiegelman D, Colditz GA, Willett WC. Body size and fat distribution
as predictors of coronary heart disease among middle-aged and
older US men. Am J Epidemiol 1995; 141: 1117 ± 1127.
6 Jousilahti P, Tuomilehto J, Vartiainen E, Pekkanen J, Puska P.
Body weight, cardiovascular risk factors and coronary mortality. 15-y follow-up of middle-aged men and women in eastern
Finland. Circulation 1996; 93: 1372 ± 1379.
7 Seidell JC, Verschuren M, van Leer EM, Kromhout D. Overweight, underweight, and mortality. A prospective study of 48
287 men and women. Arch Intern Med 1996; 156: 958 ± 963.
8 Troiano RP, Frongillo EA, Sobal J, Levitsky DA. The relationship between body weight and mortality: a quantitative analysis of combined information from existing studies. Int J
Obes Relat Metab Disord 1996; 20: 63 ± 75.
9 Shaper AG, Wannamethee SG, Walker M. Body weight:
implications for the prevention of coronary heart disease,
stroke, and diabetes mellitus in a cohort study of middle
aged men. Br Med J 1997; 314: 1311 ± 1317.
10 Expert Panel on the Identi®cation, Evaluation, and Treatment
of Overweight and Obesity in Adults. Executive summary of
the clinical guidelines on the identi®cation, evaluation, and
treatment of overweight and obesity in adults. Arch Intern Med
1998; 158: 1855 ± 1867.
11 Calle EE, Thun MJ, Petrelli JM, Rodriguez C, Heath CW.
Body-mass index and mortality in a prospective cohort of US
adults. New Engl J Med 1999; 341: 1097 ± 1105.
12 Manson JAE, Stampfer MJ, Hennekens CH, Willett WC.
Body weight and longevity; a reassessment. JAMA 1987;
257: 353 ± 358.
13 Solomon CG, Manson JAE. Obesity and mortality: a review of
the epidemiologic data. Am J Clin Nutr 1997; 66(Suppl):
1044 ± 1050.
14 World Health Organization. Obesity: preventing and managing the global epidemic. Geneva: WHO, 1997.
International Journal of Obesity
BMI, physical inactivity, low physical ®tness and mortality
N Haapanen-Niemi et al
1474
15 Paffenbarger RS, Hyde RT, Wing AL, Hsieh C. Physical
activity, all-cause mortality, and longevity of college alumni.
New Engl J Med 1986; 314: 605 ± 613.
16 Ekelund LG, Haskell WL, Johnson JL, Whaley FS, Criqui
MH, Sheps DS. Physical ®tness as a predictor of cardiovascular mortality in asymptomatic North American men: the
Lipid Research Clinics Mortality Follow-up Study. New Engl
J Med 1988; 319: 1379 ± 1384.
17 Leon AS, Connett J. Physical activity and 10.5 y mortality in
the Multiple Risk Factor Intervention Trial (MRFIT). Int J
Epidemiol 1991; 20: 690 ± 697.
18 Hein HO, Suadicani P, Gyntelberg F. Physical ®tness or
physical activity as a predictor of ischaemic heart disease? A
17-y follow-up in the Copenhagen Male Study. J Intern Med
1992; 232: 471 ± 479.
19 Sandvik L, Erikssen J, Thaulow E, Erikssen G, Mundal R,
Rodahl KI. Physical ®tness as a predictor of mortality among
healthy, middle-aged Norwegian men. New Engl J Med 1993;
328: 533 ± 537.
20 Lakka TA, VenaÈlaÈinen JM, Rauramaa R, Salonen R, Tumilehto J, Salonen JT. Relation of leisure-time physical activity
and cardiorespiratory ®tness to the risk of acute myocardial
infarction in men. New Engl J Med 1994; 330: 1549 ± 1554.
21 Haapanen N, Miilunpalo S, Vuori I, Oja P, Pasanen M.
Characteristics of leisure time physical activity associated
with decreased risk of premature all-cause and cardiovascular
disease mortality in middle-aged men. Am J Epidemiol 1996;
143: 870 ± 880.
22 US Department of Health and Human Services. Physical
activity and health: a report of the Surgeon General. Atlanta,
GA: US Department of Health and Human Services, Centers
for Disease Control and Prevention, National Center for
Chronic Disease Prevention and Health Promotion, 1996.
23 Williamson DF, Madans J, Anda RF, Kleinman JC, Kahn HS,
Byers T. Recreational physical activity and 10-y weight
change in a US national cohort. Int J Obes Relat Metab
Disord 1993; 17: 279 ± 286.
24 DiPietro L. Physical activity, body weight, and adiposity: an
epidemiologic perspective. Exerc Sport Sci Rev 1995; 23:
275 ± 303.
25 DiPietro L, Kohl HW, Barlow CE, Blair SN. Improvements in
cardiorespiratory ®tness attenuate age-related weight gain in
healthy men and women: the Aerobics Center Longitudinal
Study. Int J Obes Relat Metab Disord 1998; 22: 55 ± 62.
26 Haapanen N, Miilunpalo S, Pasanen M, Oja P, Vuori I.
Association between leisure time physical activity and 10-y
body mass change among working-aged men and women. Int J
Obes Relat Metab Disord 1997; 21: 288 ± 296.
27 Coakley EH, Rimm EB, Colditz G, Kawachi I, Willett WC.
Predictors of weight change in men: results from the Health
Professionals Follow-Up Study. Int J Obes Relat Metab
Disord 1998; 22: 89 ± 96.
28 Lee CD, Jackson AS, Blair SN. US weight guidelines: is it
also important to consider cardiorespiratory ®tness? Int J Obes
Relat Metab Disord 1998; 22(Suppl 2): 2 ± 7.
29 Barlow CE, Kohl HW, Gibbons LW, Blair SN. Physical
®tness, mortality and obesity. Int J Obes Relat Metab Disord
1995; 19(Suppl 4): 41 ± 44.
International Journal of Obesity
30 Miilunpalo S, Vuori I, Oja P, Pasanen M, Urponen H. Selfrated health status as a health measure: the predictive value of
self-reported health status on the use of physician services and
on mortality in the working-age population. J Clin Epidemiol
1997; 50: 517 ± 528.
31 Willlett WC. Weight loss in the elderly: cause or effect of poor
health? Am J Clin Nutr 1997; 66: 737 ± 738.
32 Ainsworth BE, Haskell WL, Leon AS, Jacobs DR, Montoye
HJ, Sallis JF, Paffenbarger RS. Compendium of physical
activities: classi®cation of energy costs of human physical
activities. Med Sci Sports Exerc 1993; 25: 71 ± 80.
33 Cox DR. Regression models and life tables (with discussion).
J R Stat Soc (B) 1972; 34: 187 ± 220.
34 Dorn JM, Schisterman EF, Winkelstein W, Trevisan M. Body
mass index and mortality in a general population sample of
men and women: the Buffalo Health Study. Am J Epidemiol
1997; 146: 919 ± 931.
35 Durazo-Arvizu RA, McGee DL, Cooper RS, Liao Y, Luke A.
Mortality and optimal body mass index in a sample of the US
population. Am J Epidemiol 1998; 147: 739 ± 749.
36 Stevens J, Cai J, Pamuk ER, Williamson DF, Thun MJ, Wood
JL. The effect of age on the association between body mass
index and mortality. New Engl J Med 1998; 338: 1 ± 7.
37 Powell KE, Thompson PD, Caspersen CJ, Kendrick JS.
Physical activity and the incidence of coronary heart disease.
A Rev Public Health 1987; 8: 253 ± 287.
38 Berlin JA, Colditz GA. A meta-analysis of physical activity in
the prevention of coronary heart disease. Am J Epidemiol
1990; 132: 612 ± 628.
39 Miller TD, Balady GJ, Fletcher GF. Exercise and its role in the
prevention and rehabilitation of cardiovascular disease. Ann
Behav Med 1997; 19: 220 ± 229.
40 Blair SN, Kampert JB, Kohl HW, Barlow CE, Macera CA,
Paffenbarger RS, Gibbons LW. In¯uences of cardiorespiratory
®tness and other precursors on cardiovascular disease and
all-cause mortality in men and women. JAMA 1996; 276:
205 ± 210.
41 Sobolski J, Kornitzer M, de Backer G, Dramaix M, Abramowicz M, Degre S, Denolin H. Protection against ischemic heart
disease in the Belgian Physical Fitness Study: physical ®tness
rather than physical activity? Am J Epidemiol 1987; 125:
601 ± 610.
42 Pate RR, Pratt M, Blair SN, Haskell WL, Macera CA,
Bouchard C, Buchner D, Ettinger W, Heath GW, King AC,
Kriska A, Leon AS, Marcus BH, Morris J, Paffenbarger RS,
Patrick K, Pollock ML, Rippe JM, Sallis J, Wilmore JH.
Physical activity and public health: a recommendation from
the Centers for Disease Control and Prevention and the
American College of Sports Medicine. JAMA 1995; 273:
402 ± 407.
43 Fletcher GF, Balady G, Froelicher VF, Hartley LH, Haskell
WL, Pollock ML. Exercise standards: a statement for health
care professionals from the American Heart Association.
Circulation 1995; 91: 580 ± 615.
44 NIH Consensus Development Panel on Physical Activity and
Cardiovascular Health. Physical activity and cardiovascular
health. JAMA 1996; 276: 241 ± 246.