© International Epidemiological Association 1999
International Journal of Epidemiology 1999;28:911–918
Printed in Great Britain
Social background, adult body-height
and health
Karri Silventoinen,a Eero Lahelmaa and Ossi Rahkonenb
Study
objective
To study the socio-demographic determinants of body-height and the bearing of
these determinants on the association between body-height and health among
Finnish adults.
Data and
Method
Cross-sectional population survey including questions on social background,
body-height and health, and retrospective questions on childhood living conditions.
The data derive from a representative Survey on Living Conditions collected by
Statistics Finland in 1994. The response rate was 73%. Male and female respondents >20 years were included in the analysis (N = 8212). Statistical methods
include regression analysis and logistic regression analysis.
Results
Body-height was strongly associated with year of birth, region, childhood living
conditions and education among adult men and women. Body-height was also
associated with limiting long-standing illness and perceived health as below good.
Tall men had the best health and short men the poorest health. Among women
the association of body-height with health differed from men, as tall women
showed high levels of limiting long-standing illness, notably musculo-skeletal
diseases. Adjusting for the background variables weakened but did not abolish
the association between poor health and short stature among men and women.
Conclusions Short stature is associated with poor health among Finnish men and women. A
non-linear association among women was found for musculo-skeletal diseases.
The studied social background factors explained only little of the association
between body-height and health.
Keywords
Body-height, health, social background
Accepted
3 March 1999
Final body-height is achieved as a result of the combination of
genetic and environmental factors.1 The impact of environmental factors is seen as an increase in average body-height
over the developed world during this century2,3 as well as
systematic differences in body-height between population
subgroups.4,5 It has been suggested that body-height can be
used as an indicator of fetal and childhood living conditions.6
From the epidemiological viewpoint an interesting issue is the
relationship between short stature and poor health. This association has been found for various domains of health including
mortality and morbidity, particularly cardiovascular diseases
and perceived health.4,7–9 In addition to poor nutrition,10
increased energy consumption during adolescence,11 maternal
undernutrition12 and maternal smoking during pregnancy,13
diseases14 and even psycho-social factors15 can put growth at
risk and contribute to final body-height.
a Department of Public Health and b Department of Social Policy, University
of Helsinki, Helsinki, Finland.
Reprint requests to: Karri Silventoinen, Department of Public Health, PO Box
41, 00014 University of Helsinki, Finland. E-mail: karri.silventoinen@
helsinki.fi
The study by Forsdahl in 197716 suggested that poor childhood living conditions would be a risk factor for adult health,
particularly mortality from cardiovascular diseases. Several
subsequent studies17–19 have reported associations between
childhood environment and health. Recently Barker et al.20–22
have emphasized the impact of undernourishment during
fetal life and infancy as a risk factor of cardiovascular diseases.
Barker23 suggests that permanent changes in blood pressure
and cholesterol level due to undernutrition could lie behind this
association. Barker’s hypothesis has, however, been criticized
methodologically24,25 as well as empirically as new evidence
has contrasted with that of previous studies.26,27
However, childhood is not only biologically but also a psychosocially vulnerable period in the human development. Health
behaviours and attitudes are likely to have a bearing on the
association between early environment and later health.28
A shortcoming in many epidemiological studies is that bodyheight is used as an indicator of childhood environment without sufficient attention to factors contributing to the association
between body-height and adult health. Also in many countries
genetic variation between areas and ethnic groups can confound
the observed differences. A particular problem in morbidity
911
912
INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
studies is that tall body-height is a potential risk factor for
musculo-skeletal diseases.29 This should be noted particularly
when the functional consequences of illness are considered.
Thus it is not self-evident that childhood environment lies
behind the association between body-height and health.
This study examines the associations between body-height,
social background and current health among Finnish men and
women. For this purpose we first examine the association of
adult body-height with a number of demographic and social
environmental factors indicating early living conditions. Secondly, we examine the impact of these factors on the
association between body-height and adult health.
Data and Methods
Sample
The data derive from the Finnish 1994 Survey on Living
Conditions collected by the government statistical authorities,
Statistics Finland. The number of respondents was 8650 and the
response rate was 74% for women and 72% for men. The
sampling represents satisfactorily the non-institutional Finnish
population aged >15. Subjects >20 years are included in the
analysis and the final number of subjects is 8212.
Self-reported body-height
Body-height was elicited by asking ‘How tall are you in centimetres?’ Comparing self-reports and measured body-height has
shown that especially poor, old and short men tend to slightly
overestimate their body-height. Differences in the overestimation between the socio-demographic groups are not more than
one or two centimetres. Among women similar but smaller
differences in overreporting are seen.30 If overreporting affects
the results the impact is likely to be conservative.
Measurement of health status
Two health indicators, limiting long-standing illness and perceived health, are used. The general question concerning
long-standing illness reads: ‘Do you have any long-standing
illness, disability or infirmity?’ If the answer is yes, a follow-up
question is asked: ‘Does your illness/disability restrict your
work or does it limit your daily activities (gainful employment, housework, schooling, studying)?’ The response alternatives were: ‘A great deal’, ‘To some extent’ and ‘Not at all’.
Diseases reported have been classified according to major disease
groups. Comparing self-reported diseases with those diagnosed
by a physician has shown relatively good agreement for cardiovascular diseases but poorer agreement for musculo-skeletal
diseases.31
Perceived health was elicited by asking the respondent to
describe his/her general health as ‘Very good’, ‘Good’, ‘Average’, ‘Poor’ or ‘Very poor’. Only 92 respondents reported their
health as very poor and this category was combined with poor
health. In prospective studies poor perceived health has been
found to predict mortality which is likely to indicate the good
reliability of this health indicator.32
Background variables
Place of residence was divided first into Southern, Western,
Eastern and Northern part of Finland. Secondly, a division was
made by urbanization level: urbanized, semi-urbanized and
rural municipalities. Mother tongue was classified into Finnish,
Swedish or an other language.
Childhood living conditions were elicited by two questions on
economic difficulties and alcohol problems in the childhood
family. The questions read: ‘When you think about your childhood years, that is the years before the age of 16, did your
family have long-lasting economic difficulties?’, and ‘Did a
member of your family have problems due to alcohol?’ A third
indicator of childhood living conditions is age when the
respondent had started work. This was studied by asking when
did the respondent first have a job which lasted at least one
year. The variable was divided into four categories: ,15 years,
15–16 years, 17–18 years and >19 years. Young age at starting
work reflects the low socioeconomic status of parents but
physically heavy work, common especially in Finnish
countryside, may also contribute to the growth velocity.
Information on education derives from a national register
of educational degrees. Educational attainment (ISCEDclassification) was categorized into three levels. ‘Higher’ equals
a university degree or an examination in another higher educational institution, requiring a total of >13 years of education.
‘Secondary’ equals secondary school plus vocational training, or
the matriculation examination, requiring an average total of
10–12 years of education. ‘Basic’ equals compulsory education
or less i.e. a maximum of 9 years of education in all.
Statistical methods
Mean body-height in Table 1 is age-standardized except for
body-height by year of birth. Direct age-standardization was
carried out using 5-year age groups. The male study population
was used as the standard population for men and the female
population for women.
Associations between childhood living conditions and bodyheight were analysed by using ordinary regression analysis with
dummy variables by including all variables into the model
simultaneously. Associations between body-height and health
were analysed by using logistic regression analysis. For this
purpose the outcome variables, limiting long-standing illness
and perceived health, were dichotomized. Those who reported
an illness which limited daily life at least to some extent had a
limiting long-standing illness. For perceived health we combined the categories ‘very good’ and ‘good’ as well as categories
‘average’, ‘poor’ and ‘very poor’.
The association between body-height and health may be nonlinear and body-height was classified into three categories
estimated by using a formula: mean body height ±1 * standard
deviation (= 170 cm and 183 cm for men, 157 cm and 169 cm
for women). For the same reason year of birth was included as
a classified variable divided into one-year age groups. Results
are presented as odds ratios (OR) by using average body-height
as the reference category (OR = 1.00). Models were fitted using
the GLIM statistical package.33
In the modelling a four-step strategy was utilized. First the
base model includes only year of birth. At the second step,
region of residence, urbanization level and mother tongue are
included. These factors may include genetic variation, which
can affect the results. Region of residence is included as
information on the region of birth is lacking in the data. Urbanization level is included to control the impact of migration.
If regional differences are caused by selective migration this
HEIGHT AND HEALTH
913
Table 1 Proportions of respondents, age-standardizeda mean body-heights (cm) and standard deviations by background variables. Men and
women
Men
Women
%
Mean
SD
%
Mean
SD
1973–1964
17.8
179.24
6.67
16.6
166.15
5.57
1963–1954
22.3
178.50
5.89
23.1
165.15
5.59
1953–1944
24.8
176.68
5.71
23.1
163.75
5.02
1943–1934
16.9
175.41
5.94
15.3
162.14
5.32
1933–1924
12.5
173.54
6.01
12.0
161.12
5.79
5.7
172.18
8.38
9.9
160.43
7.26
Year of birth
1923–
Region of residence
Southern
46.6
177.53
6.37
49.1
163.91
5.43
Western
28.1
176.46
5.94
27.2
163.63
6.03
Eastern
14.1
175.04
5.47
13.1
162.08
5.38
Northen
11.2
175.19
6.54
10.7
162.24
5.52
Urban
55.3
176.97
6.25
59.4
163.56
5.51
Semi-urban
15.4
176.46
6.69
14.4
163.80
5.51
Rural
19.3
175.78
5.80
26.3
162.67
5.99
Finnish
92.8
176.45
6.23
93.9
163.30
5.68
Swedish
6.7
178.17
5.73
5.7
164.84
5.63
Other
0.5
–
–
0.4
–
–
No
73.7
176.87
6.17
73.1
163.58
5.60
Yes
24.6
175.77
6.19
25.3
162.77
5.91
1.7
–
–
1.6
–
–
No
86.7
176.69
6.26
82.9
163.48
5.71
Yes
12.7
175.55
5.63
16.5
162.85
5.29
0.6
–
–
0.6
–
–
Level of urbanization
Mother tongue
Economic difficulties
Missing
Alcohol problems
Missing
Age at first job (years)
>19
43.4
177.10
6.17
43.3
163.68
5.44
17–18
21.4
176.25
6.39
22.7
163.17
5.57
15–16
23.7
175.87
5.62
21.6
162.68
5.70
,15
5.9
175.41
6.24
5.6
163.03
6.13
Missing
6.0
–
–
6.8
–
–
Higher
21.4
178.22
5.85
17.8
163.84
4.86
Secondary
41.8
176.87
5.86
43.9
163.45
5.55
Basic
36.8
175.35
6.16
38.3
162.67
5.86
No illness
43.6
177.14
6.06
38.1
163.45
5.27
Non-limiting illness
27.3
176.67
6.22
28.2
163.70
5.48
Limiting illness
20.6
175.74
6.30
25.4
163.37
5.84
8.5
175.72
7.00
8.3
163.02
6.52
Very good
20.3
177.89
6.00
21.8
163.92
5.46
Good
42.1
177.28
6.29
41.0
163.94
5.43
Average
30.2
176.09
6.41
30.4
163.26
5.59
7.4
175.53
6.20
6.8
163.95
5.26
4076
176.55
6.57
4131
163.38
5.99
Education
Long-standing illness
Illness limiting to a great extent
Perceived health
Poor or very poor
All
a Except year of birth.
914
INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
should be seen as differences in mean body-height between
urban and rural areas. At step three indicators of childhood
living conditions are added to the model. At the final step four,
educational level is included. As the social status of the parents
is unavailable education also reflects the socioeconomic
background in addition to the respondent’s current social status.
Analyses were also carried out including each variable in the
model step-by-step but this did not change the results.
Results
Background variables
The average body-height among Finns has increased constantly
during this century (Table 1). The difference between birth
cohorts born before 1924 and in 1964–1973 was 7.06 cm
among men and 5.72 cm among women. In the model (Table 2)
adjusting for all variables the regression coefficient was 0.12
among men and 0.10 among women. This means about one
centimetre growth in average body-height per decade.
Body-height differences by region and mother tongue were
clear. Body-height was tallest in the Southern part and shortest
in the Eastern part of the country both among men and women.
The difference was 1.99 (95% CI : 1.32–2.64) cm among men
and 1.54 (95% CI : 0.99–2.09) cm among women after adjustment for the other background variables. Body-height among
the Swedish-speaking minority was taller than among the
Finnish-speaking majority. After adjustment the difference was
1.14 (95% CI : 0.36–1.92) cm among men and 0.86 (95%
CI : 0.14–1.59) cm among women.
Childhood living conditions were associated with bodyheight. Men and women who had had economic difficulties in
their childhood family were shorter than others. After adjustment for the other variables this difference was 0.59 (95%
CI : 0.14–1.04) cm among men and 0.48 (95% CI : 0.09–0.87)
cm among women. The variable ‘alcohol problems in the
family’ was also statistically significant among men but not
among women. The difference among men was 0.86 (95%
CI : 0.27–1.45) cm. Age at first job was not statistically significant when the other variables were adjusted for. Among men,
however, the difference between the categories 15–16 years and
>19 years (0.27–1.45 cm), and the linear trend for the variable,
were statistically significant.
Education also showed a strong and linear association with
body-height. The age-adjusted body-height was greatest among
Table 2 Regression analysis of body-height. Men and women
Men
Constant
Women
Parameter coefficient
Standard error
Parameter coefficient
Standard error
183.5
0.50
168.9
0.44
0.1199
0.0071
0.1001
0.0058
Year of birth
Regression line
Region of residence
Southern
0
0
Western
–0.64
0.24
–0.15
0.21
Eastern
–1.99
0.33
–1.54
0.28
Northern
–1.82
0.33
–1.52
0.29
Level of urbanization
Urban
0
0
Semi–urban
–0.36
0.28
0.31
0.25
Rural
–0.35
0.25
–0.33
0.22
1.14
0.40
0.86
0.37
–3.13
1.55
1.01
1.39
0.23
–0.48
0.30
–0.39
Mother tongue
Finnish
Swedish
Other
0
0
Economic difficulties
No
0
Yes
–0.59
0
0.20
Alcohol problems
No
0
Yes
–0.86
0
0.24
Age at first job (years)
>19
0
0
17–18
–0.35
0.26
–0.13
15–16
–0.63
0.27
–0.29
0.23
0.24
,15
–0.66
0.41
–0.35
0.37
Education
Higher
0
0
Secondary
–0.87
0.28
–1.07
0.29
Basic
–2.04
0.30
–1.70
0.29
HEIGHT AND HEALTH
men and women with the highest education and shortest
among those with basic education only. The adjustment narrowed these differences slightly. When the other background
variables were adjusted for the difference between the highest
and the lowest educational categories was still 2.04 (95%
CI : 1.45–2.63) cm among men and 1.70 (95% CI : 1.13–2.27)
cm among women. Thus differences in height between
educational groups were larger than for any other background
variable, excluding year of birth.
Health
Among men the association between body-height and limiting
long-standing illness as well as perceived health was linear
(Table 1). Men without illness or with very good perceived
health were tallest. In a parallel way average body-height was
lowest for those with long-standing illness limiting them a great
deal, and those with poor perceived health. The gap between
the tallest and the shortest body-height was 1.42 cm by limiting
long-standing illness and 2.36 cm by perceived health.
However, among women the association was not linear. The
tallest body-height was among those with non-limiting longstanding illness and those with good perceived health. Shortest
women were those who reported long-standing illness limiting
them a great deal, and those who perceived their health as
average. The difference between the greatest and the shortest
body-height was 0.68 cm for both limiting long-standing illness
and perceived health.
Table 3 summarizes results from the logistic regression model
for limiting long-standing illness. Among men the association
between body-height and morbidity was linear. Tall men had
the best and short men the poorest health. In the first model
adjusted for year of birth only, the OR for having a limiting
long-standing illness for short men was 1.42 (95% CI : 1.17–
1.72) but for tall men 0.69 (95% CI : 0.53–0.88). Adjusting for
additional variables the difference narrowed but remained
statistically significant. In the final model the OR was 1.30 (95%
CI : 1.07–1.59) for short and 0.75 (95% CI : 0.58–0.97) for tall
915
men. Adding childhood living conditions to the model had the
strongest impact on the differences.
Both short and tall women reported more limiting longstanding illness compared to women with average body-height.
In the first model the OR for short women was 1.28 (95%
CI : 1.06–1.57) and 1.22 (95% CI : 1.00–1.49) after adjustment.
Among tall women the OR was 1.16 (95% CI : 0.95–1.42) in
the first model increasing slightly to 1.21 (95% CI : 0.98–1.48)
when all variables were adjusted for. The health difference for
short women was statistically significant in all models. For tall
women the difference was not statistically significant in any
model but was close to the 95% limit of statistical significance.
Similarly among men childhood living conditions had the
strongest impact on the association between body-height and
limiting long-standing illness.
Table 4 shows corresponding models for perceived health.
Among men the association was similar to that for limiting
long-standing illness. The OR for short men was 1.47 (95%
CI : 1.22–1.78) but for tall men 0.76 (95% CI : 0.59–0.97) in the
first model compared to men with average body-height. In
the final model, adjusting for all background variables, the OR
were 1.39 (95% CI : 1.15–1.68) and 0.81 (95% CI : 0.63–1.04)
respectively. Short women had somewhat poorer perceived
health than other women. The difference was not, however,
statistically significant in any model. Tall women had perceived
health similar to women with average stature. Among men as
well as women, childhood living conditions and education had
the strongest impact on the association between body-height
and perceived health.
Finally, the association between body-height and major disease
groups was studied. Results are shown only for cardiovascular
and musculo-skeletal diseases, for which statistically significant
associations were found. Among men the association was
significant only for cardiovascular diseases (Table 5). In this
group the OR among short men was 1.43 (95% CI : 1.15–1.76)
adjusting only for year of birth. The OR decreased to 1.32 (95%
CI : 1.06–1.65), when the other variables were adjusted for.
Table 3 Logistic regression analysis of limiting long-standing illness, odds ratios (OR) with 95% CI for short and tall men and women compared
to average stature (OR = 1.0)
Men
Model
Women
Short
Tall
Short
Tall
Year of birth
1.42 (1.17–1.72)
0.69 (0.53–0.88)
1.28 (1.06–1.57)
1.16 (0.95–1.42)
1 + Region + Urbanization +
Mother tongue
1.39 (1.15–1.69)
0.70 (0.54–0.90)
1.27 (1.04–1.55)
1.16 (0.95–1.43)
2 + Economic difficulties +
Alcohol problems + Age of first job
1.33 (1.10–1.62)
0.72 (0.56–0.94)
1.24 (1.01–1.51)
1.18 (0.96–1.45)
3 + Education
1.30 (1.07–1.59)
0.75 (0.58–0.97)
1.22 (1.00–1.49)
1.21 (0.98–1.48)
Table 4 Logistic regression analysis of perceived health, odds ratios (OR) with 95% CI for short and tall men and women compared to average
stature (OR = 1.0)
Men
Model
Women
Short
Tall
Short
Tall
Year of birth
1.47 (1.22–1.78)
0.76 (0.59–0.97)
1.21 (0.99–1.47)
1.00 (0.80–1.24)
1 + Region + Urbanization +
Mother tongue
1.47 (1.21–1.77)
0.77 (0.60–0.99)
1.18 (0.97–1.44)
1.01 (0.81–1.25)
2 + Economic difficulties +
Alcohol problems + Age of first job
1.44 (1.19–1.75)
0.78 (0.61–1.01)
1.16 (0.95–1.42)
1.01 (0.81–1.25)
3 + Education
1.39 (1.15–1.68)
0.81 (0.63–1.04)
1.13 (0.93–1.38)
1.04 (0.83–1.29)
916
INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
Table 5 Logistic regression analysis of cardiovascular diseases, odds ratios (OR) with 95% CI for short and tall men and women compared to
average stature (OR = 1.0)
Men
Model
Women
Short
Tall
Short
Tall
Year of birth
1.43 (1.15–1.76)
0.92 (0.67–1.26)
1.08 (0.87–1.34)
0.93 (0.71–1.23)
1 + Region + Urbanization +
Mother tongue
1.40 (1.13–1.73)
0.93 (0.68–1.28)
1.04 (0.84–1.29)
0.95 (0.72–1.26)
2 + Economic difficulties +
Alcohol problems + Age of first job
1.33 (1.07–1.65)
0.96 (0.70–1.33)
1.03 (0.83–1.28)
0.95 (0.72–1.26)
3 + Education
1.32 (1.06–1.65)
0.98 (0.71–1.35)
1.01 (0.81–1.26)
0.97 (0.74–1.29)
Table 6 Logistic regression analysis of musculo-skeletal, odds ratios (OR) with 95% CI for short and tall men and women compared to average
stature (OR = 1.0)
Men
Model
Women
Short
Tall
Short
Tall
Year of birth
1.26 (1.02–1.57)
1.01 (0.76–1.35)
1.40 (1.14–1.72)
1.24 (0.98–1.59)
1 + Region + Urbanization +
Mother tongue
1.25 (1.00–1.55)
1.01 (0.76–1.35)
1.39 (1.13–1.71)
1.26 (0.99–1.61)
2 + Economic difficulties +
Alcohol problems + Age of first job
1.21 (0.97–1.51)
1.03 (0.77–1.38)
1.37 (1.11–1.69)
1.29 (1.01–1.64)
3 + Education
1.18 (0.95–1.47)
1.06 (0.79–1.41)
1.33 (1.07–1.64)
1.32 (1.03–1.69)
Among women the association was significant only for
musculo-skeletal diseases (Table 6). For limiting long-standing
illness the association was U-shaped; that is, both tall and short
women showed poor health. Adjusting for year of birth only the
OR was 1.40 (95% CI : 1.14–1.72) among short and 1.24 (95%
CI : 0.98–1.59) among tall women. After adjustment the OR for
short women was 1.33 (95% CI : 1.07–1.64) and for tall women
1.32 (95% CI : 1.03–1.69).
Discussion
As expected, year of birth was the most important background
variable associated with body-height. The results show a secular
trend of body-height in Finland similar to other western
countries2–4 and consistent with previous Finnish studies.34
The rapid rise in the standard of living in Finland during this
century is likely to lie behind this development.
The differences between the South-Western and NorthEastern parts of Finland are very clear. Previous studies, based
on military registers since the 17th century, show that a similar
regional pattern in body-height has existed for at least 200
years.34–36 The regional differences in the standard of living or
selective migration may partly explain the regional differences.
However, the level of urbanization was not statistically significantly associated with body-height.
Genetic differences form another potential explanation. There
are inconsistent indications about genetic variation in Finland.
Kajanoja37 discovered differences between the Eastern and
Western parts of Finland in the distribution of blood type and
colour blindness, which indicate genetic differences. By contrast
Nevanlinna38 found no regional differences when using blood
markers, suggesting that the Finnish population is genetically
fairly uniform.
Another variable which may reflect the genetic impact is
mother tongue. According to Virtaranta-Knowles et al.39 the
Swedish speaking population differs genetically from the
majority. In this study mother tongue had an independent
association with body-height. Previous studies have reported
similar results.34
Childhood living conditions showed strong associations with
body-height. Previous studies have reported similar results from
other countries.40–42 It is, however, noteworthy that the variable ‘alcohol problems in the childhood family’ was not statistically significant among women, but only among men. Among
men these differences were larger than those for economic
difficulties in childhood. Slight differences were also found by
the variable ‘age when started work’ for men but not women.
Because women’s puberty starts earlier than men’s, strenuous
work started at the same age may affect the growth of men
more strongly than that of women.
Education was the second most important background
variable associated with body-height after year of birth. We had
no information on parents’ socioeconomic status and own education also reflects the intergenerational social background. However, there are studies showing a positive relationship between
social mobility and body-height.43,44 Causes for this association
are broadly unknown but it is possible that health in childhood
is a partial explanation.
A common feature in the associations between body-height
and the background variables was a similar pattern for men
and women. An exception was alcohol problems which was
statistically significant among men but not among women.
The differences according to the background variables were,
however, larger among men. Taken together all background
variables explained 18% of the variance of body-height among
men and 16% among women. Previous studies have reported
similar sex differences which may be caused by genetic
factors.45,46
The association between body-height and health among men
and women was a linear one. Using both limiting long-standing
illness and perceived health as health indicators short men
showed poorer and tall men better health than men with
HEIGHT AND HEALTH
average stature. Adjusting for the background variables
weakened the associations but they remained statistically
significant.
With regard to specific disease groups, an association was
found between short stature and cardiovascular diseases. This is
in accordance with previous studies which have suggested an
association between childhood living conditions and cardiovascular diseases.7,19,24,25 Cardiovascular diseases may explain
the poorer health of short men but not why tall men had better
than average health.
Among women the relationship was U-shaped for limiting
long-standing illness because tall women also showed poorer
health. No association could be found for poor perceived health
and tall stature. A similar U-shaped association was found for
musculo-skeletal diseases among women. Previous studies29
have suggested that tall stature is a risk factor for low-back
conditions. However, the physiological mechanisms between
musculo-skeletal diseases and short stature are not known. The
association between poor health and short stature among
women may reflect poor adult and childhood socioeconomic
conditions. Additionally musculo-skeletal diseases may contribute to body-height. For cardiovascular diseases an association
similar to that for men could not be found for women, possibly
due to better average health and smaller number of cases.
To sum up, the social background factors included in the
analysis showed only a small impact on the association between
body-height and health although they did show a strong association with body-height itself. The fact that elements reflecting
childhood living conditions are associated with adult bodyheight does not necessarily mean that these factors lie behind
the association found between body-height and health. This
complicates the interpretation of the association between bodyheight and health as being due to the studied childhood living
conditions. In future research additional childhood variables as
well as measurement of detailed diseases are needed.
917
10 Ruel M, Rivera J, Habicht J, Martorell R. Differential response to early
nutrition supplementation: long-term effects on height at adolescence. Int J Epidemiol 1995;24:404–12.
11 Post B, Kemper H, Welten D, Coudert J. Dietary pattern and growth
of 10–12-years-old Bolivian girls and boys: relation between altitude
and socioeconomic status. Am J Hum Biol 1997;9:51–62.
12 Kusin J, Kardjati S, Houtkooper J, Renqvist U. Energy supplemen-
tation during pregnancy and postnatal growth. Lancet 1992;
340:623–26.
13 Butler N R, Goldstein H. Smoking in pregnancy and subsequent child
development. Br Med J 1973;4:573–75.
14 Malleson P. Pain syndromes, disability and chronic disease in child-
hood. Curr Opin Rheumatol 1991;3:860–66.
15 Skuse D, Albanese A, Stanhope R et al. A new stress-related syndrome
of growth failure and hyperphagia in children, associated with reversibility of growth-hormone insufficiency. Lancet 1996;348:353–58.
16 Forsdahl A. Are poor living conditions in childhood and adolescence
an important risk factor for arteriosclerotic heart disease? Br J Prev Soc
Med 1977;31:91–95.
17 Nieto J, Szklo M, Comstock G. Childhood weight and growth rate as
predictors of adult mortality. Am J Epidemiol 1992;136:201–13.
18 Rahkonen O, Lahelma E, Huuhka M. Past or present? Childhood
living conditions and current socioeconomic status as determinants of
adult health. Soc Sci Med 1997;44:327–36.
19 Forsén T, Eriksson G, Tuomilehto J et al. Mother’s weight in pregnancy
and coronary heart disease in a cohort of Finnish men: follow up
study. Br Med J 1997;315:837–40.
20 Barker D, Gluckman P, Godfrey K et al. Fetal nutrition and cardio-
vascular disease in adult life. Lancet 1993;341:938–41.
21 Barker D, Martyn C. The maternal and fetal origins of cardiovascular
disease. J Epidemiol Community Health 1992;46:8–11.
22 Barker D, Osmond C, Golding J. Height and mortality in the counties
of England and Wales. Ann Hum Biol 1990;17:1–6.
23 Barker D. Mothers, Babies, and Disease in Later Life. Plymouth: BMJ
Publishing Group, 1994.
24 Elford J, Whincup P, Shaper A. Early life experience and adult cardio-
vascular disease: longitudinal and case-control studies. Int J Epidemiol
1991;20:833–44.
References
1 Sinclair D. Human Growth after Birth. London: Oxford University Press,
1989.
2 Murata M, Hibi I. Nutrition and the secular trend of growth. Horm Res
1992;38(Suppl.1):89–96.
3 Tanner J. Growth as a measure of the nutritional and hygienic status
of a population. Horm Res 1992;38(Suppl.1):106–15.
4 Macintyre S. A review of the social patterning and significance of
measures of height, weight, blood pressure and respiratory function.
Soc Sci Med 1988;27:327–37.
5 Eveleth P, Tanner J. Worldwide Variation of Human Growth. Cambridge:
Cambridge University Press, 1976.
6 Fogel W. New sources and new techniques for the study of secular
25 Elford J, Shaper A, Whincup P. Early life experience and cardio-
vascular disease-ecological studies. J Epidemiol Community Health
1992;46:1–8.
26 Baker D, Illsley R, Vågerö D. Today or in past? The origins of
ischaemic heart disease. J Public Health Med 1993;15:243–48.
27 Christensen K, Vaupel J, Holm N, Yashin A. Mortality among twins
after age 6: fetal origins hypotesis versus twin method. Br Med J
1995;310:432–36.
28 Wadsworth M. Health inequalities in the life course perspective.
Soc Sci Med 1997;44:859–69.
29 Heliövaara M, Mäkelä M, Knekt P et al. Determinants of sciatica and
low-back pain. Spine 1991;16:608–14.
30 Boström G, Diderichsen F. Socioeconomic differentials in misclassi-
trends in nutritional status, health, mortality, and the process of
aging. Historical Methods 1993;26:5–43.
fication of height, weight and body mass index based on questionnaire data. Int J Epidemiol 1997;26:860–66.
7 Notkola V. Living conditions in childhood and coronary heart disease
31 Heliövaara M, Aromaa A, Klaukka P et al. Reliability and validity of
in adulthood. A mortality and morbidity study in two areas of
Finland. Societas Scientiarum Fennica 29, Helsinki, 1985.
32 Idler E, Benyamin Y. Self-rated health and mortality: a review of
8 Nyström Peck M, Vågerö D. Adult body height, self perceived health
interview data on chronic diseases. J Clin Epidemiol 1993;46:181–91.
twenty-seven community studies. J Health Soc Behav 1997;38:21–37.
and mortality in Swedish population. J Epidemiol Community Health
1989;43:380–84.
33 Aitkin M, Anderson D, Francis B, Hinde J. Statistical Modelling in GLIM.
9 Leon D, Davey Smith G, Shipley M, Strachan D. Adult height and
34 Kajanoja P. A study in the morphology of the Finns and its relation to
mortality in London: early life, socioeconomic confounding, or
shrinkage? J Epidemiol Community Health 1995;49:5–9.
the settlement of Finland. Annales Academiae Scientiarum Fennicae Series
A Medica 1971, 146.
Oxford: Oxford Statistical Science Series, 1990.
918
INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
35 Kajava Y, Finne J. Mitteailungen über die Körpergrösse des finnischen
41 Reading R, Raybould S, Jarvis S. Deprivation, low birth weight and
Mannes Ende des 18. und Anfang des 19. Jahrhunderts. Annales
Academiae Scientiarum Fennicae Medica Serie A 1926;25:5.
children’s height: a comparison between rural and urban areas.
Br Med J 1993;307:1458–62.
36 Westerlund F. Studier i Finlands antropologi II. Kropslängden hos
42 Cernerud L, Elfving J. Social inequality in height: a comparison
Finlands befolkning. Fennia 1900;18(2):32–89.
37 Kajanoja P. A contribution to the physical anthropology of the Finns.
Variation of the ABO, Rhesus, MN, P and Lewis blood group
frequencies, PTC taste ability and colour blindness. Annales Academiae
Scientiarum Fennicae Medica Series A 1972, 153.
38 Nevanlinna H. The Finnish population structure. A genetic and
genealogical study. Hereditas 1972;71:195–236.
between 10-year-old Helsinki and Stockholm children. Scand J Soc Med
1995;23:23–27.
43 Nyström Peck M. Childhood environment, intergenerational mobility
and adult health—evidence from Swedish data. J Epidemiol Community
Health 1992;46:71–74.
44 Cernerud L. Height and social mobility: a study of the height of 10
year olds in relation to socio-economic background and type of formal
schooling. Scand J Soc Med 1995;23:28–31.
39 Virtaranta-Knowles K, Sistonen P, Nevanlinna H. A population
45 Bielicki T. Physical growth as a measure of economic well-being of
genetic study in Finland: comparison of the Finnish- and Swedishspeaking populations. Hum Hered 1991;41:248–64.
populations: the twentieth century. In: Falkner F, Tanner J (eds).
Human Growth: A Comprehensive Treatise. 2nd Edn Vol. 3. New York:
Plenum Press, 1986.
40 Kuh D, Wadsworth M. Parental height: childhood environment and
subsequent adult height in a national birth cohort. Int J Epidemiol
1989;18:663–68.
46 Kuh D, Power C, Prodgers B. Secular trend in social class and sex
differences in adult height. Int J Epidemiol 1991;20:1001–09.