American Journal of Epidemiology
Copyright © 2003 by the Johns Hopkins Bloomberg School of Public Health
All rights reserved
Vol. 157, No. 6
Printed in U.S.A.
DOI: 10.1093/aje/kwf219
Body Mass Index in Adolescence in Relation to Total Mortality: 32-Year Follow-up
of 227,000 Norwegian Boys and Girls
Anders Engeland1, Tone Bjørge2, Anne Johanne Søgaard1, and Aage Tverdal1
1
2
Division of Epidemiology, Norwegian Institute of Public Health, Oslo, Norway.
Department of Pathology, The Norwegian Radium Hospital, Oslo, Norway.
Received for publication July 2, 2002; accepted for publication September 12, 2002.
A large number of studies have explored the relation between body mass index (BMI) and mortality in adults.
The relation between BMI in adolescence and mortality has been investigated to a lesser extent. It has been
suggested that all-cause mortality is elevated among those who were overweight during adolescence, but the
limitation of previous studies has been study size. The present study explored this relation in a Norwegian cohort
of 227,003 boys and girls, aged 14–19 years, whose height and weight were measured during tuberculosis
screening in 1963–1975. These persons were followed for an average of 31.5 years (about 7.2 million personyears). A total of 7,516 deaths were registered. Multivariate Cox proportional hazards regression models were
used in the analyses. An increasing risk of death by increasing BMI in adolescence was observed. Mortality
among males whose baseline BMI was between the 85th and 95th percentiles and above the 95th percentile in
the US reference population was 30% and 80% higher, respectively, than that among those whose baseline BMI
was between the 25th and 75th percentiles. The corresponding rates among females were 30% and 100%. The
excess mortality among adolescents whose BMI was high was not clearly manifested before they reached their
thirties. Hence, BMI in adolescence is predictive of adult mortality.
adolescence; body mass index; cohort studies; mortality
Abbreviations: BMI, body mass index; CDC, Centers for Disease Control and Prevention; NCHS, National Center for Health
Statistics.
Overweight and obesity in childhood, adolescence, and
adulthood is an increasing problem in many countries. A
large number of studies have explored the relation between
body mass index (BMI) and mortality, but, in most of these
studies, height and weight have been measured in adults (1,
2). As a consequence, knowledge is limited about the
connection between obesity in adolescence and later
mortality (3–5).
The Third Harvard Growth Study included height and
weight measurements for 3,000 schoolchildren during 1922–
1935. A total of 508 lean or overweight adolescents aged 13–
18 years were followed for more than 50 years with regard to
death (3), and overweight in adolescence was associated
with increased mortality. In a long-term follow-up of Dutch
men aged 18 years, mortality 20–30 years later was 50
percent higher among those whose BMI ((weight in kg)/
(height in meters)2) was higher than 25 compared with men
whose BMI was 19.0–19.9 (6). Hoffmans et al. (6) also
found increased mortality among men whose BMI was less
than 18, which they ascribed to impaired health status.
It is known that obesity in adolescence also has other negative effects (5). Adolescent obesity has been shown to be
associated with early maturation, increased truncal deposition of fat (7), and lasting social effects on self-esteem and
body image (5, 7, 8). Obesity in childhood/adolescence also
seems to be an important predictor of adult obesity (5, 9),
although research that includes long-term follow-up data is
lacking (10).
BMI is not a perfect measure of adiposity in adolescents,
but it has been shown to be a valid measure of fatness in
adolescents (11). In addition, a workshop on childhood
obesity convened by the International Obesity Task Force in
1997 concluded that BMI offers a reasonable measure of
fatness in children and adolescents (12). Since height and
Correspondence to Dr. Anders Engeland, Division of Epidemiology, Norwegian Institute of Public Health, P.O. Box 4404 Nydalen, N-0403 Oslo,
Norway (e-mail: [email protected]).
517
Am J Epidemiol 2003;157:517–523
518 Engeland et al.
weight measurements are simple and inexpensive to collect
and often have been a routine part of health examinations,
BMI can be calculated in many epidemiologic studies.
There is no international agreement on an appropriate definition of obesity for adolescents (2), who are growing and
are in various stages of maturation. Hence, age- and sexspecific growth curves need to be used to define overweight
and risk for overweight. On the basis of a proposal from the
International Obesity Task Force workshop (13), Cole et al.
(14) proposed age- and sex-specific cutoff points to define
overweight and obesity in adolescents. These cutoff points
were linked to the adult categories for overweight (BMI, 25–
30) and obesity (BMI >30) and may be used in international
comparisons of the prevalence of overweight and obesity. In
the United States, the Centers for Disease Control and
Prevention (CDC; Atlanta, Georgia) at the National Center
for Health Statistics (NCHS; Hyattsville, Maryland) has
created growth charts for children and adolescents up to the
age of 20 years based on data from US health examinations
(15). The CDC/NCHS guidelines for adolescents suggest
using age- and sex-specific BMI to identify adolescents at
the upper end of the distribution as being “at risk for overweight” (BMI, 85th–94th percentiles) and overweight (BMI
≥95th percentile). In addition to age- and sex-specific
growth curves for the 85th and 95th percentiles, CDC/NCHS
also provides growth curves for the 3rd, 5th, 10th, 25th, 50th,
75th, 90th, and 97th percentiles (16).
The aim of the present study was to explore the relation
between BMI and total mortality in a cohort of more than
200,000 Norwegian boys and girls aged 14–19 years at
measurement.
MATERIALS AND METHODS
Subjects
During 1963–1975, height and weight were measured as
part of a screening program to detect tuberculosis in the
general Norwegian population (17–19). This material
(which included 1.7 million persons aged ≥15 years) has
been described previously, and the impact of height and
weight on morbidity and mortality has been reported (17,
18). However, because of a relatively short follow-up period,
the relation between BMI in adolescence and mortality could
not be analyzed extensively at that time. The mass examination was compulsory for persons aged 15 years or older, but
height and weight were also measured for some persons less
than age 15 years. The attendance rate was about 85 percent
for persons aged 15 years or older (17). Body weight (kilograms) was measured by using scales that were calibrated
regularly and was noted to the nearest half kilogram. Body
height was measured and noted to the nearest centimeter.
Height was measured without shoes, and weight was
measured with the subject wearing light clothing.
In the present study, all persons measured at age 14–19
years were included, except 2,333 for whom the measurements were not performed according to the protocol (for
example, they were wearing shoes), persons who declined to
be measured, persons who were disabled, or women who
claimed to be pregnant. The first measurement for each
person was included. Altogether, 227,048 persons were
eligible for the analysis. BMI was defined as (weight in kg)/
(height in meters)2. Information on covariates other than sex,
age, time of measurement, and area of residence was not
available.
All residents of Norway are assigned a unique 11-digit
identification number. By linkage to the Death Registry at
Statistics Norway, it was possible to follow almost all
persons in the present study from date of measurement until
emigration, death, or June 30, 2001. A relatively small
number of persons (n = 29) was lost to follow-up. Furthermore, 16 were excluded because their measurements were
taken after the end of follow-up (when day and month of
measurement were missing, the date was set as June 30).
Statistical analysis
Multivariate Cox proportional hazards regression models,
with time since measurement as the time variable, were fitted
to obtain relative risk estimates of dying (20). It was
assumed that the hazard function for a person with a covariate vector x = (x1, x2, ... ,xp)′ could be expressed by h(t; x) =
h0(t) × exp(x′ × β), where h0(t) represents the hazard function for a person with covariate values all equal to zero, and
β = (β1, β2, ... ,βp)′ is a vector of regression coefficients. The
first measurement obtained at age 14–19 years was used. In
the analyses, the following three categorized variables were
included:
1. Age at measurement: 14–16 years, 17–19 years
2. Year of birth: 1943–1949, ≥1950
3. BMI at baseline: followed the guidelines from CDC/
NCHS (15, 16) by using percentiles in a US reference
population: <3rd, 3rd–4th, 5th–9th, 10th–24th, 25th–
74th, 75th–84th, 85th–94th, ≥95th
The proportionality assumption in the Cox model was
assessed by inspecting log-minus-log plots, results from
stratified analyses, and results from separate analyses for
different time periods after measurement. The analyses were
performed by using the statistical program package SPSS
(21). Analyses were performed separately for each sex.
Results were presented as relative risks, with 95 percent
confidence intervals, of dying.
RESULTS
A total of 227,003 persons (115,141 boys and 111,862
girls) (mean age, 17.0 years) were followed for on average
31.5 years (range, 0–38 years), comprising 7,157,147
person-years (table 1). Among these persons, 7,516 deaths
were observed. Mean age at death was 36.8 years for males
and 39.8 years for females. The overall mean BMI was 21.1
in boys and 21.4 in girls (table 2). Sixty-three percent of the
boys and 65 percent of the girls had a BMI between the 25th
and 75th percentiles in the US reference population, whereas
5.2 percent of the males and 8.0 percent of the females had a
BMI above the 85th percentile. The distribution of BMI in
two age groups is shown in figures 1 and 2. The distribution
for those aged 17–19 years shifted to the right compared with
that for those persons aged 14–16 years. By June 30, 2001,
Am J Epidemiol 2003;157:517–523
BMI in Adolescence in Relation to Mortality 519
TABLE 1. Number of observed person-years and deaths in a study of the association of body mass index in
adolescence with total mortality, Norway, 1963–1975
Boys
Variable
Girls
No. of
deaths
No. of
person-years
Death
rate*
No. of
deaths
No. of
person-years
Death
rate*
0–4
591
574,017
103
149
557,420
27
5–9
542
569,898
95
142
552,763
26
10–14
513
565,830
91
218
549,220
40
15–19
614
561,739
109
275
546,243
50
20–24
846
556,714
152
472
542,939
87
25–29
1,106
518,954
213
613
507,902
121
856
276,165
310
579
277,344
209
14–16
2,511
1,876,891
134
1,166
1,825,199
64
17–19
2,557
1,746,425
146
1,282
1,708,632
75
No. of years since measurement
≥30
Age (years) at measurement
Year of birth
1943–1949
1,348
844,205
160
743
860,719
86
≥1950
3,720
2,779,111
134
1,705
2,673,112
64
BMI† category‡
<3rd
127
73,314
173
28
40,753
69
3rd–4th
78
54,055
144
25
29,790
84
5th–9th
207
149,119
139
63
92,195
68
10th–24th
746
576,270
129
272
40,2631
68
25th–74th
3,141
2,301,251
136
1,493
2,282,278
65
75th–84th
412
280,262
147
305
402,653
76
85th–94th
266
151,816
175
216
248,779
87
91
37,229
244
46
34,751
132
5,068
3,623,316
140
2,448
3,533,831
69
≥95th
Total
* No. of deaths per 100,000 person-years.
† BMI, body mass index ((weight in kg)/(height in meters)2).
‡ Percentiles in a US reference population (15).
214,600 persons (94.5 percent) were still alive and living in
Norway. The death rate was about two times higher for
males than for females—irrespective of BMI level.
Table 3 shows the relative risks of dying according to
different categories of baseline BMI adjusted for other
factors. Mortality was elevated in the groups with the highest
BMI. Compared with the referent group (BMI, 25th–74th
percentiles in the reference population), mortality in the
highest BMI group (≥95th percentile) was 80 percent and
100 percent higher among males and females, respectively.
A tendency toward increased mortality among the thinnest
persons was also observed.
The association between baseline BMI and mortality was
also analyzed separately in different time periods after
measurement. In these analyses, the three categories with the
lowest baseline BMI values were combined, and the two
categories with the highest baseline BMI values were
combined. The excess mortality in adolescents whose baseline BMI was above the 85th percentile in the US reference
Am J Epidemiol 2003;157:517–523
population was not significant in the first 10 years after
measurement (data not shown). Thereafter, the excess
mortality in this group was about 50 percent (both sexes
combined) compared with those whose baseline BMI was
between the 25th and 75th percentiles in the US reference
population. The same analysis was performed for intervals
of attained age instead of time since measurement (table 4).
At age 20–29 years, a modest, nonsignificant, increased risk
was observed for those persons whose baseline BMI was
above the 85th percentile in the reference population
compared with those whose baseline BMI was between the
25th and 75th percentiles. At age 30–58 years, the excess
risk was more than 50 percent for both males and females.
DISCUSSION
In this study, we explored the relation between BMI and
mortality in a cohort of 227,003 Norwegian boys and girls
followed for an average of 31.5 years. An increasing risk of
520 Engeland et al.
TABLE 2. Mean values of study characteristics by category of body mass
index, Norway, 1963–1975
Mean value§
Sex and BMI*
category†
Percentage of
subjects‡
Weight
(kg)
Height
(cm)
BMI
Age at
measurement
(years)
Male
<3rd
2.1
50.8
172.9
16.9
17.4
3rd–4th
1.5
53.4
173.1
17.7
17.4
5th–9th
4.2
55.1
173.7
18.2
17.3
10th–24th
16.0
58.0
174.0
19.1
17.2
25th–74th
63.4
64.8
174.5
21.2
16.9
75th–84th
7.7
73.0
174.6
23.8
16.7
85th–94th
4.2
79.2
174.7
25.9
16.7
≥95th
1.0
93.4
174.6
30.5
16.7
100.0
64.4
174.4
21.1
17.0
<3rd
1.2
44.0
164.3
16.3
17.4
3rd–4th
0.9
46.2
164.6
17.0
17.3
5th–9th
2.6
47.7
164.6
17.6
17.3
10th–24th
11.5
50.3
164.6
18.5
17.2
25th–74th
64.6
56.8
164.4
21.0
17.0
75th–84th
11.3
64.6
163.9
24.0
16.8
85th–94th
7.0
70.9
163.6
26.5
16.8
≥95th
1.0
85.1
163.3
31.9
16.7
100.0
57.7
164.3
21.4
17.0
Total
Female
Total
* BMI, body mass index ((weight in kg)/(height in meters)2).
† Percentiles in a US reference population (8).
‡ Percentages do not total 100 because of rounding.
§ Additional mean values (both sexes): year of birth, 1952; year of adolescent
measurement,1969.
FIGURE 1.
Distribution of body mass index among study subjects aged 14–16 years, Norway, 1963–1975.
Am J Epidemiol 2003;157:517–523
BMI in Adolescence in Relation to Mortality 521
FIGURE 2.
Distribution of body mass index among study subjects aged 17–19 years, Norway, 1963–1975.
death by increasing BMI in adolescence was observed.
Mortality among males whose baseline BMI was between
the 85th and 95th percentiles and above the 95th percentile in
the US reference population was 30 percent and 80 percent
higher, respectively, than that among those persons whose
baseline BMI was between the 25th and 75th percentiles.
The corresponding numbers for females were 30 percent and
100 percent.
Use of Cox proportional hazards regression models
assumes that relative risks do not vary by time. In the present
study, a markedly lower relative risk of dying was observed
in the first 10 years after measurement than later for persons
whose BMI was above the 85th percentile in the reference
population. This finding violates the proportionality assumption in the Cox model. However, because of a low number of
deaths during the first 10 years, the impact of the first 10
TABLE 3. Relative risks of death, with 95% confidence intervals, obtained from a multivariate
analysis including age at measurement, year of birth, and body mass index, Norway, 1963–1975
Boys
Girls
Variable
RR*
95% CI*
RR
95% CI
14–16
1.00
Referent
1.00
Referent
17–19
1.12
1.06, 1.19
1.21
1.11, 1.32
1943–1949
1.00
0.93, 1.07
0.99
0.90, 1.10
≥1950
1.00
Referent
1.00
Referent
<3rd
1.27
1.06, 1.52
1.05
0.72, 1.53
3rd–4th
1.05
0.84, 1.32
1.29
0.87, 1.92
5th–9th
1.02
0.88, 1.17
1.04
0.81, 1.33
10th–24th
0.95
0.87, 1.03
1.04
0.91, 1.18
25th–74th
1.00
Referent
1.00
Referent
75th–84th
1.08
0.97, 1.20
1.15
1.01, 1.30
85th–94th
1.29
1.14, 1.47
1.31
1.14, 1.52
≥95th
1.82
1.48, 2.25
2.03
1.51, 2.72
Age at measurement (years)
Year of birth
BMI* category†
* RR, relative risk; CI, confidence interval; BMI, body mass index ((weight in kg)/(height in meters)2).
† Percentiles in a US reference population (15).
Am J Epidemiol 2003;157:517–523
522 Engeland et al.
TABLE 4. Relative risks of death, with 95% confidence intervals, obtained from a multivariate
analysis by body mass index in adolescence and attained age, Norway, 1963–1975*
Sex and
attained age
(years)
BMI† in adolescence‡
<25th
25th–74th
≥85th
75th–84th
RR†
95% CI†
RR
95% CI
RR
95% CI
RR
95% CI
20–29
1.07
0.93, 1.24
1.00
Referent
0.97
0.77, 1.22
1.17
0.91, 1.51
30–39
1.06
0.93, 1.21
1.00
Referent
1.00
0.81, 1.24
1.60
1.30, 1.96
40–49
0.92
0.83, 1.03
1.00
Referent
1.11
0.94, 1.30
1.41
1.18, 1.68
50–58
0.84
0.64, 1.11
1.00
Referent
1.47
1.03, 2.12
2.22
1.52, 3.25
20–29
0.78
0.56, 1.10
1.00
Referent
0.96
0.67, 1.37
1.20
0.82, 1.75
30–39
1.17
0.95, 1.46
1.00
Referent
1.29
1.01, 1.63
1.29
0.98, 1.70
40–49
1.03
0.87, 1.21
1.00
Referent
1.14
0.95, 1.36
1.44
1.19, 1.73
50–58
1.03
0.72, 1.48
1.00
Referent
1.10
0.74, 1.65
1.89
1.30, 2.75
Male
Female
* Adjusted for age at measurement and year of birth.
† BMI, body mass index ((weight in kg)/(height in meters)2); RR, relative risk; CI, confidence interval.
‡ Categorized according to percentiles in a US reference population (15).
years on the overall relative risk of dying was minor when
the whole period was analyzed. Excluding the first 10 years
of observation led to a slightly higher relative risk of dying
for persons with the highest baseline BMI.
One of the major strengths of this study is the large number
of persons included, who were recruited from the general
population. The subjects were recruited from the national
tuberculosis screening program, which was compulsory for
persons aged 15 years or older. The measurements were
performed in a standardized way. Follow-up of the study
subjects was almost complete until the end of follow-up; of
the 227,048 persons eligible for the study, 97.8 percent were
registered as either being alive by the end of follow-up or
having died during follow-up. A total of 2.2 percent had
emigrated, and only 33 (for four of them, the date on which
they were lost was known) persons were lost to follow-up.
Obesity in childhood/adolescence has many negative
health consequences, which have been divided into the categories of immediate physical and social, intermediate
(cardiovascular risk factor levels, persistent obesity into
adulthood), and long-term (adult morbidity and mortality)
(5). In the present study, the impact of adolescent BMI on
long-term mortality was explored.
Because mortality is low during adolescence as well as
young and middle-aged adulthood, a very large number of
persons and a long follow-up are necessary to observe a
sufficiently large number of deaths to obtain precise estimates. Both conditions were met in the present study.
However, although follow-up in the present study was up to
38 years, the oldest persons were only aged 58 years when
exiting the study. Hence, the excess mortality we observed
among persons with a high BMI in adolescence shows
excess mortality at middle age.
Studies including BMI in adolescence have used different
definitions of overweight and obesity (5). In the present
study, we decided to use age- and sex-specific growth charts
from CDC/NCHS to group the adolescents by BMI. Another
alternative was to use the growth charts from Cole et al. (14).
However, those growth charts correspond to BMI levels of
25 and 30 at age 18 years only. The CDC/NCHS growth
charts made it possible to develop more detailed categories.
However, the present material was also analyzed by using
the growth curves from Cole et al. (14). Mortality among
boys whose BMI corresponded to an adult BMI of 25–30
and to an adult BMI of more than 30 was 25 percent (95
percent confidence interval: 13, 38) and 125 percent (95
percent confidence interval: 79, 185), respectively, higher
than among those whose BMI corresponded to an adult BMI
of less than 25. The corresponding figures for girls were 27
percent (95 percent confidence interval: 12, 43) and 114
percent (95 percent confidence interval: 61, 182).
After 55 years of follow-up of 508 persons, Must et al. (3)
observed excess mortality among males, but not females, who
were overweight (BMI, >75th percentile in the US reference
population) in adolescence compared with those who were
lean (BMI, 25th–49th percentiles in the reference population).
The observed increased risk of death was independent of adult
BMI. In the present study, excess mortality was found for both
sexes among persons whose BMI in adolescence was above
the 85th percentile in a US reference population compared
with those whose BMI was lower (25th–74th percentiles).
Among persons whose BMI was above the 95th percentile in
the reference population, the mortality rate was more than 80
percent and 100 percent higher among boys and girls, respectively, than among persons whose BMI was between the 25th
and 75th percentiles.
A weakness of the present study was our lack of information on other potential confounders besides age, sex, and
birth year regarding the association between BMI and
mortality. Other potential confounders include smoking
habits, social class, and physical activity. Inclusion of
smoking status in the regression model did not significantly
Am J Epidemiol 2003;157:517–523
BMI in Adolescence in Relation to Mortality 523
change the estimated relative risks in the study by Must et al.
(3), but the overall impact of these factors on the association
between BMI and mortality is unclear. However, smoking
habits, social class, and physical activity in adulthood might
be influenced by BMI in adolescence; hence, these factors
might partially be intermediate variables.
In a cohort of Dutch men aged 18 years, Hoffmans et al.
(6) found increased mortality among those whose BMI was
25 or higher compared with those whose BMI was lower.
However, the increase in mortality was evident only after a
follow-up of more than 20 years (6). The necessity of having
a long follow-up period when studying the association
between BMI and mortality was also made clear in the
present study. The excess mortality among overweight
adolescents was not apparent the first 10 years after
measurement. However, both males and females whose BMI
was above the 85th percentile in the US reference population
during adolescence already evidenced excess mortality in
their thirties compared with those whose BMI was between
the 25th and 75th percentiles.
In adolescence and young adulthood, mortality is low, and
a large proportion of the deaths are due to accidents or
suicides. After age 30 years, a larger proportion of the deaths
are due to illnesses that may be influenced by BMI.
However, since no information on cause of death was available in this study, we could not perform analyses on different
causes of death. It is also possible that adverse health effects
of obesity during adolescence are connected to long-term
obesity. We had no information on how old the study
subjects were at onset of obesity.
In summary, this study showed that overweight adolescents have increased long-term mortality. The excess
mortality is not clearly manifested before they reach their
thirties. Even though BMI is not regarded as an ideal
measure of adiposity in adolescents, it is predictive of adult
mortality.
REFERENCES
1. The BMI in Diverse Populations Collaborative Group. Effect of
smoking on the body mass index-mortality relation: empirical
evidence from 15 studies. Am J Epidemiol 1999;150:1297–
308.
2. World Health Organization Consultation on Obesity. Preventing and managing the global epidemic: report of a WHO Consultation on Obesity, Geneva, Switzerland, June 3-5, 1997.
Geneva, Switzerland: World Health Organization, 1998:1–276.
Am J Epidemiol 2003;157:517–523
3. Must A, Jacques PF, Dallal GE, et al. Long-term morbidity and
mortality of overweight adolescents. A follow-up of the Harvard Growth Study of 1922 to 1935. N Engl J Med 1992;327:
1350–5.
4. Dietz WH, Robinson TN. Use of the body mass index (BMI) as
a measure of overweight in children and adolescents. J Pediatr
1998;132:191–3.
5. Must A, Strauss RS. Risks and consequences of childhood and
adolescent obesity. Int J Obes Relat Metab Disord 1999;
23(suppl 2):S2–11.
6. Hoffmans MDAF, Kromhout D, de Lezenne Coulander C. The
impact of body mass index of 78,612 18-year-old Dutch men on
32-year mortality from all causes. J Clin Epidemiol 1988;41:
749–56.
7. Dietz WH. Health consequences of obesity in youth: childhood
predictors of adult disease. Pediatrics 1998;101(suppl):518–25.
8. Gortmaker SL, Must A, Perrin JM, et al. Social and economic
consequences of overweight in adolescence and young adulthood. N Engl J Med 1993;329:1008–12.
9. Whitaker RC, Wright JA, Pepe MS, et al. Predicting obesity in
young adulthood from childhood and parental obesity. N Engl J
Med 1997;337:869–73.
10. Parsons TJ, Power C, Logan S, et al. Childhood predictors of
adult obesity: a systematic review. Int J Obes Relat Metab Disord 1999;23(suppl 8):1–107.
11. Pietrobelli A, Faith MS, Allison DB, et al. Body mass index as
a measure of adiposity among children and adolescents: a validation study. J Pediatr 1998;132:204–10.
12. Bellizzi MC, Dietz WH. Workshop on childhood obesity: summary of the discussion. Am J Clin Nutr 1999;70:173S–5S.
13. Dietz WH, Bellizzi MC. Introduction: the use of body mass
index to assess obesity in children. Am J Clin Nutr 1999;70:
123S–5S.
14. Cole TJ, Bellizzi MC, Flegal KM, et al. Establishing a standard
definition for child overweight and obesity worldwide: international survey. BMJ 2000;320:1240–3.
15. Kuczmarski RJ, Ogden CL, Grummer-Strawn LM, et al. CDC
growth charts: United States. Adv Data 2000;314:1–27.
16. National Center for Health Statistics. CDC growth charts:
United States, August 2002. (http://www.cdc.gov/growthcharts/).
17. Waaler HT. Height, weight and mortality. The Norwegian
experiment. Acta Med Scand 1984;suppl 679:1–56.
18. Tverdal A. Body mass index and incidence of tuberculosis. Eur
J Respir Dis 1986;69:355–62.
19. Bjartveit K. The National Health Screening Service: from fight
against tuberculosis to many-sided epidemiological activities.
(In Norwegian). Nor Epidemiol 1997;7:157–74.
20. Cox DR, Oakes D. Analysis of survival data. London, United
Kingdom: Chapman and Hall Ltd, 1984.
21. SPSS Inc. SPSS for Windows, release 10.0.7. Chicago, IL:
SPSS Inc, 2000.