Vol. 25, No. 4
Printed In Great Britain
International Journal of Epidemiology
O International EpWemiologlcal Association 1996
Relationship between Morbidity and
Extreme Values of Body Mass Index
in Adolescents
AYALA LUSKY,* VITA BARELL,* FLORA LUBIN," GIORA KAPLAN,* VERONIQUE LAYANI,*
ZIPORA SHOHAT,* BOAZ LEV+ AND MICHAEL WIENER*
Lusky A (Health Services Research Unit, Gertner Institute, Sheba Medical Center, Tel Hashomer 52621, Israel),
Barell V, Lubln F, Kaplan G, Layani V, Shohat Z, Lev B and Wiener M. Relationship between morbidity and extreme
values of body mass index in adolescents. International Journal of Epidemiology 1996; 25: 829-834.
Background. Although the association between overweight and cardiovascular risk factors is well documented in crosssectional and longitudinal studies, reports of adolescent morbidity associated with underweight in industrialized countries
are rare.
Methods. This population-based study includes approximately 110 000 17 year old Israeli Jewish males who underwent
routine physical examination at army induction centres. Computerized data tapes include overall health profiles, specification of physical and mental conditions, and height and weight measurements. Medically significant conditions are those
with sufficient severity to preclude service in a combat unit.
Results. Functional limitation is more prevalent at both extremes of the body mass index (BMI) distribution: 149.5/1000
among severely underweight individuals and 164.3/1000 among severely overweight subjects. Overweight was associated with hypertension (14 9/1000 among the severely overweight), as well as joint conditions of the lower extremities,
mainly hip, ankle and knee disorders. Functional disorders associated with underweight are bronchial and lung conditions, including asthma (14.2 and 18.9/1000 in the mildly and severely underweight), scoliosis, intestinal conditions and
emotional disorders (mainly neurosis).
Conclusions. Both under- and overweight are associated with morbidity at age 17. Intervention programmes should
begin at an early age.
Keywords. BMI, overweight, underweight, medical conditions, adolescents
associated with an increased risk of morbidity and
mortality from coronary heart disease and colorectal
cancer, and it was a better predictor of morbidity than
overweight in adulthood.
Most studies which relate intercurrent morbidity and
mortality to obesity in childhood and adolescence focus
on the association between obesity and cardiovascular
risk factors.14"21 Obese children are at greater risk for
hypertension,16 total plasma cholesterol, l6 ' 18 diabetes 22
and orthopaedic conditions.19-20 There are only a few reports of morbidity associated with underweight at childhood and adolescence in industrialized societies.23'24
The Israeli male birth cohort of 17 year olds represents an ideal population to provide an estimate of the
prevalence of under- and overweight and of related
morbidity, based on a nationwide cohort. The data include individual measurements on height, weight, and a
wide range of medical conditions, as well as functional
status.
The aim of the present study was to determine
whether 17 year old Israeli Jewish males with extreme
Physiological childhood and the growth process tend to
end by age 18. Population studies show that weight at
this age represents the ideal reference of individual's
weight to be maintained throughout adult life.1
Overweight, a common disorder among adults in
western countries, is often accompanied by significant
morbidity, such as hypertension,2'3 lipid disturbances,3'4
diabetes3 and osteoarthritis.5 About 80% of obese male
adolescents become obese adults6 and, as such, have
a higher risk of future morbidity and mortality.7"13
Hoffman found higher mortality rates for coronary
heart disease in adulthood among individuals with a
BMI 3=25 at age 18, while the leanest group showed
excess cancer mortality. In a longitudinal study,
Must9 showed that overweight in adolescent males was
• Health Services Research Unit, Geruier Institute, Sheba Medical
Center, Tel Hashomer 52621, Israel.
** Department Of Clinical Epidemiology, Gertner Institute, Sheba
Medical Center, Israel.
* Israel Defense Force Medical Corps.
829
830
INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
BMI values have a higher prevalence of functional
problems than those within the normal range, and to
describe the medical conditions associated with
underweight and overweight in this population.
METHODS
Study Population
All Jewish 17 year old males in Israel are assessed for
military service and undergo routine physical examination at five regional induction centres, including
measurements of height and weight. These draft boards
carried out medical examinations of all recruits. Data
were analysed separately for each induction centre to
evaluate possible confounding between geographical
distribution of ethnic groups and regional draft boards.
No bias was ascertained.
This study is based on re-analysis of data on approximately 110 000 Jewish Israeli male inductees, only
10% of whom were foreign-born.
Ethnicity was defined by country of origin of the
paternal line, i.e. the father's country of birth if foreignborn, or the origin of the paternal grandfather when the
father was also Israeli-born. In all, 19 ethnic origin
groups were analysed—Bulgaria, Egypt, Ethiopia,
Germany, Hungary, India, Iran, Iraq, Libya, Morocco,
Poland, Romania, Russia, Syria, Turkey, Tunisia, USA,
Yemen and at least three generations Israeli-born.
These were re-grouped into three major subgroups
according to their BMI distribution:
• Those whose BMI distribution was shifted to the left
of the national distribution (the light group—India,
Yemen and Ethiopia).
• Ethnic groups whose BMI distribution was shifted to
the right of the national distribution (the heavy group—
Hungary, Romania, Russia, Libya and Tunisia).
• All other countries whose BMI distribution was similar to the national BMI distribution (the normal group).
Relative Weight
Measurement of weight and height are taken routinely
at induction centres. Relative weight was calculated by
body mass index (BMI = weight[kg]/height [m2]).
Inductees were classified into five weight categories
according to their ethnic-specific BMI distribution:
severe underweight—BMI < 5th percentile; mild underweight—5th < BMI < 15th; normal weight—15th
=s BMI =s 85th percentiles; mild overweight—BMI
85th < BMI *s 95th percentile; and severe overweight
—BMI > 95th percentile.
Health assessment was based on a standard and systematic clinical examination of each recruit independent of all anamnestic and anthropometric information.
TABLE 1 Prevalence and odds ratio (OR) of overall significant
functional limitation by weight category
Weight category
Normal
Mild underweight
Severe underweight
Mild overweight
Severe overweight
Prevalence (per 1000)
103.5
119.5
149.5
116.6
164.3
OR
95% CI
.0
.18
.52
.14
.70
(1.10-1.25)
(1.40-1.64)
(1.07-1.21)
(1.57-1.83)
Severity levels for all conditions are pre-defined in a
standard code book and the single most severe condition determines the overall level of physical fitness and
whether the recruit is qualified for combat or for military service at all. Conditions were considered medically significant if they were of sufficient severity to
disqualify the inductee for service in a combat unit.
Conditions presented in this study include: hypertension, lower limb problems (knee and hip joint, meniscus
and ankle), scoliosis, lung conditions, gastrointestinal
disease, and neuroses. Definitions are given in the
Appendix.
Statistical Analysis
Prevalence rates (per 1000 individuals) for overall
functional limitation and for selected medical conditions were evaluated by the five weight categories. The
association between medical conditions and weight was
assessed by calculating odds ratios (OR) and 95%
confidence intervals (CI).25
RESULTS
Extremes of BMI and General Functional Limitation
The overall prevalence of significant functional limitation, i.e. that which prevents assignment to active
combat duty, was more prevalent at both extremes of
the BMI distribution (Table 1). The functional limitation rate among mildly underweight individuals was
119.5/1000 compared to 103.5/1000 among those with
normal weight, while the severely underweight reached
a rate of 149.5/1000. Recruits who were mildly overweight had a rate similar to those mildly underweight,
while the severely overweight reached a rate of 164.3/
1000 (OR = 1.70). All OR were significant at P < 0.05.
Association between Overweight and Specific Medical
Conditions
A positive association was found between hypertension
and BMI (Figure 1). Recorded prevalence of hypertension increased as BMI increased, and exceeded 5/1000
831
MORBIDITY AND EXTREME BMI
rate per 1000
17
18
19
20
21 22 23 24 25 26
BMI'[weight(kg)/height(m*)]
27
28
29 30*
—— smoothed
FIGURE 1 Prevalence of hypertension by body mass index (BMI = weight
ikgVheightlm2})
at BMI of 26. Hypertension by weight categories, as
well as other conditions associated with overweight, are
presented in Table 2. Prevalence of hypertension was
14.9/1000 severely overweight subjects and 3.2 for the
mildly overweight, compared to 1.1 for normal weight
individuals (OR = 13.1 and 2.8, respectively). Joint
conditions of the lower extremities were also more prevalent among severely and mildly overweight recruits
(13.6 and 9.1/1000 respectively) mainly due to hip
disorders (OR = 3.9), ankle disorders (OR = 2.3), knee
problems (OR = 1.7) and meniscal tears (OR = 2.4,
borderline significance).
Association between Underweight and Specific
Medical Conditions
Functional disorders associated with underweight are
presented in Table 3. The most significant functional
disorders were bronchial and lung conditions, including
asthma. The severely underweight subjects had a
prevalence rate of 18.9, compared to 14.2 in the mildly
TABLE 2 Prevalence and odds ratio (OR) for selected medical conditions associated with overweight
Medical conditions
Weight category
Hypertension
Normal
Mild overweight
Severe overweight
Normal
Mild overweight
Severe overweight
Normal
Mild overweight
Severe overweight
Normal
Mild overweight
Severe overweight
Normal
Mild overweight
Severe overweight
Normal
Mild overweight
Severe overweight
Lower extremities:
Hip
Meniscus
Ankle
Knee
Prevalence (per 1000)
1.1
3.2
14.9
5.4
9.1
13.6
0.8
1.1
3.1
0.8
1.0
2.0
1.8
3.2
4.0
1.5
26
2.6
OR
95% CI
1.0
28
13.1
1.0
1.7
2.5
1.0
1.4
3.9
1.0
1.2
2.4
1.0
1.8
2.3
1.0
1.7
1.7
(1.7-3.9)
(9.2-17 1)
_
(1.3-2.1)
(1.9-3.2)
(0.5-2.2)
(1.8-5.9)
(0.4-1.9)
(0.9-3.9)
(1.1-2.5)
(1.3-3.3)
(1.0-2.4)
(0.8-2.7)
832
INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
TABLE 3 Prevalence and odds ratio (OR) for selected medical conditions associated with underweight
Medical conditions
Bronchial/lung
Scoliosis
Intestinal
disease'
Neurosis
1
Weight category
Prevalence (per 1000)
OR
Normal
Mild underweight
Severe underweight
Normal
Mild underweight
Severe underweight
Normal
Mild underweight
Severe underweight
Normal
Mild underweight
Severe underweight
10.8
14.2
189
1.0
1.4
2.2
0.4
0.9
2.2
10.9
14.1
14.4
1.0
1.3
1.8
1.0
1.4
2.3
1 0
2 1
5.0
1.0
1.3
1.3
95% CI
-
(1.1-1.6)
(1.4-2.1)
-
(0.6-2.2)
(0.9-3.7)
(0 6-3.6)
(1.7-8.4)
_
(1.1-1.5)
(1.0-1.6)
Predominately coeliac and Crohn's disease.
underweight and 10.8 in the normal weight individuals.
Scoliosis was also more prevalent among the severely
underweight (2.2/1000, borderline significance), as
well as chronic intestinal conditions (2.2), mainly
coeliac and Crohn's disease.
Emotional disorders, particularly neuroses, were also
associated with underweight. A prevalence rate of 14/
1000 among the mild and severely underweight subjects
was found, compared to 10.9 among normal weight
recruits.
DISCUSSION
This cross-sectional population-based study shows an
association between current morbidity and over- or
underweight status among male adolescents, and
identifies the different health conditions contributing to
excess morbidity at each extreme of the BMI scale. The
prevalence of pathology among adolescents associated
with the entire spectrum of weight has not been
previously available.
The BMI has been accepted as the most satisfactory
scale of obesity based on weight and height.26'27 It is
highly correlated with weight and independent of height.
However, it cannot distinguish between the contribution to body weight of fat tissue and that of muscle,
bone and water. This may introduce bias and cause
underestimation of the risk ratios. However, prediction
of ponderosity in middle age males from BMI early in
life has been shown to be reliable.28
Five regional medical draft boards carried out medical examinations of all recruits including evaluation of
weight and height. Inter-draft board differences in
techniques and results were negligible, and eliminates a
possible bias.
In this study, overweight in 17 year old males was
significantly associated with intercurrent hypertension
and with lower limb joint disorders. Measurement
artifacts, such as the use of regular cuffs for obese individuals, might artificially raise blood pressure levels
and may have been present. However, ergometric measurement was part of the basic hypertension evaluation,
thus reducing the chance of false positive results.
The direction of the association between overweight
and joint conditions is unclear. Severely overweight
individuals may have an excess burden on joints, thereby worsening existing problems. Conversely, people
with painful joint conditions may decrease their
physical activity, and hence gain weight.
In this study underweight was found to be associated
with excess bronchial conditions, severe scoliosis
(borderline significance), neurosis and intestinal malabsorptive disorders. The association between underweight and intestinal disorders found in this study may
be a consequence rather than an aetiological effect.
The higher prevalence of morbidity at extreme weight
categories at age 17 shown in this study strengthens the
importance of weight and BMI as a readily available
marker of morbidity. The identification of extreme
values of weight at a young age may enable targeting
individuals with potentially severe morbidity which
can be successfully prevented, or may ameliorate the
effects of current morbidity. Preventive health programmes are often initiated in adulthood, possibly
too late in life to reduce the onset of chronic disease.
The accumulation of additional risk factors as age
increases causes difficulties in preventing the onset of
disease. An early marker for morbidity may lead to
improved effectiveness of primary as well as secondary
prevention.
MORBIDITY AND EXTREME BM1
A number of caveats: In contrast to a standard
morbidity survey with research definitions, this study is
a standardized mass clinical examination designed to
identify combat fitness among military recruits, as well
as to identify conditions present prior to recruitment
which may be adversely affected by field conditions,
and thus prevent inappropriate placement during military service. Several of the reported diseases are based
on the individual report of a specific problem, and additional examinations are conducted for confirmation.
Underestimation of morbidity may therefore occur
among those individuals who fail to report disease
(false negatives). However, most conditions are evaluated for each individual. The prevalence reported in this
study is based on the medical conditions reported and
on judgement of the severity of this condition. In this
case a health problem could be considered more serious
if the individual was obese as well, which will lead
to overestimation of the prevalence of the health conditions. However, health assessment was based on a
standard and systematic clinical examination of each
recruit, independent of all anamnestic and anthropometric information. Severity levels for all conditions
are pre-defined in a standard code book and the single
most severe condition determines the overall level of
physical fitness and whether the recruit is qualified for
combat or for military service at all. In addition, diagnoses are based on clinical examinations, independent of the desire of the inductee to serve in the army.
It should be noted that in Israel reluctance to serve
in the army is relatively rare, and attempts to simulate morbidity, and thus elude service, are generally
identified.
In order to negate the influence of ethnicity on the
relationship between BMI and health conditions, each
recruit was categorized according to his specific ethnic
group BMI distribution. The association between BMI
and health conditions held true within each ethnic
group.
In conclusion, the strength of this study is that results
are based on nationwide cohorts systematically examined, with anthropometric and medical information
available, rather than a selected population of 17 year
old males. The association between morbidity and overand underweight at age 17 suggests that intervention
programmes developed to prevent or reduce morbidity
should be targeted to adolescents and young adults.
REFERENCES
1
2
Recommended Dietary Allowances, 10th Edn. Washington DC:
National Academy Press, 1989.
Dustan H P. Obesity and hypertension. Ann Intern Med 1985;
103: 1047-^9.
3
833
Van Itallie T B. Health implications of overweight and
obesity in the United States. Ann Intern Med 1985; 103:
983-88.
4
Glueck C J, Taylor H L, Jacobs D, Morrisson J A, Beaglehole R,
Williams O D. Plasma high-density lipoprotein cholesterol:
association with measurements of body mass. The Lipid
Research Clinics Program Prevalence Study. Circulation
1980; 62: IV-62-69.
5
Davis M A, Neuhaus J M, Ettinger W H, Mueller W H. Body fat
distribution and osteoarthntis. Am J Epidemiol 1990; 132:
701-07.
6
Guo S S, Roche A F, Cameron Chumlea W, Gardner J D,
Siervogel R M. The predictive value of childhood body
mass index values for overweight at age 35 y. Am J Clin
Nutr 1994, 59: 810-19.
7
Hoffmans M D A F, 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.
8
Hoffmans M D A F, Kromhout D, de Lezenne Coulander C.
Body mass index at the age of 18 and its effects on 32-yearmortality from coronary heart disease and cancer. J Clin
Epidemiol 1989, 42: 513-20.
9
Must A, Jacques P F, Dallal G E, Bajema C J, Dietz W H. Longterm morbidity and mortality of overweight adolescents.
A follow-up of the Harvard growth study of 1922-1935.
N EnglJ Med 1992; 327: 1350-55.
10
Mossberg H O. 40-year follow-up of overweight children. The
Lancet 1989; I: 491-93.
" Lew E A, Garfinkel L. Variations in mortality by weight
among 750,000 men and women. J Chron Dis 1979; 32:
563-76.
12
Rabkin S W, Mathcwson F A L, Hsu P H. Relation of body
weight to development of ischemic heart disease in a cohort of young North American men after a 26-year observation period: The Manitoba Study. Am J Cardiol 1977;
39: 452-58.
13
Hubert H B, Feinleib M, McNamara P M, Castelli W P. Obesity
as an independent risk factor for cardiovascular disease: A
26-year follow-up of participants in the Framingham heart
study. Circulation 1983; 67: 968-77.
14
Gilbert T J, Percy C A, Sugarman J R, Benson L, Percy C.
Obesity among Navajo adolescents: relationship to dietary
intake and blood pressure Am J Dis Child 1992; 146: 28995.
13
Gutin B, Basch C, Shea S el al. Blood pressure, fitness, and
fatness in 5- and 6-year-old children. JAMA 1990; 264:
1123-27.
16
Williams D P, Going S B, Lohman T C d a / Body fatness and
risk for elevated blood pressure, total cholesterol, and
serum lipoprotein ratios in children and adolescents. Am J
Public Health 1992; 82: 358-63.
17
Smoak C G, Burke G L, Webber L S, Harsha D W, Srinivasan
S R, Berenson G S. Relation of obesity to clustering of
cardiovascular disease risk factors in children and young
adults: The Bogalusa Heart Study. Am J Epidemiol 1987;
125: 364-72.
" Resnicow K, Morabia A. The relation between body mass index
and plasma total cholesterol in multiracial sample of US
schoolchildren. Am J Epidemiol 1990; 132: 1083-90.
" Dietz W H, Gross W L , Kirkpatrick J A. Blount disease (tibia
vara): another skeletal disorder associated with childhood
obesity. J Pedialr 1982; 101: 735-37.
834
20
21
22
23
24
INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
Kelsey J L, Acheson R M, Keggi K J. The body build of patients
with slipped capital femoral epiphysis. Am J Dis Child
1972; 124:276-81.
Shohat M, Shohat T, Kedem R, Mimouni M, Danon Y L. Childhood asthma and growth outcome. Arch Dis Childhood
1987; 62: 63-65.
Deschamps I, Giron B J, Lestradet H. Blood glucose, insulin,
and free fatty acid levels during oral glucose tolerance tests
in 158 obese children. Diabetes 1977; 26: 8 9 - 9 3 .
Booth I W . T h e Nutritional Consequences of Gastrointestinal
Disease in Adolescence. Ada Paedialr Scand 1991; 373
(Suppl. 1): 91-102.
Kraemer R, Rudeberg A, Hadora B, Rossi E. Relative underweight in cystic fibrosis and its prognostic value. Acta
Paedialr Scand 1978; 67: 33-37.
APPENDIX
Deflnitions of Medical Conditions used in this Study
Hypertension.
Measurement of diastolic blood pressure >95 mmHg or systolic pressure > 150 mmHg, not
receiving treatment and with a hypertensive response
to ergometry, or hypertension requiring continuous
medical treatment.
Knee. Radiological changes and objective physical
findings following trauma or surgery to the knee.
Hip joint. Joint changes with limitation of movement
in all directions, osteoarthrosis, or following surgery of
hip with x-ray changes.
Meniscus. Tear with recurrent clinical symptoms or
tear of ligaments affecting joint stability.
Ankle.
Following ankle trauma or surgery with
movement limitation, recurrent sprain, or arthrodesis
of the ankle.
25
Fliess L J. Statistical Methods for Rates and Proportions. 2nd
Edn. New York, John Wiley & Sons Inc., 1981, p. 6 3 .
26
Gray D S, Fujiaka K. Use of relative weight and body mass
index for the determination of adiposity. J Clin Epidemiol
1991; 44: 545-50.
27
Roche A F, Siervogel R M, Chumlea W C, Webb P. Grading
body fatness from limited anthropometric data. Am J Clin
Nutr 1981; 34: 2831-38.
21
Casey V A, Dwyer J T, Coleman K A, Valadian I. Body mass
index from childhood to middle age: a 50-y follow-up.
Am J Clin Nutr 1992; 14: 14-18.
(Revised version received January 1996)
Scoliosis. Only if severe or associated with frequent
back pain or respiratory difficulties.
Asthma.
Bronchitis.
sema.
Chronic asthma requiring continuous therapy.
Chronic bronchitis with signs of emphy-
Neuroses. Including phobia, and obsessive reactive
depression.
Overall significant functional limitation. Combination of any medical condition (including all others not
mentioned above) which precluded service in a combat
unit. The overall Israeli Defense Force fitness profile is
set at the most severe score assigned for any condition.