is a dynamic statement of the general health of that child. Meas-
urements should be performed often and accurately to detect
alterations from physiologic growth. Although any single point
on the growth chart is not very informative, when several growth
points are plotted over time, it should become apparent whether
that individual’s growth is average, a variant of the norm, or
pathologic. Somatic growth and maturation are influenced by
several factors that act independently or in concert to modify an
individual’s genetic growth potential. Linear growth within the
first 2 y of life generally decelerates but then remains relatively
constant throughout childhood until the onset of the pubertal
growth spurt. Because of the wide variation among individuals in
the timing of the pubertal growth spurt, there is a wide range of
physiologic variations in normal growth. Nutritional status and
heavy exercise training are only 2 of the major influences on the
linear growth of children. In the United States, nutritional
deficits result from self-induced restriction of energy intake.
That single factor, added to the marked energy expenditure of
training and competition for some sports, and in concert with
the self-selection of certain body types, makes it difficult to
identify the individual factors responsible for the slow linear
growth of some adolescent athletes, for example, those who par-
take in gymnastics, dance, or wrestling.
The longitudinal growth of an individual child
Am J Clin Nutr
potential, athlete, nutrition, gymnastics, wrestling, nutritional
dwarfism, athletic training
Growth, adolescent development, growth
A child’s growth can be compared with that of his or her peers
by referring to the norm on an appropriate growth chart. More
important, the longitudinal measurements of a child’s growth are
a dynamic statement of his or her general condition or health.
Tanner (1) has proposed that children be measured accurately
to identify individuals or groups of individuals within a commu-
nity who require special care, to identify illnesses that influence
growth, or to determine an ill child’s response to therapy. The
linear growth of a child-adolescent athlete may also reflect the ade-
quacy of energy intake for a particular training regimen. Meas-
urement of growth may also be used as an index of the general
health and nutrition of a population or subpopulation of children.
By definition, normal (physiologic) growth encompasses the
95% CI for a specific population. Most children and adoles-
cents who have a normal growth pattern but who remain below
the lower 2.5 percentile (approximately ?2.0 SD) are other-
wise normal. The farther an individual’s growth falls below the
?2.0 SD mark, the more likely he or she is to have a condition
that is keeping him or her from reaching the genetically deter-
mined height potential.
Cross-sectional data are derived from the measurements of
many children at various ages and are generally used to derive
standard growth charts. However, individual children do not
necessarily grow according to these standard curves. Longitu-
dinal growth charts derived from growth points of the same
child over time more accurately describe the growth pattern of
an individual. In adolescence, there may be quite large devia-
tions from the derived percentile lines, depending on the timing
and tempo of the pubertal growth spurt. An average pubertal
growth pattern is built into the percentiles derived from cross-
sectional data, but virtually no one individual adheres strictly
to that pattern.
The growth of children should be measured periodically and
accurately. Two common devices are adequate for such measure-
ments and were described by Rogol and Lawton (2).
Neonates and infants
Small inaccuracies in length measurement can easily affect a
child’s percentiles on growth curve charts. Infants should be
placed with the top of the head against the fixed headboard of the
measurement device and with the eye-ear plane perpendicular to
the base of the device (Figure 1). The child’s knees must be flat
against the table and the footboard moved until the soles of the
feet are against it, with the toes pointing up.
Am J Clin Nutr 2000;72(suppl):521S–8S. Printed in USA. © 2000 American Society for Clinical Nutrition
Growth and pubertal development in children and adolescents:
effects of diet and physical activity1–4
Alan D Rogol, Pamela A Clark, and James N Roemmich
1From the University of Virginia Health Sciences Center Charlottesville.
2Presented at the workshop Role of Dietary Supplements for Physically
Active People, held in Bethesda, MD, June 3–4, 1996.
3Supported in part by grant HD32631 from the National Institutes of Health.
4Address reprint requests to A Rogol, Insmed Incorporated, 800 East
Leigh Street, Richmond, VA 23219 . E-mail: firstname.lastname@example.org.
by on March 5, 2008
Children and adolescents
Growth in children older than 2 y is measured with the child
standing. A diurnal variation of <0.7 cm in height may occur in
these children, whose height is greatest upon arising. Children
should be measured without shoes while standing against the
vertical plane to which the measuring tape is attached. The
child’s heels, buttocks, shoulders, and back of the head should be
touching the wall (Figure 2). The eye-ear plane should be per-
pendicular to the wall and the feet, including the heels, should be
flat on the floor. With the child in this position, the right-angle
device is lowered until it touches the top of the head, and the
height is recorded on the appropriate physical growth curve.
An important parameter of growth is height velocity, which
can be derived from measurements taken every 3–4 mo in infants
and every 6 mo in older children. Because children often have
growth spurts, yearly growth velocities are usually more accu-
rately determined by taking yearly measurements rather than by
“annualizing” the growth velocity from intervals shorter than 1 y.
Growth velocity in children has a wide normal range, accord-
ing to the percentile along which a child is growing. Children
growing along the third percentile average 5.1 cm of growth per
year, whereas those growing along the 97th percentile average
6.4 cm/y for boys and 7.1 cm/y for girls during childhood to
maintain growth along one of the percentiles on the growth curve
(3). To maintain growth along the 10th percentile for height, a
child must grow at the 40th percentile for velocity, whereas to
maintain growth along the 90th percentile for height, a velocity
at the 60th percentile is required. This implies that a child who
persistently grows at the 10th percentile for velocity will pro-
gressively cross percentiles downward on the standard height
curve. Some children have a small increase in growth velocity at
approximately 6–7 y (midgrowth spurt), but this is not a consis-
tent finding, and the gain in height is generally of small magni-
tude (4). Seasonal variations in growth have been noted in some
children. Linear growth tends to be greater in the spring than in
the fall, but weight gain is greater in the fall months. These
trends emphasize the need for repeated measurements during a
year to accurately assess a child’s growth pattern.
Linear growth and physical maturation are dynamic processes
encompassing molecular, cellular, somatic, and organismal
changes. Traditionally, stature has been primarily used for growth
assessment, but changes in body proportion and composition are
also essential elements of growth, especially of maturation.
Growth standards have been derived for several populations and
parameters within a population and are often codified into a
series of growth charts. The following discussion emphasizes the
genetic, nutritional, hormonal, and physical activity factors that
might alter the growth process.
PHYSIOLOGIC GROWTH PATTERNS
Although any single point on a growth chart is not very
informative, when several growth points are plotted over time,
it becomes apparent whether an individual’s growth is average,
a variant of the norm, or pathologic (growth failure). The point
at which an individual is placed at any given time can be
related to the height age, or the age at which that child’s height
would be at the 50th percentile. This point indicates the mean
age of children of that measured height in the normal popula-
tion. The height age is determined from the growth chart by
drawing a line parallel to the chronologic age axis from the
child’s plotted point to the 50th percentile and then a perpen-
dicular line to the horizontal axis. The intersection of the latter
line with the age axis is the height age.
Growth in several dimensions shows a significant family
resemblance. Adult stature, tempo of growth, timing and rate of
sexual development, skeletal maturation, and dental develop-
ment are all significantly influenced by genetic factors (5) and
estimates of genetic transmissibility range from 41% to 71% (4).
Adult stature is best correlated with calculations of midparental
height (the difference in the mean adult heights of the parents),
but the polygenic mode of inheritance of height results in greater
variation in the size of children born to parents of disparate
heights than in children of parents who are both of medium
height (6). Mature height can be predicted on the basis of mid-
parental height. The adjusted midparental height (target height)
is calculated by adding 13 cm (the difference between the 50th
percentiles for adult men and women) to the mother’s height (for
boys) or subtracting 13 cm from the father’s height (for girls)
522SROGOL ET AL
FIGURE 1. Growth measurement length of an infant. Reprinted with
permission from reference 2.
FIGURE 2. Growth measurements: height of a child or adolescent.
Reprinted with permission from reference 2.
by on March 5, 2008
and then taking the mean of the height of the same-sex parent
and the adjusted height of the opposite-sex parent. Adding
8.5 cm above and below the midparent target height will approx-
imate the target height range of the 3rd to 97th percentile for the
anticipated adult height for that child adjusted for his or her mid-
parental stature (genetic potential). Other methods also exist for
predicting the adult stature of an individual on the basis of math-
ematical formulations derived from the growth history of that
child or from the attained height and bone age of the child as cal-
culated from specific tables.
The overall contribution of heredity to adult size and shape
varies with environmental circumstances, and the 2 continuously
interact throughout the entire growth period. Children with sim-
ilar genotypes, who would reach the same adult height under
optimal conditions, may be differentially affected by adverse cir-
cumstances. Thus, the interaction between genetic makeup and
the environment is complex and nonadditive. The genetic control
of the tempo of growth appears to be independent of that for
body size and shape, and environmentally induced changes in
tempo do not seem to significantly alter adult height or shape (4).
FACTORS INFLUENCING SOMATIC GROWTH
Somatic growth and maturation are influenced by several fac-
tors that act independently or in concert to modify an individ-
ual’s genetic potential. For example, at birth, an infant’s size is
more dependent on maternal nutrition and intrauterine and pla-
cental factors than on genetic makeup. The correlation coeffi-
cient for adult height is only 0.25 at birth but is 0.80 by 2 y of
age (7). There is also evidence that not all genes are actively
expressed at the time of birth, which probably accounts for the
observation that the correlation between the sizes of the parents
and the child is weak during the first year of life but increases to
the adult value of 0.5 by ?18 mo of age (4).
Differences in growth and development also vary as a function
of sex and ethnic origin. Sex-specific patterns in the tempo of
growth, the timing of the adolescent growth spurt, overall size,
and the age of skeletal maturity are well known, but differences
between the sexes are apparent from the time of fetal life. At
birth, the skeletal maturation of females is 4–6 wk more
advanced than that of males, and this trend continues throughout
childhood and adolescence. Growth velocity is slightly slower in
females at birth, becomes equal at ?7 mo of age, and is then
somewhat faster until age 4 y. Thereafter, children of both sexes
grow at approximately the same rate until the adolescent growth
spurt. On average, females enter puberty 2 y earlier than males
but have a lesser peak height velocity (9 cm compared with
10.3 cm) and adult stature (8, 9). Overall size and rate of devel-
opment vary significantly among ethnic populations. Black
infants tend to be smaller at birth but experience an acceleration
of linear growth that results in greater height than in white chil-
dren during the first few years of life. Skeletal maturity in black
children, especially girls, also tends to be more advanced and the
age at peak height velocity earlier (10, 11). Black girls also tend
to be taller and heavier than white girls during puberty and have
a tendency toward greater body mass index and greater skinfold-
Growth in the first 2 y of life
Growth during the first 2 y of life is characterized by a grad-
ual deceleration in both linear growth velocity and rate of weight
gain, both of which level off at 2–3 y of age. It is during this
period that infants exhibit the pattern of growth consistent with
their genetic backgrounds. Two-thirds of all infants cross per-
centiles on the growth curve, either upward (catch-up growth) or
downward (lag-down growth) (6). Catch-up growth typically
begins within the first 3 mo and is complete by 12–18 mo,
whereas lag-down growth commences a little later and may not
be complete until 18–24 mo (6). With the exception of puberty,
the crossing of growth percentiles at any other time is cause for
concern and further evaluation.
Growth during childhood is a relatively stable process. The
infancy shifts in the growth pattern are complete and the child fol-
lows the trajectory attained previously. Until about the age of 4 y,
girls grow slightly faster than boys and both sexes then average a
rate of 5–6 cm/y and 2.5 kg/y until the onset of puberty (4). A gen-
eral rule of thumb is that a child grows 10 cm (25 inches) in the
first year of life, half that [12–13 cm (5 inches)] in the second year,
and then 5–6 cm (2.5 inches) each year until puberty. Assuming an
average birth length of 51 cm (20 inches), an average 1-y-old is
76 cm (30 inches) long, a 2 y-old is 89 cm (35 inches), a 4-y-old
is 102 cm (40 inches), and an 8-y-old is 127 cm (50 inches).
Puberty is a dynamic period of development marked by rapid
changes in body size, shape, and composition, all of which are
sexually dimorphic. The onset of puberty corresponds to a skele-
tal (biological) age of ?11 y in girls and 13 y in boys (12). On
average, girls enter and complete each stage of puberty earlier
than do boys. The timing and tempo of puberty vary widely, even
among healthy children. In determining the appropriateness of a
particular growth velocity, the child’s degree of biological matu-
ration must be considered. Skeletal or pubertal maturation may
be used to determine the child’s degree of biological develop-
ment. The bone age is determined as the mean of the skeletal
ages of several of the small bones of the hand and wrist. Puber-
tal maturation status is based on the development of breasts and
pubic hair in girls and of pubic hair and genitals in boys. This
range of normal variability is expanded to an even greater degree
by alterations in energy intake and expenditure. Although mod-
erate activity is associated with cardiovascular benefits and
favorable changes in body composition, excessive physical activ-
ity during childhood and adolescence may negatively affect
growth and adolescent development. Sports that emphasize strict
weight control and high energy output—for example, scholastic
wrestling, gymnastics, and dancing—are of particular concern
for growth disorders, although selection criteria for certain body
types make selection bias a confounding variable in assessing the
effect of training on growth and adolescent development. One
must consider that some of these change are transient, at least in
wrestlers. The same markers of growth and body composition
that are slowed during training (in season) accelerate after the
season, which permits a catch-up process to control growth and
cause no permanent growth reductions (see the section “Consti-
tutional delay of growth,” below).
One of the hallmarks of puberty is the adolescent growth
spurt. As puberty approaches, growth velocity slows to a nadir
(“preadolescent dip”) before its sudden acceleration during mid-
puberty. The timing of the pubertal growth spurt in girls is typi-
cally at Tanner breast stage 3 and does not reach the magnitude
GROWTH OF CHILDREN: DIET AND ACTIVITY523S
by on March 5, 2008
of that in boys. Girls average a peak height velocity of 9 cm/y at
age 12 and a total gain in height of 25 cm during the pubertal
growth period (13). Boys, on average, attain a peak height veloc-
ity of 10.3 cm/y 2 y later than girls, during Tanner genital stage 4,
and gain 28 cm in height (9, 13). The longer duration of prepu-
bertal growth in boys, combined with a greater peak height
velocity, results in an average adult height difference of 13 cm
between men and women. After a period of decelerating height
velocity, growth virtually ceases because of epiphyseal fusion,
typically at a skeletal age of 15 y in girls and 17 y in boys (4).
Puberty is also a time of significant weight gain; 50% of adult
body weight is gained during adolescence. In boys, peak weight
velocity occurs at about the same time as peak height velocity
and averages 9 kg/y. In girls, peak weight gain lags behind peak
height velocity by ?6 mo and reaches 8.3 kg/y at ≈12.5 y of age
(4). The rate of weight gain decelerates in a manner similar to
height velocity during the later stages of pubertal development.
Marked changes in body composition, including alterations in
the relative proportions of water, muscle, fat, and bone, are a
hallmark of pubertal maturation and result in typical female-
male differences. Under the influence of the gonadal steroid hor-
mones and growth hormone (GH), increases in bone mineral
content and muscle mass occur and the deposition of fat is max-
imally sexually dimorphic. The changes in the distribution of
body fat (central compared with peripheral, subcutaneous com-
pared with visceral, and upper compared with lower body)
results in the typical android and gynoid patterns of fat distribu-
tion of the older adolescent and adult (14).
Under the influence of testosterone, boys have a significant
increase in growth of bone and muscle and a simultaneous loss
of fat in the limbs (4). The maximal loss of fat and increase in
muscle mass in the upper arms corresponds to the time of peak
height velocity. In boys, the significant increase in lean body
mass exceeds the total gain in weight because of the concomitant
loss of adipose tissue. As height velocity declines, fat accumula-
tion resumes in both sexes but is twice as rapid in girls. As
adults, males have 150% of the lean body mass of the average
female and twice the number of muscle cells (15). The increase
in skeletal size and muscle mass leads to increased strength in
males. Both androgens and estrogens promote deposition of
bone mineral, and >90% of peak skeletal mass is present by age
18 y in adolescents who have undergone normal pubertal devel-
opment at the usual time. In girls, nearly one-third of total skele-
tal mineral is accumulated in the 3–4-y period immediately after
the onset of puberty (16, 17). Adolescents with delayed puberty
or secondary amenorrhea may fail to accrue bone mineral nor-
mally and have reduced bone mineral density as adults.
During pubertal development, interactions between GH and
the sex steroid hormones are striking and pervasive. Studies of
adolescent boys showed that the rising concentrations of testos-
terone during puberty play a pivotal role in augmenting sponta-
neous secretion of GH and production of insulin-like growth
factor I (IGF-I). The ability of testosterone to stimulate pituitary
GH secretion, however, appears to be transient and expressed
only peripubertally; GH and IGF-I concentrations decrease
significantly during late puberty and into adulthood, despite con-
tinued high concentrations of gonadal steroid hormones (18). In
contrast with testosterone, estrogen modulates GH secretory
activity in a disparate manner; low doses of estrogen stimulate
IGF-I production through enhanced GH secretion, but higher
doses inhibit IGF-I production at the hepatic level (19).
VARIATIONS OF NORMAL GROWTH
Normal variants of growth were found in 82% of children whose
height decreased at the third percentile (?2 SD) but in only 50% of
those whose height decreased at the first percentile (?3 SD) of the
mean for age (20). Assessment of skeletal maturation is perhaps the
best indicator of biological age or maturity status, because its devel-
opment spans the entire period of growth. Several methods exist for
determining the former (21–23). Each uses a single radiograph of
the left hand and wrist and makes comparisons with children of nor-
mal stature by using an atlas and scoring system. Because girls are
more developmentally mature than boys at any given chronologic
age, separate standards exist for females and males.
Familial short stature
On average, children of smaller parents will eventually attain
lesser height than children of taller parents. Because bone age
approximates chronologic age, these children usually grow at an
appropriate rate during childhood and attain sexual maturation and
pubertal growth spurt at the usual ages.
Constitutional delay of growth
A constitutional growth delay is considered to be a delay in
the tempo of growth. In this case, each calendar year is not
accompanied by a full year of growth and skeletal development,
so the individual requires more time to complete the growth
process. Most of these children will eventually have delayed
adolescence as well as delayed attainment of adult stature. Birth
history and birth length are generally normal, but the growth pat-
tern shifts downward to the lower percentiles, so that the lowest
values for growth velocity are obtained at ?3–5 y of age. There-
after, this pattern is characterized by steady progression of
growth. Because the bone age does not advance 1 y for each cal-
endar year, it progressively deviates from the chronologic age.
The height age is usually approximately the same as the bone
age, and if true, the mature height will be well within the normal
range for the appropriate population. Like familial short stature,
this pattern is often familial, and because both are relatively
common, some children will have elements of both.
NUTRITION AND GROWTH
Worldwide, the single most common cause of growth retarda-
tion is poverty-related malnutrition. In the United States, nutri-
tional growth retardation (also known as nutritional dwarfism)
and delayed pubertal development among suburban upper-
middle- and upper-class adolescents more often result from self-
induced restriction of nutrient (energy) intake. In addition to
effects on overall growth, malnutrition secondary to avoidance
of certain foods or malabsorption can lead to serious disorders,
such as osteopenia, anemia, and syndromes related to deficien-
cies of vitamins, minerals, essential fatty acids and amino acids,
and trace elements. Nutritional status also has a significant mod-
ulating effect on the timing of adolescent sexual development.
Undernutrition is associated with later age at menarche (as well
as secondary amenorrhea), whereas a moderate degree of obesity
is associated with early sexual maturation (24, 25). The growth
curves for length and weight may at first be indistinguishable
from those of children and adolescents with constitutional delay
of growth (see “Constitutional delay of growth,” above) or, more
rarely, from those of children with familial short stature.
524SROGOL ET AL
by on March 5, 2008
The diagnostic criteria for nutritional growth failure follow
those of the Wellcome Trust classification. The weight for
chronologic age is low, although there may often be a minimal
deficit in weight-for-height, as occurs in constitutional delay of
growth and adolescence or even familial short stature. Although
the specific behaviors required for the diagnosis of anorexia ner-
vosa or bulimia nervosa are absent, there is deteriorating linear
growth or delay in adolescent development associated with inad-
equate weight gain. It appears that a preoccupation with slimness
and striving for weight control, fueled by current health beliefs,
cause the retarded growth. Single growth points of underweight-
for-height are not nearly as important as longitudinal data,
because individuals with constitutional thinness may have a
weight that is >2 percentile lines below their height.
Nutritional growth retardation must be differentiated from the
variations of normal growth noted above and also from some of
the forms of inflammatory and other bowel diseases in which
growth failure, often noted by deviation from a previously
defined length and weight channel, may be well above the low-
est percentiles. Rehabilitation of nutritional growth failure or
relief of the inflammation may promote catch-up growth.
The theoretical weight, the weight deficit for that theoretical
weight, and the weight-for-height deficit should be defined (Fig-
ure 3). This is because children with familial short stature and mild
constitutional delay of growth and adolescence most often continue
to gain weight. This weight gain is either along an established per-
centile or slightly below but parallel to the lowest percentile on the
chart. The hallmark of nutritional growth delay is that weight pro-
gressively deviates from the previous channel, an observation that
underscores the importance of gathering longitudinal data.
Children with nutritional growth retardation may have reached
a new energy equilibrium phase between their genetically deter-
mined growth potential and the present energy intake, because
growth deceleration is the adaptive response to suboptimal energy
intake. This growth deceleration has limits; for example, energy
intake (and often protein) may be inadequate for such a pro-
longed period of time that energy malnutrition becomes evident.
Acutely, suboptimal intake due to illness or heavy exercise load
(see the following section) may temporarily delay growth, but
this will be followed quickly by catch-up growth. This process
must be properly distinguished from other causes of organic and
nonorganic growth retardation.
The adaptive response is marked by a decrement in basal (and
resting) metabolic rate and a decrease in protein synthesis, the
latter being an energy-intensive process. In addition, there may
be deficiencies in minerals, particularly zinc and iron, and vita-
mins. All may lead to a decrease in physical activity, which is an
attempt to decrease ongoing energy losses.
EFFECTS OF PHYSICAL ACTIVITY AND TRAINING ON
GROWTH AND ADOLESCENT DEVELOPMENT
Does physical activity, sport training, or both affect linear
growth and pubertal maturation? The literature is replete with
reports that the effects of athletic training on growth and puber-
tal development are salutary, deleterious, or nonexistent [for a
review, see Malina (27)]. However, careful appraisal of these
reports often reveals severe methodologic faults, such as lack of
consideration of interindividual variation in biological maturity
status and subject selection. Certain sports show advantages for
GROWTH OF CHILDREN: DIET AND ACTIVITY 525S
FIGURE 3. Growth pattern of nutritional dwarfing (A and B) compared with constitutional growth delay (C). A: Patient in whom body weight gain
and height progression decreased after 10 y of age. Extrapolated weight after age 14 y revealed a body weight deficit based on the previous growth per-
centile. However, there was no body weight deficit for height; with nutritional rehabilitation, there was recovery in weight gain and catch-up growth. B:
Patient with body weight deficit for height but a more marked deficit for theoretical weight. C: Patient without nutritional dwarfing. This patient, with
constitutional growth delay, showed a body weight gain consistently along the lower percentile, with no deviation in growth. Note that there was no body
weight deficit for height or for theoretical weight based on previous growth. Reprinted from reference 26, page 105, by courtesy of Marcel Dekker, Inc.
by on March 5, 2008
the early maturer, especially for males, and others, especially
gymnastics and dance, favor the later-developing female. Thus,
there is concern about the potential effects of training on the tim-
ing and progression (tempo) of puberty “caused” by participa-
tion in training and sports. Critical analysis with the biological
indicators of bone age or peak height velocity in longitudinal
study designs is required to tease out the effects of such training
on pubertal development and adult height.
Delay in growth and sexual maturation is well documented
among certain groups of elite female athletes, most notably gym-
nasts, dancers, and long-distance runners (28). The underlying
mechanisms, however, are not entirely clear, in part because of
few longitudinal data in girls. Control of growth and age at
menarche involve the complex interaction of many factors,
including the physical and metabolic demands of intensive ath-
letic training and competition.
Investigations of growth parameters in adolescent female
gymnasts consistently find these girls to be shorter and lighter
and to have a significantly lower percentage of body fat than do
age-matched control girls or athletes participating in less strenu-
ous sports, such as swimming. Girls participating in the latter
types of sports are generally taller and mature earlier than nor-
mal (28–31). Theintz et al (29) followed a cohort of adolescent
gymnasts and swimmers over an interval of 2–3 y. Training peri-
ods averaged 22 h/wk for the gymnasts and 8 h/wk for the swim-
mers. The gymnasts had significantly lower growth velocities
from skeletal age 11–13 y, showing a peak height velocity of
only 5.48 ± 0.32 cm/y compared with 8.0 ± 0.50 cm/y for the
swimmers. Over time, height SD scores decreased significantly
in the gymnasts without a change in the ratio of chronologic age
to bone age. Consequently, predicted heights of the gymnasts
decreased with time, but those of the swimmers did not change.
Lindholm et al (32) also observed slower growth velocities
among a group of adolescent female gymnasts. These girls did not
display the distinct growth spurt seen in the control group of inac-
tive girls, and 27% had adult heights that were less than expected
based on midparental height. Bernadot and Czerwinski (33) stud-
ied 2 groups of female gymnasts, one aged 7–10 y and the other
aged 11–14 y. Weight-for-age and height-for-age decreased from
the 48th percentile in the younger group to the 20th percentile in
the older gymnasts. Body fat did not differ significantly between
the age groups, and at all ages the gymnasts had significantly more
muscle mass for their size than did the control group.
Several investigations have compared age at menarche among
female athletes participating in different sports with that of the
general population. Claessens et al (34) found the median age at
menarche to be 15.6 ± 2.1 y among a group of gymnasts and
13.2 ± 1.2 y among the control population. Theintz et al (29)
observed that among a group of gymnasts and swimmers aged
12.7 ± 1.1 y, only 7.4% of the gymnasts had experienced menar-
che, in contrast with 50% of the age-matched swimmers. The
gymnasts in this study, however, had a significant delay in skeletal
age (?1.42 ± 0.99 y), but the swimmers had comparable chrono-
logic and skeletal ages. This report emphasized the importance of
the interaction between somatic growth and sexual maturation and
the interpretation of physiologically versus pathologically delayed
puberty. Baxter-Jones et al (35) reported the mean ages at menar-
che of adolescents being intensively trained in gymnastics, swim-
ming, and tennis to be 14.3, 13.3, and 13.2 y, respectively, with a
population reference value of 13.0 y. Significant delay was again
noted only among the group of gymnasts. The data for gymnasts
are replicated to a lesser degree in dancers and runners. Sports
such as swimming, speed skating, and tennis appear to have min-
imal effects on growth or age at menarche (27, 28, 35).
Although these data suggest a relation between intense athletic
training and growth and pubertal development in female gymnasts,
they are not conclusive. In interpreting growth and development
data of athletes, a host of other variables, including the intensity of
training, must also be considered. An individual’s general state of
health is critical to normal growth and development, but this is
assumed in adolescents who meet the great physical demands of
long-term training. Genetic predisposition also plays an important
role; the short stature of gymnasts is often familial (28) and a
positive correlation has been found between menarcheal age in
mothers and daughters (35). Historically, socioeconomic class and
family size have been influential; menarche occurs earlier in the
higher socioeconomic classes and in families with fewer siblings
(4). Psychologic and emotional stressors associated with years-
long training, frequent competition, maintenance of low body
weight, altered peer relations, and demands of coaches may also
influence growth and pubertal timing (28).
Nutrition, especially dieting behavior, can be a major factor for
disordered growth, particularly in sports that emphasize strict
weight control. Although the principle of a critical percentage of
body fat is no longer considered valid, the issue of energy balance
is crucial to growth and development. Intake of energy, as well as
of vital nutrients such as calcium, which is necessary for bone
mineral accrual, may be suboptimal in athletes who restrict dietary
intake during a time of increased metabolic demand. Nutter (36)
found that the desire to be thin may influence dietary patterns of
female athletes even more than do changes in exercise training.
Several investigators, for example, Warren (37), have stressed
the importance of strenuous physical training before menarche
that might cause disordered growth and adolescent development.
Younger children may be especially susceptible to the energy
demands of strenuous exercise. Although similar trends depend-
ing on type of sport are apparent, menarche is more delayed in
gymnasts than in swimmers or tennis players who began training
at a comparable age. Prior menstrual irregularity appears to be an
important risk factor for oligomenorrhea or amenorrhea in ado-
lescents who begin training after menarche.
Alterations in growth and pubertal maturation are not com-
mon among young women engaging in recreational exercise or
in adolescents who train <15 h/wk (38). The incidence of
oligomenorrhea or amenorrhea and secondary amenorrhea has
been cited as 10–40% among athletes and 2–5% among the
general population. The distinction between elite and nonelite ath-
letes is important because it pertains to training time and inten-
sity. Olympic athletes have been shown to have significantly
later menarche than high school, college, and club-level athletes
(27). The different demands of various sports also dictate the
amount of time spent in strenuous physical activity; gymnasts
and dancers far exceed swimmers and tennis players in the avail-
able studies. Catch-up growth has been reported in gymnasts
when their training is temporarily reduced or stopped (30).
However, one of the most important variables (perhaps the
single most important variable) to take into account is that of
selection bias. Body types that are most successful are selected
for particular sports. Several studies have reported gymnasts to
be smaller than their peers from a young age (27, 28). Delayed
526S ROGOL ET AL
by on March 5, 2008
menarche favors the continuation of sports such as gymnastics,
which suggests that elite gymnasts are selected in part for this
attribute. Continued participation in turn leads to more intense
training and blurring of cause and effect.
The implications of delayed menarche are directly relevant to
the accrual of bone mineral. Because >90% of the total adult bone
mass is established during the pubertal years, failure to accrue
bone mineral at a normal rate during this time may result in per-
manent deficits. Bone mineralization is a complex process influ-
enced by nutrition (especially calcium intake), weight-bearing
activity, and sex steroid hormones. Hypoestrogenism because of
pubertal delay or secondary amenorrhea can lead to low bone
mineral density despite adequate weight-bearing exercise. In a
group of female runners, Louis et al (39) found decreases in bone
mineral density in all subjects with oligomenorrhea or amenor-
rhea, whereas runners with regular menses had values within the
normal range. A low rate of bone mineral accrual has been sug-
gested as one factor contributing to skeletal injuries in gymnasts.
In general, boys who participate in sports have normal growth
rates and are normal or advanced for their state of skeletal and
sexual maturation (27). The advanced states of maturation in
male athletes may be attributed to the power and performance
advantages associated with maturation (40).
However, for sports that may create an energy drain, the
effects on growth and maturation remain inconclusive. Seefeldt
et al (41) reported that the height velocity of elite male distance
runners was equal to nonrunning control subjects during 1 y of
training. Other investigations have reported the linear growth of
male distance runners to be either slowed or advanced relative to
reference data. Unfortunately, the maturity levels of the runners,
the reference data, or both were not given for the 2 former stud-
ies, so few conclusions can be made with regard to the influence
of distance running on growth velocity.
The growth of scholastic wrestlers has also been the concern of
several investigations. American wrestlers begin losing weight to
certify for lower competitive weight classes as young as 8 y of age.
The weight is lost through dieting, severe exercise, dehydration,
and various other methods (42, 43), which has produced enough
concern to warrant both the American College of Sports Medicine
and the American Medical Association to publish position state-
ments calling for the limitation of this practice. In fact, several
authors have speculated that the growth of peripubescent wrestlers
may be slowed during the sports season. As a group, high school
wrestlers are usually shorter than average for their age (27),
although this too is probably a self-selection process for wrestling.
In a cross-sectional study, the growth patterns of 477 high
school wrestlers were compared with those of a representative
sample of adolescent males (44). The wrestler and reference
groups were not different at any age for body weight, but the
slope value for the gain in body weight was significantly greater
for the reference sample. The reference group was significantly
taller than the wrestlers after age 16.4 y, but the slope values for
gain in height were not statistically different. Slope values were
also compared for 13 other anthropometric variables, with few
notable group differences. The investigators concluded that
wrestling does not slow growth and maturation (44). However,
the study did not address whether the growth rate during the
sport season was slowed and, if so, whether there was catch-up
growth during the nontraining season.
As expected, many investigators have reported reductions in
weight, fat mass, and percentage of body fat during the wrestling
season (45, 46). However, the fat-free mass is more conserved;
most investigators report nonsignificant reductions (45, 46).
Still, the fat-free mass does not increase as one would expect for
normal pubescent males. Because arm and leg strength diminish
(45), one might suggest that statistically insignificant reductions
in these variables may be biologically relevant. After the sport
season, wrestlers experience accelerated incremental gains in
weight, fat mass, and fat-free mass (45, 47). The postseason
gains in weight may be above the 99th percentile for age. Accel-
erated postseason gains in weight, fat mass, and fat-free mass
suggest soft-tissue catch-up growth in the wrestlers. During the
sport season, changes in anthropometric measures of lean tissue,
such as mid-arm girth and lean limb cross-sectional areas
(obtained from skinfold corrected girths), also provide evidence
that despite heavy bouts of training, wrestlers can fail to accrue
lean tissue during the sport season and show an accelerated
accrual postseason (45).
A few compelling data implicate training or competition as
causal in the shorter stature and decreased body mass of some
pubertal athletes in specific sports. It appears likely that activities
such as gymnastics and dance in girls or wrestling in boys select
for those participants with desirable genetic anthropometric traits.
Added to this process are the interactions among diminished nutri-
tion and the energy drain of training. Preliminary hormonal stud-
ies cannot distinguish between constitutionally delayed puberty
and a syndrome caused by sport participation. However, studies
designed to make this distinction probably cannot be done in ado-
lescents. Investigations in adult women show that some amenor-
rheic athletes have altered pulsatile gonadotropin release, but it
has not yet been possible to separate the effect of the training itself
from nutritional and stress factors (48).
1. Tanner JM. Auxology. In: Kappy MS, Blizzard RM, Migeon CJ,
eds. The diagnosis and treatment of endocrine disorders in child-
hood and adolescence. 4th ed. Springfield, IL: Charles C Thomas,
2. Rogol AD, Lawton EL. Body measurements. In: Lohr JA, ed. Pedi-
atric outpatient procedures. Philadelphia: JB Lippincott, 1990:1–9.
3. Baumgartner RN, Roche AF, Himes JH. Incremental growth tables:
supplementary to previously published charts. Am J Clin Nutr 1986;
4. Tanner JM. Fetus into man: physical growth from conception to
maturity. Cambridge, MA: Harvard University Press, 1989.
5. Sinclair D. Human growth after birth. London: Oxford University
6. Smith DW. Growth and its disorders. Philadelphia: WB Saunders,
7. Tanner JM, Healy MJR, Lockhart RD, et al. Aberdeen Growth Study,
I: the prediction of adult body measurement from measurements
taken each year from birth to five years. Arch Dis Child 1956;31:372.
8. Marshall WA, Tanner JM. Variations in patterns of pubertal changes
in girls. Arch Dis Child 1969;44:291–303.
9. Marshall WA, Tanner JM. Variations in patterns of pubertal changes
in boys. Arch Dis Child 1970;45:13–23.
10. Berkey CS, Wang X, Dockery DW, Ferris B. Adolescent height
growth of US children. Ann Hum Biol 1994;21:435–42.
GROWTH OF CHILDREN: DIET AND ACTIVITY527S
by on March 5, 2008
11. Malina RM. Skeletal maturation studied longitudinally over one Download full-text
year in American whites and Negroes six through thirteen years of
age. Hum Biol 1970;42:377–90.
12. Tanner JM, Whitehouse RH, Marshall WA, Carter BS. Prediction of
adult height, bone age, and occurrence of menarche, at age 4 to 16 with
allowance for midparental height. Arch Dis Child 1975;50:14–26.
13. Kelch RP, Beitins IZ. Adolescent sexual development. In: Kappy
MS, Blizzard RM, Migeon CJ, eds. The diagnosis and treatment of
endocrine disorders in childhood and adolescence. 4th ed. Spring-
field, IL: Charles C Thomas, 1994:193–234.
14. Johnston FE. Developmental aspects of fat patterning. In: Hernan-
dez M, Argente J, eds. Human growth: basic and clinical aspects.
Amsterdam: Elsevier, 1992:217–26.
15. Cheek DB. Body composition, hormones, nutrition and adolescent
growth. In: Grumbach MM, Grave GD, Mayer FE, eds. Control of
the onset of puberty. New York: John Wiley & Sons, 1974:424–47.
16. Bonjour J, Theintz G, Buchs B, Slosman D, Rizzoli R. Critical years
and stages of puberty for spinal and femoral bone mass accumula-
tion during adolescence. J Clin Endocrinol Metab 1991;73:555–63.
17. Slemenda CW, Reister TK, Hui SL, Miller JZ, Christian JC, John-
ston CC. Influence on skeletal mineralization in children and ado-
lescents: evidence for varying effects of sexual maturation and
physical activity. J Pediatr 1994;125:201–7.
18. Martha PM Jr, Rogol AD, Veldhuis JD, Kerrigan JR, Goodman DW,
Blizzard RM. Alterations in the pulsatile properties of circulating
growth hormone concentrations during puberty in boys. J Clin
Endocrinol Metab 1989;69:563–70.
19. Ho KY, Evans WS, Blizzard RM, et al. Effects of age and sex on the
24-hour profile of growth hormone secretion in man: importance of
endogenous estradiol concentrations. J Clin Endocrinol Metab 1987;
20. Lacy KA, Parkin JM. The normal short child: community study of
children in Newcastle-upon-Tyne. Arch Dis Child 1974;49:417–24.
21. Greulich WW, Pyle SI. Radiographic atlas of skeletal development
of the hand and wrist. 2nd ed. Stanford, CA: Stanford University
22. Roche AF, Chumlea WC,Thissen D. Assessing the skeletal maturity of
the hand-wrist: Fels method. Springfield, IL: Charles C Thomas, 1988.
23. Tanner JM, Whitehouse RH, Cameron N, Marshall WA, Healy MJR,
Goldstein H. Assessment of skeletal maturity and prediction of adult
height. 2nd ed. New York: Academic Press, 1983.
24. Epstein LH, Wing RR, Valaski A. Childhood obesity. Pediatr Clin
North Am 1985;32:363–79.
25. Forbes GB. Influence of nutrition. In: Forbes GB, ed. Human body
composition: growth, aging, nutrition and activity. New York:
26. Lifshitz F. Pediatric endocrinology: a clinical guide. New York:
Marcel Dekker, Inc, 1996.
27. Malina RM. Physical activity and training: effects on stature and the
adolescent growth spurt. Med Sci Sports Exerc 1994;26:759–66.
28. Malina RM. Physical growth and biological maturation of young
athletes. Exerc Sport Sci Rev 1994;22:389–433.
29. Theintz GE, Howald H, Weiss U, Sizonenko PC. Evidence for a
reduction of growth potential in adolescent female gymnasts. J Pedi-
30. Theintz GE. Endocrine adaption to intensive physical training dur-
ing growth. Clin Endocrinol 1994;41:267–72.
31. Constantini NW, Warren MP. Menstrual dysfunction in swimmers; a
distinct entity. J Clin Endocrinol Metab 1995;80:2740–4.
32. Lindholm C, Hagenfeldt K, Ringertz BM. Pubertal development in
elite juvenile gymnasts: effects of physical training. Acta Obstet
Gynecol Scand 1994:73:269–73.
33. Bernadot D, Czerwinski C. Selected body composition and growth
measures of junior elite gymnasts. J Am Diet Assoc 1991;91:29–33.
34. Claessens AL, Malina RM, Lefevre J, et al. Growth and menarcheal
status of elite female gymnasts. Med Sci Sports Exerc 1992;24:
35. Baxter-Jones ADG, Helms P, Baines-Preece J, Preece M. Menarche
in intensively trained gymnasts, swimmers and tennis players. Ann
Hum Biol 1994;21:407–15.
36. Nutter J. Seasonal changes in female athlete’s diets. Int J Sports
37. Warren MP. The effects of exercise on pubertal progression and repro-
ductive function in girls. J Clin Endocrinol Metab 1980;51:1150–7.
38. Bonen A. Recreational exercise does not impair menstrual cycles: a
prospective study. Int J Sports Med 1992;13:110–20.
39. Louis O, Demeirleir K, Kalender W, et al. Low vertebral bone den-
sity values in young non-elite female runners. Int J Sports Med
40. Roemmich JN, Rogol AD. Physiology of growth and development:
its relationship to performance in the young athlete. Clin Sports
41. Seefeldt V, Haubenstricker J, Branta CF, Evans S. Physical charac-
teristics of adult distance runners. In: Brown EW, Branta CF, eds.
Competitive sports for children and youth. Champaign, IL: Human
42. Steen SN, Brownell KD. Patterns of weight loss and regain in
wrestlers: has the tradition changed? Med Sci Sports Exerc
43. Williams MH. Exercise effects on children’s health. Gatorade Sport
Science Institute. Sports Science Exchange 1993;4(3).
44. Housh TJ, Johnson GO, Stout J, Housh DJ. Anthropometric growth
patterns of high school wrestlers. Med Sci Sports Exerc 1993;25:
45. Roemmich JN, Sinning WE. Sport-seasonal changes in body com-
position, growth, power and strength of adolescent wrestlers. Int J
Sports Med 1996;17:92–9.
46. Hughes RA, Housh TJ, Johnson GO. Anthropometric estimations of
body composition in wrestlers across a season. J Appl Sport Sci Res
47. Tipton CM, Tcheng TK. Iowa Wrestling Study: weight loss in high
school wrestlers. JAMA 1970;214:1269–74.
48. Loucks AB,Verdun M. Slow restoration of LH pulsatility by refeeding
in energetically disrupted women. Am J Physiol 1998;275:R1218–26.
528S ROGOL ET AL
by on March 5, 2008