ResearchGate

See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/10764200

Body Composition in Division I Football Players

Article  in  The Journal of Strength and Conditioning Research · June 2003

DOI: 10.1519/1533-4287(2003)017<0228:BCIDIF>2.0.CO;2 · Source: PubMed

CITATIONS READS
48 1,521

4 authors, including:

Mary Margaret Noel Jaci L Vanheest

Michigan State University University of Connecticut
51 PUBLICATIONS   296 CITATIONS    78 PUBLICATIONS   1,496 CITATIONS   

SEE PROFILE SEE PROFILE

C. D. Rodgers
University of Saskatchewan
47 PUBLICATIONS   682 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Project Issues in Nutrition, FP Essentials View project

All content following this page was uploaded by Mary Margaret Noel on 28 November 2017.

The user has requested enhancement of the downloaded file.

 ,1

-
:
-
Journal of Strength and Conditioning Research, 2003, 17(2), 228 237 J
tv» r.v U K rotegted
4
© 2003 National Strength & Conditioning Association Notice: Tms :

by copyright iaw t £ > tie 17 U.S. Code)

T *

Body Gomposition in Division I Football Players !

!
!
•
i

MARY BARTH NOEL,1 JACI L. VANHEEST,2 PHIL ZANETEAS,3 AND

CAROL D. RODGERS4 ¥

department of Family Practice, College of Human Medicine, Michigan State University, East Lansing ,
Michigan 48824; 2Human Performance Laboratory, Department of Kinesiology, NEAG School of Education,
University of Connecticut, Storrs, Connecticut 06269; 3Zionsville, Indiana 46077; 4Faculty of Physical
Education and Health, The University of Toronto, Toronto, Ontario, Canada.

!
i
r
ABSTRACT several years. Much of this focus has been directed ,1
i

This study assessed body composition o( Division I football toward sports such as gymnastics, diving, wrestling,
players ( n = 69) and compared the findings with previously long-distance running, and other aesthetic sports i

reported data to ascertain whether the increase in player to - where a lower body mass has been perceived as a ;

tal body mass that has been observed over the past 10 years
has been accompanied by an increase in body fat. Body com-
position was determined by hydrostatic weighing and the
measurement of skinfold thicknesses. Total body mass, skin
fold thicknesses, and body fat were greater in the current
players than in players in studies conducted in the early
1980s and early 1990s. Body fat varied significantly across
playing position, with the defensive backs, offensive backs,
and receivers being the leanest and the offensive linemen
means of enhancing performance. It is important to
note/ however, that issues related to low body mass
represent only one end of the body mass continuum
- encountered in sport. In contrast, sports such as foot- f
ball face the challenge of increasing body mass as a !
•
'
means of enhancing performance. Although anecdotal
observation suggests that over the past 10 years in -f
football a '"bigger is better" concept has pervaded :)

and tight ends the most fat. There was no significant rela- through all playing positions at all competitive levels, $
tionship between body composition and playing year or there is limited evidence to support this claim. Fur- !

scholarship status, nor were any differences observed be- thermore, and perhaps of more importance, it is not <\; ;

tween ethnic groups. Of important clinical relevance was the clear whether this increase in body size has been the 'I
finding that the linemen (offensive, defensive) and tight ends result of an increase in lean body mass or of an in- .3
;
}.

were on average greater than 25% body fat, the borderline crease in body fat. It is logical to assume that an in- 3
for obesity in this age group. Much of this fat was deposited i

in the abdominal region, a significant finding when one con- crease in body mass accompanied by an increase in
siders the high correlation between abdominal obesity and fat-free mass would be particularly important in en- i

ischemic heart disease and stroke. The current findings sug- hancing the performance of players at a number of i
1
gest that more attention needs to be given to the nature of playing positions. However, as previous work has in- V.
A
the increase in body mass being achieved by today's football dicated (1, 5, 6, 10), should bus increase in body •

*
player to minimize long-term negative health consequences, weight be the result of an increase in body fat, the . > $j
and the findings reemphasize the need identified in earlier
studies of the importance of detraining programs for these
potential negative effects on performance and long- .
athletes.
Key Words: obesity, body density, body fat, skinfold
thickness, health implications, detraining programs
^
term heal i implications cannot be overlooked. It was
therefore the purpose of this study (a) to evaluate body
composition across playing positions in a group of
NCAA Division I football players (Michigan State Uni-
i

versity), and (b) to cojripare body composition of cur-

Reference Data: Noel, M.B., J.L. VanHeest, P. Zaneteas, rent collegiate football players with that of players in |
and C.D. Rodgers. Body composition in Division I
football players. J. Strength Cond . Res. 17( 2 ):228~237. the early 1980s (20) and early 1990s (2) to determine i

2003. if players in positions that have tended to focus on an

increase in size over the past several years also dem-
onstrate an increase in body fat percentage. Finally, the
skinfold data has been used to calculate body density
Introduction or body fat percentage using a number of different I
T^he integral relationship between performance in calculation formulas. These data were used to ascer-
"

A sport7Todyweighl7 ahd overalItealth ofthe ath“

“ " " ”
taamthe-value-of- using- skinfold - measures to evaluate
lete has received considerable attention over tire past
,
differences in body fat both within and across playing
228 /
^

position, as well as with respect to classifying players
into general population health risk categories based on
percentage body fat.
Methods
Subjects
-
Sixty nine players (35 white; 34 African American)
from a Division I football team volunteered to partic-
ipate in this study. All players were actively engaged
in fall / winter season practices, including a regular
maintenance strength training program, at the time of
evaluation. In accordance with University Guidelines
for the Use of Human Subjects, all measurement pro-
cedures and potential risks were verbally explained to
each participant prior to obtaining written informed
~ ~
consent ” ^
Subjects were asked to report for evaluation twice
within a 2-week period, both times in a state of normal
hydration to ensure optimal measurements. During
the initial visit, each subject's height (Stadiometer,
Country Technology, Gay Mills, WI) and body weight
(Health-O-Meter, Country Technology) were deter-
mined for the calculation of body mass index (BMI,
calculated as weight divided by height squared,
kg m-2). Skinfold thickness at 8 sites was also assessed
*
during this session. Body density was determined by
hydrostatic weighing technique during the second
evaluation session. However, because only 25 of the
players were willing to participate in this portion of
the study, between-position differences were not cal-
culated because of the limited sample size in each
group. However, the average data by position and a
comparison of these data with those of previous stud-
ies is provided.
Determination of Skinfold Thickness
Skinfold thickness at 8 sites (triceps, biceps, subscap-
ular, chest, abdomen, superior iliac, thigh, and calf )
was measured according to the guidelines established
by Jackson and Pollock (9). The selected sites were an-
atomically marked and identified on the right side of
the body prior to measurement. The measurement
procedure involved grasping the skinfold between the
thumb and forefinger, the subsequent placement of the
calipers 1 cm below and perpendicular to the fold, re-
moval of the fingers, and application of caliper pres-
sure at the site for approximately 2 seconds. The thick-
ness of the site was then read from the caliper mea-
suring device, recorded in millimeters to the nearest
0.5 mm, and the calipers released. Two measurements
were taken at each site, and the average of the 2 read
ings was recorded. If the 2 measures differed by more
- these formula, specifically the general population for-
than 2 mm, a third measurement was taken, and the
2 closest were then averaged and recorded as the final
value-AlLmeasurements-were~taken using Lange skin - composition. parameters...acxoss.. playing position and
fold calipers (Cambridge Scientific Instruments, Cam -
Body Composition in Football 229
bridge, MD), and all were taken by the same investi-
gator to control for intertester variability.
Determination of Body Density by Hydrostatic
Weighing
Upon arrival at the laboratory each subject was asked
to change into swim trunks and to void the bladder
and the bowel. Body mass "in air / on land" was then
determined using a standard calibrated weighing bal-
ance. The subjects were then instructed to enter the
water tank (temperature = 35° C) and clear skin, hair,
and clothing of trapped air to avoid inaccurate mea-
sure due to trapped air pockets. The breathing appa-
ratus (mouthpiece and associated tubing) and nose
clips were then placed on the subject, and the subject
was given an opportunity to practice breathing
through the apparatus until a comfortable breathing
pattern was achieved. The subject was then instructed-
to descend to the weighing sear / scale underwater.
Once submerged, the subject sat quietly underwater,
breathing normally. At the first signal, the subjects ex-
haled as much air as was comfortable and then held
their breath for 4-6 seconds, during which time un-
derwater body weight was determined . At the second
signal, the subjects began breathing again from the re-
breathing bag (1.5 L; 100% 02) for a total of 4 cycles
so that residual volume could be determined, as per
the methods of Rahn et al. (16). Upon completion of
the rebreathing process, a third signal was sounded,
and the subjects were instructed to ascend to the sur-
face. The entire testing procedure was repeated 3 times
to ensure accurate data. If a deviation of more than +1
occurred between trials, a fourth measure was taken.
Body density (g- mh1) was calculated.
Determination of Body Composition
Several formulas have been used to determine body
density in football players as well as the general pop-
ulation using skinfold thickness measures. The pri-
mary differences among these formulas involve the
different skinfold site thicknesses and the number of
skinfold sites used in the calculation. In light of these >
'
differences and the health implications of fat depot #
distribution, calculations of standard parameters of
body composition (body density [BD], fat-free mass
[PPM], and body fat percentage) were done using 6
different formulas to emphasize the impact of these 2 •
variables (number and site of skinfold thickness mea-
sure) on interpreted measures of body composition in
this particular population. However, BD, FFM and
body fat percentage data calculated from only 1 of
mula of Sloan (19), which has been shown to be highly
reliable with muscular male athletes, will be used to
discuss the comparative differences in these body !
,
with the previously reported data from players of the

230 Noel/ VanHeest, Zaneteas, and Rodgers
1980s. Percentage body fat was determined using the
Siri (18) formula for white subjects and the formula of
Schutte et al. (17) for African American players when
data were analyzed across ethnic groups.
Statistical Analyses
For the purpose of data analyses, subjects were
grouped according to playing position as per Wilmore
and Haskell (28) ( offensive linemen / tight ends; defen-
sive linemen; offensive backs / receivers; defensive
backs; linebackers; quarterbacks and kickers), ethnicity
(white, African American), and years played (1, 2, 3,
4). Sum of skinfolds, skinfold thickness at each of the
8 measurement sites, body density, FFM, and body fat
percentage was compared across position group and
within each subcategory using analysis of variance
'
techniques (23). Significant pairwise differences were
determined using the Scheffe post hoc test. The level
values presented are expressed as mean ± standard
.
of significance was set at p 0.05 in all instances All
error of the mean.
Expermimental Approach to the Problem
In order to determine whether the increase in body
mass that has been observed in Division I football
players has been accompanied by an increase in body
fat or by an increase in FFM, we ascertained the body
composition of a Division I football team and com-
pared the findings with those reported previously in
the literature. Because not all the earlier data had been
derived using methods similar to those used in the
current study, percentage body fat values reported in
the 1984 study of Smith and Mansfield (21) were re-
calculated using the reported mean skinfold measures
so that more accurate comparisons could be made. The
current study also determined body composition us-
ing both hydrostatic weighing and a variety of differ-
ent skinfold equations. The reasons for this were two-
fold. First, access to hydrostatic weighing is often lim-
ited, and therefore the skinfold technique is a more
commonly used method of body fat determination. By
ascertaining whether the measures obtained from the
skinfold technique are similar to those obtained from
hydrostatic weighing and whether there is a trend in
the disparity between the 2 methods, the strengths
and. weaknesses of the skinfold technique for this spe-
cific population group can be more readily defined. By
comparing the data from several skinfold formulas
that use a different number or a different set of sites ,
of skinfold measures, the strength and weaknesses of
each of the specific formulas, as applied to this specific
population, can similarly be determined. This also en-
ables one to determine whether those formulas that
include specific skinfold measurement sites, particu-
larly those sites that have been shown to be related to
- long-term health- risk-,- may- be- of - more-usefulness -in-
" “
assessing the body composition of football players
1
from a health vs. performance perspective. Finally, the
l
players in this study have been divided across 6 stan-
dard playing position groups, and the data have been
compared among these groups to determine whether
differences in body composition exist across playing
positions. This form of data analysis has also enabled i
us to ascertain whether changes that have occurred
over time in 1 playing position are consistent with
those observed for other playing positions, or whether )
these changes are restricted to playing positions that
depend highly on body mass for effective perfor- i
mance.
Results
Physical characteristics of the subjects across playing i :
position group are described in Table 1 (top half ). The !
offensive linemen / tight ends were significantly taller .I
'
than players of all other positions except the defensive
linemen. The defensive linemen were significantly tall-
er than the offensive backs / receivers and the defensive
backs. Both the offensive linemen / tight ends and the 1
defensive linemen weighed significantly more than the i.
:! I
offensive backs / receivers, the defensive backs, and the
quarterbacks / kickers. The offensive linemen / tight
ends had a significantly greater BMI and ponderal in-
dex than the defensive backs, offensive backs / receiv-
ers, and the quarterbacks / kickers. 4
For comparative purposes, Table 1 ( bottom half ) 2
has been constructed using the data presented by
*
Smith and Mansfield (21) characterizing these same I
parameters in NCAA Division I players of the early $ [
1980s.
As Table 1 demonstrates, the age, height, BMI, and :
ponderal index of NCAA Division I football players I
has remained relatively consistent across time. In con- $ :
trast, body mass has increased approximately 10 kg in $
the offensive linemen / tight ends and the linebackers I
I
since the early 1980s, with players of the other playing *
positions showing only subtle body mass changes (3 I
kg) over this same time period. The sum of skinfolds 3 i
$
for 7 skinfold sites that were common between, the 2 I \
sets of data has also increased during this same time I
£
period for all players except the linebackers, where a
decrease in the total sum was evident. I j
Figure 1 displays skinfold thickness (in millime- $
ters) for each of the 8 sites measured in the current J
study, across playing position. Only the subscapularis,
abdomen, and tricep sites demonstrated a difference
!%
in thickness across position. Skinfold thickness at each i
of these 3 sites was significantly greater in the offen- I
sive linemen / tight ends than in the offensive backs / -iW
receivers. The offensive linemen / tight ends and defen- g
sive linemen each had significantly thicker abdominal
.3
'
and subscapularis sites than the defensive backs. Ab-
I
*
—
dominaL-thickness- was-significantly greater in the de-
fensive linemen than in the offensive backs / receivers.
1
f
%
i
 Body Composition in Football 231

.
Table 1 Subject physical characteristics and sum of skinfolds.

Offensive Offensive
linemen / Defensive backs / Quarterbacks /
tight ends linemen receivers Defensive backs Linebackers kickers Pooled mean
Current study:};
Age (y) 19.7 ± 0.3 19.4 ± 0.3 19.5 ± 0.3
'

19.8 ± 0.3 20.2 ± 0.5 19.3 ± 0.6 19.6 ± 0.1

( n = 16) ( n = 14) ( n = 15) ( n = 12) (n = 6) ( n = 6) ( n = 69)
Height (cm) 190.6 ± 1.2 190.1 ± 1.9 177.3 ± 1.4*+ 182.4 ± 1.1*+ 188.3 ± 1.2* 181.2 ± 2.8* 185.1 ± 0.9
Body mass (kg) 122.4 ± 4.7 110.5 ± 5.4 86.6 ± 2.9*+ 86.9 ± 1.3*+ 104.1 ± 4.7 84.8 ± 1.8*+ 101.2 ± 2.5
BMI 33.6 ± 1.0 30.8 ± 1.6 27.6 ± 1.0* 26.7 ± 0.5* 29.4 ± 1.3 25.8 ± 0.5* 29.4 ± 0.6
Ponderal index 10.13 ± 0.16 10.72 ± 0.37 11.18 ± 0.17* 11.43 ± 0.11* 10.82 ± 0.25 11.50 ± 0.10* 10.96 ± 0.19
Sum of skinfolds 160.8 ± 9.8 150.0 ± 18.0 92.7 ± 11.0* 88.8 ± 6.2* 110.2 ± 14.7 100.42 ± 14.7 121.6 ± 6.5
(mm)§ :
Percentage body fat See Table 3
Smith and Mansfield (21)
Age (y)

Height (cm )
Body mass (kg)
19.4
( n = 20)
190.6
110.2
19.5
( n = 11)
193.0
111.3
.
20.8
( n = 14)
182.1
89.1
(n
20.0
= 10)
183.9
83.2
( n 9)
186.3
-
19.6

94.6 '
19.5
( « = 4)
186.1
84.5
19.8
( n = 68)
187.4
98.5
BMI|| •
30.33 29.88 26.87 24.6 27.25 24.39 28.05
Ponderal index|| 10.60 10.68 11.26 11.77 11.19 11.82 11.02
Sum of skinfolds 107.7 86.6 53.8 54.4 131.1 62.7 78.3
(mm)§
Percentage body fatf 19.4 13.9 7.8 7.3 11.5 10.9 11.8
:
J Values are expressed as means ± standard error of the mean.
§ Measurements of 7 sites: chest, subscapularis, triceps, suprailiac, abdominal, thigh, calf .
|| As calculated from mean height and weight values.
*1 As presented in Table 1 of Smith and Mansfield (21).
* p < 0.05 vs. offensive linemen / tight ends.
+ p < 0.05 vs. defensive linemen.

Total skinfold thickness across the 8 sites was signifi-

cantly greater in the offensive linemen/ tight ends
(170.7 mm ± 10.1 mm) than in the offensive backs /
45 - receivers (99.9 mm ± 12.0 mm) and the defensive
backs (95.3 mm ± 6 . 7 mm).

E
E.
40 -
35 -
30 -
fc Using the equation of Sloan (19), body density,
FFM, and body fat percentage were calculated for each
player and compared across playing position (Table 2).
25 -
Body density was significantly less in the offensive '/Si-
a
.E
20 -
linemen / tight ends than in the defensive backs. TTie
15 -
(0
defensive linemen also had a significantly lower b dy
10 -
5-
density when compared with the offensive backs / re-
ceivers and the defensive backs. FFM was greatest in
^
0 *
the offensive linemen / tight ends, a difference that was
significant when compared with the FFM values of the
offensive backs, defensive backs, and quarterbacks /
<3 kickers. Body fat percentage (27.4% ± 3.6%) was sig-
nificantly greater in the defensive lineman than in the
Offensive Linemen/Tight Ends m Defensive Linemen
defensive backs (15.2% ± 1.0%). There was no signif-
M Offensive Backs/Receivers Defensive Backs
icant difference in body fat percentage among the of-
H Linebackers ffl QBs and Kickers
J fensive linemen / tight ends, offensive backs / receivers,
.
Figure 1 Skinfold thicknesses (mm) across playing linebackers, or quarterbacks / kickers.
positiona+-8-selected-cirLthropometric- sites. Data- are Forxompamtiva purpqses,. body-density, FFM, and
^

presented as means ± standard error of the mean. percentage body fat of players of the early 1980s were

.
i
 232 Noel, VanHeest, Zaneteas, and Rodgers

Table 2. Body density ( BD), fat-free mass (FFM), and body fat percentage across playing position.

Calculated from Smith

and Mansfield (21)
Current study;): mean data
Player position BD . FFM % fat BD FFM % fat
Offensive linemen / tight ends 1.0421 ± 0.004 90.57 ± 3.0 25.4 ± 2.0 1.0581 •
90.57 '
17.81
Defensive linemen 1.0395 ± 0.008 77.96 ± 2.4* 27.4 ± 3.6 1.069 97.06 12.79
Offensive backs / receiver 1.0664 ± 0.005t 71.15 ± 1.0* 17.0 ± 1.9 1.0796 81.54 8.49 ;
Defensive backs
*
1.0685 ± 0.002*t 73.6 ± 0.9* 15.2 ± l.Ot 1.0802 '
76.32 8.27
Linebackers 1.0543 ± 0.010 81.87 ± 3.3 20.6 ± 4.5 1.0721 83.53 11.70
Quarterb acks / kickers 1.0644 ± 0.005 71.39 ± 2.0 . 15.8 ± 1.7 1.0796 77.33 8.49
Pooled mean 1.0544 ± 0.005 71.39 ± 2.0 15.8 ± 1.7 1.0704 86.25 12.44

$ Values are expressed as means ± standard error of the mean. i

*p 0.05 vs. offensive linemen / tight ends. ;
fp 0.05 vs. defensive linemen. i
;
!
i
/
Table 3. Body fat percentage across different calculation formulas and across playing position. *

Offensive
linemen / Defensive Offensive backs / Quarterbacks /
Formula tight ends linemen Defensive backs
J
receivers Linebackers kickers
Hydrostatic weighing1 21.07 ± 2.37 16.3 ± 0.84 7.48 ± 2.48 8.24 ± 0.97 N/A 10.45 ± 1.05
( n = 8) (n = 5) ( n = 2) (n = 7) ( n = 2)
Dumin and Rahaman 26.39 ± 0.87 23.62 ± 1.82 18.75 ± 1.54 17.88 ± 0.86 19.93 ± 2.01 19.15 ± 1.88 i
(7)
Forsyth and Sinning (8) 43.35 ± 3.6 J
•
38.40 ± 5.22 20.27 ± 3.53 17.84 ± 1.07 25.26 ± 5.84 21.13 ± 3.90
Jackson and Pollock 21.89 ± 2.02 20.86 ± 3.53 8.99 ± 2.12 8.99 ± 1.19 13.91 ± 4.39 11.81 ± 3.08 3.
Sloan (19) 25.12 ± 1.96 26.56 ± 3.78 14.30 ± 2.06 13.30 ± 0.10 19.73 ± 4.41 15.11 ± 2.03 $
White et al. (25) 21.60 ± 1.40 19.80 ± 2.60 12.50 ± 1.80 11.30 ± 0.80 13.80 ± 2.50 12.50 ± 1.70 o
<v

Wilmore and Behnke 23.82 ± 0.84

, 21.66 ± 1.38 16.20 ± 1.09 16.45 ± 0.54 18.49 ± 1.73 17.85 ± 1.88 u
;

(27)
Pooled mean 25.79 ± 8.15 23.62 ± 7.33 13.83 ± 4.90 13.39 ± 4.09 18.54 ± 4.27 15.45 ± 4.03 15
$
1 i

calculated using positional group mean data provided
in a similar study, published in 1984, by Smith and
Mansfield (21) (Table 2). As is evidenced by these data,
players in all 6 playing position categories have dem-
onstrated a change in each of these 3 body composi-
tion parameters across time, irrespective of playing
position. In general, players of the current study had
a lower body density and a higher percentage body
fat than those players in the same playing position in
the early 1980s. FFM has decreased (defensive line
men, offensive backs / receivers, quarterbacks / kickers)
or showed relatively little change (offensive linemen /
tight ends; linebackers) over the same time period.
The impact of the different methods of calculating
body density on percentage body fat is displayed in Ta
ble 3. In all instances, the lowest body fat percentage was
the percentage calculated using the body density mea
sure determined by hydrostatic weighing, although one
"
must be cautious -in-interpreting this -finding-in-light-of
' ~“
'
I
groups. The Forsyth and Sinning (8) equation yielded the !
lowest body density and the highest percentage body fat
for all players except the defensive linemen, often yield-
ing a measure at least 2 times greater than the percent-
J
,s |
age body fat calculated from hydrostatic weighing. The 3
range in body fat percentage across the 8 different cal- i
K
culation methods was between 9.6% (defensive tUcks;
)
range 8.24-17.88%) and 24.47% (offensive linemen / tight
ends; range 18.9-43.4%). When one considers the num-
- ber and location of skinfold sites used to ascertain body k
*
density, those equations that used one of either the ab- 5
'- j
dominal (Wilmore and Behnke [27]) or subscapularis h
(Sloan [19]) site or both sites (Forsyth and Sinning [8])
tended to result in a lower body density and therefore :
- a higher percentage body fat. The offensive linemen /
tight ends had the highest calculated percentage body
- fat for 6 of the 7 methods of body composition deter-
mination that were used. Similarly, the defensive line -

the limited sample size for several of the player position

— man-had- the.Jowest-percentage body fat except when
the method of White et al. (25) was used.
^

!

Research has indicated that both BMI and body fat
percentage, particularly depot distribution, are affect-
ed by ethnic origin. Data were therefore further ana-
lyzed to determine whether these general population
differences also prevail within an athletic population.
There was no significant effect of player race on total
skinfold thickness, skinfold thickness at any of the 8
sites, body fat percentage, or BMI. The white players
were, however, significantly taller (187.9 cm ± 1.2 cm)
than African American players (182.2 cm ± 1.2 cm).
When the data were further analyzed across playing
position within each specific ethnic group, however,
different positional patterns did become evident. For
instance, the height of the white players did not differ
across position. However, in the African American
subjects, the defensive linemen (189.79 cm ± 1.57 cm)
and the linebackers (187.54 cm ± 1.69 cm) were sig-
nificantly taller than the offensive backs / receivers
(177.02 cm ± 1.37 cm). BMI was significantly greater
in the offensive linemen / tight ends /33.45 cm ± 1.04
cm) than the defensive backs (25.18 cm ± 0.51 cm)
among white subjects, whereas no differences existed
in this parameter across position group in the African
American subjects. Position group differences in body
fat percentage using skinfold assessment differed only
when using the Wihnore and Behnke (27) formula (de-
fensive linemen 22.43% ± 0.08% vs. offensive backs /
receivers 16.37% ± 1.25%) in the African American
subjects, whereas with the white subjects, position
group differences were noted only when using the For
syth and Sinning formula (offensive linemen / tight
ends 43.21% ± 3.77% vs. defensive back 18.53% ±
1.32%). This finding must, however, also be tempered
with the disparity in sample size among the different
playing positions across the 2 ethnic groups.
To determine whether the duration of exposure to
an NCAA Division I playing environment, with its
specific positional player training programs, had any
effect on body composition, the data were also ana
lyzed across player year. There was no significant ef
fect of playing year on BMI, body density, or body fat
percentage (data not shown).
Data were also grouped according to scholarship
status,, irrespective of playing position, to ascertain
whether the body composition of scholarship players
differed from those of players who might be assumed
to be of lesser talent in light of their lack of scholarship
status. ‘ BMI and body mass (BM) were significantly
-
greater in. the scholarship athletes . (n = 43; BMI =
30.51 ± 0.69; BM 106.49 kg ± 3.22 kg) than in the
nonscholarship athletes ( n = 26; BMI = 27.52 ± 0.95;
BM = 92.31 kg ± 3.31 kg). Body fat percentage, when
determined by underwater weighing only, was also
significantly greater in scholarship athletes ( n
- &4Z%...±. 3.41%)Jhan inmQnscholarship athletes (re ~
1 ,
— 9;
15; 11.85% ± 1.51%).
Body Composition in Football 233
Discussion
The current study was designed primarily to ascertain
whether the emphasis on body size that has pervaded
the sport of football over the past 2 decades has been
accompanied by a parallel increase in fat-free body
mass or by the more negative consequence of an in-
crease in body fat percentage in NCAA Division I foot-
ball players. A secondary purpose of the study was to
compare the use of different skinfold equations with
respect to their body fat percentage predictive value,
both within and across playing positions, as well as
their ability to classify players into health risk-related
categories of body fat (e.g., poor, fair, average, and ex-
cellent).
Division I football players
_ in this study were on
~
average imilaf in height but heavier m iotarbody -
^
' “ ' ” ^ '
mass than players of a similar competitive playing lev -
el in the early 1980s. This finding is similar to that of
Olson and Hunter (14), who compared the 1973-1974
and 1983-1984 height / weight patterns in players of
this same playing division, as well as to those reported
by Oppliger et al. (15) comparing Division I-A players
and Snow et al. (22) in a study comparing the body
composition of professional football players over the
past 20 years. The current data, in conjunction with
the similarly reported trends across multiple playing
levels over the past 20 years, supports the contention
of a continuing emphasis on increasing body size
among players in this sport.
- When an increase in total body mass is accompa -
nied by a parallel increase in FFM there is often a con -
comitant positive effect on player performance (5, 6),
and health risks are minimal, if not nonexistent. It is
important to note, however, that our data indicate that
this increase in total body mass may not necessarily
have been accompanied by an increase in FFM. A com -
parison of the FFM (pooled mean) of the players in
- the current study with those of the Smith and Mans
field study (21) would suggest that this increase in to-
-
- tal body mass has been accompanied by a lack of
change, or even a decrease, in FFM. This is particularly
evident in the linemen (offensive and defensive ) and
the linebackers. Failure to increase FFM in conjuration
with an increase in total body mass would suggest
that the increase in total body mass was primarily due
to an increase in body fat. This speculation is further
supported by a comparison of the skinfold thickness
data and the percentage body fat measures of ’ the
Smith and Mansfield (21) study with those of the cur
rent study. Mean pooled skinfold thickness of the 6
sites that were common to the 2 studies (subscapularis,
abdomen, triceps, suprailiac, thigh, and calf ) indicated
-
that both total skinfold thickness and the thickness of
each of the individual sites were greater in the current
study than in the earlier study. This difference also
persisted across all playing positions. Furthermore, the
234 Noel, VanHeest, Zaneteas, and Rodgers
reported range in body fat was from 7.3% to 13.9% in
the Smith and Mansfield (21) study, whereas in the
current study the average ranged from 15.2% to 25.4%.
Although one might argue that this approximately
twofold increase is due to the different body fat for-
mulas used between the 2 studies, when body fat per-
centage was recalculated with the mean raw scores re-
ported in the Smith and Mansfield study (see Table 3)
usmg the same formula as was used in the current
study, this difference in the range of scores was only
modestly reduced. Collectively, these data would sug-
gest that the particular group of Division I football
players in the current study were not only "bigger,"
they were also fatter than previous players of a similar
competitive playing level. A study by Bale et al._ (1) of
NCAA Division -II players in the early1990s_also noted
that the body fat percentage of the heavyweight play-
ers (primarily heavy linebackers) in their study was
significantly greater than that which had been report - A secondary purpose of this study was to deter-
ed in studies conducted in the 1970s (ll, 20) and early
1980s (25). More recently, in a study examining pro-
fessional football players, Snow et al. (22) noted that
although the percentage body fat of the defensive
backs and linebackers was similar to that reported by
Wilmore et al. in 1976 (29), body fat percentage had
increased an average of 9.8% in the offensive linemen /
tight ends across this same time period . It is interest-
ing to note that previous work by Wickkiser and Kelly
(26) found that both the player and coach tend to es-
timate the mean optimal player competitive body
mass to be between 6 and 9 pounds heavier than den-
sitometric analysis indicates as optimal. Although in
the present study we did not discuss optimal playing
body mass with the coaches and players, it would'
seem that based on the discrepancy between total
body mass and current optimal body fat levels for
males of this age group (11-13% = "good"; 7), this
focus on a greater total body mass regardless of body
composition continues to persist.
As has been demonstrated in previous studies (26),
a significant difference in percentage body fat was
found to exist across playing positions in the Division
I football players examined in the present study. Body
density was greatest in the defensive backs, followed
closely by the offensive backs / receivers and quarter
backs. Consequently, this group was also the leanest
group of players. The defensive linemen and the of-
fensive linemen / tight ends were the most fat, with
body fat percentages of 27.4 and 25.4, respectively.
This rank order across playing position is consistent
with that observed previously in Division I and II
players (2, 21, 25, 26) and in professional football play-
ers (22, 28) and suggests that the association between
player size and performance responsibilities at each
particular playing position have remained relatively
^
c6hstffin er the pasrseveraldecades;
'
^
* ” "
——
The average body fat values observed in this study
ranged from 6% (linebackers) to 10% (offensive line-
men / tight ends) greater than the "ideal" values by po-
sition recommended previously by Crews and Mea-
dors (6). In addition to the potential negative impact
of this increase in body fat on playing performance (1,
5, 6), what is of perhaps more critical importance are
the health implications of this increase in body fat. The
defensive linemen and the offensive linemen / tight
ends are on average greater than 25% body fat, the
borderline for obesity in men of this age group. More- [
over, as is evident from the differences across playing
position in skinfold site thickness, much of this fat is
deposited in the abdominal and subscapularis regions. :
This finding is of significant medical importance when
one considers cross-sectional studies that have shown
abdominal obesity to be strongly correlated with risk
f-
factors for ischemic heart disease, stroke/ and non-in-
sulin-dependent diabetes mellitus (4).
mine the strength of several different formulas that can
be used to determine percentage body fat with respect
to both the actual percentage body fat calculated when
using that particular formula and its ability to quan-
tify the trend in body composition differences com-
monly observed in football across the different playing
positions. The formulas that were chosen varied in the
number of skinfold measures that were required and
the skinfold sites that were used.
The mean calculated percentage body fat varied
widely across formula and across playing position.
This was most likely because of the difference in lo-
cation of the body fat deposits between positional
" *
groups and the different emphasis of particular fat de-
posit locations within the formulas. For the linemen
•
(offensive and defensive) and defensive backs in par-
ticular, the formula of Forsyth and Sinning (8) resulted
in a twofold higher calculated level of body fat than
was observed for these same players when body den- a
sity, and subsequently percentage body fat, were de-
termined using the hydrostatic weighing technique.
All other formulas differed no more than 7% from hy- .V
drostatic weighing for these same 3 player groups., Al-
f
V
though one might first look upon this disparity among
methods and across the different formulas as an Im- I i
- mediate reason to negate the use of the Forsyth and
Sinning formula in this particular population group, it
is of interest to note that an important difference of
this formula is the inclusion of the abdominal skinfold
site measurement, a site of fat deposition that has been l-
->
highly associated with an increased risk of obesity, I
cardiovascular disease, and diabetes (3, 4). When this !
is considered in conjunction with the fact that regard-
less of playing position, the site of greatest fat depo-
sition in this population of young male athletes was
the abdominal site, the health implications of this
— -trend- to -inerease body- mass cannot be ignored.
In light of the limited sample size in several of the
i

position groups, it is difficult to make comparisons
among the various formulas used for body fat deter-
mination with the measure attained using hydrostatic
weighing, the traditional "gold standard" method . In
the positions where 50% or more of the players did
undergo hydrostatic weighing (offensive linemen / re-
ceivers and defensive backs) the accuracy of the skin-
fold measures was dependent both on the method
used to calculate percentage body fat and the player
position. For instance, for the offensive linemen, with
the exception of the Forsyth and Sinning formula, all
the other formulas calculated percentage body fat
within + 5% of that determined using hydrostatic
weighing. In contrast, when one examines the data for
the defensive backs, 3 of the formulas (Durnin and
Rahaman [7], Forsyth and Sinning [8], and Wilmore
and Behnke [27]) yielded calculated body fat percent-
ages approximately 2 times greater than that deter-
mined by hydrostatic weighing, whereas the other 3
formulas yielded measures equal to ( Jackson and Pol-
'*
lock [9]) or within +5% (Sloan [19], White et al. [25]).
Thus it would appear that the accuracy of each of the
skinfold formulas is also position dependent. Previous
work by White et al. ( 25) found nonsignificant differ-
ences between measures from hydrostatic weighing
and . the Sloan skinfold method, although this same
study also identified higher correlative values when
using position-specific equations for determining lean
body mass and body density in backs and in linemen.
Snow et al. (22) also observed a difference across play-
ing position in the accuracy of skinfold body fat per-
centage measures vs. hydrostatic weighing; however,
their data indicated that the skinfold method resulted
in a significant underestimation of body fat percentage
as body fat percentage increased. This finding is con-
tradictory to what was noted in the current study, al-
though it is important to note that the Snow et al. (22)
! study used the Jackson and Pollock (9) skinfold equa-
tion, not the Sloan (19) equation that was used in this
study to determine percentage body fat, and it also
examined players from the National Football League,
# not college-level players. It must also be remembered
that many of the skinfold equations that are used do
not account for ethnic differences or differences in
body fat depot deposition. The position differences
that were noted in the current study with respect to
the accuracy of the skinfold equations may have been
) a function of either one, or both, of these factors. Al-
* the moderate health risk category. In reviewing the
though the actual percentage body fat calculated dif -
fered across the 6 formulas, the trends across playing
position remained relatively constant: the lineman (de-
fensive and offensive) and linebackers were the fattest
and the defensive backs the leanest. Moreover, when
using a general classification scheme to describe level
of fatness relative to age-related standards, the offen-
si-ve-linemen-wouldr-have-been categorized as "poor"
using 4 of the 6 calculation formulas that were exam-
d-.rV,
'
- • •
Body Composition in Football 235
ined and the defensive backs as "good ' (in 3 formulas)
7
or "average" (in 3 formulas). This would suggest that,
although the specific percentage of body fat deter-
mined by skinfold measures may vary widely across
formulas, the general classification of players into cat-
egories of fatness and across playing position remains
relatively consistent.
Although body density is known to be different
between white and African American individuals (17),
there was no significant effect of player race on body
density in the current study. This failure to find a dif-
ference is not entirely unexpected because we used the
Schutte et al. equation (17), an equation designed to
account for the proposed ethnic difference in lean
body mass, to determine body density in our African
American players. What was of interest to note, how-
ever, was that, despite research indications to the com
trary (2, 24), we failed to find any significant difference
in skinfold site thicknesses between the 2 primary eth-
nic groups (white, African American) in this study.
This was likely because of the relatively uniform na-
ture of body types across the 2 ethnic groups that were
examined in the study or because of the uneven dis-
tribution of African Americans and whites among the !
different playing positions. Despite these findings,
however, the known differences across racial groups
in cardiovascular disease risk and its relationship to
obesity and mortality (24) make further investigation
in this area essential.
In the nonathletic population, BMI is highly cor-
related with other measurements for obesity and rep-
resents a practical means to determine the presence of
obesity. However, because of the muscular nature of
most athletes, BMI is typically overlooked as a prac-
tical test for the sports medicine physician. The over-
weight BMI is defined as 27.7 or greater for men (13),
which correlates with a body weight 20% above the
desirable value. It was interesting to note that this
standard was met by all but 2 position groups (defen-
sive backs, quarterbacks / kickers) including 2 that
were considered to be lean (offensive backs, lineback- ,
ers) based on their skinfold percentage Those individ-
uals having a BMI of 30 to 35 are considered at mod-
erate health risk, and those with a BMI of 35 to *0 are
considered to be at high health risk (a BMI greater
than 40 represents a very high health risk ) (3). Based
on these criteria, the linemen are at the high end of
skinfold data of this group, it is hard to represent this
group as only overmuscular when accounting for their
BMI. Indeed, they fit the criteria for being overfat as
well. For a group as young as the linemen (average
age 19 years), they can be considered as falling into i
the moderately high-risk category while approaching
the high-risk category for weight-related medical prob-
— lems.-These_.data - continue to support . the inadequacy
, .
of BMI as a measurement of health risk in muscular
236 Noel, VanHeest, Zaneteas, and Rodgers

populations. However, the initial usefulness of BMI as study also revealed that players ,in these 3 positional :
groups were categorized as fair or "poor" with re-
// /

a screening device to identify potentially overweight

individuals should not be ignored, particularly when
BMI exceeds 30.
In conclusion, the results of this study suggest that
the increase in total body mass that has been observed
in Division I football players over the past several
years has also been accompanied by an increase in
body fat percentage, with the ultimate result being
that some players are not only overweight by usual
standards but overfat as well. Although the implica -
tions from these conclusions are significant from a
functional performance perspective, they are most sig-
nificant from the standpoint of the athlete's overall
health and medical prognosis. The issue of optimal
playing total body mass and body fat for football play-
ers (especially the linemen) represents a potentially
spect to body fat percentage when compared with
general population norms for individuals of the same .
age. Because individuals in these categories are at a
substantially increased risk for obesity, diabetes, and
cardiovascular disease, among these players one might
conclude that the continuing emphasis on body mass
in the sport of football has had a significant negative
-
effect on long term health for individuals in these
playing positions. If continued emphasis is to be
placed on increasing the body mass of players in these
position, more specific programs that focus on increas -
ing FFM, in contrast to body mass in general, must be
developed. Additionally, players need to receive more
education regarding nutrition and detraining to min-
-
imize the long term health cohsequences of this big
// “

!
problematic medical development in that the increased is better" trend.
size of the players may be followed by a concurrent
increase in body fat. At some point the question of References
long-term implications of the overweight football play-
.?

er needs to be considered in the development of op- 1. BALE, P., E. COLLEY, J.L. MAYHEW, EC. PIPER, AND J.S. WARE.
timal weight / body fat guidelines for each position, as
well as its influence on position-specific performance.
Detraining programs, such as those described by Krae- 2.
mer (12) and reemphasized again by Snow et al. ( 22)
must become a reality for all players, but in particular
for players in these higher health risk playing posi-
. BjORNTORP, P. Regional patterns of fat distribution. Ann. Intern.
tions. Within this context, the maximization of sport-
3
specific function should not represent a antithetical ob- 4.
jective, and the long-term health considerations of the
athlete must always be at the forefront of all training
program design. Finally, it must be considered that 5 .
there is probably a point of diminishing returns,
Anthropometric and somatotype variables related to strength
in American football players. J. Sports Med. Phys . Fitness 34:383-
389. 1994.
BARKER, M., T.J. WYATT, R.L. JOHNSON, M.H. STONE, H.S.
.
O'BRYANT, C POE, AND M. KENT. Performance factors, psycho-
logical assessment, physical characteristics, and football play- i
ing ability. J. Strength Cond. Res. 7:224-233. 1993.
Med . 103:1052-1062. 1985 .
BJORNTORP, P. The associations between obesity, adipose tissue
distribution and disease. Acta Med. Scand . 723(Suppl.):121-134.
1988.
BURKE, E.J., E. WINSLOW, AND W.V. STRUBE. Measures of body
composition and performance in major college football players.
^ .
/. Sports Med. 20:173-180 1980.
5

where increasing largeness past a certain point is not 6. .

CREWS, T.R., AND W.J MEADORS. Analysis of reaction time,
helpful in football skill because an increase in muscle speed and body composition of college football players. J.
mass will no longer occur, but rather an increase in Sports Med. 18:169-174. 1978. <
7. DURNIN, J.V.G.A., AND M.M. RAHAMAN. The assessment of the
body fat, where the medical risks undoubtedly out- amount of fat in tire human body from measurements of skin-
weigh the athletic prowess. .
fold thickness. Br. J. Nutr. 21:681-689 1967. i
J
8. FORSYTH, H.L., AND W.E. SINNING. The anthropometric esti- V
Practical Applications mation of body density and lean body weight of male athletes.
.
Med Sci . Sports Bxerc. 5:174-180. 1973.
IV

Although a greater body mass may be beneficial to the 9.
positional playing demands of football linemen and
10 .
linebackers, the current data suggest that this tendency
JACKSON, A .S., AND M.L. POLLOCK. Generalized equations for
-
predicting body density of men. Br. }. Nutr. 40:497 504. 1978.
KELLY, J.M., AND J.D. WICKKISER. For "ideal football weight"
.
assess body fat, not poundage. Phys. Sportsmed. 3:38-42 1975. V

to increase body mass among players in these posi - 11. .

tions has been accompanied by a significant increase
in body fat percentage over the past 2 decades. Al-
12.
though this study did not . directly assess whether this
increase in body fat has been accompanied by a 13.
change in performance ability, previous work showing
a strong positive correlation between FFM and perfor-
mance would certainly suggest that the increase in
14.
body mass that was observed in this study, because it
KOLLIAS, JV E R. BLJSKIRK, E.T. HOWLEY, AND J.L. LOOMIS. Car-
diorespiratory and body composition measurements of a select
:<
i
group of high school football players. Res. Q. 43:472-478.1972.
KRAEMER, W.J. Detraining the "bulked up" athlete: Prospects a
for lifetime health and fitness. NSCA J. 5(2):10-12. 1983. 1>
NATIONAL CENTER FOR HEALTH STATISTICS. Anthropometric and
Prevalence of Overweight, United States 1976-1980 (U.S. Depart- 1;
ment of Health and Human Services publication no. 87-1688),
P
Hyattsville, MD: National Center for Health Statistics, 1987.
OLSON, J.R., AND G.R. HUNTER. A comparison of 1974sand 1984
player sizes, and maximal strength and speed efforts for Di-
S?: •

is in the form of body fat as opposed to FFM, is not vision I NCAA universities. NSCA J. 6(6):26-28. 1985. I
- .
likely-to - enhance- the--playing-ability of these -players - T5. — OPPLIGER; -R.-Av - D.H;- NIELSEN> A.C. SHETLER, E.T. CROWLEY, %
Moreover, and of perhaps greater importance, this AND J.P. ALBRIGHT. Body composition of collegiate football *
4
J
 o 1 ^
Body Composition in Football 237

players: Bioelectrical impedance and skinfolds compared to hy - as predictors of mortality in black and white men. Am. }. Epi -
drostatic weighing. / Orthop. Sports Phys. Ther. 15:187 192.1992.
-
25. .
-
demiol 135:1137 1146. 1992.
WHITE, J , J.L. MAYHEW, AND F.C. PIPER. Prediction of body
16. RAHN, H., W.O. FENN, AND A.B. Oris. Daily variations of vital
capacity, residual air, and expiratory reserve including a study -
composition in college football players. J. Sports Med. 20:317 i
of the residual air method. Am. J. Physiol . 1:725 736. 1949.
- .
324 1980.
17. SCHUTTE, J.E., E.J. TOWNSEND, ]. HUGG, R.F. SHOUP, R M. MAL . - 26. WICKKISER, J.D., AND J.M. KELLY. Tlie body composition of a
.
INA, AND C.G. BLOMQVIST Density of lean body mass is greater
.
college football team. Med . Sci Sports Exerc. 7:199 202. 1975.
-
. 27. WILMORE, J.H., AND A.R. BEHNKE, An anthropometric estima -
-
in blacks than in whites. /. Appl Physiol 56:1647 1649. 1984
tion of body density and lean body weight in young men. J.
18. SIRI, W.E. Body Composition From Fluid Spaces and Density (Don - Appl. Physiol 27:25 31. 1969.
-
.
ner Lab. Med Physicas Report 19). Berkeley: University of Cal - 28. WILMORE, J.H., AND W.L. HASKELL. Body composition and en -
ifornia, 1956. durance capacity of professional football players. /. Appl Phy-
i

19. SLOAN, A.W. Estimation of body fat in young men. J. Appl Phy
-
- siol 33:564-567. 1972.
siol . 23:311 315. 1967. 29. .
WILMORE, J.H., R.B PARR, W.L. HASKELL, D.L. COSTILL, L J ..
20. SMITH, D.P., AND R.J. BYRD. Body composition, pulmonary MILBURN, AND R.K. KERLAN. Football pro's strengths and cv
—
function and Vo2max of college football players. J. Sports Med . — . -
weakness charted . Phys. Sportsmed 4:45 54. 1976.
-
Phys. Fitness 16:301 308. 1976.
21. SMITH, J.F., AND E.R. MANSFIELD. Body composition prediction Acknowledgments
in university football players. Med . Sci. Sports Exerc. 16:398 405 .
~
;
- . Tfre autfrors~would—like
”
to-gratefuHy-acknowledge-the-
19847
-
22. SNOW, T.K., M. MILLARD STAFFORD, AND L.B. ROSSKOPF Body . technical assistance of Mr. David Anderson (Department of i
composition profile of NFL football players. J. Strength Cond . Physical Education and Exercise Science, Michigan State
.
Res 12:146-149. 1998. [ University) and the secretarial assistance of Ms. Deb Misiak !
23. STATVIEW® SE + GRAPHICS. The Solution for Data Analysis and (Department of Family Practice, Michigan State University).
i
.
Presentation Graphics Berkeley, CA: Abacus Concepts, Inc., 1987. i

.. .
24. STEVENS, J., J.E. KEIL, P F RUST, R.R. VERDUGO, C.E DAVIS, H.A . Address correspondence to Dr. Carol D. Rodgers, [
.
TYROLER, AND P C. GAZES. Body mass index and body girths carol.rodgers@utoronto.ca.

i
t
\
i
: !
(

:
i

"i

View publication stats