Foster C. Monitoring training in athletes with reference to overtraining syndrome. Med Sci Sports Exerc. 1998; 30(7), 1164-1168. Groll A, Tutz G. Variable selection for generalized linear mixed models by L1-penalized estimation. Technical Report Number 108, Ludwig-Maximilians-University, 2011. Guyon E, Pommeret D. Imputation by PLS regression for linear mixed model. Journal SFdS 2011, 152. Hastie T, Tibshirani R, Friedman J. The elements of statistical learning. Data mining, inference and prediction. Springer Series in Statistics, 2nd edition, 2009. Hellard P, Avalos M, Lacoste L, Barale F, Chatard J-C, Millet GP. Assessing the limitations of the Banister model in monitoring training. 1 Sports Sci. 2006; 24(5):509-20. Huang Z, Kurobe K, Nishiwaki M, Ozawa G, Tanaka T, Taguchi N, Ogita F. Relationships between propelling efficiency and swimming performance in elite swimmers. In: Xlth International Symposium for Biomechanics and Medicine in Swimming, Oslo, June 16-19, 2010. Mujika I, Busso T, Lacoste L, Barale F, Geyssant A, Chatard JC. Modeled responses to training and taper in competitive swimmers. Med Sci Sports Exerc. 1996; 28(2):251-8. Olbrecht J, Madsen 0, Mader A, Liesen H, Hollmann W. Relationship between swimming velocity and lactic concentration during continuous and intermittent training exercises. lnt 1 Sports Med. 1985; 6. Sauerbrei W, Royston P, Binder H. Selection of important variables and determination of functional form for continuous predictors in multivariable model building. Stat Med. 2007; 26(30):5512-28. Schelldorfer J, Buhlmann P, De Geer SV. Estimation for high-dimensional linear mixed-effects models using l1penalization. Scand J Stat. 2011;38(2):197-214. Toussaint HM, Beek PJ. Biomechanics of competitive front crawl swimming. Sports Med. 1992; 13. Wallace LK, Slattery KM, Coutts AJ. A comparison of methods for quantifying training load: relationships between modelled and actual training responses. Eur 1 Appl Physio/. 2014; 114(1). Zamparo P, Bonifazi M, Faina M, Milan A, Sardella F, Schena F, Capelli C. Energy cost of swimming of elite longdistance swimmers. Eur 1 Appl Physiol. 2005; 94(5-6):697-704.

Relationship between body composition and competition performance in swimming 1

3

2

2

2 Megan E Shephard ' , Kellie R Pritchard-Peschek , Tina L Skinner , Kate A Bolam 3 2 Queensland Academy of Sport, School of Human Movement Studies, The University of Queensland, Eiite Sports Division, Swiss Federal Institute of Sport, Switzerland 1

Keywords: elite swimming, body composition, DXA

Introduction In swimming, a multitude of physical, physiological, biomechanical and psychological parameters influence performance. While body composition is believed to be an important contributor to performance in swimming, with some coaches placing a strong emphasis on it during the training preparation phase, previous research is limited and inconclusive (Stager 1984; Siders et al. 1993; Carter & Ackland 1994; Anderson et al. 2008). Anthropometrical data of swimmers competing in the 1991 World Championships reported the best swimmers in all strokes were taller and had longer limb lengths (Carter & Ackland 1994). In that study the best performers in most strokes also had lower proportional skinfold thicknesses (Carter & Ackland 1994). In comparison to elite athletic groups in other sports such as running however, swimmers appear to have higher levels of body fat (Thorland et al. 1983). A study of adolescent female swimmers reported that while the faster swimmers had greater fat-free mass, there was no difference in body fat measurements compared with the slower swimmers (Stager et al. 1984). lt has been previously suggested that a certain level of body fat may be useful for swimmers, enhancing buoyancy and body position in the water, or by providing rounded

516

BMS2014-PROCEEDINGS

body surfaces which are more favourable for streamlining with less drag characteristics (Stager et al. 1984; Bixler 2005). Body composition changes over the competitive season have been previously reported (Meleski & Malina 1985; Siders et al. 1993; Anderson et al. 2008). Siders et al. (1993} reported a decrease in fat mass and an increase in fat-free mass, along with a slight increase in body mass in female collegiate swimmers over a season; however these changes in body fat were not associated with performance. In contrast, over a training season, an increase in lean mass in elite male swimmers and a reduction in skinfold in elite female swimmers has been demonstrated to correlate with enhanced swimming performance (Anderson et al. 2008). Since these studies the use of dual-energy X-ray absorptiometry (DXA) has become a common tool to comprehensively measure body composition in athletes. The purpose of this study was to evaluate the relationship between competition performance and body composition using DXA and other standard measurement practices in a group of elite swimmers.

Methods The study sample consisted of twenty-six elite swimmers; 9 male (age 22.7 ± 2.7 y; mean± SD) and 17 female (age 22.2 ± 2.8 y). All swimmers were, at a minimum, of national finalist standard with the cohort including Olympic and World Championship medalists. Over a 2 year period the body composition of the swimmers was measured in the weeks leading into, or following, the major longcourse domestic competition (National Championships) and/or international competition (i.e. Olympic Games or World Championships) (19 ± 19 d before/after the competition; mean ± SD). A total of 59 performances with corresponding body composition measures were used in the investigation (2.3 ± 1.2 per swimmer). Performance times for each swimmer were recorded as the fastest swim in their best individual competitive event at each major competition. The swimmers' performance times were then converted to a FINA point score (http://www.fina.org) in order to normalise the competition performance across the group. Body composition measures of height, body mass, body mass index (BM I) and sum of 7 skinfolds (SF7} were recorded by an accredited anthropometrist in accordance with the recommended methods of the International Society for the Advancement of Kinanthropometry (Stewart et al. 2011}. Dual energy x-ray absorptiometry (DXA) (Hologic Discovery W, Waltham MA, USA) was used to determine total lean mass (LM), total fat mass (FM), and %body fat (%BF). Measurements were made between 0600 and 0800 h with the swimmers presenting in a rested (no exercise prior) and overnight fasted state. All DXA scans were performed and analysed by the one trained technician and the positioning was standardised within the scanning area. The swimmers were instructed to present in a euhydrated state, and hydration status was determined from assessing the specific gravity of the first void urine sample using an automated refractometer (Pen-Urine S.G, Atago, Tokyo, Japan). Body mass was measured to the nearest 0.05 kg with digital standing scales (Model UC-321, A&D, Tokyo, Japan). Skinfold thickness was measured using calibrated Harpenden skinfold calipers (Baty International, West Sussex, UK). Measurements were taken from seven sites: triceps, subscapular, biceps, supraspinale, abdomen, front thigh, and medial calf. The anthropometrist's typical error of measurement for the sum of seven skinfolds was 0.5 mm or 0.7%. A subsection of 19 swimmers (8 male, 11 female) who underwent body composition assessments near the beginning of the season and near the main domestic competition (i.e. national championships) were analysed for the relationship between competition performance and the change in body composition measures over the domestic training preparation. A total of 26 performances with corresponding change in body composition measures over the domestic training preparation were evaluated {1.4 ± 0.5 per swimmer). Early season measures were performed 130 ± 37 d before the competition, while measurements towards the end of training preparation were collected 20 ± 9 d before the competition. Data was analysed with SPSS (version 7.5, SPSS, Inc., Chicago, IL). Descriptive statistics (means± SD) were calculated for all performance and body composition measures. Pearson correlations were

BMS20 14-PROCEEDINGS

517

performed to determine the relationship between competition performance (FINA point score) and body composition measures. The subsection data evaluating the change in body composition measures from early to late in the season was converted to a percent change before correlations were performed. Paired t-tests were performed to analyse the change in body composition measures from early to late season. Magnitudes of all the correlations were interpreted using the following thresholds: 0.00-0.10 trivial, 0.10-0.30 small, 0.30-0.50 moderate, and >0.50 large {Cohen 1988). In all statistical analyses the 0.05 level of significance was adopted.

Results Body composition data of the swimmers close to the major competition is shown in Table 1. The swimmers presented for body composition assessment in a moderately dehydrated state as determined by their urine specific gravity of 1.022 ± 0.01. For all swimmers combined there was a strong relationship between SF7 and DXA estimated fat mass (r = 0.84). Table 1

Descriptive data of competition performance and body composition variables measured near the main competition (mean 19 ± 19 d)

Performance-FINA Point Score

All Swimmers

Male Swimmers

Female Swimmers

n=54

n=21

n=33

893 ±44

887 ±48

897 ±44

Height (m)

1.80 ±0.08

1.88 ± 0.04

1.75 ±0.05

Body Mass (kg)

73.4 ± 10.1

82.3 ± 8.8

67.8±6.1

Sum 7 Skinfolds (mm)

64.0±23.0

46.3 ± 6.9

75.4 ± 23.0

22.5 ± 2.0

23.2 ± 1.9

22.1± 2.0

Total Mass {kg)

72.5 ± 10.1

81.3 ± 8.7

66.8±6.0

Body Mass Index {kg.m"

2

)

DXA Estimates:

lean Mass {kg)

55.5 ± 9.7

65.9 ± 6.4

49.0±3.8

Fat Mass {kg)

14.4±3.2

12.4 ± 2.3

15.6 ± 3.1

Percent Body Fat {%)

20.1 ±4.7

15.2 ± 1.6

23.2 ± 3.0

Data are presented as mean± SD

The relationship between competition performance and body mass, lean mass and fat mass for the male and female swimmers are shown in Figure 1. Overall, as a combined group there was only trivial to small relationships between the swimmers' competition performance and the body composition measures (range: r = 0.17 to 0.00). The male swimmers had a large positive relationship between performance and body mass (r =0.58; p = 0.006}, fat mass (r = 0.58; p = 0.006), lean mass (r = 0.55; p = 0.01), height (r = 0.54; p =0.01}, and BMI (r = 0.50; p = 0.02). There was also a moderate positive relationship which trended towards significance for the male swimmers between performance and% body fat (r =0.41; p =0.06). While there was no statistically significant relationship between SF7 and competition performance for the male swimmers (r =0.19; p =0.42). Comparatively, for the female swimmers there was a large positive relationship between competition performance and height (r = 0.50; p =0.01) and a moderate negative relationship approached significance between performance and %body fat (r = -0.32; p = 0.07}. However, there was no statistically significant relationship between competition performance and body mass (r = 0.10; p =0.59), lean mass (r =0.28; p = 0.11), BMI (r =-0.24; p = 0.19), fat mass (r = -0.15; p = 0.39) and skinfolds (r =-0.16; p = 0.37) for the female swimmers.

518

BMS20 14-PROCEEDINGS

26 24

• • • • • • • • o• •

22 20 OD

2::. ~ ~

"'

~

:;;

18 16

u..

••

0

14 12

0 -

10

0

... •

•••

·--- ----

-o_tJ.-000 oe • 0 ---• 0 • Q:> o o•

8 80 0

0

75

0

70 OD 2::.

65

~

V>

"' ~ c "' ~

-- -

0

60 0

-

-

--00

0

-

_o-o-C:

o.--

0

0

•

0

0

••

55

Female Swimmers Female Regression Line

0 -

Male Swimmers -

Male Regression Line

50 45

# 40

• •• •

100

--o ---00--- ~·

85 2::. 80

>-

-o 0 CO

_-o-0

70

• •

65 60

o.•• ....

800

820

840

a-•.1\c.

•

55 50 780

... .. . • • 0

V> V>

75

0--

0

OOo

OD

"'

o o--

0

90

~

0

0

95

860

••

•

•••• •

880

900

920

940

960

980 1000

FIN A Point Score

Figure 1

Relationship between competition performance (FINA point score) and body mass, lean mass and fat mass for male and female swimmers

The changes in body composition measures from early to late season are presented in Table 2. There were only trivial to small relationships between competition performance and the change in all body composition measures in both male and females over the training preparation (r ~ 0.18; p > 0.05). The relationship between competition performance and the percent change in SF7 from early to late in the training preparation is shown in Figure 2.

BMS20 14-PROCEEDINGS

519

Change in key body composition measures of the swimmers from early to late in the domestic training season

Table 2

Male Swimmers (n

=12)

Female Swimmers (n

=14)

Absolute Change

%Change

Absolute Change

Body Mass

-2.6 ± 1.4 kg*

-3.1 ± 1.6%

-0.8 ± 1.7 kg

-1.3

Sum 7 Skinfolds

-11.0 ± 5.5 mm*

-19.1 ± 9.1%

-8.0 ± 8.6 mm*

-10.1 ± 9.0%

Lean Mass

-0.9 ± 0.9 kg*

-1.3 ± 1.3%

0.2 ±0.6kg

0.5

Fat Mass

-1.7 ± 1.2 kg*

-11.8 ± 8.4%

-1.0 ± 1.4 kg*

-6.5 ±8.7%

%Change

± 2.7%

DXA Estimates:

± 1.4%

*Denotes a significant change (p < 0.5). Data are presented as mean± SD

Female Swimmers • - - Female Regression Line Male Swimmers 0 Male Regression Line

10

0

•

*"'

•

•

0

""0

~

;.;z

0

-10

• • • •

0

V)

•

0

•

"E::::>

V)

£ -20 g},

-------------L---~-----0 0