In certain sports (e.g., wrestling, light weight crew, and running) body weight is very important. Practices such as acute thermal dehydration, fasting, fluid restriction, or other weight loss practices have been used by athletes in the past as a method to alter weight before competition (3,21,24,29). This is an obvious concern because these practices may cause poor athletic performance and can be harmful to the body; leading to chronic disease if the practice is sustained (20), or even death (4). To reduce acute weight loss practices, athletic trainers and coaches may monitor body composition on a regular basis. Additionally, governing bodies such as the National Collegiate Athletic Association (NCAA) have set new guidelines such as minimum weight assessment in wrestling, which is based on body composition (19). The regulatory guidelines follow the American College of Sports Medicine (ACSM) (21) and the American Medical Association (2) recomendations requiring body composition assessment before the competitive season (18). Since the NCAA implemented their program in 1998 (11), data indicate that it has been effective in reducing unhealthy weight cutting behaviors and promoting competitive equity (23).
Recently, the NCAA added air-displacement plythesmography (ADP) (19) to hydrostatic weighing (HW) and skinfolds (SKs) as acceptable methods for determining body composition. Air-displacement plythesmography is a densitometric method that relies on the measurement of mass and volume to calculate body density (26). Currently, the only commercially available instrument using ADP is the BOD POD® Body Composition System (Life Measurement Inc., Concord, CA, USA). The BOD POD® was developed using the same theory as HW, but it uses air as the displacement medium instead of water. The NCAA approved ADP after the accuracy, precision and bias were evaluated by Utter (27) using HW as the criterion method. Validations of body composition methods against densitometry using HW have been performed in many athletes (6-8,16,22,25,26). For example, the Lohman SK equation (17) for predicting body composition has been crossvalidated using HW in several samples (6-8,26). More recently, as criterion reference standards improve, both the HW and Lohman SK methods have been validated against the 4-compartment model (5,9).
In all previous validation studies for ADP (5-9,16,17,25-27), the calibration protocols, manufacturer guidelines, and approved procedures were carefully followed, and participants were reported to be cooperative and followed strict preparticipation and data collection protocols. This is important because previous studies that have examined ADP found that if the manufacture's recommended protocols regarding body temperature (12), clothing (13,28), and hair (15) were not followed, results could be altered from an underestimation of 1.8% body fat to a 9.0 overestimation of percent body fat compared to the control. Unfortunately, we are aware of, or suspect that some athletes may employ covert methods during body composition testing in an attempt to alter their percent body fat result to gain a competitive advantage or be allowed to compete at an unhealthy weight. “Why would an athlete want to intentionally misrepresent, or more importantly, overestimate their percent body fat? Quite simply, an overestimation of their body fat makes it appear that they have more weight to lose under the NCAA rule (for wrestlers) or an institutional, team or coaches guidelines.”
This has major implications on both the health of athletes and competitive fairness in some NCAA sports. However, to our knowledge, no previous study has evaluated intentional covert manipulation of body composition measurements by ADP. Therefore, in this study, we sought to examine covert behaviors that athletes might employ to purposely alter the results from ADP.
Experimental Approach to the Problem
This study used an initial measurement for each subject following the manufacturer's protocol allowing for each altered measurement to be compared this control measure. To our knowledge, this is the first study evaluating intentional covert manipulation of body composition measurements by ADP. In addition, this novel approach investigates intentional manipulations actually being used and provided to the investigators by athletes.
We recruited 25 varsity athletes at the University of Wisconsin—Madison who participated in wrestling, light weight crew, or track using a combination of flyers, trainer referrals, and word of mouth. These individuals were healthy young adults between the ages of 18 and 25. There were 16 women (crew n = 14, crosscountry n = 2) and 9 men (wrestling n = 7, crosscountry n = 2). The Human Subjects Committee of the University of Wisconsin approved this protocol and informed consent was obtained from the subject prior to participation.
After providing written informed consent, subjects saw a study physician to undergo a prestudy physical examination and provide a medical history to ensure that they were eligible for the study. Subjects would have been excluded if they had any history of chronic illness or disease (i.e., kidney, liver, heart, blood vessel disease, diabetes, and high blood pressure) or any current unstable medical or psychological problems.
Subjects reported to the laboratory after a minimum 2-hour fast to have their body composition measured by ADP (Bod Pod®, Life Measurement Inc., Concord, CA, USA, software version 2.3). All testing was preceded by the manufacturer specified calibration routine using a 50-L calibration cylinder. All measures were completed within a period of 4 hours while the subjects fasted. Subjects removed jewelry and watches, voided if necessary and wore minimal clothing (spandex tights and, for women, a sports bra) and a tight fitting swim cap to cover hair per the standardized manufacturer's guidelines. After entering subject information and completing a weight measurement, subjects completed a control measurement following the manufacture's protocol. Each Bod Pod measurement involves 2 components, the body volume and thoracic gas volume measurement. The first 2 tests are completed to determine body volume. If the first 2 tests were not in close-enough agreement per the manufacture's specifications, the instrument would prompt for a third test. Once the body volume measurement was completed, a thoracic lung volume test was completed for lung volume correction. For this measurement, tidal breathing is monitored, and once the breathing pattern is established, the airway is occluded at approximately midexhalation for a brief period while the subject gives 3 gentle puffs. A time-correlated analysis of airway pressure and the tidal breathing record is used to calculate the thoracic gas volume, which accounts for the isothermal air in the lungs in the determination of body volume and density.
Four additional measurements were then performed with covert altered breathing techniques where the subjects were instructed to employ high lung and low lung volume during the body volume measurement and then the thoracic gas measurement, respectively. For subjects to obtain high lung volume, they were instructed to cognitively inhibit full exhalation and take shallow breaths. For the first altered breathing technique measurement, subjects were instructed to breath normally (control) during the body volume measurement and then to maintain a high lung volume during the thoracic gas volume measurement (HLV-TG). For the next measurement, the subject was instructed to employ the same high lung volume breathing technique for the duration of the body volume determination, while breathing normally (control) during the thoracic gas measurement (HLV-BV).
For the third and fourth altered breathing technique measurements, the subjects were instructed to maintain a low lung volume by cognitively inhibiting full inspiration and then taking shallow breaths to maintain the low lung volume. For the third additional measurement, the body volume was measured while the subject used normal breathing (control breathing) and then the subject employed the low lung volume breath technique during the thoracic gas volume measurement (LLV-TG). For the last altered breathing technique measurement, the subject was instructed to employ the low lung volume breathing technique for the duration of the body volume measurement, but normal (control) breathing during the thoracic gas measurement (LLV-BV). For both the control and altered breathing techniques, measurements were repeated if they did not pass the manufacturer's merit test. Measurements were only used in the analysis if they passed the vendor's merit test.
In addition, we also tested 4 other covert manipulations by subjects including (a) tapping their feet to introduce movement, (b) cupping their hands an inch away from their legs during the body volume measurement in an effort to increase isothermal air volume, (c) sitting in a cold room to cool skin temperature before the ADP measurement, and (d) sitting under heat lamps for 20 minutes to increase skin temperature before the ADP measurement. Subjects tapped their feet and cupped their hands by their body during the body volume measurement. For all these last 4 measurements, the thoracic lung volume measurement was entered from the value determined during the thoracic gas measurement during the control measurement at the beginning of the study.
A power analysis indicated that a sample size of 25 subjects was recommended if we were to allow for a 15% sample loss and yet be able to detect an effect size of 2% fat between the altered breathing tests and the control measurement. We selected the value of 2 percentage point difference because it is comparable to the test-retest precision of ADP, and thus, it is a value that would neither grossly inflate technical error nor be so small that it would be hard to detect (10). Test-retest precision was not measured as part of this study, but preliminary testing found this to be 0.9% fat for young adults. The Cohen's d standardized effect size (absolute effect size/SD) used in our power analysis was 2.2.
A repeated analysis of variance (ANOVA) was performed to test whether the within-subject differences in percent body fat between any of the covert methods and the control method was statistically significant. This analysis indicated the within-subject percent body fat differed from the control condition by treatment and sex. Dunnett's test for multiple comparisons against a single control was then used for post hoc analyses. Sex was only significant when the HLV-BV condition was compared with control, and thus, data were merged across sexes except for this condition. Bland-Altman (1) comparisons, used to compare 2 measurements techniques involving the same data, were performed to determine if the effects of the subjects' behaviors during the measurement varied as a function of the percent body fat or total body mass. Pearson product-moment correlation analysis was used to examine the relationship between the results of the control measure and the altered techniques. A p ≤ 0.05 was required for statistical significance.
Twenty-five subjects completed the protocol. Four subjects were removed from the sample, 3 distance runners and 1 rower, because of a missing result from 1 or more of the testing conditions, resulting in a sample size of 21. A description of the 21 subjects is presented in Table 1. The within-person difference between the covert method and the control method is presented with SDs in Table 2.
Although men were significantly taller, leaner, and heavier compared to the women (p < 0.001 for height, body mass, and percent body fat), sex differences were significant for only 1 of the within-subject difference between altered techniques and the control measure (HLV-TG, p = 0.15; LLV-TG, p = 0.68; HLV-BV, p = 0.25; LLV-BV, p = 0.50). Lung volume estimates averaged 3.7 ± 1.0, 4.8 ± 1.8, and 2.8 ± 1.3 L for control, HLV-TG, and LLV-TG, respectively.
An ANOVA compared within-subject percent body fat calculated from each of the 4 altered breathing techniques, HLV-TG, LLV-TG, HLV-BV, and LLV-BV. Treatment and gender were significant (p < 0.001 and p = 0.02, respectively). All were significantly different from control for percent body fat in women and all but HLV-BV were significantly different from control in men (Table 2). The HLV-TG led to an overestimation of percent body fat (p < 0.001) as did LLV-BV (p < 0.001) in comparison to the control measurement. The LLV-TG (p < 0.001) and HLV-BV (p < 0.001) resulted in an underestimation of percent body fat compared with the control measurement. Gender differences were not significant for the within-subject difference between altered techniques and the control measure; therefore, all further data were reported with men and women combined Within-subject percent body fat calculated from the measurements that involved tapping of the feet, cupping the hands, cold and heat treatments were not significantly different from control percent body fat (Figure 1).
Table 2 provides the variation around the null value when comparing the 4 altered breathing patterns to control. Figure 1 illustrates the variation around the null value when comparing the trapping, cupping, cold and heat alterations to control. No bias with increasing fatness was detected when comparing the trapping, cupping, cold and heat alterations to control (Figures 2-5). In addition, no bias with increasing body mass was detected when comparing the altered techniques to control (Figure 6). The mean difference between percent body fat from each covert breathing technique and percent body fat from control ± 2SD of the corresponding mean difference are shown on these Bland-Altman (1) comparisons (Figures 2-6).
The between-subject variance for each method and the correlation between methods can be used to describe the effects of a systematic behavior on precision of ADP. Measures were strongly correlated (r > 0.85), and there was no detectable increase in the SD (Table 2).
This study evaluated covert subject actions on percent body fat by ADP. We did not validate ADP, and there is no claim that the percent body fat value determined by the standard measurement is accurate, but rather we examined how covert actions may change that value. The techniques HLV-TG and LLV-BV overestimated the percent body fat of the subjects, as both a group mean and individuals, compared to the control measure. The LLV-TG and HLV-BV techniques underestimated the percent body fat compared to the control measurement for the individual and group mean values. The other altered behaviors did not introduce a detectable change in the estimated percent body fat.
With ADP, pressure-volume relationships are used to estimate body volume and calculate body density, which is then used to estimate fat mass, fat-free mass, and percent body fat. This volume measurement is based on the principle that the placement of an object in the ADP chamber reduces the internal volume of the chamber by displacement. When that object is a human body, however, 2 types of chamber air need to be considered in the pressure-volume relationship, isothermal air, and adiabatic air. Isothermal air is considered to be at a constant temperature. The pressure-volume relationship under this condition is described by Boyles' law in which the quantity of air compressed will decrease its volume proportionally to the increasing pressure. In contrast, the temperature of adiabatic air changes as its volume changes during compression and expansion as described by Poisson's law, which is 40% more compressible than adiabatic air (14).
The air in ADP is evaluated as adiabatic air, but the air close to the body surface and in the lungs is isothermal air. The measurement calculation accounts for the air close to the body by calculating a surface air artifact assuming the subjects are wearing minimal, tight fitting clothing, and a swim cap to avoid trapping additional air by the body surface that would behave in an isothermal manner. The calculations must also account for lung volume. Unlike underwater weighing, lung volume is not measured as part of the body volume during an ADP measurement. Instead, both lung volume and air outside the body are measured as the nondisplaced air volume inside the chamber. The thoracic gas volume, however, is isothermal air and is added to body surface air to calculate total isothermal air. Unlike hydrostatic body volume measurements, the thoracic air volume is not measured as part of the body volume, but, it is a vital part of the calculation of the body volume because it is used in the vendor's isothermal air correction. A large thoracic volume results in a large isothermal air artifact correction. Because of this, thoracic volume is measured right after the test to account for the average lung volume.
By altering the subjects breathing pattern, an artifact is introduced into the correction for the assumed portion of total chamber air volume that is isothermal air. During the HLV-TG and LLV-BV techniques, isothermal air is overaccounted for leading to an increased apparent percent body fat. For the LLV-TG and HLV-BV techniques, the opposite was observed, that is, isothermal air is underaccounted for leading to a decreased body fat value.
These covert alterations in breathing technique lead to the artifacts in percent body fat because of the volitional alteration of the assessed thoracic gas volume. These artifacts (altered minus control) averaged between −3.4 ± 2.6 (LLV-TG), −2.1 ± 2.3 (HLV-BV), 2.2 ± 2.1 (LLV-BV), and 3.7 ± 2.1 (HLV-TG) percent body fat. Of interest for this study were the techniques that allowed participants to artificially increase their percent body fat including HLV-TG (3.7 ± 2.1 percent body fat) and LLV-BV (2.2 ± 2.1 percent body fat).
In the interpretation of these results, it will be helpful to consider the context in which the instrument is being used and a practical example. A wrestler undergoing minimum weight assessment may employ methods to fool the technician with the goal to increase their percent body fat reading. If successful, this overestimate of percent body fat will lower their minimum weight. For example, if a wrestler weighing 72 kg employed the HLV-TG technique during testing and thus produced an overestimate of his or her percent body fat by 3.7%, it would allow that wrestler to lower their minimum weight by 2.7 kg (5.9lbs). If the same wrestler also employed the LLV-BV technique during testing, the resulting overestimate of 2.2% body fat would allow them to lower their minimum weight by an additional 1.6 kg (3.5 lbs). A combination of the LLV-BV and HLV-TG techniques could be additive because they are separate components of the ADP procedure. For this example, the summation effect of the covert actions could overestimate percent body fat by 5.9%, corresponding to a 4.3 kg (9.5 lb) underestimation of minimum weight.
Because we found significantly different results with altered breathing, one might consider using a prediction equation for gas volumes instead of measured values. In fact, a study by McCrory et al. (18) found no significant difference between ADP TG measured and TG predicted when using the Bod Pod programed Crapo equation. However, although no significant mean difference was found, the considerable variation suggests individual values may vary beyond acceptable limits. Additionally, athletes could alter their breathing pattern during body volume measurements leading to inaccurate results. Further study is warranted to address this issue and other covert attempts to manipulate body composition test results.
An invalid minimum weight may be problematic from both a health and competitive fairness standpoint. An overestimation of minimum weight would place an athlete in an inappropriate weight class, thus removing the competitive fairness of weight classes. An underestimation of minimum weight suggests an athlete could safely lose additional weight and potentially lead to rapid weight loss strategies previously described. This could provide a false sense of safety and potentially lead to the rapid weight cutting behaviors previously documented (20,21,29) and reduce the NCAA rule's effectiveness in reducing unhealthy behaviors and promoting competitive equity (23). Both an under and overprediction of minimum weight are concerns, because protecting the health and safety of the student athlete is the goal of the NCAA rule and the purpose of the body composition measurement.
The high correlation between the covert techniques and the control measure indicates that these techniques may be hard to detect. For competitive fairness and for the health of the athletes, coaches, trainers and technicians administering an ADP body composition test should be aware of the potential effects altered breathing techniques could have on the measurement outcome. During testing they need to be vigilant watching for altered breathing patterns that may alter body composition results.
We thank the athletes that took the time to participate in the study and the sports conditioning coaches, athletic trainers, and head coaches who allowed them time to participate. The authors have no professional relationships with the companies or manufacturers in this study and the results do not constitute endorsement, or lack thereof, of any product by the authors, the University of Wisconsin or the National Strength and Conditioning Association. No funding received from National Institutes of Health; Welcome Trust; Howard Hughes Medical Institute); and other for this work.
1. Altman, DG and Bland, JM. Measurement in medicine: The analysis of method comparison studies. Statistican
32: 307-317, 1983.
2. American Medical Association. Wrestling and weight control. JAMA
201: 541-543, 1967.
3. Bartok, C, Schoeller, DA, Clark, RR, Sullivan, JC, and Landry, GL. The effect of dehydration on wrestling minimum weight
assessment. Med Sci Sports Exerc
36: 160-167, 2004.
4. Center for Disease Control and Prevention. Hyperthermia and dehydration-related deaths associated with intentional rapid weight loss in three collegiate wrestlers—North Carolina, Wisconsin, and Michigan, November-December 1997. JAMA
279: 824-825, 1998.
5. Clark, RR, Bartok, C, Sullivan, JC, and Schoeller, DA. Minimum weight
prediction methods cross-validated by the four-component model. Med Sci Sports Exerc
36: 639-647, 2004.
6. Clark, RR, Kuta, JM, and Oppliger, RA. The Wisconsin wrestling minimal weight project: Cross-validation of prediction equations. Pediat Exerc Sci
4: 117-127, 1992.
7. Clark, RR, Kuta, JM, Sullivan, JC, Bedford, WM, Penner, JD, and Studesville, EA. A comparison of methods to estimate minimal weight in high school wrestlers. Med Sci Sports Exerc
25: 151-158, 1993.
8. Clark, RR, Oppliger, RA, and Sullivan, JC. Cross-validation of the NCAA method to predict body fat for minimum weight
in collegiate wrestlers. Clin J Sport Med
12: 285-290, 2002.
9. Clark, RR, Sullivan, JC, Bartok, C, and Schoeller, DA. Multicomponent Cross-Validation of Minimum Weight
Predictions for College Wrestlers. Med Sci Sports Exerc
35: 342-347, 2003.
10. Collins, AL and McCarthy, HD. Evaluation of factors determining the precision of body composition
measurements by air displacement plethysmography. Eur J Clin Nutr
57, 770-776, 2003.
11. Committee refines wresting safety rules. NCAA News
35: 1, 1998.
12. Fields, DA, Higgins, PB, and Hunter, GR. Assessment of body composition
by air-displacement plethysmography: Influence of body temperature and moisture. Dyn Med
3: 3, 2004.
13. Fields, DA, Hunter, GR, and Goran, MI. Validation of the BOD POD
with hydrostatic weighing: Influence of body clothing. Int J Obes Relat Metab Disord
24: 200-205, 2000.
14. Heymsfield, S. Human Body Composition
. Champaign, IL. Human Kinetics, 2005. pp. 17-34.
15. Higgins, PB, Fields, DA, Hunter, GR, and Gower, BA. Effect of scalp and facial hair on air displacement plethysmography estimates of percentage of body fat. Obes Res
9: 326-330, 2001.
16. Housh, TJ, Johnson, GO, Kenny, KB, Cisar CJ, Hughes RA, Kenney KB, McDowell SL, and Lundvall P. Validity of anthropometric estimations of body composition
in high school wrestlers. Res Q Exerc Sport
17. Lohman, TG. Skinfolds and body density and their relation to body fatness: A review. Hum Biol
53: 181-225, 1981.
18. McCrory, M, Mole, P, Gomez, T, Dewey, K, and Bernauer, E. Body composition
by air-displacement plethysmography by using predicted and measured thoracic gas volumes. J Appl Physiol
84: 1475-1479, 1998.
19. National Collegiate Athletic Association. 2007-2008 Wrestling Weight Certification Program
. Indianapolis, IN: National Collegiate Athletic Association, 2007. pp. 8-12.
20. Oppliger, RA and Bartok, C. Hydration testing of athletes. Sports Med
32: 959-971, 2002.
21. Oppliger, RA, Case, HS, Horswill, CA, Landry, GL, and Shetler, AC. American College of Sports Medicine Position Stand: Weight loss in wrestlers. Med Sci Sports Exerc
28: 135-138, 1996.
22. Oppliger, RA, Nielsen, CH, and Vance, CG. Wrestlers' minimal weight: anthropometry, bioimpedance, and hydrostatic weighing compared. Med Sci Sports Exerc
23: 247-253, 1991.
23. Oppliger, RA, Utter, AC, Scott, JR, Dick, RW, and Klossner, D. NCAA rule change improves weight loss among national championship wrestlers. Med Sci Sports Exerc
38: 963-970, 2006.
24. Steen, SN and Brownell, KD. Patterns of weight loss and regain in wrestlers: Has the tradition changed? Med Sci Sports Exerc
22: 762-768, 1990.
25. Thorland, WG, Johnson, GO, Cisar, CJ, and Housh TJ. Estimation of minimal wrestling weight using measures of body build and body composition
. Int J Sports Med
8: 365-370, 1987.
26. Thorland, WG, Tipton, CM, Lohman, TG, Bowers RW, Housh TJ, Johnson GO, Kelly JM, Oppliger RA, and Tcheng TK. Midwest wrestling study: Prediction of minimal weight for high school wrestlers. Med Sci Sports Exerc
23: 1102-1110, 1991.
27. Utter, AC, Goss, FL, Swan, PD, Harris, GS, Robertson, RJ, and Trone, GA. Evaluation of air displacement for assessing body composition
of collegiate wrestlers. Med Sci Sports Exerc
35: 500-505, 2003.
28. Vescovi, JD, Hildebrandt, L, Miller, W, Hammer, R, and Spiller, A. Evaluation of the BOD POD
for estimating percent fat in female college athletes. J Strength Cond Res
16: 599-605, 2002.
29. Webster, S, Rutt, R, and Weltman, A. Physiological effects of a weight loss regimen practiced by college wrestlers. Med Sci Sports Exerc
22: 229-234, 1990.
Keywords:Copyright © 2011 by the National Strength & Conditioning Association.
body composition; adiposity; minimum weight; Bod Pod