Secondary Logo

Journal Logo

Evaluation of the BOD POD® for assessing body fat in collegiate football players

COLLINS, MITCHELL A.; MILLARD-STAFFORD, MELINDA L.; SPARLING, PHILLIP B.; SNOW, TERESA K.; ROSSKOPF, LINDA B.; WEBB, STEPHANIE A.; OMER, JAY

Medicine & Science in Sports & Exercise: September 1999 - Volume 31 - Issue 9 - p 1350-1356
Special Communications: Technical Note
Free

Evaluation of the BOD POD® for assessing body fat in collegiate football players. Med. Sci. Sports Exerc., Vol. 31, No. 9, pp. 1350-1356, 1999.

Purpose: The purpose of this investigation was to evaluate the accuracy of a new air displacement plethysmograph, BOD POD® Body Composition System, for determining %fat in collegiate football players.

Methods: Body fatness was estimated from body density (Db), which was measured on the same day using the BOD POD and hydrostatic weighing (HW) in 69 Division IA football players. In addition, 20 subjects were whole body scanned using dual-energy x-ray absorptiometry, DXA (Lunar DPX-L) to assess total body mineral content and %fat. Mineral content and HW determined Db were used to compute %fat from a three-component model (3C; fat, mineral, and residual).

Results: Test-retest reliability for assessing %fat using the BOD POD (N = 15) was 0.994 with a technical error of measurement of 0.448%. Mean (± SEM) Db measured with the BOD POD (1.064 ± 0.002 g·cc−1) was significantly greater (P < 0.05) than HW (1.060 ± 0.002 g·cc−1), thus resulting in a lower %fat for the BOD POD (15.1 ± 0.8%) compared with HW (17.0 ± 0.8%). Similar results (N = 20) were found for DXA (12.9 ± 1.2%) and the 3C (12.7 ± 0.8%) where %fat scores were significantly higher (P < 0.05) than scores determined using the BOD POD (10.9 ± 1.0%).

Conclusions: Db measured with the BOD POD was higher than the criterion HW, thus yielding lower %fat scores for the BOD POD. In addition, BOD POD determined %fat was lower than DXA and 3C determined values in a subgroup of subjects. Assessment of %fat using the BOD POD is reliable and requires minimal technical expertise; however, in this study of collegiate football players, %fat values were underpredicted when compared to HW, DXA, and the 3C model.

Exercise Research Laboratory, Department of Health and Performance Sciences, Georgia Institute of Technology, Atlanta, GA 30332-0110; Department of Health, Physical Education, and Sport Science, Kennesaw State University, Kennesaw, GA 30144-5591

Submitted for publication May 1998.

Accepted for publication February 1999.

Address for correspondence: Dr. Mitchell A. Collins, Department of HPS, Kennesaw State University, 1000 Chastain Road, Kennesaw, Georgia 30144-5591. E-mail: mcollins@kennesaw.edu.

Increased body fatness (excess weight) has been negatively associated with physical performance, particularly when the activity involves weight-bearing exercise (4,25). In sports where athletes are highly mesomorphic and of variable body size (e.g., American football), the measurement of body fat is extremely valuable when making body weight recommendations to players. For this reason, numerous studies have documented the %fat levels of football players (3,13,19,22,26,27) and have attempted to develop time and cost-effective methods other than hydrostatic weighing (HW) to estimate body fatness. Therefore, a method that could provide a quick and accurate assessment of %fat would be beneficial for monitoring large groups of athletes. Several studies have examined the accuracy of skinfolds (3,13,19,22,26,27), anthropometry (22,26,27), bioelectrical impedance (19), and near-infrared spectrophotometry (13) for conveniently assessing %fat in football players; however, these methods have greater prediction errors compared with laboratory assessments such as HW or dual energy x-ray absorptiometry (DXA).

A new air displacement plethysmograph, BOD POD Body Composition System (Life Measurement Instruments, Concord, CA), was developed for assessing %fat using a fast and relatively simply procedure (5). Although the use of air plethysmography for assessing body composition is not a new procedure, earlier methodologies yielded results that were not acceptable in terms of accuracy (6,8,24). From these earlier works, it became apparent that one had to account for changes in temperature, gas composition, and the impact of isothermal versus adiabatic conditions in the test chamber for valid results. Gundlach and Visscher (11) described a system that attempted to control for these conditions. Although their results had less error than previous methods, the meticulous nature of the procedure proved impractical for most applications. The BOD POD is the first commercial application of air plethysmography for densitometric analysis of body composition.

Air displacement plethysmography is a densitometric method that estimates the fat and fat-free mass via a two-component model. Historically, the criterion method for determining body fatness using a two-component model has been HW (9). Similar to HW that measures body volume by water displacement, this plethysmographic method measures body volume via air displacement. Once body volume is determined, body density (Db) is derived from the mass/volume ratio. The BOD POD requires less technical expertise than HW and the use of air instead of water as the fluid medium provides a more comfortable environment for most subjects. There are little data on the validity and reliability of the BOD POD. In one study of a heterogeneous group of 68 subjects, it was reported that the BOD POD was both reliable and valid in comparison to HW (16). However, Heyward (12) has recently stated that more research is warranted in diverse subgroups of the population to determine the validity of the BOD POD before it can replace HW. Because collegiate and professional sports teams (e.g., Buffalo Bills) are using the BOD POD, we felt it was important to investigate the validity of the device in athletic populations.

In muscular athletes such as football players, the density of the fat-free mass may be different than the underlying assumption of a two-component model where the density is assumed to be 1.1 g·cc−1(18,21). This constant is derived from assumed fractional components and respective densities comprising the fat-free mass (i.e., water, mineral, and residual) (2). Prior et al. (20) found that both athletes and nonathletes had a density and composition of the fat-free mass than differed from the assumed values used to estimate %fat from Db. Body composition assessment utilizing a three-component model (i.e., fat, mineral, and residual) would thus reduce the number of assumptions used when estimating %fat, resulting in a more accurate estimation compared to a two-component model (15). However, no study to date has compared the BOD POD with a multicomponent model of body composition. Therefore, the primary purpose of this study was to determine the accuracy of the BOD POD Body Composition System for determining Db and %fat in collegiate football players when compared with the criterion measure, HW. A secondary purpose was to compare body composition estimates from the BOD POD with those from DXA and a three-component model.

Back to Top | Article Outline

METHODS

Subjects. Sixty-nine Division IA collegiate football players participated in the study. The physical characteristics of the subjects are presented in Table 1 according to player position. Thirty-two of the subjects were Caucasian and 37 were African-American. All testing for this project was completed in conjunction with the routine battery of assessments that are performed annually for all football players (with the exception of DXA). A small group of subjects (N = 20) volunteered to participate in the DXA testing. Because DXA testing involves exposure to radiation, this aspect of the study was reviewed and approved by the Institutional Review Board. These subjects gave written consent in accordance with the policies established by the American College of Sports Medicine for the use of human subjects.

TABLE 1

TABLE 1

Procedures. All testing was completed during a single session with subjects reporting to the laboratory in a 3-h post prandial, postexercise, and normally hydrated state. Each subject completed the BOD POD test immediately followed by HW. Because of the impact of hair and clothing on the BOD POD assessments, all subjects wore a Lycra swim cap and Lycra shorts during the testing session. It should be noted that the Lycra shorts were a slight deviation from the manufacturer's recommendation of wearing a Lycra competition swimsuit. All DXA testing was conducted on the same day for those subjects who participated.

BOD POD. The BOD POD assesses %fat by measuring body volume using air displacement plethysmography. This system uses a single fiberglass structure that has two chambers separated via a fiberglass seat. The front chamber is the test chamber and the rear chamber is the reference chamber. There is a volume-perturbing element (diaphragm) mounted between the two chambers. As the diaphragm oscillates because of computer control, complementary volume perturbations occur in the two chambers. Using the application of basic gas laws within the chamber, pressure fluctuations that occur as a result of the volume changes are used to determine the chamber air volume. Chamber air volume is determined both with and without the subject seated in the test chamber. Thus, body volume is the difference between the two measures. The BOD POD was calibrated before each test using a two-point calibration method with volumes of 0 and 50 L (manufacturer's calibration cylinder).

An important consideration for the application of air displacement plethysmography is the recognition of how air behaves when compressed under isothermal versus adiabatic conditions. Therefore, during testing it is important to account for the impact that clothing, hair, skin surface area, and thoracic gas volume (VTG) all make on the measurement of the test chamber volume. Subjects are tested wearing minimal clothing and a swim cap to compress the hair. A correction is made for surface area artifact that is computed-based on height and mass of the subject (5). VTG is measured during the test or can be estimated based on the subject's height and age. VTG is the average volume of air in the lungs and thorax during normal tidal breathing. Measurement of VTG employs basically the same plethysmographic technique used for measuring body volume. To assess compliance with the VTG measurement procedure, a figure of merit is computed. Merit is a mathematical figure based on the relationship between the airway pressure curve and the chamber pressure curve (5). A merit less than 1.0 served as the criterion measure for validity. In addition, airway pressure is measured that reflects the maximum airway pressure generated during the "puffing" technique of the VTG measurement procedure. When the merit (≥1.0) or airway score (>35 mm Hg) is too high, the VTG measurements must be repeated. A more comprehensive description of the basis for the technology, system design, and the operating principles of the BOD POD has been previously published (5).

Db was computed using BOD POD determined body volume and mass. Db was used to predict %fat using the equation of Siri (21). To assess the reliability of the BOD POD, 15 subjects volunteered to repeat the BOD POD test. The second test was performed during the same session with approximately 5 min separating the tests.

Hydrostatic weighing. HW was performed using a custom built, stainless steel tank to measure body volume based on Archimedes' principle. Body mass in air was measured using a calibrated Chatillon electronic platform scale to the nearest 0.01 kg. Weight under water was measured at residual volume using a Chatillon autopsy scale to the nearest 0.025 kg. Residual volume was measured simultaneously using the standard oxygen-rebreathing nitrogen-dilution technique modified from Goldman and Buskirk (10). Nitrogen was measured using a Med Science 505 nitralizer. (St. Louis, MO). Db was computed using mass and volume with corrections for water density, residual lung volume, and gastrointestinal tract gas volume (0.1 L). Db was converted to %fat using the equation of Siri (21). Our previously published test-retest reliability of HW (N = 16) for assessing %fat was 0.998 with a technical error of measurement of 0.144% (23).

Dual-energy x-ray absorptiometry. In a subgroup of 20 subjects (mean ± SD age = 19.6 ± 1.1 yr; height = 182.6 ± 5.4 cm; mass = 90.8 ± 13.0 kg), total body mineral content and %fat were determined from whole body scans using a Lunar DPX-L DXA (Madison, WI; software version 1.3Z; medium mode, 3000 μA). It was not feasible to scan all subjects because of the limited dimensions of the scanning bed in relation to the size of most subjects and the constraints of time in the athletes' schedules. To ensure quality control, the DXA unit was calibrated on a daily basis using the standard calibration block provided by the manufacturer. The calibration block was made of a thermoplastic acrylic resin that contained three bone-equivalent chambers filled with hydroxyapatite. Our test-retest reliability of the DXA (N = 7) for assessing %fat was 0.995 with a technical error of measurement of 0.402%.

In addition to DXA determined values, %fat was computed based on a three-component model (3C; fat, mineral, and residual) from Lohman (15) and modified by Modlesky et al. (18): (Equation) where m is the mineral fraction of the body mass and Db is the density of the body from HW. This equation is based on the assumption that the water-to-protein ratio (73.8/19.4%) is constant. Because DXA was assumed to measure bone mineral ash, total body mineral was estimated by multiplying the ash by 1.2741 (2).

Statistical analysis. Statistical analyses were done with SAS for Windows version 6.12 (SAS Institute, Inc., Cary, NC). The reliability data for the BOD POD were analyzed using a paired t-test. Linear regression analysis and Pearson correlation coefficients were computed. In addition, technical error of measurement was computed, which reflects the standard deviation of the within-subject variance.

For %fat comparisons between HW and the BOD POD, the data were analyzed using a paired t-test. Linear regression analysis and Pearson correlation coefficients were computed. Agreement between the two methods for %fat estimations was determined using a Bland-Altman plot (1).

For the subgroup of subjects, the data were analyzed using a single factor ANOVA with repeated measures and Tukey post hoc tests. Linear regression analysis and Pearson correlation coefficients were computed. An alpha level of 0.05 was used for all significance testing.

Back to Top | Article Outline

RESULTS

A scatter plot of the reliability data for the BOD POD is presented in Figure 1. Mean (± SEM) %fat for trial 1 (13.5 ± 1.5%) was not significantly (P > 0.05) different than Trial 2 (13.4 ± 1.4%). The correlation coefficient between the two trials was 0.994 with a technical error of measurement of 0.448%.

Figure 1-A

Figure 1-A

To assess compliance with the procedure for measuring VTG, a figure of merit and an airway score were determined. Mean (±SEM) figure of merit (N = 69) was 0.36 ± 0.16, and the airway score was 19.5 ± 0.8. After repeated trials (≥5 trials), four subjects failed to achieve a figure of merit less than 1.0. Because these subjects failed to achieve the criterion for compliance, estimated VTG was used to determine the Db for these subjects for the purpose of comparing the BOD POD with HW. Their data were excluded from all analyses comparing measured and estimated VTG.

Relationship between %fat using estimated versus measured VTG is illustrated in Figure 2. There was a strong correlation (r = 0.986) between BOD POD determined %fat using estimated VTG and measured VTG. The slope of the regression equation was significantly (P < 0.05) less than 1.0. The mean (±SEM) %fat (N = 65) using measured VTG (15.1 ± 0.9%) was significantly (P < 0.05) less than %fat using estimated VTG (15.6 ± 0.9%). The correlation between estimated and measured VTG was 0.207 (P = 0.10) with a SEE of 0.65 L. The mean measured VTG of 3.998 L (range = 2.897-6.304 L), was significantly (P < 0.05) less than the estimated value mean, 4.342 L (range = 3.647-4.977 L).

Figure 2

Figure 2

The BOD POD and HW are both based on a two-component model where %fat is estimated from Db. In this study, mean (± SEM) Db measured with the BOD POD (1.064 ± 0.002 g·cc−1) was significantly greater (P < 0.05) than HW (1.060 ± 0.002 g·cc−1). Because comparisons among body composition techniques are typically made based on %fat, we elected to report our findings as %fat. However, any similarities or differences between the BOD POD and HW are due to the ability of each technique to accurately measure Db. A scatter plot illustrating the agreement between %fat assessed using HW and BOD POD is presented in Figure 3. Mean (± SEM) %fat determined using the BOD POD (15.1 ± 0.8%) was significantly less (P < 0.05) than found for HW (17.0 ± 0.8%). The slope for the regression equation was significantly (P < 0.05) less than 1.0. For over 60% of the subjects, the differences in %fat scores between HW and the BOD POD were within 2 percentage points. A scatter plot of the relation between %fat from HW and BOD POD values using estimated VTG is presented in Figure 4. Mean (± SEM) %fat from the BOD POD (15.6 ± 0.9%) based on estimated VTG was significantly lower (P < 0.05) than those found for HW (17.0 ± 0.8%).

Figure 3-A

Figure 3-A

Figure 4-A

Figure 4-A

To evaluate the individual agreement between HW and BOD POD determined %fat, a Bland-Altman plot is presented in Figure 5. There was no relationship (P > 0.05) between mean %fat and the difference between HW and the BOD POD. Therefore, the accuracy of the BOD POD to assess %fat does not appear to be influenced by the relative body fatness of the subject.

Figure 5-C

Figure 5-C

To further test the accuracy of the BOD POD, %fat was assessed using DXA and a 3C model (fat, mineral, and residual) in a subgroup of subjects (N = 20). Scatter plots illustrating the agreement between %fat determined from the BOD POD and %fat from DXA and the 3C model are presented in Figure 6. To facilitate comparison, data are also presented for HW for the same subjects. Mean (±SEM) %fat from the BOD POD (10.9 ± 1.0%) was significantly (P < 0.05) less than values for HW (13.3 ± 0.9%), DXA (12.9 ± 1.2%), and the 3C model (12.7 ± 0.8%). For 45-50% of the subjects, the differences in %fat scores between the three methods and the BOD POD were within 2 percentage points.

Figure 6-A

Figure 6-A

Back to Top | Article Outline

DISCUSSION

Body composition can be estimated from a two-component model based on Db. Use of this model to estimate %fat is dependent on accurate measures of Db. For densitometric methods, HW has historically been the standard for comparison (9). Because air displacement plethysmography (BOD POD) is a densitometric method, it would seem prudent to validate this method using HW. Because both methods measure Db, differences in %fat would reflect Db variations. In this study, we used %fat from HW as a criterion measure to validate %fat values in a group of 69 collegiate football players. In addition, a subgroup of 20 subjects were studied to validate %fat values from the BOD POD against DXA and 3C determined %fat values based on three-component models.

In a group of 15 subjects, test-retest BOD POD determinations of %fat were made to assess reliability. There was excellent agreement between the duplicate measures. These data agree with the reliability data previously reported for the BOD POD (16). They reported a between-trial SD of 0.4%fat, which is the same as found in the present study. In comparison with other measures of %fat, we found a strong correlation coefficient between trials for the BOD POD that was comparable to our data for HW and DXA. In addition, the technical error of measurement for the BOD POD was low with a value of 0.402% that was consistent with our data and those reported for HW (0.42%) and skinfolds (0.61%) (28). Based upon the findings of this study, the test-retest reliability for %fat assessment using the BOD POD was very good.

The VTG must be considered for the measurement of Db. The BOD POD is designed to allow for measurement of VTG or estimation based on subject height and age; therefore, it was of interest to compare measured versus estimated VTG and to determine the impact of VTG on %fat values. For the majority of the subjects, compliance with the VTG measurement procedure was achieved within three trials. After five repeated trials, four subjects were unable to comply with the procedures for VTG measurement and were excluded from this portion of the data analyses. Noncompliance in these subjects was probably related to an inability to perform the "puffing" technique without simultaneously executing a Valsalva maneuver. Surprisingly, there was considerable variability between measured and estimated VTG. On the average, measured VTG values were lower than the predicted values. In addition, the relation between measured and estimated VTG was weak (r = 0.207) with a large SEE of 0.65 L. The impact of VTG on Db values was less than 0.002 g·cc−1, which corresponded to 0.5%fat with estimated VTG yielding higher values compared with measured VTG. McCrory and colleagues (17) compared estimated and measured VTG in a heterogeneous population and found no significant difference between the methods, with a mean difference of 54 mL, which resulted in a %fat difference of only 0.2%. Although we found a significant difference between %fat using measured and predicted VTG, unlike McCrory et al. (17), the effect was still in the direction observed by these investigators (i.e., when VTG is over predicted, %fat is overestimated). It is unclear what accounts for the differences between the two studies, especially because our subjects met the manufacturer's criteria for compliance with the VTG procedure. In the present study, it is apparent that using estimated VTG slightly enhanced the agreement between the BOD POD and our other measures of %fat. This is not meant to imply that estimated VTG is more accurate than measured VTG, but only to recognize that the impact on %fat values is quite small. When compared with HW, the use of an estimated VTG contributes less potential error (less than one-half the effect) on measurement of %fat via BOD POD than observed for the same error using an estimated residual volume for HW (17). Because estimated VTG tended to enhance the agreement and there is an additional cost for the disposable breathing tube along with more time required for the procedure, we recommend the use of predicted VTG when using the BOD POD to track body composition in athletic populations.

To validate the accuracy of the BOD POD to estimate %fat, comparisons were made with HW. There was a strong relationship (r2 = 0.89) between the BOD POD and HW determined %fat, but the BOD POD systematically yielded lower %fat values. McCrory et al. (16) also reported a strong agreement between the BOD POD and HW determined %fat in a heterogeneous group of subjects. The r2 was 0.93 with an intercept and slope that were not significantly different from 0 and 1, respectively. Although our r2 and slope were quite similar to those reported by McCrory et al. (16), our intercept (3.33 vs 1.86) was slightly larger. To our knowledge, the only differences between the studies were the type of subjects used and the timing of residual lung volume measurement during HW. In the present study, the subjects had a more similar somatotype (football players); consequently, they were all male, younger, heavier, and slightly taller than subjects from the McCrory et al. study (16). The manufacturer designed the system to accommodate large individuals up to 165 kg of mass and 218 cm of height. We found no relation (r = −0.16, P = 0.2) between body mass and the agreement between methods, suggesting that body size alone does not explain the divergent findings. During the present study, residual lung volume was measured simultaneously during HW, whereas in the McCrory et al. (16) study residual volume was measured in a separate procedure on land. We cannot determine which of these variables might be more influential with regard to the differences found between the studies. A difference in clothing from the manufacturer's suggestion may also be involved. However, it was not practical to expect subjects to wear a tight-fitting Lycra competition swim suit for all tests. It is unclear in the McCrory et al. (16) study whether the male subjects complied with this clothing specification.

To further assess the validity of the BOD POD to assess %fat, the SEE was computed. Lohman (14) suggested that a SEE less than 3% indicates good accuracy of a new method. McCrory et al. (16) reported a SEE of 1.81%, which is slightly less than the 2.2% found in our study. Our SEE was quite low in comparison with other field methods of estimating %fat in football players, such as skinfolds (2.3-3.5%), bioelectrical impedance (3.9-6.0%), and near-infrared spectrophotometry (4.1%) (3,13,19).

The agreement between %fatHW and %fatBOD POD was determined by examining the differences among individual scores. The standard deviation of the differences was 2.2% with a range of −5.4 to 7.5%. McCrory et al. (16) reported that 75% of their subjects had %fat values that were within ± 2% of the mean difference (0.3%) between methods. Our mean difference was much larger (2.0%), and approximately 61% of our subjects fell within ± 2%. Based on the Bland-Altman plot, there was no systematic difference in the agreement between the two methods across a range of body fatness. These findings are consistent with those reported by McCrory and colleagues (16).

Because the BOD POD measures Db, we chose to use HW as our criterion that measures Db and is based on a two-component model. However, multi-component models that measure mineral may improve on the accuracy of a two-component model. Friedl et al. (7) reported that %fat from DXA (18.0%) and 3C (18.3%) was slightly larger than values for HW (17.3%). To further assess the validity of the BOD POD, comparisons were made with %fat values from DXA and 3C in a subgroup of 20 subjects. Our %fat via HW was similar to DXA and slightly higher than 3C, but the %fat via the BOD POD was lower than DXA and 3C. This is consistent with the total subject population, where the %fat from BOD POD was lower than HW. Based on the results from this subgroup, the BOD POD also yielded lower estimates of %fat when compared with other measures of body composition (DXA and 3C).

In conclusion, Db values measured using the BOD POD were systematically higher than the criterion HW determined values, thus yielding lower %fat values for the BOD POD. This trend was consistent in the subsample of athletes when comparing %fat for the BOD POD to DXA and 3C determined values. Assessment of body composition using the BOD POD is very reliable and has these advantages: easy operation, relatively quick procedure, fast results, and comfortable for the subject. However, some limitations include the rigid clothing specifications, chamber size (i.e., the morbidly obese might not be accommodated), and potential difficulty in obtaining VTG. In this study of collegiate football players, %fat scores were observed to be slightly lower than values for HW, DXA, and 3C. Thus, at this time, it is premature to recommend replacement of HW with air displacement plethysmography when determining Db for research or clinical purposes. More research is clearly needed to determine the validity of this device in additional subgroups of the population.

Back to Top | Article Outline

REFERENCES

1. Bland, J. M., and D. G. Altman. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1:307-310, 1986.
2. Brozek, J., F. Grande, J. T. Anderson, and A. Keys. Densitometric analysis of body composition: revision of some quantitative assumptions. Ann. N. Y. Acad. Sci. 110:113-140, 1963.
3. Clark, R. R., J. M. Kuta, and J. C. Sullivan. Cross-validation of methods to predict body fat in African-American and Caucasian collegiate football players. Res. Q. Exerc. Sport 65:21-30, 1994.
4. Cureton, K. J., P. B. Sparling, B. W. Evans, S. M. Johnson, U. D. Kong, and J. W. Purvis. Effect of experimental alterations in excess weight on aerobic capacity and distance running performance. Med. Sci. Sports 10:194-199, 1978.
5. Dempster P., and S. Aitkens. A new air displacement method for the determination of human body composition. Med. Sci. Sports Exerc. 27:1692-1697, 1995.
6. Fomon, S. L., R. L. Jesson, and G. M. Owen. Determination of body volume from infants by a method of helium displacement. Ann. N. Y. Acad. Sci. 110:80-90, 1963.
7. Friedl, K. E. J. P. Deluca, L. J. Marchitelli, and J. A. Vogel. Reliability of body-fat estimations from a four-compartment model using density, body water, and bone mineral measurements. Am. J. Clin. Nutr. 55:764-770, 1992.
8. Gnaedinger, R. H., E. P. Reinike, A. M. Pearson, W. D. Van Huss, J. A. Wessel, and H. J. Montoye. Determination of body density by air displacement, helium dilution, and underwater weighing. Ann. N. Y. Acad. Sci. 110:96-108, 1963.
9. Going, S. Densitometry. In: Human Body Composition, A. F. Roche, S. B. Heymsfield, and T. G. Lohman (Eds.). Champaign, IL: Human Kinetics, 1996, pp.3-23.
10. Goldman, R. F., and E. R. Buskirk. Body volume measurement by underwater weighing: description of a method. In: Techniques for Measuring Body Composition, J. Brozek and A. Henschel (Eds.). Washington, DC: National Academy of Science, 1961, pp. 77-89.
11. Gundlach, B. L., and G. J. W. Visscher. The plethysmometric measurement of total body volume. Hum. Biol. 58:783-799, 1986.
12. Heyward, V. H. Practical body composition assessment for children, adults, and older adults. Int. J. Sports Nutr. 8:285-307, 1998.
13. Houmard, J. A., R. G. Israel, M. R. McCammon, K. F. O'Brien, J. Omer, and B. S. Zamora. Validity of a near-infrared device for estimating body composition in a college football team. J. Appl. Sport Sci. Res. 5:53-59, 1991.
14. Lohman, T. G. Dual energy X-ray absorptiometry. In: Human Body Composition, A. F. Roche, S. B. Heymsfield, and T. G. Lohman (Eds.). Champaign, IL: Human Kinetics, 1996, pp. 63-78.
15. Lohman, T. G. Applicability of body composition techniques and constants for children and youths. In: Exercise and Sport Sciences Reviews, K. Pandolf (Ed.). New York: Macmillan, 1986, pp. 325-357.
16. McCrory, M. A., T. D. Gomez, E. M. Bernauer, and P. A. Mole. Evaluation of a new air displacement plethysmograph for measuring human body composition. Med. Sci. Sports Exerc. 27:1686-1691, 1995.
17. McCrory, M. A., P. A. Molé, T. D. Gomez, K. G. Dewey, and E. M. Bernauer. Body composition by air-displacement plethysmography by using predicted and measured thoracic gas volumes. J. Appl. Physiol. 84:1475-1479, 1998.
18. Modlesky, C. M., K. J. Cureton, R. D. Lewis, B. M. Prior, M. A. Sloniger, and D. A. Rowe. Density of the fat-free mass and estimates of body composition in male weight trainers. J. Appl. Physiol. 80:2085-2096, 1996.
19. Oppliger, R. A., D. H. Nielsen, A. C. Shetler, E. T. Crowley, and J. P. Albright. Body composition of collegiate football players: bioelectrical impedance and skinfolds compared to hydrostatic weighing. J. Sports Phys. Ther. 15:187-192, 1992.
20. Prior, B. M., E. M. Evans, C. M. Modlesky, M. A. Sloniger, R. D. Lewis, and K. J. Cureton. Density and composition of the fat-free mass in athletes. Med. Sci. Sport Exerc. 29:S54, 1997.
21. Siri, W. E. Body composition from fluid spaces and density: analysis of methods. In: Techniques for Measuring Body Composition, J. Brozek and A. Henschel (Eds.). Washington, DC: National Academy of Science, 1961, pp. 223-244.
22. Smith, J. F., and E. R. Mansfield. Body composition prediction in university football players. Med. Sci. Sports Exerc. 16:398-405, 1984.
23. Sparling, P. B., M. Millard-Stafford, L. B. Rosskopf, L. J. Dicarlo, and B. T. Hinson. Body composition by bioelectrical impedence and densitometry in black women. Am. J. Hum. Biol. 5:111-117, 1993.
24. Taylor, A., Y. Aksoy, J. W. Scopes, G. Du Mont, and B. A. Taylor. Development of an air displacement method for whole body volume measurement in infants. J. Biomed. Engl. 7:9-17, 1985.
25. Vanderburgh, P. M., and T. Edmonds. The effect of experimental alterations in excess mass on pull-up performance in fit young men. J. Strength Cond. Res. 11:230-233, 1997.
26. White, J., J. L. Mayhew, and F. C. Piper. Prediction of body composition in college football players. J. Sports Med. 20:317-324, 1980.
27. Wickkiser, J. D., and J. M. Kelly. The body composition of a college football team. Med. Sci. Sports 7:199-202, 1975.
28. Wilmore, K. M., P. J. McBride, and J. H. Wilmore. Comparison of bioelectrical impedance and near-infrared interactance for body composition assessment in a population of self-perceived over-weight adults. Int. J. Obes. 18:375-381, 1994.
Keywords:

AIR DISPLACEMENT PLETHYSMOGRAPHY; %FAT; DXA; BODY COMPOSITION; THREE-COMPONENT MODEL

© 1999 Lippincott Williams & Wilkins, Inc.