SPECIAL COMMUNICATIONSMethods

Comparison of the BOD POD with the four-compartment model in adult females

DAVID A. FIELDS, G. DENNIS WILSON, L. BRUCE GLADDEN, GARY R. HUNTER, DAVID D. PASCOE, andMICHAEL I. GORAN

Division of Physiology and Metabolism, Department of Nutrition Sciences, and Department of Human Studies, Universityof Alabama at Birmingham, Birmingham, AL; Department of Health and Human Performance, Auburn University,Auburn, AL; Department of Preventive Medicine, Institute for Preventive Research, University of Southern California, LosAngeles, CA; and Department of Internal Medicine, Center for Human Nutrition, Washington University, St. Louis, MO

ABSTRACT

FIELDS, D. A., G. D. WILSON, L. B. GLADDEN, G. R. HUNTER, D. D. PASCOE, and M. I. GORAN. Comparison of the BODPOD with the four-compartment model in adult females. Med. Sci. Sports Exerc., Vol. 33, No. 9, 2001, pp. 1605–1610. Purpose: Thisstudy was designed to compare the accuracy and bias in estimates of total body density (Db) by hydrostatic weighing (HW) and theBOD POD, and percent body fat (%fat) by the BOD POD with the four-compartment model (4C model) in 42 adult females.Furthermore, the role of the aqueous and mineral fractions in the estimation of body fat by the BOD POD was examined. Methods:Total body water was determined by isotope dilution (2H20) and bone mineral was determined by dual-energy x-ray absorptiometry.Db and %fat were determined by the BOD POD and HW. The 4C model of Baumgartner was used as the criterion measure of bodyfat. Results: HW Db (1.0352 g·cm�3

) was not statistically different (P � 0.35) from BOD POD Db (1.0349 g·cm�3). The regression

between Db by HW and the BOD POD significantly deviated from the line of identity (Db by HW � 0.90 � Db by BOD POD �

0.099; R2 � 0.94). BOD POD %fat (28.8%) was significantly lower (P � 0.01) than %fat by the 4C model (30.6%). The regressionbetween %fat by the 4C model and the BOD POD significantly deviated from the line of identity (%fat by 4C model � 0.88 � %fatby BOD POD � 5.41%; R2 � 0.92). BOD POD Db and %fat showed no bias across the range of fatness. Only the aqueous fractionof the fat-free mass (FFM) had a significant correlation with the difference in %fat between the 4C model and the BOD POD.Conclusion: These data indicate that the BOD POD underpredicted body fat as compared with the 4C model, and the aqueous fractionof the FFM had a significant effect on estimates of %fat by the BOD POD. Key Words: BODY COMPOSITION, PLETHYSMOG-RAPHY, BODY DENSITY

The ability to accurately assess the level of fatness inpersons has major health consequences because ofthe association between obesity and conditions such

as hypertension, insulin resistance, dyslipidemia, and hy-perinsulinemia (26). A multicompartment approach wouldbe ideal in determining body composition because the in-dividual constituents of the fat-free mass (FFM) are mea-sured by independent techniques (1,19). However, the use ofa multicompartment approach is impractical for most re-search and clinical settings because of cost constraints. As aresult, most laboratories have used hydrostatic weighing(HW). Unfortunately, HW is impractical in certain popula-tions (children, obese, elderly), and for many subjects it isconsidered to be time consuming (typically six to eight trialsof ~30 min) and intimidating (water submersion), thus lim-

iting its usefulness. In 1995, Dempster and Aitkens (7)demonstrated an alternative technique (BOD POD) derivedfrom plethysmographic principles developed by others(9,12,14,15). Plethysmography determines body volume onthe basis of the pressure/volume relationship. Boyle’s lawexplains this relationship in an isothermal testing chamberas: PV � k, where k is the proportionality constant (36). Ifthe testing chamber temperature is not constant (adiabatic),Poisson’s law describes the pressure/volume relationship:PV� � k, where � is the ratio of the specific heat of the gasat constant pressure to that at constant volume (32,33). Theliterature comparing BOD POD and HW data is equivocal,as four studies show no significant difference between theline of identity and the regression between HW and theBOD POD (3,11,24,25), whereas two studies report a sig-nificant difference between the line of identity and theregression between the BOD POD and HW (5,22). To ourknowledge, no study has attempted to validate the BODPOD with the four-compartment model (4C model) inadults. Therefore, the purpose of this study was threefold:

0195-9131/01/3309-1605/$3.00/0MEDICINE & SCIENCE IN SPORTS & EXERCISE®

Copyright © 2001 by the American College of Sports Medicine

Submitted for publication June 2000.Accepted for publication November 2000.

1605

first, to compare body density (Db) from the BOD PODagainst Db from HW; second, to compare percent body fat(%fat) by the BOD POD against %fat by the 4C model; andthird, to investigate how variations in the aqueous andmineral factions of the FFM affect %fat estimates of theBOD POD.

METHODS

Experimental Design

An overview of the study design is presented in Table 1.All subjects came to the Division of Physiology and Me-tabolism at the University of Alabama at Birmingham(UAB) at 6:00 a.m. in a fasted state. After a baseline urinesample was obtained, the subjects were asked to ingest acontainer that had ~10 g of deuterium. The %fat was thendetermined by either dual-energy x-ray absorptiometry(DXA) or the BOD POD (the order was randomized). Afterthe first test (either DXA or BOD POD, depending on theorder), a 45-min urine sample was obtained. HW was al-ways performed after the DXA and the BOD POD becausepreliminary data have suggested that an increase in bodytemperature and moisture can affect %fat estimates by theBOD POD. After the completion of HW, the subjects wereasked to wait quietly so a 3- and 4-h urine sample could beobtained.

The study sample included 42 adult females, of whom 39were Caucasian and 3 were African American from theBirmingham area. All subjects gave their informed consentbefore participation in the study. Subjects who indicatedapprehension about head submersion in the hydrostaticweighing tank before testing were excluded from the study.Approval for the use of human subjects was obtained fromthe Institutional Human Subject Review Board from bothAuburn University and UAB.

Protocol

Assessment of bone mineral content. DXA wasused to measure bone mineral content (BMC) (LunarDPX-L densitometer, Lunar Radiation Corp., Madison,WI). The total dose of radiation to the subject was less thana typical chest radiograph. Whole-body scans were analyzedusing the adult medium mode for all subjects (DPX-L ver-sion 3.6z). The repeat measures between days for BMCderived from DXA in eight healthy females had an intraclasscorrelation of R � 0.99 and a standard deviation of 48 g.

Assessment of total body density by HW. Db wasestimated by HW with simultaneous measurement of residual

lung volume by using the closed-circuit oxygen dilution tech-nique (35). Underwater weight was measured to the nearest50 g in a stainless steel tank in which the subject was sus-pended from a LCL 20 Shear Beam Load Cell (Omega, Stan-ford, CT) calibrated from 0 to 10,000 g. After one practice trial,underwater weight and residual lung volume were measuredsimultaneously in five consecutive trials. The average of mul-tiple trial densities within 0.001 g·cm�3

were used. The repeatmeasures between days for Db derived from HW in eighthealthy females had an intraclass correlation of R � 0.99 anda standard deviation of 0.002 g·cm�3

.Assessment of total body density and %fat by

plethysmography (BOD POD). Whole-body air dis-placement was evaluated with the BOD POD version 1.69(Body Composition System, Life Measurement Instruments,Concord, CA). The BOD POD is a single, “egg-shaped” unitconsisting of two chambers: a testing chamber where thesubject sits, and a reference chamber where the breathingcircuit, pressure transducers, and electronics are stored (7).The testing procedure involved several steps. First, calibra-tion was conducted before the subject’s entry into the BODPOD. Calibration involved the computation of the ratio ofthe pressure amplitudes (reference chamber and testingchamber) for an empty chamber and a known volume(49.860 L). The software calculated a regression equationbetween the testing chamber volume and the ratio of thepressure amplitudes (7). Essentially, the relationship is lin-ear for any testing chamber volume and the ratio of thepressure amplitudes (7,33). After the calibration was com-pleted and the procedures fully explained, the subject wasgiven a nose clip and swim cap (worn to minimize isother-mal air trapped within the hair). The subject entered theBOD POD for two trials of approximately 45 s each. Duringthis stage, the subject’s uncorrected body volume (Vbraw)was determined with the testing chamber door being openedbetween trials. If both volumes were within 150 mL, thenthe two trials were averaged. However, if the volumes werenot within 150 mL, a third trial was performed and the twovolumes that were the closest were averaged. The last stepinvolved the measurement of the thoracic gas volume (VTG).This stage required the subject to sit quietly in the BODPOD and breathe through a disposable tube and filter thatwas connected to the reference chamber in the rear of theBOD POD. After four or five normal breaths, the airwaywas occluded during midexhalation and the subject wasinstructed to make two quick light pants. VTG was consid-ered successful when the four criteria were met. First, themerit had to be � 1. The merit is a theoretical value thatdemonstrates subject compliance during the measurement ofthe VTG. A merit of � 1 indicates perfect agreement be-tween chamber pressure and the airway pressure in the tube,with a merit � 1 demonstrating poor compliance (usuallybecause of leaking of air around the mouth) (7). Second,airway pressure had to be below 35 cm H2O. Third, tidalvolume had to be between 0.40 and 0.70 L. Finally, themeasured VTG had to be within 0.70 L of the predicted VTG

(27). Db from the BOD POD was calculated as follows: Db� M/(Vbraw � 0.40VTG � SAA), where SAA and 0.40VTG

TABLE 1. Testing schedule.

Activity Time (Begin) Time (End)

Collect baseline urine 6:00 a.m. 6:05 a.m.Administration of deuterium 6:10 a.m. 6:15 a.m.DXA 6:20 a.m. 6:50 a.m.Collect �45-min urine sample 6:55 a.m. 7:05 a.m.BOD POD 7:15 a.m. 7:45 a.m.Hydrostatic weighing 8:00 a.m. 8:45 a.m.Collect 1 urine 9:10 a.m. 9:15 a.m.Collect 1 urine 10:10 a.m. 10:15 a.m.

1606 Official Journal of the American College of Sports Medicine http://www.acsm-msse.org

are used to correct for the isothermic conditions within thechamber and M is the mass of the subject. The %fat wasthen calculated using the Siri equation (31). The repeatmeasures between consecutive days for Db derived from theBOD POD in eight healthy females (all wearing a one-pieceswimsuit) had an intraclass correlation of R � 0.98 and astandard deviation of 0.006 g·cm�3

.Assessment of total body water by isotope dilu-

tion. Total body water (TBW) was determined using 2H20(deuterated water). The dosing procedure followed the tech-nique developed by Schoeller (29) and Schoeller et al. (30).After an initial urine sample was collected (baseline), thesubject was weighed in minimal clothing and a 10-g dose of2H20 was given. Approximately 45 min later the subjectvoided, and 3- and 4-h postdosing urine samples were thenobtained. Deuterium was prepared by the zinc reductionmethod described by Kendall and Coplen (21). All sampleswere analyzed on an OPTIMA (Micromass, Inc., Beverly,MA) mass spectrometer in triplicate. The calculation of2H2O dilution space was calculated from the enrichment of2H2O in the body at zero time by extrapolation of the logenrichment versus time plots back to zero time (6) using thefollowing equation (29): Dilution space (L) � d/20.02 ·18.02 · 1/R · E, where d is grams of 2H2O given, R is thestandard ratio of 2H/1H (0.00015576), and E is enrichmentof 2H2O at the extrapolated zero time (the percent abovebackground). Total body water was determined by takingthe mean of the zero-time isotope dilution space for 2H2Oand dividing by 1.04 (to correct for the exchange withnonaqueous tissues) (30).

Calculation of %fat using the 4C model. The %fatby the 4C model was determined by the Baumgartner et al.(1) equation: % body fat � 205 (1.34/Db � 0.35 A � 0.56M � 1), where Db is total body density from HW, and A andM are the fractions of body mass that are aqueous andmineral, respectively. Although the derivation of this modelcame from subjects 65–94 yr of age, 4C model equations are(almost) free of assumptions and are theoretically derived. Itwould be almost impossible to derive new 4C model equa-tions for every population group, and we cannot think of anyreason why the equation by Baumgartner et al. (1) would notbe applicable to our study population.

Statistical Analysis

To determine the accuracy of Db and %fat estimates bythe BOD POD, linear regression analysis was used. Theaccuracy of Db by the BOD POD was determined by Dbmeasured by the BOD POD as the independent variable andDb by HW as the dependent variable. The accuracy of %fatby the BOD POD was evaluated with %fat by the BODPOD as the independent variable and %fat measured by the4C model as the dependent variable. If the slope was notsignificantly different from 1 and the intercept not signifi-cantly different from 0, the estimates of Db and %fat de-rived from the BOD POD were not considered significantlydifferent from the dependent variable. Potential bias in Dband %fat estimates by the BOD POD were examined using

residual plots. This analysis examines the discrepancy be-tween techniques across the range of fatness. Also, the R2

and the standard error of the estimate (SEE) for Db and %fatby the BOD POD were calculated. Paired t-tests were usedto compare group means. Additionally, the difference be-tween estimates of %fat by the BOD POD and estimates of%fat by the 4C model were calculated and regressed on theaqueous (AFFM) and mineral (MFFM) fractions of the DXAFFM. The pure or total error between the BOD POD andHW with the 4C model was also calculated:

�� �Y1–Y2)2

n(1)

where Y1 is %fat assessed by the 4C model and Y2 is %fatassessed by either the BOD POD or HW (20). Statisticalsignificance was set at P � 0.05.

RESULTS

The physical characteristics and group mean estimates of%fat by the different techniques for the 42 females arepresented in Table 2. The AFFM is similar to what wasreported by Hewitt et al. (16) (71%) and Bergsma-Kadijk etal. (2) (72%), but not work by Visser et al. (34) (74%). TheMFFM in this study (5.6%) is lower than the 6.8% and 7.5%reported in the same population (2,34).

FIGURE 1—The regression of Db by the BOD POD against Db by HWin the total sample of 42 subjects. The dotted line is the line of identity(regression slope � 1 and regression intercept � 0). The slope wassignificantly different from 1 and the intercept was significantly dif-ferent from 0.

TABLE 2. Descriptive characteristics for subjects and body composition variables.

Variable Mean � SD Range

Age (yr) 32.8 11.0 19–54Body weight (kg) 63.4 13.0 43.6–115.0Height (cm) 167.5 7.0 147.2–183.4Body density (g�cm–3)

HW 1.0352 0.019 0.992–1.073BOD POD 1.0349 0.021 0.993–1.094

4C model % fat 30.6 9.2 15.9–50.5BOD POD % fat* 28.8 10.1 5.2–51.3Total body water (L) 30.8 4.4 22.9–42.9Bone mineral (kg) 2.5 0.5 1.8–3.6Mineral fraction of FFM 0.056 0.005 0.047–0.072Aqueous fraction of FFM 0.714 0.042 0.639–0.821

* Significantly different from 4C model % fat (P � 0.01).

FOUR-COMPARTMENT MODEL IN ADULT FEMALES Medicine & Science in Sports & Exercise� 1607

The total error for both the BOD POD and HW was 2.3%body fat and 2.4% body fat, respectively. Paired t-test re-sults revealed no significant difference (P � 0.35) betweenHW Db (1.0352 g·cm�3

) and BOD POD Db (1.0349g·cm�3

); this represents a difference of 0.14% body fat. Theregression for Db by HW versus Db by the BOD PODsignificantly deviated from the line of identity because of anintercept significantly different from 0 and a slope signifi-cantly different from 1, where Db by HW � 0.099 � 0.90� (Db by BOD POD) (Fig. 1). Db by the BOD PODexplained 94% of the variance in Db by HW, whereas theSEE was 0.005 g·cm�3

. A residual plot was performed todetermine if bias existed between the Db by the BOD PODand the Db by HW (Fig. 2). The Db by the BOD PODshowed no bias across the range of fatness as indicated bya nonsignificant correlation (r � 0.26; P � 0.09). Theaccuracy of the %fat from the BOD POD was examined byregression analysis in each individual. The regression of%fat assessed by the 4C model versus %fat assessed by theBOD POD significantly deviated from the line of identitybecause of an intercept that was significantly different from

0 and a slope significantly different from 1, where %fat bythe 4C model � 5.41 � 0.88 � (%fat from the BOD POD)(Fig. 3). The %fat by the BOD POD explained 95% of thevariance in %fat by the 4C model, whereas the SEE was2.7% body fat. A residual plot was performed to determineif bias existed between estimates in %fat between the BODand the 4C model (Fig. 4). BOD POD %fat did not exhibitany bias across the range of body fatness as indicated by anonsignificant correlation (r � 0.27; P � 0.09).

The magnitudes of the difference between the BOD PODand 4C model estimates of %fat had a significantly positivecorrelation (r � 0.51; P � 0.01) with the fraction of FFMthat was water (AFFM), as shown in Figure 5. However, thefraction of the FFM that was mineral (MFFM) had a negativeassociation with the difference between the BOD POD and4C model estimates of %fat (r � �0.17; P � 0.27) (Fig. 6).Additionally, HW showed the same trends (not presented).

DISCUSSION

This study examined the accuracy of and bias in mea-surements of %fat as assessed by the BOD POD relative to

FIGURE 3—The regression of %fat by the BOD POD against %fat bythe 4C model in the total sample of 42 subjects. The dotted line is theline of identity (regression slope � 1 and regression intercept � 0). Theslope was significantly different from 1 and the intercept was signifi-cantly different from 0.

FIGURE 4—The residual plot where the middle dashed line representsthe mean difference between %fat by the BOD POD � %fat by the 4Cmodel and the upper and lower dashed lines represent � 2 SD from themean. No bias between the techniques was observed, as indicated by anonsignificant P value.

FIGURE 5—The regression between the difference in the %fat by theBOD POD and estimates of %fat by the 4C model against the aqueousfraction of the FFM. The solid line represents the regression for theBOD POD (r � 0.51; P < 0.01).

FIGURE 2—The residual plot where the middle dashed line representsthe mean difference between Db by BOD POD � Db by HW and theupper and lower dashed lines represents � 2 SD from the mean. No biasbetween the techniques was observed, as indicated by a nonsignificantP value.

1608 Official Journal of the American College of Sports Medicine http://www.acsm-msse.org

the 4C model in 42 females. Additionally, the relative frac-tions of the FFM that are water and mineral and its effect on%fat estimates using the BOD POD was investigated. Thisis significant because this is the first study to compare theBOD POD with the 4C model in adults. However, Fieldsand Goran (10) demonstrated that the BOD POD couldaccurately, precisely, and without bias estimate %fat inchildren 9–14 yr old using the 4C model as the criterionmethod.

Four studies have demonstrated excellent agreement be-tween the BOD POD and HW, with the mean difference in%fat ranging from 0.05–1.0% (3,11,24,25). However, threestudies have shown that the BOD POD significantly under-predicted %fat by ~2% as compared with HW (5,8,22).Caution should be used in interpreting these results(3,5,8,11,22,24,25) because the BOD POD and HW arederived from two-compartment model (2C model) assump-tions; thus, the BOD POD was not validated against amulticompartment approach.

The major finding of this study is that the BOD PODunderestimated %fat in females of varying age and fatnesswhen compared with the 4C model. Our data comparing theestimates of %fat by the BOD POD with the 4C model anddata making the same comparisons with HW and the 4Cmodel are somewhat similar (1,4,13,34). This would bereasonable because both methods measure total body den-sity; essentially the only difference is that HW measuresbody volume by water displacement and the BOD PODmeasures body volume by air displacement. Group meanestimates in %fat by the BOD POD were 1.8% lower thanthe 4C model, whereas regression analysis indicated a mod-erate degree of agreement (2.7%) on an individual basis(SEE). Others have found group mean estimates in fatnessbetween HW and the 4C model to be relatively small

(~1.5%) and %fat estimates on an individual basis to besomewhat higher (~4%) (4,13,34). Additionally, studies inchildren (16,28) and older females have shown these groupmean differences to be somewhat larger (6–10%). Thislarge discrepancy in %fat in these two populations mostlikely is attributable to deviations from the basic assump-tions in the 2C model (1,16,17,23). To date, this is the firststudy to compare BOD POD estimates of %fat with the 4Cmodel in an adult population.

The difference in estimates of %fat between the BODPOD and the 4C model were significantly related to theaqueous fraction of the FFM. Visser et al. (34) and Baum-gartner et al. (1) found the measurement of the aqueous andmineral fractions of the FFM increased the accuracy of the4C model while taking into account biological variationresulting from aging, disease, ethnicity, and training status.However, the mineral fraction in those two studies contrib-uted minimally to the model. Our data showed a statisticallysignificant positive relationship between the difference in%fat by the BOD POD and the 4C model and the aqueousfraction of the FFM. This is in agreement with other studies(2,16), but not work by Baumgartner et al. (1) and Visser etal. (34). Our data demonstrated a negative relationship be-tween the mineral fraction of the FFM and the differencebetween %fat by the BOD POD and the 4C model; this hasbeen observed by others making the comparison with HWand the 4C model (2). However, not all studies have re-ported a negative relationship (1,34). The role of the aque-ous and mineral fractions of the FFM and their relationshipsin explaining differences between %fat by a 2C model andthe 4C model are debatable; however, it would appear the4C model does increase accuracy in estimating body com-position in a wide range of populations (1,4,13,18,26,33).

CONCLUSION

The BOD POD demonstrated a moderate degree of indi-vidual variation in %fat estimates as compared with the 4Cmodel in adult females. Additionally, %fat estimates by theBOD POD were affected by the aqueous fraction of theFFM, thus highlighting the importance of a multicompart-ment approach in the evaluation of body compositionanalysis.

We would like to thank the Division of Physiology and Metabo-lism, Department of Nutrition Sciences at the University of Alabamaat Birmingham for making available lab equipment and labresources.

Address for correspondence: David A. Fields, Ph.D., Departmentof Internal Medicine, Center for Human Nutrition, Washington Uni-versity, Campus Box 8031, St. Louis, MO 63110-1010; E-mail:dfields@im.wush.edu.

REFERENCES

1. BAUMGARTNER, R. N., S. B. HEYMSFIELD, and S. LICHTMAN. Bodycomposition in elderly people: effect of criterion estimates onpredictive equations. Am. J. Clin. Nutr. 53:1345–1353, 1991.

2. BERGSMA-KADIJK, J. A., B. BAUMEISTER, and P. DEURENBERG. Mea-surement of body fat in young and elderly women: comparison

between a four-compartment model and widely used referencemethods. Br. J. Nutr. 75:649–657, 1996.

3. BIAGGI, R. R., M. W. VOLLMAN, M. A. NIES, et al. Comparisonof air-displacement plethysmography with hydrostatic weigh-ing and bioelectrical impedance analysis for the assessment of

FIGURE 6—The regression between the difference in the %fat by theBOD POD and estimates of %fat by the 4C model against the mineralfraction of the FFM. The solid line represents the regression for theBOD POD (r � �0.17; P � 0.27).

FOUR-COMPARTMENT MODEL IN ADULT FEMALES Medicine & Science in Sports & Exercise� 1609

body composition in healthy adults. Am. J. Clin. Nutr. 69:898 –903, 1999.

4. CLASEY, J. L., J. A. KANALEY, L. WIDEMAN, et al. Validity ofmethods of body composition assessment in young and older menand women. J. Appl. Physiol. 86:1728–1738, 1999.

5. COLLINS, M. A., M. L. MILLARD-STAFFORD, P. B. SPARLING, et al.Evaluation of the BOD POD for assessing body fat in collegiatefootball players. Med. Sci. Sports Exerc. 31:1350–1356, 1999.

6. COWARD, W. A., A. M. PRENTICE, P. R. MURGATROYD, et al. Humanenergy metabolism: physical activity and energy expenditure mea-surements in epidemiological research based upon direct and in-direct calorimetry. Euro-Nutr. Ser. 5:126–128, 1984.

7. DEMPSTER, P., and S. AITKENS. A new air displacement method forthe determination of human body composition. Med. Sci. SportsExerc. 27:1692–1697, 1995.

8. DEWIT, O., N. J. FULLER, M. S. FEWTRELL, M. ELIA, and J. C. K.WELLS. Whole body air displacement plethysmography comparedwith hydrodensitometry for body composition analysis. Arch. Dis.Child. 82:159–164, 2000.

9. FALKNER, F. An air displacement method of measuring body vol-ume in babies: a preliminary communication. Ann. N. Y. Acad. Sci.110:75–79, 1963.

10. FIELDS, D. A., and M. I. GORAN. Evaluation of research based bodycomposition techniques in children using the 4Com model as acriterion method. J. Appl. Physiol. 89:613–620, 2000.

11. FIELDS, D. A., G. R. HUNTER, and M. I. GORAN. Validation of theBOD POD with hydrostatic weighing: influence of body clothing.Int. J. Obes. 24:200–205, 2000.

12. GNAEDINGER, R. H., E. P. REINEKE, A. M. PEARSON, W. D. VAN

HUSS, J. A. WESSEL, and H. J. MONTOYE. Determination of bodydensity by air displacement, helium dilution, and underwaterweighing. Ann. N. Y. Acad. Sci. 110:96–108, 1963.

13. GORAN, M. I., M. J. TOTH, and E. T. POEHLMAN. Assessment ofresearch-based body composition techniques in healthy elderlymen and women using the 4Com model as a criterion method. Int.J. Obes. 22:135–142, 1998.

14. GUNDLACH, B. L., H. G. M. NIJKRAKE, and J. G. A. J. HAUTVAST. Arapid and simplified plethysmometric method for measuring bodyvolume. Hum. Biol. 52:23–33, 1980.

15. GUNDLACH, B. L., and G. J. W. VISSCHER. The plethysmometricmeasurement of total body volume. Hum. Biol. 58:783–799, 1986.

16. HEWITT, M. J., S. B. GOING, D. P. WILLIAMS, and T. G. LOHMAN.Hydration of the fat-free body mass in children and adults: impli-cations for body composition assessment. Am. J. Physiol. 265:E88–E95, 1993.

17. HEYMSFIELD, S. B., S. LICHTMAN, R. N. BAUMGARTNER, et al. Bodycomposition of human: comparison of two improved four-com-partment models that differ in expense, technical complexity, andradiation exposure. Am. J. Clin. Nutr. 52:52–58, 1990.

18. HEYMSFIELD, S. B., J. WANG, J. KEHAYIAS, S. HESHKA, S. LICHTMAN,and R. N. PIERSON, JR. Chemical determination of human bodydensity in vivo: relevance to hydrodensitometry. Am. J. Clin. Nutr.50:1282–1289, 1989.

19. HEYMSFIELD, S. B., J. WANG, S. LICHTMAN, Y. KAMEN, J. KEHAYIAS,and N. PIERSON. Body composition in elderly subjects: a criticalappraisal of clinical methodology. Am. J. Clin. Nutr. 50:1167–1175, 1989.

20. KATCH, F. I., and V. L. KATCH. Measurement and prediction errorsin body composition assessment and the search for the perfectprediction equation. Res. Q. Exerc. Sport 51:249–260, 1980.

21. KENDALL, C., and T. B. COPLEN. Multisample conversion of waterto hydrogen by zinc for stable isotope determination. Anal. Chem.57:1437–1440, 1985.

22. LEVENHAGEN, D. K., M. J. BOREL, D. WELCH, et al. A comparisonof air displacement plethysmography with three other techniquesto determine body fat in healthy adults. JPEN J. Parenter. EnteralNutr. 23:293–299, 1999.

23. LOHMAN, T. G. Applicability of body composition techniques andconstants for children and youths. Exerc. Sport Sci. Rev. 14:325–357, 1986.

24. MCCRORY, M. A., T. D. GOMEZ, E. M. BERNAUER, and P. A. MOLE.Evaluation of a new air displacement plethysmograph for mea-suring human body composition. Med. Sci. Sports Exerc. 27:1686–1691, 1995.

25. NUNEZ, C., A. J. KOVERA, A. PIETROBELLI, et al. Body compositionin children and adults by air displacement plethysmography. Eur.J. Clin. Nutr. 53:382–387, 1999.

26. PI-SUNYER, X. Medical hazards of obesity. Ann. Intern. Med.119:655–660, 1993.

27. QUANJER, P. H., J. STOCKS, G. POLGAR, M. WISE, J. KARLBERG,and G. BORSBOOM. Compilation of reference values for lungfunction measurements in children. Eur. Respir. J. 2:184S–261S,1989.

28. ROEMMICH, J. N., P. A. CLARK, A. WELTMAN, and A. D. ROGOL.Alterations in growth and body composition during puberty, I:comparing multicompartment body composition models. J. Appl.Physiol. 83:927–935, 1997.

29. SCHOELLER, D. A. Hydrometry. In: Human Body Composition,A. F. Roche, S. B. Hemsfield, and T. G. Lohman (Eds.). Cham-paign, IL: Human Kinetics, 1996, pp. 25–43.

30. SCHOELLER, D. A., E. VAN SANTEN, W. PETERSON, W. DIETZ, J.JASPAN, and P. D. KLEIN. Total body water measurement in humanswith 18O and 2H labeled water. Am. J. Clin. Nutr. 33:2686–2693,1980.

31. SIRI, W. E. Body composition from fluid spaces and density:analysis of methods. In: Techniques for Measuring Body Compo-sition, J. Brozek and A. Hencshel (Eds.). Washington, DC: Na-tional Academy of Sciences/National Research Council, 1961, pp.223–224.

32. SLY, P. D., C. LANTERI, and J. H. T. BATES. Effect of the thermo-dynamics of an infant plethysmograph on the measurement ofthoracic gas volume. Pediatr. Pulmonol. 8:203–208, 1990.

33. TAYLOR, A., J. W. SCOPES, G. DU MONT, and B. A. TAYLOR.Development of an air displacement method for whole bodyvolume measurement of infants. J. Biomed. Eng. 7:9–17, 1985.

34. VISSER, M., D. GALLAGHER, P. DEURENBERG, J. WANG, R. PIERSON,and S. B. HEYMSFIELD. Density of fat-free body mass: relationshipwith race, age, and level of body fatness. Am. J. Physiol. 272:E781–E787, 1997.

35. WILMORE, J. H. A simplified method for the determination ofresidual lung volume. J. Appl. Physiol. 27:96–100, 1969.

36. ZUMDAHL, S. S. Gases. In: Chemistry. Lexington, MA: DC Heathand Company, 1989, pp. 178–180.

1610 Official Journal of the American College of Sports Medicine http://www.acsm-msse.org