Share this article on:

Assessing body composition before and after resistance or endurance training


Medicine & Science in Sports & Exercise: May 1997 - Volume 29 - Issue 5 - p 705-712
Special Communications: Methods

This study's purpose was to determine the validity of near-infrared interactance (NIR) and bioelectric impedance (BIA) in tracking changes in body composition over 12 wk of either a high intensity endurance (ET) or resistance(RT) training program in nondieting weight-stable untrained males. Prior to and following the control or training period, each subject completed a series of body composition analyses including hydrostatic weighing (HW) with a measurement of residual volume; anthropometric measurements including height, weight, skinfold, and girth; BIA measurement; and NIR measurements. Based on the HW results, there were no significant body composition changes in the control group. For the ET group, a significant decline in relative body fat resulted from a reduction in fat weight (FW) with no change in fat-free weight(FFW). In the RT group, both a significant decline in FW and an increase in FFW contributed to this group's decline in relative body fat. Tracking changes in relative body fat, FW, and FFW, skinfolds agreed reasonably well with HW in all groups while BIA and NIR did not always track body composition changes well. For example, SF and BIA were significantly correlated with the changes in FFW (HW = + 4.1%, SF = + 4.5%, BIA = + 3.1%, NIR = - 0.7%) observed in the RT group compared to HW (SF: r-value = 0.45, SEE = 2.5; BIA: r = 0.33, SEE = 3.4) while the NIR measurements were nonsignificant (r = 0.09, SEE = 5.0). Interestingly, NIR underestimated the gain in FFW in the resistance trained group while BIA underestimated the changes in relative body fat, FW, and FFW in the endurance trained group. Based on these results, BIA and NIR appear not to be appropriate measurement tools for tracking body composition changes in endurance and resistance training individuals respectively.

The Human Performance Laboratory, East Tennessee State University, Johnson City, TN 37614-70654; The Department of Kinesiology and Health Education, The University of Texas at Austin, TX 78712

Submitted for publication March 1995.

Accepted for publication January 1997.

The authors would like to thank our volunteers for their dedication and quality training efforts on this project. This work was supported by a grant from the Canyon Ranch Foundation.

Address for correspondence: Craig E. Broeder, Ph.D., East Tennessee State University, The Human Performance Lab, Box 70654, Johnson City, TN 37614-0654.

Bioelectrical impedance analysis (BIA) and near-infrared interactance (NIR) are two body composition assessment techniques used today in a variety of settings (e.g., hospitals, wellness fairs, and local health clubs). These techniques are very portable, easy to use, and reduce the overall time of assessing relative body fat in the general public. Previous studies have investigated the use of BIA in determining a person's body composition profile (10,14,18,24). The results of these studies are sometimes conflicting regarding the accuracy of BIA for predicting fat-free weight and relative body fat. However, when measurement techniques are carefully standardized, both fat-free weight and relative body fat can be consistently measured with SEEs ranging between 2.1% and 2.7% (12).

Data from a limited number of validation studies on the NIR body composition assessment technique show results that are not very encouraging. For example, Davis et al. (3) and Davis and Paynter(4) reported excellent reliability and good validity using a Futrex-5000 NIR device. However, data were limited to only those subjects whose Futrex-computed relative body fat agreed with other previously validated methods, severely biasing the final statistical comparisons. In contrast, other investigators have reported Futrex-computed relative body fat estimates to be inaccurate in either healthy male and female Caucasians(8,15,20) or in overweight and obese subjects (24).

In many health club or hospital based fitness facilities, BIA or NIR assessment techniques are used in determining the effects of an exercise based training program on a person's body composition profile. However, studies have not clearly determined if BIA or NIR can accurately detect subtle changes in body composition following training. Thus, this study's purpose was to determine the validity of NIR and BIA in tracking changes in body composition before and after 12 wk of either a high intensity endurance or resistance training program in nondieting weight-stable males.

Back to Top | Article Outline


Subject Recruitment and Treatment Group Descriptions

Sixty-four male volunteers (ages 18-35 yr) were recruited from the student population at The University of Texas at Austin and from the surrounding community. All methods and procedures had been previously approved by the Institutional Review Board at The University of Texas at Austin. All subjects read and signed the subject consent form and completed a medical history form prior to the start of the study. No subject had any significant change in body weight over the 6-month period prior to the start of the study. Subjects were randomly assigned to one of three groups: a control group (C, N = 20); a resistance-trained group (RT, N = 22); and an endurance-trained group (ET, N = 22) for a 12-wk intervention period. Of these 64 original volunteers, 23 subjects are not included in the final data analysis making the total sample size 41 subjects (c, N = 15; rt; N = 12; et, N = 14). Of those 23 subjects not included in the final analysis, 14 subjects dropped out because of lack of time and/or commitment to the program and three subjects dropped out as a result of injury. The remaining six subjects were not included in the data set because of technical problems with the Futrex-5000 NIR device.

Back to Top | Article Outline

Pre- and Post-Treatment Measurements

Prior to and following the control or training period, each subject completed a series of body composition analyses including hydrostatic weighing(HW) with a measurement of residual volume; anthropometric measurements including height, weight, skinfold, and girth; BIA using the Valhalla 1990B Bio-Resistance Body Composition Analyzer (Valhalla Scientific, San Diego, CA); and NIR using the Futrex-5000 device (Futrex, Inc., Gaithersburg, MD). In addition, each subject completed two graded treadmill tests to determine˙VO2max, two one-repetition maximum (1-RM) strength tests for six different lifts, and a 3-d dietary and activity log prior to and following the control or training period.

Back to Top | Article Outline

Body Composition Analysis

To minimize measurement error, all subjects were measured by the same investigator for each assessment technique prior to and following either the control or treatment periods (JHW, anthropometric measurements and NIR; LS, BIA; JV, residual volume (RV); and CEB, HW). All subjects were measured at the same time of day during both the pre- and post-treatment measurements, with each subject abstaining from exercise 48 h prior to testing. Subjects were also encouraged to stay well hydrated (i.e., drink an 8-ounce glass of water 2 h prior to testing) and to completely fast thereafter before each body composition assessment session.

Hydrostatic weighing to determine body density was performed as described by Behnke and Wilmore (1). Relative body fat was calculated using the equation of Siri (19) for Caucasians and Schutte et al. for blacks (17). Residual lung volume was determined using the O2 dilution method according to Wilmore et al. (23).

Anthropometric measurements included height to the nearest 0.1 cm; weight to the nearest 0.01 kg; skinfold measurements at the subscapular, triceps, chest, medial suprailiac, anterior-suprailiac, abdomen, and thigh to the nearest 0.1 mm; and circumferences at the neck, chest, upper arm, forearm, waist, hip, thigh, and calf to the nearest 0.1 cm. All anthropometric measurements were taken in duplicate according to Lohman et al.(13). Body mass index was calculated from height and weight (BMI = wt(kg)/ht(m) 2). Skinfold measurements were taken in duplicate. If the difference between two trials exceeded 1.0 mm, a third trial was performed and the closest two trials were averaged. Body density was then determined using the sum of seven-site equation by Jackson and Pollock (9). Relative body fat was calculated as previously described in the hydrostatic weighing procedures.

The BIA instrument was calibrated before each testing session as described by the manufacturer. Electrode placement was standardized according the manufacturer's guidelines. All subjects were tested in the supine position following a 5-min rest period. Resistance values were then used in the following gender specific equation provided by the manufacturer(21): Equation

All near infrared procedures were standardized according to the manufacturer's instructions (6). Each subject's relative body fat was determined by a multiple regression equation using sex, weight, height, and daily physical activity level (F.I.T.). Each subject's planned daily physical activity level was determined using the F.I.T. equation by self assessment of exercise frequency (F), intensity (I), and minutes of activity(T). Using the numerical F.I.T. value, scores were converted to a final score for the physical activity component of the multiple regression equation (SeeTable 1).

Back to Top | Article Outline

Training Program

After completing all pre-treatment tests, each subject participated for 12 wk in either the control, resistance, or endurance training treatment protocol. Control group subjects were instructed not to change either their activity or dietary habits during the 12-wk interim period between pre- and post-measurements. A more detailed explanation of each respective training regiment, as well as the protocols for the determination of ˙VO2max and each subject's 1-repetition maximum lifts, have been published(2).

Resistance-trained group. Each subject in the resistance-trained group participated in a 12-wk training program specifically designed to increase strength and fat-free mass. Each subject performed heavy resistance training (80% of each subject's one-repetition maximum lift for each exercise movement) with a combination of free weights and various accommodating resistance machines, 4 d·wk-1 (i.e., Monday, Tuesday, Thursday, and Friday). The strength-training program included the following exercise movements: bench press, parallel dip, and behind-the-neck press, upright rows, triceps press-down, leg press, leg extension, leg curl, lat pushdown, barbell curl, and abdominal crunch. These exercises were divided into six movements per training day and took approximately 1 h. Upper body exercises were performed on Monday and Thursday while lower body exercises were performed on Tuesday and Friday. Abdominal crunches were performed at the end of every workout day. For safety while performing free weight exercises, subjects were divided into groups so that each person lifting had at least one spotter. During the first 2 wk of the resistance training program, subjects performed 10-12 repetitions per set, three sets per exercise movement after a brief stretching and warm-up period. The resistance was set on each exercise so that subjects became fatigued between 10 and 12 repetitions. This 2-wk period was designed to prepare subjects for the remaining 10 wk of the program. During the next 10 wk upper and lower body exercises were performed with the weight established on each set so that failure to lift the weight occurred between 10 and 12 repetitions on the first set, 8 and 10 repetitions on the second set, and 6 and 8 repetitions on the third set. Resistance was increased for each exercise movement during the program based on the number of repetitions required to promote fatigue in each subject. Abdominal crunches were performed for one set until failure or until the subject reached 50 repetitions. When a subject could do 50 repetitions of the abdominal crunch, weight was added by having the subject hold a 2.27-11.33-kg weight over his chest while performing each crunch.

Endurance trained group. Each subject in the endurance-trained group participated in a walk and/or jog program 4 d·wk-1 (i.e., Monday, Tuesday, Thursday, and Friday) for 12 wk. All exercise training sessions were monitored by a trained exercise leader, and each subject was instructed to gradually increase his exercise duration and intensity so that a new training goal was reached every 4 wk. By the end of the fourth week, subjects were exercising for 40 min at a minimum intensity of 75% of treadmill determined maximal heart rate. After the eighth wk, each person was exercising for 50 min at an intensity of between 75 and 85% of his maximal heart rate. In weeks 8-12, subjects also included in their regimen fartlek-type interval training, (i.e., duration between 2 to 5 min) at > 90% of their respective maximal heart rates.

To accurately determine that each subject maintained a predetermined target heart rate, the subject periodically wore a telemetric heart rate monitoring system developed by CIC Inc. (Port Washington, NY). The determination of each subject's initial target heart rate was estimated from a linear regression model using steady-state oxygen consumption and heart rate values achieved during each stage of the pre-test determination of maximal oxygen consumption. This procedure allowed the endurance program exercise leaders to establish safe and effective cardiovascular training prescriptions for these previously untrained subjects. For both training groups make-up sessions were on Wednesday and Saturday.

Back to Top | Article Outline

Statistical Analysis

An ANOVA, with repeated measures design (treatment group × time), was used to determine the effects of each 12-wk training and control period on body composition for the hydrostatic weighing results. When there was a significant difference between groups or between pre- to post-measurements, a multiple-comparison post hoc procedure was used for orthogonal comparisons with the P-level set at ≤ 0.05 and. The delta values were analyzed using a two-way repeated measure factorial analysis (treatment group × measurement procedure as the repeated measure). When significant ANOVA differences were identified, orthogonal post hoc multiple comparisons were used to identify specific differences. Regression analysis was used for determining how well each body composition assessment technique followed changes in FW and FFW after each respective treatment intervention. All statistical calculations were performed using SuperANOVA and Statview 512+ by Abacus Concepts, Inc. (Calabasa, CA).

Back to Top | Article Outline


The initial physical characteristics of the subjects are presented inTable 2. There were no statistical differences among groups for any of the listed variables prior to the control or training periods. ˙VO2max significantly increased 11.2% in the ET group while the RT and C groups did not significantly change after respective treatments. In contrast, only the RT trained group significantly increased following training on the 1-RM trials (21.6%). A detailed description of the training results for ˙VO2max and strength changes in the endurance and resistance trained groups has been published (2).

Following the 12-wk intervention period, only the ET and RT groups showed changes in body composition. Based on the hydrostatic weighing results, there were no significant body composition changes in the control group following the 12-wk period of normal activity. In contrast, both the ET and RT groups showed a significant decline in relative body fat following training(Table 3). For the ET group, the significant decline in relative body fat resulted from a significant reduction in total body fat with no change in fat-free weight. In the RT group, as expected by the training program designed, both a significant decline in total fat weight and an increase in fat-free weight contributed to this group's decline in relative body fat. Fat-free weight significantly increased 4.1% from pre-training values.

The linear regression analyses for all subjects, showing the relationship between the hydrostatic weighing relative body fat value and the value determined by skinfold, BIA, and NIR methods are shown inTable 4. The results indicate that the skinfold measurements had the highest correlation coefficient and smallest SEE compared with the hydrostatic weighing results for both the pre- and post-treatment measurements. During the pre-treatment period, BIA was a stronger predictor of relative body fat than the NIR technique. In contrast, following the 12-wk treatment period, BIA and NIR showed a similar relationship with hydrostatic weighing while NIR had the smaller SEE. It is important to note that the BIA SEE were large (± 4.5 and ± 4.3 for the pre- and post-treatment periods, respectively).

The repeated measures ANOVA and multiple-comparison post hoc results indicated that both the Jackson and Pollock skinfold equation and NIR assessment technique (9) significantly underestimated relative body fat when compared with hydrostatic weighing during both the pre- and post-treatment periods for all subjects combined. For the BIA assessment technique, there were no significant differences found during either the pre- or post-treatment periods for relative body fat when compared with hydrostatic weighing (Table 4). It is important to note that while the entire group mean for BIA was not significantly different from HW, for specific individuals the technique was not an accurate measure of relative body fat. For example, BIA consistantly underestimated one subject's relative body fat during his pre- and post-treatment measurements from the endurance training group. For this individual, in absolute terms BIA underestimated relative body fat 13.2% and 10.4% for the pre-and post-treatment, respectively. For these absolute differences, relative body fat was underestimated 51% and 43% in relative terms for the pre- and post-treatment periods, respectively. Similar individual differences were also observed in NIR measurements.

In regard to tracking changes in relative body fat, FW and FFW, the SF-derived values agreed reasonably well with HW. However, BIA and NIR did not always track these body composition parameters well in the specific training groups and significant differences were found compared with the HW results. For example, when pre- to post-treatment FFW values from the resistance trained group (Fig. 1, Table 5) were used, no significant differences were observed between the changes measured by hydrostatic weighing, skinfolds, and BIA. (4.1%,4.5%, and 3.1%, respectively). In addition, regression analyses indicated that the SF and BIA FFW delta values (change in FFW after each respective 12-wk treatment) were significantly correlated with the hydrostatic weighing FFW delta value for the RT group (SF: r = 0.45, SEE = 2.5, P < 0.05; BIA: r = 0.33, SEE = 3.4, P < 0.05). In contrast, the NIR delta value was not significantly correlated with the HW delta value. In addition, the change observed in FFW following training was significantly less using NIR than the other methods. In fact, according to the NIR technique, FFW significantly declined 0.47 kg in absolute terms following the 12-wk resistance training program. As a result, the change in relative body fat measured using NIR was in the positive direction (0.7%), while the other techniques showed relative body fat was lower following resistance training. NIR measurements also indicated that the control subjects had a significant decline in relative body fat while the other procedures indicated no significant change in body composition over the 12-wk treatment period.

On the other hand, BIA agreed reasonably well with the HW measurements in detecting changes in body composition in the resistance trained group and the lack of changes in the control group. However, BIA was significantly different from each of the other assessment techniques in determining changes in relative body fat, FW, and FFW in the endurance trained group. Following endurance training, BIA underestimated the changes in relative body fat (BIA =- 0.4%; HW = -1.9%). Consequently, both FW and FFW declined 0.51 kg and 0.50 kg, respectively, according to the BIA measurements, while the HW measurements indicated that FW decreased 1.66 kg and FFW increased 0.66 kg.

Back to Top | Article Outline


The purpose of this study was to determine how well SF, BIA, and NIR assessment techniques detected subtle changes in relative body fat, FW, and FFW following either a high intensity resistance or endurance training program when compared with hydrostatic weighing. The body composition changes according to the hydrostatic weighing results in the control, endurance, and resistance trained groups were as expected based on each treatment protocol. There were no significant changes in the control group while both exercise groups showed significant declines in relative body fat. Exercise resulted in maintenance of FFW in the endurance trained group, while FFW increased significantly in the resistance trained group. Regarding the ability of SF, BIA, and NIR to track body composition changes following exercise training, the SF technique resulted in similar changes in FW and FFW when compared with the hydrostatic weighing. In contrast, BIA and NIR were not always able to detect the body composition changes induced by training. For example, NIR measurements were unable to detect the significant increase in FFW that occurred in the resistance trained group. Eckerson et al.(5) found that following 12 wk of resistance training, NIR showed statistically significant but low correlations with FFW changes determined by hydrostatic weighing in 21 college-aged Caucasian males. In the current study, while HW, SF, and BIA detected a similar FFW increase for the resistance trained group of 2.5 kg, 2.9 kg, and 2.0 kg, respectively, the disparity between FFW estimates by NIR and HW were very apparent. In contrast to the HW results, FFW significantly decreased 0.5 kg according to the NIR results. Collectively, the results of the Eckerson et al.(5) and the current study suggest that NIR using the Futrex 5000 may not be a sensitive technique for determining FFW changes in an individual involved in resistance training. In addition, NIR measurements showed a significant decline in relative body fat and FW with a corresponding increase in FFW in the control group which was not in agreement with any of the other techniques used. In contrast, NIR assessment was able to detect the magnitude of changes in body composition in the endurance trained group, while BIA showed significantly lower relative body fat and FW changes following training compared with the hydrostatic weighing.

It is important to note that when HW density is used with a two-compartment model, the noted increase in FFW is often underestimated. However, a recent study by Nelson et al. (16) showed that 24-hr urinary creatinine, regional CAT scan, hydrostatic weighing, total body water, and whole-body potassium counting were all sensitive measures for detecting small significant changes in FFM when subjects remained weight stable between pre-treatment and post-treatment measurements. In this comprehensive study specifically designed to determined which body composition techniques could detect changes in soft tissue after resistance training in older women best, a 3.3% (1.3 kg) increase in FFM was observed and detected well by hydrostatic weighing. These findings are very similar to the present study in which the resistance trained group increased their FFW by approximately 3.0% (2.5 kg) according to hydrostatic weighing with no significant changes in body weight. Thus, it appears that HW is a valid method for determining small but significant changes in FFW in weight-stable individuals as in the present study. Still, others have suggested that the inclusion of total-body water with total-body density from hydrostatic weighing in a three-compartment body composition model provides the best sensitivity to small changes in FFW(7).

What factors prevent NIR body composition analyses from being a valid body composition assessment tool? One major potential source of error appears to be the inability of a subject to accurately estimate activity level using the F.I.T. equation provided by the manufacturer. Wilmore et al.(24) clearly showed that the F.I.T. self-estimation of activity level could lead to large errors in relative body fat, FW, and FFW estimates by NIR. In fact, according to them, body fat estimates can be over-or underestimated two-fold from the true value because of errors in activity level reporting. In the Wilmore et al. study(24), of all the variables used in the multiple regression equation for NIR measurement procedures, replication of a person's self-assessed activity level showed the lowest correlation between trials 1 and 2 (r = 0.73). Specifically, exercise intensity was the most difficult of the variables to replicate. Their subjects found it difficult to determine their true activity level because of problems associated with estimating walking or jogging pace, or activities like resistance training which was not included as an activity choice. In addition, it is possible that following resistance training more than one measurement site is needed to accurately detect FFW changes which cannot be detected by a bicep measurement. Measuring only at the biceps certainly is not a true representation of total body FFW changes resulting from increases in total body skeletal muscle mass that can accompany high intensity resistance training. However, previous research has shown that NIR multiple site measurements do not improve the accuracy of the technique in assessing body composition and therefore would most likely have little value in tracking FFW changes that can occur during resistance training(15).

In regard to the BIA underestimations of relative body fat, FW, and FFW changes following endurance training, increases in total body water may have played a role in this group of subjects. One of the most important adaptations that occurs with endurance training is an increase in plasma volume increasing stroke volume and thus cardiac output. Previous studies have shown that BIA is affected greatly by changes in hydration status and that overhydrating a person while in the resting state results in an overestimation of a person's true relative body fat and FW levels (12). In an ongoing study in our laboratory investigating the effects of electrolyte solutions on BIA measurements (M. Saunders, manuscript in review) overhydration with distilled water or an electrolyte solution increased BIA relative body fat values 26.5% and 17.0%, respectively, compared with pre-treatment normally hydrated values. Since the subjects in the current study did show a significant increase in their aerobic fitness level (11.2%) and training occurred in a hot climate, it is safe to assume that increases in plasma volume, total body water, and electrolyte changes may have affected the BIA measurements, leading to an underestimation of the amount of FW lost and FFW gained following endurance training. These results indicate that besides standardizing fluid intake before each BIA measurement, a correction factor may be needed to account for changes in body fluid content following endurance training activities in longitudinal type studies.

Several studies have shown that NIR has the largest total error (4.2%) and random error (8.4%) when compared with BIA, SF, and HW(20,22). The results of the current study agree well with these findings. Prior to starting the training program, the NIR had the lowest correlation value when compared with HW (r = 0.53). Interestingly, following the 12 wk of each respective treatment, the correlation between NIR and HW significantly improved and compared favorably with the other assessment techniques. In fact, the SEE decreased pre-treatment to post-treatment from 4.4% to 2.5%, respectively. The NIR pre-treatment SEE was similar to that reported by others (8,24) while the post-treatment SEE is the lowest reported value to date. The NIR correlation value compared with HW for relative body fat during the pre-treatment was also low compared with previous studies (3,4,11,24) while the post-treatment value is similar to the Wilmore et al. study(24) but lower than other studies(3,4,11). One explanation for the improved correlation and reduced SEE between HW and NIR during the post-treatment measurement may be related to the subject's training experience. At the start of the study, all subjects were sedentary and had little or no experience rating exercise intensity. After 12 wk of formal exercise with a trained exercise leader, each subject was more likely to accurately rate his exercise intensity level. Thus, reducing the overall measurement error related to each subject's self report of exercise intensity in the NIR measurement.

In conclusion, SF measurements were able to track subtle changes in body composition following 12 wk of either ET or RT. However, the SF measurements, using the Jackson and Pollock equation (9), underpredicted each group's relative body fat when compared with HW. Use of the NIR device for assessing body composition changes after resistance training programs is not appropriate. The potential for inaccurate results regarding the determination of changes in body composition following extended exercise training periods is high. The NIR procedure using a single bicep measurement is not an accurate method for assessing FFW changes that accompany high intensity resistance training. As previous research has suggested(24), strong consideration should be given to delete or modify the activity pattern scales used in the current NIR multiple regression equation. Finally, BIA impedance measurement may not be a valid method for assessing changes in body composition when individuals make significant changes in plasma volume and/or total body water stores. Further research is needed to determine what effects plasma volume changes might have on the validity of BIA body composition assessment. Based on these findings, professionals assessing body composition with NIR and BIA for longitudinal purposes must understand the limitations of these devices and carefully standardize all pre-testing and testing procedures to minimize test error. When possible, skinfold measurements should also be used to help determine body composition changes following extended periods of weight loss and/or exercise training programs.

Figure 1-Correlations and SEEs for the fat-free weight deltas following 12 wk of resistance training. UW, underwater weighing; SF, skinfolds; BIA, bioelectric impedance; NIR, near-infrared interactance; NS, nonsignificant. ¥, significantly different from underwater weighing results. †, significantly different from skinfold and bioelectric impedance results.

Figure 1-Correlations and SEEs for the fat-free weight deltas following 12 wk of resistance training. UW, underwater weighing; SF, skinfolds; BIA, bioelectric impedance; NIR, near-infrared interactance; NS, nonsignificant. ¥, significantly different from underwater weighing results. †, significantly different from skinfold and bioelectric impedance results.

Back to Top | Article Outline


1. Behnke, A. R. and J. H. Wilmore. Evaluation And Regulation Of Body Build And Composition. Englewood Cliffs: Prentice-Hall, Inc., 1974.
2. Broeder, C. E., K. A. Burrhus, L. S. Svanevik, and J. H. Wilmore. The effects of either high-intensity resistance or endurance training on resting metabolic rate. Am. J. Clin. Nutr. 55:802-810, 1992.
3. Davis, P. O., C. O. Dotson, and P. D. Manny. NIR evaluation for body composition analysis (Abstract). Med. Sci. Sports Exerc. 20(Suppl):8, 1988.
4. Davis, P. O. and L. Paynter. Evaluation of a Commercial Near-Infrared Instrument for Body Composition Analysis. Falls Creek: Human Performance Centers, 1987.
5. Eckerson, J., J. Stout, T. Housh, and G. Johnson. Validity of skinfold, bioelectrical impedance, and near-infrared interactance equations for assessing changes in fat-free weight (Abstract). Med. Sci. Sports Exerc. 25(Suppl):59, 1993.
6. Futrex. Futrex-5000 users manual. Gaithersburg, MD: Futrex, Inc., 1987.
7. Goran, M. I. and E. T. Poehlman. Endurance training does not enhance total energy expenditure in health elderly persons. Am. J. Physiol. 263: E950-E957, 1992.
8. Israel, R. G., J. A. Houmard, K. F. O'Brien, M. R. McCommon, B. S. Zamora, and A. W. Eaton. Validity of a near-infrared spectrophotometry device for estimating human body composition. Res. Q. Exerc. Sport 60:379-383, 1989.
9. Jackson, A. S. and M. L. Pollock. Practical assessment of body composition. Physican Sportsmed. 13:76-90, 1985.
10. Jackson, A. S., M. L. Pollock, J. E. Graves, and M. T. Mahar. Reliability and validity of bioelectrical impedance in determining body composition. J. Appl. Physiol. 64:529-534, 1988.
11. Laverty, M. A., V. J. Paolone, C. S. O'Shea, and Z. V. Kendrick. Comparison of near infrared interactance to densitometry and skinfold estimations of percent body fat (Abstract). Med. Sci. Sports Exerc. 21(Suppl):102, 1989.
12. Lohman, T. G. Advances in Body Composition Assessment (Monograph Number 3). Champaign: Human Kinetics Publishers, 1992.
13. Lohman, T. G., A. F. Roche, and R. Martorell.Anthropometric Standardization Reference Manual. Champaign: Human Kinetics Books, 1988.
14. Lukaski, H. C. Methods for the assessment of human body composition: tradional and new. Am. J. Clin. Nutr. 46:537-556, 1987.
15. McLean, K. P. and J. S. Skinner. Validity of Futrex-5000 for body composition determination. Med. Sci. Sports Exerc. 24:253-258, 1992.
16. Nelson, M. E., M. A. Fiatarone, J. E. Layne, et al. Analysis of body-composition techniques and models for detecting change in soft tissue with strength training. Am. J. Clin. Nutr. 63:678-686, 1996.
17. Schutte, J. E., E. J. Townsend, J. Hugg, R. F. Shoup, R. M. Malina, and C. G. Blomqvist. Density of lean body mass is greater in blacks than in whites. J. Appl. Physiol. 56:1647-1649, 1984.
18. Segal, K. R., B. Gutin, E. Presta, J. Wang, and T. B. Van Itallie. Estimation of human body composition by electrical impedance methods: a comparative study. J. Appl. Physiol. 58:1565-1571, 1985.
19. Siri, W. E. Body composition from fluid spaces and density. In: Techniques For Measuring Body Composition, J. Brozek and A. Henschel (Eds.). Washington, D.C.: National Academy of Sciences National Research Council, 1961.
20. Stout, J. R., J. M. Erikson, T. J. Housh, G. O. Johnson, and N. M. Betts. Validity of percent body fat estimations in males.Med. Sci. Sports Exerc. 26:632-636, 1994.
21. Valhalla, I. 1990A. Bio-Resistance Body Composition Analyzer Operation Manual. San Diego: Valhalla Scientific, 1986.
22. Vehrs, P. R., J. D. George, C. Payne, J. Peugnet, G. R. Bryce, G. W. Fellingham, and A. G. Fisher. Reliability of four methods of body composition assessment. Med. Exerc. Nutr. Health 3:2-8, 1994.
23. Wilmore, J. H., P. A. Vodak, R. B. Parr, R. N. Girandola, and J. E. Billing. Further simplification of a method for determination of residual lung volume. Med. Sci. Sports Exerc. 12:216-218, 1980.
24. Wilmore, K. M., P. J. McBride, and J. H. Wilmore. Comparison of bioelectric impedance and near-infrared interactance for body composition assessment in a population of self-perceived over-weight adults.Int. J. Obes. 1994:375-381, 1994.


©1997The American College of Sports Medicine