Air displacement plethysmography for estimating body composition changes with weight loss in middle-aged Japanese men.

Abstract

To examine the degree to which air displacement plethysmography (ADP) can track body composition changes in response to weight loss in obese Japanese men.
50 men, aged 30-65 years with a mean BMI of 30 kg/m(2), were included in a 3-month weight loss program. Percentage of fat mass (%FM) was determined by dual energy X-ray absorptiometry (DXA) and ADP at baseline and month 3.
With 6.2 ± 4.3 kg of weight loss, %FM, as determined by DXA and ADP, significantly decreased by 3.9 ± 2.9% and 3.9 ± 3.3% respectively. There was no mean difference for change (Δ) in %FM between the two methods. DXA-derived Δ%FM significantly correlated with Δ%FM determined by ADP (R(2) = 0.48, p < 0.01). Furthermore, the Bland-Altman plots demonstrated no systematic bias for Δ%FM (r = -0.20, p = 0.17). However, %FM by ADP (r = 0.42) at baseline and Δ%FM by ADP (r = -0.54) were significantly correlated to the differences between Δ%FM by DXA and ADP.
These results suggest that ADP is comparably accurate for evaluating Δ%FM when compared with DXA, although there were proportional biases.

Full text

Page 1

1
Title: Air displacement plethysmography for estimating body composition changes with weight
loss in middle-aged Japanese men

Short running head: Body composition change by air displacement method

Section: Original Article

Summary:
Aim: To examine the degree to which air displacement plethysmography (ADP) can track body
composition changes in response to weight loss in obese Japanese men. Method: Fifty men,
aged 30–65 yr, with a mean body mass index of 30 kg/m2, received a 3-month weight-loss
program. Percentage of fat mass (%FM) was determined by dual energy X-ray absorptiometry
(DXA) and ADP at baseline and month 3. Results: With 6.2 ± 4.3 kg of weight loss, %FM, as
determined by DXA and ADP, significantly decreased by 3.9 ± 2.9% and 3.9 ± 3.3%,
respectively. There was no mean difference for change (Δ) in %FM between the two methods.
DXA-derived Δ%FM significantly correlated with Δ%FM determined by ADP (R2 = 0.48, P <
0.01). Furthermore, the Bland-Altman plots demonstrated no systematic bias for Δ%FM (r =
-0.20, P = 0.17). However, %FM by ADP at baseline and Δ%FM by ADP were significantly
correlated to the differences between Δ%FM by DXA and ADP (r = 0.42, -0.54, respectively).
Conclusion: These results suggest that ADP is comparably accurate for evaluating Δ%FM
determined by DXA, although there were proportional biases.

Keywords: air displacement plethysmography; dual energy X-ray absorptiometry; percentage
of fat mass; weight loss; obesity

Page 2

2
1. Introduction
The obesity epidemic has become one of the biggest public health concerns worldwide
including within Asian countries [1,2], because of its close relationship to chronic diseases such
as hypertension, dyslipidemia, diabetes and cardiovascular disease [3,4]. Since obesity is
characterized and defined as the excess accumulation of body fat, accurate measurements of
body composition are of utmost importance for public health and clinical perspectives. The
ability of health professionals to accurately estimate body composition changes in obese
individuals is also critical for determining the effectiveness of weight loss or management
strategies.
Several body composition methods are available, which differ in terms of their theoretical
bases and scientific assumptions, as well as their cost, complexity, and subject acceptability. For
example, hydrostatic weighting (HW) has long been considered the gold standard for measuring
body composition [5]. However, it requires many methodological assumptions and
inconveniences that limit its usefulness and widespread application. Consequently, dual energy
X-ray absorptiometry (DXA) has emerged as one of the most accurate techniques to assess body
composition. DXA uses a 3-compartment model to provide an estimate of bone mineral content,
fat and lean soft tissue. DXA is easier to administer than HW and correlates well with other
measures of body composition such as HW, total body water, and multi-compartment
approaches [6,7]. DXA can also detect body composition changes due to weight loss [8].
Air displacement plethysmography (ADP) has emerged as a fairly novel technique that
provides a body composition estimate from body density using a 2-compartment model, as does
HW [9]. The biggest difference between ADP and HW is that ADP does not require participants
to exhale maximally while submerged under water. Therefore, ADP seems much more feasible
and comfortable than HW. Cross-sectional studies showed that percentage of fat mass (%FM)
estimated from ADP highly correlate with those from DXA [10,11].
A few studies have demonstrated that ADP was sensitive enough to detect moderate body
composition changes compared to DXA over various time periods (from 2 to 16 months) and
percent weight loss (from 4.7% to 7.0%), primarily among Caucasian adults [11-13]. Asians,
including Japanese, have greater %FM than Caucasians for the same BMI and age [14].
Therefore, it is important to determine whether ADP is a valid method of tracking body

Page 3

3
composition changes in Japanese men.
The purpose of this study, then, was to examine whether ADP could accurately track body
composition changes in response to weight loss in obese Japanese men, using DXA as a
reference method. Better understanding of this issue will be useful for designing longitudinal
studies regarding body composition, as well as in providing clinically-relevant information for
health professionals.

2. Participants and Methods
Participants
We recruited participants from communities around the University of Tsukuba for 3-month
weight-loss programs through local newspaper advertisements and study flyers. Eligibility
criteria included the following: 1) men, 2) aged 30–65 yr, 3) a body mass index (BMI) of greater
than 25 kg/m2 according to the domestic obesity guideline [15] and 4) no history of
cardiovascular disease. In Japan, despite the fact that only 2–3% has been characterized as
having a BMI ≥ 30 kg/m2 , the prevalence of metabolic disorders are relatively high [16,17].
Thus, the cutoff value for the definition of obesity for Asian populations was used [15]. After an
orientation session, 54 men participated in the weight-loss programs in 2007. The men were
non-randomly assigned to receive either a dietary program (n = 30) or an exercise program (n =
24). These two weight-loss methods were employed to achieve a wide range of %FM changes.
We excluded four men with incomplete assessments after the programs. Thus, data from the
remaining 50 men (dietary program: n = 26; exercise program: n = 24) were analyzed. We fully
explained the purpose and design of the study to each participant before they gave written
informed consent. The research protocol was approved by the institutional review board at the
University of Tsukuba, and thus meets the standards of the Declaration of Helsinki.

Weight-loss programs
The dietary program consisted of weekly group-based 90-min instructional sessions for 3
months, as well as individual counseling by trained staff. In each session, the participants
received lectures on low-calorie diets and eating behavior. We instructed them to consume a
well-balanced 1680 kcal diet per day. The exercise program consisted of 90-min sessions of

Page 4

4
aerobic exercise three times per week for 3 months. Each exercise session began with 15–30
minutes of warm-up activities such as stretching. The session was followed by 30–60 minutes of
brisk walking performed outdoors in the campus of the University of Tsukuba and finished with
15–30 minutes of cool-down activities.

Energy intake and expenditure
Energy intake in kilocalorie was assessed before and during a 2-week period (weeks 9 and
10) of both weight-loss programs by 3-day weighed dietary records, and dietary recall
interviews for each participant were performed by skilled dieticians. The dietary data for each
participant were analyzed by using commercially-available computer software (Excel Eiyo-kun,
Kenpakusha, Tokyo, Japan). Total energy expenditure was also assessed by a validated uniaxial
accelerometer (Lifecorder-EX; Suzuken, Nagoya, Japan). The accelerometer was firmly
attached to the participant’s clothing (belt or waistband) during all waking hours (except while
bathing and sleeping) 2 weeks before starting the 3-month programs, for baseline examination,
and throughout the programs. Forty-two men (23 for the dietary group and 19 for the exercise
group) successfully completed the above measurements before and throughout the 3-month
programs.

Measurement procedures
We instructed the participants not to participate in vigorous physical activity or alcohol
consumption within 24 h prior to body composition measurement, as well as to refrain from
eating or drinking for 2 h prior to the measurements. They removed all metal objects and jewelry
during the measurements. We assessed height to the nearest 0.1 cm using a wall-mounted
stadiometer. Body weight was measured to the nearest 0.01 kg using DXA equipment. We then
calculated BMI as body weight in kilograms divided by squared height in meters. Waist
circumference was measured to the nearest 0.1 cm at the level of the umbilicus using a flexible
retractable fiberglass tape measure. We always implemented DXA measurement first, and then
ADP within 30 min. All measurements were conducted in the same order at baseline and at
month 3.

Page 5

5
Dual energy X-ray absorptiometry
We measured %FM by DXA (e.g., %FMDXA) using a Lunar DPX-NT densitometer (Lunar,
Madison, WI). We calibrated the densitometer every day according to manufacturer’s
recommendations. The densitometer calculated soft tissue mass, including fat and lean tissue
mass, from the ratio of mass-attenuation coefficients (R-value) at 40–50 keV and 80–100 keV.
We defined fat-free mass as lean tissue mass plus bone mineral content. During each
measurement, the participants remained motionless in the supine position while the scanning
arm passed over his body in parallel 1-cm strips. To minimize technical error, the same examiner
operated the densitometer and positioned the participant at baseline and month 3. The CVs
in %FM with repeated examinations were less than 2% in our laboratory (n = 33).

Air displacement plethysmography
We measured %FM by ADP (e.g., %FMADP) using a Bod Pod (Life Measurement Inc.,
Concord, CA) according to manufacturer’s instructions. The physical concepts and operational
principles of ADP were described elsewhere [10]. Before each measurement, an examiner
calibrated the Bod Pod six consecutive times to estimate the mean volume of the chamber with
the 49.490 L cylinder. The participant wore trunks and a swim cap to minimize clothing and
compress air, thereby estimating volume variations. Then, the participant was weighed on an
electronic scale which was calibrated in advance with two standard 10-kg weights. The
participant sat quietly in the chamber while their raw body volume (Vb) was measured
consecutively until two values within 150 mL were obtained. We estimated thoracic gas volume
(Vtg) by having participants perform the panting maneuver according to manufacturer’s
instruction until a successful measurement (the merit and airway value was less than 1 or 35 cm
H2O) was obtained. Briefly, the participant wore a nose clip and breathed normally for three
breathing cycles through a tube connected to the internal system. At the midpoint of an
exhalation, the airway tube was momentarily occluded. And, the examiner signaled him to
perform three small puffs of air into the tube while maintaining a tight seal around the end of the
tube. We used the measured Vtg to calculate a corrected Vb (raw Vb minus Vtg) and computed
body density as body weight divided by the corrected Vb. We calculated %FMADP using the
Siri’s equation [18]. The CVs in %FM with repeated examinations were less than 5% in our

Page 6

6
laboratory (n = 8).

Statistical Analysis
We chose DXA, which is not a recognized gold standard, as the reference method, as it has
been considered a reasonable alternative to a multi-compartment approach [8]. After stratifying
by program, we applied a paired Student’s t-test to compare the mean %FM from DXA and
ADP at each time point, and to assess the difference between measurements at baseline and
month 3. A simple linear regression analysis examined the proportion of %FMDXA at each time
point could be accounted for by %FMADP. Bland-Altman plots [19] assessed the potential bias
between DXA and ADP at each time point in wide ranges of body composition and the changes
(Δ). We tested the null hypothesis that the slope was equal to zero in the models to consider the
presence of bias. In addition, Pearson’s correlation coefficients were calculated to explain the
variance of the difference between Δ%FM by DXA and ADP. A P value less than 0.05 was
regarded as statistically significant. We performed all statistical analyses using SAS, version 9.1
(SAS Institute, Inc., Cary, NC).

3. Results
Table 1 describes baseline characteristics, and compares body composition results obtained
using DXA and ADP at baseline and month 3. Attendance rates were 77.5% for dietary sessions
and 86.8% for exercise sessions. The weight-loss programs significantly reduced body weight,
waist circumference, body volume and %FM, and significantly increased body density in both
programs and in the combined data. Fat-free mass measured by DXA was significantly
decreased in dietary program and combined data. In contrast, bone mass remained unchanged
for the all groups. There were no significant differences in %FM at both time points or in
Δ%FM in response to weight loss between the two methods.
Energy intake was significantly decreased from 2062 ± 236 to 1567 ± 285 kcal/day in the
dietary program. Total energy expenditure significantly increased from 2422 ± 242 to 2662 ±
284 kcal/day during the exercise program. In the combined data, there were significant
reductions in energy intake (from 2172 ± 351 to 1841 ± 496 kcal/day) and significant increase in
total energy expenditure (from 2462 ± 230 to 2585 ± 263 kcal/day) through the weight-loss

Page 7

7
programs.
Table 2 summarizes the simple regression analyses and the Bland-Altman plots for %FM at
baseline and month 3 using DXA as the reference method. Coefficients of determination (R2)
were moderate to high (≥ 0.7) for %FM at baseline. After weight loss, the R2 for %FM slightly
increased (≥ 0.8). The Bland-Altman plots demonstrated no significant bias in %FM at both
time points.
Fig. 1 displays the results of simple regression analyses for Δ%FM by DXA and ADP. The
R2 coefficients for Δ%FM were moderate (≥ 0.4). Almost all plots for Δ%FM were within the
95% prediction intervals.
Fig. 2 depicts the results of the Bland-Altman plots for assessing bias in the estimation of
Δ%FM between the two methods. The plots demonstrated no significant bias in Δ%FM. In
addition, almost all individual plots for Δ%FM in both programs were within the 95% limits of
agreement.
Fig. 3 shows relation of Δ%FM determined with DXA minus Δ%FM estimated from ADP
to baseline %FM determined by ADP and the Δ%FM determined by ADP. ADP-derived %FM
at baseline proved to be positively related to the differences between Δ%FM by DXA and ADP
(r = 0.42, P < 0.01). Similarly, ADP-derived Δ%FM was inversely associated with the
difference (r = -0.54, P < 0.01).

4. Discussion
The present study examined whether ADP could track body composition changes with
weight loss using DXA as the reference method in middle-aged Japanese men. We found that
ADP with the use of Siri’s equations tracked Δ%FM in a manner similar to that derived by DXA.
We also demonstrated a uniform distribution of measurement errors (DXA minus ADP) in
Δ%FM across the range of Δ%FM in the Bland-Altman analysis. However, we also showed
that %FM from individuals with larger %FM by ADP and/or with greater Δ%FM is
overestimated as compared to DXA method.
To date, few studies have examined the ability of ADP to accurately estimate body
composition change over a period of weight loss [11-13]. In 22 obese men and women with a
mean BMI of 30 kg/m2, Weyers et al. reported no significant mean differences in Δ%FM

Page 8

8
between DXA and ADP using Siri’s equation after an 8 week moderately energy restricted diet
[11]. The R2 between DXA and ADP in their study was 0.44 for Δ%FM (P < 0.05). Frisard et al.
also reported that R2 between DXA and ADP exhibited 0.24–0.28 for Δ%FM (Brozek’s and
Siri’s equations) in 56 obese men and women [12]. Elberg et al. revealed good agreement of
Δ%FM between DXA and ADP with the use of Siri’s equation in natural changes over 1 year in
86 boys and girls (R2 = 0.59) [20]. Moreover, a recent study of Minderico et al. demonstrated no
significant differences in Δ%FM between DXA and ADP in 93 obese women [13]. They
described an R2 of 0.76 for Δ%FM between DXA and ADP. The current study also
demonstrated no significant differences in Δ%FM. R2 between DXA and ADP were 0.48 for
Δ%FM (P < 0.01). Our results are consistent with these prior studies.
Of the four previous studies, two examined the presence of systematic bias in the estimation
of Δ%FM by using a Bland-Altman approach. Minderico et al. indicated no significant bias,
suggesting the ability of ADP to assess Δ%FM against the wide spread variations (P = 0.67; r
value not shown) [13]. In contrast, Elberg et al. reported that ADP underestimated Δ%FM as
the %FM gain was increased in children (r = 0.35, P < 0.01) [20]. In the current study, we found,
in the Bland-Altman approach, no significant bias in Δ%FM, suggesting that the ability of ADP
to track %FM (r = 0.20, P = 0.17). These mixed results might be accounted for by the amount of
Δ%FM during follow-up periods. Along with Minderico et al., we conducted weight-loss
intervention studies that produced large Δ%FM (mean changes of greater than 3%) [13], while
Elberg et al. detected less than 1% of Δ%FM [20]. Similar results were also found when
comparing DXA to bioelectrical impedance, potassium counting and deuterium dilution [12,21].
However, our data also suggest that %FM from individuals with larger %FM by ADP and/or
with greater Δ%FM by ADP are overestimated as compared to DXA method. Thus, results of
similar interventions might be interpreted with cautions. To our knowledge, this is the first study
assessing the ability of ADP to detect Δ%FM in Asians populations using DXA as the reference
method. Asians have a generally greater %FM than Caucasians for the same BMI and age [14].
Therefore, it is important to examine the validity of ADP for body composition change in Asian
populations.
It is worth addressing the advantages of ADP uses in research, clinical, and public health
settings. ADP can measure a large number of participants and a wide range of participant types.

Page 9

9
It also has the advantages of lower initial investment and no radiation exposure, even though the
radiation exposure of DXA is extremely low. Since the current study revealed that ADP can
track Δ%FM with no significant bias, ADP can be applied to longitudinal intervention studies
around body composition to estimate Δ%FM as well as weight management strategies in
clinical or public health settings. Even though researchers and practitioners should consider the
possibility of a ± 2% standard error of estimate may exist in the estimation of Δ%FM when the
participants experienced approximately 7% of weight loss.
This study has several strengths to address. First, the study population was Asian men.
Previous studies with similar designs all included mainly Caucasians men and women [11-13],
whereas our study enrolled a Japanese population. Our results were mostly consistent with
previous studies, so the ethnic difference may not contribute to, or hinder, the ability of ADP to
track body composition changes. Second, this study elicited enough weight loss to investigate
the study purpose. Previous studies with similar designs produced 4.7–7.0% weight loss [11-13].
Our study also obtained 7.0% weight loss (approximately 10% weight loss in the dietary
program), allowing us to utilize one of the widest ranges of weight loss of all published studies.
However, there were also several limitations to the current study. First, we did not validate
ADP compared to the multi-compartment models. Obviously, DXA is not a gold standard and
seems to include certain limitations such as an ability to detect lean tissue mass changes in
regions with large areas of bone pixels. In addition to this, weight loss resulted in slight (1-2 %)
reduction in hydration fraction of fat-free mass according to the previous studies [22,23]. This
would potentially increase %FM determined by DXA. However, Nord and Payne reported
DXA was relatively insensitive to tissue hydration status compared to other clinical methods
using classical 2-compartment model [24]. Taken together, it is difficult to dismiss a possibility
that DXA itself could be the cause of the observed bias. To correct the bias, further
investigations would be warranted by using multi-compartment models. Second, we were
unable to consider any gender-specific or age-related differences, since our study recruited only
middle-aged men. Therefore, it remains unclear whether our results can be extrapolated to
women and older adults. In addition, we acknowledge our relatively small sample size as a
limitation.
In summary, the current study indicated that the mean changes in %FM were similar

Page 10

10
between the two methods investigated. DXA- and ADP-derived Δ%FMs were highly correlated,
after an approximately 7% weight loss in obese Japanese men. In the Bland-Altman approach,
measurement errors (DXA minus ADP) in Δ%FM were uniformly distributed across the ranges
of body composition changes. However, %FM from individuals with larger %FM by ADP
and/or with greater Δ%FM might be overestimated as compared to DXA method.

Literature cited
[1] Caterson ID, Hubbard V, Bray GA, Grunstein R, Hansen BC, Hong Y, Labarthe D,
Seidell JC, Smith SC: Prevention Conference VII: Obesity, a worldwide epidemic related
to heart disease and stroke: Group III: worldwide comorbidities of obesity. Circulation
2004;110:e476-e483.
[2] Yoshiike N, Seino F, Tajima S, Arai Y, Kawano M, Furuhata T, Inoue S: Twenty-year
changes in the prevalence of overweight in Japanese adults: the National Nutrition Survey
1976-95. Obes Rev 2002;3:183-190.
[3] Ishikawa-Takata K, Ohta T, Moritaki K, Gotou T, Inoue S: Obesity, weight change and
risks for hypertension, diabetes and hypercholesterolemia in Japanese men. Eur J Clin
Nutr 2002;56:601-607.
[4] Cui R, Iso H, Toyoshima H, Date C, Yamamoto A, Kikuchi S, Kondo T, Watanabe Y,
Koizumi A, Wada Y, Inaba Y, Tamakoshi A: Body mass index and mortality from
cardiovascular disease among Japanese men and women: the JACC study. Stroke
2005;36:1377-1382.
[5] Roubenoff R, Kehayias JJ, Dawson-Hughes B, Heymsfield SB: Use of dual-energy x-ray
absorptiometry in body-composition studies: not yet a "gold standard". Am J Clin Nutr
1993;58:589-591.
[6] Wellens R, Chumlea WC, Guo S, Roche AF, Reo NV, Siervogel RM: Body composition
in white adults by dual-energy x-ray absorptiometry, densitometry, and total body water.
Am J Clin Nutr 1994;59:547-555.
[7] Kohrt WM: Preliminary evidence that DEXA provides an accurate assessment of body
composition. J Appl Physiol 1998;84:372-377.
[8] Evans EM, Saunders MJ, Spano MA, Arngrimsson SA, Lewis RD, Cureton KJ:

Page 11

11
Body-composition changes with diet and exercise in obese women: a comparison of
estimates from clinical methods and a 4-component model. Am J Clin Nutr 1999;70:5-12.
[9] McCrory MA, Gomez TD, Bernauer EM, Molé PA: Evaluation of a new air
displacement plethysmograph for measuring human body composition. Med Sci Sports
Exerc 1995;27:1686-1691.
[10] Fields DA, Goran MI, McCrory MA: Body-composition assessment via air-displacement
plethysmography in adults and children: a review. Am J Clin Nutr 2002;75:453-467.
[11] Weyers AM, Mazzetti SA, Love DM, Gómez AL, Kraemer WJ, Volek JS: Comparison
of methods for assessing body composition changes during weight loss. Med Sci Sports
Exerc 2002;34:497-502.
[12] Frisard MI, Greenway FL, Delany JP: Comparison of methods to assess body
composition changes during a period of weight loss. Obes Res 2005;13:845-854.
[13] Minderico CS, Silva AM, Teixeira PJ, Sardinha LB, Hull HR, Fields DA: Validity of
air-displacement plethysmography in the assessment of body composition changes in a
16-month weight loss program. Nutrition & Metabolism 2006;3:32.
[14] Levitt DG, Heymsfield SB, Pierson RN, Shapses SA, Kral JG: Physiological models of
body composition and human obesity. Nutrition & Metabolism 2009;6:7.
[15] The Examination Committee of Criteria for 'Obesity Disease' in Japan; Japan Society for
the Study of Obesity: New criteria for 'obesity disease' in Japan. Circ J 2002;66:987-992.
[16] Kuzuya T: Prevalence of diabetes mellitus in Japan compiled from literature. Diabetes
Res Clin Pract 1994;24 Suppl:S15-S21.
[17] Sakata K, Labarthe DR: Changes in cardiovascular disease risk factors in three Japanese
national surveys 1971-1990. J Epidemiol 1996;6:93-107.
[18] Siri WE: Body composition from fluid spaces and density: analysis of methods. 1961.
Nutrition 1993;9:480-491.
[19] Bland JM, Altman DG: Statistical methods for assessing agreement between two methods
of clinical measurement. Lancet 1986;1:307-310.
[20] Elberg J, McDuffie JR, Sebring NG, Salaita C, Keil M, Robotham D, Reynolds JC,
Yanovski JA: Comparison of methods to assess change in children's body composition.
Am J Clin Nutr 2004;80:64-69.

Page 12

12
[21] Pødenphant J, Gotfredsen A, Engelhart M, Andersen V, Heitmann BL, Kondrup J:
Comparison of body composition by dual energy X-ray absorptiometry to other estimates
of body composition during weight loss in obese patients with rheumatoid arthritis. Scand
J Clin Lab Invest 1996;56:615-625.
[22] Jebb SA, Siervo M, Murgatroyd PR, Evans S, Frühbeck G, Prentice AM: Validity of the
leg-to-leg bioimpedance to estimate changes in body fat during weight loss and regain in
overweight women: a comparison with multi-compartment models. International Journal
of Obesity 2007;31:756-762.
[23] Ritz P, Sallé A, Audran M, Rohmer V: Comparison of different methods to assess body
composition of weight loss in obese and diabetic patients. Diabetes Res Clin Pract
2007;77:405-411.
[24] Nord RH, Payne RK: Dual-energy X-ray absorptiometry vs underwater weighing
comparison of strengths and weaknesses. Asia Pac J Clin Nutr 1995;4:173-176.

Page 13

13
Tables
Table 1 Comparison of body composition results obtained using dual-energy X-ray absorptiometry (DXA) and air displacement
plethysmography (ADP) at baseline and month 3 among middle-aged obese Japanese men.
Dietary program (n = 26)
Variables Baseline Month 3 Change Baseline
Age, yr 47.1 (7.9)
Height, cm 170.1 (5.1)
Body weight, kg
87.6 (11.2) 78.7 (11.2) -8.8 (3.6)e 87.8 (10.1) 84.4 (9.2) -3.4 (2.9)e 87.7 (10.6) 81.5 (10.6) -6.2 (4.3)e
BMI, kg/m2 30.2 (3.2) 27.1 (3.1) -3.1 (1.3)e 30.4 (3.4) 29.2 (3.0) -1.2 (1.0)e 30.3 (3.2) 28.1 (3.2) -2.2 (1.5)e
Overweighta, % 46.2 53.8 58.3
Obeseb, % 53.8 15.4 41.7
Waist circumference, cm
102.1 (7.8) 92.9 (8.8) -9.2 (4.5)e 102.4 (7.8) 97.9 (6.9) -4.5 (2.8)e 102.3 (7.7) 95.3 (8.3) -7.0 (4.4)e
Fat-free massc, kg 57.1 (6.4) 55.4 (5.7) -1.7 (1.6)e 58.8 (5.5) 58.5 (5.1) -0.3 (1.5) 57.9 (5.9) 56.9 (5.6) -1.0 (1.7)e
Bone massc, g 3046 (475) 3065 (466) 19 (126) 3074 (363) 3113 (394)
Body volumed, L 85.5 (11.6) 75.8 (11.7) -9.7 (4.1)e 85.1 (10.7) 81.1 (9.6) -4.1 (3.2)e 85.3 (11.0) 78.3 (10.9) -7.0 (4.6)e
Body densityd, kg/L 1.02 (0.01) 1.03 (0.01) 0.01 (0.01)e 1.03 (0.01) 1.03 (0.01) 0.01 (0.01)e 1.02 (0.01) 1.03 (0.01) 0.01 (0.01)e
Percentage of fat mass, %
DXA 34.5 (4.4) 29.1 (6.0) -5.4 (3.0)e 32.8 (4.1) 30.4 (4.9) -2.4 (1.8)e 33.7 (4.3) 29.7 (5.4) -3.9 (2.9)e
ADP 34.9 (4.6) 29.7 (5.4) -5.2 (3.3)e 31.8 (4.6) 29.2 (4.8) -2.6 (2.8)e 33.4 (4.8) 29.5 (5.1) -3.9 (3.3)e
Δ between DXA and ADP -0.4 (2.8) -0.6 (2.6) -0.2 (2.2) 1.0 (3.0)
Abbreviations: Δ, differences; BMI, body mass index. Values are means (SD), aBMI of 25-30 kg/m2, bBMI of greater than 30 kg/m2,
cMeasured by DXA, dMeasured by ADP, eSignificantly different from baseline.

Exercise program (n = 24)
Month 3
48.5 (9.4)
170.0 (6.3)

Combined (n = 50)
Month 3
47.8 (8.6)
170.1 (5.6)
Change Baseline


Change






62.5
33.3


52.0
48.0
58.0
24.0


39 (113) 3060 (421) 3088 (429) 29 (119)

1.2 (3.0) 0.2 (2.8) 0.3 (2.9) 0.2 (2.9) 0.0 (2.5)

Page 14

14
Table 2 Summary of simple regression analysis and Bland and
Altman approach for percentage of fat by air displacement
plethysmography (ADP)ct baseline and month 3 compared to
dual-energy X-ray absorptiometry (DXA).

Baseline Month 3
Simple regression analysis

Slope

0.71 0.91
Intercept 9.97 2.94
R 0.79 0.85
R2 0.63 0.72
SEE 2.62 2.92

P
< 0.01 < 0.01
Bland and Altman approach
Biasa 0.25 0.25
95% LoAb -5.52, 6.02 -5.48, 5.98
R -0.18 0.13

Abbreviations: SEE, standard error of estimate, LoA, limits of
agreement. aMean difference between DXA and ADP (i.e., DXA
minus ADP), bUpper and lower 95% LoA.
P
0.20 0.37

Page 15

15
Legends for figure

Figure 1
Simple regression analysis for changes (Δ) in percentage of fat mass (%FM) determined by
dual-energy X-ray absorptiometry (DXA) and air displacement plethysmography (ADP). The
bold solid line indicates the regression line between Δ%FM from DXA and ADP. The dashed
lines represent 95% prediction intervals.
●: participants engaged in dietary
program, ○:
participants engaged in exercise program, SEE: standard error of estimate.

Figure 2
Bland-Altman plots for the systematic bias in the estimation of changes (Δ) in percentage of fat
mass (%FM) between dual-energy X-ray absorptiometry (DXA) and air displacement
plethysmography (ADP). The middle solid line indicates the mean difference between Δ%FM
from DXA and ADP. The upper and lower dashed lines represent limits of agreement (± 2 SD
from the mean). ●: participants engaged in dietary program, ○: participants engaged in exercise
program.

Figure 3
Relation of changes (Δ) in percentage of fat mass (%FM) determined with dual-energy X-ray
absorptiometry (DXA) minus the Δ%FM estimated from air displacement plethysmography
(ADP) to baseline %FM determined by ADP (left) and the Δ%FM determined by ADP (right).
The middle solid line indicates the mean difference between Δ%FM estimated from DXA and
ADP. The upper and lower dashed lines represent limits of agreement (± 2 SD from the mean).
●: participants engaged in dietary program, ○: participants engaged in exercise program.