# Letter to the editor-in-chief.

In the April issue of this journal Heiss et al. (1) reported on the effect of a 1 litre fluid load on body composition measured by air displacement plethysmography (ADP) and bio-electrical impedance. In table 2 they reported mean body densities before fluid load of 1.087 g/[cm.sup.3] and 1.082 g/[cm.sup.3] after the ingestion of 1 litre of water. Based on the SIRI-equation (2) for converting body density into a percentage body fat (PF) these densities correspond to PF's of 5.4% and 7.5% respectively which seems very unlikely given their sample was composed of 29 women and 5 men, and not for example elite male athletes. It also is in contrast with the respective mean PF's of 23.6% and 24.7% reported in the same table.

Apart from these unlikely values, in their discussion they state that the density of water is 1.0 g/[cm.sup.3] ... , halfway between the assumed densities of fat and lean mass,...' and therefore '...the water load should not have influenced body density'. This statement is only correct if the density of their subjects before the fluid load was about 1.0000 g/[cm.sup.3], corresponding to a PF of 45%. In fact it is relatively straightforward to derive mathematically what 'really' happens to a subject's body density when 1 litre of water is 'added' and how this would translate into a change in PF when using ADP and the SIRI-equation, as the assumption of this equation are no longer valid: adding 1 litre of water changes the assumed composition of fat free mass (FFM) in the SIRI-equation.

In the SIRI equation it is assumed that the fat mass (FM) (triglycerides) has a density of 0.9 g/[cm.sup.3]. The fat-free mass (FFM) is composed of 73.8% of water, 6.8% of mineral and 19.4% of protein with respective densities (d) of 0.994 g/[cm.sup.3], 3.038 g/[cm.sup.3] and 1.34g/[cm.sup.3]. Entering these values in equation 1 and solving for dFFM results in the density of FFM (dFFM) being 1.1 g/[cm.sup.3].

FFM/dFFM = [Mwater/dwater] + [Mprot/dprot] + [M mineral/d mineral] (eq.1)

If an average subject, based on the values in table 2 of Heiss et al. (1) has 23.6% fat and weighs 67.6 kg before the fluid load, one obtains values of 51.646kg for FFM and 15.954kg of FM. Entering these values with their respective densities in equation 2 and solving for 'BD' results in a body density (BD) of 1.0451845 g/[cm.sup.3] for this subject.

M/BD = [FM/dFM] + [FFM/dFFM] (eq.2)

Assuming the constitution of FFM in the SIRI-equation is indeed correct for this subject his/her FFM is composed out of 38.115 kg of water, 10.019 kg of protein and 3.512kg of minerals. Now what happens when this subject ingests 1 litre of water, which we assume to be at body temperature? As water is part of the FFM the FFM of our subject will increase with 0.994 kg as will the body weight, resulting in a body mass of 68.594 kg, a FFM of 52.640 kg and an unchanged FM of 15.954kg. The resulting actual PF now is [(15.954/68.594).sup.*] 100=23.3% vs 23.6% before the fluid load. The composition of this subject's FFM has now changed and consists of 39.109 kg of water (=74.295% of FFM), 10.019 kg of protein (=19.033% of FFM) and 3.512 kg of minerals (=6.672% of FFM). Substituting the appropriate values in equation 1 and solving for dFFM results in a dFFM of 1.0971705 g/[cm.sup.3], as opposed to 1.1 g/[cm.sup.3] before the fluid load. If one solves equation 2 for density in the fluid-loaded situation, taking into account the changed density of FFM and the changed FFM one arrives at an 'actual' body density of 1.0439553 g/[cm.sup.3]. If one, erroneously, applies the SIRI equation, which assumes dFFM=1.1g/[cm.sup.3] to this actual body density this results in a PF of 24.2%. If volume measurement by ADP would be perfect, the actual body density would indeed equal 1.0439553 g/crm for the above subject. The 'actual' change in PF by ingesting 1 litre of water of -0.3% fat is however registered as a change of +0.6% fat if one still assumes that dFFM=1.1g/[cm.sup.3] by using the SIRI-equation.

It is regrettable that the reported values in table 2 are so unlikely, but indeed a rise in PF is registered in the fluid-load condition as is predicted by the above calculations. This merely reflects the limitations of assessing body composition in a 2 component model, dividing body mass into fat mass and fat free mass, with the assumption of a constant composition of the fat-free mass. In my opinion the actual 'test' for the accuracy of ADP in detecting this change in body composition lies in establishing how well the changed density after fluid load is registered compared to what 'really' happens to it and should not be confused with the inherent limitations of a 2-component model of body composition.

Sincerely,

Maarten.W. Peeters, Ph.D.

ACKNOWLEDGMENTS

M.W. Peeters is supported as a postdoctoral fellow by the Research Foundation Flanders.

Address for correspondence: Peeters M.W., Ph.D., Department of Biomedical Kinesiology, Research Centre for Exercise and Health, Faculty of Kinesiology and Rehabilitation Sciences Katholieke Universiteit Leuven, Leuven, Belgium, Phone: +32 16 32 90 85, Fax: +32 16 32 91 97, e-mail: maarten.peeters@faber.kuleuven.be

REFERENCES

(1.) Heiss CJ, Gara N, Novotny D, Heberle H, and Morgan L, Stufflebeam J, Fairfield M. Effect of a 1 liter fluid load on body composition measured by air displacement plethysmography and bioelectrical impedance. JEPonline 2009; 12(2):1-8.

(2.) Siri WE. Body composition from fluid spaces and density: analysis of methods. In: Brozek J, Henschel A, editors. Techniques for Measuring Body Composition. Washington DC: National Academy of Sciences/National Research Council, 1961; 223-34.

Response to Dr. Peeter's Letter

Maarten Peeters brought up two points regarding our article published in the April 2009 issue of this journal on the effect of a 1 litre fluid load on body composition measured by air displacement plethysmography (ADP) and bio-electrical impedance.

First, he pointed out that the body densities reported in table 2 were not congruent with the percent body fat reported for air displacement plethysmography. In fact, the body volumes and body densities reported in both tables one and two were in error. The tables should have included the following values.

Dr. Peeters also provided an excellent mathematical explanation of what happens to body density when a liter of fluid is added. In our discussion we presented an error in logic regarding this effect. We appreciate Dr. Peeters' astute observation of our data and correction of our flawed statement. We regret the errors.

The purpose of the study was to determine if a 1 liter fluid load would influence body composition results using three different devices commonly used by clinicians and practitioners in the field. Despite the errors in the manuscript, the results remain the same. The practitioner using these devices needs to know that results may be influenced a small amount by the recent consumption of a liter of fluid.

Sincerely,

Cindy Heiss, Ph.D., RD

Metropolitan State College of Denver, Department of Health Professions, Denver, CO
```Table 2. Body composition before and after a 1 liter fluid load

Measure                Pre fluid

Mean  [+ or -]  SD

HBIA (1)

%fat                   20.2 [+ or -] 8.7
Lean mass (kg)         51.6 [+ or -] 12.5
Fat mass (kg)          14.0 [+ or -] 9.7

A-LBIA (2)

%fat                   23.6 [+ or -] 7.7
Lean mass (kg)         49.7 [+ or -] 11.3
Fat mass (kg)          17.5 [+ or -] 9.8

%fat                   23.6 [+ or -] 9.1
Lean mass (kg)         50.9 [+ or -] 8.9
Fat mass (kg)          16.7 [+ or -] 9.8
Body volume(l)         64.8 [+ or -] 15.0
Body density (g/ml)   1.046 [+ or -] 0.020

Measure                Post fluid            P-value

Mean  [+ or -]  SD

HBIA (1)

%fat                   20.6 [+ or -] 8.6     0.001
Lean mass (kg)         51.5 [+ or -] 12.6    0.63
Fat mass (kg)          14.0 [+ or -] 9.6     0.63

A-LBIA (2)

%fat                   24.3 [+ or -] 7.9     0.007
Lean mass (kg)         51.2 [+ or -] 9.1     0.16
Fat mass (kg)          17.3 [+ or -] 8.8     0.85

%fat                   24.7 [+ or -] 9.0     0.001
Lean mass (kg)         51.4 [+ or -] 8.9     0.28
Fat mass (kg)          17.6 [+ or -] 9.7     0.001
Body volume(l)        65.82 [+ or -] 15.0    0.001
Body density (g/ml)   1.043 [+ or -] 0.020   0.001

(1) HBIA: Hand-to-Hand Bioelectrical Impedance
Analysis; N=34

(2) A-LBIA: Arm-to-Leg Bioelectrical Impedance
Analysis; N=35

(3) ADP: Air displacement plethysmography using
the BOD POD: N=35

Table 3. Body composition before and after
a 10 minute interval with no fluid load
(control)

Measure                 Time 0

Mean  [+ or -]  SD

HBIA(1)

%fat              25.5 [+ or -] 6.9
Lean mass (kg)    52.4 [+ or -] 10.0
Fat mass (kg)     18.4 [+ or -] 7.6

A-LBIA (2)

%fat              27.2 [+ or -] 8.0
Lean mass (kg)    51.0 [+ or -] 11.5
Fat mass (kg)     19.6 [+ or -] 7.4

%fat              30.4 [+ or -] 9.9
Lean mass (kg)    48.9 [+ or -] 11.9
Fat mass (kg)     21.9 [+ or -] 10.5
Body volume(l)    68.8 [+ or -] 15.0
Body Density     1.031 [+ or -] 0.022
(g/ml)

Measure                 Time 1          P-value

Mean  [+ or -]  SD

HBIA (1)

%fat              26.0 [+ or -] 7.0     0.008
Lean mass (kg)    52.1 [+ or -] 10.1    0.01
Fat mass (kg)     18.7 [+ or -] 8.0     0.01

A-LBIA (2)

%fat              26.2 [+ or -] 6.8     0.22
Lean mass (kg)    51.7 [+ or -] 11.5    0.20
Fat mass (kg)     18.6 [+ or -] .4      0.22

%fat              30.6 [+ or -] 9.8     0.64
Lean mass (kg)    48.8 [+ or -] 11.7    0.55
Fat mass (kg)     22.0 [+ or -] 10.5    0.57
Body volume(l)    68.8 [+ or -] 15.0    0.43
Body Density     1.030 [+ or -] 0.022   0.72
(g/ml)

(1) HBIA: Hand-to-Hand Bioelectrical
Impedance Analysis; N=10

(2) A-LBIA: Arm-to-Leg Bioelectrical
Impedance Analysis; N=10