T. KAMO1,2, H. ISHII1, D. TAKAHASHI1, K. IWAGAYA3, T. ISHIDA1, Y. NISHIDA1
1. Division of Rehabilitation Sciences, Seirei Christopher University, Shizuoka, Japan; 2. Nursing home, Sakuranosono, Shizuoka, Japan; 3. Care house, Dainiadonaikan, Shizuoka, Japan.
Corresponding author: Tomohiko Kamo, 3453 Mikatahara-machi, Kitaku, Hamamatsu-shi, Shizuoka, Japan. Tel: +81-090-7682-1192, e-mail:firstname.lastname@example.org
J Frailty Aging 2014;3(4):211-215
Published online December 9, 2014, http://dx.doi.org/10.14283/jfa.2014.26
Background: Body composition is an important component of health related fitness. Near-infrared spectroscopy (NIRS) is a non-invasive, simple and rapid method of assessing body fat percentage. However, it is unknown whether NIRS can accurately estimate FFM in community-dwelling frail elderly. Objectives: This study aimed to compare NIRS with bioelectrical impedance analysis (BIA) in FFM measurement. Design: Cross- sectional study. Setting: Shizuoka, Japan. Participants: The study population comprised 53 community-dwelling frail elderly (15 men, 38 women; mean age 84.8±6.4 years; body mass index 19.7±3.5 kg/m2). Measurement: FFM and percentage fat mass (%FM) were estimated using a NIRS device at two sites (biceps and calf) and compared to body composition measured by BIA. Simple linear regression and Bland–Altman analyses were used to determine agreement between the methods. Results: FFM determined by BIA highly correlated with that determined by NIRS at both the biceps and calf (r=0.92 for both; p<0.001). The correlation coefficients for %FM estimated by NIRS were slightly lower (r=0.70 for biceps; r=0.66 for calf). In NIRS assessments, systematic biases were found for %FM but not for FFM. Conclusion: NIRS has significant potential for body composition analysis. Further comparative and longitudinal studies need to be conducted using an agreed reference analysis method to find a simple and more suitable method that can be applied among the community-dwelling frail elderly.
Key words: Body composition, near-infrared spectroscopy, Bland-Altman analysis.
Body composition is an important component of health- related fitness. In the older population, changes in body composition, especially declines in fat free mass, are relevant research topics especially in the context of the study of physical performance. Several studies have shown inconsistent results examining the capacity of muscle and/or fat mass of predicting physical performance among older adults (1). Lee et al. reported that higher body fat was associated with functional disability in an older Asian population (2). Other studies (3) found that low fat-free mass is an independent predictor of functional disability. In addition, both low fat-free mass and high fat mass were both independent predictors of functional disability in community-dwelling Caucasian men and women (4).
There are several FFM measurement methods available, but a gold standard tool has not yet been developed for the highly heterogeneous older adult population. Dual-energy X-ray absorptiometry (DXA) is frequently used to assess FFM, but it is not ideal for the routine assessment of large populations. Bioelectrical impedance analysis (BIA) is also a commonly used method for estimating body composition; it is non-invasive, portable, quick, and inexpensive (5). Indeed, previous studies have shown that there is a strong correlation between the BIA parameters and skeletal muscle measurements at both arms (6) and legs (7). In addition, Genton et al. reported that the BIA accurately predicts FFM in elderly between 65 and 94 years of age (8). Other studies also found that BIA has
a high reliability and accuracy in community-dwelling older persons (9, 10). Nevertheless, the BIA method has limitations because fat tissue may hold some amount of water (11) and may overestimate the true muscle mass. In particular, there are a large proportion of older adults who present abnormal distribution of body water (e.g. edema).
A method with promising potential for use in older adults is near-infrared spectroscopy (NIRS). Similar to BIA, NIRS is also a non-invasive, simple and rapid method for assessing body fat percentage. The NIRS method is based on the principles of light absorption and reflection (12). Only a few seconds are required to enter participant’s data into the mini- computer and obtain the NIRS measurement from a placed at the subject’s arm. NIRS is a cheaper and easier to use and transport compared to BIA. In addition, NIRS can be measured also in people with edema of the leg. NIRS has been reported to have a high reliability and accuracy (13-15). For example, Josse et al. reported that the correlation coefficient between fat mass as predicted by the Fourier transformed NIRS method and DXA was 0.95 (P<0.001)(16). Unfortunatel, there has been no study to date investigating the validity of NIRS compared to BIA in community-dwelling frail elderly. In older adults, in which body composition and hydration status frequently change with aging (17), whether NIRS can accurately measure body composition remains unanswered. The purpose of this study was to investigate the validity of NIRS for FFM estimation by comparing its results with those obtained by BIA in community-dwelling frail elderly.
The study population consisted of 53 community-dwelling frail elderly (15 men, 38 women, aged ≧65) who were eligible for long-term care insurance, lived in Iwata City, and were provided various home care services from the Iwata City Health Care Service Foundation for Older People. Prior to the data collection, oral or written informed consent was obtained from the participants or their relatives or legal guardians in cases in which geriatric residents were incapable of providing such consent.
The following inclusion criteria were adopted in the present study:
– Frailty status as certified by the long-term care insurance service;
– Age of 65 years and older;
– Living in the community.
The exclusion criteria were:
– Unstable cardiac conditions (such as ventricular dysrhythmias, pulmonary edema) or other musculoskeletal conditions;
– Hyperhydration; Implanted defibrillators.
We used a commercially available NIRS device (Fitness Analyzer BFT-3000, Kett Electrical Laboratory, Tokyo, Japan). The device is based on technology from the United States Department of Agriculture, with an NIR measurement estimating range between 2.5 and 50.0% (17). NIRS is based on light absorption and reflection using near-infrared light emission (18). The main body of the device is connected via a light cable to a microphone-size light-emitting sensor. The NIR sensor window is equipped with a light shield prior to placing it on the mid-upper arm to ensure that no external light interferes with the estimation of body fat percentage. The operator is then required to input the subject’s data into the mini-computer, following which the results of NIR measurement are obtained while the sensor remains on the participant’s arm. In our study, NIR was measured by placing the BFT-3000 sensor on the upper arm of the non-dominant hand of the participants. The process was repeated for the lower leg. FFM and %FM were obtained at the distal biceps (5 cm from the olecranon; NIRSbicep) and at the proximal calf (5 cm from the caput fibulae; NIRScalf). NIRS measurements were performed by a single trained physical therapist and completed within a few minutes.
Body composition was also estimated using a segmental multifrequency BIA instrument that operated at frequencies of 5, 50, and 250 kHz (ioi353s, Owa Corporation, Tokyo, Japan). Participants removed their socks, stood on two metallic electrodes on the floor scale barefoot and held metallic grip
electrodes placed in the palm of the hand, with the fingers holding the handrails.
All data were normally distributed and expressed as means
± standard deviations. Differences between men and women were analyzed using the unpaired t-test. Bland–Altman analysis was used to determine agreement between the body composition variables assessed by criterion methods (19). Pearson’s correlation coefficient was calculated to determine the relationship between variables. A p value of <0.05 was considered statistically significant. Statistical analysis was performed using SPSS statistical software (ver. 19.0 for Windows; IBM SPSS Japan, Tokyo, Japan).
Table 1 presents the general characteristics of the participants. In total, 53 elderly residents (15 men, 38 women) with a mean age of 84.8 ± 6.4 years (66–98 years) were enrolled in the study. The mean BMI was 19.7 ± 3.5 kg/m2 (range 12.8 – 32.3 kg/m2). There were significant differences in age, height, weight, FFM, %FM between men and women, but not for BMI.
Table 2 summarizes the correlation coefficients between body composition parameters assessed by NIRS and BIA. FFM determined by BIA highly correlated with that determined by NIRS at both the biceps and calf (r=0.92 for both; p<0.001). The correlation coefficients for %FM estimated by NIRS at both biceps and calf were slightly lower than that by BIA (r=0.70 and 0.66, respectively).
The limits of agreement for NIRS (mean bias and 95% confidence interval [95%CI]) were narrow for FFM (mean
bias -0.7, 95%CI: −4.9 to 3.5 kg for biceps; mean bias -1.0,
95%CI: −5.8 to 3.8 kg for calf) and wider for %FM (mean bias 0.2, 95%CI: −10.0 to 10.4 for biceps; mean bias 0.8, 95%CI:
−10.2 to 11.8 for calf). The mean differences between FFMBIA and FFMNIRS at both biceps and calf were not significant (mean difference -0.70 kg for biceps; mean difference -1.0 for calf). Compared to BIA, systematic biases were found for the assessment of %FM, but not of FFM using NIRS (biceps and calf). NIRS overestimated %FM when it was low and underestimated %FM when it was high (Figure 1). The bias between FFM assessed by BIA and NIRS was inversely related to BMI (r=−0.33, p<0.05 for biceps; r=−0.45, p<0.01 for calf) and positively associated with height when measured at the biceps (r=0.38, p<0.01). The bias between %FM assessed by BIA and NIRS was inversely related to height (r=−0.33, p<0.05 for biceps) and positively associated with BMI when measured at the calf (r=0.31, p<0.05).
This study compared two of the routinely accessible methods for predicting body composition (i.e. BIA and NIRS) in the assessment of FFM in community-dwelling and assisted-living elderly. FFM estimates were comparable between NIRS and BIA with no significant differences. This finding is consistent with data previously reported by Yoshimatsu et al. (21). Moreover, FFM measurement obtained by BIA and NIRS confirmed to be highly correlated (21). A recent longitudinal study analyzing NIRS concluded that the method has a high degree of reproducibility and may be suitable for detecting longitudinal changes in body fat (22).
In the present study, the NIRS data were obtained at the distal biceps and proximal calf. With regard to FFM and
%FM, the NIRS data showed a good correlation coefficient with the BIA data. Previous studies have demonstrated higher correlations with %FM at thinner adipose sites than at thicker adipose sites when using the BFT-2000 (23) and Futrex 5000
(15) NIRS devices. Inconsistent associations at the various sites might simply be the result of differences in the depth of penetration of the infrared radiation. These results therefore suggest that it might be preferable to perform measurements at sites with minimal subcutaneous fat.
Our study confirmed the bias associated with increasing
%FM. NIRS devices tend to overestimate %FM at a low %FM and underestimate it at a high %FM (Figure 1). In contrast, no systematic bias was found for FFM estimation using NIRS. The results of the present study, in which the participants had a high amount of adiposity, suggest the possibility of underestimating body fat percentage using NIRS rather than BIA. This tendency of NIRS to underestimate fat measurements has been previously observed in a comparison with standard hydrodensitometry in healthy individuals (24). In that study, the underestimation of NIRS tended to be enhanced as the proportion of fat increased. Furthermore, when a separate small group of obese women with BMI of >50 kg/m2 were studied,
the underestimate was as much as 16% against the comparison technique. In a validation study using healthy subjects, estimation of body fat using NIRS was overestimated for leaner subjects and underestimated for more obese subjects compared with that using hydrodensitometry (25). It is clear then that subcutaneous fat may affect the results obtained by NIRS.
In the present study, the mean BMI of men and women was 19.2±2.7 kg/m2 and 20.0±3.7 kg/m2, respectively; no participant had a BMI higher than 35 kg/m2. Previous studies of older Japanese individuals reported a BMI ranging from 19.9 to 23.3 kg/m2 (11, 26). Thus, the impact of subcutaneous fat on NIRS data might be minimal, suggesting NIRS as a valid method for assessing their percentage of body fat and fat mass.
The present study had several limitations. First, we cannot consider that one method of body composition analysis is more accurate than the other because there was no recognized gold standard for comparison. Second, most study participants were women. Moreover, the sample size was small and studies with a larger number of participants are required to confirm and extend our findings.
In conclusion, NIRS results are comparable with the more sophisticated BIA method and potentially suitable for use in body composition analysis. Further comparative and longitudinal studies are needed to find a simple and more suitable method to use in community-dwelling frail elderly. This could be aided by seeking agreement on a suitable reference method of body composition analysis. In the meantime, we propose that NIRS may have merit as a portable, quick and easy to perform method of assessing body composition.
1. Visser M, Harris TB, Langlois J, et al. Body fat and skeletal muscle mass in relation to physical disability in very old men and women of the Framingham heart study J Gerontol A Biol Sci Med Sci 1998;53:214–221.
2. Lee JS, Auyeung TW, Kwok T, Lau EM, Leung PC, Woo J. Associated factors and health impact of sarcopenia in older Chinese men and women: a cross-sectional study. Gerontology 2008;53:404–410.
3. Reid KF1, Naumova EN, Carabello RJ, Phillips EM, Fielding RA. Lower extremity muscle mass predicts functional performance in mobility-limited elders. J Nutr Health Aging 2008;12:493–498.
4. Broadwin J1, Goodman-Gruen D, Slymen D. Ability of fat and fat-free mass percentages to predict functional disability in older men and women. J Am Geriatr Soc 2001;49:1641-1645.
5. Tanimoto Y, Watanabe M, Higuchi Y, Hirota C, Kono K. Evaluation of the best indicator of muscle mass in community-dwelling elderly persons. Nippon Ronen Igakkai Zasshi 2008;45:213-219.
6. Miyatani M, Kanehisa H, Fukunaga T. Validity of bioelectrical impedance and ultrasonographic methods for estimating the muscle volume of the upper arm. Eur J App Physiol 2000;82:391-396.
7. Miyatani M, Kanehisa H, Masuo Y, Ito M, Fukunaga T. Validity of estimating limb muscle volume by bioelectrical impedance. J Appl Physiol 2001;91:386-394.
8. Genton L, Karsegard VL, Kyle UG, et al. Comparison of four bioelectrical impedance analysis formulas in healthy elderly subjects. Gerontology 2001;47:315-323.
9. Aleman-Mateo H, Rush E, Esparza-Romero J, et al. Prediction of fat-free mass by bioelectrical impedance analysis in older adults from developing countries: A cross-validation study using the deuterium dilution method. J
Nutr Health Aging 2010;14:418-426.
10. Bertoli S1, Battezzati A, Testolin G, Bedogni G. Evaluation of air- displacement plethysmography and bioelectrical impedance analysis vs dual-energy x-ray absorptiometry for the assessment of fat-free mass in elderly subjects. Eur J Clin Nutr 2008;62:1282-1286.
11. Yamada Y, Kimura M, Nakamura E, Masuo Y, Oda S. Limb muscle mass decrease with aging in Japanese men and women aged 15-97 yr. Jpn J Phys Fitness Sports Med 2007;56:461-472.
12. Conway JM, Norris KH, Bodwell CE. A new approach for the estimation of body composition: infrared interactance. Am J Clin Nutr 1984;40:1123- 1130.
13. Conway JM, Norris KH, Bodwell CE. A new approach for the estimation of body composition: infrared interactance. Am J Clin Nutr 1984;40:1123- 1130.
14. Sawai S, Shirayama M, Mutoh Y, Miyashita M. Estimation of body fat by near infrared spectroscopic technique. Jpn J Phys Fitness Sports Med 1990;39:155-163.
15. Fuller NJ, Dewit O, Wells C. The potential of near infrared interactance for predicting body composition in children. Eur J Clin Nutr 2001;55:967- 972.
16. Josse AR, Azizian H, French SB, Kramer JK, Phillips SM. Body fat content determination in premenopausal, overweight, and obese young women using DXA and FT-NIR. Obesity (Silver Spring) 2011;19:1497- 502.
17. Waki M1, Kral JG, Mazariegos M. Relative Expansion of Extracellular Fluid in Obese vs. Nonobese Women. Am J Physiol 199;261:199-203.
18. Futrex-5000a/ZL User’s Manual: Body Fat Fitness Computer version 8.4;
Futrex inc., Galtherburg, MD; 1996.
19. Lukaski HC. Methods for the assessment of human body composition: traditional and new. Am J Clin Nutr 1987;46:537-556.
20. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;301–310.
21. Yoshimatsu T, Yoshida D, Shimada H, Komatsu T, Harada A, Suzuki
T. Relationship between near-infrared spectroscopy, and subcutaneous fat and muscle thickness measured by ultrasonography in Japanese community-dwelling elderly. Geriatr Gerontol Int 2013;13:351-357.
22. Kalantar-Zadeh K, Block G, Kelly MP, Schroepfer C, Rodriguez RA, Humphreys MH. Near infra-red interactance for longitudinal assessment of nutrition in dialysis patients. J Ren Nutr. 2001;11:23-31.
23. Yasukawa M, Horvath SM, Oishi K, Kimura M, Williams R, Maeshima
T. Total body fat estimations by nearinfrared interactance, A-mode ultrasound, and underwater weighing. Appl Human Sci 1995;14:183-189.
24. Elia M, Parkinson SA, Diaz E. Evaluation of near infrared interactance as a method for predicting body composition. Eur J Clin Nutr 1990;44:113- 121.
25. Brooke-Wavell K, Jones PR, Norgan NG, Hardman AE. Evaluation of near-infrared interactance for assessment of subcutaneous and total body fat. Eur J Clin Nutr. 1995;49:57-65.
26. Yamada Y, Schoeller DA, Nakamura E, Morimoto T, Kimura M, Oda
S. Extracellular water may mask actual muscle atrophy during aging. J Gerontol A Biol Sci Med Sci 2010;65:510-516.