Professional Documents
Culture Documents
Segmental Bioelectrical Impedance Analysis An Update
Segmental Bioelectrical Impedance Analysis An Update
CURRENT
OPINION Segmental bioelectrical impedance analysis:
an update
Leigh C. Ward
Purpose of review
Bioelectrical impedance analysis is a popular, noninvasive and practical method for assessment of body
composition. The last decade has seen the development of impedance analyzers designed to assess the
composition of body segments as well as the whole body. This review outlines the theoretical basis for
segmental impedance analysis, validity and use in practice.
Recent findings
Segmental impedance analysis tends to underestimate fat-free mass and overestimate fat mass when
compared to reference techniques, although the magnitude of these differences can be small. Performance
is improved with population-specific prediction equations; algorithms in-built into instrument firmware should
not be relied upon. Prediction of whole-body composition from the sum of the individual segments, although
theoretically preferable, shows little advantage over whole body wrist to ankle impedance approaches.
Prediction of appendicular skeletal muscle mass, although promising, requires further research. The use of
measured impedance data directly as indices of composition, rather than for prediction, has not found
extensive application in nutritional research despite its success in other fields.
Summary
Segmental bioimpedance techniques have advanced substantially in recent years due to availability of
simple-to-use analyzers and simplified measurement protocols. The method has been well validated and
increasingly adopted in nutritional and clinical practice. Segmental impedance, like conventional whole
body impedance approaches, provides indirect prediction of body composition whose accuracy is yet to
achieve that of reference techniques such as magnetic reference imaging. This lack of accuracy, however,
is outweighed by the method’s practicality of use in many settings.
Keywords
body composition, eight-electrodes, fat-free mass, fluid distribution, segmental bioelectrical impedance,
skeletal muscle
either directly or indirectly, these compartments. analyzers [5 ]. Despite the wide variety of equipment
Garrow [1], while reviewing the available tech- and measurement methodologies now available, the
niques for measuring body composition, stated that basic measurement principles remain the same.
the ideal method should ‘be relatively inexpensive,
require little inconvenience for the individual, be
operated by unskilled technicians and yield highly
School of Chemistry and Molecular Biosciences, The University of
reproducible and accurate results’. Few methods Queensland, St. Lucia, Brisbane, Australia
meet these requirements, one being bioelectrical Correspondence to Associate Professor Leigh C. Ward, School of
impedance analysis (BIA). Chemistry and Molecular Biosciences, The University of Queensland,
BIA first came to prominence in the mid-1980s St Lucia, Brisbane QLD 4072, Australia. Tel: +61 7 3365 4633; fax: +61
with the first commercial availability of impedance 7 3365 4699; e-mail: l.ward@uq.edu.au
analyzers and following the seminal publication of Curr Opin Clin Nutr Metab Care 2012, 15:424–429
Lukaski et al. [2]. Since then there has been rapid DOI:10.1097/MCO.0b013e328356b944
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Segmental bioelectrical impedance analysis: an update Ward
1363-1950 ß 2012 Wolters Kluwer Health | Lippincott Williams & Wilkins www.co-clinicalnutrition.com 425
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Assessment of nutritional status and analytical methods
important outcome of this study was that the equi- component model of body composition. They
potential approaches overestimates limb resistance reported high correlations between the two
by between 3 and 8% when compared to the paired methods for percentage total body fat with no sig-
electrode approach. This implies that the equipo- nificant difference in group mean values for men
tential method determines the resistance of a but significant underestimation (2.5–3.0%) by BIA
‘virtual’ limb that includes a small contribution in women. The standard errors of estimation ranged
of the trunk. Nevertheless, the convenience of from 2.8 to 3.5% body fat.
stand-on impedance devices that do not require A mean difference in whole-body FFM, com-
the individual to undress outweighs this small puted from segmental analyses, of 1.6 kg and wide
source of error. Furthermore, as long as these over- limits of agreement (1.8 to 4.9 kg) between BIA and
estimations are reasonably constant between DXA was reported by Chao et al. [14] in young men.
individuals, the consequential error in predicting This difference was decreased to 0.04 kg (3.2 to
segmental FFM is automatically corrected for 3.2 kg) when a population-specific segmental
when the method is calibrated against techniques equation was used.
&&
such as dual-energy X-ray absorptiometry (DXA) Ling et al. [15 ] compared segmental BIA with
or MRI. DXA in 484 middle-aged men and women. They
reported excellent agreement (intraclass correlation
coefficients >0.93) between the two methods.
PREDICTING BODY COMPOSITION FROM Difference in whole body lean mass was only
SEGMENTAL IMPEDANCE 0.2 kg in normal weight individuals, although lean
MEASUREMENTS mass was underestimated by 1.24 and 4.1 kg in the
Impedance measurements are most appropriately overweight and obese, respectively. In contrast,
used to predict body fluid volumes or immediately body fat mass was overestimated in the overweight
derived values such as FFM. Consequently, predic- and obese by 1.98 and 4.85 kg, respectively. Limits of
tions of segmental composition are typically derived agreement (2 SD) were approximately more or less
using an algorithm based on Eq. (1), wherein v than 4.5 kg (approximately 10 and 25% of lean
represents the volume of tissue water or FFM. In and fat masses, respectively) in both men and
practice, Eq. (1) is rarely solved directly. Instead, Eq women suggesting that, although performing well
(1) is ‘calibrated’ empirically by regressing the impe- at a group level, segmental BIA may have limited
dance quotient (l2/R) against measurements of value for predicting body fat in an individual.
segment composition obtained from DXA or MRI, The segmental approach for predicting whole
the slope of the regression line equating to the body composition has also been applied to children
‘apparent’ resistivity. [16]. FFM was significantly underestimated
In segmental analysis, length should be that of (P < 0.003) by BIA compared to FFM measured by
the segment, however, this is rarely used; instead, deuterium oxide dilution, irrespective of sex and
segment length is assumed to be proportional to age, in a large group of Gambian children aged from
height and this approach is used in eight-electrode 5 to 18 years. Although the differences were signifi-
style analyzers. The validity of this assumption cant, the absolute magnitude was relatively small,
across all populations is open to question and approximately 4%.
may contribute to the relatively poor performance Despite the main advantage of segmental BIA
of segmental impedance in some populations. being the ability to predict the composition of
An additional problem presents itself when the individual body segments, there have been
calculating total body composition using Eq. (3). relatively few studies validating this approach.
&&
In most studies, the resistivity of segments is Ling et al. [15 ] observed that BIA generally
assumed to be constant. Owing to the different underestimated the lean masses of all body seg-
tissue composition of body segments, this is ments (arms, legs and trunk) when compared to
unlikely to be the case, and Zhu et al. [12] have DXA measurements. These underestimates were
noted that the use of segment-specific resistivity smaller for the arms (approximately 5%) than legs
values can improve prediction of fluid volumes in (7%) and, particularly, the trunk (12%). Similar
dialysis patients. observations were made by Chao et al. [14] who
determined the limits of agreement as approxi-
mately more or less than 0.6 kg (20%) and 1.9 kg
SEGMENTAL IMPEDANCE ANALYSIS AND kg (17%) for the arms and legs, respectively. Mally
BODY COMPOSITION &
et al. [17 ] also found underestimation (6–18%) of
Gibson et al. [13] validated, in a large multiethnic appendicular lean mass in older men but overesti-
study, two octapolar analyzers against a four- mated in the arms of women (13%); the reverse
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Segmental bioelectrical impedance analysis: an update Ward
was found for fat mass, particularly for truncal fat measurements have come to prominence in a num-
(60% overestimation). ber of fields. The measurement of impedance of a
As the prevalence of obesity has increased across sublimb segment, the calf, has become the method
the world, interest has focussed on measuring body of choice when impedance is used to monitor fluid
fat mass rather than the FFM. It is easy to estimate fat status during dialysis [24,25]. This technique is more
mass from BIA-derived measures of whole body FFM sensitive and precise than whole body or limb
by subtracting FFM from total body mass. This is measurements. Typical coefficients of variation
readily achieved since most commercially available were reported as less than 0.5%. The technique
segmental analyzers are incorporated into a digital has recently been adapted to assess body compo-
stand-on weighing scale. Determination of segment sition of dialysis patients based on measurement of
weight is, however, extremely difficult, yet, most arm impedance to predict muscle and subcutaneous
analyzers provide estimates of segmental fat mass. adipose tissue [26]. This technique should find wider
The methods by which this is achieved are largely application in nutritional investigations that
unknown as the algorithms used are proprietary to involve changes in body water volumes or distri-
the manufacturers [18]; however, it is likely that bution between cellular compartments or body
they are based on a notional ‘weight of the segment’ regions [27].
derived from the predicted segment FFM and per- Highly localized impedance measurements
centage of fat mass derived from regression formulae focussing on small segment volume are also being
determined in independent DXA studies of body used to provide qualitative information on the com-
composition [19]. position and architecture of muscle in a variety of
&&
Leahy et al. [20 ] in a large study of young adults disease states [28]. This approach, known as electri-
(167 women and 236 men aged 18–29 years) com- cal impedance myography, promises to be useful as
pared fat mass measurements predicted by segmen- an indicator of disease state and efficacy of treat-
tal BIA with DXA measurements. Total body fat was ment. To date, its application has been primarily to
underestimated by approximately 2%, although this evaluate neuromuscular disease, but because it pro-
increased to more than 4% in men and women with vides information on the composition of muscle
greater than 24.6 and 32% body fat, respectively. Fat including the presence of edema, fat and connective
mass was significantly overestimated in the trunk in tissue, it is likely to be of value in nutritional studies.
men (2.1 kg, 31%) but underestimated in women The prediction of body composition from impe-
(0.4 kg, 5.6%). There were also significant differ- dance data includes many potential sources of error.
ences for appendicular fat, overestimating leg fat Recognition of this has led some to question the
mass by as much as 1.6 kg (31.4%) in men. These value of predicting body composition parameters
authors urged caution in the use of BIA for segmen- per se and, alternatively, to propose that clinically
tal body composition analysis, particularly for the useful information can be equally gained from
trunk. consideration of the measured impedance alone
&& &
In an earlier review of segmental impedance [29 ,30 ]. This approach has been successfully used
analysis in this journal, De Lorenzo and Andreoli to monitor the development of postsurgical edema
[21] speculated that segmental BIA techniques may localized to the leg [31]. Accumulation of extra-
be used to quantify regional skeletal muscle mass, cellular fluid in lymphoedema of the arm or leg is
replacing imaging techniques such as MRI or com- routinely monitored using impedance indices only
&
puted tomography. Unfortunately, this has not [32,33 ], although prediction of absolute fluid
transpired to be the case. Various authors have volumes is possible [34]. The application of this
developed methods for prediction of skeletal muscle method to assess edema from other causes including
volume (e.g. [22]). Biases in prediction ranged from nutritional perturbation is yet to be realized.
2 to 6% compared to standard error of approxi- Segmental BIA has only recently begun to move
mately 9% when predicting whole-body skeletal into routine clinical practice. Determination of fluid
muscle volume from wrist–ankle impedance distributions in body segments has been shown to
measurements. Despite the promise of these studies, be useful in palliative care of patients with cancer
little work appears to have been undertaken recently [35] as prognostic indicators with the potential to
[23]. The ability of segmental BIA to predict skeletal predict survival time.
muscle mass specifically would extend significantly Patients undergoing peritoneal dialysis, at risk of
the utility of the technique and be of great value and developing protein–energy wasting, routinely have
relevance to those wishing to monitor sarcopenia in nutritional assessments. Segmental BIA measure-
the aged. ments are strongly correlated with DXA measure-
Segmental impedance measurements are not ments in such patients [36,37] and may be a more
confined to whole limbs. Localized impedance practical alternative.
1363-1950 ß 2012 Wolters Kluwer Health | Lippincott Williams & Wilkins www.co-clinicalnutrition.com 427
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Assessment of nutritional status and analytical methods
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Segmental bioelectrical impedance analysis: an update Ward
25. Zhu F, Kuhlmann MK, Kotanko P, et al. A method for the estimation of 31. Codognotto M, Piazza M, Frigatti P, Piccoli A. Influence of localized edema on
hydration state during hemodialysis using a calf bioimpedance technique. whole-body and segmental bioelectrical impedance. Nutrition 2008;
PhysiolMeasur 2008; 29:S503–S516. 24:569–574.
26. Carter M, Zhu F, Kotanko P, et al. Assessment of body composition in dialysis 32. Ward LC, Dylke E, Czerniec S, et al. Confirmation of the reference
patients by arm bioimpedance compared to MRI and 40K measurements. impedance ratios used for assessment of breast cancer-related lymphede-
Blood Purif 2009; 27:330–337. ma by bioelectrical impedance spectroscopy. Lymphat Res Biol 2011;
27. Zhu F, Leonard EF, Levin NW. Extracellular fluid redistribution during hemo- 9:47–51.
dialysis: bioimpedance measurement and model. Physiol Measur 2008; 33. Ward L, Winall A, Isenring E, et al. Assessment of bilateral limb lymphedema
29:S491–S501. & by bioelectrical impedance spectroscopy. Int J Gynecol Cancer 2011;
28. Rutkove SB. Electrical impedance myography: background, current state, and 21:409–418.
future directions. Muscle Nerve 2009; 40:936–946. This article describes the use of impedance indices determined in body segments
29. Piccoli A. Bioelectric impedance measurement for fluid status assessment. for the assessment of fluid accumulation in lymphedema. This method could be
&& Contrib Nephrol 2010; 164:143–152. adopted for monitoring edema in other clinical conditions.
This article reviews the use measured impedance data in the assessment of fluid 34. Ward LC, Czerniec S, Kilbreath SL. Quantitative bioimpedance spectroscopy
status rather than their conversion via prediction equations to absolute measures for the assessment of lymphoedema. Breast Cancer Res Treat 2009;
of body composition. This approach avoids invoking the assumptions involved in 117:541–547.
predictive methods. 35. Crawford GB, Robinson JA, Hunt RW, et al. Estimating survival in patients
30. Lukaski HC, Piccoli A. Bioelectrical impedance vector analysis for assess- with cancer receiving palliative care: is analysis of body composition using
& ment of hydration in physiological states and clinical conditions (Chapter 16). bioimpedance helpful? J Palliat Med 2009; 12:1009–1014.
In: Preedy V, editor. Handbook of Anthropometry. Springer: London, UK; 36. Fürstenberg A, Davenport A. Assessment of body composition in peritoneal
2012. pp. 287–306. dialysis patients using bioelectrical impedance and dual-energy x-ray absorp-
This article reviews the use measured impedance data in the assessment of fluid tiometry. Am J Nephrol 2011; 33:150–156.
status rather than their conversion via prediction equations to absolute measures 37. Fürstenberg A, Davenport A. Comparison of multifrequency bioelectrical
of body composition. This approach avoids invoking the assumptions involved in impedance analysis and dual-energy X-ray absorptiometry assessments in
predictive methods. outpatient hemodialysis patients. Am J Kidney Dis 2011; 57:123–129.
1363-1950 ß 2012 Wolters Kluwer Health | Lippincott Williams & Wilkins www.co-clinicalnutrition.com 429
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.