Open Journal of Nephrology, 2013, 3, 194-204
Published Online December 2013 (
Open Access OJNeph
Ratio of Capacitance/BMI Reflects Deficit in Nutritional
Concentration While CH2 Reflects Total Nutritional
Deficit in CAPD Patients and General Population *#
Keng-Hee Koh1,2†, Hin-Seng Wong1,3
1Department of Nephrology, Kuala Lumpur Hospital, Kuala Lumpur, Malaysia
2Department of Medicine, Miri Hospital, Miri, Malaysia
3Department of Nephrology, Selayang Hospital, Selayang, Malaysia
Received September 20, 2013; revised October 15, 2013; accepted November 12, 2013
Copyright © 2013 Keng-Hee Koh, Hin-Seng Wong. This is an open access article distributed under the Creative Commons Attribu-
tion License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly
Traditionally phase angle was the best predictor in BIA for nutrition and survival in dialysis population. We recently
showed that normalized bioimpedance indices are a better risk discriminator for dialysis patients and the general popu-
lation. We hereby aimed to explore discriminating factors behind them. Methods: We assessed the body capacitive
index (BCI = Capacitance × Height2/Weight); body resistive index (BRI = Resistance × Weight/Height2); and also, CH2
(= Capacitance × Height2) which represents total body capacitive volume in physics. We initially performed BIA for
206 female, 116 male healthy volunteers, followed by, prospective study for 128 CAPD patients (47 diabetes mellitus
(DM), 81 non-DM; 59 male, 69 female) for >2 years. Results: Moderately good negative correlation of albumin and
BCI (r = 0.533, p < 0.001) with linear regression (BCI = 8.780 0.184 × [albumin], R2 = 0.339, p < 0.001) was shown
in CAPD patients. BCI and CH2 were much higher in CAPD patients in comparison to healthy volunteers (3.4 ± 0.1 vs
2.0 ± 0.0 nFm2/kg, p < 0.001 and 203 ± 8 vs 125 ± 1 nFm2, p < 0.001, respectively). In age and gender adjusted logistic
regression model of 128 CAPD patients versus 322 healthy volunteers, the best risk discriminator was BCI (X2 = 165.6),
followed by CH2 (X2 = 140), phase angle (X2 = 59.3) and BRI (X2 = 52.2). Thirty five (27.3%) patients died during the
study period (Fatal cause: infection (54%), cardiovascular (26%)). In Cox regression, CH2 (2
= 32.4) was the best
predictor for all-cause mortality, followed by BCI (2
= 27.7) and phase angle (2
= 19.3). Conclusion: The phase
angle was a compound parameter of the body capacity index (BCI) and body resistive index (BRI). BCI has a moder-
ately good negative correlation with albumin and this supports its role in reflecting the severity of malnutrition in
CAPD patients. CH2 represents total nutrition deficit and thus the major risk indicator for the survival of CAPD pa-
Keywords: Bioimpedance Analysis; Continuous Ambulatory Peritoneal Dialysis;
Comparison with General Population; Nutrition; Survival
1. Introduction
Bioimpedance analysis (BIA) has been recognized as a
powerful tool to assess nutrition and hydration status in
patients on CAPD [1-4] with standardized methodology
[5,6] and various resistive indexes have been described
[7-9]. We have recently shown that normalized capaci
tive indices (e.g., Capacitance/Body Mass Index and Ca-
pacity × Height2) carried more predictive value than
non-normalized parameter like the phase angle [10]. Here
we aim to further explore the physics rationale of such
parameters by analyzing our cohort of CAPD patients
and the general population.
Gender differences in non normalized bioimpedance
parameters (e.g., phase angle and resistive index) have
been reported [11,12]. However we have demonstrated
that these gender differences disappeared when normal-
*Conflict of interest statement: None declared.
#This paper has not been published previously in whole or part, except
in abstract format; Grant support: none.
Corresponding author.
K.-H. KOH, H.-S. WONG 195
ized bioimpendance parameters (e.g., body capacitive
index) were used [10]. The higher body resistive index
(BRI) in female normal subjects is probably due to the
higher percentage of fat which resulted in the lower pha-
se angle [10].
This current study serves to further verify the roles of
varying normalized bioimpedance parameters with vari-
ous analytical models in line with current thought of
translating physiological science into clinical practice
[13], as well as exploring the underlying risk discrimi-
nating factor.
2. Methods
2.1. Study Design and Population
We initially surveyed the bioimpedance profile of the
general population by recruiting 322 healthy volunteers
as representation of normal population in the neighbour-
hood of our hospital. We then prospectively studied all
stable CAPD patients (a total of 128 patients) in Kuala
Lumpur Hospital. These patients were performing 4 ex-
changes a day with 3 exchanges during daytime and a
long dwell at night. The study subjects were followed up
for 2.2 to 2.3 years.
2.2. Bioelectric Impedance Analysis (BIA)
BIA was performed at time of enrolment, with Bioelec-
trical Impedance Analyzer, Maltron Bioscan 916 v3, with
single frequency 50 kHz with alternating sinusoidal cur-
rent, 0.7 mA on all CAPD patients with emptied perito-
neal cavity in tetrapolar placement on the hand and foot.
After the patient drained off the dialysate and was in a
supine position for at least 10 minutes, the standard tetra-
polar electrodes were placed on the dorsum of the wrist
and anterior aspect of the ankle on the left side of the bo-
dy [1]. Bioimpedance measurements obtained at the time
of enrollment were used in the predictive model for this
The clinical value of BIA indexes has been postulated
in previous study [7,8]. We derived body resistive index
(BRI), body capacitive index (BCI), CH2 or H2/XC with
the formulae from A pp en dix A.
Fat free mass, fat mass and fat percentage were de-
rived with equation of Fat Mass = Weight Fat Free
Mass and fat percentage = Fat Mass / Weight × 100%, in
which calculation of fat free mass were as described in
past literature by Kotler et al. [14].
2.3. Statistical Methods
The statistical data were analyzed using Microsoft excel
and SPSS (Statistical package for Social Science. SPSS
Inc. 233 South Wacker Drive, 11th Floor, Chicago, Illi-
nois 60606-6307). Kolmogorov Smirnov test was ini-
tially used to determine whether the data is in statistical
normal distribution. Logarithm transformation would be
performed to achieve statistical normal distribution in
data with non-normal distrubution. Parametric test would
be performed in data with normal distribution. In pa-
rameter that could not achieve normal distribution even
with logarithm transformation with low number of ana-
lyzed subjects, non-parametric test would be utilized.
Logistic regression was used to identify the best risk
discriminator among BIA parameters between Malaysian
normal population and CAPD cohort in Table 1.
All means were presented with ±standard error of
mean (± SEM). Demographic features, blood investiga-
tions, peritoneum equilibrium test, anthropometry and
BIA parameters of CAPD patients were studied to iden-
tify the risk factors for mortality in Table 2. Univariate
analysis was performed with parametric test (e.g., student
t-test, ANOVA) for survival comparison in data with
statistical normal distribution, whereas non-parametric
test (e.g., Mann Whitney U test) would be used in others.
Factors that significantly affect the predictor and survival
were analyzed with ANCOVA test.
In study of correlation relationship between data,
Spearman correlation was used as non-parametric test.
Cox regression was performed to analyse the significance
of the relevant clinical parameters in survival (Table 3).
This was a prospective observational study of clinical
practice, applying BIA and other investigations in CAPD
patients. The protocol was in consistence with the princi-
ples of the Declaration of Helsinki as amended in Tokyo
(1975), Venice (1983), and Hong Kong (1989) [15].
3. Results
3.1. Healthy Volunteers
We enrolled 322 healthy volunteers, i.e., 206 female and
116 male healthy volunteers. They consist of 178 Malays,
89 Chineses, 54 Indians and 2 of other ethnicity. Their age
ranged from 16 to 68 years old with the median of 38 years.
3.2. BIA Parameters of CAPD Patients and
Correlation with Serum Albumin
We subsequently enrolled 128 CAPD patients with eth-
nicity composed of 64 Malays, 50 Chinese and 14 Indi-
ans. Aetiology of renal failure included diabetes mellitus
(43 patients); chronic glomerulonephritis (27 patients);
hypertension (10 patients); obstructive uropathy (6 pa-
tients); adult polycystic kidney disease (1 patient); con-
current diabetes nephropathy and obstructive uropathy (4
patients); while 37 patients have unknown cause of renal
For the CAPD patients, albumin level has the best
correlation with BCI (r = 0.533, p < 0.001), followed by
phase angle (r = 0.520, p < 0.001), capacitance (r =
0.457, p < 0.001), XC/H (r = 0.426, p < 0.001) and CH2
(0.386, r < 0.001).
Open Access OJNeph
Open Access OJNeph
Table 1. Logistic Regression Models with various BIA parameters adjusted for age and gender for CAPD patients versus
Normal Population.
95% CI
Parameters per Unit increment Odds ratio Lower boundUpper bound p-value X2
BIA parameter alone
Phase angle per 1˚ increase 0.428 0.335 0.547 <0.001 59.3
Body Capacitive Index per 1 nFm2/kg increase 19.028 9.168 39.490 <0.001 165.6
CH2 per 1 nFm2 increase 1.047 1.035 1.059 <0.001 140.0
Body Resistive Index per 1 kg/cm2 increase 0.010 0.003 0.041 <0.001 52.2
Combined Model 1
Phase angle per 1˚ increase 1.353 1.159 1.581 <0.001 8.2
Body Capacitive Index per 1 nFm2/kg increase 34.854 14.320 84.833 <0.001
Combined Model 2
Phase angle per 1˚ increase 1.258 1.070 1.480 <0.001 4.0
CH2 per 1 nFm2 increase 1.054 1.040 1.068 <0.001 84.8
Combined Model 3
Phase angle per 1˚ increase 0.505 0.400 0.637 <0.001 37.7
Body Resistive Index per 1 kg/cm2 increase 0.021 0.005 0.089 <0.001 30.6
Combined Model 4
Phase angle per 1˚ increase 1.482 1.240 1.772 <0.001 13.9
Body Capacitive Index per 1 nFm2/kg increase 15.585 6.046 40.171 <0.001 46.2
CH2 per 1 nFm2 increase 1.030 1.015 1.045 <0.001 16.4
Combined Model 5
Phase angle per 1˚ increase 5 1 17 0.019 29.7
Body Capacitive Index per 1 nFm2/kg increase 4280 127 144139 <0.001 115.8
Body Resistive Index per 1 kg/cm2 increase 40770 228 7285483 <0.001 31.8
Abbreviation: CI, confidence interval.
Linear regression yielded: BCI = 8.780 0.184 × [al-
bumin], R2 = 0.339, p < 0.001.
3.3. Comparison of BIA Profile between Healthy
Volunteers and CAPD Patients
Body capacity index and CH2 were much higher in
CAPD patients in comparison to healthy volunteers [3.4
± 0.1 vs 2.0 ± 0.0 nFm2/kg, (p < 0.001) and 203 ± 8 vs
125 ± 1 nFm2, (p < 0.001) respectively]. Figures 1 and 2
demonstrated the difference in BCI and CH2 between
healthy volunteers and CAPD patients.
In logistic regression with 128 CAPD patients versus
322 healthy volunteers (Table 1), age and gender ad-
justed BIA parameter alone model showed that BCI has
the highest risk discrimination, followed by CH2, phase
angle and finally BRI. Phase angle have odds ratio per 1˚
increase of less than 1 because higher phase angle predict
better nutritional status. However when combined with
body capacitive index (BCI) and CH2 in combined model
1 and 2, the odds ratio of phase angle became more than
1. This is because the function of body resistive index is
unmasked for phase angle, when a combined model built
with body capacitive index. In contrast, combined model
5 showed that X2 value of phase angle was minimized
and became insignificant when combined with body ca-
pacitive index and body resistive index because its func-
tions in assessing capacitance and resistance of the body
were replaced by BCI and BRI.
3.4. Survival Analysis
35 (27.3%) patients died during the study period. Infec-
tive (54%) and cardiovascular diseases (26%) were the
main cause of death. The detail causes of death included:
cardiac events, 6 patients; cerebral events, 3 patients;
K.-H. KOH, H.-S. WONG 197
Table 2. Univariate analysis Comparison of clinical parameters during enrolment for survival.
Mean p-value p-value
Survived Died Non-adjusted Adjusted#
Age year 48.7 ± 1.5 55.4 ± 2.3 0.018 0.141
Duration of dialysis year 1.9 ± 1.1 2.2 ± 1.1 0.467
BMI kg/m2 24.3 ± 0.5 24.7 ± 0.7 0.716
Albumin g/L 30.1 ± 0.5 27.7 ± 0.9 0.016 0.028##
TSF cm 1.86 ± 0.10 1.82 ± 0.19 0.852
MAC cm 28.2 ± 0.5 29.1 ± 1.0 0.442
MAMC cm 22.4 ± 0.4 23.4 ± 0.6 0.190
AMA cm2 40.7 ± 1.3 44.1 ± 2.2 0.201
Body Resistive Index kg/cm2 1.325 ± 0.032 1.294 ± 0.034 0.506
Body Capacitive Index* nFm2/kg 2.855 3.985 <0.001 0.004
CH2* nFm2 168.70 244.73 <0.001 0.001
Phase angle** ˚ 4.86 3.58 <0.001
Fat Percentage** % 30.2 25.9 0.506
Mean are expressed with ±standard error of mean. Student t-test was used for comparison except in data with **. *Geometric means were shown and compared.
**Medians were shown and Mann Whitney U test were performed. Anthropometry was measured using triceps skin fold thickness (TSF), mid arm circumfer-
ence (MAC), mid arm muscle circumference (MAMC), and arm muscle area (AMA) and calculated using formulae of MAMC = MAC л.TSF; and AMA =
(MAC л.TSF)2 / 4л; #Univariate adjusted analysis with age, DM status and albumin via ANCOVA were performed for factor with normal distribution, if the
unadjusted analysis by t-test demonstrated significant differences. ##Albumin analysis was adjusted with DM status and age.
Table 3. Cox Regression Survival Hazard Ratio Model with various BIA parameters for CAPD patients.
95% CI 95% CI
Parameters per Unit increase Hazard ratio Lower boundUpper boundp-valueHazard ratioLower bound Upper boundp-value
Model with phase angle Model with Capacitive Index
Age per 1 year increase 1.022 0.996 1.050 0.0971.030 1.002 1.059 0.036
Diabetes Status DM:nonDM 1.402 0.680 2.891 0.3591.597 0.769 3.315 0.209
Albumin per 1 g/L increase 0.968 0.909 1.032 0.3210.954 0.894 1.019 0.160
Phase angle per 1˚ increase 0.454 0.310 0.664 <0.001
Capacitive Index per 1 nFm2/kg increase 1.389 1.139 1.695 0.001
Overall model* 2
X = 19.3 0.001
4 = 27.7 <0.001
Model with XC/H Model with CH2
Age per 1 year increase 1.024 0.997 1.053 0.0871.028 1.000 1.058 0.053
Diabetes Status DM:nonDM 1.187 0.561 2.511 0.6541.275 0.592 2.743 0.535
Albumin per 1 g/L increase 0.961 0.901 1.026 0.2360.942 0.891 0.995 0.034
XC/H per 1 /cm increase 0.000 0.000 0.012 <0.001
CH2 per 1 nFm2 increase 1.006 1.003 1.009 <0.001
Overall model* 2
X = 17.5 0.002
4 = 32.4 <0.001
Abbreviation: CI, confidence interval; *Overall model tested were with Omnibus test of Model Coefficients. Note: CH2 per 1 nFm2 increase is replaceable by
CH2 per 100 nFm2 increase with hazard ratio of 1.753 (CI: 1.302 - 2.368, p < 0.001), or H2/XC in the unit of cm2/ with hazard ratio of 1.002 (CI: 1.001 - 1.003,
p < 0.001).
peritonitis, 7 patients; other infection, 12 patients; ma-
lignancy, 1 patient; other causes of death, 2 patients; un-
known cause of death, 4 patients. Twenty out of 47 pa-
tients with diabetes died (43%) in comparison to 15 out
of 81 non diabetic patients (19%) (p = 0.003). There was
no gender predisposition for survival in this cohort with
52 out of 69 female patients (75%) survived and 41 out
of 59 male patients (69%) survived (p = 0.458).
Adjusted univariate survival analyses were performed
with age, diabetes mellitus status and albumin as in Ta-
ble 2, for those parameters which had significant differ-
ence between the survival and fatal patients. We have
Open Access OJNeph
95% Confidential Interval for BCI (nF.m2/kg)
malefemale malefemale
CAPD patientsnormal population
Figure 1. BCI of CAPD patients versus normal population.
95% Confidential Interval for C.H2 (nFm2)
malefemale malefemale
CAPD patientsnormal population
Figure 2. CH2 of CAPD patients versus normal population.
previous reported the absent of correlation between sur-
vival and blood pressure, lipid profile, calcium, phos-
phate, ALP, iPTH, Hb, ESR, Kt/V, creatinine clearance,
urea clearance and ultrafiltration in this cohort [8] and
hence have limited the univariate analysis to salient fac-
tors, nutritional anthropometry assessment and bioim-
pedance. Surviving patients had significantly lower BCI
(Figure 3), CH2 (Figure 4) and phase angle.
We built 4 essential predictive models for survival
prediction with Cox regression analysis. Table 3 showed
that the overall survival of this cohort was best fitted into
models with CH2 (or H2/XC), with highest X2 value fol-
lowed by BCI, phase angle and XC/H.
4. Discussion
In order to improve the evaluation of nutritional and hy-
dration status of the chronic dialysis patients, BIA has
been advocated for both chronic ambulatory peritoneal
dialysis (CAPD) patients [1,10,16,17] and hemodialysis
patients [18-20]. Phase angle has been shown to be pre-
dictive of survival in dialysis population [21]. We have
demonstrated that newer normalized bioimpedance pa-
rameters have better survival predictive value compared
to phase angle [10]. However, the underlying factors
influencing these normalized parameters and its physics
rationale is still unsettling.
We evaluated these two parameters: body capacitive
index (BCI), which is the product of capacitance and
Height2/Weight, i.e., ratio of capacitance over body mass
index. In Appendix A, we demonstrated that it represents
ε/D, i.e., the ratio of permittivity of dielectric in the body
over density of body; and body resistive index (BRI),
which is the product of resistance and Weight/Height2.
Open Access OJNeph
K.-H. KOH, H.-S. WONG 199
95% Confidentia l Interval for BCI ( nFm 2/ kg)
FatalSurvived FatalSurvived
Figure 3. BCI at enrolment in Survived versus Fatal CAPD patients with 2 years follow-up.
95% Confidential Interval for C.H2 (nFm2)
FatalSurvived FatalSurvived
Figure 4. CH2 at enrolment in Survived versus Fatal CAPD patients with 2 years follow-up.
The phase angle is actually the arctangent of 1/ωCIRI.
We also showed that CH2 represents εV, the total body
capacitive volume in physics. BCI reflects nutritional
deficit concentration, i.e., both nutritional and hydration
status while CH2 reflects total nutritional deficit. (Note
that H2/XC = ωCH2 to compare the result with previous
study of general population in Germany [7])
BCI correlated with albumin, better than all other pa-
rameters. Linear regression of BCI and albumin sug-
gested that the presence of albumin and other unmeas-
ured substance reduce the capacitive indexes. Previous
literature has discussed regarding the survival predictive
value of albumin [17] and the potential relation of hypo-
albuminaemia with overhydration [16]. In current study,
although BCI has higher predictive value than albumin
alone in predictive model, we are still not able to differ-
entiate the causal-result relationship of fluid and nutrition
with BCI.
Nonetheless, higher BCI in CAPD patients reflected
their higher nutritional deficit concentration while higher
CH2 reflected their higher total nutritional deficit. Logis-
tic regression of various BIA parameters showed that
body capacitive index is the main risk discriminator be-
tween CAPD patients and general population. This
showed the marked change in the deficit in nutrition con-
centration with disease occurrence. The combined model
of bioimpedance parameters revealed the interesting un-
derlying interplay between BCI, BRI and phase angle.
The hidden BRI property of phase angle unmasked and
its risk discriminating property disappeared, when model
involved phase angle and BCI. Meanwhile, the hidden
BCI property of phase angle unmasked when it is in
Open Access OJNeph
combined model with BRI. CH2 is the second risk dis-
criminator as total nutritional deficit would be increased
when end stage renal failure set in.
Both CH2 and BCI were the key survival predictors for
CAPD patients as demonstrated by Cox regression mod-
els (Table 3). Total nutrition deficit (with the marker of
CH2) and nutritional deficit concentration (BCI) of a
CAPD patient predicts his/her survival. Therefore, we
proposed their assessment in line with noble clinical
opinion of systematic nutritional assessment with bioim-
pedance for end stage renal failure patients [13,22-24].
Our previous study demonstrated that both BCI and
CH2 were better risk indicators than phase angle because
of the gender effect on the latter which actually makes it
less discriminatory [10]. In addition, because BRI was
non-significantly higher in survived than diseased pa-
tients, they reduced the phase angle of the survived pa-
tients mathematically and thus limited the survival pre-
dictability of the phase angle.
In short, we showed the better correlation of albumin
and BCI in comparison to other bioimpedance parame-
ters. And we also demonstrated the better risk discrimi-
natory effect of BCI in comparison to phase angle with
logistic regression which is not presented in our last pa-
per [10].
Besides, we reported the use of single frequency BIA
in determining the nutritional and hydration status of the
patient as well as survival prediction [7,25-27]. We sug-
gest further exploration in regard to resistive and capaci-
tive indices [25,26] and extension of research into hae-
modialysis [28], HIV [29] and other diseases and healthy
control populations. Certainly further study is needed to
reaffirm the clinical role of BCI and CH2. These are in
line with current opinion on BIA research [22,30].
Unlike other sophisticated BIA derived parameters,
BCI, CH2 and BRI are practical factors that easily meas-
ured and derived from height, weight, capacitance and
resistance of the patient, in routine clinical assessment.
Nevertheless, just like other BIA parameters, e.g., phase
angle, the baseline normal reference of BCI and CH2 are
needed for one to confidently conclude the extent of nu-
tritional deficit. Gaining normal reference for various
ethnic and disease population is one of the main hurdles
in bioimpedance clinical use and we have shared this
concern to the fellow researchers [31].
Recently, Zhu F et al. has significant break through
with innovative advance using segment-specific resistiv-
ity to derive accurate water distribution information
[32,33]. At the same time, phase angle of various body
compartment was explored by Nescolarde L [34]. We
hereby hope to propose that future research possibly
should also look into segment-specific capacitive permit-
tivity. One might be able to postulate higher significance
in nutritional assessment and risk discriminating effect
with trunk capacitive index.
5. Conclusions
In summary, BCI represents nutritional deficit concen-
tration with good negative correlation with albumin and
was the main risk discriminator for CAPD patients ver-
sus general population. On the other hand, CH2 repre-
sents total nutrition deficit and thus the major risk indi-
cator for the survival of CAPD patients. And, the tradi-
tionally measured phase angle was a compound parame-
ter for BCI and BRI.
Further research for BCI, BRI and CH2 in other
healthy community and other disease groups are needed
before we could draw a firm conclusion to their specific
role in physiology and clinical management.
6. Acknowledgements
We thank all clinical staffs of Kuala Lumpur Hospital in
managing this study cohort and accomplishing the study.
We also thank Director General of Health, Malaysian
Ministry of Health in approving the publication of this
research paper.
[1] L. T. Cheng, W. Tang and T. Wang, “Strong Association
between Volume Status and Nutritional Status in Perito-
neal Dialysis Patients,” American Journal of Kidney Dis-
eases, Vol. 45, No. 5, 2005, pp. 891-902.
[2] A. Picolli, “Bioelectric Impedance Vector Distribution in
Peritoneal Dialysis Patients with Different Hydration
Status,” Kidney International, Vol. 65, No. 3, 2004, pp.
[3] F. Zhu, G. Wystrychowski, T. Kitzler, S. Thijssen, P.
Kotanko and N. W. Levin. “Application of Bioimpedance
Techniques to Peritoneal Dialysis,” Contributions to Ne-
phrology, Vol. 150, 2006, pp. 119-128.
[4] G. A. Young, J. D. Kopple and B. Lindholm, “Nutritional
Assessment of Continuous Ambulatory Peritoneal Dialy-
sis Patients: An international Study,” American Journal
of Kidney Diseases, Vol. 17, No. 4, 1991, pp. 462-471.
[5] U. G. Kyle, I. Bosaeus, A. D. De Lorenzo, P. Deurenberg,
M. Elia, J. M. Gómez, B. L. Heitmann, L. Kent-Smith, J.
C. Melchior, M. Pirlich, H. Scharfetter, A. M. Schols and
C. Pichard, “Composition of the ESPEN Working Group.
Bioelectrical Impedance Analysis, Part I: Review of Prin-
ciples and Methods,” Clinical Nutrition, Vol. 23, No. 5,
2004, pp. 1226-1243.
[6] U. G. Kyle, I. Bosaeus, A. D. De Lorenzo, P. Deurenberg,
M. Elia, J. M. Gómez, B. L. Heitmann, L. Kent-Smith,
J.-C. Melchior, M. Pirlich, H. Scharfetter, A. M. W. J.
Schols and C. Pichard, “Bioelectrical Impedance Analy-
Open Access OJNeph
K.-H. KOH, H.-S. WONG 201
sis—Part II: Utilization in Clinical Practice,” Clinical
Nutrition, Vol. 23, No. 6, 2004, pp. 1430-1453.
[7] A. Bosy-Westphal, S. Danielzik, R. P. Dörhöfer, A. Pic-
coli and M. J. Muller, “Patterns of Bioelectrical Imped-
ance Vector Distribution by Body Mass Index and Age:
Implications for Body-Composition Analysis,” The Ame-
rican Journal of Clinical Nutrition, Vol. 82, No. 1, 2005,
pp. 60-68.
[8] R. F. Kushner, D. A. Schoeller, C. R. Fjeld and L. Dan-
ford, “Is the Impedance Index (ht2/R) Significant in Pre-
dicting Total Body Water?” The American Journal of
Clinical Nutrition, Vol. 56, No. 5, 1992, pp. 835-839.
[9] H. C. Lukaski, P. E. Johnson, W. W. Bolonchuk, G. I.
Lykken, “Assessment of Fat-Free Mass Using Bioelectri-
cal Impedance Measurements of the Human Body,” The
American Journal of Clinical Nutrition, Vol. 41, No. 4,
1985, pp. 810-817.
[10] K. H. Koh, H. S. Wong, K. W. Go and Z. Morad, “Nor-
malized Bioimpedance Indices Are Better Predictors of
Outcome in Peritoneal Dialysis Patients,” Peritoneal Di-
alysis International, Vol. 31, No. 5, 2011, pp. 574-582.
[11] M. Dittmar, “Reliability and Variability of Bioimpedance
Measures in Normal Adults: Effects of Age, Gender, and
Body Mass,” American Journal of Physical Anthropology,
Vol. 122, No. 4, 2003, pp. 361-370.
[12] M. Arroyo, A. M. Rocandio, L. Ansotegui, H. Herrera, I.
Salces and E. Rebato, “Comparison of Predicted Body
Fat Percentage from Anthropometric Methods and from
Impedance in University Students,” British Journal of
Nutrition, Vol. 92, No. 5, 2004, pp. 827-832.
[13] P. A. Fein, G. Gundumalla, A. Jorden, B. Matza, J. Chat-
topadhyay and M. M. Avram, “Usefulness of Bioelectri-
cal Impedance Analysis in Monitoring Nutrition Status
and Survival of Peritoneal Dialysis Patients,” Advances in
Peritoneal Dialysis, Vol. 18, 2002, pp. 195-199.
[14] D. P. Kotler, S. Burastero, J. Wang and R. N. Pierson Jr.,
“Prediction of Body Cell Mass, Fat-Free Mass, and Total
Body Water with Bioelectrical Impedance Analysis: Ef-
fects of Race, Sex, and Disease,” American Journal of
Physical Anthropology, Vol. 64, Suppl. 3, 1996, pp. S489-
[15] “Declaration of Helsinki,” Bulletin of the Pan American
Health Organization, Vol. 24, 1990, pp. 606-609.
[16] C. H. Jones, L. Wells, J. Stoves, F. Farquhar and G.
Woodrow, “Can a Reduction in Extracellular Fluid Vol-
ume Result in Increased Serum Albumin in Peritoneal
Dialysis Patients?” American Journal of Kidney Diseases,
Vol. 39, No.4, 2002, pp. 872-875.
[17] D. G. Struijk, R. T. Krediet, G. C. Koomen, E. W.
Boeschoten and L. Arisz, “The Effect of Serum Albumin
at the Start of Continuous Ambulatory Peritoneal Dialysis
Treatment on Patient Survival,” Peritoneal Dialysis In-
ternational, Vol. 14, No. 2, 1994, pp. 121-126.
[18] N. W. Levin, F. Zhu, E. Seibert, C. Ronco and M. K.
Kuhlmann, “Use of Segmental Multifrequency Bioimpe-
dance Spectroscopy in Hemodialysis,” Contributions to
Nephrology, 2005, Vol. 149, 2005, pp. 162-167.
[19] P. L. Cox-Reijven, J. P. Kooman, P. B. Soeters, F. M.
Van der Sande and K. M. Leunissen, “Role of Bioim-
pedance Spectroscopy in Assessment of Body Water
Compartments in Hemodialysis Patients,” American Jour-
nal of Kidney Diseases, Vol. 38, No. 4, 2001, pp. 832-838.
[20] A. E. Jabara and R. L. Mehta, “Determination of Fluid
Shifts during Chronic Hemodialysis Using Bioimpedance
Spectroscopy and an In-Line Hematocrit Monitor,”
ASAIO Journal, Vol. 41, No. 3, 1997, pp. M682-M687.
[21] R. Mushnick, P. A. Fein, N. Mittman, N. Goel, J.
Chattopadhyay and M. M. Avram, “Relationship of Bio-
electrical Impedance Parameters to Nutrition and Survival
in Peritoneal Dialysis Patients,” Kidney International,
Vol. 64, Suppl. 87, 2003, pp. S53-S56.
[22] M. C. Barbosa-Silva, A. J. Barros, J. Wang, S. B. Heyms-
field and R. N. Pierson Jr., “Bioelectrical Impedance
Analysis: Population Reference Values for Phase Angle
by Age and Sex,” The American Journal of Clinical Nu-
trition, Vol. 82, No. 1, 2005, pp. 49-52.
[23] K. Kalantar-Zadeh, T. A. Ikizler, G. Block, M. M. Avram
and J. D. Kopple, “Malnutrition-Inflammation Complex
Syndrome in Dialysis Patients: Causes and Conse-
quences,” American Journal of Kidney Diseases, Vol. 42,
No. 5, 2003, pp. 864-881.
[24] G. M. Chertow, J. M. Lazarus, N. L. Lew, L. Ma and E.
G. Lowrie, “Bioimpedance Norms for the Hemodialysis
Population,” Kidney International, Vol. 52, No. 6, 1997,
pp. 1617-1621.
[25] M. K. Kuhlmann, F. Zhu, E. Seibert and N. W. Levin,
“Bioimpedance, Dry Weight and Blood Pressure Control:
New Methods and Consequences,” Current Opinion in
Nephrology & Hypertension, Vol. 14, No. 6, 2005, pp.
[26] A. Piccoli, G. Pastori, M. Guizzo, M. Rebeschini, A.
Naso and C. Cascone, “Equivalence of Information from
Single versus Multiple Frequency Bioimpedance Vector
Analysis in Hemodialysis,” Kidney International, Vol. 67,
No. 1, 2005, pp. 301-313.
[27] R. N. Baumgartner, W. C. Chumlea and A. F. Roche,
“Bioelectric Impedance Phase Angle and Body Com-
position,” The American Journal of Clinical Nutrition,
Vol. 48, No. 1, 1998, pp. 16-23.
[28] G. M. Chertow, D. O. Jacobs, J. M. Lazarus, N. L. Lew
and E. G. Lowrie, “Phase Angle Predicts Survival in
Hemodialysis Patients,” Journal of Renal Nutrition, Vol.
7, No. 4, 1997, pp. 204-207.
[29] M. Ott, H. Fischer, H. Polat, E. B. Helm, M. Frenz, W. F.
Caspary and B. Lembcke, “Bioelectrical Impedance
Open Access OJNeph
Open Access OJNeph
Analysis as Predictor of Survival in Patients with Human
Immunodeficiency Virus Infection,” Journal of Acquired
Immune Deficiency Syndromes and Human Retrovirology,
Vol. 9, No. 1, 1995, pp. 20-25.
[30] P. Kotanko, N. W. Levin and F. Zhu, “Current State of
Bioimpedance Technologies in Dialysis,” Nephrology
Dialysis Transplantation, Vol. 23, No.3, 2008, pp. 808-
[31] K. H. Koh, H. S. Wong, K. W. Go and Z. Morad, “The
Implication of Bioimpedance Analysis in Peritoneal Di-
alysis patients,” Asian Chapter Newsletter, International
Society of Peritoneal Dialysis, Vol. 9, No. 1, 2011, pp. 2-
[32] F. Zhu, M. K. Kuhlmann, G. A. Kaysen, S. Sarkar, C.
Kaitwatcharachai, R. Khilnani, L. Stevens, E. F. Leonard,
J. Wang, S. Heymsfield and N. W. Levin, “Segment-
Specific Resistivity Improves body Fluid Volume Esti-
mates from Bioimpedance Spectroscopy in Hemodialysis
Patients,” Journal of Applied Physiology, Vol. 100, No. 2,
2006, pp. 717-724.
[33] F. Zhu, P. Kotanko, G. J. Handelman, J. G. Raimann, L.
Liu, M. Carter, M. K. Kuhlmann, E. Seibert, E. F. Leo-
nard and N. W. Levin, “Estimation of Normal Hydration
in Dialysis Patients Using Whole Body and Calf Bioim-
pedance Analysis,” Physiological Measurement, Vol. 32,
No. 7, 2011, pp. 887-902.
[34] L. Nescolarde, J. Rosell-Ferrer and T. Doñate, “Relation-
ship between Segmental and Whole-Body Phase Angle in
Peritoneal Dialysis Patients,” Physiological Measurement,
Vol. 29, No. 9, 2008, pp. N49-N57.
K.-H. KOH, H.-S. WONG 203
Appendix A
For a given electrical conductor undergoing direct cur-
rent, resistance is proportionately correlated with length
and inversely proportionately correlated with area. It is
generally hard to convert the cylinder model to human
body. However, we could deduce their relationship with
length (height) and weight with below approximation (7)
with modification :
Weig ht
Weig ht
kH D
Weight or
whereby R represents resistance; ρ, apparent resistivity of
the conductor; A, Area; L, Length, converting to H,
height in subsequent equations; D, density of the body;
W, weight; BMI, body mass index; k, converting factor
of body; And, Body resistive index (BRI) reflects the
product of resistivity of the body and density of body.
In same simulation model, during the alternative cur-
rent, we assume that the cell membrane and other protein
substance can act as a capacitor due to its dielectric
property, where the A is area of cell membrane, L is the
thickness of cell membrane plus body compartments, and
ε is the permittivity of cell membrane. The reactance was
calculated from wrist to ankle bioimpedance measure-
ment which reflects all parallel connected cells in the
Therefore, the reactance of the capacitor is reflected
by the below equations:
whereby C, capacitance; f, frequency of applied alterna-
tive current; XC, reactance.
For a given capacitor consisting of 2 metal plates, the
capacitance could be measured by:
whereby C represents capacitance; ε, permittivity of the
dielectric between the two plates; A, Area; L, Distance
between these plates.
Therefore, we could derive body capacitive index
(BCI), as well as C·H2 for human in BIA with the below
22 22
Weigh tAV
 
Let 2
 (4)
Besides, from Equation (3),
CH k
whereby V represents volume of body; D, density of
body; H, height; k, converting factor of body.
Thus, BCI represents the ratio of capacitive permittiv-
ity over body density, while CH2 represents the total
body capacitive volume in physics.
It is worth pointing out that the renowned terms of
height2/reactance (5) could be derived with:
 (6)
Thus, the clinical implication of height2/reactance
(H2/XC) is equivalent to CH2.
However, Height2/resistance is actually mathemati-
cally representing V/ρ:
whereby R represents resistance; H, height; V, body
volume; ρ, resistivity of body.
Phase angle (α) is defined mathematically as arc tan-
gen of reactance over resistance, being measured in de-
Reactance 1
ArcTan ArcTan
From Equation (1),
From Equation (4),
Weight BCI
Putting the above Equations (9) and (10) into Equation
(8), we get:
Open Access OJNeph
Take note that resistance for measurement of BRI in
our current study is taken from direct current. Because of
mathematically low phase angle range for all subjects
and thus cosine is closed to 1, the body resistance in al-
ternative current was thus not differed much from resis-
tance in direct current.
Putting 2
and into Equa-
tion (11)
ArcTan ArcTan
Inserting these into Equation (8) with info from Equa-
tion (9) and (10), we get
 (13)
Therefore, CR, the product of capacitance and resis-
tance is representing ερ, i.e., the product of resistivity of
body and permittivity of body dielectric.
Open Access OJNeph