LYMPHATIC RESEARCH AND BIOLOGY

Volume 00, Number 00, 2016 Original Article

ª Mary Ann Liebert, Inc.
DOI: 10.1089/lrb.2016.0004

Distribution of Extracellular Fluid in Legs

with Venous Edema and Lymphedema

Kotaro Suehiro, MD, Noriyasu Morikage, MD, Osamu Yamashita, MD, Takasuke Harada, MD,
Koshiro Ueda, MD, Makoto Samura, MD, Yuya Tanaka, MD, Yuriko Takeuchi, MD,
Kaori Nakamura, and Kimikazu Hamano, MD

Abstract

Background: This study aimed to determine the specific fluid distribution pattern in legs with lymphedema (LE)
in comparison to normal legs and legs with venous edema (VE) using bioelectrical impedance analysis (BIA).
Methods and Results: BIA was performed in 47 patients with lymphedema (LE; 63 legs), 33 patients with
venous edema (VE; 60 legs), and 33 normal subjects (N; 66 legs). The ratio of intracellular fluid (ICF)
resistance (Ri) to extracellular fluid (ECF) resistance (Re) of a whole leg normalized to the right arm (Ri/Re leg),
a surrogate parameter for ECF/ICF, and the Ri/Re of the thigh and calf without normalization (Ri/Re thigh, Ri/Re calf,
respectively) were obtained. Increases in Ri/Re leg (N 2.5 – 0.7, VE 3.9 – 3.7, LE 3.7 – 1.5), Ri/Re thigh
(N 1.8 – 0.5, VE 3.2 – 3.8, LE 3.8 – 1.9), and Ri/Re calf (N 2.6 – 0.6, VE 4.6 – 1.7, LE 4.4 – 2.2) were confirmed
in VE and LE compared to normal subjects. However, the ratios of Ri/Re calf to Ri/Re thigh in normal subjects and
those with untreated VE/LE were all *1.5 (N 1.5 – 0.3, VE 1.5 – 0.7, LE 1.6 – 0.7).
Conclusions: Fluid content was increased in legs with VE and LE compared to normal subjects, while the mode
of gravitational fluid distribution was similar among all legs. Thus, no specific finding for LE was confirmed.

Introduction Generally, the symptoms in legs with venous edema (VE)

B ioelectrical impedance analysis (BIA) has been
used for the assessment of arm lymphedema (LE) since
the early 1990s1,2 and currently is also used for leg LE.3–5
The magnitude of the electrical impedance is inversely pro-
portional to the volume of fluid in the tissue.6 Accordingly,
the ratio of extracellular fluid (ECF) to intracellular fluid
(ICF) volumes can be expressed as the inverse ratio of each
resistance, that is, Ri (the impedance of ICF) and Re (the
impedance of ECF). The increase in this Ri/Re has been
considered as a sensitive diagnostic criterion for early LE,7
because the changes predominantly or solely occur in ECF
volume in the early stage of LE development. However, as
LE progresses further, the fluid increases in protein content
with cellular infiltration, eventually developing tissue fibrosis
and fat deposition. As a result, the overall limb volume may
continue to increase, but fluid content decreases proportio-
nately.6 Furthermore, LE is often complicated by other types
of edema caused by venous insufficiency, cardiac/renal fail-
ure, hypoalbuminemia, side effects of medications, and so on.
Therefore, increased fluid content alone is not a sufficient
explanation for LE.
are evident in the lower part of the leg, but swelling of the
thigh can be evident, particularly in secondary LE. In con-
trast, we previously reported that subcutaneous echo-free
space observed with ultrasonography, which indicated local
fluid accumulation in subcutaneous tissue, was similarly
distributed according to gravity both in legs with VE and LE.8
Thus, it is unclear whether there is a specific fluid distribution
pattern for LE when compared to legs with other types of
edema or normal legs. The current study attempted to answer
this question.
Patients and Methods
This study was approved by the Institutional Review Board
of Yamaguchi University Hospital (Ube, Yamaguchi, Japan).
All participants gave informed consent before enrollment.
The subjects included in the current study were 33 patients
with VE and 47 patients with leg LE who attended our clinic
between April 2015 and September 2015. Thirty-three heal-
thy volunteers were also assessed. The diagnoses were made
by patient histories, physical examinations, blood tests, chest
radiograms, and electrocardiograms, to exclude edemagenic

Division of Vascular Surgery, Department of Surgery and Clinical Science, Yamaguchi University Graduate School of Medicine,
Yamaguchi, Japan.

1
2 SUEHIRO ET AL.

Table 1. Subject Characteristics

Normal (N = 33) Venous edema (N = 33) Lymphedema (N = 47) p
a a
Age, years; median (range) 35 (23–79) 74 (46–86) 70 (31–88) <0.0001
Height, cm; median (range) 160 (143–180) 151 (143–169) 155 (140–176) <0.0001
Weight, kg; median (range) 58 (45–105) 57 (43–98) 55 (40–88)a,b <0.01
BMI, kg/m2; median (range) 23 (19–34) 25 (19–41) 22 (15–31)b <0.001
Sex (male:female) 12:21 5:28 9:39 N.S.
Dominant side (right:left) 31:2 30:3 43:5 N.S.
Edema laterality (unilateral:bilateral) 6:27 31:16 <0.0001
Legs (N = 66) (N = 60) (N = 63)
Diagnosis PVI 18 Primary
lymphedema 4
PTS 3 Secondary
lymphedema 44
FVI 39 Stage I 3
Stage II 54
Stage III 6
Edema side (dominant:nondominant) 29:31 31:32 N.S.
Treatment (untreated:treated) 39:21 14:49 <0.0001
a
p < 0.05 versus normal.
b
0.05 versus venous edema.
BMI, body mass index; FVI, functional venous insufficiency; PTS, post-thrombotic syndrome; PVI, primary venous insufficiency.

conditions due to systemic diseases. A duplex venous ultra-

sound was performed to find any reflux in the deep veins,
saphenous veins, accessory saphenous veins, and perforators.
Subcutaneous tissue ultrasonography8,9 was also performed
to find changes in echogenicity that possibly correlated with
inflammatory change and echo-free space that indicated ac-
cumulation of free fluid. Absence of any abnormal findings
in the above examinations was confirmed in normal subjects.
In this study, edema found in the legs with primary venous
insufficiency, post-thrombotic syndrome, and functional ve-
nous insufficiency (FVI) were all regarded as VE. FVI was
defined as below:
 No reflux or occlusion in the deep veins, saphenous
veins, accessory saphenous veins, or perforators on du-
plex venous ultrasound
 Confirmed severe gait disturbance and/or prolonged
sitting due to joint problems, general weakness, obe-
sity, and so on
 No other edemagenic conditions, for example, cardiac/
hepatic/renal failure
In the current study, LE complicated by apparent venous
disorders was excluded. The clinical severity of LE was
judged according to the Consensus Document of the Inter-
national Society of Lymphology10 as follows:
Stage 0: A latent or subclinical condition in which limb
swelling is not yet evident.
Stage I: An early accumulation of fluid that subsides with
limb elevation.
Stage II: Tissue swelling that is not reduced by limb elevation FIG. 1. Electrode positions in Protocol 3. The electrode
alone. Pitting manifests in earlier stage II, but the limb may or position at the ankle was the same with that used in the
may not pit in later stage II as excess fat and fibrosis supervene. standard equipotential electrode arrangement. At the knee
Stage III: Lymphostatic elephantiasis in which pitting can level, the middle of the electrode was placed at the level of
upper edge of patella, but it was attached slightly medial to
be absent and trophic skin changes such as acanthosis, further the patella. At the groin level, the electrode was placed just
deposition of fat and fibrosis, and watery overgrowths have below the crease of the front aspect of the groin. However,
developed. in the subjects with small stature, this was sometimes nee-
An asymptomatic contralateral leg with unilateral LE ded to be shifted proximally to keep interelectrode distance
might be regarded as Stage 0 LE because the leg could have >20 cm.
BIOELECTRICAL IMPEDANCE ANALYSIS IN LEG EDEMA
FIG. 2. Correlation between Ri/Re of a leg normalized to
the contralateral leg (Ri/Re leg (cont)) and that normalized to
the right arm (Ri/Re leg (rarm)) in patients with unilateral leg
edema and normal subjects.
reduced lymph transport due to intrapelvic lymph node dis-
section in secondary LE or possible congenital abnormality
in primary LE. However, they were excluded from the
analysis because these legs had no edema. Participant char-
acteristics are summarized in Table 1.
A bioimpedance spectrometer (U-400; Impedimed Ltd,
Brisbane, Australia) was used for BIA. All assessments were
performed between 11 a.m. and 3 p.m. Assessment within
1 hour after a meal and/or exercise was avoided. Participants
were set supine in a room maintained at 25C, and the skin
was degreased using alcohol swabs before the attachment of
electrodes. Specially designed electrodes that could be used
both for current drive and voltage sensing were attached
FIG. 3. (A) Correlation between the extracellular impedances of a whole leg (Re leg) and the sum of the extracellular
impedances of thigh (Re thigh) and calf (Re calf). (B) Correlation between the intracellular impedances of a whole leg (Ri leg)
and the sum of the intracellular impedances of thigh (Ri thigh) and calf (Ri calf).
3
on the dorsal wrists and the anterior surface of the leg at
three levels (ankle, upper edge of the patella, groin) for three
protocols as below:
Protocol 1: This was performed in patients with unilateral
leg edema and normal subjects only. The impedance in a leg
with edema (or the right leg in normal subjects) normalized to
the contralateral normal leg (or the left leg in normal sub-
jects) was obtained.
Protocol 2: This was performed in all participants. The im-
pedance in each leg normalized to the right arm was obtained.
Protocol 3: This was performed in all participants. The
impedances in the thigh and calf without normalization were
separately obtained.
Protocol 1 and 2 were performed using the standard
equipotential electrode arrangement. Detailed electrode po-
sitions in Protocol 3 are demonstrated in Figure 1.
Statistical analyses
The results are expressed as mean – standard deviation or
count, unless otherwise indicated. Simple linear regression
analysis was used to study the correlation between Ri/Re
obtained by different methods and to test the correlation
between Ri/Re and the sum of Ri/Re for the thigh and calf. To
test for differences in subject characteristics and differences
in Ri/Re, Ri, and Re in legs with various conditions, the
Kruskal–Wallis test was used. The Mann–Whitney U-test
was used for multiple comparisons. Statistical analyses were
performed using JMP 11.0 (SAS Institute, Cary, NC). A
p-value <0.05 was considered significant.
Results
The currently used BIA device was originally designed to
calculate Ri and Re of a leg normalized to the contralateral
leg, which is essentially applicable only to unilateral edema.
To obtain Ri and Re of both legs, we tried to use the right arm
for normalization and to verify whether the method is valid. Ri/
Re of a leg normalized to the contralateral leg (Ri/Re leg (cont))
and that normalized to the right arm (Ri/Re leg (rarm)) were
4 SUEHIRO ET AL.

FIG. 4. (A) Ri/Re leg, Re leg, and Ri leg in normal legs, legs with venous edema, and lymphedema. (B) Ri/Re calf, Re calf, and
Ri calf in normal legs, legs with venous edema, and lymphedema. (C) Ri/Re thigh, Re thigh, and Ri thigh in normal legs, legs with
venous edema, and lymphedema. (D) Ratios of Ri/Re calf to Ri/Re thigh in normal legs, legs with venous edema, and
lymphedema. Ri/Re rarm, Re rarm, Ri rarm, Ri/Re calf, Re calf, Ri calf, Ri/Re thigh, Re thigh, and Ri thigh: as in Figures 1 and 2. N:
normal legs, Vu: legs with untreated venous edema, Vt: legs with treated venous edema, Lu: legs with untreated lym-
phedema, Lt: legs with treated lymphedema. *:p < 0.05 versus normal legs. {: p < 0.05 between treated and untreated legs.
BIOELECTRICAL IMPEDANCE ANALYSIS IN LEG EDEMA
compared in patients with unilateral leg edema (tested
leg = diseased leg, n = 37) and in normal subjects (tested
leg = right leg, n = 33). A good linear correlation between
Ri/Re leg (cont) and Ri/Re leg (rarm) was obtained both in patients
with unilateral leg edema (R2 = 0.97, p < 0.0001) and in nor-
mal subjects (R2 = 0.84, p < 0.0001) (Fig. 2). The regression
lines obtained from these legs were very similar. Accordingly,
it seemed reasonable to compare edematous legs and normal
legs directly using Ri/Re leg (rarm) instead of Ri/Re (cont). Among
the current study groups, no significant differences were
observed in terms of sex and limb dominance, while signif-
icant differences were observed in terms of age, physique,
edema laterality, and treatment status, which might affect the
results according to a report by Ward et al.3 However, we
decided to use Ri/Re leg (rarm) as a guide for fluid volume status
regardless of these factors, because these were directly and/or
indirectly related to the nature of the diseases.
Next, we measured the extracellular and intracellular im-
pedances of thigh (Re thigh, Ri thigh) and calf (Re calf, Ri calf),
respectively, and correlated them with those of the whole leg
(Re leg, Ri leg). Theoretically, Re leg and Ri leg each should be
equal to the sum of Re thigh and Re calf, and Ri thigh and Ri calf,
respectively. This was roughly true both in the thigh
(R2 = 0.74, p < 0.0001) and calf (R2 = 0.35, p < 0.0001), al-
though a certain range of deviation was observed (Fig. 3).
Therefore, it seemed acceptable to divide the whole leg im-
pedance into the impedances of thigh and calf for analysis.
BIA in legs with VE, LE, and normal legs
The results showed that Ri/Re leg, a surrogate parameter for
ECF/ICF of a whole leg, was increased both in VE and LE
compared to that in N (N 2.5 – 0.7, VE 3.9 – 3.7, LE
3.7 – 1.5), and this was true regardless of treatment status
(Fig. 4A left). Since Ri leg was similar among these legs
(Fig. 4A right), this increase in Ri/Re leg was considered to be
caused by a decrease in Re leg (Fig. 4A middle). When the
thigh and calf were separately evaluated, the same phenom-
enon was confirmed in the calf. Namely, Ri/Re calf was in-
creased in legs with edema (N 2.6 – 0.6, VE 4.6 – 1.7, LE
4.4 – 2.2); this was true regardless of treatment status (Fig. 4B
left), due to a decrease in Re calf (Fig. 4B middle), while Ri calf
in these legs was similar (Fig. 4B right). In the thigh, an
increase in Ri/Re thigh was again confirmed in legs with edema
(N 1.8 – 0.5, VE 3.2 – 3.8, LE 3.8 – 1.9). However, this was
slightly affected by treatment status (Fig. 4C left). Re thigh was
decreased in treated VE, and a further decrease was observed
in treated/untreated LE (Fig. 4C middle). In contrast to the
calf, Ri thigh was increased both in VE and LE compared to N
(Fig. 4C right). The ratios of Ri/Re calf to Ri/Re thigh (C/T ratio)
were compared and were almost the same in N and untreated
VE/LE. A significant decrease in C/T ratio was observed
between treated and untreated LE, but this was not the case in
VE (Fig. 4D).
Discussion
The increase in Ri/Re in the whole leg, thigh, and calf could
be demonstrated in VE and LE. Moreover, it was found that
C/T ratio was similarly about 1.5 in N and untreated VE/LE.
This indicated that distribution of fluid in a leg according to
gravity was constant regardless of fluid status in the leg.
5
Accordingly, we could not find any particular feature in LE in
terms of fluid distribution pattern.
In the current study, C/T ratio was decreased by treatment
in LE, but not in VE. We previously reported that a greater
decrease of fluid in the calf, which was confirmed as de-
creased subcutaneous echo-free space, was found compared
to that in the thigh in treated LE.9 This might be related to
hosiery design, with relatively insufficient interface pressure
at the thigh that could cause ‘‘the toothpaste in the toothpaste
tube’’ phenomenon, in which fluid was squeezed by a higher
compression below to a lower compression area above. In
contrast, this decrease in C/T ratio was not observed in VE.
One possible explanation is the nature of the disease and its
treatment. With compression therapy, most VE is resolved;
thus, diligent patients did not appear in the clinic because
they could manage the symptoms by themselves. Since LE
caused persistent leg swelling, the more diligent the patient,
the more regularly he/she appeared. Therefore, it was spec-
ulated that less compliant patients and/or nonresponders had
VE, while more diligent patients had LE.
Interestingly, Ri was similarly increased in the thigh with
VE and LE. It is known that fluid with a relatively high
protein content accumulates in LE10 and that chronic in-
flammation is ongoing in legs with LE and VE.11 These might
have changed the electrolyte concentration and/or capacitive
resistance of the cell membrane, resulting in a change in
the coefficient of resistivity.12,13 Further investigation will be
needed to clarify this issue.
Limitations
Since our clinic manages a limited number of patients with
leg edema, we could only assess a small numbers of legs,
particularly those with untreated LE. However, as LE treat-
ment is gaining popularity, many patients had already started
some form of compression therapy before they visited our
clinic. A study, including a larger number of patients, will be
needed to draw definite conclusions.
Conclusion
BIA demonstrated that fluid content was increased both in
legs with VE and LE, compared to N, while the mode of
gravitational fluid distribution was similar among all legs.
Thus, no specific finding for LE was confirmed using BIA in
the present study.
Author Disclosure Statement
No competing financial interests exist.
References
1. Ward LC, Bunce IH, Cornish BH, Mirolo BR, Thomas BJ,
Jones LC. Multi-frequency bioelectrical impedance aug-
ments the diagnosis and management of lymphoedema in
post-mastectomy patients. Eur J Clin Invest 1992; 22:751–
754.
2. Ward LC. Bioelectrical impedance analysis: Proven utility
in lymphedema risk assessment and therapeutic monitoring.
Lymphat Res Biol 2006; 4:51–56.
3. Ward L, Winall A, Isenring E, et al. Assessment of bilateral
limb lymphedema by bioelectrical impedance spectros-
copy. Int J Gynecol Cancer 2011; 21:409–418.
6 SUEHIRO ET AL.
4. Schoeller DA, Alon A, Manekas D, et al. Segmental bio-
impedance for measuring amlodipine-induced pedal edema:
A placebo-controlled study. Clin Ther 2012; 34:580–592.
5. Ward LC, Dylke E, Czerniec S, Isenring E, Kilbreath SL.
Reference ranges for assessment of unilateral lymphedema
in legs by bioelectrical impedance spectroscopy. Lymphat
Res Biol 2011; 9:43–46.
6. Ward LC, Czerniec S, Kilbreath SL. Quantitative bioim-
pedance spectroscopy for the assessment of lymphoedema.
Breast Cancer Res Treat 2009; 117:541–547.
7. Cornish BH, Thomas BJ, Ward LC, Hirst C, Bunce IH. A
new technique for the quantification of peripheral edema
with application in both unilateral and bilateral cases. An-
giology 2002; 53:41–47.
8. Suehiro K, Morikage N, Murakami M, et al. Subcutaneous
tissue ultrasonography in legs with dependent edema and
secondary lymphedema. Ann Vasc Dis 2014; 7:21–27.
9. Suehiro K, Morikage N, Murakami M, et al. A study on
increase in leg volume during complex physical therapy for
leg lymphedema using subcutaneous tissue ultrasonogra-
phy. J Vasc Surg Venous Lymphat Disord 2015;3:295–302.
10. International Society of Lymphology. The diagnosis and
treatment of peripheral lymphedema: 2013 Consensus Docu-
ment of the International Society of Lymphology. Lymphol-
ogy 2013; 46:1–11.
11. Bergan JJ, Schmid-Schonbein GW, Smith PD, Nicolaides
AN, Boisseau MR, Eklof B. Chronic venous disease. N Engl
J Med 2006; 355:488–498.
12. Deurenberg P. Limitations of the bioelectrical impedance
method for the assessment of body fat in severe obesity.
Am J Clin Nutr 1996; 64(3 Suppl):449S–452S.
13. Thomas BJ, Cornish BH, Ward LC, Jacobs A. Bioimpe-
dance: Is it a predictor of true water volume? Ann N Y
Acad Sci 1999; 873:89–93.
Address correspondence to:
Kotaro Suehiro, MD
Division of Vascular Surgery
Department of Surgery and Clinical Science
Yamaguchi University Graduate School of Medicine
1-1-1 Minamikogushi, Ube
Yamaguchi 755-8505
Japan
E-mail: ksuehiro-circ@umin.ac.jp