The understanding of fluid changes during hemodialysis (HD) is essential fo
r reducing complications as well as efficacy of the procedure. Bioimpedance
spectroscopy provides a noninvasive method of measuring total body water (
TBW), the distribution of intra (ICF) and extracellular (ECF) fluids, and t
heir changes during HD. Segmental bioimpedance may be used to measure the s
ame fluid shifts but from different body segments; the technique has previo
usly been shown to compare well with whole body measures. It is possible th
at fluid shifts occur differently in different body compartments during HD.
Based on previous hemodynamic studies we postulated that during HD ultrafi
ltration (UF) the body attempts to preserve its central blood volume (cardi
opulmonary circulation plus great vessels), and thus fluid shifts would be
greater from the periphery than from central compartments. To test this hyp
othesis, segmental bioimpedance (Xitron Technologies, San Diego, CA) was pe
rformed on 11 subjects undergoing HD where ECF and ICF values were obtained
from the legs, arms and trunk before and after a period of UF. Blood volum
e change (Delta BV%) was also followed using an on-line optical hematocrit
(Hct) sensor (Crit-Line monitor, In-Line Diagnostics, UT) where Delta BV% =
Delta .BV% (1 - Hct(1)/Hct(0)) x 100 (Hct(0) = baseline Hct; Hct(1) = post
ultrafiltration Hct). The UF of 2.0 L +/- 0.79 L (M +/- SD) over 75 minutes
was associated with a Delta BV% of -9.43% +/- 3.6% (M +/- SD), a significa
nt (Student's paired t-test) reduction in total body (TB) ECF (p < 0.02), a
weak correlation in reduction in TBW (p = 0.09) but not in TB ICF. The ECF
reductions from the trunk, legs, and arms were all significant (minimum p
< 0.02); no ICF changes from these compartments were significant. The amoun
t of ECF reduction was greater from the legs (0.7 L +/- 0.6 L) than the arm
s (0.12 L +/- 0.08 L) and trunk (0.2 L +/- 0.2 L) (all M +/- SD). Multiple
regression analysis showed that TB ECF changes correlated strongly with leg
(r = 0.94, p < 0.001) and arm (r = 0.72, p = 0.002) ECF changes but not wi
th trunk changes. Delta BV% correlated weakly with leg (r = 0.45, p = 0.08)
and arm (r = 0.42, p = 0.10) ECF changes but not with the trunk. As the De
lta BV% represents the net volume change between OF and plasma water refill
ing, this indicates that plasma water is being removed more from the periph
eral compartments than from the trunk. These data suggest that plasma refil
ling during HD to preserve central blood volume is more dynamic from the le
g ECF than from elsewhere and may, in turn, explain the frequent occurrence
of leg cramps during and after hemodialysis.