J Appl Physiol
91: 733–736, 2001.
Validation of equilibration and chromium reduction
methods for deuterium measurements of fluid volumes
T. HEDESTIG, A. EBBERYD, AND S. G. E. LINDAHL
Department of Anaesthesia and Intensive Care, Karolinska Hospital
and Institute, S-171 76 Stockholm, Sweden
Received 16 October 2000; accepted in final form 10 April 2001
body water; stable isotopes; mass spectrometry
physiologically well regulated although life threatening water changes are not
uncommon. During surgery and in connection with
trauma, fluid therapy and balance are vital for patient
outcome. Although infusion therapy is equally important at all ages, it is at the extremes of age, i.e., in
neonates and in the elderly, that fluid balance often is
a difficult and delicate matter in clinical practice for
both in-hospital treatment and prehospital care. It is
therefore of great value to be able to accurately measure changes in body water content.
The latest development of gas isotope ratio mass
spectrometers (GIRMS) using stable isotopes as tracers has opened up new possibilities for the clinical use
of this highly technological device. Measurements with
THE VOLUME OF BODY WATER IS
Address for reprint requests and other correspondence: T. Hedestig, Dept. of Anesthesia and Intensive Care, Karolinska Hospital,
S-171 75 Stockholm (E-mail: [email protected]).
http://www.jap.org
isotope tracers by using a dilution principle are well
known, and deuterium is a widely used isotope for
measurements of total body water (TBW). Analyses of
deuterium/hydrogen (D/H) have previously been done
by reducing water samples, with uranium or zinc, to H2
and then analyzing for D/H (1). An alternative method
is the use of equilibration (EQ) between samples of
water and gaseous H2 followed by analyses of D/H in
the equilibrated H2 (8). A promising new technique for
measurement of hydrogen isotopes using chromium
reduction (CR) has recently been developed (4). This
new CR method also presents possibilities for measurements of small samples in all kinds of body liquids. The
aim of this study was to evaluate the CR method and
compare it to the EQ technique. Precision, accuracy,
and variability for measurements of TBW were tested
in a body compartment model.
METHOD
Experimental design. Different weights of ordinary tap
water were contained in water tanks to cover a simulated
range of total body water volumes from neonates to adults.
The following “body water” volumes were used: 1.8 ⫻ 103 ml,
3.0 ⫻ 103 ml, 15 ⫻ 103 ml, 30 ⫻ 103 ml, 45 ⫻ 103 ml, and 54 ⫻
103 ml. A high-precision digital scale (Stathmos, Vaxjo, Sweden) was used and had an exactitude of ⫾10 g in the weight
range between 0 and 120 kg.
A high-precision laboratory scale was used for weighing
the correct amounts of tracer (0.1 g/kg 99.98%) to be used.
The deuterium was diluted in tap water and added to the
water tanks. The dose-containing bottle was rinsed with tap
water, and this water was also added to the tanks. Double
samples were taken from background, tap water, diluted
dose, and postdose fluids. Each sample was divided into four
parts and processed in parallel. On the basis of the result of
these four measurements, a mean was calculated for each
sample. The final value was then taken as the arithmetic
mean of the two original duplicate samples. Because the
different fluid volumes were weighed and because the density
of water is 0.99 kg/dm3, the phrases volume and weight have
been used interchangeably in the study.
EQ technique. This system is a fully automatic EQ system
(Thermo Finnigan MAT, Bremen, Germany). Glass bottles
(20 ml) were filled with the water samples, and platinum
catalysts (Hokko sticks) were placed in the bottles before
The costs of publication of this article were defrayed in part by the
payment of page charges. The article must therefore be hereby
marked ‘‘advertisement’’ in accordance with 18 U.S.C. Section 1734
solely to indicate this fact.
8750-7587/01 $5.00 Copyright © 2001 the American Physiological Society
733
Downloaded from http://jap.physiology.org/ by 10.220.32.247 on June 18, 2017
Hedestig, T., A. Ebberyd, and S. G. E. Lindahl. Validation of equilibration and chromium reduction methods for
deuterium measurements of fluid volumes. J Appl Physiol
91: 733–736, 2001.—Determinations of fluid volumes are of
importance for correct treatment of patients subjected to
shock and trauma. Gas isotope ratio mass spectrometry
(GIRMS) is an advanced method for analysis of stable isotopes. These can be used as tracers for measurement of
various fluid volumes. In the current in vitro study, deuterium was used to determine different volumes of water simulating a range of body fluid volumes from neonates to
adults. A high-precision scale gave control weights (i.e., volumes), and two methods, equilibration (EQ) and chromium
reduction (CR), were compared by use of a GIRMS. The
coefficient of variation was ⬍1% when using both EQ (0.45%)
and CR (0.79%). The variability was greater at small volumes, and, when regression equations for the relation between measured and calculated volumes were used as formulas, the deviation was 0.4% using EQ and 2.8% using CR at
the volume of 1,000 ml. At larger volumes, the deviation
when using CR approached 1%. These variations are better
than previously published data using other methods. It was
concluded that GIRMS is a suitable technique for fluid volume determinations in neonates as well as in adult patients,
using deuterium as a tracer. EQ and CR methods were both
regarded to give acceptable variabilities in this in vitro study.
GIRMS may in the future increasingly be used clinically for
accurate measurements of body fluid volumes.
734
DEUTERIUM MEASUREMENTS OF FLUID VOLUMES
they were connected to the analyzing rack. The rack was
mounted on a sled that enabled the bottles to be immersed in
a water bath. The sled was shaken using an eccentric drive.
The temperature was maintained constant at 18°C ⫾ 0.05°C.
The temperature was chosen to minimize condensation inside the bottles. The sample bottles were connected via a
short capillary to a manifold with a pneumatic valve. The line
was evacuated by using a rotary vacuum pump. Then the
bottles were filled with EQ gas. The EQ pressure (0.4–0.6
bar) gave an appropriate pressure in the inlet system after
EQ. The EQ time was set to 120 min.
After EQ, the sample gas was sequentially transferred to
the inlet system of the mass spectrometer. The transfer line
was immersed in a cold trap to prevent water vapor from
entering the inlet system. The isotope ratio measurements
were carried out on a Delta plus mass spectrometer (ThermoQuest). The H3⫹ factor was measured, and the contribution to the mass 3 ion current was quantitatively corrected.
CR method. Chromium reduces water to hydrogen gas at
⬎700°C according to the reaction
The reaction quartz tube was filled with ⬃50 g of chromium
powder with a layer of quartz wool. The reaction furnace was
connected on-line with the GIRMS. Samples are loaded into
0.75-ml vials in the tray of an auto sampler A200s. The
sample, 2 ␮l, was injected by the auto injector into the
reaction furnace at 900°C, by use of a gas-tight microsyringe.
The sample quickly evaporated and was reduced almost
instantly at the contact with chromium. The hydrogen gas
flows into the inlet system because of the pressure gradient.
Measurements start after pressure EQ between the reaction
furnace and the variable volume bellows. Reaction and EQ
times are 80 s from injection to start of measurement. In this
study, an automatic H-device was used (ThermoQuest).
Apparatus. GIRMS is an analytical technique for accurate
and precise measurements of stable isotopes. The mass spectrometer consists of an inlet system, an ion source, a magnet,
and a collector. The entire analyzing process is completely
controlled by a computer. Samples must be in a gaseous state
before they are introduced into the ion source. For hydrogen
isotope measurements, physiological fluids must be converted to hydrogen gas. This gas sample and a reference gas
are usually introduced into the ion source through the inlet
system. The inlet system matches the pressures between
sample and reference gas. The gas molecules are presented
sequentially through capillary leak lines into the ion source,
which is under high vacuum. In the source, the gas molecules
are ionized by electrons from a filament wire. Some of the
molecules lose an outer electron and become positively
charged. Repelling electrodes force these ions through a series of focusing lenses into the analyzing part of the mass
spectrometer. A magnet is used to deflect the molecular ions
according to their masses. Each mass has different ion beams
and separated into specific collectors. A detectable electron
current is discharged, and the ratio of these ion currents is
measured and directly related to the isotope ratio.
The isotope enrichment is expressed as the difference
between the observed ratio of the sample and the reference
gas in delta per mille.
Calculations. The dilution spaces were calculated from the
equation
N⫽
共WA/a兲共␦a ⫺ ␦t兲f
共␦s ⫺ ␦p兲
J Appl Physiol • VOL
RESULTS
The average coefficient of variation within subjects
was 0.49% for EQ and 0.79% for CR, indicating a high
reproducibility with both techniques. Corresponding
precisions were 0.98 and 1.58%, respectively.
Variabilities between calculated and measured
weights. The highest variability when using EQ was
measured at the lower volume range (Table 1, Fig. 1A).
The variability when using CR was somewhat greater
(Table 1, Fig. 1B). The differences between EQ and CR
were, however, not statistically significant.
Correlation between calculated and measured
weights. The relationship between calculated and measured volumes, using EQ, was highly correlated (R2 ⫽
1.0) and could be expressed by the equation Vcalculated ⫽
0.433 ⫹ 1.004 ⫻ Vmeasured (r ⫽ 1.0). The same relationship, using CR, was described according to the expression Vcalculated ⫽ ⫺37.31 ⫹1.01 ⫻ Vmeasured (r ⫽ 1.0),
where V is volume. Regression lines for EQ and CR are
presented in Fig. 2A and 2B.
DISCUSSION
The most important findings in this study were that
GIRMS, using deuterium as a tracer, resulted in preTable 1. Comparisons of various fluid volumes,
measured, and calculated with deuterium as tracer
Measured
Volume, ml
Calculated
Volume, ml
1,803
3,022
15,000
30,000
45,000
54,000
1,831
3,040
15,014
30,116
45,206
54,191
Mean ⌬%
SD
Range
29
16
28
134
116
223
75
48
62
326
308
489
1.56
0.59
0.09
0.39
0.46
0.35
109
133
199
435
1903
542
0.46
0.80
⫺0.14
0.81
1.15
0.78
EQ
CR
1,803
3,024
15,000
30,000
45,000
54,000
1,811
3,048
14,972
30,243
45,519
54,423
41
53
90
177
689
268
Two analyzing methods, equilibration (EQ) and chromium reduction (CR), were used; n ⫽ 10 in each volume group. SD, standard
deviation between measured and calculated volumes; Range, gap
between minimal and maximal values of differences (in ml); mean
⌬%, mean percentage difference between calculated and measured
volumes.
91 • AUGUST 2001 •
www.jap.org
Downloaded from http://jap.physiology.org/ by 10.220.32.247 on June 18, 2017
2Cr ⫹ 3H2O 3 Cr2O3 ⫹ 3H2
where N is dilution space, expressed in grams, W is the mass
of water used to dilute the dose, A is the dose administered,
a is the mass of the dose used in preparing the diluted dose,
f is the fractionation factor for the physiological sample, ␦a is
the measured value for the diluted dose, ␦t is the value for the
tap water used in the dilution, ␦s is the value for the postdose physiological sample, and ␦p is the value for the predose
physiological sample (5).
Statistics. Mean volumes (⫾ SD) were used, and results
were evaluated by regression analysis and Mann-Whitney
U-test. A P value ⬍0.05 was considered to indicate statistical
significance.
DEUTERIUM MEASUREMENTS OF FLUID VOLUMES
735
Fig. 1. Differences between calculated and measured volumes, expressed in ⌬%, for various volume sizes. Individual values are presented using equilibration (EQ; A) and chromium reduction (CR; B).
cise fluid volume determinations of small as well as
large fluid spaces and that the reproducibility was
0.49% (mean) when using EQ and 0.79% when using
CR. In addition, the CR technique requires samples of
only 2 ␮l compared with the EQ technique in which 5
ml are needed. This opens up possibilities for the use of
CR in neonates not only for single samples but also for
serial tests and invites a more frequent clinical use of
this technique whenever accurate determinations of
body fluid volumes are needed.
The difference in reproducibility between EQ and CR
techniques was small. It can therefore be argued that
the techniques are interchangeable. The intercept,
from regression analysis of ⫺37 when using the CR
technique compared with 0.43 using the EQ technique,
may result in an unacceptable deviation between measured and calculated volumes when using CR. Thus
the error caused by the intercept is insignificant for
large volumes, which means that deviations from measured volumes are almost identical with the slope of
the regression lines, i.e., 1.004 using EQ and 1.01 using
the CR technique, which means a mean variability of
0.4% in the adult using EQ and of 1.0% using CR. On
the basis of these results, the two techniques, EQ and
CR, are both acceptable and fulfill high demands on
precision and reproducibility in large fluid volumes.
Even for small volumes of 1,000 ml, the CR technique
results in a deviation of 3%, which is acceptable.
J Appl Physiol • VOL
Fig. 2. Relationships between measured and calculated fluid volumes obtained by EQ (A) and CR (B) techniques are shown. The
regression equation for EQ determinations was Vcalculated ⫽ 0.433 ⫹
1.004 ⫻ Vmeasured (r ⫽ 1.0) and, for CR analysis, Vcalculated ⫽ ⫺37.31
⫹1.01 ⫻ Vmeasured (r ⫽ 1.0), where V is volume.
91 • AUGUST 2001 •
www.jap.org
Downloaded from http://jap.physiology.org/ by 10.220.32.247 on June 18, 2017
GIRMS is a technically advanced method, expensive
to purchase, maintain, and run. Also, it requires a
specially trained operator. It can be questioned
whether this currently is an apparatus for clinical use.
Alternatives may be more versatile, rapid, portable,
and easy to use bedside. Bioimpedance (BIA) is one
alternative. BIA measurements of TBW have betweenmeasurement variations of 2–4 liters in an adult patient, i.e., a variability between 5 and 10% at a volume
of 40 liters (6). BIA can therefore only detect relatively
large intraindividual changes and may be best suited
for quantification of group changes. Another alternative method is infrared spectrophotometric determinations of deuterium. This approach, however, is not as
precise as GIRMS, with a mean precision of 2.5% using
repeated measurements of the same sample (7). The
stable isotope, 18O, could also be used for determinations of body fluid spaces. It has been shown that the
deuterium dilution space is ⬃2% larger than that with
18
O in adult patients and ⬃3% larger in premature
infants (9). This is probably because deuterium participates in the nonaqueous exchange to a greater extent
than does 18O (2). This is explained by the fact that the
water-to-protein ratio is on the order of 3:1 in adults
and 5:1 in neonates (3). Thus there are advantages
with use of 18O determinations of body fluid volumes.
736
DEUTERIUM MEASUREMENTS OF FLUID VOLUMES
J Appl Physiol • VOL
more time and larger sample volumes. The potential
for the use of stable isotopes such as deuterium and 18O
as tracers in future studies of neonates and adults,
focused on various body water spaces, body composition, and energy expenditure with relation to anesthesia, surgery, and intensive care, is promising.
The study was supported by grants from the Swedish Medical
Research Council (10401).
REFERENCES
1. Bigeleisen J, Perlman ML, and Prosser HC. Conversion of
hydrogenic materials to hydrogen for isotopic analysis. Anal
Chem 24: 1356–1357, 1952.
2. Culebras JM and Moore FD. Total body water and the exchangeable hydrogen. I. Theoretical calculation of nonaqueous
exchangeable hydrogen in man. Am J Physiol Regulatory Integrative Comp Physiol 232: R54–R59, 1977.
3. Fomon SJ, Hascke F, Ziegler EE, and Nelsson SE. Body
composition of reference children from birth to age 10 years.
Am J Clin Nutr 35: 1169–1175, 1982.
4. Gehre M, Hoefling R, Kowski P, and Strauch G. Sample
preparation device for quantitative hydrogen isotope analysis
using chromium metal. Anal Chem 68: 4414–4417, 1996.
5. Halliday D and Miller DG. Precise measurement of total body
water using tracer quantities of deuterium oxide. Biomed Mass
Spectrom 4: 82–87, 1977.
6. Kushner RF. Bioelectric impedance analysis: a review of principles and applications. J Am Coll Nutr 11: 199–209, 1992.
7. Lukash HC and Johnson PE. A simple, inexpensive method of
determining total body water using a tracer dose of D2O and
infrared absorption of biological fluids. Am J Clin Nutr 41:
363–370, 1992.
8. Ohsumi T and Fujino H. Isotope exchange technique for preparation of hydrogen gas in mass spectrometric D/H analysis of
natural waters. Annal Sci 2: 489–490, 1986.
9. Racette SB, Schoeller DA, Luke AH, Shay K, Hnilicka J,
and Kushner RF. Relative dilution spaces of 2H- and 18Olabeled water in humans. Am J Physiol Endocrinol Metab 267:
E585–E590, 1994.
10. Scholler DA, Colligan AS, Shriver T, Avak H, and BartokOlsson C. Use of an automated chromium reduction system for
hydrogen isotope ratio analysis of physiological fluids applied to
doubly labelled water analysis. J Mass Spectrom 35: 1128–1132.
91 • AUGUST 2001 •
www.jap.org
Downloaded from http://jap.physiology.org/ by 10.220.32.247 on June 18, 2017
These measurements of 18O require the same GIRMS
technique as with deuterium, but 18O is more expensive. 18O is, however, an excellent alternative to deuterium for body fluid measurements. The difference
between 18O and deuterium measurements is systematic, and deuterium is often the tracer used. In this in
vitro study using deuterium and fluids that did not
contain any proteins, there was a low variability of
⬍1% and a good precision when the clinically more
versatile CR technique was applied. Taken together,
the deuterium approach for body fluid volume determinations is acceptable for studies in humans.
It was interesting to note that deviations from measured volumes were in most cases positive, indicating
that the scale may underestimate volumes. Also the
variability was somewhat greater when using CR compared with EQ. The differences were small, however,
and most probably of minor clinical importance although all EQ determinations of small volumes were
positive, indicating a systematic deviation that should
be elucidated in future studies. Reduction of water by
the use of chromium instead of zinc, uranium, and
perhaps also manganese is advantageous because the
other reduction reagents are known to give a larger
variability (4). This is thought to be due to a less
precise determination when analyzed water contains
impurities, which always occur in biological samples.
Hence, the CR technique seems to be acceptable for
measurements of TBW in adult humans; it requires
smaller sample volumes (2 ␮l) compared with EQ (5
ml), is faster, and may not be disturbed by fluid samples such as blood, plasma, urine, and spinal fluid that
contain organic material (10).
In conclusion, both small and large fluid volume
determinations using GIRMS with deuterium as a
tracer were reproducible with low coefficients of variation; they are also precise and accurate. CR of water
seems to be acceptable for clinical use compared with
other reductive metals. The EQ technique requires