27

Clinica Chimica Acta, 92 (1979) 27-32

@ Elsevier/North-Holland Biomedical Press

CCA 9847

A GAS CHROMATOGRAPHIC METHOD FOR DETERMINATION OF

ACETATE LEVELS IN BODY FLUIDS

MORTEN KVEIM a and JAN ERIK BREDESEN b**

a Institute for Surgical Research, University of Oslo, Rikshospitalet, Oslo and b Division of
Clinical Pharmacology and Toxicology, The Central Laboratory, Ulledl Hospital, Oslo
(Norway)

(Received July 5th, 1978)

Summary

A simple, sensitive and accurate gas chromatographic method for the deter-
mination of acetate in plasma and urine is described. Following precipitation of
proteins by perchloric acid and removal of potassium perchlorate by precipita-
tion with KOH, acetic acid is quantitated by injection of the acidified sample
on a highly polar column. A flame ionization detector is used. Propionic acid is
added to the sample initially as internal standard. The sample volume required
for the analysis is 0.5 ml.
The standard curve is linear in the range from 0.01 to 10 mmol/l. Recovery
of added acetate is complete, and the coefficient of variation is less than 4%
even at low concentrations.
The concentration of free acetate in plasma from 10, and urine from 4, healthy
humans ranged from 0.04 to 0.07 and 0.09 to 0.24 mmol/l, respectively.

Introduction

Free acetate is a physiological component of body fluids. Although several

gas liquid chromatographic (GLC) methods for the quantitative determination
of acetate in aqueous solution have been reported [l--8], many aspects of its
metabolism, quantitative as well as qualitative, still remain obscure. This is
partly due to difficulties in connection with the determination of acetate levels
in biological materials.
Many gas chromatographic methods of acetate analysis include a preliminary
clean up step [4,6,9-111 which makes the methods both time consuming and
less accurate. Adsorption of the acids to the column is another problem of such

* To whom correspondence should be addressed.

28

methods, resulting in tailing and irregular shape of the acid peaks, as well as
ghosting [8,12,13].
While tailing of the peaks is directly visible, ghosting is much more difficult
to detect. Water injections are usually used to detect ghosting, but it has been
shown that water does not always elute adsorbed acids [14]. Addition of
formic acid to the carrier gas is a simple and effective treatment recommended
to suppress ghosting and tailing [ 1,3,5,12].
The present report describes a method where a clean up step is avoided and
where ghosting is suppressed by adding formic acid to’ the carrier gas. The
present method is similar to that of Wadke and Lowenstein [7] and Desch and
Descomps [8], but represents some additional advantages making it more
accurate in the low concentration area, as well as more, practical and suitable
for serial analysis. In principle, the biological fluid is deproteinized, acidified
and injected into the gas chromatograph.

Material and methods

Reagents
Sodium propionate (Sigma Chem. Co) 10 mmol/l solution, perchloric acid
20 g/l, KOH 10 mol/l, thymol blue 1 mol/l, HCl 10 mol/l, FFA-free albumin
(Sigma Chem. Co.) were used.

Procedure
To 0.5 ml plasma or urine is added 25 ~1 of 10 mmol/l sodium propionate
solution as internal standard and 25 ~1 of thymol blue (1 mol/l) as an indicator.
The proteins are then precipitated by the addition of 50 ~1 of perchloric acid
(20 g/l). After centrifugation (3000 r.p.m., 5 min) the supernatant is with-
drawn and made alkaline by the addition of KOH (10 mol/l) to pH 10 (blue
colour of the indicator). The insoluble potassium perchlorate is removed by
centrifugation (3000 r.p.m., 5 min) and the supematant is acidified with HCl
(10 mol/l) to pH 2 (red colour of the indicator), before injection into the gas
chromatograph.

Samples of analysis
A standard curve has to be made for each series of analyses. Albumin solu-
tions, 6 g/100 ml, containing different acetate concentrations (0.05, 0.1, 0.5,
1 .O, 5.0 and 10.0 mmol/l, as well as blank solution) are prepared and treated as
described. For the determination of. recovery and accuracy albumin standards
of 0 .l and 5 .O mmol/l acetate concentrations were used.

Gas chromatography
A gas chromatograph (Varian 2000) with flame ionization detector was used.
The oven temperature was 130°C and the temperature, of both the injector and
detector was 150°C. Separation was carried out in a glass column (0.4 cm i.d. X
100 cm) packed with 10% free fatty acid phase (FFAP) on Chromosorb W AW
80-100 mesh (Supelco Inc.). The following columns were also tested for the
analysis of free fatty acids: 3% Carbowax 20 M/0.5% H3P04 on 60/80 Carbo-
pack B (Supelco Inc.), 10% SP 1200/l% H3P04 on 80/100 Chromosorb W AW
 29

(Supelco Inc.). The carrier gas was nitrogen, acidified by passage through an
atmosphere saturated with formic acid vapour. The carrier gas flow was 30 ml/
min, the hydrogen flow 30 ml/min, and the air flow 300 ml/min. Under these
conditions acetic acid eluted after 1 min, and the internal standard (propionic
acid) after 1.5 min. 2 mol/l hydrochloric acid was always injected prior to each
sample injection. If ghost peaks appeared that might interfere with the acetate
or propionate peaks, the injection of acid was repeated until the ghost peaks
disappeared. In the rare case of a persistent ghost peak the analysis was inter-
rupted and the first 10 cm of the column material changed. Peak heights rather
than areas were used for quantitative determinations.

Results

Fig. 1 shows a typical gas chromatograni obtained by analysis of a normal

plasma sample. Fig. 2 shows a typical standard curve (left), and a recovery
curve of acetate added to normal plasma (right). The peak height ratios
between acetate and internal standard, plotted against the acetate concentra-
tion of albumin standards gave a straight line in the concentration range tested
(0.01-10 mmol/l). The standard curve passed through the origin. The recovery
of added acetate was quantitative as shown in Fig. 2 and Table I. The coeffi-
cient of variation was determined in two different concentration areas (0.1
mmol/l and 5.0 mmol/l) and was less than 4% even at the low concentrations of
acetate (Table I).

prop

I
relative response
-/prop
a.0

ac
t

Fig. 1. A typical gas chromatogram

1.0

kL
0.5
mmolfl

1.0
///I 0.S
mmol/l

1.0

of a normal plasma sample. Experimental conditions as described in

text.

Fig. 2. (Left) A typical standard curve showing the linear relationship between relative response acetate/
propionate and acetate concentration. (Right) Recovery of acetate added to normal plasma. Experimental
conditions as described in text.
30

TABLE I

THE COEFFICIENT OF VARIATION AND RECOVERY OF THE METHOD

The calculations are made on the basis of 20 parallel observations in two different concentration areas.

Acetate concn. Coeff. of variation Mean recovery

(mmol/l) (Per cent) (per cent)
n = 20 n = 20

0.10 23.3 93
5.0 +3.1 100

Acetate concentrations in venous plasma samples from 10 normal persons

(2 males, 8 females) ranged from 0.04 to 0.07 mmol/l with a mean of 0.06
mmol/l. Freshly voided urine from 4 normal persons contained 0.09 to 0;024
mmol/l.
Storage of heparinized blood samples for 30 h at +4”C had no influence on
plasma acetate concentrations. A 40% increase was observed, however, after
storage of the blood sample at room temperature for 30 h prior to separation.
Reanalysis of plasma kept at -20°C for 5 months revealed no change in acetate
concentration. Alkaline, deproteinized samples could be stored in the refrigera-
tor for at least three days without influencing the acetate content.

Discussion

Most gas chromatographic methods include a preliminary separation step in

order to get rid of non-volatile components of the deproteinized plasma. This
has been accomplished by micro-distillation as suggested by Bartley [4], and
later modified by Frijhlich and Wieland [ 61 by steam distillation [9], by thin-
layer chromatography [ll], or by extraction with toluene [lo]. In common
with Wadke and Lowenstein [7] we omitted such a step in the present method.
Friihlich and Wieland argue that avoidance of a clean-up step would be accom-
panied by contamination of the column system with resulting instability and
prolongation of the injection intervals [6].. This is not a problem with the
present method, provided that the first 10 cm of the column is changed
frequently. Ghost peaks are further reduced to a minimum by saturation of the
carrier gas by formic acid vapour [1,2,12]. ‘Normally ghost peaks which inter-
fere with acetate and propionate are undetectable. If ghost peaks appear after
repeated injection of acidified water one should not hesitate to change the
packing material.
In a previous paper [15] dealing with the acetate metabolism another gas
chromatographic method was used. This procedure included a concentration
steG at alkaline pH’ at a temperature of 120°C. During this step glucose was
degraded, not only to yield the wel known browning products (formic acid,
levulinic acid, f&fural derivatives, etc.) [16], but also acetate, which was
formed in large amounts (Kveim, M., unpublished data). Due to this unfor-
tunate and unexpected decomposition reaction the concentration step was
avoided in the Ijresent method, ‘and no interference from glucose was seen.
Meanwhile Wadke and Lowenstein [7] published a method .similar to the
 31

method described here. Our method has, however, definite advantages. By

using perchloric acid as a protein precipitant instead of trichloroacetic acid, the
constant impurity peak coinciding with the acetate peak [7] was avoided and
corrections became unnecessary. Moreover, in studies on acetate metabolism
one regularly has’ to assay concentrations of acetate below 0.1 mmol/l. As
claimed by Wadke and Lowenstein [ 71 the use of perchloric acid as a protein
precipitant gives a number of peaks which may substantially interfere with the
acetate as well as the propironate detection. By removing the perchloric acid as
insoluble potassium perchlorate, however, this problem of interference was
solved.
Like Wadke and Lowenstein [7] we have also tried 10% SP 1200/l% H3P04
on 80/100 Chromosorb W AW (Supelco Inc.). We have further tried 3% Carbo-
wax 20 M/0.5% H3P04 on 60/80 Carbopack B (Supelco Inc.), which has been
recommended for the chromatography of acetic acid in water. Both materials
showed ghost peaks in the critical region of the chromatogram. In contrast, the
use of 10% FFAP on 80-100 Chromosorb W (Supelco Inc.) did not result in
such peaks. This packing material was therefore also chosen for the present
method.
The acetate peak of the plasma sample chromatogram coincided with the
acetate peak of the standard solution chromatogram in the various chromato-
graphic systems. This proves for practical purposes the true identity of the
plasma acetate peak. Mass spectrometry was also tried in order to support this,
but the quantities of acetate in the unconcentrated plasma sample turned out
to be too small for the techniques available to us.
The reproducibility, sensitivity and recovery data obtained by the present
method are comparable to, or better than, those of methods previously
reported [3,5-8,10,13,17,18]. In particular, our method allows exact mea-
surement of relatively small changes in the lower physiological concentration
area. This must be considered in the light of the simplicity of the present
method. Like Frijhlich and Wieland [6] we have found propionate suitable as
internal standard. This compound, as well as methylmalonic acid which is
readily converted into propionic acid, are both absent from the plasma except
in extremely rare cases of specific metabolic disorders [ 11,191.
The present method has been used in experimental studies on acetate metab-
olism and has been found suitable for the determination of plasma acetate
levels both in the physiological range (about 0.05 mmol/l) and at high concen-
trations (up to 10 mmol/l).

Acknowledgement

This study was supported through grants from the Norwegian Medical Depot.

References

1 Ackmen. R.G. (1972) J. Chromatogr. Sci. 10, 560

2 Ackman. R.C. and Sipos, J.C. (1964) J. Chromatogr. 13.337
3 Annison. E.F. (1954) Biochem. J. 56.670
4 Bartle~. W. (1953) Biochem. J. 53.305
5 Ciaranfi, E. and Fonnesu. A. (1952) Biochem. J. 60.698
32

6 Fr6hlich. J. and Wieland, 0. (1968) Z. Klin. Chem. 4,277

7 Wadke. M. and Lowenstein. J.M. (1976) Methods Ensymol. 36.307
8 Desch. G. and Descomps. B. (1977) CBn. Chim. Acta 76,193
9 Perry, T.L.. Hansen. S.. Diamond. S., BulBs. B.. Mok. C. and Melancon, S.B. (1970) Clin. Chim. Acta
29,369
10 Richards. R.G.. MendenhaB, C.L. and MacGee. J. (1978) J. Lipid Res. 16.395
11 Stokke, 0.. Jellum. E., Eldjam, L. and Schnitler. R. (1973) Clin. Chim. Acta 45.391
12 Ceddes. D.A.M. and Gilmour. M.N. (1970) J. Chromatogr. Sci. 6.394
13 Gonzales, F.M.. Pearson, J.E.. Garbus. S.B. and Holbert. R.D. (1974) Trans. Am. Sot. Artif. Intern.
Org. 20,169
14 Van Eenaeme, C., Bienfait. J.M. and Pondant. A. (1974) J. Chromatogr. Sci. 12.398
15 Kveim. M. and Nesbakken. R. (1976) Trans. Am. Sot. Artif. Intern. Org. 21.138
16 Groves, M.J. (1973) Parenteral Products; the Preparation and Quality Control of Products for Injec-
tion. p. 43. Heinemann. London
17 Lundquist. F., Fugmann. U. and Rasmussen, H. (1961) Biochem. J. 80,393
18 Medsihradsky. F. and Lamprecht. W. (1965) Hoppe-Seylers Z. Physiol. Chem. 343.36
19 Hommes. F.A., Kuipers, J.R.C.. Elema, J.D.. Jansen, J.F. and Jonxis, J.H.P. (1968) Pediatr. Res. 2.
519