Gas-Chromatographic Assay for Thiopental in ... - Clinical Chemistry

CLIN.CHEM. 27/1, 113-115 (1981)
Gas-Chromatographic Assay for Thiopental in Plasma, with Use of a
Nitrogen-Specific Detector
Donald Jung, Michael Mayersohn,’ and Donald Perrier
An accurate, sensitive, and specific gas-liquid-chromatographic
procedure is described for determining concentrations
of thiopental in human plasma. After a double
extraction of 0.2 or 1.0 mL of plasma containing phenobarbital
as an internal standard, thiopental and the internal
standard are derivatized in a polar non-aqueous solvent
system with iodomethane. The reaction mixture is then
evaporated, the residue reconstituted with ethyl acetate,
and 20 tL injected into a 3% OV-17 column of a gas
chromatograph equipped with a nitrogen-phosphorus
detector. Linearity and reproducibility over the concentration
range 25 tg/L to 10 mg/L in plasma are excellent.
The sensitivity and wide range of linearity exhibited by this
method permit thorough characterization of the disposition
of thiopental after the usual induction doses of 3-4 mg/kg
of body weight.
Additional Keyphrases: barbiturates . drug assay
anesthetics
Thiopental [5-ethyl-5’-( 1 -methylbutyl) -2-thiobarbituric
acidj is an ultra-short-acting barbiturate used to induce anesthesia
in man and animals. Numerous methods of analysis
for thiopental have been reported in the literature. Most lack
the necessary specificity or sensitivity, or both, for completely
describing the disposition kinetics of the drug in humans.
Oroszlan and Maengwyn-Davies (1) ‘reported a nonspecific
spectrophotometric assay involving three different wavelengths
and having a sensitivity of 1 mg/L. A specific spectrophotofluorometric
method (2) has a reported sensitivity
of 0.1 mg/L and requires 3 mL of plasma. Several “high-performance”
liquid-chromatographic assays for thiopental in
plasma have also been reported (3, 4). The sensitivity of these
assays ranges between 0.1 and 1.0 mg/L. Numerous gas-liquid-chromatographic
(GLC) procedures for barbiturates other
than thiopental have been reported (5-9). A GLC method (10)
with use of a flame ionization detector has a reported sensitivity
of 0.2 mg/L, but we found it to be nonspecific for thiopental,
owing to thepresence of an interfering peak from the
methylated oxy analog and metabolite of thiopental (pentobarbital).
A specific and apparently sensitive assay was reported
by Van Hamme and Ghoneim (11) to have a “low-end
sensitivity” of

nobarbital per liter). The contents of the tube were vortexmixed
for 15 s and centrifuged (3000 rpm, 5 mm). The organic
(upper) phase was then transferred to a 15-mL test tube
containing 1 mL of 2.5 mol/L NaOH. The solution was vortex-mixed
for lOs, and centrifuged (3000 rpm, 1 mm), and the
hexane layer aspirated. To the basic aqueous extract was
added 4 mL of drug-free hexane and the solution was vortex-mixed,
centrifuged, and aspirated as before. To the
washed basic aqueous extract we added 1 mL of 3 molfL HCI
and 4 mL of hexane. The contents of the tube were vortexmixed
and centrifuged as before, and the organic layer was
separated and evaporated under a stream of nitrogen at 60#{176}C
in a Reacti-vial (Pierce Chemical Co., Rockford, IL 61105).
Methylation was performed by adding 500 zL of acetone,
100 L of iodomethane, and about 5 mg of sodium carbonate
to the dried hexane extract, then incubating the mixture in
a vortex evaporator at 60 #{176}C for 30 mm. The supernate then
was evaporated as before and reconstituted into 20 L of ethyl
acetate by vortex-mixing; 2 tL was injected into the gaschromatographic
column. Retention times for thiopental and
phenoharbital are 1.7 and 1.3 mm, respectively. The standard
curves were constructed by plotting thiopental/phenobarbital
peak height ratios vs thiopental concentrations.
For standard curves for higher thiopental concentrations,
200 L of plasma containing 0, 0.5, 1, 2.5, 5, and 10 mg of
thiopental per liter was added to 200 tL of 1.5 mol/L NaH2-
P04 and 200 zL of 2-propanol, and the mixture was extracted
with 1 mL of hexane containing 8 mg of phenobarbital per
liter. The extraction procedure at these higher plasma thiopental
concentrations was performed as for the lower concentrations,
except that the basic aqueous extract was not
washed with hexane. Methylation was done as before.
Analytical
Variables
Reproducibility. The reproducibility of the method was
determined from results for repeated injections. Concentrations
of 50 and 500 g/L in plasma for the low-end standard
curve and 1 and 10 mgfL for the high-end standard curve were
prepared and each solution was injected six times into the
chromatograph.
Extraction efficiency. We determined the efficiency of the
extraction of thiopental from plasma by adding known
amounts of thiopental, equivalent to concentrations of 50 and
500 g/L in plasma, to the hexane layers after extraction of
drug-free plasma as previously described. The concentrations
of thiopental in these samples were compared with the same
concentrations of thiopental that had been added directly to
plasma and extracted as previously described.
Stability. Concentrations of 50 and 500 ig of thiopental per
liter were prepared as described above, each in 10 mL of
plasma. One-milliliter or 200-tL samples from each concentration
were placed into separate glass tubes and stored at -20
#{176}C. Analyses were performed once a week for eight weeks. On
each day of analysis a standard curve in plasma was constructed
and the thiopental concentrations of the stored
samples were determined by comparing their thiopental/
phenobarbital peak height ratios with those of the standard
curve.
Data
Analysis
With all standard curves, a straight-line fit of the data was
performed by least-squares linear regression analysis. All results
are expressed as mean ± standard deviation (SD).
Results and Discussion
The use of a nitrogen-specific detector in analyses for barbiturates
has been described by many workers. Such studies
show that these compounds are not readily gas-chromatographed
without prior derivatization, especially when nano-
lS
I IS I
T
HH1HI
I I , I r I I , , I
2 1 02 1 0 2 1 021 0
A
B
C 0
Fig. 1. Gas-chromatographictracingsobtainedfor human
plasma
A: Extractof1 mL ofdrug-free plasma with phenobarbital (IS) as internal Standard.
B: Extractof 1 mL of plasma with 50 ig of thiopental (T) added perliter.
C: Extract of 200 zL of drug-free plasma with phenobarbital (IS). D: Extract of
200 iL of plasma supplemented with 500 .sg of thiopental (7) per liter
gram amounts are injected into the columns. In addition,
chromatography of the free drug does not take full advantage
of the sensitivity of a nitrogen-phosphorus detector. When
barbiturates are methylated by flash-heater methylating
techniques with trimethylanilinium hydroxide or tetramethylammonium
hydroxide, the nitrogen detector detects
large peaks in the chromatogram, caused by the derivatizing
agents at the sensitivities required for the analysis. Dunges
and Bergheim-Irps (13) reported a method in which an alkyl
iodide and sodium carbonate in a polar nonaqueous solvent
system were used to prepare the alkyl derivative. This procedure
allowed better control of reaction conditions and
avoided pyrolytic formation of trialkylamines, thus permitting
measurement of thiopental concentrations in the microgram
per liter range.
Figure 1 depicts representative chromatograms for the high
(500 g/L) and low (50 tg/L) concentrations. The methylated
oxy analog of thiopental, N,N-dimethylpentoharbital, elutes
with the solvent front in our system.
The standard curves constructed for the peak height ratios
of thiopental to internal standard were all linear and reproducible.
Seven standard curves for thiopental in plasma with
concentrations of 0, 25, 50, 100, 250, and 500 g/L at the low
concentrations and 0, 0.5, 1, 2.5, 5, and 10 mg/L at the high
concentrations were made at different times during two weeks.
The slopes of the plot of peak height ratio of thiopental to
internal standard vs thiopental concentration in plasma for
the 42 points at the low and high concentrations were 0.00585
± 0.00021 L/g and 0.523 ± 0.013 L/mg, respectively. The
coefficients of variation (CV) of the slopes were 3.6 and 2.5%,
Is
T
IS
114 CLINICAL CHEMISTRY, Vol. 27, No. 1, 1981

20
I0
5.
11.I0.5
here is necessary to characterize completelythe pharmacokinetics
of thiopental after the usual 3-4 mg/kg induction
doses. An illustration of the latter application is shown in
Figure 2, a plot of thiopental serum concentrations as a
function of time after the intravenous administration of a 4.8
mg/kg dose of thiopental sodium to one subject. The assay
sensitivity permitted characterization of the concentrationtime
course for two days; the data were best described by a
triexponential function with a terminal elimination half-life
of 8.6 h.
Supported in part by a grant from the Office of the Provost for
Graduate Studies and Health Sciences, University of Arizona.
0.1
0.05
N
N,
) 10 20 30 40 50
Time,h
Fig. 2. Thiopental concentration in serum as a function of time
in a subject who received an intravenous thiopental sodium dose
of 4.8 mg/kg.
The solid line represents
a non-linearregression fit of the data
and correlation coefficients of 0.98 and 0.99 were obtained for
the low-end and high-end standard curves, respectively. The
corresponding intercepts were -0.090 (SD 0.049) tg/L and
-0.168 (SD 0.062) mg/L.
The CVs for thiopental in plasma for the low-end standard
curve were 6.0% (51.7 ± 3.1 tg/L) and 3.9% (507.1 ± 19.8
tg/L), respectively. At the high concentrations, the coefficients
of variation for thiopental were 10.5% (1.05 ± 0.11
mg/L) and 3.4% (10.14 ± 0.34 mg/L), respectively.
Analytical recovery of thiopental from plasma averaged 89.9
(SD 6.6)% and 82.8 (SD 2.2)% at 50 and 500 sg/L, respectively.
No detectable deterioration in thiopental was observed over
an eight-week study period. The 50 and 500 g/L solutions
yielded concentrations of 49.9 (SD 4.6) tg/L and 497.5 (SD
29.2) zg/L, respectively.
Diazepam, fentanyl, and morphine-drugs frequently used
in conjunction with thiopental-were tested for potential
interference with the thiopental assay. These drugs are bases
and are not extracted in our procedure, so, as expected, we saw
no interfering peaks at the retention times of thiopental or its
internal standard, phenobarbital.
Our method is more sensitive and specific than other published
methods. The 25 tg/L sensitivity of the assay described
References
I. Oroszlan, S. I., and Maengwyn-Davies, G. D., Ultraviolet photometric
assay of thiopental and pentobarbital in blood and plasma. J.
Am. Pharm. Assoc., Sci. Ed. 49,507-509(1960).
2. Dayton, P. G., Perel, J. M., Landrau, M. A., et al., The relationship
between binding of thiopental to plasma and its distribution into
adipose tissue in man, as measured by a spectrophotofluorimetric
method. Riochem. Pharmacol. 16, 2321-2336 (1967).
3. Blackman, C. L., and Jordan, C. J., Analysis of thiopentone in
plasma by high-performance liquidchromatography. J. Chromatogr.
145, 492-495 (1978).
4. Masoud, A. N., Scratchley,G. A., Stohs, S. J., and Wingard, D. W.,
Simultaneous determination of lidocaine (lignocaine) and thiopental
in plasma using high pressure liquid chromatography. J. Liq. Chromat
ogr. 1,607-616 (1978).
5. Braddock, L. I., and Marec, N., Gas chromatographic analysis of
submicrogram quantities of barbiturates using a flame ionization
detector. J. Gas Chromatogr. 3,274-277 (1965).
Ii. l)unges, W., Bergheim-Irps, E., Straub, H., and Kaiser, R. E.,
Microtechniques for the gas chromatographic determination of barbiturates
in small blood samples. J. Chromatogr. l45, 265-274
(1978).
7. Greely, R. H., New approach to derivatization and gas-chromatographic
analysis of barbiturates. (‘un. Chem. 20, 192-194 (1974).
8. MacGee, J., Rapid identification and quantitative determination
of barbiturates and glutethimide in blood by gas-liquid chromatography.
(‘ha. (‘hem. 17, 587-591 (1971).
9. Fiereck, E. A., and Tietz, N. W., A gas-chromatographic method
b>r separating and measuring barbiturates and glutethimide in blood.
(‘un. (‘hem. l7, 1024-1027 (1971).
10. Becker, K. E., Jr., Gas chromatographic assay for free and total
plasma levels of thiopental. Anesthesiology 45, 656-660 (1976).
II. Van Hamme, M. J., and Ghoneim, M. M ., A sensitive gas chromatographic
assay for thiopentone in plasma. Br. J. Anaesth. .50,
143-145 (1978).
12. Sennello,L. T., and Kohn, F. B., Gas chromatographic determination
of thiopental in plasma using an alkali flame ionization detector.
Anal. Chem. 46, 752-755 (1974).
13. Dunges, W., and Bergheim-Irps, E., A new methylation method
for the gas chromatography of barbituric acids. Anal. Lett. 6, 185-195
41973).
CLINICALCHEMISTRY, Vol.27,No. 1,1981 115