Journal of Analytical Toxicology, Vol. 22, September 1998
A New Specific Method to Detect Cyanide in Body
Fluids,EspeciallyWhole Blood,by Fluorimetry
D. Felscher* and M. Wulfmeyer
Institut f~r Rechtsmedizin des Universit~tsklinikums "Cad Gustav Carus" der Technischen Universit~t Dresden, Fetscherstrage74,
01307 Dresden, Germany
[Abstract
[
This study shows a simple, rapid, and specific method for the
quantitative determination of cyanide ion in body fluids, especially
blood, by fluorimetry. It is based upon the transformation of
cyanide ion into hydrocyanic acid, which then reacts with
2,3-naphthalenedlaldehyde and taurine in a self-contained system.
The 1-cyano-2-benzoisoindole derivate thus formed is suitable for
fluorimetric measurement (~EX= 418 nm; ~EM= 460 nm). The
fluorescence intensity can be determined by spectrophotometry
or by high-performance liquid chromatography (HPLC) with
fluorescence detection. The detection limit is 0.002 pg/mL.
Linearity was excellent from 0.002 to 1 pg/ml, for
spectrophotometry and from 0.002 to 5 pg/mL for HPLC
with fluorescence detection. The coefficient of variation for
repeatability was 8% or less. Thiocyanate and sulfide did not
interfere, even at high concentrations (200 pg/mL). The method
was applicable to whole blood, so it should be suitable for both
clinical and forensic purposes.
graphic methods such as high-performance liquid chromatography (HPLC)with fluorimetric detection (17,18) and gas chromatography (GC) with nitrogen-phosphorus (NPD) (10,1820), thermionic (21-24), or electron-capture detection (ECD)
(25,26) were used.
The method suggested here is based upon the transformation
of cyanide by acidificationinto hydrocyanic acid, followedby its
reaction with 2,3-naphthalenedialdehyde and taurine (17) in a
self-contained system. In contrast to Sano et al. (17), the preparation of the samples can take place in a single step because the
release of the cyanide (in 20-mL headspace vials) and the subsequent derivatization to a 1-cyano-2-benzoisoindole compound (in a 2-mL tube) is realized by a spatial separation.
The cyanide concentration is determined by HPLCwith fluorimetric detection or by direct measurement of the fluorescence with a fluorescence spectrometer. The method is suitable
for a rapid and exact determination of cyanide in body fluids,
and its employment for the diagnosis of a cyanide poisoning is
possible in a toxicology emergency.
Introduction
Material and Methods
Cyanidepoisonings are rather common because humans are
exposed to cyanide from the consumption of plants such as almonds or cassava that contain cyanide (1), from voluntary ingestion of acetonitrile and alkaline cyanide (2,3), and from
exposure and accidental inhalation of toxic gases in several
branches of industry (4) or in fires (5,6). Over the past few
years, hydrogen cyanide has been investigated as the second
most important gas forensically, the most forensically important gas being carbon monoxide.
In all these cases the blood determination of cyanide for an
exact diagnosis is important in a concentration ranging from
0.1 to 5 lag/mL.The followingmethods have been employedfor
this purpose: spectrophotometry (9,10), fluorimetry (11-14),
electrochemistry (15,16), and chromatography. Chromato* Author to whom correspondenceshould be addressed.
Reagents
All chemicals were analytical reagent grade. Solutions were
prepared with deionizedwater. A 10raM stock standard solution
of cyanide was prepared with potassium cyanide (Merck,Darmstadt, Germany) in 0.1M sodium hydroxide.
The concentration of the stock standard solution was calibrated by silver nitrate according to Liebig's method (27).
Working standard solutions were prepared by dilution of the
stock solution with water. A phosphate buffer (pH 8.0) was prepared by mixing 969 mL of a sodium hydrogen phosphate
solution (11.878 g/L) and 31 mL of a potassium dihydrogen
phosphate solution (9.073 g/L). A stock 4mM solution of 2,3naphthalenedialdehyde (NDA,Aldrich-Chemie,Steinheim, Germany) was prepared in methanol. This solution was diluted
with the phosphate buffer to give a lmM solution and stored in
Reproduction (photocopyin 8) of editorial content of this journal is prohibited without publisher's permission.
363
Journal of Analytical Toxicology, Vol. 22, September 1998
an amber glass bottle at 4~ The diluted NDAsolution could be
used for at least seven days. Taurine (Aldrich-Chemie, Steinheim, Germany) was used as a 5raM solution in the phosphate
buffer. The taurine solution could be used for at least four
weeks when stored at 4~
Equipment
HPLC method. The HPLC system was a Hewlett-Packard
model HP]090 equipped with a fluorescence detector model
F1080 (~.Ex= 418 nm; ~EM= 460 nm; Merck, Darmstadt, Germany) and a 5-tJm Hypersil ODS RP18 column (100 mm • 2.1mm i.d., Hewlett-Packard, Waldbronn, Germany). The column
temperature was ambient, and the flow rate was 0.4 mL/min.
Results were processed using a Merck-Hitachi D2500 reporting
integrator.
Fluorescence spectroscopic method. A fluorescence spectrometer Perkin-Elmer LS 50B with a 0.3-mL precision quartz
glass cuvette Suprasil (pathlength 0.5 cm) was used, operating
at 418-nm excitation and 460-nm emission. The attenuation
was calculated as 9.
Changes in incubation time (from 15 to 180 min) and temperature (from 20 to 120~ provided no improvement. At
room temperature, the intensity of the fluorescence began to
fail 120 min after the end of the reaction time.
Possible interferences of thiocyanate (up to 100 pg/mL) and
sulfide concentrations (up to 300 pg/mL) were studied. Samples
of deionized water spiked with various concentrations of
cyanide (0.05, 0.1 and, 0.5 pg/mL) and of thiocyanate or sulfide
were studied. The fluorescence intensities determined with and
without thiocyanate and sulfide were identical, indicating that
thiocyanate and sulfide did not interfere (Table II).
The determinations were repeated using whole bloodsamples
(with heparin as anticoagulant) spiked with cyanide. The reproducibility of curve I (0.25 pg/mL cyanide in whole blood) is
characterized by a coefficient of variation of 6%. For curve 2
(0.5 lag/mL cyanide in whole blood), we found a coefficient of
variation of 8%. The recoveries were about 100% for the prepared whole blood samples. Whole blood samples and samples
of deionized water gave identical standard curves.
HPLC method
Procedure
The retention time of the 1-cyano-2-substituted benzoisoA 2-inK sample (standard or heart blood) was pipetted into
indole derivate was about 1.7 min, and the resulting peak was
20-mL headspace vials (Perkin-Elmer). Sulfuric
acid (10% w/w; 0.5 mL) containing 200 mg
NaHSO4 was added.
Ill
Special tubes (2 mL) containing 200 pL each of
5~
diluted NDA solution and taurine solution (5mM)
were immediately placed into the headspace vials
with a pair of tweezers to avoid evaporation of the
HCN in the process of development. The 20-mL
~--~
Blood[CN ]-+ H~S0,-'-* HCN ~ [CN]headspace vials were stoppered using a Teflon-faced
NDA + Taurine
9' I I
1
butyl rubber septum and an aluminum crimp-top
cap (Figure 1). The sample was vortex mixed for 5 s.
Sealed vials were kept for 120 rain at 35~ to
Q
2h
~-benzoisoindol~
permit any CN- to defuse into the NDAand taurine.
Figure 1. Sample preparation.
Results
Table I. Calibration Curves for Determination of Cyanide by Fluorimetry
Fluorimetric method
Calibration curves were made using deionized
water spiked with various concentrations of cyanide
(curve 1: range of 0 to 0.5 lJg/mL in steps of 0.05
lag/mL; curve 2: range of 0 to 1.0 lag/mL in steps of
0.1 lag/mL). Table I shows the concentration
ranges, the equations, and the correlation coefficients (n = 10) for both determinations, curve I and
curve 2. Linearity was excellent in both ranges.
The results of the calibration were compared with
two reference materials with argentometrically
determinated cyanide concentrations of 0.25 and
0.75 lag/mL in deionized water.
The coefficient of variation (CV)for repeatability
was less than 8% (n = 10). The recovery (0.25
tag/mL in curve 1 and 0.5 pg/mL in curve 2) was
between 80 and 90%.
364
Concentration
ranges (pg/mL)
Slit width
(nm)
Attenuation
Curve 1: 0.002-0.500
Curve 2:0.002-1.000
"r0/9
10/9
9
9
Equation
Correlation
coefficient(r)
(n = 10)
y = 1036x + 8.4556
y = 1064x + 1.2
0.999
0.999
Table II. Fluorescence Intensity of Various Cyanide Concentrations
Determined With and Without Thiocyanate and Sulfide
Cyanide Relativefluorescence Relativefluorescence Relativefluorescence
concentrations intensityof the
intensity of the
intensityof the cyanide(pg/mL)
cyanidesample(9) thiocyanatesample (9)
sulfidesample(9)
0.05
0.1
0.5
1.00
1.00
1.00
1.01
0.99
0.99
0.99
1.00
0.98
Journal of Analytical Toxicology, Vol. 22, September1998
uniform and symmetrical with no tailing (Figure 2).
Calibration curves were made using normal blood and water
spiked with various concentrations of cyanide. Three calibration
curves where made for defined ranges of cyanide concentrations: high (curve 3, 0 to 5.00 IJg/mL cyanide), beginning toxic
(curve 4, 0 to 1.00 IJg/mL cyanide), and physiological levels of
cyanide (curve 5, 0 to 0.05 IJg/mL cyanide). Injection volume
and filter factors were different in the three ranges. Each of the
calibration curves was linear in the defined range under these
conditions (Table III).
The intra-assay CVwas 2--8% (n = 10). For curve 3, there is
a CVof 4.1% at 0.025 IJg CN-/mLwith a standard deviation (SD)
of 0.001 IJg CN-/mL.In the case of curve 4, there is a CV of 3.1%
at 0.5 I~gCN-/mL with an SD of 0.015 IJg CN-/mL, and in curve
5, there is a CV of 2.0% at 2.50 IJg CN-/mL with an
SD of 0.05 IJg CN-/mL. The results of the calibra0
1oo
tion were compared with reference solutions with
argentometrically determinated cyanide concentrations of 0.025, 0.500, and 2.5 IJg/mLin deionized
SO
water. The coefficient of variation for repeatability
was between 2 and 8%.
SO
Specially prepared samples were determined by
HPLC because the retention time of various spikes
40
in HPLC determination makes a differentiation of
various fluorescending substances possible. Therefore, samples of deionized water spiked with various
20concentrations of sodium thiocyanate, thiosulfate,
sulfide, sulfite, potassium hexacyanoferrate-(II),
and hexacyanoferrate-(III) were prepared. Samples
. . . .
0:5
"
t
. . . .
115 "
;~
215 "
"
with concentrations of 50 and 100 Iag/mL each of
Time (min)
the substances were analyzed using HPLC. We difFigure 2. Chromatogram of the 1-cyano-2-benzoisoindolederivative (retention time, 1.764
ferentiated these determinations in the following
rain). The flow ratewas 0.4 mL/min, deionized water 92.0%/acetonitrile 8.0%.
manner: samples were prepared as mentioned previously, but substances were added directly into
the special tubes (2 mL) containing 200 IJL each of
Table III. Calibration Curves for Determination of Cyanide by HPLC
diluted NDA solution and taurin solution (5mM).
Injection Correlation
Results can be summarized as follows:neither conConcentration
Filter
volume coefficient(r)
centration of 50 nor of 100 IJg/mL each of the derange (pg/mL)
Equation
factor
(pl)
(n = 10)
terminated anions produced any fluorescence when
Curve 3:0.002-5.000
y = 210.57x + 0.319
4
20
0.998
samples were prepared using our method. Analyzed
Curve 4:0.002-1.000
y= 5.7818x + 0.4455
6
10
0.996
after direct addition to NDA solution and Taurin
Curve 5:0.002-0.050
y = 4.2971 x + 1.3905
8
3
o.990
solution, only S- and SO32- (100 IJg/mL) showed a
peak in the chromatogram. The respective retention times for sulfite and sulfide were about 1.9
Table IV. Reaction of Several Anions with 2,3-Naphthalenedialdehyde
and 2.0 min (Table IV).
and Taurine
These HPLC results showed that a direct determination of cyanide by fluorimetry is possible beConcentration Retentiontime Relative
Fluorescending
Anion
(pg/mL)
(rain)
peak area
substance
cause the fluorescence intensity is not influenced
by thiocyanate, thiosulfate, sulfide, sulfite, hexaCyanide
0.1
1.764
1
1-cyano-2-benzoisocyanoferrate-(II), and hexacyanoferrate-(III).
indole derivate
The accuracy of both method was compared using
Thiocyanate
50
cyanide samples (0.1 IJg/mL). The results of five
100
samples each determined by direct fluorimetry and
Thiosulfate
50
HPLC were analyzed by F-test using MS Excel 5.
100
The results can be summarized as follows: with a
Sulfide
50
likelihood of error at p = 5%, the differences of the
0.3
Sulfide derivate*
100
2,0'
analyzed cyanide concentrations are not significant;
Sulfite
50
the accuracy of both methods is excellent.
100
1.9'
0.2
Sulfite derivate*
0
...........
TM
Hexacyanoferrate-(ll) 50
100
Discussion
Hexacyanoferrate-(lll) 50
100
* Detected
only
after direct addition to
NDA
and taurine
solution.
Because the concentration of cyanide in red
blood cells is 98% and only 2% in plasma (28),
365
Journal of Analytical Toxicology,Vol. 22, September1998
whole blood, which should not be coagulated, was used for the
determination. The cyanide half-life is about 1 h, so the sample
should be taken and analyzed as quickly as possible. Heart
blood is usually used for the determination of carbon monoxide;
therefore, we also used the same sample material for cyanide
determination. The described procedure is convenient because
the reactions are carried out in a single step. The determination
of the cyanide concentration can be realized either by measurement of the fluorescence intensity after HPLC or directly by
fluorescence spectroscopic method. The limit of detection (peak
area or value of fluorescence intensity three times higher than
background level) was found to be 0.002 IJg cyanide/mL.
Conclusion
The aim of this study was to develop a new procedure for the
fluorimetric detection of cyanide in whole blood. The preparation of the sample before analysis is simple and rapid. The
method is specific for cyanide and does not show any interference from thiocyanate and sulfide. The sample material for the
forensic determination of cyanide was heart blood, but samples
of venous blood can also be analyzed for the determination in
emergency toxicology.
References
1. M.J. Ellenhorn and D.G. Barceloux. Cyanide. In Medical Toxicology, Diagnosis and Treatment of Human Poisoning. Elsevier,
New York, NY, 1988, pp 829-835.
2. A. Pasi, M. Morath, and H. Hartmann. Die Cyanvergiftung: Forensisch-toxiko[ogische Beobachtungen bei der Untersuchung von 54
F~llen todlicher Intoxikationen, Z. Rechtsmed. 95:35-43 (1985),
3, R. Hartung, Cyanides and nitriles, In Party's Industrial Hygiene and
Toxicology, 3rd ed., vol. 2C, G.D. Clayton and F.E. Clayton, Eds.
Wiley Interscience, New York, NY, 1982, pp 4845-4900.
4. K.H. Ricken. Blaus~urevergiftung~in arbeitsmedizinischer Notfall. Zbl. Arbeitsmed. 29:212-215 (1979).
5. L. yon Meyer, G. Drasch, and G. Kauert. Zur Bedeutung der Bildung yon Cyanwasserstoff bei Br~nden. Z. Rechtsmed. 84" 69-73
(1979).
6. H. KOhler-Schmidt, G. Bohn, and F. Fechner. Nachweis der kombinierten Blausaure-Kohlenmonoxid-lntoxikation bei Zimmerbr~nden. Beitr. Gerichtl. Med. 46:243-245 (1986).
7. M. Daunderer. Lethal smoke poisoning with cyanides from slow
burns. Fortschr. Med. 38:81-96 (1979).
8. 1-. Sigrist and R. Dirnhofer. Zur Entstehung der kombinierten inhalatorischen Blausaure-Kohlenmonoxid-Vergiftung. Arch. Krirninol. 163:145-59 (1979).
9. P. Lundquist and B. Sorbo. Rapid cyanide determination of toxic
cyanide concentration in blood. Clin. Chem. 35:617-19 (1989).
10. Z.L. Cardeal, J.R Gallet, A. Astier, and D. Pradeau. Cyanide assay:
statistical comparison of a new gas chromatographic calibration
method versus the classical spectrophotometric method. J. AnaL
ToxicoL 19:31-34 (1995).
366
11. A. Ganleloo, G.E. Isom, R.L. Morgan, and J.L. Way. Fluorometric
determination of cyanide in biological fluids with p-benzoquinone. Toxicol. Appl. Pharmacol. 55' 103-107 (1980).
12. O. Suzuki, H. Hattori, M. Oya, and Y. Katsumata. Direct fluorimetric determination of cyanide in human materials. Forensic Sci.
Int. 19:189-195 (1982).
13. P. Lundquist, H. Rosling, and B. S~rbo. Determination of cyanide
in whole blood, erythrocytes, and plasma. Clin. Chem. 31:
591-595 (1985).
14. P. Lundquist, H. Rosling, B. SOrbo, and L. Tibbling. Cyanide concentrations in blood after cigarette smoking, as determined by a
sensitive fiuorimetric method. Clin. Chem. 33:1228-1230 (1987).
15, J.O. Egekeze and FW. Oehme. Direct potentiometric method for
the determination of cyanide in biological materials. J. Anal. Toxicol. 3:119-124 (1979),
16. J.L. Way. Cyanide intoxication and its mechanism of antagonism.
Ann. Rev. Pharmacot. ToxicoL 24:451-481 (1984).
17. A. Sano, M. Takezawa, and S. Takitani. Spectrofluorirnetric determination of cyanide in blood and urine with naphthalene-2,3-dialdehyde and taurine. Anal. Chim. Acta 225:351-358 (1989).
18. A. Sano, N. Takimoto, and S. Takitani. High-performance liquid
chromatographic determination of cyanide in human red blood
cells by precolumn fluorescence derivatization. J. Chromatogr.
582:131-135 (1992).
19. F. McAuley and D.S. Reive. Rapid quantitation of cyanide in
blood by gas chromatography. J. Anal. Toxicol. 7:213-215 (I 983).
20. J. Zamecnik and J. Tam. Cyanide in blood by gas chromatography with NP detector and acetonitrile as internal standard. Application on air accident fire victims. J. Anal. Toxicol. 11:47-48
(1987).
21. R.W, Darr, T.L. Capson, and F.D. Hileman. Determination of hydrogen cyanide in blood using gas chromatography with alkali
thermionic detection. Anal. Chem. 52:1379-I 381 (1980).
22. K, Funazo, M, Tanaka, and T. Shono. Methylation of inorganic anions for gaschromatographic determination. J. Chromatogr. 211:
361-368 (I 981 ).
23, K. Funazo, K. Kusano, H. Wu, M. Tanaka, and T. Shono. Trace determination of cyanide by derivatization and flame thermoionic
gas chromatography. J. Chromatogr. 245: 93-I O0 (I 982).
24. D. Felscher and H. Lowe. Beitrage zur gaschromatographischen
Bestimmung von toxikoJogisch re[evanten Cyanid- und Thiocyanatkonzentrationen mitte[s Dampfraumanalyse. Krim. Forensische Wiss. 75:165-166 (1990).
25. M. Odoul, B. Fouillet, B. Nouri, R. Chambon, and P. Chambon.
Specific determination of cyanide in blood by headspace gas
chromatography. J. Anal. Toxicol. 18:205-207 (1994).
26. S.-H. Chen, S.-M. Wu, H.-S. Kou, and H.-L. Wu. Electon-capture
gas chromatographic determination of cyanide, iodide, nitrite,
sulfide, and thiocyanate anions by phase-transfer-catalyzed derivatization with pentafluorobenzyl bromide. J. Anal. Toxicol. 18"
81-85 (1994).
27. L. Medicus, bearbeitet von W. Poethke. KurzeAnleitung zur Massanalyse. T. Steinkopf, Dresden, Germany, 1952, pp 279-281.
28. A. van Heijst and M. Geldmacher-von Mallinckrodt. Klinische
Interpretation der Cyanidkonzentration im Blut. In KlinischToxikologische Analytik bei akuten Vergiftungen und Drogenmifgbrauch, H.J. Gibitz and M. Geldmacher-von Mallinckrodt,
Eds., VCH Verlagsgesellschaft, Weinheim, Germany, 1989, pp
196-198.
Manuscript received July 16, 1997;
revision received December 4, ] 997.