Determination of Paracetamol (Acetaminophen)

by HPLCwith Post-ColumnEnzymatic
Derivatization and FluorescenceDetection

2001, 54, 163-167

J. Meyer/U. Karst*
Anorganisch-Chemisches Institut,Westf0hsche Wilhelms-U niversit0t MO nster,Wilhelm-Klemm-Str. 8, 48149 MOnster, Germany;
E-Maih uwe.karst@u ni-m uenster.cle

Dedicated to Professor Dr. Heinz Engelharclt on the occasion of his 65 th birthday.

trokinetic capillary chromatography

KeyWords (MEKC) with UV detection [7], capillary
electrophoresis with UV, MS [8], and
Column liquid chromatography N M R detection [9], and HPLC with
Post-colu m n derivatization photochemical post-column derivatiza-
Fluorescence detection tion and UV detection [10]. Recent HPLC
Paracetamol methods with UV detection have used de-
Urine tection wavelengths of 254 [11], 248 [12],
and 240 nm [ 13]. The concentration range
for calibration of these methods starts at
Summary 5 mg L 1, in accordance with the low sen-
sitivity mentioned above, because of the
A novel method is described for the determination of paracetamol (acetaminophen; N-ace@- unfavourable spectral properties of para-
4-aminophenol) in urine. After reversed-phase HPLC separation, paracetamol is oxidized by cetamol. Long retention times (approx.
H202 with horseradish peroxiclase catalysis. Detection is performed fluorimetrically at an exci- 19 min) are also necessary, to prevent coe-
tation wavelength of 329 nm and an emission wavelength of 435 nm. Urine samples were lution of matrix constituents [13]. For
spiked with paracetamol, diluted, and injected directly without further pretreatment. Under some applications, immunoassay (e. g. the
these conditions, the limit of detection was 2 x 10-s mol L 1, and the limit of quantification TDx Assay) must be used to achieve ap-
was 7 x 10-s mol L 1. The method was validated by l",,vodifferent approaches based on HPLC propriate selectivity and sensitivity [14].
with UV-Vis detection. In this paper, we report on a new meth-
od based on HPLC separation coupled
with enzymatically catalysed post-column
oxidation of the analyte, the reaction pro-
Introduction paracetamol [1 5], convenient proce- duct of which is then detected fluorimetri-
dures based on the oxidation of paraceta- cally. Analogous chemical principles were
Paracetamol (acetaminophen; N-acetyl-4- mol and subsequent photometric or used previously for the determination of
aminophenol) is a popular antipyretic and fluorimetric detection of one of the pro- atmospheric peroxides by oxidative cou-
analgesic drug in widespread use world- ducts are only applicable for less-demand- pling of phenolic peroxidase substrates
wide. It is known to undergo a variety of ing matrices containing high concentra- [15] the peroxides were separated by
oxidation reactions under the action of a tions of the drug. Human urine or plasma HPLC and determined by post-column
wide range of oxidants including iodoso- samples cannot be analysed by these tech- derivatization (PCD) with peroxidase as
benzene [1], potassium hexacyanoferra- niques. HPLC separation can afford the catalyst and p-hydroxyphenylacetic acid
te(III) [2, 3], complexes of iron(III) with desired selectivity, but the intrinsic spec- (pHPA) as phenolic substrate [16]. In this
1,10-phenanthroline [4], and H202 in the tral properties of the analyte (maximum latter example the oxidant (co-substrate
presence of horseradish peroxidase [5, 6]. absorbance below 250 nm) do not enable of the enzyme) is quantified, whereas in
The peroxidase-catalysed reaction yields a high sensitivity and selectivity, and coelu- our modified scheme the phenol (sub-
mixture of paracetamol oligomers [5], tion of typical matrix constituents with strate of the peroxidase) is the analyte.
some of which are fluorescent [6]. the analyte is likely.
Although all these oxidation reactions Recent developments in the field of se-
have been used for the determination of paration techniques include micellar elec-

Original Chromatographia 2001, 54, August (No. 3/4) 163

0009-5893/00/02 163- 05 $ 03.00/0 9 2001 Friedr. Vieweg & Sohn Verlagsgesellschaft mbH
Table I. Acetonitrile-water binary gradient profile used with system A. The flow rate was 0.7 mL many) protected by a guard column 8 mm
min 1. long but otherwise identical. The mobile
Time (min) 0 5 7 8.5 9 10 (stop) phase was a mixture of deionized water
Concentration of CH3CN (%, v/v) 20 20 100 100 20 20 (400 mL), phosphoric acid (85% aqueous,
40 ixL), and acetonitrile (80 mL). Elution
was performed isocratically at a flow rate
of 1 mL min 1.

~
Injection
./valve The injection volume was 20 ixL for all
chromatographic measurements. The wa-
Computer velength for UV-Vis spectrometric detec-
Analytical .~ tion was set to 254 nm.
column t
Sample Treatment
UV/vis detector Fluorescence ] Tablets 'Dolomo TN' from Klinge Phar-
detector
ma (Munich, Germany) were used for
LJ~ ~ Mixing T = 35 ~ measurements on spiked urine samples.
./tee One tablet with a stated content of 250 mg
paracetamol was dissolved in 50 mL deio-
Oven nized water in an ultrasonic bath. The re-
sulting turbid mixture was filtered and di-
Figure 1. Schematicdiagramof chromatographicsystemA withpost-columnderivatization. luted to 100 mL in a volumetric flask. This
solution was diluted by factors of 103 and
Experimental flow cell and focussing optics for flow- 104 and these solutions were used to pre-
through measurements obtained from pare (i) a mixture of diluted tablet solution
Chemicals Polytec (Waldbronn, Germany). Com- (500 ixL), human urine (50 ixL), and deio-
pounds were separated on a 150mm x nized water (450 ixL), and (ii) a mixture of
Paracetamol, ammonium chloride (ACS 4.6 mm i. d., 5-1xmparticle, 100-A pore size diluted tablet solution (500 ixL) and deio-
reagent), N-[2-hydroxyethyl]piperazine- Discovery ClS column (Supelco, Deisen- nized water (500 ixL). Thus two solutions
N'-[2-ethanesulphonic acid] (HEPES) so- hofen, Germany) protected by a guard were obtained with an expected concentra-
dium salt (99.5%), microperoxidase (MP- column 20 mm long but otherwise identi- tion of 1.25 mg L 1 (8.3 ixM)paracetamol,
11) sodium salt (90%), and horseradish cal. The profile of the gradient used with two others with an expected concentration
peroxidase (HRP) [EC 1.11.1.7] (116 units system A is shown in Table I. of 125 ixg L 1 (0.83 ixM). This would cor-
mg 1, RZ 1.3) were obtained from Sigma The post-column instrumentation used respond to unspiked urine samples with a
(Deisenhofen, Germany). Hydrogen per- with system A was a Knauer (Berlin, Ger- respective concentration of 25 mg L 1 or
oxide (35%) was from Aldrich (Steinheim, many) 64 HPLC pump (pump 1), a Sie- 2.5 mg L 1.
Germany). Acetonitrile (gradient grade) mens (Karlsruhe, Germany) prototype A patient will suffer from paracetamol
was purchased from SDS (Peypin, HPLC pump (pump 2), two mixing tees intoxication if the concentration in his
France). Ammonia solution (25%, analy- from Upchurch (Oak Harbor, WA, USA), blood is > 2 0 0 m g L 1. Approximately
tical grade) and phosphoric acid (85%, an 18-m knitted Teflon capillary (pre- 10%o of this concentration will be found
purum DAB) were supplied by Grfissing pared as described elsewhere [17]) with an unchanged in the urine [18]. In our experi-
(Filsum, Germany). Methanol (gradient inner diameter of 0.3 mm, and a non-com- ments, therefore, the high concentration
grade) and acetic acid (micro select) were mercial oven to heat the reaction capillary spiked urine sample simulates a real sam-
from Fluka (Neu-Ulm, Germany). to 35~ HRP (10mgL 1) in NH3- ple from a person with paracetamol intox-
NH4C1 buffer (pH 9.5, 0.1 M) was deliv- ication, whereas the low-concentration
ered at a flow rate of 0.4mLmin 1 by sample is a factor of 10 more dilute than
Instrumentation pump 1 and H202 (10mM) was delivered that.
at a flow rate of 0.05mLmin 1 by pump
Fluorescence spectra were recorded with a 2. A schematic diagram of the arrange-
Shimadzu (Duisburg, Germany) R F ment of the chromatographic system is Results
5301-PC spectrofluorophotometer. presented in Figure 1.
Chromatography was performed with System B comprised an LC-10AS Two different methods were used for de-
two systems. System A comprised two pump, SPD-M10Avp diode array detec- termination of paracetamol. The first used
LC-10AS pumps, SPD-10AVvp UV-Vis- tor, SIL-10A autosampler, and CBM-10A chromatographic system A and detection
detector equipped with a high-pressure controller unit with Class LC-10 Version was performed photometrically (before
flow cell, and CBM-10A controller unit 1.6 software, all from Shimadzu. Com- PCD) and fluorimetrically (after PCD).
with software Class LC-10 Version 1.6, all pounds were separated on a 70mm x This was a method newly developed. The
from Shimadzu (Duisburg, Germany), 3 mm i.d. ChromCart cartridge packed second uses chromatographic system B,
and an Aminco Bowman AB2 spectro- with 5-1xm particle, 100-A pore size Nu- with photometric evaluation only. This
fluorophotometer equipped with a 8-1xL cleosil Cs (Macherey-Nagel, Diiren, Ger- method, which was developed for the de-

164 Chromatographia 2001, 54, August (No. 3/4) Original

termination of paracetamol in pharma-
ceutical formulations, was used as the re-
ference method in this work.
Paracetamol is oxidized to one or more
fluorescent compounds by hydrogen per-
oxide in the presence of horseradish per-
oxidase [5, 6]. If the HPLC mobile phase
conditions after PCD are simulated, the
fluorescence spectrum in Figure 2 is ob-
tained. The excitation maximum was
found to be at 329 nm and the emission
maximum at 435 nm. The wavelengths for
fluorescence detection were, therefore, set
to these values. To achieve appropriate
sensitivity the bandpass was set to 16 nm
for both excitation and emission, and the
high voltage of the photomultiplier tube
was set to 600 V. Figure 2. Fluorescence spectrum of paracetamol after peroxidase-catalysed oxidation. Paracetamol
in double-distilled water (10 5 M, 20 ~tL), CH3CN (0.14mL), double-distilled water (0.56mL), HRP
PCD reagent (0.4mL), and H202 PCD reagent (50 ~tL) were mixed in a cuvette and the spectrum
was recorded after reaction for 2 min.
Optimization
Chromatographic system A was opti-
mized in several steps. The Discovery col-
umn furnished favourable peak shapes for
the polar paracetamol. The gradient
shown in Table I was found to be most
suitable for post-column derivatization. It
should be noted that paracetamol is still
eluted under isocratic conditions in all
chromatograms in this work the acetoni-
trile concentration was increased to re-
move less polar residues from the column
before the start of the next chromato-
gram.
PCD was optimized with regard to
PCD reagent composition, buffer, type of
peroxidase, length of the reaction loop
(reaction time), oven temperature, and Figure 3. Optimization of the flow rate of the HRP post-column reagent.
flow rates. Optimization was conducted
using a 10 5M paracetamol solution (pre-
pared from pure paracetamol and deio- therefore, used for all subsequent experi- The flow rates of the H202 and the
nized water) by successive variation of ments. HRP PCD reagents were varied indepen-
one condition at a time, and evaluation of Replacement of H R P by microperoxi- dently, from 0.1 to 0.6mLmin 1 for the
the height of the fluorescence peak. dase (MP-11) yielded unsatisfactory re- HRP reagent and from 0.05 to 0.15mL
Mixing of the H202 and peroxidase sults peak heights were reduced by a fac- min 1 for the H202 reagent. The respec-
PCD reagents by use of one reagent pump tor of approximately 40 compared to tive flow rates o f 0 . 4 m L m i n 1 (Figure 3)
only did not yield stable fluorescence sig- H R P solution. This led to the conclusion and 50 ixLmin 1 were found to give the
nals, because of decomposition of the that under the conditions of these mea- strongest signals. The increased peak
post-column reagent. A system with two surements the microperoxidase does not height obtained for H R P reagent flow
reagent pumps was, therefore, set up as catalyse the oxidation efficiently. rates below 0.3mLmin 1 is assumed to
shown in Figure 1. Reaction loops 8, 18, and 28m long be because of the accompanying increase
When HEPES buffer (pH 8, 0.1 M) was were tested; optimum results were ob- in reaction time.
used for the HRP reagent solution the tained with the 18-m loop; this was
fluorescence signal from a 10 5M parace- equivalent to a reaction time of - 1.1 min
tamol solution decreased from 130 mV to (taking into account the flow rates of the Spiked UrineSamples
less than 4mV in the course of 18 injec- eluent and the post-column reagents after
tions over a period of 233 min, whereas optimization). For all methods, calibration was per-
use of NH3-NH4C1 buffer resulted in The oven temperature was varied be- formed in the concentration range be-
stable peak heights for the period of time tween 30 and 40 ~ in steps of 5~ 35 ~ tween 0.5 and 20 ixM. Each standard solu-
under consideration. This buffer was, was found to yield the best results. tion was prepared from pure paracetamol

Original Chromatographia 2001, 54, August (No. 3/4) 165

Table II. Data from six point calibration, with limits of detection (LOD) and quantification (LOQ). The calibration function was lOgl0(Y) = A +
B • lOgl0(X). Errors are given as the standard deviation of the mean (n = 3).

Method Intercept, A Slope, B Regression LOD LOQ

coefficient
New method UV-Vis detection (no PCD) 10.5 • 0.2 1.08 • 0.04 0.997 5 • 10 7 M 1.7 • 10 6 M
Newmethod fluorescence detection (PCD) 12.05• 1.17 • 0.999 2 • 10 8M 7 • 10 8M
Reference method 10.00 • 0.03 0.989 • 0.005 0.999 4 • 10 8 M 1.3 • 10 7 M

TaMe III. Results from analysis of the real sample (diluted pharmaceutical and spiked urine samples). Errors are given as the standard deviation of the
mean (n = 3).

New method Reference method

UV-Vis detection (no PCD) Fluorescence detection (PCD)
High concn, no matrix a (mg L 1) 1.18 • 0.05 1.17 • 1.051 •
Recovery (%) 94% 94% 84%
High concn, urine matrix b (mg L l) 1.35 • 0.09 1.15 • 2.33 •
Recovery (%) 104% 92% 186%
Low concn, no matrix b (mg L 1) 0.10 • 0.02 0.119 • 0.005 0.115 • 0.003
Recovery (%) 80% 95% 92%
Low concn, urine matrix b (mg L 1) 0.15 • 0.03 0.10 • 0.01 Evaluation not possible
Recovery (%) 150% 84%
a Expected value 1.25 mg L 1; b Expected value 0.125 mg L 1 (125 ~tgL 1).

200 of p a r a c e t a m o l was almost three times

t h a t f r o m the strongest m a t r i x interferent.
175

150
/
Discussion
125

A new, highly selective, a n d sensitive

.-~ l o o
m e t h o d has been developed for the deter-
75-
m i n a t i o n of p a r a c e t a m o l in urine. Because
p o s t - c o l u m n derivatization (PCD) with
50- subsequent fluorescence detection adds to
the selectivity of c h r o m a t o g r a p h i c separa-
25-
,-,' ;
tion, the new m e t h o d is preferable to U V -
0-
Vis detection for the analysis of urine sam-
i i i i i i I ples. In c o n t r a s t to k n o w n H P L C proce-
0 1 2 3 4 6 7 8
dures [11 13] with U V detection, the low-
t (min)
er end of the linear range using the new
Figure 4. Chromatograms obtained from analysis of the spiked urine sample: solid line = sample P C D m e t h o d starts at 101xgL 1, com-
without matrix; dotted line = sample with urine matrix. pared with 5 m g L 1, whereas elution a n d
selective detection of the analyte can be
a n d deionized w a t e r a n d analysed in tri- baselines were o b t a i n e d for the other two achieved within 5 m i n , even in the pre-
plicate. The calibration data, limits of de- methods. The urine m a t r i x interferes se- sence of a urine matrix. P C D with fluores-
tection (LOD), a n d limits of quantifica- verely with UV-Vis detection in b o t h in- cence detection yields reliable results for
tion (LOQ) are s u m m a r i z e d in Table II. stances. The signal height for the m o s t low c o n c e n t r a t i o n s of p a r a c e t a m o l in this
P C D with fluorescence detection gave the strongly a b s o r b i n g m a t r i x c o m p o n e n t s difficult matrix. Simple dilution by a fac-
best limits of detection a factor of two was approximately ten times larger t h a n tor of 20 is sufficient for the d e t e r m i n a t i o n
higher for the reference method. C h r o m a - t h a t for the high c o n c e n t r a t i o n of parace- of 2 . 5 m g L 1 p a r a c e t a m o l in urine, a n d
t o g r a m s o b t a i n e d f r o m the high-concen- tamol. Because of coelution of m a t r i x the LOQ a n d lack of interference indicate
t r a t i o n diluted drug solution a n d f r o m c o m p o n e n t s , evaluation of the reference t h a t even lower c o n c e n t r a t i o n s m i g h t be
spiked urine, using P C D with fluorescence m e t h o d was possible only for samples accessible by the new technique.
detection, are depicted in Figure 4. The re- w i t h o u t biological matrices. F o r matrix- A n attractive option remains for
sults o b t a i n e d f r o m all the samples are free samples excellent c h r o m a t o g r a m s further optimization of the system. Incor-
listed in Table III. were o b t a i n e d by use of the reference p o r a t i o n a n d i m m o b i l i z a t i o n of the perox-
A very u n s t e a d y baseline was o b t a i n e d m e t h o d , only few m a t r i x c o m p o n e n t s idase in a p o s t - c o l u m n reactor m i g h t re-
w h e n the new m e t h o d was used with U V - fluoresce at the detection wavelengths sult in better signal stability a n d simplify
Vis detection (because of the low flow-cell used. W h e n the new m e t h o d with P C D the p o s t - c o l u m n i n s t r u m e n t a t i o n [19, 20].
volume a n d pressure changes induced by a n d fluorescence detection was used the
the p o s t - c o l u m n p u m p s ) whereas stable signal height from the high c o n c e n t r a t i o n

166 C h r o m a t o g r a p h i a 2001, 54, A u g u s t (No. 3/4) Original

References son, I.D.; Bailey, N.J.; Scarfe, G.B., Ni- [16] Hellpointner, E.; G~ib, S. Nature 1989,
cholson, J.K.; Lindon, J.C. Analyst 1998, 337, 631 634.
123, 2835 2837. [17] Engelhardt, H.; Lillig, B. Chromatogra-
[1] Krishna, K.V.; Archana, J. Talanta 1985, [10] Di Pietra, A.M.; Gatti, R.; Andrisano, V.; phial986,21,136 142.
32, 238 240. Cavrini, V. J. Chromatogr. A 1996, 729, [18] Forth, W.; Henschler, D.; Rummel, W.;
[2] Martinez Calatayud, J.; Gomez Benito, C. 355 361. Starke, K.; Eds. Allgemeine und spezielle
Anal. Chim. Aeta 1990,231,259 264.
[11] Lau, G.S.N.; Critchley, J.A.J.H.J. Pharm. Pharmakologie und Toxikologie, Spek-
[3] Murillo Pttlgarin, J. A.; Garcia Bermejo Biomed. Anal. 1994,12, 1563 1572. trum Akademischer Verlag, Heidelberg,
L.F. Anal. Chim. Aeta 1996, 333, 59 69. [12] Goicoechea, A.G.; Dealda, M.J.L.; Vila- 1998.
[4] Carmona, M.; Silva, M.; P6rez-Bendito, jato, J.L.J. Liq. Chromatogr. 1995, 18, [19] Li, Y.-Z.; Townshend, A. Anal. Chim.
D. Anal. Chim. Aeta 1989,218, 312 333.
3257 3268. Aeta 1998,359, 149 156.
[5] Potter, D.W.; Miller, D.W.; Hinson, J.A. [13] Di Girolamo, A.; O'Neill, W.M.; Wainer, [20] Sakuragawa, A.; Taniai, T.; Okutani, T.
J. Biol. Chem. 1985,260, 12174 12180. Anal. Chim. Aeta 1998, 374, 191 200.
I.W.J. Pharm. Biomed. Anal. 1998, 17,
[6] McCormick, C.P.; Shihabi, Z.K.J. Liq. 1191 1197.
Chromatogr. 1990, 13, 1159 1171. [14] Stork, C.M.; Rees, S.; Howland, M.A., Received: Oct 23, 2000
[7] Kunkel, A.; W~itzig, H. Eleetrophoresis Kaplan, L.; Goldfrank, L.; Hoffmann, Revised manuscripts
1999,20, 2379 2389. R.S.J. Toxieol. Clin. Toxieol. 1996, 34, received: Jan 29
[8] Heitmeier, S.; Blaschke, G. J. Chroma- 157 162. and Feb 15, 2001
togr. B1999, 721,93 108.
[15] Guilbault, G.G.; Brignac, P.J.; Juneau, Accepted: Apr 9, 2001
[9] Schewitz, J.; Gfr6rer, P.; Pusecker, K.; M. Anal. Chem. 1968, 40, 1256 1263.
Tseng, L.H.; Albert, K.; Bayer, E.; Wil-

Original Chromatographia 2001, 54, August (No. 3/4) 167