Determination of ciprofloxacin and enrofloxacin in edible
animal tissues by terbium-sensitized luminescence†
J. A. Hernández-Arteseros, R. Compañó and M. D. Prat*
Departamento de Química Analítica, Universitat de Barcelona, Diagonal 647, 08028
Barcelona, Spain. E-mail: [email protected]
Received 3rd July 1998, Accepted 29th September 1998
A terbium-sensitized luminescence method is described for the determination of the sum of residues of
enrofloxacin and its major metabolite ciprofloxacin in edible animal tissues. Several parameters affecting both
detection and extraction were studied. Analytes were extracted from spiked samples of chicken and trout tissues
with pH 7.4 buffer–dichloromethane. The organic extract was evaporated and the residue dissolved in aqueous
nitric acid and defatted with hexane. Determination was carried out directly in the aqueous phase (in a micellar
medium). The calibration curves were linear up to 75 mg l21. The detection limit was 3.5 mg kg21 (for a 5 g
sample) and the repeatability was 7.0% (n = 7). The sensitivity was similar for both quinolones and therefore
calibration can be carried out with either ciprofloxacin or enrofloxacin. In any case, the differences were <10%.
Introduction
Experimental
Fluoroquinolones are broad-spectrum antibacterial agents
which are widely used as veterinary drugs in food-producing
animals. Their misuse has led to the drawing up of enforcement
regulations. For example, the European Community (EC) has
fixed a maximum residue limit (MRL) in edible animal products
for some fluoroquinolones, such as enrofloxacin (ENR) and its
metabolite ciprofloxacin (CIP). In this case, the MRL for the
sum of ENR and CIP has been settled at 30 mg kg21 in several
edible animal tissues.1 Therefore, the development or improvement of analytical methods for monitoring their levels in farm
animals and their primary products is of interest.
Current methods of the analysis of quinolones are based on
liquid chromatography (LC), mainly with fluorimetric detection.2–4 Some procedures have been developed to detect several
naturally fluorescing quinolones2,3 and other methods use postcolumn derivatization to form fluorescent compounds.4
Luminescence spectroscopy offers other possibilities for
sensitive and selective detection, such as the use of lanthanidesensitized luminescence. Several compounds are known to have
efficient lanthanide-sensitizing characteristics and can be
detected by time-resolved luminescence.5,6 The main advantages of this technique include long wavelength emission and
long luminescence lifetimes, resulting in the elimination of
scattering interference and a significant decrease in the
background fluorescence. Since quinolones have suitable
functional groups to form stable complexes with these metal
ions, intramolecular energy transfer between the quinolone and
the lanthanide ion can occur.7–11 In this study we investigated
the potential of terbium-sensitized luminescence spectroscopy
for the determination of the sum of ciprofloxacin and enrofloxacin in edible animal tissues (chicken and trout) without
chromatographic separation. An extraction procedure compatible with the terbium-based detection system is proposed. The
detection limits are low enough to determine concentrations
below the permissible MRL in animal products. The proposed
method is rapid and makes use of small volumes of organic
solvents, which are costly and hazardous substances.
Materials and reagents
† Presented at the Third International Symposium on Hormone and
Veterinary Drug Residue Analysis, Bruges, Belgium, June 2–5, 1998.
Ciprofloxacin hydrochloride and enrofloxacin (Fig. 1) standards were kindly supplied by Cenavisa (Reus, Spain). Stock
standard solutions (100 mg l21) of the quinolones were
prepared by dissolving the compounds in 0.01 m nitric acid and
were stored in dark glass bottles at 4 °C. Working standard
solutions were freshly prepared by dilution with 0.01 m nitric
acid.
A 1 g l21 Tb(iii) solution was prepared by dissolving the
appropriate amount of TbCl3·6H2O (Alfa, Karlsruhe, Germany)
in 0.01 m aqueous nitric acid. The solution was stored in
polyethylene bottles.
Buffer solutions of 0.25 m formic acid–NaOH, 0.25 m acetic
acid–sodium acetate, 0.25 m succinic acid–NaOH and 0.1 m
diethylmalonic acid–NaOH were used.
A stock standard solution of 0.2 m sodium lauryl sulfate
(SLS) (Merck, Darmstadt, Germany) was prepared.
Doubly de-ionized water (Milli-Q; Millipore, Molsheim,
France) with a resistivity of 18.2 MW cm21 was used
throughout. All other reagents and solvents were of analytical
reagent grade.
All glassware used for experiments was soaked in 10% nitric
acid for 24 h and rinsed with doubly de-ionized water.
Apparatus
Luminescence measurements were performed using a PerkinElmer (Beaconsfield, Buckinghamshire, UK) LS-50 fluores-
Fig. 1 Structural formulae of the quinolones studied.
Analyst, 1998, 123, 2729–2732
2729
cence spectrophotometer equipped with a pulsed xenon lamp
(60 Hz) and a 10 mm quartz cell. Excitation and emission slits
were set to 5 and 20 nm, respectively. The instrument was set in
the phosphorescence mode and a delay time (td) of 0.03 ms and
a gate time (tg) of 5 ms were used. Terbium(iii) luminescence
was detected at an emission wavelength of 549 nm with an
excitation wavelength of 276 nm.
A Radiometer (Copenhagen, Denmark) PHM 84 pH meter
equipped with an Orion (Boston, MA, USA) 81-02 Ross
combination electrode was used for pH measurements.
A Breda Scientific rotary shaker (Breda, Netherlands) and a
Heraeus Christ centrifuge (Osterode am Harz, Germany) were
used to carry out extractions. A rotary Resona Technics LABO
ROTA S300 evaporator (Gossau, Switzerland) was used to
remove the extracting solvent.
Samples
Trout and chicken tissues used for the preparation of spiked
samples were obtained from Navarra Food (Yusa, Spain) and
from the Laboratori de Salut Pública de la Generalitat de
Catalunya (Barcelona, Spain), respectively. The samples
chosen contained neither CIP nor ENR. Skin and bones were
removed before grinding the muscle. Minced muscle was stored
at 220 °C and each sample was thawed before analysis.
Procedure
Weigh 5.00 g of thawed sample in a 30 ml centrifuge tube. Add
1.5 ml 0.1 m diethylmalonic acid buffer (pH 7.4) and 20 ml of
CH2Cl2. Agitate for 10 min then centrifuge for 10 min at 3500
rpm. Transfer the organic phase into a 50 ml heart-shaped flask.
Rinse the sample with another 10 ml portion of CH2Cl2.
Centrifuge again for 10 min at 3500 rpm. Combine the two
organic extracts and add 1 ml of 0.5 m aqueous NaCl in 0.01 m
HNO3. Evaporate under vacuum in a rotary evaporator at room
temperature until only aqueous phase remains (about 7 min).
Defat by extraction with 10 ml of hexane.
Transfer 0.5 ml of the aqueous phase into a 10 ml calibrated
flask. Add 1.5 ml of Tb(iii) solution, 0.5 ml of SLS and 5 ml of
buffer solution (0.25 m acetic acid–sodium acetate, pH 6.0) and
dilute with water. Measure the emission intensity at 549 nm
using an excitation wavelength of 276 nm.
To optimize the detection system, the effects of chemical and
instrumental variables on the luminescence intensity were
examined at a quinolone level of 200 mg l21. Measurements
were performed at the same excitation wavelength for both
quinolones (276 nm). As can be seen in Fig. 3, the luminescence
intensity was at a maximum in the pH range 5.7–6.3 for ENR
and 5.7–6.8 for CIP. We found that the luminescence intensity
was greatly affected by the buffer composition. The variation
shown in Fig. 3 was obtained with formic and acetic acid
buffers. Although succinic acid and diethylmalonic acid buffers
possess a higher buffer capacity at pH values above 5.8, a
decrease in the emission occurred when these dicarboxylic acids
were used. Therefore, the optimum response was found with an
acetic acid buffer at pH 6.
The luminescence intensity increased on addition of SLS at
concentrations above its critical micellar concentration (c.m.c.),
but at higher SLS concentrations no significant effect was
found. Moreover, if the SLS concentration was below its c.m.c.,
a precipitate appeared due to Tb(LS)3.12 The c.m.c. for SLS
decreases with increasing ionic strength (I) (from 8.1 3 1023 m
at I = 0 to 1.4 3 1023 m at I = 0.1 m). Hence the use of
concentrated buffer solutions (0.12 m) leads to micellar
solutions at a concentration of SLS of 1022 m.
Studies on the effect of terbium concentration showed that
the response reaches a maximum at about 150 mg l21.
Instrumental parameters, such as td, tg and slits, were also
studied. A delay time of 0.03 ms is sufficient to eliminate
scattering emission. The optimal gate time was found to be 5 ms
and at higher values no increase in emission was observed.
Optimum conditions for the detection system are summarized
in Table 1.
There is a linear relationship between luminescence and
concentration for CIP and ENR up to 75 mg l21 (the highest
concentration tested). The limits of detection and quantification
were calculated as the concentrations corresponding to a signal
of 3 and 10 times the standard deviation of 10 blanks,
respectively. As can be seen in Table 2, there is little difference
Results and discussion
Luminescence studies
Ciprofloxacin and enrofloxacin show native fluorescence, with
excitation and emission wavelengths around 275 and 440 nm,
respectively. Addition of terbium ions to a quinolone solution
results in the formation of complexes that absorb energy at the
characteristic wavelength of the organic ligands and emit
radiation at the characteristic wavelength of Tb3+, 549 nm (Fig.
2). It was observed that luminescence was increased by the
addition of micelle forming surfactants such as sodium lauryl
sulfate. Rayleigh scattering, which is significant at wavelengths
near 549 nm (lexc = 276 nm) can be eliminated by using a
suitable delay time.
Fig. 2 Scheme of the terbium-sensitized luminescence process. An
asterisk indicates an excited state.
2730
Analyst, 1998, 123, 2729–2732
Fig. 3 Variation of the emission luminescence for the terbium complexes
of (-) CIP and (:) ENR versus pH in SLS medium.
Table 1 Parameters studied and optimum values selected in terbiumsensitized luminescence
Parameter
Range studied
Optimum value
pH
2.5–7.0
[Tb3+]/mg l21
[SLS]/m
Delay time/ms
Gate time/ms
Excitation wavelength/nm
Emission wavelength/nm
Excitation slit/nm
Emission slit/nm
4–300
0–0.08
0–12
1–10
240–400
400–700
2.5–15
5–20
6 (acetic acid–acetate
buffer)
150
0.01
0.03
5
276
549
5
20
between the CIP and ENR slopes and this is important for the
joint determination of both quinolones.
Optimization of the extraction procedure
Selection of extracting agent. The extraction of quinolones
from biological matrices has been achieved using different
methods. Extraction of the anionic or cationic forms with watermiscible organic solvents in acidic or basic media is the most
common approach.2,3,13,14 Partition equilibria between aqueous
buffer solutions and non-water-miscible solvents, such as
chloroform, dichloromethane (DCM) or ethyl acetate, has also
been used.15,16 In these cases the extracted species is the
uncharged neutral form in equilibrium with the zwitterionic
form.
In order to investigate the compatibility of extraction systems
with the terbium-sensitized detection method, the following
extracting agents were tested: (a) acidic ethanol, (b) acidic
acetonitrile, (c) basic acetonitrile, (d) basic acetone, (e) ethyl
acetate–buffer solution (pH 7.4) and (f) dichloromethane–
buffer solution (pH 7.4).17 Samples of chicken tissues containing no quinolone were extracted with each solvent system and
the analytes were added to the extracts. Extraction methods (a)–
(e) did not provide satisfactory results, since addition of Tb(iii)
to extracts caused precipitation, which hindered further determination. The precipitation was attributed to the presence of
significant concentrations of phosphate in the chicken tissues,
which was co-extracted with the analytes when polar solvents
were used. Phosphate forms a highly insoluble compound with
Tb(iii). Although phosphate interference could be eliminated by
means of an ion exchange clean-up step, it would increase the
complexity of the analysis. When extraction was performed
with DCM no positive reaction to phosphate was found.
Moreover, the amount of fat extracted was less than with the
other extracting agents. Therefore, DCM–diethylmalonic acid
buffer (pH 7.4) was chosen to carry out the extraction.
Distribution of analytes in DCM–water. In order to find the
optimum conditions for the extraction of CIP and ENR, the
distribution of the analytes in a DCM–water system was
studied. With this aim, individual standard solutions of CIP (or
ENR) were diluted with buffer solutions and then shaken with
DCM. After separation, the concentration in both phases was
determined by measuring the absorbance at 271 nm and
comparing it with that of standard solutions in water and DCM,
respectively. The pH of the aqueous phase, volume ratio of
organic solvent to aqueous phase, equilibration time and
method of agitation were optimized.
The acid–base characteristics of the analytes render extraction strongly pH dependent. Owing to the carboxylic group and
the ammonium of the piperazine ring, the quinolones studied
can be present in aqueous solution as cationic, anionic or
intermediate forms. Since the carboxylic group is stronger than
the ammonium group, the intermediate form is a zwitterion and
maximum extraction is achieved at pH values around 0.5 (pK1
+ pK2), at which zwitterionic species prevail in the aqueous
phase and can be extracted as neutral molecules. The isoelectric
point can be calculated from the pKa values for CIP and ENR
available in the literature (pK1,CIP = 6.2, pK2,CIP = 8.5, pK1,ENR
= 6.2, pK2,ENR = 7.8).18 These values led to an optimum pH of
7.4 for CIP and 7.0 for ENR. The distribution of analytes was
independent of pH between 7.0 and 8.0.
Rotary agitation and sonication were tested and sonication
was ruled out because decomposition of enrofloxacin in DCM
medium was observed.
It was found that the equilibration time is not an important
factor when working with standard solutions, since equilibrium
was reached in a few seconds.
As can be seen in Fig. 4, the recovery of ciprofloxacin
increased continuously with increasing DCM : water ratio from
0.1 to 8 but, in contrast, enrofloxacin reached a maximum
recovery (80%) at a volume ratio of the organic and aqueous
phases of about 1.5 and no further increase was observed. Hence
the extraction conditions described under Procedure (20 + 10 ml
of DCM) led to recoveries of about 90%.
Solvent change. As DCM is not a suitable solvent for
terbium-sensitized luminescence detection, a solvent change to
water was essential. First, back-extraction of the analytes in an
acidic aqueous solution (0.01 m nitric acid) was attempted but
no satisfactory results were obtained. Although the method
proved to be suitable for ENR, no recovery of CIP was obtained.
Neither longer equilibration times (up to 1 h) nor higher
concentrations of HNO3 (up to 1 m) succeeded in backextracting CIP from the organic phase. This surprising result
was not easily explained but it could be related to the differences
in solvation of CIP and ENR by DCM. These differences must
be due to the presence of an ethyl chain on the nitrogen of the
piperazine ring in ENR. Rotary evaporation under vacuum and
dissolution of the residue in water proved to be suitable.
Although evaporation to dryness led to some losses of analytes,
this could be avoided by adding a small volume of an aqueous
solution containing 0.01 m nitric acid and 0.5 m NaCl to the
extract prior to evaporation.
Analysis of chicken and trout tissues
Prior to the analysis of animal tissues, the matrix effect on the
luminescent response was investigated. Calibration graphs were
established from standard solutions containing extracts from
different analyte-free samples, and calibration data for each
compound in the two different matrices studied are given in
Table 3. The calibration curves are linear up to 75 mg l21 (the
highest concentration tested). The sensitivities are similar for
both compounds but they depend on the matrix composition:
Table 2 Figures of merit of terbium-sensitized luminescence in a pure
water matrix
Parameter
CIP
ENR
Calibration linea
r
LOD/mg l21
LOQ/mg l21
RSD (n = 5) at 30 mg l21 (%)
3.44c + 18.80
0.9998
0.8
2.7
2.0
2.94c + 19.45
0.9997
0.9
3.2
1.8
a
Concentration c expressed in mg l21.
Fig. 4 Plots of recoveries of (-) CIP and (:) ENR versus VCH2Cl2/VH2O
and VCH2Cl2/Vtotal.
Analyst, 1998, 123, 2729–2732
2731
Table 3
Figures of merit of total procedure (extraction + detection). LOD and LOQ were calculated taking recoveries into account
Trout
Chicken
Parameter
CIP
ENR
CIP
ENR
Calibration linea
r
Recovery (%)
LOD/mg kg21
LOQ/mg kg21
RSD (n = 7) at 30 mg kg21 (%)
2.47c + 44.51
0.9998
61.9
2.9
9.7
6.6
2.29c + 43.58
0.9998
67.1
2.7
9.2
6.9
2.11c + 34.61
0.9994
47.4
3.8
12.7
7.2
2.08c + 32.88
0.9991
53.8
3.4
11.4
6.9
a
Concentration c expressed in mg l21.
pure water (Table 2 and Fig. 3), chicken extract or trout extract.
These results suggest that quantification should be carried out
by using matrix-matched standards or the standard additions
method. However, the latter is more tedious and therefore less
practical for large numbers of samples. It should be noted that
data obtained from different chicken (or trout) samples showed
no significant differences.
To evaluate the recovery of CIP and ENR from edible animal
tissues, the proposed method was applied to the analysis of
spiked chicken and trout tissues. Samples not containing any of
the compounds of interest were analysed after additions of
known amounts of CIP or ENR. Spiking was performed on each
portion of sample by adding 10 ml of aqueous standard solution
to the weighed sample and leaving it to stand for 30 min in the
dark before the analysis.
Replicate experiments on samples spiked at four levels (from
10 to 100 mg kg21) for each fluoroquinolone demonstrated that
the recovery is independent of the initial concentration.
Moreover, it was found that both analytes were recovered in
similar yields. The mean recoveries from trout tissues were 62%
for CIP and 67% for ENR, whereas lower values (about 50%)
were obtained from chicken tissues (Table 3). The repeatability
was 6.7 and 7.0% for CIP and ENR, respectively. Although
better recoveries have been reported for ENR,14 for CIP our
recoveries are similar to those found in the literature.2,15
The low recovery from chicken tissues, which could be due to
the sticky texture of the tissue, indicates that the sample
treatment should be improved and further work is now being
focused in this direction.
Since both analytes show similar calibration graphs and
extraction recoveries, the method is appropriate to determine
the sum of CIP and ENR residues in treated animals. The total
amount of analyte (CIP + ENR) can be calculated from
calibration curves obtained with either CIP or ENR. In any case,
the differences between the results obtained are <10%.
Navarra Food for trout samples. Financial support from the
Comissionat per Universitats i Recerca de la Generalitat de
Catalunya (CIRIT SGR97-394) is gratefully acknowledged.
J. A. Hernández Arteseros also thanks CIRIT for an FI
scholarship.
References
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Acknowledgements
The authors thank the Laboratory de Salut Pública de la
Generalitat de Catalunya for providing chicken samples and
2732
Analyst, 1998, 123, 2729–2732
17
18
EEC Commission Regulation (EEC) No. 2901/93 of 23/10/93,
European Economic Community, Brussels, 1993.
R. Charrière, W. Leiser and R. Dousse, Trav. Chim. Aliment. Hyg.,
1996, 87, 223.
T. J. Strelevitz and M. C. Linhares, J. Chromatogr. B, 1996, 675,
243.
H. Scholl, K. Schmidt and B. Weber, J. Chromatogr., 1987, 416,
321.
Y. Y. Xu and I. A. Hemanilä, Talanta, 1992, 39, 759.
A. Rieutord, P. Prognon, F. Brion and G. Mahuzier, Analyst, 1997,
122, 59R.
A. Rieutord, L. Vázquez, M. Soursac, P. Prognon, J. Blais, P. Bourget
and G. Mahuzier, Anal. Chim. Acta, 1994, 290, 215.
T. J. Wenzel, K. Zomlefer, S. B. Rapkin and R. H. Keith, J. Liq.
Chromatogr., 1995, 18, 1473.
S. Panadero, A. Gómez-Hens and D. Pérez-Bendito, Anal. Chim.
Acta, 1995, 303, 39.
M. Rizk, F. Belal, F. A. Aly and N. M. El-Enany, Anal. Lett., 1997,
30, 1897.
C. J. Veiopoulou, P. C. Ioannou and E. S. Lianidou, J. Pharm.
Biomed. Anal., 1997, 15, 1839.
M. Amin, K. Harrington and R. von Wandruszka, Anal. Chem., 1993,
65, 2346.
M. Juhel-Gaugin and J. P. Abjean, Chromatographia, 1998, 47,
101.
A. Rogstad, V. Hormazábal and M. Yndestad, J. Liq. Chromatogr.,
1991, 14, 521.
A. Anadón, M. R. Martínez Larrañaga, M. J. Díaz, P. Bringas, M. A.
Martínez, M. L. Fernández-Cruz, M. C. Fernández and R. Fernández,
Am. J. Vet. Res., 1995, 56, 501.
J. M. Degroodt, B. Wyhowski de Bukanski and S. Srebrnik, J. Liq.
Chromatogr., 1994, 17, 1785.
M. Lizondo, M. Pons, M. Gallardo and J. Estelrich, J. Pharm.
Biomed. Anal., 1997, 15, 1845.
I. Turel, N. Bukovec, and E. Farkas, Polyhedron, 1996, 15, 269.
Paper 8/05160G