Alternatives for sample pre-treatment and HPLC determination of Ochratoxin A in red wine using fluorescence detection.
ABSTRACT Ochratoxin A is a mycotoxin widely studied due to its nephrotoxic, immunotoxic, teratogenic and carcinogenic effects. The European Commission has fixed maximum limits for Ochratoxin A in wines and in other foods. In order to determine Ochratoxin A levels in red wine, the present paper contrasts and discusses the results of a systematic study of analytical parameters for sample pre-treatment using different immunoaffinity cartridges as well as C-18 cartridges with three solvent combinations. The direct injection of wine into two types of C-18 chromatographic columns (conventional packed column and monolithic column) is evaluated as screening method. In all cases, the analysis was carried out using HPLC with fluorescence detection. The results show statistical differences when 3 types of immunoaffinity columns were used, while higher recoveries were obtained for C-18 cartridges using acetonitrile as extraction solvent. Repeatability and accuracy of immunoaffinity and C-18 sample pre-treatment were statistically comparable (alpha=0.05). Their sensitivity was also comparable, although more favorable detection limits were obtained using the immunoaffinity treatment (0.01 microg L(-1)) in comparison with C-18 treatment (0.09 microg L(-1)). Considering the maximal allowed concentration of Ochratoxin A in wine (2.00 microg L(-1)), both methods are suitable for its determination in wine. Both methods were applied to determine this toxin in 154 wine samples, and the quantitative results demonstrated statistic comparability (alpha=0.05). These results were also confirmed from the qualitative point of view using a GC-MS method. To find an easy screening method, based on a recent publication, a monolithic HPLC column and 2 conventional packed columns were tested for Ochratoxin A determination in real wine samples by direct injection, without previous clean-up. The results show that this procedure is not useful at the concentration levels usually found in wine and although shorter time is required when using the monolithic columns even with the chromatographic analysis. Finally, based on the results, it was concluded that the combination of C-18 cartridges with conventional particle packed columns and HPLC-FLD is the most appropriate alternative for Ochratoxin A analysis in wine. Indeed, considering cost, sensitivity and selectivity, this method can be used in broad prospective programs.
- SourceAvailable from: Michele Solfrizzo[show abstract] [hide abstract]
ABSTRACT: The main source of ochratoxin A (OTA) in the wine food chain is the infection of grapes by "black aspergilli" in the field. OTA-producing black aspergilli include principally Aspergillus carbonarius, followed by A. niger and possibly A. tubingensis. They are opportunistic fungi that develop particularly on damaged berries at ripening, although they may occur and form OTA on grapes from veraison to harvest. Climatic conditions (high humidity and temperature) and geographical location are important factors favouring OTA accumulation in grape berries. The severity of aspergillus rot is influenced by excessive irrigation and rainfall prior to harvest, which causes berry splitting. In addition, berry wounds caused by insect attack provide preferential entries for black aspergilli. High OTA levels occur in grapes severely damaged by the grape moth, Lobesia botrana, particularly in Mediterranean areas. Some grape varieties display greater susceptibility to aspergillus rot due to intrinsic genetic characteristics and bunch conformation (i.e. compact>sparse). Control measures for toxigenic mycoflora in the vineyards must consider these critical control points. Proper fungicidal and insecticidal treatments can reduce OTA contamination. Nevertheless, knowledge about the fate of OTA and its distribution in wine and winery by-products is important to manage OTA risk in contaminated stock. In our wine-making experiments, only 4% of the OTA present in grapes remained in the wine--the majority is retained in pressed grape pomaces. OTA concentration remained unchanged in wine after a 1-year aging as well as in all liquid fractions collected during vinification (i.e. must, free run wine, and wine after first and second decantation). Activated carbon can reduce OTA levels in wine but negatively affects wine quality.Food Additives and Contaminants - Part A Chemistry, Analysis, Control, Exposure and Risk Assessment 03/2008; 25(2):193-202.
- [show abstract] [hide abstract]
ABSTRACT: Ochratoxin A (OTA) is a secondary fungal metabolite produced by several moulds, mainly by Aspergillus ochraceus, A. carbonarius, A. niger and by Penicillium verrucosum. The present work shows the results of comparative studies using different procedures for the analysis of OTA in maize bread samples. The studied analytical methods involved extraction with different volumes of PBS/methanol, different extraction apparatus, and clean-up through immunoaffinity columns. The separation and identification were carried out by high-performance liquid chromatography with fluorescence detection. The optimized method for analysis of OTA in maize bread involved extraction with PBS:methanol (50:50), and clean-up with IAC column. The limit of quantification was 0.033ngg(-1). Recoveries ranged from 87% to 102% for fortifications at 2.000 and 0.500ngg(-1), respectively, within-day R.S.D. of 1.4% and 4.7%. The proposed method was applied to 15 samples and the presence of OTA was found in nine samples at concentrations ranging from nd to 2.650ngg(-1).Talanta 10/2007; 73(2):246-50. · 3.50 Impact Factor
- [show abstract] [hide abstract]
ABSTRACT: This study investigated the impact of skin damage on Aspergillus carbonarius colonization and ochratoxin A (OTA) production in grapes at different temperatures and relative humidity. Four ochratoxigenic A. carbonarius strains isolated from wine grapes were used to inoculate artificially damaged and undamaged table grapes. Grapes were stored at three levels of relative humidity (80%, 90% and 100%) and at two temperatures (20 and 30 °C). After seven days, the infection percentage of A. carbonarius was recorded and OTA accumulation in berries was analysed. Damaged grapes were more commonly infected and development of colonies was higher than in undamaged ones; consequently more OTA was detected in the former treatment. Temperature and relative humidity had significant influences on both infection and toxin content. The amount of OTA detected at 30 °C was higher than at 20 °C in most of the treatments. The highest relative humidity (100%) led to maximum amounts of OTA while no significant differences were found between 90% and 80% in the OTA content. The implementation of preventive measures in order to minimise berry damage in the field by controlling pathogenic fungi and insects during grape growing and removing visibly damaged grapes at harvest may significantly reduce OTA contamination in grapes.Food Control. 01/2007;
Analytica Chimica Acta 660 (2010) 119–126
Contents lists available at ScienceDirect
Analytica Chimica Acta
journal homepage: www.elsevier.com/locate/aca
Alternatives for sample pre-treatment and HPLC determination of Ochratoxin A
in red wine using fluorescence detection
Catherine Tessini, Claudia Mardones∗, Dietrich von Baer, Mario Vega, Erika Herlitz,
Roberto Saelzer, Jorge Silva, Olga Torres
Departamento de Análisis Instrumental, Departamento de Bromatología, Facultad de Farmacia, Universidad de Concepción, Casilla 237, Concepción, Chile
a r t i c l ei n f o
Received 21 August 2009
Received in revised form 5 November 2009
Accepted 9 November 2009
Available online 24 November 2009
High performance liquid
Solid phase extraction
Gas chromatography-mass spectrometry
a b s t r a c t
Ochratoxin A is a mycotoxin widely studied due to its nephrotoxic, immunotoxic, teratogenic and car-
cinogenic effects. The European Commission has fixed maximum limits for Ochratoxin A in wines and in
other foods. In order to determine Ochratoxin A levels in red wine, the present paper contrasts and dis-
cusses the results of a systematic study of analytical parameters for sample pre-treatment using different
of wine into two types of C-18 chromatographic columns (conventional packed column and monolithic
column) is evaluated as screening method. In all cases, the analysis was carried out using HPLC with
fluorescence detection. The results show statistical differences when 3 types of immunoaffinity columns
were used, while higher recoveries were obtained for C-18 cartridges using acetonitrile as extraction
solvent. Repeatability and accuracy of immunoaffinity and C-18 sample pre-treatment were statistically
comparable (˛=0.05). Their sensitivity was also comparable, although more favorable detection lim-
its were obtained using the immunoaffinity treatment (0.01?gL−1) in comparison with C-18 treatment
methods are suitable for its determination in wine. Both methods were applied to determine this toxin
in 154 wine samples, and the quantitative results demonstrated statistic comparability (˛=0.05). These
results were also confirmed from the qualitative point of view using a GC–MS method. To find an easy
screening method, based on a recent publication, a monolithic HPLC column and 2 conventional packed
columns were tested for Ochratoxin A determination in real wine samples by direct injection, without
previous clean-up. The results show that this procedure is not useful at the concentration levels usually
found in wine and although shorter time is required when using the monolithic columns even with the
chromatographic analysis. Finally, based on the results, it was concluded that the combination of C-18
cartridges with conventional particle packed columns and HPLC–FLD is the most appropriate alternative
for Ochratoxin A analysis in wine. Indeed, considering cost, sensitivity and selectivity, this method can
be used in broad prospective programs.
© 2009 Published by Elsevier B.V.
Ochratoxin A (OTA) is a mycotoxin commonly present in wine
and other foods . It is a toxic secondary metabolite produced by
ence in foods has been studied extensively because OTA is known
to have nephrotoxic, immunotoxic, teratogenic and carcinogenic
effects [3,4]. For these reasons, the European Commission has fixed
maximum limits for OTA in wines and other foods . OTA can be
present on the grape and transferred to wine during the fermenta-
tion process. The maximum permitted level for white, rosé and red
wines is 2.00?gL−1.
∗Corresponding author. Tel.: +56 412204598.
E-mail address: email@example.com (C. Mardones).
The International Organization of Vine and Wine (OIV) in its
Resolution OENO 16/2001 established as analytical method for
OTA quantification in wine, its pre-treatment on immunoaffin-
ity column (IAC) and HPLC separation with fluorescence detection
(FLD) . Several studies include the determination of this inju-
rious compound in wine and other matrixes using this method
[2,7–9], obtaining good sensitivity and especially a high selectivity
[6,7,10]. When using this method to analyze OTA in wine, typical
three principal IACs are commercially available for this purpose, no
systematic information about possible differences in the recovery
obtained with these devices is available.
The utilization of HPLC–MS/MS with different sample treat-
ments has also been studied [12–14]. Even when this approach
achieved excellent selectivity, its detection limits are lower than
those obtained using FLD . Another proposed alternative
0003-2670/$ – see front matter © 2009 Published by Elsevier B.V.
C. Tessini et al. / Analytica Chimica Acta 660 (2010) 119–126
is solid phase microextraction (SPME) using a polydimethyl-
siloxane/divinylbenzene (PDMS/DVB) fiber for sample extraction,
coupled with HPLC–FLD. However, although sufficient selectiv-
ity was obtained, a lower sensitivity than the IAC treatment was
observed . Other techniques, like Capillary Electrophoresis,
have been evaluated, obtaining unfavorable sensitivities [16–19].
Consequently, this technique is not appropriate to determine OTA
in wine, considering the concentration levels commonly found
[15,20]. GC–MS has been described for qualitative analysis, and is
a good alternative for confirmatory analysis of OTA in wine .
C-18 solid phase extraction (SPE) cartridges has been proposed
using different solvents for OTA extraction from wine [20,22]. Sáez
et al.  published a comparative study including different SPE
cartridges, showing that OTA treatment on C-18 cartridges with
methanol–acetic acid (99.5:0.5) as elution solvent provided good
hand, Leitner et al.  compared SPE with C-18 column and IAC
treatment using HPLC–MS/MS, showing that the analytical data
obtained with C-18 SPE combined with LC–MS/MS detection and
immunoaffinity extraction combined with FLD offered comparable
good results in the sub-ppb concentration level. This publication
indicated that a high selectivity of either the sample clean-up or
alternatively the detection system is equally well-suited to guar-
antee an accurate OTA analysis in wine. However, in the last case,
due to high equipment costs.
In the search of inexpensive, fast methods for OTA detection in
wine, Tafuri et al.  recently published a method for its direct
determination in wine without sample pre-concentration and
clean-up, using HPLC–FLD with a monolithic column. The method
was evaluated with spiked samples between 0.1 and 1?gL−1of
OTA, and according to the authors, obtained results are compara-
ble to those obtained with the IAC method. Still, the publication
does not present a chromatogram of a real wine sample analyzed
using the proposed method .
The present work contrasts the analytical parameters for differ-
conventional particle packed and monolithic C-18 columns. The
relevant results are also confirmed by GC–MS with previous OTA
derivatization. The main aim of this work is the systematic com-
parison of sample pre-treatment methods for OTA determination
by fluorescence detection, considering their analytical parame-
ters, costs, time and results obtained for a larger number of real
wine samples. The evaluation of direct injection of wine using a
monolithic column and conventional particle packed column are
also included and contrasted in order to evaluate their sensitiv-
ity, selectivity, analysis time and costs for broad OTA detection
2. Materials and methods
The HPLC analyses were carried out on a Merck Hitachi HPLC
system equipped with a Hitachi L-2200 auto-sampler, a Hitachi
L-2130 pump and a Hitachi L-7485 Fluorescence detector (Merck
KG, Darmstadt, Germany). The data was processed with an Inter-
active Graphics Software, version 6.20 from Varian Inc. (Palo Alto,
The GC–MS analysis were performed on a HP 6890 Series gas
chromatograph Hewlett-Packard (Palo Alto, CA, USA) equipped
The system was controlled by a HP ChemStation G1701AA, version
2.2. Material, reagents and standards
The OTA (N-[(3R)-(5chloro-8-hydroxy-3-methyl-1-oxo-7-
isochromanyl)carbonyl]-l-phenyl-alanine) standard was obtained
from Sigma (St. Louis, MO, USA). NaHCO3, NaCl, methanol,
acetonitrile (both HPLC grades) and polyethylenglycol (PEG)
were purchased from Merck (Darmstadt, Germany). OchraTest
immunoaffinity cartridges were obtained from VICAM Cultivating
Success through Science®(Watertown, MA, USA) (Cartridge A),
OchraStarTMCOIAC 2000 was purchased in Romer Labs®Diag-
nostic GMBH (Tulln, Austria) (Cartridge B) and OchraPrep®was
provided by R. Biopharm Rhone Ltd. (Glasgow, Scotland) (Cartridge
C). C-18 Sep-Pak Vac cartridges of 500mg/3mL were obtained
from Waters (Milford, MA, USA) and Accu Bond II C-18 cartridges
of 500mg/3mL were provided by Agilent (California, USA). All
aqueous solutions were prepared in 18m? de-ionized water
from a Millipore Milli-Q water purification system (Bedford, MA,
2.3. HPLC method
The HPLC method was based on the official method of OIV
(Resolution OENO 16/2001) with some variation. In the present
research, five different HPLC columns were evaluated. The first
three were provided by Waters (Milford, MA, USA): (1) Symmetry®
C-18 of 5?m, 4.6mm×150mm (Column A); (2) YMC-Pack ODS-
A of 5?m, 4.6mm×150mm (Column B); (3) YMC-Pack ODS-A of
5?m, 4.6mm×250mm (Column C). The fourth, Onyx monolithic
C-18 column of 3mm×100mm (Column D), was obtained from
Phenomenex (Torrance, CA, USA) and the fifth, Chromolith®Per-
in Merck (Darmstadt, Germany). In all cases, a Symmetry®C-18
pre-column of 5?m, 3.9mm×20mm provided by Waters was
used. The mobile phase was isocratic at 1mLmin−1, constituted
by acetonitrile:formic acid:water (49.5:49.5:1) v/v. The detection
was by fluorescence at 333 and 460nm as excitation and emis-
were assayed; before injection, all samples were filtered through a
0.45?m membrane filter.
2.4. GC–MS method
Based on the method described by Soleas et al. , the
chromatographic separation was performed on a HP-5 MS col-
umn (5% phenyl/95% dimethylsiloxane copolymer-bonded) of
30mm×0.25mm I.D., 0.25?m film thickness and as mobile phase,
electronic helium, grade 6.0 (99.9999%), supplied by Air Liquide
energy of 70eV. The ionization source was set at 230◦C and ana-
lyzer temperature at 150◦C. The single ion monitoring (SIM) mode
was selected for qualitative analysis and the following mass frag-
ments were monitored for OTA: 528, 529, 530, 531, 532, 604, 606
and 619m/z. The initial temperature was 150◦C, with an increase
of 8◦Cmin−1up to 220◦C (ramp 1), 25◦Cmin−1up to 250◦C (ramp
2), and a final clean-up at 290◦C for 5min. The injection port tem-
perature was 260◦C and the injection volume was 2?L using the
2.5. Sample pre-treatments
2.5.1. SPE procedures using C-18 cartridges
ously by two different authors [20–22] were evaluated. In all cases,
the cartridges were previously washed with 5mL of methanol fol-
lowed by 5mL of water. The extracts were evaporated at room
C. Tessini et al. / Analytica Chimica Acta 660 (2010) 119–126
temperature under nitrogen stream and re-dissolved in 500?L of
Procedure A: 10mL of wine and 10mL of 1% PEG–5% NaHCO3
solution were mixed, 5mL of this solution were pre-concentrated
in the C-18 cartridge, followed by a wash step with 5mL of water.
acid 99.5:0.5 v/v .
Procedure B: As procedure A, but using 2mL of acetonitrile as
extraction solvent .
Procedure C: 5mL of wine were diluted with 5mL of water and
pre-concentrated on the C-18 cartridge. A washing step using 2mL
of water followed by 2mL of methanol:water solution (60:40%
v/v) was used. The extraction was carried out also with 2mL of
methanol/glacial acetic acid 99.5:0.5 v/v .
2.5.2. Immunoaffinity cartridge pre-treatment
The pre-treatment was carried out following the method
injection volumes were evaluated in order to find the best compro-
mise between chromatographic efficiency and sensitivity .
2.5.3. Direct injection without pre-concentration and clean-up
Based on a method recently published by Tafuri et al. ,
the direct injection on the HPLC without any clean-up or pre-
concentration step was tested using monolithic and conventional
packed C-18 HPLC columns under the same separation conditions
of the OIV method .
2.5.4. Extraction and derivatization for GC–MS analysis
An aliquot of 200?L of wine were extracted twice using 500?L
of dichloromethane followed by vortex homogenization for 3min.
The organic phase was separated by centrifugation and evapo-
rated under nitrogen stream. The dry extract was derivatized using
BSTFA:ethyl acetate solution 50:50% v/v at 70◦C over 2h in a ther-
mostatic bath. Two microliters of the product were injected by
splitless mode to the GC–MS .
2.6. Wine samples
Commercial wine samples were purchased in retail. Samples
produced in Chile, Argentina and France were used in this study.
The wines were opened, filtered and they were subjected to the
different treatments described before. For spiked samples, differ-
ent volumes of OTA standard solution dissolved in methanol were
previously added to the wine samples.
3. Results and discussion
3.1. Optimization and internal validation of reference method
3.1.1. HPLC columns
The OIV method specifies that the chromatographic separa-
tion of OTA in wine must be carried out using a C-18 column of
150mm×4.6mm and packing material of 5?m particle size. In
Fig. 1. Chromatograms obtained for 2.0?gL−1of OTA standard using different columns: Symmetry®C-18 column of 5?m, 4.6mm×150mm (Column A); YMC-Pack ODS-A
column of 5?m, 4.6mm×150mm (Column B); YMC-Pack ODS-A column of 5?m, 4.6mm×250mm (Column C); Onyx monolithic C-18 column of 3mm×100mm (Column
D); Chromolith®Performance RP-18e column of 3mm×100mm (Column E) purchased from Merck (Darmstadt, Germany).
C. Tessini et al. / Analytica Chimica Acta 660 (2010) 119–126
the present work, different C-18 columns were tested, including
2 columns with this dimension, as well as a longer one (25cm)
of the same manufacturer and packing material and 2 monolithic
less back pressure and allowing higher flow rates, reducing time
analysis. Fig. 1 presents the chromatograms obtained for 2.0?gL−1
of OTA standard injected on different columns. The Columns A and
B have the same length and particle size, but Column B duplicates
the chromatographic efficiency for OTA, measured as HEPT (HEPT=
0.001mm) and resolution than Column A (HEPT=0.002mm)
although the retention time for OTA on Column B is 30% longer.
The 25-cm column with the same packing material and manufac-
turer (Column C) shows higher efficiency (HEPT=0.0005mm), and
resolution (Rs>2) but longer retention time for OTA (12.3min).
In the case of monolithic Columns D and E, using the same chro-
matographic conditions, their retention time was shorter (1.5min)
but was combined with a significantly lower chromatographic effi-
ciency (HEPT=0.008mm), decreasing its resolution and increasing
the risk of interferences and false positives. In order to evaluate
this, a red wine sample spiked with 2.00?gL−1of OTA was pre-
treated in IAC and the separation using a packed C-18 column and
a monolithic column was compared. The results presented in Fig. 2
show that the monolithic column provides a very fast chromatog-
raphy, although the OTA peak appears mounted on the shoulder of
the interfering matrix peak and with a low resolution. It has to be
kept in mind that this chromatogram was obtained spiking with
the maximal concentration of OTA allowed in wine and that the
interference is more critical at lower OTA concentrations, such as
those commonly found in real wine samples when OTA is detected.
3.1.2. Incidence of sample injection volume into HPLC
pre-treated sample. However, due to the different loading capacity
Fig. 2. Chromatographic separation of red wine added with 2.00?gL−1of OTA by
IAC–HPLC–FLD using: (A) Waters Symmetry®C-18, 4.6mm×150mm (Column A)
and (B) Onyx monolithic C-18, 3.0mm×100mm (Column D).
Volumes injection effect on chromatographic efficiency and sensitivity of the chro-
matographic separation of OTA.
Volumes (?L)Efficiency (N)LODa(?gL−1) Slope calibration curve
aCalculated considering the sy/xof calibration curve obtained at concentration
close to LOD level.
of each column, this parameter was evaluated, considering sen-
sitivity, detection limit and efficiency, in order to find the best
compromise between these analytic properties. These results are
summarized in Table 1. As expected, higher injection volumes
produce lower chromatographic efficiency, attributable to extra-
column peak broadening. This result is relatively independent of
OTA concentration because the concentration levels of this com-
pound in wine are low and therefore they do not generate a
saturation effect of the column with the consequent peak broad-
ening. Increasing the injection volume will increase sensitivity
(calibration curve slopes), but this effect is irrelevant at concen-
trations close to the detection limit because this parameter is more
affected by the noise than by the injection volume. Considering
these results, the optimized injection volume was 50?L of sample,
which was different from the injection volume specified in the OIV
3.1.3. Selection of IAC
OTA pre-concentration and cleanup of the wine sample; however
there are at least 3 different IAC cartridge providers. A statistical
analysis of the recovery using the 3 types of IAC is presented in
Table 2. Recovery percentages of wine spiked with 1.98?gL−1of
OTA using the 3 different IAC columns were determined. For this
of these results were demonstrated (˛=0.05) by comparison of
mean values using a one-way ANOVA (p<0.05) test. The results
showed higher recoveries using the IAC B in comparison with IAC
A and IAC C.
3.1.4. Analytical parameters of IAC–HPLC–FLD method
The analytic method was validated using IAC B for sample
pre-treatment, 50?L of injection volume, and HPLC Column A.
Previously analyzed OTA-free red wine was spiked in a concen-
tration range of 0.04–3.00?gL−1of OTA. Each spiked wine sample
was processed in triplicate by the overall method, including the
IAC pre-treatment and HPLC analysis. Each injection was car-
ried out in triplicate. The limit of detection (LD) was determined
using the method described by Miller et al. . The analytical
parameters of the method are presented in Table 3. Their exac-
titude was also evaluated using non-contaminated wine samples
spiked with different OTA concentrations. Our study also consid-
ered concentration levels below those specified in the EU 96/23/CE
directive  because OTA in wine is generally detected in that
OTA recovery from red wine using different IAC types.
Cartridge Analysis 1
68.7 ± 0.3
95.3 ± 0.7
82.7 ± 0.5
75.5 ± 0.3
87.0 ± 0.3
84.7 ± 1.3
60.3 ± 0.3
88.5 ± 0.9
78.2 ± 0.9
68 ± 7a
90 ± 4a
81 ± 3a
aStandard deviation between extractions.
C. Tessini et al. / Analytica Chimica Acta 660 (2010) 119–126
Analytical parameters of OTA determination in red wine by IAC–HPLC–FLD and C-
Linear regression modelIACC-18
Correlation coefficient (r)
Determination coefficient (R2)
Limit of detection (?gL−1)
Limit of quantification (?gL−1)
Number of observations
concentration range. The samples were analyzed and results are
summarized in Table 4. They show that the method is suitable
for quantitative OTA determination in wine in a concentration
range between 0.04 and 3.00?gL−1. The results are in agree-
ment with the results of the collaborative trial presented in
the Resolution OENO 16/2001 about OTA determination in wine
. This method was used as reference method for the present
3.2. Evaluation of sample pre-treatment method with SPE C-18
cartridges and HPLC–FLD
3.2.1. Statistic comparison of different extraction procedures
using C-18 cartridges
Based on the described SPE methods for OTA pre-treatment
with C-18 cartridges, the recoveries for red wine spiked with
2.00?gL−1of OTA were statistically analyzed. Each described
tively 77.7±1.4%, 89.4±0.6% and 93.2±0.4% for procedures A, B
and C. The statistic difference between recoveries was demon-
strated (˛=0.05) using one-way ANOVA test, showing p value
below 0.005. According to these results, the best extraction alter-
native using SPE was procedure C.
3.2.2. Optimization and analytical parameters of the
The method using sample pre-treatment on C-18 cartridges
was optimized with red wine spiked with 0.2?gL−1of OTA using
procedure C and HPLC analysis. Different injection volumes were
studied showing that injection of 20 or 50?L of spiked wine, pro-
duced the complete separation of OTA peak from the continuous
peaks present in the chromatogram (Rs=1.98), however this res-
olution was reduced when 90 or 100?L were injected (Rs=0.8),
attributable to extra-column peak broadening. Considering these
results, when using the SPE with C-18 cartridge, the optimal injec-
tive IAC; hence a greater chromatographic resolution is required to
OTA recovery from red wine by the optimized IAC–HPLC–FLD.
aNot detected. Each extraction was also injected by triplicate.
OTA recovery from red wine by SPE C-18–HPLC–FL.
prevent interferences causing false positives. The method’s analyt-
ical parameters are presented in Table 3.
The exactitude of the C-18 method was determined in the same
way as for the reference method (IAC) in order to obtain the ana-
lytical parameters for the statistical comparison of both methods.
Table 5 presents OTA recovery from red wine spiked with different
OTA concentrations treated by procedure C and HPLC–FLD. It can
be observed that the recoveries for all the studied concentrations
were higher than 90%, with repeatability within the range fixed by
the European Directive .
3.3. Statistic comparison of C-18 method with reference method
with and without addition of OTA. The comparison of the analyti-
cal parameters is presented in Table 6. A significant difference was
its obtained by the two methodologies, where the most favorable
by the higher selectivity obtained with the IAC pre-treatment. Still,
determination in wine according to the European Directives .
Considering that the detection limit (0.09?gL−1) was less than
10% of the maximal allowed concentration in wine (2.00?gL−1)
and that the cost of a C-18 cartridge is between 20% and 25% of
an IAC device, this alternative would be very useful for broad OTA
detection programs for national or international coverage.
Both methods were applied to 154 red wine samples. Special
emphasis was put in the analysis of non-contaminated samples
in order to evaluate the possibility of false positives with the C-
Comparison of analytical parameters of IAC and C-18 sample pre-treatments and
Analytical parameters Pre-treatment
% Recovery (red wine)
Significant difference were obtained for values with different superscript (˛=0.05).
C. Tessini et al. / Analytica Chimica Acta 660 (2010) 119–126
Fig. 3. GC–MS chromatogram (A) and MS spectrum (B) of a wine sample with 0.05?gL−1of OTA.
18 method due to its lower selectivity. With the IAC method and
SPE C-18 method, 127 samples were below LD, and the SPE C-
18 method presented total absence of false positives. In the case
of samples where traces or quantifiable levels of OTA were found
(27 wine samples corresponding to spiked or to naturally contam-
inated samples), the results were statistically analyzed with the
t-test for pair-wise samples (˛=0.05), finding no significant dif-
ference (p=0.08). These 27 samples, where OTA was detected or
quantified, were also submitted to a confirmatory analysis using
the GC–MS method. In all samples where a peak signal attributed
to OTA was detected by application of C-18 method, its presence
was confirmed by GC–MS. Fig. 3A presents the chromatogram of a
wine sample where OTA was detected and Fig. 3B shows the mass
spectrum for this sample’s OTA peak.
3.4. Evaluation of direct injection of wine using different
In order to find an easy screening method, based on a recent
publication , the determination of OTA using HPLC monolithic
columns and conventional packed columns, without IAC or SPE
sample pre-treatment, was evaluated for real wine samples. Fig. 4
presents the chromatograms of a red wine sample spiked with
2.00?gL−1of OTA and injected directly after filtration into the
HPLC. As can be observed, the separation of the OTA signal using
the monolithic column is very poor, whereas better results are
The evaluation of the direct determination of OTA using differ-
ent columns from the quantitative point of view is summarized in
Table 7. The results show very unfavorable detection limits when
monolithic columns and conventional packed columns of 150mm
concentration in wine. It is important to consider that a longer col-
umn provides greater resolution to separate OTA signal from the
wine matrix interference and can better tolerate a larger injection
volume. These conditions contribute to more favorable detection
limits; however, the results show that it was not useful at the OTA
concentration levels usually found in wine, at least if fluorescence
detection is used. It is possible that with increased selectivity of
MS or MS/MS detection, this limitation can be overcome, although
the MS/MS detection is accompanied by much higher acquisition
and operation costs. Consequently, its use in a broad OTA detection
program in wine is more complex, although it could be reserved as
a confirmatory tool for positive samples.
C. Tessini et al. / Analytica Chimica Acta 660 (2010) 119–126
Analytical and chromatographic parameters obtained with different HPLC columns used for OTA determination in wine without sample pre-treatment.
9.04 ± 0.03
10.16 ± 0.04
9.63 ± 0.02
8.99 ± 0.02
23.72 ± 0.12
−2.90 ± 0.93
−0.49 ± 0.11
−2.40 ± 0.15
−0.77 ± 0.09
−2.60 ± 0.16
In all cases the injection volume was 20?L, except C* for which it was 50?L.
m: mVsL?g−1, y: mV, x: ?gL−1, R2: determination coefficient.
injected directly into the HPLC without sample treatment. (A) Onyx mono-
lithic C-18, 3.0mm×100mm column (Column D) and (B) YMC-Pack ODS-AM,
4.6mm×250mm column (Column C).
4. Chromatogramofwinesamples addedwith 2.00?gL−1
Several methods have been described for OTA determination
in wine using HPLC–FLD, where the IAC treatment is an adequate
alternative due to its high selectivity and sensitivity but with a
very high cost. The choice of a high efficiency HPLC column and
an optimal injection volume are essential to achieve adequate ana-
lytical parameters. Other more economic alternatives for sample
pre-treatment, such as SPE using a less selective C-18 phase, show
appropriate analytical parameters making them feasible for use in
broad OTA detection programs in wine. No significant differences
are observed in comparison with IAC sample pre-treatment, espe-
cially when both methods are systematically applied to a great
number of wine samples. Using this optimized C-18 method, no
false positives are observed for red wines. The absence of false pos-
itives using C-18 cartridges was confirmed by GC–MS. Considering
cedure opens an interesting alternative for broad OTA detection
programs, where positive samples can be confirmed by GC–MS or
HPLC–MS, if necessary. In this context, the combination of C-18
cartridges with conventional particle packed columns is presently
the most efficient option. If in the next years, new ultra high reso-
lution HPLC instruments become more widespread, the proposed
methodology can be transferred with minor adaptations.
Sample pre-treatment is essential for HPLC–FLD OTA analy-
sis at levels usually found in wine. In this context, based on our
results, the use of a method without sample pre-treatment is not
recommended due to the lack of resolution with interfering com-
pounds present in red wine, especially when using a monolithic
HPLC column, with lower chromatographic efficiency. These inter-
fering compounds should to be previously removed by IAC or
C-18-cartridges in order to obtain appropriate detection limits and
avoid false positives.
The authors thank the financial support received from Servi-
cio Agrícola y Ganadero de Chile, Fondo del Patrimonio Sanitario
(SAG Grant C4-94-14-31) and the Consorcio Tecnológico Empre-
sarial para la Vid y el Vino (Vinnova), whose members are Innova
and Universidad de Concepción. The authors also thank the winer-
ies that provided samples as well as the Dirección de Investigación
and Dirección de Postgrado of Universidad de Concepción for their
financial support, respectively for the participation of C. Mardones
in In Vino Analytica Scientia 2009 and for the postgraduate fel-
lowship of C. Tessini in the Programa de Doctorado en Ciencias y
Tecnología Analítica, Universidad de Concepción.
 A. Visconti, G. Perrone, G. Cozzi, M. Solfrizzo, Managing ochratoxin A risk in the
grape-wine food chain, Food Addit. Contam. 25 (2008) 193–202.
 C. Juan, C.M. Lino, A. Pena, J.C. Moltó, J. Ma˜ nes, I. Silveira, Determination of
73 (2007) 246–250.
 N. Bellí, S. Marín, I. Coronas, V. Sanchis, A.J. Ramos, Skin damage, high temper-
ature and relative humidity as detrimental factors for Aspergillus carbonarius
infection and ochratoxin A production in grapes, Food Control 18 (2007)
 A.E. Pohland, S. Nesheim, L. Friedman, Ochratoxin A: a review, Pure Appl.
Chem. 64 (1992) 1029–1046, http://news.reseau-concept.net/images/oiv uk/
Client/Resolution Oeno EN 2001 16.pdf.
 Commission Regulation (EC) No. 123/2005 of 26 January 2005 amending Reg-
ulation (EC) No. 466/2001 as regards ochratoxin A.
 Resolution OENO 16/2001 determination of Ochratoxin A by column of
immuno-affinity, Organisation Internationale de la Vigne et du Vin, France.
 B. Zimmerli, R. Dick, Determination of ochratoxin A at the ppt level in
human blood, serum, milk and some foodstuffs by high-performance liquid
chromatography with enhanced fluorescence detection and immunoaffinity
column cleanup methodology and Swiss data, J. Chromatogr. B 666 (1995)
 R. Serra, C. Mendonc ¸a, L. Abrunhosa, A. Pietri, A. Venancio, Determination of
ochratoxin A in wine grapes: comparison of extraction procedures and method
validation, Anal. Chim. Acta 513 (2004) 41–47.
 K.Mu˜ noz,M.Vega,G.Rios,S.Mu˜ noz,R.Madariaga,PreliminarystudyofOchra-
toxin A in human plasma in agricultural zones of Chile and its relation to food
consumption, Food Chem. Toxicol. 44 (2006) 1884–1889.
 L. Monaci, F. Palmisano, R. Matrella, G. Tantillo, Determination of ochratoxin
A at part-per-trillion level in Italian salami by immunoaffinity clean-up and
matogr. A 1090 (2005) 184–187.
C. Tessini et al. / Analytica Chimica Acta 660 (2010) 119–126
 A. Visconti, M. Pascale, G. Centoz, Determination of ochratoxin A in wine by
means of immunoaffinity column clean-up and high-performance liquid chro-
matography, J. Chromatogr. A 864 (1999) 89–101.
 E. Beltrán, M. Ibá˜ nez, J. Vicente Sancho, F. Hernández, Determination of
mycotoxins in different food commodities by ultra-high-pressure liquid chro-
matography coupled to triple quadrupole mass spectrometry, Rapid Commun.
Mass Spectrom. 23 (2009) 1801–1809.
 J.D. Di Mavungu, S. Monbaliu, M.L. Scippo, M. Scippo, G. Maghuin-Rogister,
Y. Schneider, Y. Larondelle, A. Callebaut, J. Robbens, C. Van Peteghem, S. De
Saeger, LC–MS/MS multi-analyte method for mycotoxin determination in food
supplements, Food Addit. Contam. Part A 26 (2009) 885–895.
toxin A in wine, Anal. Chim. Acta 453 (2002) 33–41.
A in wine at sub ng/mL levels by solid-phase microextraction coupled to liquid
chromatography with fluorescence detection, J. Chromatogr. A 1115 (2006)
 E. González-Pe˜ nas, C. Leache, A. López de Cerain, E. Lizarraga, Comparison
between capillary electrophoresis and HPLC-FL for ochratoxin A quantification
in wine, Food Chem. 97 (2006) 349–354.
 M. Valcárcel, L. Arce, A. Rios, Coupling continuous separation techniques to
capillary electrophoresis, J. Chromatogr. A 924 (2001) 3–30.
 R. Pe˜ na, M.C. Alcaraz, L. Arce, A. Ríos, M. Valcárcel, Screening of aflatoxins
in feed samples using a flow system coupled to capillary electrophoresis, J.
Chromatogr. A 967 (2002) 303–314.
 S. Almeda, L. Arce, M. Valcarcel, Combined use of supported liquid membrane
and solid-phase extraction to enhance selectivity and sensitivity in capillary
electrophoresis for the determination of ochratoxin A in wine, Electrophoresis
29 (2008) 1573–1581.
 M.J. Hernández, M. García-Moreno, E. Durán, D. Guillén, C. Barroso, Valida-
tion of two analytical methods for the determination of ochratoxin A by
cence detection in must and sweet wines from Andalucía, Anal. Chim. Acta 566
 G.J. Soleas, J. Yan, D.M. Goldberg, Assay of ochratoxin A in wine and beer
by high-pressure liquid chromatography photodiode array and gas chro-
matography mass selective detection, J. Agric. Food Chem. 49 (2001) 2733–
 J.M. Sáez, A. Medina, J.V. Gimeno-Adelantado, R. Mateo, M. Jiménez, Com-
parison of different sample treatments for the analysis of ochratoxin A in
must, wine and beer by liquid chromatography, J. Chromatogr. A 1029 (2004)
 A. Tafuri, G. Meca, A. Ritiene, A rapid high-performance liquid chromatogra-
phy with fluorescence detection method developed to analyze Ochratoxin A in
wine, J. Food Protect. 71 (2008) 2133–2137.
Prentice Hall, Madrid, 2000, pp. 125–128.
 Implementing Council Directive 96/23/EC concerning the performance
of analytical methods and the interpretation of results, 2002/657/EC.
 Commission Regulation (EC) No. 401/2006, laying down the methods of
sampling and analysis for the official control of the levels of mycotox-
ins in foodstuffs. http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=OJ:L: