Food

Chemistry
Food Chemistry 97 (2006) 555–562
www.elsevier.com/locate/foodchem

Analytical, Nutritional and Clinical Methods

Determination of acrylamide in potato chips by a reversed-phase

LC–MS method based on a stable isotope dilution assay
José A. Rufián-Henares, Francisco J. Morales *

Consejo Superior de Investigaciones Cientı́ficas, Instituto del Frı́o, José Antonio Nováis 10, E-28040 Madrid, Spain

Received 21 February 2005; received in revised form 10 May 2005; accepted 11 June 2005

Abstract

Potato-based products represent an important part of the daily intake of food-derived acrylamide, mainly on adolescent
population from western countries. A reversed-phase liquid chromatography-mass spectrometry based on a stable isotope dilution
assay was used for acrylamide analysis. Aqueous sample extraction, cleaning with Carrez solution and solid phase extraction with
methanol was applied. The ratio potato/NaCl solution is critical during extraction where the optimum ratio is 0.125 g/ml NaCl 2 M
solution. The use of virgin olive oil, as retaining matrix, during methanol desiccation was critical to achieve high recoveries. The
method performance was validated for limit of detection (23.2 lg/kg) and quantitation (91.8 lg/kg), linearity (r > 0.999,
25–1000 lg/kg), recovery (98.8%). The method was applied on commercial potato chips; the intra-day repeatability was set at
3.9% and values were corrected with a labeled internal standard (13C3-acrylamide). No significant differences on the acrylamide
content were observed between industrial-scale and local-scale processed potato chips.

2005 Elsevier Ltd. All rights reserved.

Keywords: Acrylamide; Potato chips; Stable isotope dilution analysis; LC–MS

1. Introduction tional Food Administration, 2002; Tareke, Rydberg,

Karlsson, Eriksson, & Tornqvist, 2002). Besides, Mot-
Acrylamide is a useful industrial chemical that was la- tram and Wedzicha (2002) showed how acrylamide
beled as a probable human carcinogen by the Interna- could be formed from food components during heat
tional Agency for Research on Cancer (IARC). In this treatment as a result of the Maillard reaction, likely
way, the contamination of drinking water or plants asparagine and glucose. Asparagine, a major amino acid
grown hydroponically has been the driving force in the in potatoes and cereals, is a crucial participant in the
past to develop methods for acrylamide monomer anal- formation of acrylamide by Maillard reaction at temper-
ysis (Castle, Campos, & Gilbert, 1991; Hashimoto, atures above 100 C (Friedman, 2003). Since potato
1976). Things changed recently, when in April 2002, products are especially rich in asparagine and reducing
researchers from the University of Stockholm and the sugars, it is now thought that this Maillard reaction is
Swedish National Food Administration (NFA) reported most likely responsible for the majority of the acrylam-
the presence of acrylamide in a wide range of fried and ide found in potato chips and French fries.
oven-cooked foods, most notably in potato chips and The potential health risk of food acrylamide has been
French fries, at levels of 224–3700 lg/kg (Swedish Na- considered by number of government agencies and
national authorities (Food Standards Agency, 2002;
Scientific Committee on Food, 2002). Subsequently,
*
Corresponding author. Tel.: +34 91 5492300; fax: +34 91 5493627. all available data on acrylamide have been reviewed
E-mail address: [email protected] (F.J. Morales). at international level, e.g., FAO, WHO, JIFSAN

0308-8146/$ - see front matter
2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.foodchem.2005.06.007
556 J.A. Rufián-Henares, F.J. Morales / Food Chemistry 97 (2006) 555–562

Workshop- identifying and listing a number of research potato chips as a major source of acrylamide on the diet.
gaps and priorities (FAO/WHO consultation, 2002; JIF- In this way, a robust and sensitive LC–MS method was
SAN, 2002). Among these, the development and valida- used taking into account different approaches described
tion of sensitive and reliable analytical methods for the previously in the literature. Effectiveness of the proce-
quantification of acrylamide in different food matrices dure was evaluated and applied on the study of the con-
was considered as essential (Pittet, Périsset, & Oberson, tent of acrylamide on potato chips.
2004).
Because of its high water solubility and high reactiv-
ity (Mottram & Wedzicha, 2002) and also, because of 2. Materials and methods
the lack of a chromopher group, acrylamide is not easy
to detect (Jezussek & Schieberle, 2003). At present, sev- 2.1. Samples
eral analytical methods are available for determining
acrylamide in foods and the majority are classical meth- Experiments were conducted with a series of commer-
ods based on high performance liquid chromatography cial potato chips (39 brands from 34 producers) ran-
(LC) or gas chromatography (GC) techniques (Andrze- domly purchased on different supermarkets (n = 27)
jewski, Roach, Gay, & Musser, 2004; Barber, Hunt, and fried-potato shops (n = 12). Samples (400–800 g)
LoPachin, & Ehrlich, 2001; Bologna, Andrews, Barve- were thinly sliced to assure a homogeneous distribution
nik, Lentz, & Sojka, 1999; Castle, 1993; Jezussek & Schi- of hot-spots. A portion (200 g) was distributed in four
eberle, 2003; Kawata et al., 2001; Pittet et al., 2004; containers and stored under vacuum and light protected
Tareke, Rydberg, Karlsson, Eriksson, & Törnvqist, at 4 C until analysis.
2000; Tekel, Farkas, & Kovác, 1989; US EPA, 1996).
To increase the selectivity and also the sensitivity in 2.2. Chemicals and materials
GC analysis, bromination of the double bond in combi-
nation with electron capture detection was previously [13C3]-acrylamide (isotopic purity 99%) was from Cam-
applied (Hashimoto, 1976); later, this method was bridge Isotope Labs (Andover, MA, USA). Acrylamide
improved by using either methacrylamide or N,N-dim- (99%), potassium ferrocianide (Carrez I), zinc acetate
ethylacrylamide as internal standards and mass-spec- (Carrez II)-both analytical-reagent grade- and sodium
trometry (MS) as detection method (Ahn et al., 2002; chloride were from Sigma–Aldrich (St. Louis, MO,
Jung, Choi, & Ju, 2003; Tareke et al., 2000; Tareke USA). Acetic acid (ultrapure grade) was from Merck
et al., 2002) However, several groups also described (Darmstadt, Germany). Methanol and acetonitrile
methods to quantify acrylamide by direct GC–MS mea- (HPLC grade) were from Scharlau (Barcelona, Spain).
surements without bromination where the loss of acryl- The solid-phase extraction (SPE) cartridges Isolute
amide at the injection port should be evaluated Multimode (500 mg, 3 ml) were from IST (Hewgoed,
(Biedermann, Biedermann-Brem, Noti, & Grob, 2002; Mid-Glamorgan, UK), reversed-phase Oasis HLB
Clarke, Kelly, & Wilson, 2002). (200 mg, 6 ml) and mixed mode cation exchange car-
In a recent assessment of the performance of 37 labo- tridge Oasis MCX (60 mg, 3 ml) were from Waters (Mil-
ratories in determining acrylamide in crisp bread, Clarke ford, MA, USA). Syringe filter units (0.45 lm, nylon)
et al. (2002) reported that the majority of laboratories were purchased from Tecknokroma (Madrid, Spain).
use either GC–MS or LC–MS with no obvious
method-dependent differences in results obtained be- 2.3. Standard and reagents
tween the two approaches. However, the main advantage
on LC–MS based methods is that acrylamide can be ana- Stock solutions of acrylamide (0.01 mg/ml) and
lyzed without prior derivatization, which considerably [13C3]-acrylamide (5 lg/ml) were prepared by dissolving
simplifies and expedites the analysis (Joint European the compounds in Milli-Q water and methanol, respec-
Commission & Swedish National Administration, tively. These solutions were then appropriately diluted
2003). Many laboratories has developed its ‘‘own’’ LC– with Milli-Q water (Millipore Corp., Madrid, Spain)
MS or LC–MS/MS technique (Andrzejewski et al., to prepare working standards at 1.0 lg/ml. All stock
2004; Ahn et al., 2002; Becalski, Lau, Lewis, & Seaman, solutions and working standards were stored light-
2003; Jezussek & Schieberle, 2003; Roach, Andrzejewski, protected in a refrigerator at 4 C for maximum 3
Gay, Nortrup, & Musser, 2003; Tareke et al., 2002) but months. New working standards were compared with
one of the limitations of these techniques is the difficulty previous one as control of quality. Daily the instrument
of applying an universal clean-up approach valid for performance and relative response of labeled (m/z 75.1)
many different food matrices that avoid interferences and unlabeled (m/z 72.1) acrylamide was verified. Carrez
from co-extractives (Presentation at HPLC, 2003). I was prepared by dissolving 15 g of potassium ferrocia-
The purpose of the following investigation was, there- nide in 100 ml of water and Carrez II by dissolving 30 g
fore, to determine the acrylamide content of commercial of zinc acetate in 100 ml of water.
 J.A. Rufián-Henares, F.J. Morales / Food Chemistry 97 (2006) 555–562 557

2.4. Sample extraction coupled to an Agilent Quadrupole mass spectrometer

(Agilent Technologies, Palo Alto, CA, USA). Analytical
A portion of the sample (1.0 g) was weighed with a separation was achieved with an Extrasyl ODS1
precision of 0.1 mg and suspended with 8 ml of sodium (20 · 0.3 cm, 5 lm; Tecknokroma, Madrid, Spain) at
chloride 2 M in polypropylene 15 ml centrifugal tubes. 32 C. Isocratic elution was achieved with a mobile
Then, the sample was spiked with 200 ll of a 5000 lg/ phase of acetic acid–methanol–Milli-Q water
ml [13C3]-acrylamide solution as internal standard and (0.1:1.0:98.9) at a flow rate of 0.4 ml min 1.
homogenized using a tube shaker. Acrylamide extrac- Electrospray ionization in the positive ionization
tion was performed by incubation in a water bath at mode was used. The MS detector operated in selected
60 ± 0.1 C for 30 min, and 10 s shaking every 10 min. ion monitoring (SIM) mode at m/z ratios of 72.1 and
In order to clarify the solution, 1 ml of each Carrez I 75.1 for acrylamide and labeled [13C3]-acrylamide,
and Carrez II were added and finally the mixture respectively. Under these HPLC conditions, acrylamide
(10 ml) was centrifuged (9000g/10 min/4 C). In this eluted at 6.8 min. The needle and cone voltages were set
way, three layers were obtained: a thin fat layer at the at 3.0 kV and 100 V, respectively. Nitrogen was used as
top, the aqueous layer in the middle and the lower pre- nebulizer gas (12.0 l/h) and the source temperature was
cipitate layer. set at 300 C.
An aliquot (3 ml) of clarified aqueous layer was
promptly removed by pipette for SPE clean-up. The pip- 2.7. Quantitation
ette was inserted through the top oil layer avoiding the
bottom solids layer with the pipette tip. If the 3 ml ali- Acrylamide was quantified using a linear calibration
quot was stored at refrigeration previous SPE clean- function that was established with standard solutions
up, no additional precipitate was observed. of acrylamide and [13C3]-acrylamide dissolved in Milli-
Q water at the concentration levels 25, 50, 100, 200,
2.5. Sample clean up 300, 400, 500, 650, 800 and 1000 lg/l. These concentra-
tion values were within the range as encountered in the
In order to clean-up the aqueous extracts, two differ- sample extracts. A calibration graph was constructed
ent SPE cartridges were assayed. Isolute Multimode was by plotting peak area ratios against the corresponding
preconditioned with 1 ml of methanol and 2 · 2 ml of ratios of analyte amounts. Thus, the acrylamide contents
water. An aliquot of the clear supernatant (0.5 ml) were in sample extracts were calculated from the calibration
passed through the cartridge. Sample (1 ml) was loaded slope and intercept value, taking into account the recov-
onto the same column and collected by pressure-induced ery calculated by means of [13C3]-acrylamide slope. In or-
flow. der to perform a good quantification, the signal-to-noise
In other way, Oasis HLB and MCX cartridges were ratio of the LC–MS peak had to be grater than 3:1.
preconditioned with 2 ml of methanol and Oasis HLB
cartridge with another 2 ml of Milli-Q water. Then, 2.8. Statistical
1.5 ml of clear supernatant were loaded onto the HLB
cartridge and it was washed with 0.75 ml of Milli-Q Statgraphics plus v.2.0 (Statistical Graphics Corp.,
water. Finally, MCX cartridge was coupled to the Rockville, MD, USA) was used for statistical analysis
HLB one and acrylamide was eluted with 3 ml of meth- of data. All analyses were performed by duplicate from
anol, which were eluted through the MCX cartridge. two independent extractions at least.
The Isolute Multimode or Oasis extracts were col-
lected in rotary evaporator heart-flasks and added with
0.4 ml of commercial virgin Spanish olive oil as retain- 3. Results and discussion
ing matrix (other substances could also be used). Olive
oil was evaluated for free acrylamide content. Then, This work describes a quantitative analytical method
the samples were evaporated to dryness on a rotary for the determination of acrylamide in potato chips in
evaporator (60 C) or on a heated-block (40 C) under order to estimate the incidence of acrylamide in the con-
a stream of nitrogen. Finally, acrylamide was dissolved sumption of potato-based products (potato crisp and
in 1 ml of mobile phase and olive oil removed with hex- French fries) in the Spanish population. The proposed
ane (3 · 3 ml). The solution was filtered through a method is comparable to other different methods
0.45 lm filter into an amberlite LC–MS vial. described in the literature (Becalski et al., 2003; Bieder-
mann et al., 2002; Roach et al., 2003) but improvements
2.6. LC–MS analysis have been made particularly regarding sample extrac-
tion and clean-up, i.e., Carrez clarification, addition of
Sample extracts and calibration standards (20 ll) olive oil after SPE purification, etc. The most efficient
were analyzed on an Agilent 1100 liquid chromatograph approached described previously by different authors
558 J.A. Rufián-Henares, F.J. Morales / Food Chemistry 97 (2006) 555–562

are evaluated and combined for acrylamide determina- 3750

tion in potato-based foodstuffs. Moreover, isotopi-
cally-labeled [13C3]-acrylamide is added to the test
portion before extraction so as to keep control on the 3000
recoveries achieved and to keep track of possible losses

Acrylamide (µ g/kg)
occurring during the whole sample pre-treatment 2250
(extraction, clean-up and concentration).

3.1. Sample extraction 1500

Extraction of free acrylamide monomer from the

food is a critical step since acrylamide may be firmly en- 750
closed and not homogeneously distributed, e.g., in dark
solids of potato chips (Biedermann et al., 2002); concen- 0
trations are likely to be particularly high in this particles.
0 0.25 0.50 0.75 1.00 1.25 1.50
In this way, method development has to overcome these
Sample amount (g)
problems that spiking experiments are not suitable to
check extraction of enclosed material. Fig. 1. Influence of sample amount over acrylamide extraction with 8
Different approaches have been described for sample ml of sodium chloride (2 M). Expected acrylamide content as dotted
line.
extraction, such as orbital shaker, ultrasonic bath, accel-
erated solvent extraction, diastase, or freezing. To check
for effectiveness of the extraction, the same sample was tion at HPLC, 2003). In order to evaluate the effectivity
incubated at different extraction times (Table 1). It was of sodium chloride on the extraction of acrylamide in
observed that 30 min gave higher acrylamide extraction potato-based products, different amounts of sample
and that prolonged time did not exert additional advan- were extracted with the same volume of sodium chloride
tages. In addition, to know if some acrylamide was re- (Fig. 1). It could be seen that higher amounts of acryl-
tained in the residue the samples were extracted a amide were extracted when higher quantities of potato
second time by swollen again the residues from the fist chips were added until a limit was reached (with 1.00 g
extraction. Extraction was considered complete at of sample). Thus, a linear correlation between acrylam-
30 min because of the second extract contained less than ide content and added potato chip was observed from
10% of the acrylamide concentration of the first extract. 0.10 to 1.00 g of sample. The amount of sample added
Because of different authors (Castle, 1993; Rosen & was adjusted such that, on the one hand, blending re-
Hellenäs, 2002) found acrylamide losses related with sulted in a paste enabling easy mixing. When mixing re-
the extraction temperature, an experiment in which a sulted in an excessively stiff paste (from 1.25 to 1.50 g of
potato chip extract added with standard acrylamide potato chip), lower acrylamide recuperation was ob-
was heated at 60 or 80 C from 20 to 60 min was carried served, 81.4% and 68.5% for 1.25 and 1.50 g, respec-
out. No significant acrylamide loss was detected and the tively. On the other hand, the sample should not turn
extraction of acrylamide was not improved (data not as liquid that solids sediment. Then, 1.00 g of sample
shown). In addition, the effect of sonication over acryl- was selected as a reasonable amount of potato chip to
amide extraction was also evaluated. Times of 1, 2 or analyze per 8 ml of NaCl solution in the extraction step.
5 min did not induce significant acrylamide losses in After sample extraction, in order to obtain a clear
both acrylamide standard solution and potato sample supernatant, it was assayed an additional freeze–
extract. unfreeze cycle (Riediker & Stadler, 2003). After centrifu-
Some authors claimed by the use of sodium chloride gation at high speed and low temperature (4 C) it was
solution during the extraction of acrylamide (Presenta- observed a cloudy supernatant. In other experiment

Table 1
Effect of the sample extraction time in a water bath at 60 C on acrylamide analysisa
Extraction time (min) Acrylamide (lg/kg) 1st extraction Recovery (%) Acrylamide (lg/kg) 2nd extraction Recovery (%) Total recovery (%)
10 506 ± 48 45.0 602 ± 32 53.5 98.5
20 879 ± 31 78.1 254 ± 28 22.6 100.7
30 1129 ± 34 100.3 n.q.b – 100.3
45 1138 ± 28 101.2 n.q. – 101.2
60 1117 ± 41 99.3 n.q. – 99.3
a
Average values from two independent analysis.
b
Not quantifiable.
 J.A. Rufián-Henares, F.J. Morales / Food Chemistry 97 (2006) 555–562 559

Carrez was added to achieve phase separation, to pre- when oil was no present. Rotary evaporation showed
cipitate proteins and other high-molecular co-extractives lower acrylamide losses than nitrogen-heat drying (Ta-
such as starch. Because of a clear supernatant was ob- ble 2) and, in addition, acrylamide was lost in lower
tained, Carrez was chosen as clarification reagent and amounts when methanol was used as solvent. Solvent re-
the freeze–unfreeze cycle was avoided. moval is a critical step and must be performed under ro-
tary evaporation and oil addition. On the one hand,
3.2. Sample clean up vastly excessive time may still result in losses of acrylam-
ide, as well as incomplete methanol removal may disturb
Preliminary trials with the proposed method by Zy- acrylamide quantification. The whole process was timed
zak et al. (2003) showed the presence of some interfering (about 4 min) and the last step kept under technical
co-extractives on the LC–MS chromatograms. There- surveillance.
fore, limit of quantification was not acceptable Finally, although Isolute Multimode and Oasis
(100 ppb) due to ionic suppression effects. In order to showed similar clean-up capacity, it were selected the
partially overcome this problem, it was found necessary Oasis cartridges because of the lowest lost of acrylamide
to purify acrylamide extracts by solid-phase extraction during solvent removal and lowest rotary evaporation
(SPE). Two different types of cartridges were assayed, times. The removed oil by hexane extraction was ana-
Isolute Multimode and the mixed Oasis HLB + MCX lyzed for acrylamide content. Thus, 2 ml of water were
procedure. The characteristic features of the Multimode added to the hexane–oil mixture, shook for 1 min and
sorbent are hydrophobic interaction (presence of C18 water extract analyzed by LC–MS. It could be observed
functional groups), strong cationic (SCX) as well as an- no acrylamide presence.
ionic (SAX) exchange. On the other hand, Oasis HLB is
a C18 cartridge and Oasis MCX a SCX one. In this way, 3.3. LC–MS
it was minimized the interfering co-extractives load with
both types of SPE cartridges, showing ‘‘cleaner’’ chro- Different columns were assayed for acrylamide analy-
matograms and higher signal responses due to less ion sis. Classical ODS-2 analytical columns showed poor
suppression effects (data not shown). separation since acrylamide co-eluted with the chro-
The removal of both water and methanol solvent matographic front. In contrast, the ODS-2 column Syn-
coming from Isolute and Oasis cartridges, respectively, ergi Hydro-RP showed an excessive retention for
was assayed by heating at 40 C under a stream of nitro- acrylamide (12 min). Because of that, different ODS-1
gen or by rotary evaporation at 60 C. Evaporation is a columns (partially deactivated) were assayed, founding
critical step because of acrylamide losses. As shown by excellent acrylamide separation (retention times ranged
Biedermann et al. (2002) acrylamide is largely lost if sol- from 6 to 8 min).
vents are completely evaporated without a residue to re- The first mobile phase assayed was the one used by
tain it. We agree with this statement since different Zyzak et al. (2003) composed of 10 mM amonium ace-
evaporation steps were performed in absence of oil tate (pH 6) with 0.1% of each acetic and formic acid.
founding important losses of acrylamide meaning a dra- This mobile phase was compared with a second mobile
matic decrease of recoveries. Thus, olive oil was studied phase composed of methanol–water (1:99) added with
as retaining residue. Table 2 shows the results obtained formic or acetic acid. Mobile phase containing acetic
for different combinations of solvent removal with or acid significantly improves the response of acrylamide
without olive oil under acrylamide loss of a standard (m/z 72.1 H+) about 1.35-fold than formic acid, and
solution. The highest acrylamide losses were obtained 4.74 times higher than the mobile phase with ammonium
acetate. Acetic acid was selected as acidic modifier.
Table 2 In order to adjust the fragmentation level it was per-
Comparison of different solvent removal procedures on the acrylamide formed a FIA assay by injecting 20 ll of both acrylam-
recovery in potato chips extractsa ide standard (1000 lg/l) and potato sample extract and
Drying method Solvent Oil Recovery (%) changing the cone voltage from 25 to 125 V. The best
Rotary evaporation Methanolb – 36.7 ± 4.8 relation fragmentation-response was found at 100 V.
Yes 96.6 ± 1.1 The ions monitored for identification and quantifica-
Waterc – 2.6 ± 5.2 tion of both analyte and internal standard were
Yes 53.6 ± 3.1 [C3H5NO]+ = 72.1 and [13C3H5NO]+ = 75.1, respec-
N2/Heat Methanolb – 8.1 ± 4.8
tively (Fig. 2). Quantitation was performed by compar-
Yes 42.4 ± 2.9
Waterc – 30.3 ± 1.7 ison with a calibration curve (25–1000 lg/l) for
Yes 59.9 ± 4.3 acrylamide and [13C3]-acrylamide, both with correlation
a
Average values from two independent analysis. coefficients of 0.9978 and 0.9998, respectively. In every
b
Extract from Oasis clean-up procedure. assay labeled acrylamide content was taken into account
c
Extract from Isolute clean-up procedure. to correct for acrylamide recovery.
560 J.A. Rufián-Henares, F.J. Morales / Food Chemistry 97 (2006) 555–562

Fig. 2. Typical chromatogram profile of acrylamide (72.1 m/z) in a potato chip extract (1126 ± 44 lg/kg). Internal standard (75.1 m/z) concentration
in sample (100 lg/kg).

3.4. Method performance 3.5. Sample analysis

A control of analytical performance was imple- Thirty-nine classical potato chips from 34 producers,
mented by the use of a reference material that was used without flavor or species added were analyzed for acryl-
as control in each series of analysis. The reference mate- amide content. Acrylamide content ranged from 211 to
rial was a potato chip sample simultaneously analyzed 5492 lg/kg with an average value of 1484 lg/kg (Table
by Analycen (Lidköping, Sweden) in order to assess 4) and a median of 1180 lg/kg. A confidence interval
the reference value (575 ± 37 lg/kg). (95%) from 1132 lg/kg to 1836 lg/kg was obtained.
A recovery assay (Table 3) was performed by adding Similar values were obtained for other authors (Becalski
0.2, 0.4, 0.6, 0.8 and 1.0 mg of acrylamide/kg potato et al., 2003; Roach et al., 2003; Rosen & Hellenäs, 2002;
chip (each two independent determinations). It was ob- Presentation at HPLC, 2003; Tareke et al., 2002) who
served a mean recovery of 98.8 ± 3.4% and a intra-day found acrylamide levels ranged between 224 and
repeatability of 3.9%. A good coefficient of correlation 3700 lg/kg. The consumption of potato chips is Spain
of 0.9978 was achieved over the whole concentration was 1.6 kg/person/year, which gives a mean acrylamide
range (200–1000 lg/kg). intake of 6.5 lg/person/day; this is an important part of
The estimation of the detection limit (LoD) and the the daily intake of food-derived acrylamide and is of
quantitation limit (LoQ) were performed according to great importance since potato chips consumption is stea-
Jezussek and Schieberle (2003) over the recovery assay. dily expanding in Spain (a 28.4% increase since 1998)
LoD is the addition value referring to the 95% confi- (Ministerio de Agricultura, 2003). In this way, the con-
dence limit of the calibration line at the zero addition le- trol of potato chips manufacture would be of interest
vel. LoQ is the addition level that lowers the 95% to lower acrylamide formation and then to lower the
confidence limit of the addition level at LoD. In this mean acrylamide intake of the Spanish population.
way the LoD and LoQ were of 23.2 and 91.8 lg/kg. If The Spanish Food National Agency (AESA) has re-
LoD and LoQ were extrapolated from the signal- cently started to estimate the incidence of acrylamide
to-noise (S/N) ratios obtained for the responses in consumption by Spanish population in order to clarify
the ion m/z 72.1, LoD could be similar but LoQ, the risk assessment for this chemical contaminant.
taking an S/N ratio of 3:1, could be reduced until A more detailed study of sample distribution shows a
69.6 lg/kg. high variability in the acrylamide content where several

Table 3
Recovery assay from potato chip reference material spiked with acrylamidea
Acrylamide added (lg/kg) Theoretical acrylamide (lg/kg) Measured acrylamide (lg/kg) Mean recovery (%)
0 575 578 ± 13 100.6
200 775 749 ± 23 96.6
400 975 1007 ± 33 103.3
600 1175 1177 ± 11 100.1
800 1375 1363 ± 12 99.1
1000 1575 1471 ± 49 93.4
a
Average values from two independent analysis.
 J.A. Rufián-Henares, F.J. Morales / Food Chemistry 97 (2006) 555–562 561

Table 4
Statistical analysis of the acrylamide content (lg/kg) of different potato chip groupsa
Group Mean ± SD Median Maximum Minimum CI n
Total 1484 ± 1086 1180 5492 211 1132–1836 39
Commercial 1543 ± 1144 1214 5492 219 1099–1986 28
Artisanal fried-potatoes 1335 ± 955 1056 3693 211 694–1977 11
Oily 1781 ± 1507 1377 5492 219 823–2738 12
Non oily 1298 ± 844 1073 4810 211 964–1632 27
Light-protected 1645 ± 1185 1191 5492 219 695–5492 24
Non light-protected 1225 ± 883 1106 3693 211 737–1715 15
a
Average values from two independent analysis.

6000
6000
6000 16
• 1094
14
5000
5000
Acrylamide (µ g/kg)

5000 •
12

frequency
4000
4000
4000 • 10

3000
3000
3000 8

781 1719 6
2000
2000
2000 < 2500
1406 2031 4
1000
1000
1000 156 469
2344 2

0 0
potato chips 1 2 3 4 5 6 7 8 9

Fig. 3. Box-and-whisker plot (a) and frequency tabulation graph (b) from acrylamide content in Spanish potato chips. Midpoints are included for
each analysis (see text for more detail).

samples are clearly over-processed (outliners from 4. Conclusions

Fig. 3). A box-and-whisker plot was used since this
graphical presentation uses a non-parametric test. In This work describes the determination of acrylamide
addition, to gain more insight on the acrylamide distri- in commercial potato chips by a confirmatory and
bution a frequency tabulation graph was plotted. The quantitative analytical method. This procedure takes
samples were divided into 8 classes (interval range of into account and evaluates different analytical ap-
312.5 lg/kg) up to 2500 lg/kg and outliners were repre- proaches described previously in the literature (Ahn
sented in group 9 (upper than 2500 lg/kg). The group et al., 2002; Andrzejewski et al., 2004; Becalski et al.,
with the highest frequency (29.5%) showed an acrylam- 2003; Biedermann et al., 2002; Mottram & Wedzicha,
ide concentration ranged between 937.5 and 1250 lg/kg, 2002). The sample clean up is based in the removal
and 10 % of the samples contain more than 2500 lg/kg. of co-extractives from the aqueous extracts by SPE car-
The high variability of the data is mainly a result of the tridges. In addition, it must be underlined the impor-
variable raw material and processing conditions applied. tance of oil addition to avoid acrylamide losses
Extent of the thermal treatments reveals as the main fac- during solvent removal, as described by Biedermann
tor, and cooking oil (expressed as peroxide value) did et al. (2002). The analyte is well retained on the LC
not affect acrylamide levels, as well as soaking (Institute column and the chromatographic step takes only
for Reference Materials & Measurements, 2004). 10 min per measurement. In addition, our lab partici-
Samples were classified according to the type of pro- pated in an international proficiency test organized
cessing (industrial vs. local fried shops); presence or not by Institute for Reference Materials and Measurements
of oil drops on the container layer (oily vs. non-oily); (IRMM) on crisp bread, obtaining a good evaluation
and if the container was light-protected or not (Table of our method (Riediker & Stadler, 2003). These data
4). It was not observed statistically significant differences are useful to evaluate the dietary intake of acrylamide
(p < 0.05) between commercial vs. artisanal fried-potato for the Spanish population for risk assessment pur-
shops. The same differences were obtained for oily vs. poses. At present, the procedure it has been evaluated
non-oily potato chips and for container-light-protected in other food matrices to fullfil the acrylamide database
potato chips vs. non-light-protected. in the Spanish diet.
562 J.A. Rufián-Henares, F.J. Morales / Food Chemistry 97 (2006) 555–562

Acknowledgments Jezussek, M., & Schieberle, P. (2003). A new LC/MS method for the
quantitation of acrylamide based on a stable isotope dilution assay
and derivatization with 2-mercaptobenzoic acid. Comparison with
We thank M.A. Martinez and D. Gomez for techni- two GC/MS methods. Journal of Agricultural and Food Chemistry,
cal assistance, S. Eriksson and P. Karlson (AnalyCen, 51, 7866–7871.
Sweden) for analysis of the reference potato chip. This JIFSAN. (2002). Report of the analytical working group 9, Acrylam-
research was supported by the Autonomous Community ide in food workshop, 28–30 October, Chicago.
of Madrid (CAM) under project 07 G/0030/2003 and by Joint European Commission and Swedish National Administration.
(2003). Workshop on Analytical methods for acrylamide determi-
a postdoctoral grant from Consejerı́a de Innovación, nation in Food, Geel Belgium.
Ciencia y Tecnologı́a (Junta de Andalucı́a). Jung, M. Y., Choi, D. S., & Ju, J. W. (2003). A novel technique
for limitation of acrylamide formation in fried and baked
corn chips and french fries. Journal of Food Science, 68, 1287–
References 1290.
Kawata, K., Ibaraki, T., Tanabe, A., Yagoh, H., Shinoda, A., Suzuki,
Ahn, J. S., Castle, L., Clarke, D. B., Lloyd, A. S., Philo, M. R., & H., & Yasuhara, A. (2001). Gas chromatographic–mass spectro-
Speck, D. R. (2002). Verification of the findings of acrylamide in metric determination of hydrophilic compounds in environmental
heated foods. Food Additive and Contamination, 19, 1116–1124. water by solid-phase extraction with activated carbon fiber felt.
Andrzejewski, D., Roach, J. A. G., Gay, M. L., & Musser, S. M. Journal of Chromatography A, 911, 75–83.
(2004). Analysis of coffee for the presence of acrylamide by LC– Ministerio de Agricultura, Pesca y Alimentación. (2003). La alimen-
MS/MS. Journal of Agricultural and Food Chemistry, 52, tación en España, Madrid.
1996–2002. Mottram, D. S., & Wedzicha, B. L. (2002). Acrylamide is formed in the
Barber, D. S., Hunt, J., LoPachin, R. M., & Ehrlich, M. (2001). maillard reaction. Nature, 419, 448–449.
Determination of acrylamide and glycidamide in rat plasma by Pittet, A., Périsset, A., & Oberson, J. M. (2004). Trace level of
reversed-phase high performance liquid chromatography. Journal acrylamide in cereal-based foods by gas chromatography-
of Chromatography B, 758, 289–293. mass spectrometry. Journal of Chromatography A, 1035, 123–
Becalski, A., Lau, B. P., Lewis, D., & Seaman, S. W. (2003). 130.
Acrylamide in foods: occurrence, sources, and modeling. Journal Presentation at HPLC. (2003). Solid-Phase extraction and clean-up
of Agricultural and Food Chemistry, 51, 802–808. procedures for the LC/MS determination of acrylamide in fried
Biedermann, M., Biedermann-Brem, S., Noti, A., & Grob, B. (2002). potato products. Nice, France, 15th–19th June.
Two GC/MS methods for the analysis of acrylamide in foodstuff. Riediker, S., & Stadler, R. H. (2003). Analysis of acrylamide in food by
Mitteilungen aus Lebensmitteluntersuchung und Hygiene, 93, isotope-dilution liquid chromatography coupled with electrospray
638–652. ionization tandem mass spectrometry. Journal of Chromatography
Bologna, L. S., Andrews, F. F., Barvenik, F. W., Lentz, R. D., & A, 1020, 121–130.
Sojka, R. E. (1999). Analysis of residual acrylamide in field crops. Roach, J. A. G., Andrzejewski, D., Gay, M. L., Nortrup, D., &
Journal of Chromatographic Science, 37, 240–244. Musser, M. (2003). Rugged LC–MS/MS survey analysis for
Castle, L. (1993). Determination of acrylamide monomer in mush- acrylamide in foods. Journal of Agricultural and Food Chemistry,
rooms grown on polyacrylamide gel. Journal of Agricultural and 51, 7547–7554.
Food Chemistry, 41, 1261–1263. Rosen, J., & Hellena¨s, K. E. (2002). Analysis of acrylamide in cooked
Castle, L., Campos, M. J., & Gilbert, J. (1991). Determination of foods by liquid chromatography tandem mass spectrometry.
acrylamide monomer in hydroponically grown tomato fruits by Analyst, 127, 880–882.
capillary gas chromatography–mass spectrometry. Journal of the Scientific Committee on Food, (2002). Opinion of the SCF on new
Science of Food and Agriculture, 54, 549–555. findings regarding the presence of acrylamide in food, expressed on
Clarke, D. B., Kelly, J., & Wilson, L. A. (2002). Assessment of 3 July, www.europa.eu.int.
performance of laboratories in determining acrylamide in crisp- Swedish National Food Administration. (2002). Information about
bread. Journal of AOAC International, 85, 1370–1373. acrylamide in food, 24 April, http://www.slv.se.
FAO/WHO consultation. (2002). Health implications of acrylamide in Tareke, E., Rydberg, P., Karlsson, P., Eriksson, S., & Tornqvist, M.
food, Geneva, Switzerland, 25–27 June. (2002). Analysis of acrylamide, a carcinogen formed in heated
Food Standards Agency. (2002). Study on acrylamide in food. foodstuffs. Journal of Agricultural and Food Chemistry, 50,
Background information and research findings. Press Briefing 17 4998–5006.
May, www.food.gov.uk. Food, Geel, Belgium, 28–29 April. Tareke, E., Rydberg, P., Karlsson, P., Eriksson, S., & Törnvqist, M.
Friedman, M. (2003). Chemistry, biochemistry, and safety of acryl- (2000). Acrylamide: A cooking carcinogen? Chemical Research in
amide. A review. Journal of Agriculture and Food Chemistry, 51, Toxicology, 13, 517–522.
4504–4526. Tekel, J., Farkas, P., & Kovác, M. (1989). Determination of
Hashimoto, A. (1976). Improved method for the determination of acrylamide in sugar by capillary GLC with alkali flame-ionization
acrylamide monomer in water by means of GLC with electron detection. Food Additives and Contaminants, 6, 377–381.
capture detector. Analyst, 101, 932–938. US EPA. (1996). SW 846, Method 8032A, US Environmental
IARC. (1994). Monographs on the evaluation of carcinogenic risks to Protection Agency, Washington, DC.
humans. (Vol. 60, p. 389), Lyon: International Agency for Research Zyzak, D. V., Sanders, R. A., Stojanovic, M., Tallmadge, D. H.,
on Cancer. Eberhart, B. L., Ewald, D. K., Gruber, D. C., Morsch, T. R.,
Institute for Reference Materials and Measurements. (2004). Profi- Strothers, M. A., Rizzi, G. P., & Villagran, M. D. (2003).
ciency test of the Food Safety and Quality Unit (Prof. Dr. E. Acrylamide formation in heated foods. Journal of Agricultural
Anklam; Dr. T. Wedzel). and Food Chemistry, 51, 4782–4787.