Professional Documents
Culture Documents
Food Chemistry: M.S. Altaki, F.J. Santos, M.T. Galceran
Food Chemistry: M.S. Altaki, F.J. Santos, M.T. Galceran
Food Chemistry
journal homepage: www.elsevier.com/locate/foodchem
a r t i c l e i n f o a b s t r a c t
Article history: In this work, we evaluated the occurrence of furan in brews obtained from regular, decaffeinated, and
Received 19 June 2010 instant coffee and commercial packed capsules. For this purpose, a previously developed automated
Received in revised form 22 October 2010 headspace solid-phase microextraction method coupled to gas chromatography–mass spectrometry
Accepted 22 November 2010
(HS-SPME-GC–MS) was used. Initially, the influence of HS-SPME conditions on furan formation was eval-
Available online 5 December 2010
uated. In addition, the effect of roasting conditions (temperature and time) used for coffee beans on furan
formation was also studied. We found that low temperature and long roasting time (140 °C and 20 min)
Keywords:
decreases the final furan content. Furan concentrations in regular ground coffee brews from an espresso
Food analysis
Coffee
coffee machine were higher (43–146 ng/ml) than those obtained from a home drip coffee maker (20 and
Furan 78 ng/ml), while decaffeinated coffee brews from a home drip coffee maker (14–65 ng/ml) showed a
Headspace solid-phase microextraction furan concentration similar to that obtained from regular coffee. Relatively low concentrations of this
GC–MS compound (12–35 ng/ml) were found in instant coffee brews, while commercial packed coffee capsules
showed the highest concentrations (117–244 ng/ml). Finally, the daily intake of furan through coffee con-
sumption in Barcelona (Spain) (0.03–0.38 lg/kg of body weight) was estimated.
Ó 2010 Elsevier Ltd. All rights reserved.
0308-8146/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.foodchem.2010.11.134
1528 M.S. Altaki et al. / Food Chemistry 126 (2011) 1527–1532
has been reported during HS analysis of ground green coffee beans portions of the beans were roasted to the same degree under
at a temperature as low as 40 °C (Senyuva & Goekmen, 2005). three temperature–time conditions: high temperature-short roast
Moreover, the HS method proposed by the FDA in 2004 was up- (200 °C, 6 min), intermediate temperature-intermediate roast
dated in 2006, decreasing sampling temperature from 80 to 60 °C (170 °C, 12 min) and low temperature-long roast (140 °C, 20 min).
because of the observation of furan formation at the former tem- All coffee samples were then ground at room temperature to a com-
perature (Nyman, Morehouse, McNeal, Perfetti, & Diachenko, monly commercialised particle size (200–250 lm). To determine
2006, 2008). One advantage of using HS-SPME instead of HS is that the degree of roasting, the international standard colour space
lower extraction temperatures (30 °C) can be commonly used parameters were used. These parameters, proposed by the Interna-
and consequently the generation of furan can be reduced. tional Commission on Illumination (CIE) (Papadakis, Abdul-Malek,
Here we examined the furan content in brews obtained from Kamdem, & Yam, 2000), allow the characterisation of coffee colour
regular, decaffeinated and instant coffee commercialised in Spain. by means of L⁄a⁄b⁄ parameters, where L⁄ is the luminance or light-
We used a fully automated HS-SPME-GC–MS method. In addition, ness component and a⁄ and b⁄ are the chromaticity coordinates.
the occurrence of furan in coffee brews from commercially packed These measurements were performed using a Chromameter CR-
capsules was evaluated for the first time. The influence of HS-SPME 410 spectrophotometer (Konica Minolta, Osaka, Japan).
conditions, such as temperature and time, on furan formation dur- The regular ground coffee samples (9 g) were brewed using an
ing the analysis was also studied. Moreover, the effect of roasting espresso coffee machine and a home drip coffee maker to obtain
conditions and brewing procedures on the furan content of the fi- approximately 60 ml of brew in both cases. For the coffee roasted
nal brews was examined. By means of a 24-h recall questionnaire, in the laboratory, only the automatic espresso machine was used
we also estimated furan intake from coffee consumption in a small to obtain brews. Decaffeinated ground coffees were prepared with
population sample in Barcelona (Spain). a home drip coffee maker using the same coffee/water ratio as that
used for the brewed regular ground coffee. Instant coffee was ob-
tained following the manufacturer’s recommendations, mixing
2. Materials and methods
2 g of powder with 60 g of boiling water. For preparation of Nes-
pressoÒ coffee, all espresso blends, ca. 5.5 g of coffee, were brewed
2.1. Chemicals and standards
with an Essenza Nespresso coffee machine to obtain 40 ml of cof-
fee, while for the Lungo coffee capsules, 110 ml of brew was ob-
Furan and [2H4] furan (furan-d4) at a purity >99% were pur-
tained from ca. 7 g of coffee. After preparation, all brews were
chased from Sigma–Aldrich (Munich, Germany). Water of organic
placed immediately in a closed vial with no headspace and then
trace grade, methanol of GC grade and sodium chloride of analyti-
stored at 4 °C until analysis.
cal grade were all obtained from Merck (Darmstadt, Germany).
Individual stock standard solutions of furan and furan-d4 at a con-
2.3. GC–MS conditions
centration of 12 mg/g in methanol were prepared by transferring
20 ll of pure compounds to a 2-ml amber-sealed vial previously
GC–MS analyses were performed on a Varian 3800 gas chro-
filled with methanol. Intermediate standard solutions of furan
matograph coupled to a Saturn 2200 ion-trap MS system (Varian,
and furan-d4 were prepared by weight at a concentration of
Mississauga, Canada) and equipped with a CTC Combi-Pal auto-
90 lg/g from stock standard solutions by appropriate dilution with
sampler (CTC Analytics AG, Zwingen, Switzerland), which is config-
methanol. All these solutions were stored at 0 °C and prepared
ured with a sample tray for 32 vials of 20 ml, a SPME fibre
weekly. Aqueous working standard solutions of furan and furan-
conditioning station, and a temperature-controlled single magnet
d4 at a concentration of 90 ng/g were individually prepared daily
mixer tray. Separation was performed on a BX-volatile (cyanopro-
by spiking 20 ml of water with appropriate volumes of their inter-
pylphenyl polysilphenylene-siloxane), 60 m 0.25 mm i.d., fused-
mediate standard solutions. For furan determination, eight-calibra-
silica capillary column (SGE Europe, Villebon, France) with a film
tion standard solutions at concentrations ranging from 0.05 to
thickness of 1.4 lm. The temperature programme was 35 °C (held
10 ng/g were prepared by adding in weight, through the septum,
for 2 min) to 230 °C at 20 °C/min (held for 5 min). The injector was
an appropriate amount of the aqueous working standard solution
fitted with an SPME inlet liner (i.d., 0.75 mm, SGE Europe) and a
of furan into a 20-ml HS vial. In addition, 100 ll of the furan-d4
23-gauge Merlin Micro-seal septum (Supelco, Bellefonte, PA,
aqueous working solution was added to each calibration solution
USA) and operated in splitless injection mode (3 min). Helium
to achieve a concentration of 0.9 ng/g.
was used as carrier gas at a constant flow-rate of 1.7 ml/min. The
ion trap mass spectrometer was tuned using perfluorotributyl-
2.2. Coffee sample preparation amine (FC-43) following the manufacturer’s recommendations.
Electron ionisation (70 eV) with 30 lA of emission current and
A total of 23 samples of ground coffee (15 caffeinated and 8 temperatures of 200, 80 and 280 °C were set for the trap, manifold
decaffeinated) and five samples of instant coffee, selected from and transfer line, respectively. For MS acquisition, EI full-scan
the most consumed brands in Spain, were obtained from the major mode was used over the mass range m/z 35–100 at 0.75 s/scan
markets in Barcelona (Spain). In addition, 12 NespressoÒ coffee (7 lscan per scan). Quantification of furan was performed by iso-
capsules (Nestlé NespressoÒ SA, Paudex, Switzerland) (Ristretto, tope dilution using m/z 68 and m/z 72 for furan and furan-d4,
Arpeggio, Roma, Livanto, Capriccio, Volluto, Cosi, Decaffeinato, respectively. For confirmation, m/z 39 [M–CHO]+ for furan was
Decaffeinato intense, Vivalto Lungo, Decaffeinato Lungo and Finezzo monitored. A Varian Saturn Workstation software (Version 6.42)
Lungo) were purchased. Before use, commercial coffee samples was used for control, general operation and data acquisition of
were stored at 4 °C in their sealed packages to prevent loss of furan. the GC–MS system, while the Cycle composer software (Version
To study the effect of roasting on furan formation, green coffee 1.5.3) was used to control and program the CTC Combi-Pal
beans from Brazil were purchased from a local roasting factory in autosampler.
Barcelona, Spain. Portions of these beans (ca. 100 g) were then
individually roasted at temperatures of 140, 170 and 200 °C for 2.4. Automated HS-SPME method
6 min in an electric rotating laboratory roaster (Probat-Werke Pré
1Z, Probat-Werke, Germany). Moreover, to determine the influence Furan analysis in coffee samples was performed using an
of temperature and time on the formation of furan, several automated HS-SPME method that was previously developed and
M.S. Altaki et al. / Food Chemistry 126 (2011) 1527–1532 1529
validated (Altaki et al., 2009). Briefly, an aliquot (0.5 ml) of brewed coffee brew obtained using an espresso coffee machine were ana-
coffee was quickly transferred from the storage vial, immersed in lysed in triplicate using the proposed HS-SPME method (Section
an ice/water bath (4 °C) to a 20-ml headspace vial containing a 2.4). No significant differences between the furan concentration
PTFE-coated stir bar, 2 g of sodium chloride and an adequate in the two aliquots (57 ± 4 ng/ml for 0.5 ml of coffee and
amount of water (up to 8 ml). The vial was immediately sealed 55 ± 3 ng/ml for 2 ml) were observed, thereby demonstrating that
and then spiked with 100 ll of furan-d4 aqueous working standard the amount of precursor, which is directly related to the amount
solution. For HS-SPME analysis, a 75-lm carboxen–poly- of coffee, did not affect the formation of this compound at 30 °C.
dimethylsiloxane fibre (CAR/PDMS) (Supelco, Bellefonte, PA, In addition, an aliquot of 0.5 ml of the same coffee brew was ana-
USA), which was conditioned following the manufacturer’s recom- lysed using an equilibration time of 100 min and the same extrac-
mendations (300 °C under helium flow for 1 h) was used. Before tion time (20 min). In this case, the concentration of furan
analysis, the sample vial was equilibrated for 5 min in the single (58 ± 4 ng/ml) was not statistically different to that obtained previ-
magnet mixer tray at 30 °C and 750 rpm. The fibre was cleaned ously using a 5-min equilibration at a confidence level of 95%.
for 1 min at 275 °C in a conditioning station and then immediately These findings indicate that furan is not formed during HS-SPME
exposed to the headspace of sample vial for 20 min at 30 °C. Ther- analysis at 30 °C during the equilibration and extraction. Moreover,
mal desorption of furan was accomplished by exposing the fibre in the stability of the sample in the vials before analysis was exam-
the GC injector port at 275 °C for 2 min. ined. For this purpose, four sample vials were placed in the sample
tray of the automatic injector at laboratory temperature (20 ± 2 °C)
2.5. Quality control and quality criteria and analysed following the proposed HS-SPME method after 0, 1, 6
and 15 h. For all standing times, good agreement in the results
Quality control of the HS-SPME method was performed through (RSD% < 12%) was obtained, thereby demonstrating the stability
the routine analysis of procedural blanks and quality control stan- of the coffee samples in the vials before analysis.
dards and samples to ensure the absence of contaminants and pos-
sible carryover between samples, and to assess the quality of the 3.2. Influence of the roasting process on furan formation
results. In addition, a daily sensitivity test was carried out to check
the possible changes in the absorption capacity of the SPME fibre The properties of roasted coffee, such as aroma and taste, are
and the GC–MS response. Although no certified reference materials highly influenced by the roasting degree, which particularly de-
are available, the method was assessed in an interlaboratory study pends on the temperature and time applied during the process.
on furan in baby food (z-score < 0.6), organised by the Institute for However, there is little information about the effect of roasting
Reference Materials and Measurements (IRMM) of the European conditions on furan formation. For instance, Zoller et al. (2007)
Commission’s Joint Research Centre (JRC) (JRC, 2008). The identifi- found that furan concentration increases from 22 to 1792 ng/g
cation criteria for furan were based on its retention time and the (semi-quantitative data) when the roasting temperature is in-
intensity ratios of furan and furan-d4 ions. A deviation of the ion creased from 150 to 250 °C in an open headspace vial (small-scale
intensity ratios within 20% of the mean values of the calibration in the laboratory). In our study, we addressed the effect of roasting
standards was considered acceptable. on the furan content of green beans from Brazil, which were
roasted under conditions commonly used in the coffee manufac-
2.6. Coffee consumption data and furan intake turing industry (roaster, roasting degree, cooling and grinding).
First, the effect of temperature was evaluated by roasting the cof-
Data on coffee consumption habits in Spain were obtained from fee beans in an electric rotating roaster for 6 min at 140, 170 and
the Spanish Coffee Federation (FEC) and the Spanish Ministry of 200 °C. After roasting, the beans were ground and the roasting de-
Agriculture, Fisheries and Food (MAPA) (FEC, 2007; MAPA, 2007). gree was measured (Section 2.2). Green coffee was not analysed
Furthermore, coffee consumption data were obtained using a because it is not directly consumed. An Espresso machine was used
24-h recall questionnaire on 488 participants (219 females and to prepare a brew from each sample (9 g of coffee powder in 60 ml
269 males) aged from 19 to 58 years from five zones in the city of water) and the furan content was determined in triplicate using
of Barcelona. Each participant was asked about the number of cups the proposed HS-SPME GC–MS method. Furan levels increased
of coffee had in the previous 24 h. Furan intake per day and capita from 22 ± 2 to 138 ± 8 ng/ml with the roasting degree (L⁄ from
from brewed coffee was estimated from the average amount of 47.4 to 23.16) (Table 1). These findings can be explained by the in-
coffee consumed and the minimum and maximum concentration crease in roasting temperature.
of furan detected in espresso and drip coffee brews. The combined effect of both temperature and time on furan for-
mation was evaluated by roasting several portions of Brazilian
green coffee beans to a degree (L⁄ = 21 ± 0.7), commonly used for
3. Results and discussion commercialised regular roasted coffees, but applying different
temperature–time conditions. The following three conditions were
3.1. Evaluation of furan formation during HS-SPME analysis tested: high temperature-short roast (200 °C, 6 min), intermediate
range of concentrations (0.01–10 ng/g) and low limits of detection Temperature (°C) Lightness (L⁄) b
Furan level in brewed coffee
(LODs, 0.02–0.12 ng/g) with good precision (RSD% 6 10%). How- Mean ± sd (ng/ml)c RSD (%)
ever, since the formation of furan from green coffee has been re-
140 47.4 22 ± 2 9
ported during headspace analysis at an extraction temperature of 170 39.98 53 ± 3 6
40 °C (Senyuva & Goekmen, 2005), experiments were conducted 200 23.16 138 ± 8 6
to check whether furan is effectively formed during the HS-SPME a
Roasting time 6 min. Brewed with an automatic espresso machine at same
method. The effects of extraction time and the sample size on furan conditions.
formation were evaluated at the HS-SPME extraction temperature b
L⁄ = 100 (White), 0 (Black).
c
(30 °C). For this purpose, two aliquots (0.5 and 2 ml) of a natural n = 3.
1530 M.S. Altaki et al. / Food Chemistry 126 (2011) 1527–1532
ing the results obtained using the two brewing methods and the
Furan concentration ng/g
160
138
140 same coffee/water ratio, the furan content in espresso brews was
120 always higher, between 32% and 85%, than that obtained with a
89
100 home drip coffee maker. These results can be explained by the
80 64 higher pressure applied in the espresso machine (900–1000 kPa)
60 than that in the home drip maker (100 kPa), which increases the
40 extraction of furan from the coffee powder. Another factor that
20 must be considered is the loss of furan in the cup or in the pot
0 T (ºC) by evaporation during brewing. Preparation of a cup of coffee often
140 170 200 requires 30 s using an espresso machine, while a home drip coffee
Roasted during 20 min. Roasted during 12 min. Roasted during 6 min. maker takes about 100 s. This time difference may explain the
higher furan content found in espresso coffee brews. High concen-
Fig. 1. Influence of temperature–time roasting conditions on the furan content of
trations of furan in brews prepared with an espresso machine have
brewed coffee (roasting degree, L⁄ = 21 ± 0.7).
been reported previously (Kuballa et al., 2005; Zoller et al., 2007).
Decaffeinated and instant coffee samples were also analysed.
temperature-intermediate roast (170 °C, 12 min) and low temper- Decaffeinated brews were prepared using a home drip coffee ma-
ature-long roast (140 °C, 20 min). After roasting and grinding, cof- ker while instant coffee samples were obtained following the man-
fee powder was brewed in an espresso machine to obtain 60 ml of ufacturer’s instructions (Section 2.2). Furan was found in all
coffee brew which was analysed by HS-SPME-GC–MS The lowest samples at concentration levels ranging from 14 ng/ml to 65 ng/
furan concentration (64 ± 5 ng/ml) was found for low temperature ml for decaffeinated coffees and between 12 ng/ml and 35 ng/ml
and slow roasting (140 °C and 20 min) conditions, while for inter- for instant coffee (RSD% 6 12%) (Table 3). These concentrations
mediate and quick roasting processes at higher temperatures, the
concentration of this compound increased to 89 ± 7 and
138 ± 8 ng/ml, respectively (Fig. 1). These observations could be
Table 3
attributed to the combination of two effects, namely the decrease Furan concentration (ng/ml) in decaffeinated and instant coffee brews.
in furan formation at low temperature and the increase in the loss
Coffee sample Furan concentration
of furan by evaporation during the long roast at 140 °C.
(ng/ml)
Mean ± sda RSD (%)
3.3. Influence of coffee type and brewing procedure on furan content b
Decaffeinated coffee
Natural decaffeinated coffee #1 24 ± 3 12
The effect of the brewing procedure on the furan content of the Natural decaffeinated coffee #2 28 ± 2 7
coffee brews was studied using a professional espresso coffee ma- Natural decaffeinated coffee #3 20 ± 2 12
Natural espresso coffee 35 ± 3 9
chine and a home drip coffee maker, the most common machines
Natural cream espresso coffee 65 ± 2 3
used to prepare coffee in Europe. For this study, brews obtained Natural 100% Arabica coffee 14 ± 1 8
from 15 commercial ground coffees selected from the most con- Natural low caffeine content coffee 18 ± 2 11
sumed brands in Spain were analysed in triplicate using the pro- Coffee blend (50% natural and 50% torrefacto coffee) 40 ± 2 5
posed HS-SPME GC–MS method. Furan concentration in brews Instant coffeec
Natural instant coffee #1 12 ± 1 8
from the two brewing methods ranged from 20 ± 3 to 78 ± 6 ng/ Natural instant coffee #2 20 ± 2 10
ml for brews obtained with a home drip coffee maker and between Natural decaffeinated instant coffee 28 ± 3 11
43 ± 4 and 146 ± 12 ng/ml for those obtained using espresso ma- Natural pure 100% Colombian coffee 35 ± 2 6
chine with a relative standard deviation lower than 15% (Table Natural 100% Arabica coffee 32 ± 2 6
2). These concentrations are consistent with those reported in a
n = 3 replicates.
the literature for brewed coffees of diverse origins (3–199 ng/ml) b
Brewed coffee prepared using a drip coffee maker (9 ml/60 of boiling water).
c
(EFSA, 2004; FDA, 2008; SFOPH, 2004; Zoller et al., 2007). Compar- Instant coffee prepared following the manufacturer directions.
Table 2
Furan concentration (ng/g) in brewed regular coffees obtained by drip coffee maker (filter) and espresso coffee machine.
a
Regular coffee Furan concentration in coffee brew (ng/ml)
Drip coffee maker Espresso machine
Mean ± sd RSD (%) Mean ± sd RSD (%)
Natural coffee #1 40 ± 3 8 70 ± 5 7
Natural coffee #2 29 ± 3 10 65 ± 5 8
Natural coffee #3 28 ± 3 11 43 ± 4 9
Natural coffee #4 48 ± 3 6 93 ± 6 6
Natural espresso coffee 30 ± 3 10 60 ± 5 9
Natural cream espresso coffee 39 ± 3 8 72 ± 6 8
Natural filter coffee 20 ± 3 15 50 ± 5 10
Natural Italian style coffee 43 ± 3 7 110 ± 8 7
Natural pure 100% Colombian coffee 66 ± 4 6 125 ± 6 5
Natural pure 100% African coffee 46 ± 3 7 82 ± 5 6
Natural pure 100% Brazilian coffee 78 ± 6 8 146 ± 12 8
Natural 100% Arabica coffee 24 ± 3 13 56 ± 5 8
Natural black 100% Arabica coffee 52 ± 3 6 101 ± 5 5
Coffee blend (70% natural and 30% torrefacto coffee) 35 ± 3 9 74 ± 6 8
Coffee blend (50% natural and 50% torrefacto coffee) 32 ± 3 9 59 ± 5 8
a
n = 3.
M.S. Altaki et al. / Food Chemistry 126 (2011) 1527–1532 1531
were similar to those obtained for regular ground coffee and are in 3.5. Furan intake from coffee consumption
agreement with data described in the literature (Crews & Castle,
2007; FDA, 2008). For instance, the FDA reported furan concentra- At present, the average coffee consumption in Spain reported by
tions from 34 to 53 ng/ml in decaffeinated coffee, while Crews the Spanish Coffee Federation is 599 cups of coffee per person and
et al. (2009) found levels ranging between 8 and 66 ng/ml. For in- year (FEC, 2007), which corresponds to 1.64 cups of coffee per per-
stant coffee, concentrations ranging from 1 to 51 ng/ml have been son and day. These values are consistent with data provided by the
reported (Crews & Castle, 2007; Crews et al., 2009; Zoller et al., Spanish Government (MAPA, 2007) and the International Coffee
2007). Given that decaffeinated coffee is produced from green cof- Organization (ICO, 2008), which reported an average coffee con-
fee, our results confirm that the roasting process is responsible for sumption of 4.5 kg of coffee beans per person and year for the last
the formation of furan. In addition, the slightly lower concentra- three years in Spain. On the basis of these data and on the furan
tions of furan in instant coffee can be attributed to the lower cof- content obtained in this study for espresso and drip coffee brews,
fee/water ratio (2 g coffee powder/60 ml of water) used for brew we estimated the daily furan intake to be from 1.97 to 14.4 lg/per-
preparation. son, which implies exposure of between 0.03 and 0.21 lg/kg body
weight (bw)/day of this compound, assuming an average adult
3.4. Analysis of furan in Nespresso coffee capsules weight of 70 kg. Although these data are useful to establish the
overall exposure of the total population to furan, we considered
Recent years have witnessed a large increase in the consump- it of interest to obtain a more detailed picture of coffee consump-
tion of coffee from packed capsules worldwide. This system guar- tion habits of people living and working in a large city. For this
antees fresh espresso coffee brews with a high-quality aroma purpose, a 24-h recall questionnaire was completed from a repre-
and taste, and protects coffee from the damaging effects of light, sentative sample of 488 participants aged from 19 to 58 years from
air and humidity until brewing. The capsules in this system are five districts of Barcelona (Spain) (219 females and 269 males).
filled with ground roasted coffee and are hermetically plastic- Each participant was asked about his/her coffee drinking habits,
wrapped. A specialised coffee machine is required for each capsule such as the amount of coffee (cups of coffee/day) ingested in the
brand. To determine the furan content in brews obtained with this previous 24 h, type of coffee and brewing procedure, and also some
system, eleven espresso coffee capsules from Nespresso (Table 4), personal information, such as age, sex, address, citizenship and
one of the most widely consumed coffee capsules in the world, body weight. Fig. 2 shows the results of the coffee consumption
were selected and analysed in triplicate following the proposed survey. Around 20% of the participants were not habitual coffee
method. For brewing coffee, a Nespresso coffee machine (Essenza) drinkers. The average daily coffee consumption was 3.1 and 2.1
was used to obtain 40 ml of brews for coffee blends (Ristretto, cups of coffee for males and females, respectively. This consump-
Arpeggio, Roma, Livanto, Capriccio, Volluto, Cosi, Decaffeinato and tion rate implies a daily furan intake for males from 3.6 to
Decaffeinato intense) and 110 ml for Lungo coffees (Vivalto Lungo, 26.3 lg/person and for females between 2.4 and 17.5 lg/person.
Decaffeinato Lungo and Finezzo Lungo). Furan concentrations for These findings indicate exposure to furan of 0.05 to 0.38 lg/
these coffee capsules ranged from 117 ng/ml for Capriccio to kg bw/day for males and between 0.03 and 0.25 lg/kg bw/day for
244 ng/ml for Decaffeinato intenso (Table 4). These results were females. These data are only slightly higher than those obtained
higher than those obtained in espresso brews of regular ground using the official Spanish average consumption data. These results
coffee (43–146 ng/ml), although similar coffee powder/water ra- agree with those reported by the US FDA (0.15 lg/kg bw/day)
tios were used (5.5 g/40 ml for Nespresso capsule and 9 g/60 ml (Morehouse et al., 2008) and EFSA (<0.03–1.65 lg/kg bw/day)
for espresso coffee). This observation could be attributable to two (EFSA, 2004) and are much lower than the non-observed adverse
factors, namely prevention of furan loss during handling as a result effect level (NOAEL), which was established at 2 mg/kg bw/day
of the hermetic seal of the capsules, and the higher pressure ap- for experimental animals (Moser, Foley, Burnett, Goldsworthy, &
plied by the Nespresso coffee machine (1850 kPa), which favours Maronpot, 2009). Since the acceptable daily intake (ADI) was esti-
the extraction of this compound from the ground coffee. Regarding mated to be 2 lg/kg bw/day within a safety factor of 1000 (Kuballa
the Lungo coffee capsules, furan concentrations ranged from 68 to et al., 2005), the maximum furan intake found in this study
102 ng/ml and were lower than those obtained with espresso cap- (0.38 lg/kg bw/day) was around five times lower than this value.
sules. This finding may be due to the higher dilution of the coffee Consequently, we conclude that the estimated risk from coffee
brew (7 g coffee powder/110 ml of water) and the longer time per- consumption, which is the biggest source of furan exposure for
iod of brewing, which may increase the loss of furan. adults, is low.
Table 4
Furan levels (ng/ml) in brewed coffee of Nesspreso capsules.
70
4. Conclusions FDA-United Steates Food and Drug Administration (2008). Exploratory data
on furan in food. Available from: <http://www.fda.gov/Food/FoodSafety/
FoodContaminantsAdulteration/ChemicalContaminants/Furan/ucm078439.htm>.
Furan formation was not observed during the analysis of FEC-Spanish Coffee Federation (2007). Commercialization of coffee in Spain. Available
brewed coffee by automatic HS-SPME combined with GC–MS using from: <www.federacioncafe.com/Publico/ElCafe/Comercializacion.asp>.
Ho, I.-P., Yoo, S.-J., & Tefera, S. (2005). Determination of furan levels in coffee using
a low extraction temperature (30 °C) and a short time (20 min).
automated solid-phase microextraction and gas chromatography/mass
The roasting time and temperature applied to green coffee beans spectrometry. Journal of AOAC International, 88, 574–576.
affect the final furan content of coffee. Therefore, we recommend IARC-International Agency for Research on Cancer (1995). IARC monographs on the
evaluation of carcinogenic risks to humans, dry cleaning, some chlorinated solvents
gently roasting conditions (140 °C for 20 min) to reduce furan for-
and other industrial chemicals (vol. 63, pp. 3194–3407). France: Lyon.
mation. The furan concentration levels found in regular ground ICO-International Coffee Organization. 2008. Historical Coffee Statistics. London: ICO.
coffee (20–146 ng/ml), decaffeinated (14–65 ng/ml) and instant Available online at: <http://www.ico.org/historical.asp>.
coffee (12–35 ng/ml) brews commercialised in Spain were in JRC-IRMM-Institute for Reference Materials and Measurements of the European
Commission’s Joint Research Centre. (2008). Proficiency test on the determination
agreement with those reported in the literature, showing the high- of furan in baby food. Available from: <irmm.jrc.ec.europa.eu/html/ news/news/
est concentrations in espresso brews. Nevertheless, it must be furan.htm>.
mentioned that furan concentrations in brews from Nespresso cof- Kuballa, T., Stier, S., & Strichow, N. (2005). Furan in kaffee und kaffeegetränken.
Deutsche Lebensmittel-Rundschau, 101, 229–235.
fee capsules (117–244 ng/ml) were higher than those obtained for La Pera, L., Liberatore, A., Avellone, G., Fanara, S., Dugo, G., & Agozzino, P. (2009).
regular ground coffee. The daily furan intake from coffee consump- Analysis of furan in coffee of different provenance by head-space solid phase
tion was estimated to be 0.25 and 0.38 lg/kg bw for males and microextraction gas chromatography–mass spectrometry: Effect of brewing
procedures. Food Additives and Contaminants, 26, 786–792.
females, respectively. These values are lower than the established Maga, J. A. (1979). Furans in foods. Critical Reviews in Food Science and Nutrition, 11,
acceptable daily intake of 2 lg/kg bw. 355–400.
MAPA-Spanish Ministry of Agriculture, Fisheries and Food Alimentation in Spain.
(2007). Available from: <http://www.mapa.es/es/alimentacion/pags/consumo/
Acknowledgments
consumo.htm>.
Morehouse, K., Nyman, P., Mcneal, T., Dinovi, M., & Perfetti, G. (2008). Survey of
The authors gratefully acknowledge financial support from the furan in heat processed foods by headspace gas. Chromatography/mass
spectrometry and estimated adult exposure. Food Additives and Contaminants,
Departament d’Innovació, Universitat I Empresa of the Generalitat
25(3), 259–264.
de Catalunya under Project 2009SGR1325. The authors also thank Moser, G.-J., Foley, J., Burnett, M., Goldsworthy, T.-L., & Maronpot, R. (2009). Furan-
the coffee manufacturers Unión de Industrias del café, S.L. and induced dose-response relationships for liver cytotoxicity, cell proliferation,
Café Saula, S.A., for their collaboration. and tumorigenicity (furan-induced liver tumorigenicity). Experimental and
Toxicologic Pathology, 61(2), 101–111.
Murkovic, M., & Derler, K. (2006). Analysis of amino acids and carbohydrates in
References green coffee. Journal of Biochemical and Biophysical Methods, 69, 25–32.
NTP-National Toxicology Program (2004). Furan CAS No. 110-00-9. Report on
Altaki, M. S., Santos, F. J., & Galceran, M. T. (2007). Analysis of furan in foods by carcinogens (11th ed.). US Department of Health and Human Services, Public
headspace solid-phase microextraction-gas chromatography–ion trap mass Health Service. Available from: <http://ntp.niehs.nih.gov/files/Furan1.pdf>.
spectrometry. Journal of Chromatography A, 1146, 103–109. Nyman, P.-J., Morehouse, K.-M., McNeal, T.-P., Perfetti, G.-A., & Diachenko, G.-W.
Altaki, M.-S., Santos, F.-J., & Galceran, M.-T. (2009). Automated headspace solid- (2008). Single-Laboratory validation of a method for the determination of furan
phase microextraction versus headspace for the analysis of furan in foods by gas in foods by using static headspace sampling and gas chromatography/mass
chromatography–mass spectrometry. Talanta, 78, 1315–1320. spectrometry, Part 2-low-moisture snack foods. Journal of AOAC International,
Bianchi, F., Careri, M., Mangia, A., & Musci, M. (2006). Development and validation of 91, 414–421.
a solid phase micro-extraction–gas chromatography–mass spectrometry Nyman, P.-J., Morehouse, K.-M., McNeal, T.-P., Perfetti, G.-A., & Diachenko, G.-W.
method for the determination of furan in baby-food. Journal of (2006). Single-laboratory validation of a method for the determination of furan
Chromatography A, 1102, 268–272. in foods by using static headspace sampling and gas chromatography/mass
Crews, C., & Castle, L. (2007). A review of the occurrence, formation and analysis of spectrometry. Journal of AOAC International, 89, 1417–1424.
furan in heat-processed foods. Trends in Food Science and Technology, 18(7), Papadakis, S. E., Abdul-Malek, S., Kamdem, R. E., & Yam, K. L. (2000). A versatile and
365–372. inexpensive technique for measuring colour of foods. Food Technology, 54(12),
Crews, C., Roberts, D., Lauryssen, S., & Kramerc, G. (2009). Survey of furan in foods 48–51.
and coffees from five European Union countries. Food Additives and Senyuva, H.-Z., & Goekmen, V. (2005). Analysis of furan in foods. Is headspace
Contaminants, Part B, 2(2), 95–98. sampling a fit-for purpose technique? Food Additives and Contaminants, 22,
EFSA-European Food Safety Authority (2004). Report of the Scientific Panel on 1198–1202.
Contaminants in the Food Chain on provisional findings of furan in food. EFSA SFOPH-Swiss Federal Office of Public Health (2004). Furan in food: detailed list of
Journal, 137, 1–20. <http://www.efsa.europa.eu/en/scdocs/scdoc/137.htm>. results. Available from: <www.bag.admin.ch/themen/lebensmittel/04861/
EFSA-European Food Safety Authority (2006 and 2007). Invitation to submit data on 04913/index.html?lang=de>.
furan in food and beverages. Available from: <www.efsa.europa.eu/EFSA/ Wenzl, T., Lachenmeier, D., & Gökmen, V. (2007). Analysis of heat-induced
efsa_locale-1178620753812_1178620771259.htm>. contaminants (acrylamide, chloropropanols and furan) in carbohydrate-rich
FDA-United States Food and Drug Administration. (2006). Determination of food. Analytical and Bioanalytical Chemistry, 389, 119–137.
furan in food. Available from: <http://www.fda.gov/Food/FoodSafety/ Zoller, O., Sager, F., & Reinhard, H. (2007). Furan in food: Headspace method and
FoodContaminantsAdulteration/ChemicalContaminants/Furan/ucm078400.htm>. product survey. Food Additives and Contaminants, 24(S1), 91–107.
FDA-United States Food and Drug Administration (2004). Furan in food, thermal
treatment: Request for data and information. Available from: <http://www.fda.
gov/Food/FoodSafety/FoodContaminantsAdulteration/ChemicalContaminants/
Furan/ucm078469.htm>.