- Email: [email protected]
Influence of frying fat on the formation of heterocyclic amines in fried beefburgers and pan residues
Fd Chem. Toxic. Vol. 33, No. 12, pp. 993-1004, 1995
~
Pergamon
0278-6915(95)00074-7
Ltd reserved
Copyright © 1995 ElsevierScience Printed in Great Britain. All fights 0278-6915/95 $9.50 + 0.00
Research Section
Influence of Frying Fat on the Formation of Heterocyclic Amines in Fried Beefburgers and Pan Residues M. A. E. J O H A N S S O N * , L. F R E D H O L M t , I. B J E R N E t a n d M. J A G E R S T A D * *Department of Applied Nutrition and Food Chemistry, Chemical Centre, Lund University, PO Box 124, S-221 00 Lund and iVan den Bergh Foods, PO Box 721, S-251 07 Helsingborg, Sweden (Accepted 30 Mav 1995)
Abstract--The influence of six frying fats (butter, margarine, margarine fat phase, liquid margarine, rapeseed oil and sunflower seed oil) on the formation of mutagenic/carcinogenic heterocyclic amines (HAs) during the frying of beefburgers was investigated. Frying was performed at 165 and 200":'C (i.e. under conditions that represented normal household cooking practices). The fried beefburgers and their corresponding: pan residues were purified using solid-phase extraction and analysed for HAs using HPLC with photodiode array UV and fluorescence detection. The HAs 2-amino-3,8-dimethylimidazo[4,5-f]quinoxaline (MelQx), 2-amino-3,4,8-trimethylimidazo[4,5-f]quinoxaline (DiMelQx), 2-amino-l-methyl-6phenylimidazo[4,5-b]pyridine (PhlP), 9H-pyrido[3,4-b]indole (norharman) and l-methyl-9Hpyrido[3,4-b]indole (harman) were recovered. The amount increased with the temperature, and the content of HAs in the pan residue was much higher than in the corresponding beefburger. The amounts of MelQx ranged from 0.2 to 1.6 ng/g in the beefburgers and from 0.8 to 4.3 ng/g in the pan residues. DiMelQx ranged from undetectable to 0.4 ng/g in the beefburgers and from 0.4 to 1.3 ng/g in the residues. PhlP ranged from C,.08 to 1.5 ng/g in the meat and from 0.4 to 13.3 ng/g in the residues. The total amount of HAs in meat and pan residue combined was significantly lower after frying in sunflower seed oil or margarine th~.n after frying with the other fats. The observed differences in MelQx and DiMelQx formation could be explained in terms of oxidation status (peroxide and anisidine value) and antioxidant content (vitamin A, vitamin E and tocopherols/tocotrienols) using partial least squares analysis.
INTRODUCTION Several epidemiolo~ical studies have shown a relationship between the consumption of meat and an increased risk of colorectal cancer and possibly also of breast, stomach, pancreatic and urinary bladder cancer (Norell et al., 1986; Schiffman and Felton, 1990; Steineck et al., 1990 and 1993; Willet et al., *Author for corresponAence. AV = anisidine value; DCM = dichloromethane; 4,8-DiMelQx = 2-amino-3,4,8-trimethylimidazo[4,5-f]quinoxaline; FFA = free fatty acids; GIu-P- 1 = 2-amino-6-methyl-dipyrido[l,2-a :3',2'-d]imidazole; GIu-P-2 = 2-aminodipyrido[l,2-a : 3',2'-d]imidazole; harman = l-methyl-9H-pyrido[3,4-b]indole; HA(s) = heterocycli¢ amine(s); IQ = 2-amino-3methylimidazo[4,5-/]quinoline; MelQ = 2-amino-3,8dimethylimidazo[4,5-f ]quinoline; MelQx = 2-amino3,8-dimethylimidazc,[4,5-f]quinoxaline; norharman = 9H-pyrido[3,4-b]indole; PhlP = 2-amino-l-methyl-6phenylimidazo[4,5-b]pyridine; PLS = partial least squares; PRS = propylsulfonic acid-silica gel; PV= peroxide value; c~-T = c~-tocopherol; ~-T~= ~tocotrienol; fl-T =/?-tocopherol; ,/-T = y-tocopherol; 7T 3 = y-tocotrienol; ii-T = 6-tocopherol.
Abbreviations:
1990). The formation of mutagenic/carcinogenic heterocyclic amines (HAs) during the cooking of meat and fish products may be responsible for part of the observed increased cancer risk. Since first discovered by Sugimura et al. (1977), around 20 mutagenic HAs have been isolated from cooked foods, primarily from the crust of fried meat and fish, pan residues and beef extracts (Felton and Knize, 1990; Skog, 1993). All HAs tested have shown carcinogenic effects in long-term experiments in rodents (for a review, see Ohgaki et al., 1991) and 2-amino-3-methylimidazo[4,5-f]quinoline (IQ) also induces tumours in primates (Adamson et al., 1990). In addition, a transfer of 2-amino- 1-methyl-6-phenylimidazo[4,5-b]pyridine (PhlP) to the foetus and its excretion in the breast milk of mice and rats following maternal exposure has recently been reported (Brittebo et al., 1994; Ghoshal and Snyderwine, 1993; J/igerstad et al., 1994). HAs are formed from creatine, free amino acids and monosaccharides (which occur naturally in protein-rich foods of animal origin) at normal cooking temperatures (150-250°C) (for review, see Skog, 1993). Earlier studies have shown H A formation to 993
994
M . A . E . Johansson et al.
be dependent on physical parameters, such as cooking temperature and time, cooking equipment and technique, water activity, heat transport, and chemical parameters, especially the precursors, for example, carbohydrates, free amino acids and creatine (Johansson and J/igerstad, 1994; Knize et al., 1994; Skog, 1993; Skog et al., 1995; Tikkanen et al., 1993). The role of fat in the formation of HAs has not been sufficiently clarified. The few studies that have been carried out so far on the influence of frying fat or fat content on the formation of food mutagens have yielded inconsistent results. The fat content of the meat probably has some effect on HA formation, although it is difficult to distinguish between the physical and chemical effects (Barnes et al., 1983; Holtz et al., 1985; Knize et al., 1985; Overvik et aL, 1987). Modelling experiments have shown that certain oils increase the yield of HAs, while the oxidation status of the fat and its content of antioxidants showed no effects (Johansson and JS.gerstad, 1993; Johansson et al., 1993). The use of different frying fats did not affect the mutagenic activity in pork fried at 200°C (Nilsson et al., 1986). However, at 250°C, pork fried in sunflower seed oil or butter had the highest mutagenic activity, while pork fried in lard or olive oil had the lowest. When frying lamb the highest mutagenic activity was detected when butter was used (Barrington et al., 1990). Most reported data concerning the influence of fat on the formation of HAs are based on mutagenic activity. The only published study that reported quantitative data showed that pan residues from frying beefburgers in butter contained significantly higher amounts of 2-amino-3,8-dimethylimidazo[4,5-f]quinoxaline (MelQx) and 2-amino-3,4,8-trimethylimidazo[4,5-f]quinoxaline (DiMelQx) than samples fried in vegetable oil (Johansson and J~igerstad, 1994). In the present study, we quantitatively investigated the influence of different frying fats on the formation of HAs in fried beefburgers and their corresponding pan residues. Six fats with different chemical compositions were tested during frying under well defined cooking conditions. The oxidation status and antioxidant contents in the frying fats were determined after frying and the relationship between these parameters
and HA formation was statistically evaluated using partial least squares analysis (PLS). MATERIALS AND METHODS
Chemicals
All chemicals and solvents were of HPLC or analytical grade. Water was obtained from a Milli-Q water purification system (Millipore, Bedford, MA, USA). The solvents acetonitrile, methanol, dichloromethane (DCM), tetrahydrofuran, heptane, acetone and toluene, were purchased from Merck AG (Darmstadt, Germany). Synthetic IQ, 2-amino-3,8dimethylimidazo[4,5-f]quinoline (MeIQ), MeIQx, 4,8-DiMeIQx, PhIP, 2-amino-6-methyl-dipyrido [1,2-a :3',2'-d]imidazole (Glu-P-1), 2-aminodipyrido[1,2-a : 3',2"-d]imidazole (Glu-P-2) and 1-methyl-9Hpyrido[3,4-b]indole (harman) were obtained from Toronto Research Chemicals (Downsview, Ontario, Canada) and 9H-pyrido[3,4-b]indole (norharman) from Aldrich (Steinheim, Germany). A standard reference mixture of IQ, MeIQ, MeIQx, 4,8-DiMeIQx and PhIP [5 ng of each compound/#l of methanoltriethylamine (50:50, v/v), pH 3.2] was used as a spiking solution. Caffeine (Sigma, St Louis, MO, USA) at 2.5 ng//~l methanol-triethylamine (50:50, v/v), was used as internal standard. The materials used for PRS-tandem extraction (Extrelut and BondElut, PRS and C~8) were obtained from Merck AG and Sorbent (V~stra Fr61unda, Sweden). Fractogel TSK CM650(s) was obtained from Merck AG. Synthetic ]~-carotene was from Sigma. Meat sample
Brisket of beef (Musculus pectoralis superficialis, M. pectoralis profundus and M. seratus ventrilus) obtained from a local slaughterhouse, was minced (3 ram) and homogenized making up one batch of meat. The minced beef was made into burgers (87 mm diameter, 9 mm thick, weight 85 g) using a special punch. The beefburgers were stored at - 2 0 ° C until fried. The raw beef was analysed for water, fat, protein and ash using standard methods as previously described (Holtz et al., 1985). The amount of carbohydrates was calculated as residue after accounting for water, fat, protein and ash.
Table 1. Fatty acid composition of frying fats Fatty acid content (%)
Fatty acid ~
Butter
Margarine and margarine fat phase
Liquid margarine and rapeseed oil
Sunflower seed oil
7.5 3.0 11.5 28.0 11.0 28.5 2.5 -I 7
1.3 4.6 2.0 17.9 7.1 52.3 10.2 2.9 1.7 --
---4.8 1.8 63.1 19.6 8.0 2.3 0.4
---7.0 5.0 23.0 63.0 0.5 1.5 --
Effects of frying fat on the formation of HAs
995
Table 2. Contents of HAs in fried beefburgersand pan residues MelQx (ng/g)* Frying fat
F:ying temperature ('C) Meat
Margarine Margarine fat phase Liquid margarine Butter Rapeseed oil Sunflower seed oil
165 200 200 200 200 200 200
0.2 1.0~ 1.2 ± 0.1~ 1.1 ±0.1" 1.2 + 0.2~ 1.6 ± 0. Ib 1.2 ± 0.2~
DiMelQx (ng/g)*
PhlP (ng/g)*
Total amount of NAs (ng/g)*
Pan residue Meat Pan residue Meat
Pan residue Meat
Meat + Pan residue pan residue
0.8 2.7_+0.3~b 3.5 ± 0.3~ 4.3 _+0.5d 3.2 + 0.9~ 2.8 ± 0.2~ 2.1 ± 0.4~
0.4 4.6+0.9a 13.3+3.6 b 9.7_+2.3~ 11.7+3.4 ~ 11.4_+3.6~ 2.0+0.5 a
1.6 8.6±1.0 ~ 17.8+3.3 b 15.9+2.7 b 15.1+2.6 b 14.8±4.3 b 4.7±0.9 ~
nd 0.4~ 0.2b 0.4~b 0.2~ 0.4~ 0.2b
0.4t 1.3_+0.1" 1.0~ I.I _+0.2~¢ 1.0b~ 0.8b 0.5d
0.08 0.5~ 1.5_+0.6b 1.0~b 1.0~b 1.1+0.2 ~bt~l 0.9~
0.28 1.9±0.2 ~ 3.0+0.7 ~ 2.4+0.6 "b 2.5+_0.2a~ 3.2+0.3 ¢ 2.3±0.5 ~b
1.9 10.5±1.1 ~ 20.7+3.9 b 18.3_+2.1b 17.6+2.6 b 18.2+3.6 b 7.0±1.4 ~
nd = not detected *Content as ng/g cooked meal, recovery-correctedvaluesof duplicatedeterminationsfrom duplicate fryingexperiments.SDs are not givenfor HA values ~<1 ng/g. SD correspond:~ to four determinations. ?Single value. Values in the same column that do not share a common superscript (a-d) are significantlydifferent (one-way ANOVA, followed by Duncan's test; P < 0.05).
Frying fats
in t he f r y i n g fat w a s r e c o r d e d u s i n g a c h r o m / a l u m e l
T h e f o l l o w i n g six fi'ying f a t s w e r e used: b u t t e r ( A B S v e n s k t S m r r , S w e d e n ) , m a r g a r i n e , m a r g a r i n e fa t p h a s e , l i q u i d m a r g a r i : a e , r a p e s e e d a n d s u n f l o w e r seed oil ( V a n d e n B e r g h F o o d s , S w e d e n ) . T h e m a r g a r i n e , liquid margarine and butter contained 800 fat, w h e r e a s t h e m a r g a r i l a e fat p h a s e a n d t h e oils c o n t a i n e d 1 0 0 % fat. T h e m a r g a r i n e a n d t h e m a r g a r i n e
t h e r m o c o u p l e b e f o r e a d d i n g t he b e e f b u r g e r s . F o u r b e e f b u r g e r s (340 g) w e r e fried a t a t i m e , u s i n g 50 g m a r g a r i n e o r b u t t e r , or 40 g m a r g a r i n e fa t p h a s e o r oil, r e s p e c t i v e l y . T h e fat w a s p l a c e d in a p r e h e a t e d p a n . A f t e r 2 m i n t he b e e f b u r g e r s w e r e a d d e d a n d fried for 5 m i n o n o n e side a n d 3 m i n o n t h e o t h e r . T h e t o t a l h e a t i n g t i m e o f t h e fa t w a s 1 0 m i n . A l l f r y i n g e x p e r i m e n t s w e r e r e p e a t e d once. F o r a n a l y s i s
fat p h a s e h a d a n i d e n t i c a l fat c o m p o s i t i o n , a n d c o n s i s t e d o f r a p e s e e d a n d p a l m oil (oil a n d p a r t i a l l y
of HAs,
h y d r o g e n a t e d oil) arid c o c o n u t oil. T h e m a r g a r i n e
w a t e r ( a b o u t 100 g) w a s a d d e d a n d s t i r r e d t o d i s s o l v e t h e p a n re s i due . F o r fa t a n a l y s i s , f r y i n g fa t w a s
c o n t a i n e d whey. The
liquid margarine was made
f r o m r a p e s e e d oil a n d s k i m m e d m i l k . T h e s u n f l o w e r seed oil w a s n o t d e w a x e d . T a b l e 1 s h o w s the f a t t y a c i d c o m p o s i t i o n o f the fats.
Cooking conditions T h e b e e f b u r g e r s w e r e a l l o w e d to r e a c h r o o m t e m p e r a t u r e ( a b o u t 18°C) a n d t h e n w e r e fried in a thermostat-controlled teflon-coated frying pan (235 x 235 m m ) c o n s t r u c t e d a t t h e S w e d i s h I n s t i t u t e fo r F o o d R e s e a r c h ( S I K , G o t h e n b u r g , S w e d e n ) , a t fat t e m p e r a t u r e s o f 1,55 a n d 200°C. T h e t e m p e r a t u r e
pan
residues were collected after frying,
c o l l e c t e d a n d filtered i n t o a b r o w n b o t t l e . T o o b t a i n e n o u g h m a t e r i a l for a n a l y s i s o f fat o x i d a t i o n p r o d ucts, fa t f r o m t w o o r t h r e e i d e n t i c a l f r y i n g e x p e r iments was pooled to form one sample. Samples were immediately frozen and kept at -20°C until a n a l y s e d . T h e fri e d b e e f b u r g e r s w e r e w e i g h e d a n d f r o z e n a f t e r b e i n g p h o t o g r a p h e d t o r e c o r d t he d e g r e e of browning. Before extraction, the crust (outer 2 m m f r o m e a c h side) o f t h e b e e f b u r g e r s w a s r e m o v e d u s i n g a s c a l pe l . A l l s a m p l e s , i n c l u d i n g p a n r e s i d u e s , w e r e f r e e z e - d r i e d b e f o r e e x t r a c t i o n o f H A s a n d k e p t in plastic beakers at - 2 0 ° C
until they were analysed.
Table 3. Oxidation status and antioxidant content of frying fats Tocopherol/tocotrienol (mg/100 g) Frying fat Margarine
Margarine fat phase Liquid margarine Butter Rapeseed oil Sunflower seed oil
Temperature PV (<'C) (mEq/kg) unheated 165 200 200* unheated 200 unheated 200 unheated 200 unheated 200 200* unheated 200
0.0 0.0 0.2 2.0 0.0 0.4 0.1 1.2 0.0 0.3 0.0 20.0 37.6 0.1 21.1
AV 1.3 2.2 5.7 17.1 1.0 5.3 1.5 8.8 0.1 3.1 1.6 88.8 225.1 2.4 98.0
FFA Vitamin A /J-Carotene Vitamin E (%) (mg/100 g) (mg/100 g) ~t-T ~t-T3 /J-T y-T 7-T3 6-T Total (mg/100 g) 0.05 0.29 0.22 0.11 0.1 0.23 1.1 0.27 0.33 0.34 0.02 0.26 0.14 0.03 0.23
1.13 0.80 0.62 0.83 1.14 0.58 1.12 0.63 na na na na na na na
0.2 0.15 0.11 0.12 0.19 0.09 0.15 0.09 0.27 0.09 na na na na na
16.0 12.2 11.5 14.8 16.4 10.7 20.5 14.0 2.5 1.7 20.5 0.5 nd 59.5 0.7
4.4 3.4 3.2 3.9 3.4 2.1 nd nd nd nd nd nd nd nd nd
nd nd nd nd nd nd nd nd nd nd nd nd nd 2.3 nd
19.8 14.8 13.5 18.6 18.3 12.0 32.1 22.3 nd nd 33.6 2.8 1.8 1.0 nd
3.8 2.8 2.5 3.4 3.7 2.4 nd nd nd nd nd nd nd nd nd
1.0 0.6 0.6 0.8 1.0 0.7 1.2 0.7 nd nd 1.1 0.4 nd nd nd
45.0 33.8 31.3 41.5 42.8 27.9 53.8 37.0 2.5 1.7 55.2 3.7 1.8 62.8 0.7
19.3 14.6 13.8 17.8 19.3 12.6 23.8 16.2 2.5 1.7 23.9 0.7 0.2 60.7 0.7
na = not analysed nd= not detected All values are means from duplicate frying experiments. The amounts of vitamins A, fl-carotene, tocopherol/tocotrienol and vitamin E are given as rag/100 g fat I:,hase. *Fat heated at 200'C for 10 rain without meat.
996
M.A.E. Johansson et al.
~
Caffeine (is) mAU
15
IQx
10 MelQx
,'0
,'5
2's Time (min)
I
250
I
I
I
I
350 Wavelength (nm) 300
Fig. 1. Expanded region of a chromatogram from the HPLC analysis of pan residue from frying of beefburgers in liquid margarine at 200°C. The peaks corresponding to MelQx, DiMelQx, PhlP and caffeine (internal standard, i.s.) are indicated. On-line recorded UV absorbance spectra compared with those of synthetic MelQx (lower), DiMelQx (lower) and PhlP (upper), respectively. Margarine and rapeseed oil heated without meat for 10 min at 200°C were analysed for HAs and used as negative control samples. Extraction of HAs Freeze-dried meat (about 1.5g, generally corresponding to about 6 g cooked beef) and pan residue (about 0.7 g, obtained from about 5 g cooked beef) were extracted using the method described by Gross et al. (1993). Using this method, a polar extract (containing the IQ-type of HAs and the glutamic acid pyrolysates) and a non-polar extract (containing the pyrido-indole type of HAs) are obtained. Only the polar extract was analysed in this study. To enhance the solubility of the samples and increase the extraction recovery of HAs, the samples were dissolved in sodium hydroxide and placed in a water-bath at 50"C. For the extraction of pan residues 5% phenol was added to DCM (G. A. Gross, personal communication). Additional purification of the pan residues was performed using Fractogel TSK CM650(s) to improve selectivity and avoid interfering peaks during HPLC analysis (Gross et al., 1992). Standard addition quantification was performed by spiking two of four samples with the standard reference spiking mixture mentioned above. Identification and quantification o f HAs The extracts were evaporated to dryness under a stream of nitrogen and thereafter dissolved in 100 Itl
caffeine solution, which served as an internal standard. Aliquots (90pl) of the samples were injected (Varian 9100 Autosampler) into a Varian 9010 liquid chromatograph equipped with a ToyoSoda TSK Gel ODS 80TM column (250 x 4.6 mm i.d., 5-/~m particle size, Varian, Stockholm, Sweden) and a precolumn (Supelguard LC-18-DB, 20 x 4.6 mm i.d., Labkemi, Stockholm, Sweden). The chromatographic conditions were adopted from Gross (1990) with minor modifications. The mobile phase was: solvent A and B, 10 mM triethylamine adjusted with acetic acid to pH 3.2 and 3.6, respectively, and solvent C, acetonitrile. The effluent was monitored using a photodiode array UV detector (Varian 9065, Polychrom) and a programmable fluorescence detector (Varian LC 9070). The identities of the compounds were established by comparing the retention times of the peaks with those of the corresponding spiked samples run under the same conditions. Diode array UV spectra of the peaks were compared with the spectra of the spiked samples and with library entries. The HAs were quantified by the standard addition quantification method using peak areas of the unspiked and spiked samples. The areas were compared with the areas of known amount of standards. The results were corrected for incomplete recovery. All results are expressed in ng compound per g cooked beef. The contents in the pan residues are correlated to the amount of cooked meat.
Effects of frying fat on the formation of HAs
PkLte 1. Photograph of beefburgers fried in margarine at 165°C (left) and 200°C (right).
997
Effects of frying fat on the formation of HAs
999
P ~<0.05) using the software package SPSS for MS Windows 5.0 (SPSS Inc., Chicago, USA). The correlation between HA content in the samples and oxidation status, antioxidant content and chemical composition of the frying fats was statistically evaluated using partial least squares analysis (PLS) and the software SIRIUS (Pattern Recognition System A/S, Bergen, Norway). The procedure of cross-validation was used to determine the significance of the PLS components (Wold, 1978).
Analysis o f oxidation status and antioxidant content & the fat The different fats were analysed for peroxide value (PV), anisidine value (AV), free fatty acids (FFA), vitamin E (tocopherols and tocotrienols) and vitamin A (retinol) content using standard methods (the analyses were performed at Van den Bergh Foods using methods standardized by Unilever), before being heated and after heating for 10 min at 200°C with and without beefburgers. The fl-carotene content was determined according to Sadler et al. (1990). About 10 g fat was extracted using heptane-acetone-ethanol (50:25:25, by vol) and stirred for 15 min. Water (15 ml) was added and the sample was stir'red for 10min. The non-polar phase was evaporated to dryness, redissolved and analysed using HPLC. The samples were isocratically chromatographed ,an a LiChrospher 100RP-18 column (250 x 4ram i.d., 5-/~m particle size) equipped with a precolumn, using methanoltetrahydrofuran-wa~Ier (67:27:6, by vol) as a mobile phase. The flow rate was 2 mltmin and the effluent was monitored at 4'75 nm using a UV detector.
RESULTS
HAs in beefburgers The HAs MelQx, DiMelQx, PhlP, norharman and harman were recovered in the meat samples. The amounts of MelQx, DiMelQx and PhlP detected in the beefburgers fried in different frying fats are shown in Table 2. The amounts of these HAs were very low or not detected in the beefburgers fried at 165°C. At 200°C the amounts increased, but were still low. Formation of PhlP showed a pronounced increase with temperature. A significantly higher amount (1.6 ng/g) of MelQx was found in beefburgers fried in rapeseed oil than in beefburgers fried in any of the other fats. The highest amount of DiMelQx (0.4ng/g) was found in beefburgers fried in margarine, rapeseed oil and liquid margarine, while beefburgers fried in margarine fat phase contained the highest amount of PhlP. Norharman and harman
Statistics The amounts of HAs detected in meat and pan residues after frying: beefburgers in different frying fats were statistic~Llly compared using one-way ANOVA followed by Duncan's test (significance level
180
,.! /"
I,o-|
, s . . ~. o ' ° t 160
..'.-2:: ...............
"" ";:;';"
v
o.-" •
140
.
,,..;-,:-. • .',
i
.
120
/i,. ,:',,'"/ .........
'l
:.,,.... ............,.....
t-,
i ,*"
",, ,..~
/
!
s
.,. ...
,"
,/
,,
,
• " ..........
•..
100
I
I
I
3
I
I
4
I
I
5
I
I
6
I
I
I
7 Time (rain)
I
8
I
I
9
I
I
10
I
I
II
I
I
12
Fig. 2. Time-temperature profiles during frying of eight batches of beefburgers in rapeseed oil at a fat temperature of 200°C. The temperature was measured every 30 sec using a chrom/alumel thermocouple placed in the fi'ying fat between the pan and the beefburger. The variation in temperature is due to the variations in transport of water from the meat into the pan. The beefburgers were turned at 7 min.
1000
M . A . E . Johansson et al.
were detected at low ppb levels (0.5-3.2 ng/g and 0.1-0.7ng/g, respectively) in all samples fried at 200'~C (data not shown). HAs in pan residues
As shown in Table 2, the content of HAs was generally higher in the pan residues than in the beefburgers. The amounts of MeIQx ranged from 2.1 ng/g, in pan residue from frying in sunflower seed oil, to 4.3 ng/g in liquid margarine. The highest amount of DiMeIQx was found in pan residue from frying with margarine. The amount detected in pan residue from frying in sunflower seed oil was significantly lower than after frying with any other oil. Of all the HAs detected, PhIP was present in the highest amounts in pan residues, with levels ranging from 2.0 ng/g after frying in sunflower oil to 13.3 ng/g after cooking using margarine fat phase. A low level of PhIP in pan residues was detected after frying in margarine, while frying in margarine fat phase, liquid margarine or rapeseed oil produced significantly higher amounts. Norharman and harman were detected at low ppb levels in the pan residues (data not shown). The lowest total amount of HAs in the pan residues were detected after frying in sunflower seed oil (4.7 ng/g) or margarine (8.6 ng/g). Frying in margarine fat phase, liquid margarine, butter and rapeseed oil produced significantly higher amounts. The total content of HAs in meat plus pan residue was significantly lower after frying in sunflower seed oil or margarine than after frying with the other fats. IQ, MeIQ, GIu-P-I and GIu-P-2 were not detected in any sample. No HAs were detected in margarine or rapeseed oil heated for 10 rain at 200°C without meat. Oxidation status and antioxidant content o f the frying fats
Table 3 shows the results from the analysis of the frying fats for PV, AV, FFA, vitamin A,/~-carotene and vitamin E: ~-T (~-tocopherol), c~-T3 (~-tocotrienol),/~-T, 7-T, i'-T3 and 6-T. Frying in margarine, margarine fat phase, liquid margarine and butter increased PV and AV moderately. Considerably higher PV and AV were found in rapeseed and sunflower seed oil after frying at 200°C, while PV and AV were very low after frying in margarines and butter. F F A was high in unheated butter, but did not increase after frying with meat. Generally, AV and F F A increased less than expected after frying with meat. The degree of oxidation of fats heated without meat increased, as would be expected. The vitamin A and /~-carotene contents were reduced to about half the original amounts in the heated fats. Only trace amounts of ct- and 7-tocopherol were retained in the oils after frying, whereas two-thirds of the original values remained in margarines and butter.
Correlation between HA formation, oxidation status and antioxidant content
The observed differences in MelQx and DiMeIQx formation can be explained in terms of oxidation status (PV and AV) and antioxidant content (vitamin A, vitamin E and tocopherols/tocotrienols) in the frying fat, using partial least squares analysis. The first and second PLS components of the above mentioned variables explained most of the differences in MelQx and DiMelQx formation. One PLS component explained 53% of the variation in MelQx in the pan residues, while two PLS components explained 83% of the MelQx variation in the beefburgers and 74% of the DiMelQx variation in the pan residues. However, these variables did not explain the formation of PhlP. The correlation coefficients for measured and predicted levels of MelQx were r---0.95 and r = 0.79 in the beefburgers and pan residues, respectively. The corresponding coefficient for DiMelQx in the pan residue was r = 0.89. The DiMelQx content of the beefburgers was too low for statistical evaluation. The results obtained could not be explained by differences in fat composition or any of the other parameters measured. Identification and quantification
Identification and quantification of HAs in the samples were performed by HPLC with on-line UV spectra for peak identification and purity. Peak identification was achieved by comparison of retention time and UV spectra with spiked samples and library entries. Figure 1 shows the UV chromatogram at 263 nm after HPLC analysis of a pan residue sample obtained after frying beefburgers at 200C in liquid margarine. The peaks corresponding to MelQx, DiMelQx, PhlP and caffeine (internal standard) are indicated. Figure 1 also shows the on-line recorded UV spectra of MelQx, DiMelQx and PhlP in the sample together with reference spectra. There was good correspondence between the absorbance peak shape obtained from the sample and that of the reference spectra generated using synthetic references. The average recoveries from the purification of the spiked beef samples was 71 + 14%, 63 + 14% and 66 + 16% (mean +_ SD, n = 28) for MelQx, DiMeIQx and PhIP, respectively. For the pan residues, the corresponding values were 76 + 10%, 63 +_ 12% and 51 +_ 16% (n = 28), respectively. The variations between duplicate determinations from duplicate frying experiments is shown in Table 2 (variation is not given for values less than 1 ng/g). The detection limits varied with the compound and the complexity of the sample matrix. In the samples from fried beefburgers MeIQx, DiMeIQx and PhIP could be detected and confirmed by UV absorption spectra at 0.03, 0.02 and 0.01 ng/g, respectively. Each of these three HAs could be detected and confirmed in the pan residues at 0.02 ng/g. Norharman and harman are known to appear both in the polar and
Effects of frying fat on the formation of HAs nonpolar extracts with varying degrees of recovery, and therefore no attempt was made to quantify these compounds. Chemical composition, weight loss, amount o f crust, colour, time-temperature profiles The chemical composition of the uncooked meat was: fat, 11.2%; protein, 19.1%; water, 67.9%; ash, 2.0%; carbohydrates, 0%. The weight losses during cooking were 28.4:_~2.3% ( n = 16) at 165°C and 34.5 _+ 1.4% (n = 96) at 200~'C. The amount of crust was 30.5 + 3.4% (n = 8) at 165°C and 31.6_+ 3.4% (n = 48) at 200~C. 1-he surface of the beefburgers fried at 165C was somewhat pale. The crust formed at 200~C was quite brown and would be judged as appetizing by some and too dark by others. The beefburgers fried in oil were slightly darker than the beefburgers fried in butter or margarine. Plate 1 shows a photograph of beefburgers fried in margarine at 165 and 200' C. Time-temperature profiles for each frying experiment were obtained by measuring the temperature, using a thermocouple placed between the pan and a beefburger, every 30 sec. Temperature differences were observed between the various frying experiments due to uneven transport of water from the meat into the pan. The presence of water lowered the temperature at the position of the thermocouple. Figure 2 shows time-temperature profiles obtained from frying eight ba':ches of beefburgers in rapeseed oil at 200~C. DISCUSSION
Our study is the first in which quantitative data on the influence of different frying fats on the formation of HAs in beefburgers and pan residues are reported. The type of frying fat had only minor effects on the formation of HAs irL fried beefburgers. However, it did influence the amount of HAs in the pan residue. The lowest total amount of HAs was obtained after frying in sunflower seed oil or margarine. The results are in agreement with a previous study performed at our laboratory, in which pan residues from frying beefburgers in butter were shown to contain significantly higher amounts of MelQx and DiMelQx than samples fried in vegetable oil (Johansson and J/igerstad, 1994). However, different HAs might be formed under different cooking conditions, and since there are variations in the carcinogenic potency of the different HAs, more research is needed before any particular frying fat can be recommended. The amounts of HAs are consistent with those reported previously for beefburgers cooked under similar conditions (Gross, 1990; Gross et al., 1993; Johansson and J/igerstad, 1994; Knize et al., 1994; Murray et al., 1993; Skog et al., 1995). The results of the present study, together with those of model experiments (Johansson and J/igerstad, 1993), leave no doubt that the formation of HAs is, to some extent, influenced by the frying fat.
1001
By using the PLS method, the observed variations in MelQx and DiMelQx formation can be explained in terms of antioxidant content (vitamin A, vitamin E and tocopherols/tocotrienols) and oxidation status of the frying fats (PV and AV). The content of antioxidative sterols in un-dewaxed sunflower seed oil might have had some inhibiting effects on the formation of HAs, but this was not investigated. Differences in fat composition did not explain the variation in HA formation. Interestingly, the variation in PhIP formation was not explained by the PLS components of any of the measured variables. According to the 'Las Vegas-hypothesis' IQ-compounds, such as MelQx and DiMelQx, are formed through the condensation of naturally occurring creatine and pyrazines and Strecker aldehydes, which are formed from free amino acids and monosaccharides in the Maillard reaction (J~igerstad et al., 1983). The present results show that antioxidants, such as tocopherols, tocotrienols and retinol, might reduce the formation of IQ-like HAs, probably by scavenging free radicals in the early stage of the Maillard reaction, which is known to be enhanced by free radical reactions (Namiki and Hayashi, 1981). The results of the present study are further supported by our observation of a decrease in the formation of IQx, MeIQx and DiMeIQx when antioxidants, such as y-tocopherol and /~-carotene, were incorporated into a model system containing pro-oxidants (Johansson and J~igerstad, 1995). PhIP was the major HA formed in the pan residues. The mechanism for the formation of PhIP is still not known in detail, but in contrast to the IQ-compounds, pyrazines/pyridines are not required. Obviously, the formation of PhIP is not associated with free radical reactions and is not inhibited by free radical scavenging antioxidants. The observed variations in PhlP formation after frying in different fats could not be explained in terms of any of the measured parameters and further experiments are needed to explain the variation. There was some variation between the meat and pan residues in the content of the different HAs. For example, while the highest levels of MeIQx were found in meat fried in rapeseed oil, the levels of MeIQx and other HAs in the corresponding pan residues were by no means the highest. Also, whereas for all fats the levels of MeIQx in the pan residues were two-four-fold those in the meat, the levels of PhIP in the residues were generally about 10-fold more than in meat. This might suggest that the various HAs are formed in different phases of the meat and that the production of these phases might be influenced by the particular fat used for the frying. The degree of fat oxidation during heating in thin layers, as occurs in shallow pan frying, is dependent on the content of polyunsaturated fatty acids and milk. In addition, the frying technique (e.g. temperature, heating time and thickness of the fat layer) have great influence (Fredholm et al., 1990). The frying
1002
M.A.E. Johansson et al.
fats used in the present study were chosen on the basis of their differences in fatty acid composition and other constituents, especially antioxidants and milk. Among the six investigated frying fats, margarine, butter and rapeseed oil are commonly used in cooking. Sunflower seed oil was selected because of its high content of polyunsaturated fatty acids and instability towards oxidation during heating. The liquid margarine differed from rapeseed oil by containing milk, which has been shown to exert antioxidative activity (Fredholm et al., 1990). The margarine fat phase was taken out before adding the water phase, which contained whey, during the manufacturing of margarine. Our study is one of the first reporting on fat oxidation and loss of antioxidants in the frying fat during shallow pan frying of meat. As expected, butter and the margarines showed rather modest oxidation, while rapeseed and sunflower seed oil were greatly oxidized. An AV close to 100 was obtained for both of the oils and an off-taste is generally noted at AV over 25. The loss of antioxidants was reflected by the degree of fat oxidation, and for the moderately oxidized fats (e.g. margarines and butter) about half of the content of antioxidants was retained after frying. The two oils lost their vitamin E activity completely. The presence of whey in the solid margarine had no effect on the fat oxidation or retention of antioxidants. However, the presence of milk in the liquid margarine had an obvious effect on the fat oxidation and the retention of antioxidants: the oxidation was about 10-fold lower in the liquid margarine than in the rapeseed oil. About half of the antioxidant content was retained in the liquid margarine after heating; the rapeseed oil lost all of its antioxidants. When solid margarine and rapeseed oil were heated without the presence of meat, both were more oxidized than when heated with meat. In rapeseed oil this difference was very big. As meat is more likely to contain pro- than antioxidants the differences in oxidation must be due to other factors such as fat temperature, depth of the frying fat and water evaporation-three circumstances that change considerably when beefburgers are added to the fat. These changes correspond to a decrease of 25-30°C in the frying temperature. It is also interesting to note that when the margarine was heated without meat the retention of antioxidants was much higher than after heating with meat. This was not the case when heating rapeseed oil. The retention was just as poor when heated without as with meat. Other studies concerning the role of frying fat on the formation of HAs, show inconsistent results (Barrington et al., 1990; Nilsson et al., 1986) and are chiefly based on Ames test data. However, unsaturated long-chain fatty acids, such as oleic and linoleic acid, are known to inhibit the Ames test, making the results uncertain (Hayatsu et al., 1981). Besides the chemical effects of fat shown in the present study,
the physical effects of fat on the formation of HAs is of great importance. In most reported studies, it is difficult to distinguish between the physical and chemical effects of the fat. Lower mutagenic activity was detected in pork fried without frying fat, which was explained by a more efficient heat transfer when frying fat is used (Nilsson et al., 1986). Similarly, when oven-baking meat loaves a shorter cooking time was needed to achieve a centre temperature of 7 0 C with products with a high fat content than with those containing less fat (Holtz et al., 1985). The results indicated that the lowest mutagenic activity was produced in high-fat meat which was cooked for the shortest time. Another study explained the observed reduction in mutagenic activity in high-fat products as being the result of the dilution of precursors (Knize et al., 1985). HAs are mainly formed in the crust of cooked food (Bjeldanes et al., 1983; Commoner et al., 1978). In order to increase sensitivity and lower the detection limit, we used only the crusts of the beefburgers. Consequently, the amounts of HAs might be slightly underestimated since the presence of small amounts of HAs in the inner parts of fried meat have been reported (Skog et al., 1995; Tikkanen, 1989). In agreement with earlier studies performed at our laboratory (Johansson and Jfigerstad, 1994; Skog et al., 1995), the content of HAs was higher in pan residues than in the corresponding beefburgers. In Sweden, pan residues are commonly served with fried meat dishes (Asp et al., 1992) and the contribution from pan residues should be included when the dietary intake of HAs is assessed. On the basis of our findings, the human exposure to MeIQx, DiMeIQx and PhlP from a diet consisting of 150-200 g fried meat, including gravy prepared from the pan residue, may range from less than 0.5 to 4.0/~g/person/day, depending on cooking conditions. Levels of total HA were 10-fold higher after frying at 200~C than at 165°C. Experts recommend a frying temperature of 155-175°C when frying beefburgers (K6ttinformation, 1986). In the present study, a higher frying temperature was chosen to obtain clear results regarding the role of frying fats on the formation of HAs. However, as can be seen from the photographs, the beefburgers were acceptable, although somewhat too brown and overcooked. Since domestic cooking equipment lacks thermostatically controlled frying surfaces, the temperature varies within a wide range while frying. It is almost impossible to maintain a steady desired temperature and overheating occurs often. IARC (1986) recommends a reduction in the daily intake of HAs. In order to do so, there is a need to develop temperature-controlled cooking equipment and to advise consumers on how to use it. Another important factor is to change people's attitude towards less browned fried food. It is also important to point out that a higher frying temperature does not lead to a shorter frying time (Dagerskog, 1978). The daily intake of HAs can also be
Effects of frying fat on the formation of HAs reduced simply by discarding the p a n residue after frying. In addition, efforts to reduce H A f o r m a t i o n require a better u n d e r s t a n d i n g of the reactions behind the formation. Acknowledgements~Ne gratefully acknowledge the excellent technical assistance of Ms Ingrid Berntsson and Gunnel Andersson, Department of Applied Nutrition and Food Chemistry, and Ms Eva Ahlbin, Van den Bergh Foods. This study was supported financially by the Swedish Cancer Foundation (1824-B93-12XBC) and the Swedish Council for Forestry and Agricultural Research (50.00440/91). REFERENCES
Adamson, R. H., Tl~orgeirsson U. P., Snyderwine E. G., Thorgeirsson S. S., Reeves J., Dalgard D. W., Takayama S. and Sugimura T (1990) Carcinogenicity of 2-amino-3methylimidazo[4,5-f]quinoline in nonhuman primates: induction of tumors in three macaques. Japanese Journal of Cancer Research 811, 10-14. Asp N. G., Isaksson ,~., Johansson U., JS_gerstad M., Lanke J., Mattisson I., Swederberg E. and Scherstrn B. (1992) Dietary habits among adults in Olofstrrm--a baseline study. Scandinavian Journal of Nutrition/Niiringsforskning 36, 106-114. Barnes W. S., Maher J. C. and Weisburger J. H. (1983) High-pressure liquid chromatographic method for the analysis of 2-amino-3-methylimidazo[4,5-f]quinoline, a mutagen formed during the cooking of food. Journal of Agricultural and Food Chemistry 31, 883-886. Barrington P. J., Baker R. S. U., Truswell A. S., Bonin A. M., Ryan A. J. and Paulin A. P. (1990) Mutagenicity of basic fractions derived from lamb and beef cooked by common household methods. Food and Chemical Toxicology 28, 141-146. Bjeldanes L. F., Morris M. M., Timourian H. and Hatch F. T. (1983) Effecl:s of meat composition and cooking conditions on mutagen formation in fried ground beef. Journal of Agricultural and Food Chemistry 31, 18 21. Brittebo E. B., Karlsson A. /~., Skog K. I. and J~igerstad I. M. (1994) Transfer of the food mutagen PhlP to foetuses and newborn mice following maternal exposure. Food and Chemicat Toxicology 32, 717-726. Commoner B., Vihayathil A. J., Dolora P., Nair S., Madyastha P. anti Cuca G. C. (1978) Formation of mutagens in beef arLd beef extract during cooking. Science 2111, 913 916. Dagerskog M. (1978) Stekning av livsmedel, studier av det processtekniska underlaget. Institutionsavdelningen frr livsmedelsvetenskap och SIK--Svenska Livsmedelsinstitutet. Chalmers Tekniska Hrgskola. Felton J. S. and Knize M. G. (1990) Heterocyclic-amine mutagens/carcinogens in foods. In Handbook of Experimental Pharmacology. Edited by C. S. Cooper and P. L. Grover. pp. 471-502. Springer-Verlag, Berlin. Fredholm L., Gustafi;son I.-B. and Jonsson L. (1990) Vad h~inder med matfetterna vid stekning och fritering. Ndringsforskning 35, 132-140. Ghoshal A. and Snyderwine E. (1993) Excretion of foodderived heterocyclic amine carcinogens into breast milk of lactating rats and formation of DNA adducts in newborn. Carcinogenesis 14, 2199-2203. Gross G. A. (1990) Simple methods for quantifying mutagenic heterocyclic aromatic amines in food products. Carcinogenesis 11, 1597-1603. Gross G. A., Gruter A. and Heyland S. (1992) Optimization of the sensitivity of high-performance liquid chromatography in the detection of heterocyclic aromatic amine mutagens. Food and Chemical Toxicology 30, 491~,98.
1003
Gross G. A., Turesky R. J., Fay L. B., Stillwell W. G., Skipper P. L. and Tannenbaum S. R. (1993) Heterocyclic aromatic amine formation in grilled bacon, beef and fish and in grill scrapings. Carcinogenesis 14, 2313-2318. Hayatsu H., Arimoto S., Togawa K. and Makita M. (1981) Inhibitory effect of the ether extract of human feces on activities of mutagens: inhibition by oleic and linoleic acids. Mutation Research 81, 287 293. Holtz E., Skjrldebrand C., J~igerstad M., Laser Reutersw~ird E. and Isberg P. E. (1985) Effects of recipes on crust formation and mutagenicity in mean products during baking. Journal of Food Technology 20, 57~56. IARC (1986) IARC Monographs of the Evaluation of the Carcinogenic Risk of Chemicals to Humans. Vol. 40. Some Naturally by Occurring and Synthetic Food Components, Furocomarins and Ultra Violet Radiation. International Agency for Research on Cancer, Lyon. Johansson M. and J~igerstad M. (1993) Influence of oxidized deep-frying fat and iron on the formation of food mutagens in a model system. Food and Chemical Toxicology 31, 971-979. Johansson M. A. E. and J~igerstad M. (1994) Occurrence of mutagenic/carcinogenic heterocyclic amines in meat and fish products, including pan residues, prepared under domestic conditions. Carcinogenesis 15, 1511-1518. Johansson M., Skog K. and J~igerstad M. (1993) Effects of edible oils and fatty acids on the formation of mutagenic heterocyclic amines in a model system. Carcinogenesis 14, 89 94. J~igerstad M., Laser Reutersw~ird A., Oste R., Dahlqvist A., Grivas S., Olsson K. and Nyhammar T. (1983) Creatinine and Maillard reaction products as precursors of mutagenic compounds formed in fried beef. In The Maillard Reaction in Food~ and Nutrition. Edited by G. R. Waller and M. S. Feather. pp. 507-520. ACS Symposium Series 215. American Chemical Society, Washington, DC. J~gerstad I. M., Johansson M. A., Karlsson A. ,~., Skog K. I., Oskarsson A., Palminger-Hallen I., Ivhed R. and Brittebo E. B. (1994) Dose-dependent milk transfer and tissue distribution of the food mutagen PhlP in rats and their suckling pups. Carcinogenesis 15, 2479-2484. Knize M. G., Andresen B. D., Healy S. K., Shen N. H., Lewis P. R., Bjeldanes L. F., Hatch F. T. and Felton J. S. (1985) Effects of temperature, patty thickness and fat content on the production of mutagens in fried ground beef. Food and Chemical Toxicology 23, 1035-1040. Knize M. G., Dolbeare F. A., Carroll K. L., Moore D. H. and Felton J. S. (1994) Effect of cooking time and temperature on the heterocyclic amine content of fried-beef patties. Food and Chemical Toxicology 32, 595~03. Krttinformation (1986) KrttR~itt. ISBN 91-7766-021-8. pp. 136-137. (In Swedish) Murray S., Lynch A. M., Knize M. G. and Gooderham N. J. (1993) Quantification of the carcinogens 2-amino-3, 8-dimethyl- and 2-amino-3,4,8-trimethylimidazo[4,5-f]quinoxaline and 2-amino-l-metyl-6-phenylimidazo[4,5-b]pyridine in food using a combined assay based on gas chromatography negative ion mass spectrometry. Journal of Chromatography 616, 211-219. Namiki M. and Hayashi T. (1981) Formation of novel free radical products in an early stage of Maillard reaction. In Progress in Food and Nutrition Service. Iiol. 5. Maillard Reactions in Food. Edited by C. Eriksson. pp. 81-92. Pergamon Press, Oxford. Nilsson L., Overvik E., Fredholm L., Levin 0., Nord C.-E. and Gustafsson J.-,~. (1986) Influence of frying fat on mu~agenic activity in lean pork meat. Mutation Research 171, 115-121. Norell S. E., Ahlbom A., Erwald R., Jacobson G., Lindberg-Navier I., Olin R., Trrnberg B. and Wiechel K. L. (1986) Diet and pancreatic cancer: a case-control study. American Journal of Epidemiology 124, 894-902.
1004
M. A. E. Johansson et al.
Ohgaki H., Takayama S. and Sugimura T. (1991) Carcinogenicities of heterocyclic amines in cooked food. Mutation Research 259, 399-412. Overvik E., Nilsson L., Fredholm L., Levin 0., Nord C.-E. and Gustafsson J.-/~. (1987) M utagenicity of pan residues and gravy from fried meat. Mutation Research 187, 47-53. Sadler G., Davis J. and Dezman D. (1990) Rapid extraction of lycopene and fl-carotene from reconstituted tomato paste and pink grapefruit homogenates. Journal of Food Science 55, 1460-1461. Schiffman M. H. and Felton J. S. (1990) Fried foods and risk of colon cancer. Letter to the Editor. American Journal of Epidemiology 131, 376-378. Skog K. (1993) Cooking procedures and food mutagens: a literature review. Food and Chemical Toxicology 31, 655~75. Skog K., Steineck G., Augustsson K. and Jiigerstad M. (1995) Effect of cooking temperature on the formation of heterocyclic amines in fried meat products and pan residues. Carcinogenesis 16, 861 867. Steineck G., Gerhardsson de Verdier M. and Overvik E. (1993) The epidemiological evidence concerning intake of mutagenic activity from fried surface and the risk of cancer cannot justify preventive measures. European Journal of Cancer Prevention 2, 293 300. Steineck G., Hagman U., Gerhardsson M. and Norell S. E.
(1990) Vitamin A supplements, fried foods, fat and urothelial cancer. A case-referent study in Stockholm in 1985-87. International Journal of Cancer 45, 100(~1011. Sugimura T., Nagao M., Kawachi T., Honda M., Yahagi T., Seino Y., Sato S., Matsukara N., Shirai A., Sawamura M. and Matsumoto H. (1977) Mutagens~arcinogens in food, with special reference to highly mutagenic pyrolytic products in broiled foods. In Origins of Human Cancer. Edited by H. H. Hiatt, J. D. Watson and J. A. Winsten. pp. 1561-1577. Cold Spring Harbor Laboratory, New York. Tikkanen L. (1989) Mutagenicity of cooked foods. VTT, Esboo. Tikkanen L. M., Sauri T. M. and Latva-Kala K. J. (1993) Screening of heat-processed Finnish foods for the mutagens 2- amino- 3, 8 - dimethylimidazo[4, 5 - f ]quinoxaline, 2-amino-3,4,8-trimethylimidazo[4,5-J]quinoxaline and 2amino- l-methyl-6-phenylimidazo[4,5-b]pyridine. Food and Chemical Toxicology 31, 717-721. Willett W. C., Stampfer M. J., Colditz G. A., Rosner B. A. and Speizer F. E. (1990) Relation of meat, fat and fiber intake to the risk of colon cancer in a prospective study among women. New England Journal of Medicine, 1664-1672. Wold S. (1978) Cross-validatory estimation of the number of components in factor and principal component models. Technometrics 20, 397-405.
~
Pergamon
0278-6915(95)00074-7
Ltd reserved
Copyright © 1995 ElsevierScience Printed in Great Britain. All fights 0278-6915/95 $9.50 + 0.00
Research Section
Influence of Frying Fat on the Formation of Heterocyclic Amines in Fried Beefburgers and Pan Residues M. A. E. J O H A N S S O N * , L. F R E D H O L M t , I. B J E R N E t a n d M. J A G E R S T A D * *Department of Applied Nutrition and Food Chemistry, Chemical Centre, Lund University, PO Box 124, S-221 00 Lund and iVan den Bergh Foods, PO Box 721, S-251 07 Helsingborg, Sweden (Accepted 30 Mav 1995)
Abstract--The influence of six frying fats (butter, margarine, margarine fat phase, liquid margarine, rapeseed oil and sunflower seed oil) on the formation of mutagenic/carcinogenic heterocyclic amines (HAs) during the frying of beefburgers was investigated. Frying was performed at 165 and 200":'C (i.e. under conditions that represented normal household cooking practices). The fried beefburgers and their corresponding: pan residues were purified using solid-phase extraction and analysed for HAs using HPLC with photodiode array UV and fluorescence detection. The HAs 2-amino-3,8-dimethylimidazo[4,5-f]quinoxaline (MelQx), 2-amino-3,4,8-trimethylimidazo[4,5-f]quinoxaline (DiMelQx), 2-amino-l-methyl-6phenylimidazo[4,5-b]pyridine (PhlP), 9H-pyrido[3,4-b]indole (norharman) and l-methyl-9Hpyrido[3,4-b]indole (harman) were recovered. The amount increased with the temperature, and the content of HAs in the pan residue was much higher than in the corresponding beefburger. The amounts of MelQx ranged from 0.2 to 1.6 ng/g in the beefburgers and from 0.8 to 4.3 ng/g in the pan residues. DiMelQx ranged from undetectable to 0.4 ng/g in the beefburgers and from 0.4 to 1.3 ng/g in the residues. PhlP ranged from C,.08 to 1.5 ng/g in the meat and from 0.4 to 13.3 ng/g in the residues. The total amount of HAs in meat and pan residue combined was significantly lower after frying in sunflower seed oil or margarine th~.n after frying with the other fats. The observed differences in MelQx and DiMelQx formation could be explained in terms of oxidation status (peroxide and anisidine value) and antioxidant content (vitamin A, vitamin E and tocopherols/tocotrienols) using partial least squares analysis.
INTRODUCTION Several epidemiolo~ical studies have shown a relationship between the consumption of meat and an increased risk of colorectal cancer and possibly also of breast, stomach, pancreatic and urinary bladder cancer (Norell et al., 1986; Schiffman and Felton, 1990; Steineck et al., 1990 and 1993; Willet et al., *Author for corresponAence. AV = anisidine value; DCM = dichloromethane; 4,8-DiMelQx = 2-amino-3,4,8-trimethylimidazo[4,5-f]quinoxaline; FFA = free fatty acids; GIu-P- 1 = 2-amino-6-methyl-dipyrido[l,2-a :3',2'-d]imidazole; GIu-P-2 = 2-aminodipyrido[l,2-a : 3',2'-d]imidazole; harman = l-methyl-9H-pyrido[3,4-b]indole; HA(s) = heterocycli¢ amine(s); IQ = 2-amino-3methylimidazo[4,5-/]quinoline; MelQ = 2-amino-3,8dimethylimidazo[4,5-f ]quinoline; MelQx = 2-amino3,8-dimethylimidazc,[4,5-f]quinoxaline; norharman = 9H-pyrido[3,4-b]indole; PhlP = 2-amino-l-methyl-6phenylimidazo[4,5-b]pyridine; PLS = partial least squares; PRS = propylsulfonic acid-silica gel; PV= peroxide value; c~-T = c~-tocopherol; ~-T~= ~tocotrienol; fl-T =/?-tocopherol; ,/-T = y-tocopherol; 7T 3 = y-tocotrienol; ii-T = 6-tocopherol.
Abbreviations:
1990). The formation of mutagenic/carcinogenic heterocyclic amines (HAs) during the cooking of meat and fish products may be responsible for part of the observed increased cancer risk. Since first discovered by Sugimura et al. (1977), around 20 mutagenic HAs have been isolated from cooked foods, primarily from the crust of fried meat and fish, pan residues and beef extracts (Felton and Knize, 1990; Skog, 1993). All HAs tested have shown carcinogenic effects in long-term experiments in rodents (for a review, see Ohgaki et al., 1991) and 2-amino-3-methylimidazo[4,5-f]quinoline (IQ) also induces tumours in primates (Adamson et al., 1990). In addition, a transfer of 2-amino- 1-methyl-6-phenylimidazo[4,5-b]pyridine (PhlP) to the foetus and its excretion in the breast milk of mice and rats following maternal exposure has recently been reported (Brittebo et al., 1994; Ghoshal and Snyderwine, 1993; J/igerstad et al., 1994). HAs are formed from creatine, free amino acids and monosaccharides (which occur naturally in protein-rich foods of animal origin) at normal cooking temperatures (150-250°C) (for review, see Skog, 1993). Earlier studies have shown H A formation to 993
994
M . A . E . Johansson et al.
be dependent on physical parameters, such as cooking temperature and time, cooking equipment and technique, water activity, heat transport, and chemical parameters, especially the precursors, for example, carbohydrates, free amino acids and creatine (Johansson and J/igerstad, 1994; Knize et al., 1994; Skog, 1993; Skog et al., 1995; Tikkanen et al., 1993). The role of fat in the formation of HAs has not been sufficiently clarified. The few studies that have been carried out so far on the influence of frying fat or fat content on the formation of food mutagens have yielded inconsistent results. The fat content of the meat probably has some effect on HA formation, although it is difficult to distinguish between the physical and chemical effects (Barnes et al., 1983; Holtz et al., 1985; Knize et al., 1985; Overvik et aL, 1987). Modelling experiments have shown that certain oils increase the yield of HAs, while the oxidation status of the fat and its content of antioxidants showed no effects (Johansson and JS.gerstad, 1993; Johansson et al., 1993). The use of different frying fats did not affect the mutagenic activity in pork fried at 200°C (Nilsson et al., 1986). However, at 250°C, pork fried in sunflower seed oil or butter had the highest mutagenic activity, while pork fried in lard or olive oil had the lowest. When frying lamb the highest mutagenic activity was detected when butter was used (Barrington et al., 1990). Most reported data concerning the influence of fat on the formation of HAs are based on mutagenic activity. The only published study that reported quantitative data showed that pan residues from frying beefburgers in butter contained significantly higher amounts of 2-amino-3,8-dimethylimidazo[4,5-f]quinoxaline (MelQx) and 2-amino-3,4,8-trimethylimidazo[4,5-f]quinoxaline (DiMelQx) than samples fried in vegetable oil (Johansson and J~igerstad, 1994). In the present study, we quantitatively investigated the influence of different frying fats on the formation of HAs in fried beefburgers and their corresponding pan residues. Six fats with different chemical compositions were tested during frying under well defined cooking conditions. The oxidation status and antioxidant contents in the frying fats were determined after frying and the relationship between these parameters
and HA formation was statistically evaluated using partial least squares analysis (PLS). MATERIALS AND METHODS
Chemicals
All chemicals and solvents were of HPLC or analytical grade. Water was obtained from a Milli-Q water purification system (Millipore, Bedford, MA, USA). The solvents acetonitrile, methanol, dichloromethane (DCM), tetrahydrofuran, heptane, acetone and toluene, were purchased from Merck AG (Darmstadt, Germany). Synthetic IQ, 2-amino-3,8dimethylimidazo[4,5-f]quinoline (MeIQ), MeIQx, 4,8-DiMeIQx, PhIP, 2-amino-6-methyl-dipyrido [1,2-a :3',2'-d]imidazole (Glu-P-1), 2-aminodipyrido[1,2-a : 3',2"-d]imidazole (Glu-P-2) and 1-methyl-9Hpyrido[3,4-b]indole (harman) were obtained from Toronto Research Chemicals (Downsview, Ontario, Canada) and 9H-pyrido[3,4-b]indole (norharman) from Aldrich (Steinheim, Germany). A standard reference mixture of IQ, MeIQ, MeIQx, 4,8-DiMeIQx and PhIP [5 ng of each compound/#l of methanoltriethylamine (50:50, v/v), pH 3.2] was used as a spiking solution. Caffeine (Sigma, St Louis, MO, USA) at 2.5 ng//~l methanol-triethylamine (50:50, v/v), was used as internal standard. The materials used for PRS-tandem extraction (Extrelut and BondElut, PRS and C~8) were obtained from Merck AG and Sorbent (V~stra Fr61unda, Sweden). Fractogel TSK CM650(s) was obtained from Merck AG. Synthetic ]~-carotene was from Sigma. Meat sample
Brisket of beef (Musculus pectoralis superficialis, M. pectoralis profundus and M. seratus ventrilus) obtained from a local slaughterhouse, was minced (3 ram) and homogenized making up one batch of meat. The minced beef was made into burgers (87 mm diameter, 9 mm thick, weight 85 g) using a special punch. The beefburgers were stored at - 2 0 ° C until fried. The raw beef was analysed for water, fat, protein and ash using standard methods as previously described (Holtz et al., 1985). The amount of carbohydrates was calculated as residue after accounting for water, fat, protein and ash.
Table 1. Fatty acid composition of frying fats Fatty acid content (%)
Fatty acid ~
Butter
Margarine and margarine fat phase
Liquid margarine and rapeseed oil
Sunflower seed oil
7.5 3.0 11.5 28.0 11.0 28.5 2.5 -I 7
1.3 4.6 2.0 17.9 7.1 52.3 10.2 2.9 1.7 --
---4.8 1.8 63.1 19.6 8.0 2.3 0.4
---7.0 5.0 23.0 63.0 0.5 1.5 --
Effects of frying fat on the formation of HAs
995
Table 2. Contents of HAs in fried beefburgersand pan residues MelQx (ng/g)* Frying fat
F:ying temperature ('C) Meat
Margarine Margarine fat phase Liquid margarine Butter Rapeseed oil Sunflower seed oil
165 200 200 200 200 200 200
0.2 1.0~ 1.2 ± 0.1~ 1.1 ±0.1" 1.2 + 0.2~ 1.6 ± 0. Ib 1.2 ± 0.2~
DiMelQx (ng/g)*
PhlP (ng/g)*
Total amount of NAs (ng/g)*
Pan residue Meat Pan residue Meat
Pan residue Meat
Meat + Pan residue pan residue
0.8 2.7_+0.3~b 3.5 ± 0.3~ 4.3 _+0.5d 3.2 + 0.9~ 2.8 ± 0.2~ 2.1 ± 0.4~
0.4 4.6+0.9a 13.3+3.6 b 9.7_+2.3~ 11.7+3.4 ~ 11.4_+3.6~ 2.0+0.5 a
1.6 8.6±1.0 ~ 17.8+3.3 b 15.9+2.7 b 15.1+2.6 b 14.8±4.3 b 4.7±0.9 ~
nd 0.4~ 0.2b 0.4~b 0.2~ 0.4~ 0.2b
0.4t 1.3_+0.1" 1.0~ I.I _+0.2~¢ 1.0b~ 0.8b 0.5d
0.08 0.5~ 1.5_+0.6b 1.0~b 1.0~b 1.1+0.2 ~bt~l 0.9~
0.28 1.9±0.2 ~ 3.0+0.7 ~ 2.4+0.6 "b 2.5+_0.2a~ 3.2+0.3 ¢ 2.3±0.5 ~b
1.9 10.5±1.1 ~ 20.7+3.9 b 18.3_+2.1b 17.6+2.6 b 18.2+3.6 b 7.0±1.4 ~
nd = not detected *Content as ng/g cooked meal, recovery-correctedvaluesof duplicatedeterminationsfrom duplicate fryingexperiments.SDs are not givenfor HA values ~<1 ng/g. SD correspond:~ to four determinations. ?Single value. Values in the same column that do not share a common superscript (a-d) are significantlydifferent (one-way ANOVA, followed by Duncan's test; P < 0.05).
Frying fats
in t he f r y i n g fat w a s r e c o r d e d u s i n g a c h r o m / a l u m e l
T h e f o l l o w i n g six fi'ying f a t s w e r e used: b u t t e r ( A B S v e n s k t S m r r , S w e d e n ) , m a r g a r i n e , m a r g a r i n e fa t p h a s e , l i q u i d m a r g a r i : a e , r a p e s e e d a n d s u n f l o w e r seed oil ( V a n d e n B e r g h F o o d s , S w e d e n ) . T h e m a r g a r i n e , liquid margarine and butter contained 800 fat, w h e r e a s t h e m a r g a r i l a e fat p h a s e a n d t h e oils c o n t a i n e d 1 0 0 % fat. T h e m a r g a r i n e a n d t h e m a r g a r i n e
t h e r m o c o u p l e b e f o r e a d d i n g t he b e e f b u r g e r s . F o u r b e e f b u r g e r s (340 g) w e r e fried a t a t i m e , u s i n g 50 g m a r g a r i n e o r b u t t e r , or 40 g m a r g a r i n e fa t p h a s e o r oil, r e s p e c t i v e l y . T h e fat w a s p l a c e d in a p r e h e a t e d p a n . A f t e r 2 m i n t he b e e f b u r g e r s w e r e a d d e d a n d fried for 5 m i n o n o n e side a n d 3 m i n o n t h e o t h e r . T h e t o t a l h e a t i n g t i m e o f t h e fa t w a s 1 0 m i n . A l l f r y i n g e x p e r i m e n t s w e r e r e p e a t e d once. F o r a n a l y s i s
fat p h a s e h a d a n i d e n t i c a l fat c o m p o s i t i o n , a n d c o n s i s t e d o f r a p e s e e d a n d p a l m oil (oil a n d p a r t i a l l y
of HAs,
h y d r o g e n a t e d oil) arid c o c o n u t oil. T h e m a r g a r i n e
w a t e r ( a b o u t 100 g) w a s a d d e d a n d s t i r r e d t o d i s s o l v e t h e p a n re s i due . F o r fa t a n a l y s i s , f r y i n g fa t w a s
c o n t a i n e d whey. The
liquid margarine was made
f r o m r a p e s e e d oil a n d s k i m m e d m i l k . T h e s u n f l o w e r seed oil w a s n o t d e w a x e d . T a b l e 1 s h o w s the f a t t y a c i d c o m p o s i t i o n o f the fats.
Cooking conditions T h e b e e f b u r g e r s w e r e a l l o w e d to r e a c h r o o m t e m p e r a t u r e ( a b o u t 18°C) a n d t h e n w e r e fried in a thermostat-controlled teflon-coated frying pan (235 x 235 m m ) c o n s t r u c t e d a t t h e S w e d i s h I n s t i t u t e fo r F o o d R e s e a r c h ( S I K , G o t h e n b u r g , S w e d e n ) , a t fat t e m p e r a t u r e s o f 1,55 a n d 200°C. T h e t e m p e r a t u r e
pan
residues were collected after frying,
c o l l e c t e d a n d filtered i n t o a b r o w n b o t t l e . T o o b t a i n e n o u g h m a t e r i a l for a n a l y s i s o f fat o x i d a t i o n p r o d ucts, fa t f r o m t w o o r t h r e e i d e n t i c a l f r y i n g e x p e r iments was pooled to form one sample. Samples were immediately frozen and kept at -20°C until a n a l y s e d . T h e fri e d b e e f b u r g e r s w e r e w e i g h e d a n d f r o z e n a f t e r b e i n g p h o t o g r a p h e d t o r e c o r d t he d e g r e e of browning. Before extraction, the crust (outer 2 m m f r o m e a c h side) o f t h e b e e f b u r g e r s w a s r e m o v e d u s i n g a s c a l pe l . A l l s a m p l e s , i n c l u d i n g p a n r e s i d u e s , w e r e f r e e z e - d r i e d b e f o r e e x t r a c t i o n o f H A s a n d k e p t in plastic beakers at - 2 0 ° C
until they were analysed.
Table 3. Oxidation status and antioxidant content of frying fats Tocopherol/tocotrienol (mg/100 g) Frying fat Margarine
Margarine fat phase Liquid margarine Butter Rapeseed oil Sunflower seed oil
Temperature PV (<'C) (mEq/kg) unheated 165 200 200* unheated 200 unheated 200 unheated 200 unheated 200 200* unheated 200
0.0 0.0 0.2 2.0 0.0 0.4 0.1 1.2 0.0 0.3 0.0 20.0 37.6 0.1 21.1
AV 1.3 2.2 5.7 17.1 1.0 5.3 1.5 8.8 0.1 3.1 1.6 88.8 225.1 2.4 98.0
FFA Vitamin A /J-Carotene Vitamin E (%) (mg/100 g) (mg/100 g) ~t-T ~t-T3 /J-T y-T 7-T3 6-T Total (mg/100 g) 0.05 0.29 0.22 0.11 0.1 0.23 1.1 0.27 0.33 0.34 0.02 0.26 0.14 0.03 0.23
1.13 0.80 0.62 0.83 1.14 0.58 1.12 0.63 na na na na na na na
0.2 0.15 0.11 0.12 0.19 0.09 0.15 0.09 0.27 0.09 na na na na na
16.0 12.2 11.5 14.8 16.4 10.7 20.5 14.0 2.5 1.7 20.5 0.5 nd 59.5 0.7
4.4 3.4 3.2 3.9 3.4 2.1 nd nd nd nd nd nd nd nd nd
nd nd nd nd nd nd nd nd nd nd nd nd nd 2.3 nd
19.8 14.8 13.5 18.6 18.3 12.0 32.1 22.3 nd nd 33.6 2.8 1.8 1.0 nd
3.8 2.8 2.5 3.4 3.7 2.4 nd nd nd nd nd nd nd nd nd
1.0 0.6 0.6 0.8 1.0 0.7 1.2 0.7 nd nd 1.1 0.4 nd nd nd
45.0 33.8 31.3 41.5 42.8 27.9 53.8 37.0 2.5 1.7 55.2 3.7 1.8 62.8 0.7
19.3 14.6 13.8 17.8 19.3 12.6 23.8 16.2 2.5 1.7 23.9 0.7 0.2 60.7 0.7
na = not analysed nd= not detected All values are means from duplicate frying experiments. The amounts of vitamins A, fl-carotene, tocopherol/tocotrienol and vitamin E are given as rag/100 g fat I:,hase. *Fat heated at 200'C for 10 rain without meat.
996
M.A.E. Johansson et al.
~
Caffeine (is) mAU
15
IQx
10 MelQx
,'0
,'5
2's Time (min)
I
250
I
I
I
I
350 Wavelength (nm) 300
Fig. 1. Expanded region of a chromatogram from the HPLC analysis of pan residue from frying of beefburgers in liquid margarine at 200°C. The peaks corresponding to MelQx, DiMelQx, PhlP and caffeine (internal standard, i.s.) are indicated. On-line recorded UV absorbance spectra compared with those of synthetic MelQx (lower), DiMelQx (lower) and PhlP (upper), respectively. Margarine and rapeseed oil heated without meat for 10 min at 200°C were analysed for HAs and used as negative control samples. Extraction of HAs Freeze-dried meat (about 1.5g, generally corresponding to about 6 g cooked beef) and pan residue (about 0.7 g, obtained from about 5 g cooked beef) were extracted using the method described by Gross et al. (1993). Using this method, a polar extract (containing the IQ-type of HAs and the glutamic acid pyrolysates) and a non-polar extract (containing the pyrido-indole type of HAs) are obtained. Only the polar extract was analysed in this study. To enhance the solubility of the samples and increase the extraction recovery of HAs, the samples were dissolved in sodium hydroxide and placed in a water-bath at 50"C. For the extraction of pan residues 5% phenol was added to DCM (G. A. Gross, personal communication). Additional purification of the pan residues was performed using Fractogel TSK CM650(s) to improve selectivity and avoid interfering peaks during HPLC analysis (Gross et al., 1992). Standard addition quantification was performed by spiking two of four samples with the standard reference spiking mixture mentioned above. Identification and quantification o f HAs The extracts were evaporated to dryness under a stream of nitrogen and thereafter dissolved in 100 Itl
caffeine solution, which served as an internal standard. Aliquots (90pl) of the samples were injected (Varian 9100 Autosampler) into a Varian 9010 liquid chromatograph equipped with a ToyoSoda TSK Gel ODS 80TM column (250 x 4.6 mm i.d., 5-/~m particle size, Varian, Stockholm, Sweden) and a precolumn (Supelguard LC-18-DB, 20 x 4.6 mm i.d., Labkemi, Stockholm, Sweden). The chromatographic conditions were adopted from Gross (1990) with minor modifications. The mobile phase was: solvent A and B, 10 mM triethylamine adjusted with acetic acid to pH 3.2 and 3.6, respectively, and solvent C, acetonitrile. The effluent was monitored using a photodiode array UV detector (Varian 9065, Polychrom) and a programmable fluorescence detector (Varian LC 9070). The identities of the compounds were established by comparing the retention times of the peaks with those of the corresponding spiked samples run under the same conditions. Diode array UV spectra of the peaks were compared with the spectra of the spiked samples and with library entries. The HAs were quantified by the standard addition quantification method using peak areas of the unspiked and spiked samples. The areas were compared with the areas of known amount of standards. The results were corrected for incomplete recovery. All results are expressed in ng compound per g cooked beef. The contents in the pan residues are correlated to the amount of cooked meat.
Effects of frying fat on the formation of HAs
PkLte 1. Photograph of beefburgers fried in margarine at 165°C (left) and 200°C (right).
997
Effects of frying fat on the formation of HAs
999
P ~<0.05) using the software package SPSS for MS Windows 5.0 (SPSS Inc., Chicago, USA). The correlation between HA content in the samples and oxidation status, antioxidant content and chemical composition of the frying fats was statistically evaluated using partial least squares analysis (PLS) and the software SIRIUS (Pattern Recognition System A/S, Bergen, Norway). The procedure of cross-validation was used to determine the significance of the PLS components (Wold, 1978).
Analysis o f oxidation status and antioxidant content & the fat The different fats were analysed for peroxide value (PV), anisidine value (AV), free fatty acids (FFA), vitamin E (tocopherols and tocotrienols) and vitamin A (retinol) content using standard methods (the analyses were performed at Van den Bergh Foods using methods standardized by Unilever), before being heated and after heating for 10 min at 200°C with and without beefburgers. The fl-carotene content was determined according to Sadler et al. (1990). About 10 g fat was extracted using heptane-acetone-ethanol (50:25:25, by vol) and stirred for 15 min. Water (15 ml) was added and the sample was stir'red for 10min. The non-polar phase was evaporated to dryness, redissolved and analysed using HPLC. The samples were isocratically chromatographed ,an a LiChrospher 100RP-18 column (250 x 4ram i.d., 5-/~m particle size) equipped with a precolumn, using methanoltetrahydrofuran-wa~Ier (67:27:6, by vol) as a mobile phase. The flow rate was 2 mltmin and the effluent was monitored at 4'75 nm using a UV detector.
RESULTS
HAs in beefburgers The HAs MelQx, DiMelQx, PhlP, norharman and harman were recovered in the meat samples. The amounts of MelQx, DiMelQx and PhlP detected in the beefburgers fried in different frying fats are shown in Table 2. The amounts of these HAs were very low or not detected in the beefburgers fried at 165°C. At 200°C the amounts increased, but were still low. Formation of PhlP showed a pronounced increase with temperature. A significantly higher amount (1.6 ng/g) of MelQx was found in beefburgers fried in rapeseed oil than in beefburgers fried in any of the other fats. The highest amount of DiMelQx (0.4ng/g) was found in beefburgers fried in margarine, rapeseed oil and liquid margarine, while beefburgers fried in margarine fat phase contained the highest amount of PhlP. Norharman and harman
Statistics The amounts of HAs detected in meat and pan residues after frying: beefburgers in different frying fats were statistic~Llly compared using one-way ANOVA followed by Duncan's test (significance level
180
,.! /"
I,o-|
, s . . ~. o ' ° t 160
..'.-2:: ...............
"" ";:;';"
v
o.-" •
140
.
,,..;-,:-. • .',
i
.
120
/i,. ,:',,'"/ .........
'l
:.,,.... ............,.....
t-,
i ,*"
",, ,..~
/
!
s
.,. ...
,"
,/
,,
,
• " ..........
•..
100
I
I
I
3
I
I
4
I
I
5
I
I
6
I
I
I
7 Time (rain)
I
8
I
I
9
I
I
10
I
I
II
I
I
12
Fig. 2. Time-temperature profiles during frying of eight batches of beefburgers in rapeseed oil at a fat temperature of 200°C. The temperature was measured every 30 sec using a chrom/alumel thermocouple placed in the fi'ying fat between the pan and the beefburger. The variation in temperature is due to the variations in transport of water from the meat into the pan. The beefburgers were turned at 7 min.
1000
M . A . E . Johansson et al.
were detected at low ppb levels (0.5-3.2 ng/g and 0.1-0.7ng/g, respectively) in all samples fried at 200'~C (data not shown). HAs in pan residues
As shown in Table 2, the content of HAs was generally higher in the pan residues than in the beefburgers. The amounts of MeIQx ranged from 2.1 ng/g, in pan residue from frying in sunflower seed oil, to 4.3 ng/g in liquid margarine. The highest amount of DiMeIQx was found in pan residue from frying with margarine. The amount detected in pan residue from frying in sunflower seed oil was significantly lower than after frying with any other oil. Of all the HAs detected, PhIP was present in the highest amounts in pan residues, with levels ranging from 2.0 ng/g after frying in sunflower oil to 13.3 ng/g after cooking using margarine fat phase. A low level of PhIP in pan residues was detected after frying in margarine, while frying in margarine fat phase, liquid margarine or rapeseed oil produced significantly higher amounts. Norharman and harman were detected at low ppb levels in the pan residues (data not shown). The lowest total amount of HAs in the pan residues were detected after frying in sunflower seed oil (4.7 ng/g) or margarine (8.6 ng/g). Frying in margarine fat phase, liquid margarine, butter and rapeseed oil produced significantly higher amounts. The total content of HAs in meat plus pan residue was significantly lower after frying in sunflower seed oil or margarine than after frying with the other fats. IQ, MeIQ, GIu-P-I and GIu-P-2 were not detected in any sample. No HAs were detected in margarine or rapeseed oil heated for 10 rain at 200°C without meat. Oxidation status and antioxidant content o f the frying fats
Table 3 shows the results from the analysis of the frying fats for PV, AV, FFA, vitamin A,/~-carotene and vitamin E: ~-T (~-tocopherol), c~-T3 (~-tocotrienol),/~-T, 7-T, i'-T3 and 6-T. Frying in margarine, margarine fat phase, liquid margarine and butter increased PV and AV moderately. Considerably higher PV and AV were found in rapeseed and sunflower seed oil after frying at 200°C, while PV and AV were very low after frying in margarines and butter. F F A was high in unheated butter, but did not increase after frying with meat. Generally, AV and F F A increased less than expected after frying with meat. The degree of oxidation of fats heated without meat increased, as would be expected. The vitamin A and /~-carotene contents were reduced to about half the original amounts in the heated fats. Only trace amounts of ct- and 7-tocopherol were retained in the oils after frying, whereas two-thirds of the original values remained in margarines and butter.
Correlation between HA formation, oxidation status and antioxidant content
The observed differences in MelQx and DiMeIQx formation can be explained in terms of oxidation status (PV and AV) and antioxidant content (vitamin A, vitamin E and tocopherols/tocotrienols) in the frying fat, using partial least squares analysis. The first and second PLS components of the above mentioned variables explained most of the differences in MelQx and DiMelQx formation. One PLS component explained 53% of the variation in MelQx in the pan residues, while two PLS components explained 83% of the MelQx variation in the beefburgers and 74% of the DiMelQx variation in the pan residues. However, these variables did not explain the formation of PhlP. The correlation coefficients for measured and predicted levels of MelQx were r---0.95 and r = 0.79 in the beefburgers and pan residues, respectively. The corresponding coefficient for DiMelQx in the pan residue was r = 0.89. The DiMelQx content of the beefburgers was too low for statistical evaluation. The results obtained could not be explained by differences in fat composition or any of the other parameters measured. Identification and quantification
Identification and quantification of HAs in the samples were performed by HPLC with on-line UV spectra for peak identification and purity. Peak identification was achieved by comparison of retention time and UV spectra with spiked samples and library entries. Figure 1 shows the UV chromatogram at 263 nm after HPLC analysis of a pan residue sample obtained after frying beefburgers at 200C in liquid margarine. The peaks corresponding to MelQx, DiMelQx, PhlP and caffeine (internal standard) are indicated. Figure 1 also shows the on-line recorded UV spectra of MelQx, DiMelQx and PhlP in the sample together with reference spectra. There was good correspondence between the absorbance peak shape obtained from the sample and that of the reference spectra generated using synthetic references. The average recoveries from the purification of the spiked beef samples was 71 + 14%, 63 + 14% and 66 + 16% (mean +_ SD, n = 28) for MelQx, DiMeIQx and PhIP, respectively. For the pan residues, the corresponding values were 76 + 10%, 63 +_ 12% and 51 +_ 16% (n = 28), respectively. The variations between duplicate determinations from duplicate frying experiments is shown in Table 2 (variation is not given for values less than 1 ng/g). The detection limits varied with the compound and the complexity of the sample matrix. In the samples from fried beefburgers MeIQx, DiMeIQx and PhIP could be detected and confirmed by UV absorption spectra at 0.03, 0.02 and 0.01 ng/g, respectively. Each of these three HAs could be detected and confirmed in the pan residues at 0.02 ng/g. Norharman and harman are known to appear both in the polar and
Effects of frying fat on the formation of HAs nonpolar extracts with varying degrees of recovery, and therefore no attempt was made to quantify these compounds. Chemical composition, weight loss, amount o f crust, colour, time-temperature profiles The chemical composition of the uncooked meat was: fat, 11.2%; protein, 19.1%; water, 67.9%; ash, 2.0%; carbohydrates, 0%. The weight losses during cooking were 28.4:_~2.3% ( n = 16) at 165°C and 34.5 _+ 1.4% (n = 96) at 200~'C. The amount of crust was 30.5 + 3.4% (n = 8) at 165°C and 31.6_+ 3.4% (n = 48) at 200~C. 1-he surface of the beefburgers fried at 165C was somewhat pale. The crust formed at 200~C was quite brown and would be judged as appetizing by some and too dark by others. The beefburgers fried in oil were slightly darker than the beefburgers fried in butter or margarine. Plate 1 shows a photograph of beefburgers fried in margarine at 165 and 200' C. Time-temperature profiles for each frying experiment were obtained by measuring the temperature, using a thermocouple placed between the pan and a beefburger, every 30 sec. Temperature differences were observed between the various frying experiments due to uneven transport of water from the meat into the pan. The presence of water lowered the temperature at the position of the thermocouple. Figure 2 shows time-temperature profiles obtained from frying eight ba':ches of beefburgers in rapeseed oil at 200~C. DISCUSSION
Our study is the first in which quantitative data on the influence of different frying fats on the formation of HAs in beefburgers and pan residues are reported. The type of frying fat had only minor effects on the formation of HAs irL fried beefburgers. However, it did influence the amount of HAs in the pan residue. The lowest total amount of HAs was obtained after frying in sunflower seed oil or margarine. The results are in agreement with a previous study performed at our laboratory, in which pan residues from frying beefburgers in butter were shown to contain significantly higher amounts of MelQx and DiMelQx than samples fried in vegetable oil (Johansson and J/igerstad, 1994). However, different HAs might be formed under different cooking conditions, and since there are variations in the carcinogenic potency of the different HAs, more research is needed before any particular frying fat can be recommended. The amounts of HAs are consistent with those reported previously for beefburgers cooked under similar conditions (Gross, 1990; Gross et al., 1993; Johansson and J/igerstad, 1994; Knize et al., 1994; Murray et al., 1993; Skog et al., 1995). The results of the present study, together with those of model experiments (Johansson and J/igerstad, 1993), leave no doubt that the formation of HAs is, to some extent, influenced by the frying fat.
1001
By using the PLS method, the observed variations in MelQx and DiMelQx formation can be explained in terms of antioxidant content (vitamin A, vitamin E and tocopherols/tocotrienols) and oxidation status of the frying fats (PV and AV). The content of antioxidative sterols in un-dewaxed sunflower seed oil might have had some inhibiting effects on the formation of HAs, but this was not investigated. Differences in fat composition did not explain the variation in HA formation. Interestingly, the variation in PhIP formation was not explained by the PLS components of any of the measured variables. According to the 'Las Vegas-hypothesis' IQ-compounds, such as MelQx and DiMelQx, are formed through the condensation of naturally occurring creatine and pyrazines and Strecker aldehydes, which are formed from free amino acids and monosaccharides in the Maillard reaction (J~igerstad et al., 1983). The present results show that antioxidants, such as tocopherols, tocotrienols and retinol, might reduce the formation of IQ-like HAs, probably by scavenging free radicals in the early stage of the Maillard reaction, which is known to be enhanced by free radical reactions (Namiki and Hayashi, 1981). The results of the present study are further supported by our observation of a decrease in the formation of IQx, MeIQx and DiMeIQx when antioxidants, such as y-tocopherol and /~-carotene, were incorporated into a model system containing pro-oxidants (Johansson and J~igerstad, 1995). PhIP was the major HA formed in the pan residues. The mechanism for the formation of PhIP is still not known in detail, but in contrast to the IQ-compounds, pyrazines/pyridines are not required. Obviously, the formation of PhIP is not associated with free radical reactions and is not inhibited by free radical scavenging antioxidants. The observed variations in PhlP formation after frying in different fats could not be explained in terms of any of the measured parameters and further experiments are needed to explain the variation. There was some variation between the meat and pan residues in the content of the different HAs. For example, while the highest levels of MeIQx were found in meat fried in rapeseed oil, the levels of MeIQx and other HAs in the corresponding pan residues were by no means the highest. Also, whereas for all fats the levels of MeIQx in the pan residues were two-four-fold those in the meat, the levels of PhIP in the residues were generally about 10-fold more than in meat. This might suggest that the various HAs are formed in different phases of the meat and that the production of these phases might be influenced by the particular fat used for the frying. The degree of fat oxidation during heating in thin layers, as occurs in shallow pan frying, is dependent on the content of polyunsaturated fatty acids and milk. In addition, the frying technique (e.g. temperature, heating time and thickness of the fat layer) have great influence (Fredholm et al., 1990). The frying
1002
M.A.E. Johansson et al.
fats used in the present study were chosen on the basis of their differences in fatty acid composition and other constituents, especially antioxidants and milk. Among the six investigated frying fats, margarine, butter and rapeseed oil are commonly used in cooking. Sunflower seed oil was selected because of its high content of polyunsaturated fatty acids and instability towards oxidation during heating. The liquid margarine differed from rapeseed oil by containing milk, which has been shown to exert antioxidative activity (Fredholm et al., 1990). The margarine fat phase was taken out before adding the water phase, which contained whey, during the manufacturing of margarine. Our study is one of the first reporting on fat oxidation and loss of antioxidants in the frying fat during shallow pan frying of meat. As expected, butter and the margarines showed rather modest oxidation, while rapeseed and sunflower seed oil were greatly oxidized. An AV close to 100 was obtained for both of the oils and an off-taste is generally noted at AV over 25. The loss of antioxidants was reflected by the degree of fat oxidation, and for the moderately oxidized fats (e.g. margarines and butter) about half of the content of antioxidants was retained after frying. The two oils lost their vitamin E activity completely. The presence of whey in the solid margarine had no effect on the fat oxidation or retention of antioxidants. However, the presence of milk in the liquid margarine had an obvious effect on the fat oxidation and the retention of antioxidants: the oxidation was about 10-fold lower in the liquid margarine than in the rapeseed oil. About half of the antioxidant content was retained in the liquid margarine after heating; the rapeseed oil lost all of its antioxidants. When solid margarine and rapeseed oil were heated without the presence of meat, both were more oxidized than when heated with meat. In rapeseed oil this difference was very big. As meat is more likely to contain pro- than antioxidants the differences in oxidation must be due to other factors such as fat temperature, depth of the frying fat and water evaporation-three circumstances that change considerably when beefburgers are added to the fat. These changes correspond to a decrease of 25-30°C in the frying temperature. It is also interesting to note that when the margarine was heated without meat the retention of antioxidants was much higher than after heating with meat. This was not the case when heating rapeseed oil. The retention was just as poor when heated without as with meat. Other studies concerning the role of frying fat on the formation of HAs, show inconsistent results (Barrington et al., 1990; Nilsson et al., 1986) and are chiefly based on Ames test data. However, unsaturated long-chain fatty acids, such as oleic and linoleic acid, are known to inhibit the Ames test, making the results uncertain (Hayatsu et al., 1981). Besides the chemical effects of fat shown in the present study,
the physical effects of fat on the formation of HAs is of great importance. In most reported studies, it is difficult to distinguish between the physical and chemical effects of the fat. Lower mutagenic activity was detected in pork fried without frying fat, which was explained by a more efficient heat transfer when frying fat is used (Nilsson et al., 1986). Similarly, when oven-baking meat loaves a shorter cooking time was needed to achieve a centre temperature of 7 0 C with products with a high fat content than with those containing less fat (Holtz et al., 1985). The results indicated that the lowest mutagenic activity was produced in high-fat meat which was cooked for the shortest time. Another study explained the observed reduction in mutagenic activity in high-fat products as being the result of the dilution of precursors (Knize et al., 1985). HAs are mainly formed in the crust of cooked food (Bjeldanes et al., 1983; Commoner et al., 1978). In order to increase sensitivity and lower the detection limit, we used only the crusts of the beefburgers. Consequently, the amounts of HAs might be slightly underestimated since the presence of small amounts of HAs in the inner parts of fried meat have been reported (Skog et al., 1995; Tikkanen, 1989). In agreement with earlier studies performed at our laboratory (Johansson and Jfigerstad, 1994; Skog et al., 1995), the content of HAs was higher in pan residues than in the corresponding beefburgers. In Sweden, pan residues are commonly served with fried meat dishes (Asp et al., 1992) and the contribution from pan residues should be included when the dietary intake of HAs is assessed. On the basis of our findings, the human exposure to MeIQx, DiMeIQx and PhlP from a diet consisting of 150-200 g fried meat, including gravy prepared from the pan residue, may range from less than 0.5 to 4.0/~g/person/day, depending on cooking conditions. Levels of total HA were 10-fold higher after frying at 200~C than at 165°C. Experts recommend a frying temperature of 155-175°C when frying beefburgers (K6ttinformation, 1986). In the present study, a higher frying temperature was chosen to obtain clear results regarding the role of frying fats on the formation of HAs. However, as can be seen from the photographs, the beefburgers were acceptable, although somewhat too brown and overcooked. Since domestic cooking equipment lacks thermostatically controlled frying surfaces, the temperature varies within a wide range while frying. It is almost impossible to maintain a steady desired temperature and overheating occurs often. IARC (1986) recommends a reduction in the daily intake of HAs. In order to do so, there is a need to develop temperature-controlled cooking equipment and to advise consumers on how to use it. Another important factor is to change people's attitude towards less browned fried food. It is also important to point out that a higher frying temperature does not lead to a shorter frying time (Dagerskog, 1978). The daily intake of HAs can also be
Effects of frying fat on the formation of HAs reduced simply by discarding the p a n residue after frying. In addition, efforts to reduce H A f o r m a t i o n require a better u n d e r s t a n d i n g of the reactions behind the formation. Acknowledgements~Ne gratefully acknowledge the excellent technical assistance of Ms Ingrid Berntsson and Gunnel Andersson, Department of Applied Nutrition and Food Chemistry, and Ms Eva Ahlbin, Van den Bergh Foods. This study was supported financially by the Swedish Cancer Foundation (1824-B93-12XBC) and the Swedish Council for Forestry and Agricultural Research (50.00440/91). REFERENCES
Adamson, R. H., Tl~orgeirsson U. P., Snyderwine E. G., Thorgeirsson S. S., Reeves J., Dalgard D. W., Takayama S. and Sugimura T (1990) Carcinogenicity of 2-amino-3methylimidazo[4,5-f]quinoline in nonhuman primates: induction of tumors in three macaques. Japanese Journal of Cancer Research 811, 10-14. Asp N. G., Isaksson ,~., Johansson U., JS_gerstad M., Lanke J., Mattisson I., Swederberg E. and Scherstrn B. (1992) Dietary habits among adults in Olofstrrm--a baseline study. Scandinavian Journal of Nutrition/Niiringsforskning 36, 106-114. Barnes W. S., Maher J. C. and Weisburger J. H. (1983) High-pressure liquid chromatographic method for the analysis of 2-amino-3-methylimidazo[4,5-f]quinoline, a mutagen formed during the cooking of food. Journal of Agricultural and Food Chemistry 31, 883-886. Barrington P. J., Baker R. S. U., Truswell A. S., Bonin A. M., Ryan A. J. and Paulin A. P. (1990) Mutagenicity of basic fractions derived from lamb and beef cooked by common household methods. Food and Chemical Toxicology 28, 141-146. Bjeldanes L. F., Morris M. M., Timourian H. and Hatch F. T. (1983) Effecl:s of meat composition and cooking conditions on mutagen formation in fried ground beef. Journal of Agricultural and Food Chemistry 31, 18 21. Brittebo E. B., Karlsson A. /~., Skog K. I. and J~igerstad I. M. (1994) Transfer of the food mutagen PhlP to foetuses and newborn mice following maternal exposure. Food and Chemicat Toxicology 32, 717-726. Commoner B., Vihayathil A. J., Dolora P., Nair S., Madyastha P. anti Cuca G. C. (1978) Formation of mutagens in beef arLd beef extract during cooking. Science 2111, 913 916. Dagerskog M. (1978) Stekning av livsmedel, studier av det processtekniska underlaget. Institutionsavdelningen frr livsmedelsvetenskap och SIK--Svenska Livsmedelsinstitutet. Chalmers Tekniska Hrgskola. Felton J. S. and Knize M. G. (1990) Heterocyclic-amine mutagens/carcinogens in foods. In Handbook of Experimental Pharmacology. Edited by C. S. Cooper and P. L. Grover. pp. 471-502. Springer-Verlag, Berlin. Fredholm L., Gustafi;son I.-B. and Jonsson L. (1990) Vad h~inder med matfetterna vid stekning och fritering. Ndringsforskning 35, 132-140. Ghoshal A. and Snyderwine E. (1993) Excretion of foodderived heterocyclic amine carcinogens into breast milk of lactating rats and formation of DNA adducts in newborn. Carcinogenesis 14, 2199-2203. Gross G. A. (1990) Simple methods for quantifying mutagenic heterocyclic aromatic amines in food products. Carcinogenesis 11, 1597-1603. Gross G. A., Gruter A. and Heyland S. (1992) Optimization of the sensitivity of high-performance liquid chromatography in the detection of heterocyclic aromatic amine mutagens. Food and Chemical Toxicology 30, 491~,98.
1003
Gross G. A., Turesky R. J., Fay L. B., Stillwell W. G., Skipper P. L. and Tannenbaum S. R. (1993) Heterocyclic aromatic amine formation in grilled bacon, beef and fish and in grill scrapings. Carcinogenesis 14, 2313-2318. Hayatsu H., Arimoto S., Togawa K. and Makita M. (1981) Inhibitory effect of the ether extract of human feces on activities of mutagens: inhibition by oleic and linoleic acids. Mutation Research 81, 287 293. Holtz E., Skjrldebrand C., J~igerstad M., Laser Reutersw~ird E. and Isberg P. E. (1985) Effects of recipes on crust formation and mutagenicity in mean products during baking. Journal of Food Technology 20, 57~56. IARC (1986) IARC Monographs of the Evaluation of the Carcinogenic Risk of Chemicals to Humans. Vol. 40. Some Naturally by Occurring and Synthetic Food Components, Furocomarins and Ultra Violet Radiation. International Agency for Research on Cancer, Lyon. Johansson M. and J~igerstad M. (1993) Influence of oxidized deep-frying fat and iron on the formation of food mutagens in a model system. Food and Chemical Toxicology 31, 971-979. Johansson M. A. E. and J~igerstad M. (1994) Occurrence of mutagenic/carcinogenic heterocyclic amines in meat and fish products, including pan residues, prepared under domestic conditions. Carcinogenesis 15, 1511-1518. Johansson M., Skog K. and J~igerstad M. (1993) Effects of edible oils and fatty acids on the formation of mutagenic heterocyclic amines in a model system. Carcinogenesis 14, 89 94. J~igerstad M., Laser Reutersw~ird A., Oste R., Dahlqvist A., Grivas S., Olsson K. and Nyhammar T. (1983) Creatinine and Maillard reaction products as precursors of mutagenic compounds formed in fried beef. In The Maillard Reaction in Food~ and Nutrition. Edited by G. R. Waller and M. S. Feather. pp. 507-520. ACS Symposium Series 215. American Chemical Society, Washington, DC. J~gerstad I. M., Johansson M. A., Karlsson A. ,~., Skog K. I., Oskarsson A., Palminger-Hallen I., Ivhed R. and Brittebo E. B. (1994) Dose-dependent milk transfer and tissue distribution of the food mutagen PhlP in rats and their suckling pups. Carcinogenesis 15, 2479-2484. Knize M. G., Andresen B. D., Healy S. K., Shen N. H., Lewis P. R., Bjeldanes L. F., Hatch F. T. and Felton J. S. (1985) Effects of temperature, patty thickness and fat content on the production of mutagens in fried ground beef. Food and Chemical Toxicology 23, 1035-1040. Knize M. G., Dolbeare F. A., Carroll K. L., Moore D. H. and Felton J. S. (1994) Effect of cooking time and temperature on the heterocyclic amine content of fried-beef patties. Food and Chemical Toxicology 32, 595~03. Krttinformation (1986) KrttR~itt. ISBN 91-7766-021-8. pp. 136-137. (In Swedish) Murray S., Lynch A. M., Knize M. G. and Gooderham N. J. (1993) Quantification of the carcinogens 2-amino-3, 8-dimethyl- and 2-amino-3,4,8-trimethylimidazo[4,5-f]quinoxaline and 2-amino-l-metyl-6-phenylimidazo[4,5-b]pyridine in food using a combined assay based on gas chromatography negative ion mass spectrometry. Journal of Chromatography 616, 211-219. Namiki M. and Hayashi T. (1981) Formation of novel free radical products in an early stage of Maillard reaction. In Progress in Food and Nutrition Service. Iiol. 5. Maillard Reactions in Food. Edited by C. Eriksson. pp. 81-92. Pergamon Press, Oxford. Nilsson L., Overvik E., Fredholm L., Levin 0., Nord C.-E. and Gustafsson J.-,~. (1986) Influence of frying fat on mu~agenic activity in lean pork meat. Mutation Research 171, 115-121. Norell S. E., Ahlbom A., Erwald R., Jacobson G., Lindberg-Navier I., Olin R., Trrnberg B. and Wiechel K. L. (1986) Diet and pancreatic cancer: a case-control study. American Journal of Epidemiology 124, 894-902.
1004
M. A. E. Johansson et al.
Ohgaki H., Takayama S. and Sugimura T. (1991) Carcinogenicities of heterocyclic amines in cooked food. Mutation Research 259, 399-412. Overvik E., Nilsson L., Fredholm L., Levin 0., Nord C.-E. and Gustafsson J.-/~. (1987) M utagenicity of pan residues and gravy from fried meat. Mutation Research 187, 47-53. Sadler G., Davis J. and Dezman D. (1990) Rapid extraction of lycopene and fl-carotene from reconstituted tomato paste and pink grapefruit homogenates. Journal of Food Science 55, 1460-1461. Schiffman M. H. and Felton J. S. (1990) Fried foods and risk of colon cancer. Letter to the Editor. American Journal of Epidemiology 131, 376-378. Skog K. (1993) Cooking procedures and food mutagens: a literature review. Food and Chemical Toxicology 31, 655~75. Skog K., Steineck G., Augustsson K. and Jiigerstad M. (1995) Effect of cooking temperature on the formation of heterocyclic amines in fried meat products and pan residues. Carcinogenesis 16, 861 867. Steineck G., Gerhardsson de Verdier M. and Overvik E. (1993) The epidemiological evidence concerning intake of mutagenic activity from fried surface and the risk of cancer cannot justify preventive measures. European Journal of Cancer Prevention 2, 293 300. Steineck G., Hagman U., Gerhardsson M. and Norell S. E.
(1990) Vitamin A supplements, fried foods, fat and urothelial cancer. A case-referent study in Stockholm in 1985-87. International Journal of Cancer 45, 100(~1011. Sugimura T., Nagao M., Kawachi T., Honda M., Yahagi T., Seino Y., Sato S., Matsukara N., Shirai A., Sawamura M. and Matsumoto H. (1977) Mutagens~arcinogens in food, with special reference to highly mutagenic pyrolytic products in broiled foods. In Origins of Human Cancer. Edited by H. H. Hiatt, J. D. Watson and J. A. Winsten. pp. 1561-1577. Cold Spring Harbor Laboratory, New York. Tikkanen L. (1989) Mutagenicity of cooked foods. VTT, Esboo. Tikkanen L. M., Sauri T. M. and Latva-Kala K. J. (1993) Screening of heat-processed Finnish foods for the mutagens 2- amino- 3, 8 - dimethylimidazo[4, 5 - f ]quinoxaline, 2-amino-3,4,8-trimethylimidazo[4,5-J]quinoxaline and 2amino- l-methyl-6-phenylimidazo[4,5-b]pyridine. Food and Chemical Toxicology 31, 717-721. Willett W. C., Stampfer M. J., Colditz G. A., Rosner B. A. and Speizer F. E. (1990) Relation of meat, fat and fiber intake to the risk of colon cancer in a prospective study among women. New England Journal of Medicine, 1664-1672. Wold S. (1978) Cross-validatory estimation of the number of components in factor and principal component models. Technometrics 20, 397-405.