Document RjLBz6m6zLeGyGXjrm4QnOj7n

REPORT -- Mutagens From Heated Chinese and U.S. Cooking Oils P. G.Shields, G.X . Xu, W.J. Blot,J . F. Fraumeni,Jr., G. E. Trivers,E. D. Pellizzari, Y. H . Qu, Y . T. Gao, C. C. Harris" Background: The lung cancer incidence in Chinese women is among the highest in the world, but tobacco smoking accounts for only a minority of the cancers. Epidemiologic investigations of lung cancer among Chinese women have implicated exposure to indoor air pollution from wok cooking, where the volatile emissions from unrefined cooking oils are mutagenic. Purpose: This study was conducted to identify and quantify the potentially mutagenic substances emitted from a variety of cooking oils heated to the temperatures typically used in wok cooking. Methods: Several cooking nils and fatty acids were heated in a wok to boiling, at temperatures (for'the cooking oils) that ranged from 240 "C to 280,"C (typical cooking temperatures in Shanghai, China). The oils tested were unrefined Chinese rapeseed, refined U.S. rapeseed (known as canola), Chinese soybean, and Chinese peanut in addition to linolenic, linoleic, and erucic fatty acids. Condensates of the emissions were collected and tested in the Salmonella mutation assay (using Salmonella typhimurium tester strains TA98 and T A W ) . Volatile decomposition products also were subjected to gas chromatography and mass spectroscopy. Aldehydes were detected using high-performance liquid chromatography and UV spectroscopy. Results: 1,3-Butadiene, benzene, acrolein, formaldehyde, and other related REPORT compounds were qualitatively and per day. As a possible etiologic factor, quantitatively detected, with emissions epidemiologic studies (9.10) of lung can- tending to be highest for unrefined cer in China have implicated indoor air Chinese rapeseed oil and lowest for pollutants from cooking-oil vapors. In peanut oil. The emission of I,3-buta- Shanghai, specifically, the use of un- diene and benzene was approximately refined rapeseed oil for wok cooking, 22-fold and 12-fold higher, respective- compared with soybean oil, was as- ly, from heated unrefined Chinese sociated with an odds ratio of 1.4 (95% rapeseed oil than from heated peanut confidence interval [CI] = 1.1-1.8) (9). oil. Lowering the cooking temperatures The ekcess risk also>vas associated with or adding an antioxidant, such as butyl- exposure to wok emissions, measured by ated hydroxyanisole, before cooking the degree of house smokiness and com- decreased the amount of these volatile plaints of eye imtation. Rapeseed oil is emissions. Among the individual fatty known in the United States as canola oil, acids tested, heated linolenic acid a refined oil with low erucic fatty acid produced the greatest quantities of 12- coctent. In China, rapeseed oils having butadiene, benzene, and acrolein. either low- or high-erucic acid content are Separately, the mutagenicity of in- used for cooking, but refined oils were dividual volatile emission condensates seldom available or used. was correlated with linolenic acid con- Condensates of volatile emissions from tent (r = .83; P = .0004). Condensates oils cooked in a wok were mutagenic in from heated linolenic acid, but not several in vitro short-term test systems linoleic or erucic acid, were highly (11). Other oils, such as peanut oil. were mutagenic. Conclusions: These studies, not. It has been found that the muta- combined with experimental and genicity of heated Chinese rapeseed oil epidemiologic findings, suggest that condensate is inhibited by the addition of high-temperature wok cooking with butylated hydroxyanisole to the oil before unrefined Chinese rapeseed oil may in- cooking or when the oil is hydrogenated. crease lung cancer risk. This study in- These findings suggest that the muta- dicates methods that may reduce that genicity is related to oxidation of fatty risk. Implications: The common use of acids. linolenic acid in particular. wok cooking in China might be an im- In Shanghai. women generally cook portant but controllable risk factor in with Chinese rapeseed oil and, to a lesser the etiology of lung cancer. In the extent, soybean oil. Customarily, about United States, where cooking oils are usually refined for purity, additional studies should be conducted to further quantify the potential risks of such methods of cooking. [J Natl Cancer Inst 87:836-841,19951 Lung cancer incidence in Chinese women is among the highest in the world (1-5). While tobacco smoking is a major risk factor for cancer, especially in men, it is not for Chinese women (4,6-8). Moreover, among Chinese women who do smoke, the tendency is to smoke much less than observed for U.S. women (6,8IO), averaging less than eight cigarettes +Afiliations of authors: P. G. Shields, G. E. Trivers, C. C. Harris (Laboratory of Human Carcinogenesis), W. J. Blot. J. F. Fnumeni, Jr. (Epidemiologyand Biostatistics Program). Division of Cancer Etiology. National Cancer Institute. Bethesda. Md. G. X. Xu. Y. H. Qu.Y. T. Gao. Shanghai Cancer Institute. Shanghai,People's Republic of China. E. D. Pellizzari, Research Triangle Institute. Re- senrch Triangle Park, N.C. Corrrspcmdrnccio: C. C. Harris. M.D.. Naiiona! Institutes of Health. Bldg. 37,Rrn. 2COI. Bethcsda. MD 20892. Scr "Notes"section following "Referenccs." * .. Journol of the National Cancer Institute. Vol. 87. No. 1 I , June 7, 1995 . 25-100 mL of cooking oil, depending on the type of food being prepared, is placed in .the wok and then heated to approximately 280 'C. This high heating temperature results in the generation of a large amount of smoke from the oil that can imtate the eyes and mucous membranes. Thus, kitchen windows are commonly left open, even in the winter. Cooking foods at lower temperatures causes the food to acquire a "bad" taste, whereas cooking at high temperatures tends to reduce this problem. In this report, we have analyzed volatile emissions from several cooking oils heated to the temperatures typically used in Chinese wok cooking. Included in these analyses is a preliminary characterization of the mutagenicity of these compounds. Methods Chemicals and cooking oils. Chinese cooking oils (unrefined Chinese rapeseed. peanut. sesame, and soybean) were purchased from retail Shanghai markets. Different lots were tested for mutagenic condensates. Unrefined Chinese rapeseed oil will be referred IO a, Chinese rapeseed oil in this report. Canola oil (Proctor and Gamble, Cincinnati, Ohio), a refined North American rapeseed oil. was purchased from a supermarket in Bethesda. Md. Linolenic acid, linoleic acid, and erucic acid were purchased from Sigma Chemical Co. (SI. Louis. Mo.). All chemicals were reagent grade. Preparation of condensates. A wok lid was modified by introducing three holes-two for the attachment of filter holders and one for a ther- mometer. Two glass filters (M9; Shanghai Huon Guan Paper Corp., Shanghai, China) were used in each filter holder. Cooking oils were placed in the wok (20 mL for volatile organic compound analysis and 100 mL for mutagenicity studies) and heated rapidly with a spherical heating mantle (loo0 mL: 380 W) until the designated temperature was reached (Chinese rapeseed: 275-280 'C [with and without butylated hydroxyanisole: Sigma Chemical Co.], canola oil: 275-280 'C. peanut oil: 260-265 'C. soybean oil: 260-265 'C. linolenic acid: 240 'C, linoleic acid: 240 'C, and erucic acid: 240 T),coin- ciding with either the release of smoke for oils. the typical Shanghai cooking temperatures. or boiling points for the fatty acids. The wok was then covered, and suction was applied throiigh the filters via a pump (20-30 L/minute). After cooking for 20 minutes. condensates from the four filters were extracted with acetone (IO mL) at room temperature. The condensate extract was then filtered with # I fil- ter paper (Whatman Ltd., Maidstone, U.K.)and evaporated at 60 'C under a stream of nitrogen. The wok was washed with liquid detergent after each use. Condensates were stored at -70 'C. Solmanella mutation assay. The assay was performed according to the method of Maron and Ames (12). Briefly. test condensates were diluted with dimethyl sulfoxide in the presence (or absence) of S9 liver homogenates (prepared from male SprapeDawley rats that had been given injections of Anxlor 1254) and preincubated (20 minutes at 37 'C) in glass culture tubes. Salmonella Fphimurium tester strains TA98 and TA104 were utilized. The bacteria were incubated under selective conditions on agar plates for 48 hours at 37 'C. Determination of volatile organic compounds. The identification of the major components in oil vapor samples was accomplished by manually elucidating their mass spectroscopic fragmentation patterns and comparing them with published spectra (13). Oil vapors were cryogenically trapped, subjected to gas chromatography (model 3700; Varian Associates. Inc., Lexington. Mass.) using thermal desorption onto a DB-624 fused silica capillary column (30 mm x 0.32 mm internal diameter. 2.7 mUminute flow rate; 240 'C) and analyzed with a magnetic sector mass spectrophotometer (model 2091: Pharmacia LKB Nuclear, Gaithersburg. Md.). Internal standatti-normalizedresponses from duplicate analyses of each standard were fit to a xcond-order calibration curve. Concentrations for I ,3-butadiene and benzene wtre calculated from the gas chromatography and mass spectroscopy responses for the analyte (normalized to the response of the internal standard) and the coefficients of the calibration curve. Larger volumes (200 mL) of the wok background and canister background samples were analyzed along with a National Institutes of Standards ad Technology standard reference mixture (SRM #1804) containing 1.3-butadiene and benzene. The recoveries were 96% and 92%. respectively. Determination of aldehyde and ketone content. Aldehydes and ketones were collected from heated oil emissions above the wok. using Sep-Pak DNPH-Silica canridges (Waters Chromatography. Millipore. Bedford, Mass.). The canridges were extracted with acetonitrile (6 mL) in a volumetric flask. Extracts were analyzed by isocratic high-performance liquid chromatography (50% acetonitrile. IO% methanol. and 40% water mobile phase on a Sphcrisorb ODS column 25 cm x 4.6 mm internal diameter) and analyzed by U V (365 nm) spectros- copy. Calibration standards were used for identification (retention time) and to construct a linear calibratior. curve. Statistical methods. Spearman's rank correlation coefficients were used to assess the relationship between linolenic acid and condensate mutagenicity. P values result from two-sided tests. Results Mutagenicity testing using the Salmonella mutation assay with tester strain TA98 and S9 liver homogenates confirmed that unrefined Chinese rapeseed condensates are mutagenic (Table I); the number of mutants (i.e., revertants) per plate increased fourfold and a doseresponse effect was observed until cytotoxicity occurred. The condensates were not mutagenic without S9 liver homogenates, indicating a requirement for metabolic activation (data not shown). Individual condensates were not mutagenic using the tester strain T.4104. with or without S9 (data not shown). Soybean oil was weakly mutagenic, but peanut oil. sesame oil, lard, and canola oil were not under these conditions. ' We hypothesized that the condensate mutagenicity was related to linolenic acid content because of previous data showing that mutagenicity could be eradicated with hydrogenation of the oil (11) and that linolenic acid is the fatty acid in cooking oil with the greatest number of Table 1. Mutagenicity of cooking oil condensates: TA9R Sulnronclla mutation assay* Dose, Chinese mg/plate rapeseed Soybean Peanut Lard Canola Sesame LiiiolelL Vehiclet 0.1 0.33 I .o 3.33 10.0 t'ositrveb. -- 19f I 42 f 9$ I I I f 16+ 157 f IO$ 29f8 44f2 136f I I 23f6 23 f 9 45f7 68fW ISf3 9f4 181 f 16 28f5 21 f S 26 f 2 25 f 3 15f3 19f3 I47 f 16 17f7 20f4 22f6 27f3 24f8 18fS I45 f 25 ''Zn ?C reven.?.its; mean of triplicates f SD. f 'q eihcle cmtr:il = dimethyl sulfoxide. 0.1 mUplate. $Positive re'L I t (>twofold increase in revertants over the vehicle control) 5Posi:ive COI :td = 2-aminoanthracene, 0.5 mglplate. 26f7 22 f 4 19f3 30f3 24fII 21 f 3 152f I 27 f5 22 f 4 23 f S 23 f4 14f8 17f2 I50f 9 - 17f4 24 f 6 33 f 5 20f2 IRf5 12f2 1 3 5 f IO Journal of thr: National Cancer Institute, Vol, 87,No. 1 I , June 7, 1995 Linolenic 24 f 3 85 f 4$ 415 f 26$ 29f 10 lOf2 OfI 123f 15 Erucic 19f2 24 f 4 24 f IO 24 f 7 22 f3 29f6 tn3f IO ___ REPORT 837 double bonds (n = 3). Further, mutagenic Chinese and nonmutagenic U.S. (canola) rawseed oits have much higher contents of linolenic acid compared with other 8 Rapeseed Oil (9%) 7 - A Peanut Oil (0%) 6 - 0 Soybean Oil (6%) nonmutagenic oils, such as peanut (9.8%, 11.3%. and 0.6%, respectively [unpub- 5- lished data]). Our tests indicated that lino- lenic acid was highly mutagenic (Table 1). Moreover, when linolenic acid was added to the peanut oil before heating, the results became positive. A correlation with the number of revertants was found for the concentration of linolenic acid present either in native oils or when 111111111111111 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 added as a supplement in different proportions to peanut oil (Fig. 1; Spearman's Final Linolenic Acid Concentration (%) rank correlation coefficient r = .83; P = .OOO4). In contrast, linoleic acid and crucic acid were not mutagenic (Table 1). Qualitative analyses of volatile organic Fig. 1. Dose-response curve showing that the concentrationof linolenic acid is correlated with the number of revertants in the Salmonella mutation assay ( r = .83; P = .OOO4). Oils were tested either alone or with the addition of increasing amounts of linolenic acid. Parentheses indicate starting concentrationof linolenic acid in native oils. Data represent thc highest number of revertants found for each of the oils tested (seeTable I). compounds from heated Chinese rape- seed, canola, soybean, and peanut oils were undertaken by gas chromatography acid, in contrast, at its boiling point studies, suggest that wok cooking with and mass spectroscopy (Fig. 2). The (240 "C) emitted sevenfold higher levels * Chinese rapeseed oil may pose a lung analysis identified 54 major and minor of 1,3-butadieneand 76-fold higher levels cancer risk in Chinese women. We o b compounds among the four oils, most of of benzene, compared with Chinese rape- served that heated cooking oils emit a which were present in all oils. There were seed oil also heated to 240 'C. Linoleic variety of toxic agents, some of which are semiquantitative and quantitative dif- acid, however, emitted only twofold known as potential or probable human or ferences in volatile organic compounds higher levels of 1.3-butadieneand 12-fold laboratory animal carcinogens (5.1420) emitted from the heated cooking oils. The higher levels of benzene, compared with or mutagens (14.2122). including 1,3- Chinese rapeseed oil tested here had the the Chinese rapeseed oil (at 240 T),butadiene, benzene, acrolein. formal- highest content of major peaks and sub- Erucic acid emitted threefold lower levels dehyde, and acetaldehyde. Some of these stantially more minor peaks. Canola oil of I ,3-butadiene and sixfold lower levels compounds may cause acute neurologic, was similar, but soybean and peanut oils of benzene. The addition of butylated respiratory. and dermatologic effects. n- had substantially less. Quantitatively, a hydroxyanisole (0.1% vol/vol) to Chinese Hexane is associated with chronic neuro- similar pattern was observed for 1,3-buta- rapeseed oil before heating lowered the logic effects (23). It is noteworthy that diene and benzene (Table 2) compared emission of total volatile organic com- some agents (e.g., acrolein and alde- with the semiquantitative differences for pounds and reduced the emissions of 1,3- hydes) cause eye irritation. coinciding total peaks. For all oils, 1,3-butadiene, butadiene by 50% and benzene by 73% with the report of eye irritation related to acetaldehyde, n-pentane, acrolein, propa- (Table 2). wok cooking and lung cancer (9). Un- nol, n-hexane, propionaldehyde, and ben- Aldehyde emissions from the cooking refined Chinese cooking oil. which has zene were observed. Of particular note, oils also were determined (Table 2). In been associated with lung cancer risk (9) the emission of 1,3-butadiene and ben- general, Chinese rapeseed oil emitted the and mutagenicity (If), had the greatest zene was approximately 22-fold and 12- most aldehydes, while peanut oi! emitted amount of volatile decomposition pro- fold higher, respectively, from heated the least. Acrolein was produced from all ducts, but levels were only somewhat unrefined Chinese rapeseed oil than from the heated oils, although there were no higher than U.S. canola oil.. Peanut oil heated peanut oil (Table 2). appreciable differences among the oils emitted substantially less. Heating the oils in a wok at lower except for peanut oil, which was substan- Our study revealed that condensates of temperatures resulted in qualitatively less tially lower. Linolenic acid produced at the volatile emissions from Chinese rape- volatile organic compounds, as evidenced least two times more acrolein than any seed oil, soybean oil, and linolenic acid by much smaller peak heights (data not oil. There were no appreciable differen- are mutagenic in the Sulmonellu mutation shown). Quantitatively, compared with ces in formaldehyde emission among the assay. Comparable data from other stud- Chinese rapeseed oil heated to 275 'C, .oils and acids. ies are limited, but volatile emissions there were fourfold and 14-fold lower from several cooked foods (e.g., beef) levels of 1,3-butadiene and onefold and Discussion sevenfold lower levels of benzene when have previously been found !o be mutagenic (2425). as were samples of res- the oils were heated at 240 'C and 185 'C, Our findings, together with car- taurant kitchen air (2627).Furthermore, respectively (data not shown). Linolenic cinogenesis bioassays and epidemiologic some studies in western countries indicate 838 REPORT Journal of the National Cancer Institute, Vol. 87, No. 11, June 7. 1995 6- .7000 6m 51)oo 4Ooo 2000 loo0 r0 ; I I I I I I I 0 10 20 30 40 50 60 70 80 Rapeseed Oil D I IIIIII 0 10 20 30 40 50 60 70 00 Canola Oil 0 10 20 30 40 50 6a 70 80 Soybean Oil Retention%me ( m i 4 0 10 20 30 40 50 60 70 80 PeanutOil Retention'lime (min) Fig. 2. Gas chromatography and mass spectroscopy analysis olatile emissions of cooking oils heated in a wok. The cooking temperatures were chosen on the basis of typical cooking temperatures in Shanghai and/or oils began to emit smoke. Chinese rapexedoil (A) and U.S.canola oil (B)were heated at 275 'C. Chinese soybean oil (C) and Chinese peanut oil @) were 260-265 'C. Smaller peaks. except for 1.3-butadiene and benzene. were not identified. The p e d s that were identified are numbered and correspond to n-butane (1). butene isomer (2). 1.3-butadiene (3). I-butene (4). acetaldehyde (5). I-pentene (6).n-pentane (7). acrolein (8). propanol (9). 1.3-pentadiene isomer (IO), I-hexene kI I),n-hexane (12). n-butanol(13). n-butanal(14). methyl vinyl ketone (IS). hnzene (16). I-hep- tene (171, n-heptane (IW,C7Hn isomers (19). C ~ H Misomers (2$, n-pentanal(21). I-pentene-3sl(22), toluene (23). I-octene(24). n-octane (25). rrons-4.octene or CRHMisomer (26). I-pentanol (27), n-hexanal (28). CeHi4 is0 r (29). cyclopentane and carbaldehyde (30).Cz-alkyl benzene (31). I-nomne (321,CRHI?isomer (33). C7HxO isomer (34A n-nonane (3% CsHtn isomer 136). cnal(37),n-butylcyclopentane (38).n-hcptanal(39),Cdi16isomer (40).rruns-2-heptanal(4I). un- saturated hydrocarbons (421, CioHin isomer (43). CIiH24 (44).octa-l,7-dienc-3-ol(45).n-octanal(46). I-decene (47). C I I His~om~ ers (48). rruns-2-gtenal (49). unsaturated hydrocarbons (50).CI1H22isomers (5 I ). C12Hzzisomers (53).and unsaturated hydrocarbons (54). Sample Rapeseed oil Rapeseed oil + buiylated hydroxyanisole, 0.1% Canola oil Soybean oil Peanut oil Peanut oil + linolenic acid. IO% Peanut oil + linolenic acid. 20% Linolenic acid Linoleic acid Tnble 2. Detection df volatile organic communds in cmkine oil emissions I. %Butadiene* I Benzene* Formaldehyde, p a t , Acetaldehyde. p@t SO4 239 I 71.2 306.9 247 663 10.0 40.0 200 664 30.2 57 450 39. I 23 203 22.8 -I23 626 49.0 725 21 200$ 3390 1wooo$ 43.3 279 23 300f 52.3 94.6 112.7 45.1 14-7.0 239.0 102.4 Acrolein. p@t 39 I .X 36.0 320.0 442.7 49.0 24-7.0 ~67.4 286.9 *Calculated using National Institute of Standards and Technolo reference standard: expressed as n a . t- = not tested. $Estimated concentration: exceedshighest standard. that restaurant cooks have elevated risks of lung cancer (28,29),even when risks are adjusted for tobacco use. aSeve I lines of evidence suEgest an importa t role for linolenic acid. Spcifically, mutagenicity was associated with linolenic acid concentration. and the mutagenicity was reduced by hydmgenation or by adding butylated hydroxy- Journal of the National Cancer Institute,Vol.87, No. I I. June 7. 1995 REPORT 839 anisole to the oil before cooking. Also, heafing pure linolenic acid resulted in the greatest production of mutagenic, volatile decomposition products. Linolenic acid is known to cause bad odors (30), which can be reduced by hydrogenation or the addition of antioxidants, and might explain the customary use of high-temperature cooking (275-280 'C) in Shanghai. However, the presence of linolenic acid does not fully explain the mutagenicity results because Chinese rapeseed oi: is mutagenic, while U.S.canola oil is not, although they have similar linolenic acid contents. Furthermore, soybean oil is weakly mutagenic even though it has less linolenic acid than canola oil. We SIISpected that the erucic acid content or other chacteristics of Chinese rapeseed plants, including additives or refining processes, may be involved, although none of these factors consistently explained the effects when we tested different lots of refined Chinese rapeseed oil or linolenic acid (data not shown). Therefore, the most plausible explanation for these findings is that canola oil contains specific inhibitors to oxidation or that Chinese rapeseed oil contains promuta- genic compounds in addition to linolenic acid. Among the chemicals detected in the cooking oil emissions, 1,3-butadiene has been classified as a probable (group 2A) human carcinogen by the International Agency for Research on Cancer ( 5 J i ) and a suspected human carcinogen by the American Conference of Governrr,ental Industrial Hygienists (20). At exposures as low as 6.5 ppm, this compound is a lung carcinogen in laboratory mice (15,16,19), although rats are typically resistant (32).The interspecies variation is likely related to differences in metabolic activation and detoxification (3.335). Epidemiologic evidence to date has Ynot documented a lung cancer risk lated to 1.3-butadiene exposure (36),although some studies suggest an excess of hematologic malignancies ( I 737). Other volatile decomposition products of the cooking oils may contribute to lung cancer risks. While benzene is an established human leukemogen (14). its association with lung cancer is not clear, although an increased risk has been reported in benzene-exposed Chinese workers (18) and experimental studies also indicate that benzene may be a lung carcinogen (38). Some studies ( 3 9 4 2 ) have related formaldehyde to lung cancer, but the evidence is inconclusive. Acrolein is mutagenic in a number of different test systems (1421,22),but human studies are lacking. Some of the chemicals detected in cooking-oii emissions are also evident in cigarette smoke. In particular, 1,3-butadiene (43) has been reported at levels ranging from 19 pg to 75 pg per cigarette in mainstream smoke, depending on whether the cigarette smoke is filtered. The amount of 1,3-butadiene generated from heating Chinese rapeseed oil under the conditions reported here, however, is approximately 0.6 pg. Therefore, for I ,3butadiene and also for benzene, acrolein, and the other detectable voldtile products, the lifetime eiposure from wok cooking appears to be substantially less than that for a heavy smoker. However, these calculations may underestimate the exposure and risk because they do not sufficiently consider peak exposures during the initial generation of volatile compounds. This study suggests several methods that could reduce exposures toLv201atileor- ganic compounds and mutagenic substances from heated cooking oils. These methods include cooking at lower temperatures, addition of antioxidants, hydrogenation during refining, and increasing ventilation during cooking. Further studies are needed to clarify the relationship of cociking oil vapors, other environmental exposures, and genetic susceptibilities to lung cancer in Chinese populations. Studies also are needed in the United States that will evaluate a lung cancer risk associated with exposure to cooking oils, especially in different ethnic and racial groups, including Americans of Chinese descent. Finally, the effects of cooking specific foods and the volatilization of other potential mutagens and carcinogens should be investigated. References ( 1 ) 'Cancer incidence in five continents. Volume V. IARC Sci Pub1 88:l-970. 1987 (2) Jin F, Devesa SS.Zheng 'A', et ai: Cancer incidence trends in urban Shanghai, 1972-1989. Int J Cancer 53:764-770. 1993 (3) Fraumeni JF Jr, Mason TJ: Cancer mortality among Chinese Americans, 1950-69. J Natl Cancer lnst 52.659-665.1974 (4) MacLennan R, Da Costa J. Day NE, et nl: Risk factors for lung cancer in Singapore Chinese,a population with high female incidence rates. Int J Cancer 20854-8Mk 1977 (5) International Agency for Research on Cancer: Cancer lncidencc in Five Continents. vol VI. Lyon: IARC, 1992.p~1-1080 (6) Mumfxd JL, He XZ, Chapman RS, et al: Lung cancer and indoor air pollution in Xuan Wei, China. Science 235:217-220. 1987 tn Hinds MW, Stemmermann GN,Yang HY, et a!: Differences in lung cancer risk from smok- ing among Japanese, Chinese and Hawaiian women in HaNaii. Int J Cancer 27:297-302. 1981 (8) Gao YT, Blot WJ, Zheng W. et al: Lung can- cer and smoking in Shanghai. Int 1 Epidemiol 17~277-280.1988 (9) Gao YT, Blot WJ. Zheng W. et al: Lung can- cer among Chinese women. Int J Cancer 40604-609,1987 (10) Wu-Williams AH, Dai XD, Blot W, et al: Lung cancer among women in northeast China. Br J Cancer 62:982-987. 1990 (11) Qu YH. Xu GX, Zhou JZ, et al: Genotoxicity of heated cooking oil vapors. Mutat Res 298:105-111,1992 (12) Maron DM. Aaes BN: Revised methods for the Sulmonclkr mutagenicity test. Mutat Res 113:173-215.1983 (13) Eight Peak Index of Kass Spectra. Mass Spectra Data Center, I974 (14) International Agency for Research on Cancer: IARC Monographs on the Evaluation of Car- cinogenic Risks to Humans. Overall Evalua- tions of Carcinogenicity: an Updating of IARC Monographs, vols 142. Lyon: IARC. 1987 (15) Melnick RL, Huff J. Matanoski GM: Car- cinogenicity of l .3-butadiene [letter] [see comment citations in Med!ine]. Lancet 340:724-725,1992 (16) Melnick RL. Huff J: 1,3-Butadiene: toxicity and carcinogenicity in laboratory animals and in humans. Rev Environ Contam Toxic01 124:lll-144,1992 (17) Landrigan PJ: Critical assessment of epidemiologic studies on the human car- cinogenicity of I .3-butadiene. Environ Health Pe~~t86143-147,1990 (18) Yin SN. Hayes RB. Linet MS: A cohort study of cancer among benzene-cxposed workers in China: overall rcsulu.Am J Ind Med.In press (19) Huff JE. Melnick RL. Solleveld HA, et al: Multiple organ carcinogenicity of 1.3- butadiene in B6C3FI mice after 60 weeks of inhalation exposure. Science 227548-549. 1985 (20) 1993-1994Thrcshrrld Limit Values for Chemi- cal Substances and Physical Agents and Biological Exposure Indices. Cincinnati,Ohio: Am ConfGoM Indust Hygienists 1 9 9 3 , 1~-124 (21) C u m RD. Ynng LL, Conklin PM. et al: Mutagenesis of xe- pigmentosum f i b blasts by m l e i n . Mutat Rex 209:17-22.1988 (22) Wilson VL. Foiles PG, Chung FL. et al: Detection of acrolein and crotonaidehyde DNA adducts in cultured human cells and canine peripheral blood lymphocytes by 32P- postlabeling and nucleotide chromatography. Carcinogenesis 121483-1J90.1991 (23) Section four: Chemical compounds 11. In Hamilton and Hardy's Industrial Toxicology, 4th ed (Finkel AJ, ed). Boston: Wright-PSG. 1 9 8 3 . ~19~7-237 (24) Rnppaport SM. McCanney MC. Wei ET: Volatilization of mutagens from beef during cooking.CancerLett 8:139-145, 1979 (25) Lofroth G. Stensman C, Brandhorst-Satrkom M: Indoor sou~cesof mutagenic aerosol par- , 840 REPORT I Journal of the National Cancer Institute. Vol. 87, No. 1 I . June 7, 1995 I ticulate matter: smoking, cooking and incense .burning. Mutat Res 261:21-28. 1991 (26) Gee A: Decrcase in essential fatty acid content of edible fate Jurh!g the frying process. Z Ernahrungswiss 2I :191-201, 1982 (27) Teschke K. Hertzman C. Van Nenen C. et al: Potential exposure of cooks to airborne mutagens and carcinogens. Environ Res 50:296308,1989 (28) Coggon D, Pannett B. Osmond C, et at: A sur- vey of cancer and occupation in young and middle aged men. 1. Cancers of the respiratory tract. Br J Ind Med 43:332-338,1986 (29) Dubrow R, Wegman DH: Cancer and occupa- tion in Massachusetts: a death certificate study. Am J Ind Med 6207-230.1984 (30) Mounts T L Odor consideratioils in the use of frying oils. J Am Oil Chem Soc 56:659-663, I979 (31) IARC Working Group on the Evaluation of Carcinogenic Risks to Humans: Occupat,ional Expnsures to Mists and Vapoun From Sirong Inorganic Acids; and Other Industrial Chemi- cals, vol54. Lyon: IARC. 1992. pp 1-336 (32) Owen PE, Glaister JR, Gaunt IF. et at: Inhala- tion toxicity studies with 1.3-butadiene. 3. Two year toxicity/carcinogenirity study :Irats. Am Ind Hyg Assoc J 48:4O7-4 I 3,1987 (33) Kohn MC, Melnick R L Species differences in the production and clearance of 1.3-butadiene m?tabolites: a mechanistic model indicates predominantly physiological, not biochemical. control. Carcinogenesis14:619-628.1993 (34) Csanady GA. Guengerich FP. Bond J A Comparison of the biotransformation of 1.3-bumdiene and its mer.Aite, butadiene monoepoxide, by hepatic and pulmonary tissues from humans. rats and mice [published erratum appears in Carcinogenesis 14:784, 19931.Carcinogenesis 13:1143-1153.1992 (35)Sabourin PJ, Burka LT. Bechtold WE. et al: Species differences in urinary butadiene metabolites: identification of 1,2-dihydroxy-4-(N- acetylcysteinyl)butane.a novel metabolite of butadiene. Carcinogenesis 13:1633-1638. I992 (36) Matanoski GM. Santos-Burga C. Schwartz L: Mortality of a cohort of workers in the styre-butadiene polymer manufacturing industry (1943-1982).Eiiviron Hcahh Perspect 86107117.1990 (37) Santos-Burga C, Matanoski GM, Zeger S. et al: Lymphohematopoietic cancer in styrene- butadiene polymenmtion workers [see comment citation in Medline]. Am J Epidemiol 136:843-854,1992 (38) Huff JE, Haseman JK. DeMarini DM. et at: Multiple-site carcinogenicity of benzene in Fixher 344 rats and BK3FI mice. Environ Health Persyect 82:125-163,1989 (39) Gardner MJ. Pannett 6.Winter PD. ei al: A rohort study of workers exposed to foimal- dehyde in the British chemical industry: an up date. Br J Ind Med 50827-834.1993 (40) Stayner LT. Elliott L, Blade L, et at: A retspcctive cohort m m l i t y study of workers exposed to formaldehyde in the garment industry. Am J Ind Med 13:667-681,198A (41) Blair A, Stewart P. OBerg M. et al: Mortality among hdustrial workers exposed to formaldehyde isre comment citaticn in Medline]. J Nail Cancer Inst 761071-1084,1%6 (42) Bertazzi PA, Pesatori AC. Rad& L, et at: Ex- posure to formaldehyde and cancer mortality in a cohon of worken producing resins. Scmd J WorkEnvim Health 12:461-468,1986 (43) Bnrnnemann KD, Kagan MR, Cox JE. et al: Analysis of 1.3-butadiene znd other selected gas-phase cornponcnu in c i g m t e mainbtrsarn and sidestmam smoke by gas chrohatugraphymass selective detection. Carcinogenesis I1:1863-1868.1990 \ Notes We thank Dr.Richard Sans for nis-eyxrtise in the Sulmonelku mutation assay. Ms. Elise Bowman-for-, assistance in the preparahi of cooking oil conden- sates, and Ms.Domthea Dud& for editorial assis- tance. Manuscript received November I I , 1994; revised March IO.1995:accepted March 27. 1995. PREDOCTORALSTIPENDS IN NUTEmON SCIENCESICANCER PREVENTION The Universityof Alabama at Birmingham NIH-funded predoctoral positions are available for pursuit of the Ph.D. in Nutrition Sciences. The training includes COUU-SCS in basic and clinical nutrition and in public health and chronic disease prevention; exposure to patients with nutritional disorders and to disease prevention projects involving nutritional intervention;mil research, which will be related to prevention of chronic disease, especially cancer. Support includes stipend, tuition and fees, supplies, and travel to scientific meetings. Applicants must be U.S. citizens or -anent residents. Minority applicants are strongly encouraged. Interested persons should send a letter describing their interest, plus curriculum vitae to Dr. Douglas C. Heimburger, Deprlrtment of Nutrition Sciences, UAB Station, Birmingham, AL 35294-3360; or call 205-9347058. The Universiwof Alabama at Birmizgham is an Aflrmative ActiodEqual Opportuniy EmpIoyer. Journal of thc National Cancer Institute,Vol. 87, No. I I , June 7. 1995 KEWRT 841