Document 10g3pJxVeQ7eeDYVY3pJjg3g5

Cancer Causes and Control, 1997,8, pp. 386-405 Oils and cancer Paige E. Tolbert (Received 13 May 1996; accepted in revised form 23 September 1996) Epidemiologic evidence on the relationship between mineral oil exposure and cancer is reviewed. The review is restricted to occupations involving substantial dermal and inhalational exposure and for which an epidemiologic literature exists: metal machining, print press operating, and cotton and jute spinning. Mineral oils are complex mixtures of aliphatic hydrocarbons, naphthenics, and aromatics, the relative distribution of which depends on the source of the oil and the method of refinement. End-use products contain a variety of additives, and contamination by other agents generally occurs during use. Suspect agents include polycyclic aromatic hydrocarbons (PAH)(par- tic u1a r1y be n z [a]p y r en e), n i t r o s a m i n es, ch 1o r i n a t ed p a r a f f i n s, 1o n g-ch a i n a1i p h a t ics, s u 1f u r, N-phenyl-2-naphthylamine, and formaldehyde. The heterogeneity of this exposure makes epidemiologic study difficult and meta-analysis inappropriate. Nonetheless, several associations emerge from the literature with varying degrees of support. There is clear evidence that early formulations of mineral oils used in cotton and jute spinning and in metal machining were carcinogenic t o the skin. Associations of mineral oil exposure with laryngeal and rectal cancer have received some support in the literature, particularly with respect t o straight oils. Evidence is suggestive that grinding operations (which can entail either mineral oil-based or ethanolamine-based fluids) are associatedwith excess risk of cancer of the esophagus,stomach,and pancreas. A number of bladder cancer case-control studies have noted an association with work as a machinist. There is limited evidence of an association with cancer of the colon, prostate, and sinonasal region. Several studies of printers have yielded positive findings for lung cancer, whereas studies in metal machinists have been generally negative. The PAH and nitrosamine content of current formulations is lower than in the past and the implications of these changes in composition to the carcinogenicity of the formulations are not yet known. Cancer Causes and Control 1997,8,386-405 K e y words: Cutting oils, cutting fluids, metalworking, mineral oil, nitrosamines, polycyclic aromatic hydrocarbons. Introduction Oils derived from petroleum or shale have been used in a wide variety of occupational settings and applications. Many of these involve enclosed systems with little opportunity for human exposure. There are several occupational environments, however, in which an oil mist is generated and in these situations the opportunities for dermal exposure or inhalation exposure, with concomitant ingestion, are substantial. Such occupations include metalworking, print press operating, and cotton and jute spinning. Workers in these occupations, particularly metalworkers, have been the focus of a number of epidemiologic investigations. Evidence regarding carcinogenic effects in these worker populations has been the subject of several reviews," including a comprehensive review conducted by a panel convened by the International Agency for Research on Cancer (IARC) in 1984.' The purpose of this paper, therefore, is to briefly summarize the findings presented in the IARC review and to present findings Dr Tolbert is with the Departments of Environmental C Occupational Health and Epidemiology, Rollins School of Public Health of Emory University. Address correspondence to Dr Tolbert, Department of Environmental & OccupationalHealth, Rollins School of Public Health of Emory University, 1518 Clifton Road, NE, Atlanta, G A 30322, USA. This review was sponsored in part by the Occupational Disease Panel of the Ontario Ministry of Labor. Additional support was provided by Dr Tolbert'sFIRST award, N I H R29 CA63622-01Al. 386 Cancer Causes and Control. Vol 8. 1997 01997 Rapid Science Publishers Oils and cancer reported in the last decade since publication of the IARC review. The IARC review concluded that: "there is sufficient evidence from studies in humans that mineral oils (containing various additives and impurities) that have been used in occupations such as mulespinning, metal machining and jute processing are carcinogenic to humans."' This conclusion was based on evidence of a strong association of mineral oil exposure with squamous cell carcinoma of the skin, primarily scrotal cancer, from early studies of workers involved in these occupations.' Mineral oil processing and production In the latter half of the 19th century, oils from oil-bearing shale largely displaced animal and vegetable oils as lubricants. Shale oil was processed primarily in Scotland and to a lesser extent in Romania.' The subsequent discovery of petroleum and the competition from this source led to the decline of the shale industry and since the 1930s, almost all of the worldwide production of lubricating oils has been from petroleum. A small amount of shale oil processing continues to occur in China, the United States, and the former Soviet republic^.^ Petroleum refining processes have changed dramatically during this century, with a trend toward increasingly refined oils containing reduced levels of polycyclic aromatic hydrocarbons (PAH) and other impurities.' Initially, acid treatment was widely used. Treatment with acid removes unstable hydrocarbons, resins, asphalts, and sulphur-, nitrogen-, and oxygen-containing compounds, improving the color, odor, and stability of the base oil. Over the past several decades, acid treatment has been largely replaced by solvent refining and/or hydrotreating. Solvent extraction removes olefins, naphthenes, and paraffins. Mild solvent extraction removes less of the aromatic content (including PAHs) than severe solvent extraction. Hydrotreating, often performed in conjunction with solvent extraction, entails catalytic hydrogenation of base oils to saturate olefins and render the oil more paraffinic. Severe hydrotreating converts more of the aromatic compounds to naphthenes. The resulting base oil is then formulated into an end-use product by the addition of various additives, described below. Exposure circumstances Metalworking In metalworking operations, fluids are directed at the working surface in a spray or stream, for the purpose of cooling and lubricating the metals being worked. Large quantities are used and generally they are collected and reused. Metalworking fluids are variously referred to as cutting oils or fluids, machining fluids, lubrication fluids, cooling emulsions or oils, grinding or drilling fluids or oils, and - as the multitude of names suggest - they com- prise a group of fluids defined more by function than by chemical composition. A wide variety of formulations has been developed for metalworking, most but not all of which contain mineral oil. Metalworking formulations can be broadly categorized into four classes': (i) Straight oils, also called neat oils, are naphthenic or paraffinic mineral oils, with additives such as sulfurized or chlorinated fats and corrosion inhibitors, and sometimes containing animal or vegetable oils. Straight oils contain no water, although small quantities of water can become entrained in the oils in the course of their reuse; entrained water can allow the growth of microorganisms, sometimes necessitating addition of biocides to the fluid. Of the various classes of metalworking fluids, straight oils have the greatest lubricity and lowest cooling capability, and therefore are preferred for use in operations that do not generate very much heat; (ii) Soluble oils, also referred to as emulsifiable, water-miscible or suds oils, are aqueous emulsions of mineral oil (generally from a more viscous stock than that used in straight oils), with added emulsifiers (e.g., petroleum sulfonates, amine soaps, sodium naphthenates), biocides (e.g., formaldehyde-releasing triazines, chlorophenols, hexachlorophene, quaternary ammonium compounds), corrosion inhibitors (e.g., amines, fatty oils, sodium nitrite, sulfurized fatty oils), antifoaming agents, dyes, and high pressure stabilizers. Solublefluids generally are purchased in concentrated form, and diluted in water five- to 50-fold. After dilution, the fluid appears milky and opaque. Soluble fluids are cheaper to use than straight oils, and have better cooling properties but lesser lubricating qualities; (iii) Synthetic fluids, also called chemical fluids, are generally alkanolamines(most often, ethanolamines) in aqueous solution, devoid of mineral oil. Additives include nitrites, nitrates, phosphates and borates for corrosion inhibition, surfactants (e.g,, complex alcohols and esters), biocides, silicone as a lubricant and antifoaming agent, and frequently dyes. They are more stable than soluble oils, and have excellent cooling properties. Synthetic fluids are obtained in concentrated form and diluted 20- to 100-fold. Although not the subject of this review because they Cancer Causes and Control. Vol 8. 1997 387 ?E. Tolbert contain no mineral oil, synthetic fluids are mentioned because in many studies of metalworkers it is difficult to distinguish among the types of fluids to which the workers were exposed and in interpreting those studies, it should be kept in mind that synthetic fluids may be responsible for any associations observed with machining fluids rather than a mineral oilbased formulation; (iv) Semi-synthetic fluids are a hybrid of soluble and synthetic fluids. They are 15 to 40 percent oil-based (pre-dilution), with glycols, polyols, and ester lubricants. Straight oils have been in use since the nineteenthcentury, whereas soluble and synthetic fluids were introduced in the 1940s when high-speed machnery made necessary the superior cooling capabilities of water-based fluids. In 1984, it was estimated that 50 percent of metalworking fluids in use were soluble oils, 45 percent were straight oils, and five percent were synthetic and semi-synthetic.' Hallock et a1 conducted an exposure reconstruction to estimate historic levels of metalworking fluids in the vicinity of machining and grinding operations at several automotive manufacturing plants. Total particulate levels were estimated to be 5.4 mg/m' in the 1960s and levels declined to approximately 1.8 mg/m' in the 1980s. More recent measurements by other investigators suggest that typical levels have declined further to less than 1mg/m3.'s6 More detailed presentations of results of monitoring machining fluid exposures in different settings are presented elsewhere."' Printing Over the last two decades, offset letterpresses have largely replaced rotary letterpresses, but for most of this century, large-scalenewspaper production made use of rotary letterpress printing machines.' The high speed rotation of rotary presses generated a mist of ink droplets. Later rotary press models were less open than earlier models, but operated at higher speeds and thus tended to generate levels of respirable particulate at least as high as the earliermodels.'a Rotary letterpresses made use of ink composed of an emulsion of carbon black in mineral oil, with additives and pigments. The carbon black is amorphous elementalcarbon created by a furnace process, the diameter of the dispersed aggregate particles ranging from 0.1-0.2 microns." Carbon black adsorbs benz[a]pyrene and other PAHs and, therefore, these agents are present as contaminants of carbon black.' Typical composition of printing ink by weight is 85 percent mineral oil, 12 percent carbon black, and three percent induline dye toner (to impart a blue hue).9 A review of ambient levels of oil mist in newspaper press rooms indicated a range of levels from 2.0 to 16.6 mg/m'." In a detailed investigation of the New York Times 388 Cancer Causes and Control. Vol 8. 1997 pressroom, the average concentration in the vicinity of modern presses was determined to be 12.2 mg/m'.'O The respirable fraction was estimated to be about 15 percent. Cotton spinning Historically, the job of mule-spinning entailed substantial exposure to mineral oil through the operation of an apparatus for spinning cotton. A mechanized rack carried the spindles back and forth to receive the cotton as it was spun. A lubricating oil was applied to the cotton and was dispersed into the air by the spindles as they revolved. In addition, a bar approximately three feet high and extending the length of the machine, became moist with the oil. As mule-spinners leaned over the apparatus to tend to the machine, the groin area of their clothing became drenched with the oil that had collected on the bar. Spindle oils were composed of shale oil until the 1870s when petroleumderived oils came into use. The petroleum-derived spindle oils were paraffinic, similar to the base oil used in metalworking formulations but usually a distillate of a somewhat lighter `cut,' ;.e., having a lower boiling point. In the past 40 years, the cotton industry has become increasingly automated and the opportunity for occupational exposure to mineral oil has greatly diminished. Mineral oil still is used occasionally as an overspray in early steps of processing (at a rate of 1 lb/1OO,OOO Ibs cotton), but its use is minimized because oil interferes with the ring spinning, air jet spinning, and open-end spinning technologies that have replaced mule-spinning (personal communication, L. Howard Olson, School of Textile and Fiber Engineering, Georgia Institute of Technology, April 1996). Jute processing Jute is treated with an emulsion of oil and water prior to spinning. Until the 1920s, whale oil was used for this purpose, but since then, mineral oil has been used. Oil is added during `batching,' the first step of processing in which the jute is softened between rollers. The ribbon then goes through `drawing,' during which it is thinned and twisted onto bobbins. Jute spindles revolve at approximately 4,000 RPM." Jute processing practices have remained essentially unchanged since the 1950s. Overall, the industry has experienced decline due to the introduction of synthetic substitutes and jute processing is now performed primarily in India and Bangladesh (personal communication, William Yaffa, President, Burlap and Jute Association, April 1996). Methodologic difficulties of studying mineral oils As an exposure for epidemiologic study, mineral oil formulations present a number of methodologic difficulties. Oils and cancel- . ' Forniulations are complex mixtures. A wide variety of formulations have been in use, some having no ingredients in common with others. With use, the fluids can undergo chemical changes such as pyrolysis. Recycling and reprocessing of fluids compound this issue. Many of the suspect carcinogens (e.g., PAH and nitrosamines) are unintentional contaminants, and, of course, are not listed in formulation descriptions. Composition of oils also varies as a function of the geographic source of the crude; with, for instance, Venezuelan crude having higher PAH content than Oklahoma crude. There have been secular changes in composition at ill-defined points in time such as the above-described changes in refining processes which have led to a reduction in PAH content in the base oils used in the end products. Finally, the settings in which substantial mineral oil exposure occurs generally entail co-exposure to other agents: in metalworking, for example, there is co-exposure to metals and abrasives, and in printing, there is co-exposure to carbon black and paper dust, making it difficult to distinguish among these agents with regard to their carcinogenic activity. Given the variability and complexity of the exposure, it is not surprising that there is a lack of consistency across epidemiologic studies of mineral oil-exposedpopulations. Further, pooling of findings across studies will lead to an underestimation of any effects due to the heterogeneous nature of the exposure. Case serieddescriptive studies The association of exposure to early lubricating oils formulated from either shale or petroleum with squamous cell carcinoma was so strong that it was apparent through simple case-series observations. Some of the first indications of carcinogenicity came from studies of the British cotton spinning industry. Workers spinning cotton became drenched in lubricants in the course of using spinning mules. A clear clustering of scrotal and other skin cancers was observed in these worker^.^'"^ Kinnear e t a1 I' extended these observations to a second textile industry, jute processing. Higher proportions of premalignant and malignant skin changes were noted among jute processers presenting at a dermatologic clinic than among other patients at the clinic.Kinnear et afl2 followed this up with a dermatologic survey of workers in the jute industry and found that spinners had the most marked degenerative changes, supporting the hypothesis that mineral oil applied to jute and dispersed during spinning was the responsible agent. Perhaps the first observation that fluids used specifically for metal working were carcinogenic to the skin was by Cruickshank and Squire.I5An excess of skin cancer was noted in metalworking fluid-exposed workers in the light engineering industries of the West Midlands area of England. A subsequent study in this region indicated that between 1950 and 1967, 187 cases of scrotal cancer occurred in this geographic area, two-thirds of which appeared to be attributable to machining fluid exposure.'' The incidence of scrotal cancer among toolsetters and machine operators in the Birmingham Cancer Registry from 1936 to 1972 was estimated to be about 17 times higher than that in the general male population." Outside England, an overrepresentation of metal workers exposed to metalworking fluids was noted in a case series of scrotal cancer in the Haute-Savoie area of France." Among 21 men with scrotal cancer, seen 1959-63 at a local hospital, all were found to have been screwcutters for over 10 years. Further investigations in the Haute-Savoie areal9confirmed this finding for scrotal and other skin cancers. Population-based and registry-based studies A number of population-based or registry-based casecontrol studies have been performed in which potential mineral-oil exposure was assessed using occupational histories. Such studies generally are limited by the fact that specificinformation on exposures (e.g., type of metal- working fluid) is not available. The studies are presented in Tables 1 through IO, and some of the more informative studies are summarized below. Findings from case-control studies nested in cohorts are presented in the next section, Cohort-Based Studies. Skin cancer An excess risk of scrotal cancer was seen in American workers exposed to metalworking fluids in a case-control study using the Connecticut Tumor Registry (United States)."The odds ratio (OR)for the association of scrotal cancer with jobs entailing machining fluid exposure was 10, and the risk persisted in the most recent time period of diagnosis, 1966-73. The population attributable risk was estimated to be 57 percent. Esophageal cancer No population- or registry-based case-control studies of esophageal cancer assessing mineral oil exposure were identified. Stomach cancer Kneller et a1 compared occupational data for 13,489 incident stomach cancer cases reported between 1980 and 1984 to the Shanghai Cancer Registry (China) to 1982 census employment data. The standardized incidence ratio (SIR) for workers in metal grinding, polishing, tool sharpening,and machine tool operation was 1.4 (P< 0.01). Other population-based studies assessing exposure to metalworking fluids or printing ink have reported risk estimates near the n~ll."~' Cancer Causes and Control. Vol8. 1997 389 I? E. Tolbert Colon cancer Case-control studies assessing exposure to metalworking fluids have reported little or no association of exposure with colon cancer risk.ZZ,Z* Rectal cancer A population-based case-controlstudy of colon and rectal cancer noted an association of rectal cancer with metalworking fluids." Type of fluid could not be assessed, but a statistically significant OR of two was observed for occupational exposure to metalworking fluids in general. Co-linearity of exposure variables precluded controlling for confounding by other exposures. Pancreatic cancer Mack and Paganini-Hill"reported a slight, nonsignificant association of employment as a machinist with risk of pancreatic cancer in a study of 3,614 incident pancreatic cancer cases diagnosed 1972-77compared with all cancers reported to the registry during that time period. Larynx cancer After controlling for smoking and alcohol use, Zagraniski et a1 26 found a statisticallysignificant association of larynx cancer with work as a machinist or metal grinder. In a second case-control study, Ahrens et a1 27 observed an association of a history of work in jobs with mineral oil exposure and larynx cancer risk that was of borderline statistical significance. Lung cancer Several case-control studies have yielded statistically significant associations of lung cancer risk with or with printing." Bladder cancer A large number of case-control studies have suggested that bladder cancer cases tend to be more iikely to have been metal machinists or engineers than controls, with ORs ranging from 1.5 to 5.0.22.11T-4h0e informativeness of these studies is limited by the fact that these job groupings are broad and encompass exposure to a variety of agents. Prostate cancer No prostate cancer case-control studies evaluating mineral oil exposure were identified. Sinonasal cancer Sinonasal cancer is a rare cancer and, therefore, case-control studies are more useful than cohort studiesin investigating this outcome. A single case-control study using the Connecticut Tumor Registry supports an association of sinonasal cancer with work entailing machining fluid exposure, with a statistically significant OR of 2.8." Cohort-based studies Findings from cohort and proportional mortality studies are summarized in Tables 1 through IO. Table 1. Results for skin cancer from epidemiologic studies of mineral oil-exposed populationsa Author (ref.) Year Location Type of study! analysis No. of exposed cases RRb (CI)' or Pvalue Study population Cohort-based studies Jarvholm et a/'@ 1985 Sweden SIRd Jarvholm & ~avenius~~ 1987 Sweden Population-based studies Roush et a/'' Connecticut 1982 (USA) SIR ORe 5 16.6 P < 0.001 Turners employed at bearing ring manufacturing plant between 1960 and 1980/squamous cell cancer of the skin 7 NA - Turners employed at bearing ring (0 expected) manufacturing plant/scrotal cancer 26 10.5 (4.0-36.9) Ever employed as toolmaker, setter, set-up man, hardener, polisher, automatic screw operator, machinist, or machine operator/scrotal cancer Adapted from NIOSH, Draft Criteria for a Recommended Standard: Occupational Exposures to Metalworking Fluids, 1996.' Only studies of incident cancer included in this table. RR = estimated relative risk. CI = 95% confidence interval. SIR = standardized incidence ratio. OR = odds ratio. 390 Cancer Causes and Control. Vol 5. 1997 Oils and cancer A? ?7 . Table 2. Results for esophageal cancer from epidemiologic studies of mineral oil-exposed populationsa ~- Author (ref.) Location Type of No. of R R ~ (CV Study population 0 Year study1 exposed ______ e Cohort-based studies analysis cases Tolbert et af 43 Michigan SMR~ 22 1.2 (0.7-1.8) Autoworkers ever exposed to straight <:+ 1992 (USA) oils, White 35 1.O (0.7-1.4) Autoworkers ever exposed to soluble L oils, White 8 1.O (0.4-1.9) Autoworkers ever exposed to synthetic oils, White 5 0.8 (0.2-1.8) Autoworkers ever exposed to straight oils, Black 10 0.7 (0.3-1.3) Autoworkers ever exposed to soluble oils, Black RR 8 1.3 (0.5-3.0) Autoworkers exposed to straight oils > 7.5 yrs 9 1.O (0.4-2.4) Autoworkers exposed to soluble oils > 20 yrs 8 1.4 (0.6-3.5) Autoworkers exposed to synthetic fluids > 8 yrs Eisen et aI4' Michigan SMR 6 1.4 (0.5-3.0) Autoworkers, Plant 111, White 1992 Sullivan et a/45.e (USA) Michigan OR' 8 3.8 (1.1-13.2) Autoworkers ever exposed to grinding submitted (USA) with synthetic fluids, 20-yr lag 9 2.5 (0.8-7.2) Autoworkers exposed to grinding with soluble fluids > 12.2 mg/m3-yrs,20-yr lag 6 0.9 (0.3-2.8) Autoworkers exposed to straight fluids Delzell et a / 5 7 1993 Michigan & .Ohio (USA) SMR , 82 > 2.5 mglm3-yrs, 20-yr lag 1.o (0.8-1.2) Automotive manufacturing workers, White 41 0.5 (0.4-0.7) Automotive manufacturing workers, Black Population-based studies Silverstein et a / 59 1988 Park et a/" 1988 Lloyd et a l e 1977 Connecticut (USA) Connecticut (USA) USA PMR~ PMR 13 6 5 1.8 (1.1-3.1) Ball bearing manufacturing workers, White 1.9 (0.7-4.0) Ball bearing manufacturing workers, White 1.7 NS Newspaper pressmen ~~~~~ a Adapted from NIOSH, Draft Criteria for a Recommended Standard: Occupational Exposures to Metalworking Fluids, 1996.' RR = estimated relative risk. CI = 95% confidence interval. SMR = standardized mortality ratio. e Nested case-control analysis of populations studied in Tolbert et OR = odds ratio. and Eisen et aL4' PMR = proportional mortality ratio. Metalworkers The most comprehensive cohort study of machining fluid-exposed workers was a study of approximately 40,000 employees who worked at least three years between 1940 and 1984 at one of three plants in Michigan (US) manufacturing automotive transmission, axle, or chassis c~mponents.~~A"n' extensive exposure reconstruction was undertaken, and each subject's exposure to the three types of fluids (in mg/m'-yrs) as well as various components and contaminants of the fluids (in years) was estimated.* Analysis proceeded from a standardized mortality analysis, using the US population as referent, to an internal analysis using Poisson regression, to nested case-control analyses for specific cancers of interest. The major positive findings in this study are related to digestivecancers and larynx cancer.For esophagealcancer, statistically significant excesses of two- to fourfold were observed among grinders relative to workers who had never worked in grinding, but a dose-response trend was not e~ident.T`~his relationship persisted in a case-control Cancer Causes and Control. Vol 8. 1997 391 i PE. Tolbert Table 3. Results for stomach cancer from epidemiologic studies of mineral oil-exposed populationsa Author (ref.) Year Location Type of study1 analysis No. of exposed cases Cohort-based studies Tolbert et a/43 Michigan 1992 (USA) SMRd RR Rotimi et a/56 1993 Decoufle5' 1978 Ohio (USA) Michigan (USA) SMR SMR Park et 1994 Jarvholm & ~avenius~~ 1987 Delzell et aI5' 1993 Leon et a/' 1994 Ohio (USA) Sweden SIRe Michigan & SMR Ohio (USA) SMR England SMR (UK) 49 99 21 9 17 25 28 10 15 17 11 6 8 123 51 10 RRb 1.1 1.2 1.3 1.1 1.O 1.2 1.2 0.8 2.5 1.3 1.7 4.4 1.1 1.1 0.9 0.7 (CI)' Study population ~~ (0.8-1.5) Autoworkers ever exposed to straight oils, White (1.0-1.5) (0.8-2.0) (0.5-2.0) (0.6-1.6) (0.7-2.0) (0.7-2.1) (0.4-1.7) (1.4-4.2) Autoworkers ever exposed to soluble oils, White Autoworkers ever exposed to synthetic oils, White Autoworkers ever exposed to straight oils, Black Autoworkers ever exposed to soluble oils, Black Autoworkers exposed to straight oils > 7.5 yrs Autoworkers exposed to soluble oils > 20 yrs Autoworkers exposed to synthetic fluids > 8 yrs Engine plant workers, White NS Ball bearing manufacturing workers, White NS (1.6-9.6) 5 or more yrs of heavy exposure to oil mist, White Automotive stamping plant (0.5-2.1) Grinders and turners at bearing ring manufacturing plant (0.9-1.4) Automotive manufacturing workers, White (0.6-1. l ) Automotive manufacturing workers, Black (0.3-1.3) Unskilled rotary press operators Proportional mortality studies Silverstein Connecticut PMR' 36 2.0 (1.4-2.7) Ball bearing manufacturing workers, White et a / 5 9 1988 (USA) 13 3.4 P < 0.001 Employed in grinding 10 or more yrs Park et a/'' Connecticut PMR 11 2.0 (1.l-3.5) Ball bearing manufacturing workers, White 1988 (USA) MOR^ 8 6.2 P = 0.05 Nested case-control study - those workers ever exposed to soluble oils Mallin et a/" 1986 Illinois (USA) SIR 6 1.9 NS Diesel engine and equipment manufacturing workers, White Lloyd et a / 6 4 USA PMR 11 1.6 NS Newspaper pressmen 1977 Case-control studies/population-based studies Kneller et a/" 1990 China SIR 191 1.4 P < 0.01 Metal grinders, polishers, tool sharpeners, machine-tool operators SIR 193 1.1 NS Toolmakers, metal patternmakers, metal workers Chow et 1994 Sweden 376 1.1 NS Toolmakers and machinists 230 I .o NS Printers Siemiatycki et a/" 1987 Montreal, Canada ORh 24 1.1 90% CI: Ever exposed to cutting oils (0.8-1.4) a Adapted from NIOSH, Draft Criteria for a Recommended Standard: Occupational Exposures to Metalworking Fluids, 1996.' b RR =estimated relative risk. CI = 95% confidence interval. d SMR = standardized mortality ratio. e SIR = standardized incidence ratio. PMR = proportional mortality ratio. MOR = mortality odds ratio. OR =odds ratio. 392 Cancer Causes and Control. Val 8. 1997 Oils and cancer . Table 4. Results for colon cancer from epidemiologic studies of mineral oil-exposed populationsa Author (ref.) Year Location Cohort-based studies Tolbert et a/ 43 1992 Michigan (USA) Eisen et 1992 Dec0ufl6~' 1978 Michigan (USA) Michigan (USA) Type of No. of study1 exposed analysis cases SMR~ RR SMR 59 116 26 8 24 32 11 22 SMR 17 7 Leon et a/' 1994 England (UK) Paganini-Hill et a/63 Los Angeles, 1980 CA (USA) SMR SMR 12 8 R R ~ (CIY Study population 0.8 (0.6-1.O) Autoworkers ever exposed to straight oils, White 0.9 (0.7-1.O) Autoworkers ever exposed to soluble oils, White 0.8 (0.5-1.2) Autoworkers ever exposed to synthetic oils, White 0.6 (0.2-1.1) Autoworkers ever exposed to soluble oils, Black 0.9 (0.6-1.5) Autoworkers exposed to straight oils 7.5 yrs 0.9 (0.4-1.9) Autoworkers exposed to soluble oils > 20 yrs 1.6 (0.8-3.4) Autoworkers exposed to synthetic fluids > 8 yrs 1.5 (0.9-2.2) Autoworkers, Plant 111, White 1.3 NS Workers at metal machining plant w/5+yrs exposure to oil mist, White 1.1 NS 5 or more yrs of heavy exposure to oil mist, White 1.6 (0.8-2.8) Unskilled rotary press operators 0.9 NS Newspaper pressmen (includes rectal cancer) Proportional mortality studies Silverstein et a/59 Connecticut 1988 (USA) Park et also 1988 Vena et a/" 1985 Connecticut (USA) New York (USA) Mallin et a/" 1986 Greene et aIg2 1979 Lloyd et a/64 1977 Illinois (USA) Washington DC (USA) USA PMR' 41 13 PMR . 15 14 8 SIR 10 42 PMR 12 1.4 (1.o-1.9) Ball bearing manufacturing workers, White 1.9 P = 0.02 Employed in grinding 10 or more yrs 1.2 (0.7-1.9) Ball bearing manufacturing workers, White 1.4 NS Engine plant workers, White 1.7 NS Employed in engine plant > 20 yrs 1.2 NS Diesel engine and equipment manufacturing workers, White 1.4 (1.l-1.8) US GPO workers 1 .o NS Newspaper pressmen Case-control studies/population-based studies Siemiatycki et a/" Montreal, 1987 Canada ORg 32 1.0 9O%CI = Ever exposed to cutting oils (0.8-1.4) Gerhardsson de Verdier et a/24 1992 Sweden OR 25 1.5 (0.8-2.8) Ever exposed to cutting fluids Brownson et aIQ3 Missouri 1989 (USA) 18 1.9 (1.O-3.3) Printing machine operators a Adapted from NIOSH, Draft Criteria for a Recommended Standard: Occupational Exposures to Metalworking Fluids, 1996.' RR = estimated relative risk. CI = 95% confidence interval. ' SMR = standardized mortality ratio. PMR = proportional mortality ratio. e SIR = standardized incidence ratio. OR =odds ratio. Cancer Causes and Control. Vol 8. 1997 393 PE. Tolbert Table 5. Results for rectal cancer from epidemiologic studies of mineral oil-exposed populationsa Author (ref.) Year Location Type of No. of study1 exposed analysis cases Cohort-based studies Tolbert et a / 4 3 Michigan 1992 (USA) SMR~ RR Eisen et a/4' 1992 Decoufle5' 1978 Leon et a/' 1994 Michigan (USA) Michigan (USA) England (UK) SMR SMR SMR 37 51 9 21 14 8 7 8 9 RRb 1.5 1.1 0.9 3.2 1.2 0.9 1.7 1.3 1.5 (CI)' (1.0-2.0) (0.8-1.4) (0.4-1.7) (1.6-6.2) (0.5-2.6) (0.4-2.0) (0.7-3.5) NS (0.7-2.9) Study population Autoworkers ever exposed to straight oils, White Autoworkers ever exposed to soluble oils, White Autoworkers ever exposed to synthetic oils, White Autoworkers exposed to straight oils > 7.5 yrs Autoworkers exposed to soluble oils > 20 yrs Autoworkers exposed to synthetic fluids > 8 yrs Autoworkers, Plant ill, White Workers at metal machining plant w/5+yrs exposure to oil mist, White Unskilled rotary press operators Proportional mortality studies Silverstein et Connecticut 1988 Park et a/" (USA) Connecticut 1988 (USA) Lloyd et a / 6 4 USA 1977 PMR' PMR PMR 14 11 4 7 1.4 (0.8-2.3) Ball bearing manufacturing workers, White 3.1 (1.5-5.5) Ball bearing plant workers, White 2.8 P e 0.05 Employed in engine plant > 20 years 1.6 NS Newspaper pressmen Case-control studies/population-based studies Gerhardsson de Sweden Verdier et a / 2 4 OR' 1992 Siemiatycki et a/'' Montreal, OR 1987 Canada 25 13 2.0 (1.0-4.2) Ever exposed to cutting oils 0.7 9O%CI = Ever exposed to cutting oils (0.4-1.0) ' a Adapted from NIOSH, Draft Criteria for a Recommended Standard: Occupational Exposures to Metalworking Fluids, 1996.' RR = estimated relative risk. CI = 95% confidence interval. SMR = standardized mortality ratio. e PMR =proportional mortality ratio. OR = odds ratio. analysis nested in this cohort that controlled for other exposures, and the association was apparent for grinding with either soluble or synthetic fluids." An association of rectal cancer with straight oils also was observed in the cohort, with a monotonically increasing dose-response trend and a statistically significant rate ratio of 3.0 for those who had held jobs entailing straight oil exposure for more than 7.5 There was a mild, nonsignifi- cant elevation in stomach cancer risk among those who had ever worked in grinding, but no dose-response trend was observed. For pancreatic cancer, Black workers exposed to soluble oils at Plants I and I1 were observed to have an SMR of 1.6 (95 percent confidence interval [CI] = 1.0-2.5). In a subsequent case-control analysis of pancreatic cancer nested in the c ~ h o r t ,a' ~n O R of 3.1 (CI = 1.3-7.6) was observed for those exposed to synthetic fluids in grinding operations for more than 1.4 mg/m'- years. A very mild but stable association of exposure to straight oils with risk of prostate cancer was noted in the cohort analysis.'] Lastly, a twofold excess of larynx cancer 394 Cancer Causes and Control. Vol 8. 1997 mortality was observed among those exposed to straight oils." This finding was pursued in a case-control study nested qfthin this cohort which, unlike the nested analyses mentioned above, entailed addition of incident cases ascertained through cancer registries and collection of information on possible non-machining fluid confounding exposures (such as solvents,acid mists, and asbestos).'* The association of straight oil exposure with larynx cancer risk persisted; a dose-response trend was fioted, with a statistically significant twofold excess among those exposed to greater than 0.5 mg/m'-years. Most noteworthy negative findingsof this cohort study were for bladder cancer and for lung cancer. With respect to bladder cancer, the study yielded no evidence of an association of bladder cancer risk with any type of machining fluid. This is of interest in light of the generally positive evidence from case-control studies, mentioned above. Regarding lung cancer, internal mortality analyses of this cohort revealed no dose-response trend for the relationship of lung cancer and straight oil exposure, and Oils and cancer Table 6. Results for pancreatic cancer from epidemiologic studies of mineral oil-exposed populationsa Author (ref.) Year Location Cohort-based studies Tolbert et a / 4 3 Michigan 1992 (USA) Eisen e t aI4' 1992 Bardin et a/ 47,e 1997 Michigan (USA) Michigan (USA) Type of study1 analysis No. of R R ~ exposed cases ~- (CIT or Pvalue SMR~ RR SMR 34 61 19 8 19 15 25 9 8 0.8 (0.6-1. l ) 0.8 (0.6-1.O) 1.o (0.6-1.6) 1.4 (0.6-2.8) 1.6 (1.O-2.5) 0.9 (0.5-1.6) I .4 (0.5-3.7) 2.0 (0.9-4.7) 0.9 (0.4-1.7) OR' 18 0.9 (0.5-1.6) 19 1.4 (0.7-2.7) 9 3.0 (1.2-7.5) Rotimi et a / 5 6 1993 Decoufl&' 1978 Ohio (USA) SMR Michigan (USA) SMR Leon et a/' England (UK) SMR 1994 Paganini-Hill et als3 Log Angeles, SMR 1980 CA (USA) Delzell et .Michigan & SMR ? 993 Ohio (USA) SMR 8 0.9 (0.4-1.8) 7 3.0 (1.2-6.2) 8 1.1 NS I1 0.3 NS 6 1.I (0.4-2.5) 5 1.3 NS 145 0.9 (0.8-1. I ) 62 1.1 (0.8-1.4) Study population Autoworkers ever exposed to straight oils, White Autoworkers ever exposed to soluble oils, White Autoworkers ever exposed to synthetic oils, White Autoworkers ever exposed to straight oils, Black Autoworkers ever exposed to soluble oils, Black Autoworkers exposed to straight oils > 7.5 yrs Autoworkers exposed to soluble oils > 20 yrs Autoworkers exposed to synthetic fluids 8 yrs Autoworkers, Plant 111, White Autoworkers, exposed to straight oiis > 2.1 mg/m3-yrs Autoworkers, exposed to grinding with soluble fluids Autoworkers, exposed to grinding with synthetic fluids > 1.4 mg/m3-yrs Engine plant workers, White Engine plant workers, Black Workers at metal machining plant w/5+yrs exposure to oil mist, White 5 or more yrs of heavy oil mist exposure, White Unskilled rotary press operators Newspaper web pressmen Automotive manufacturing workers, White Automotive manufacturing workers, Black Proportional mortality studies Silverstein et a / 5 9 Connecticut 1988 (USA) Park et a/" 1988 Vena et also 1985 Connecticut (USA) New York (USA) PMR~ MOR^ MOR PMR Mallin et a/" 1986 Illinois (USA) SIR Lloyd et a/64 1977 USA PMR 24 9 5 8 11 7 5 55 9 1.4 (1.o-2.1) Ball bearing manufacturing workers, White 3.1 P = 0.05 Employed in grinding 10 or more yrs 3.7 P c 0.05 Employed in machinery 10 or more yrs 1.I (0.6-2.2) Ball bearing plant workers, White 1.9 P c 0.05 Engine plant workers, White 2.3 P < 0.05 Employed in engine plant > 20 years 1.2 NS Diesel engine and equipment manufactbring workers, White 3.6 P e 0.05 Diesel engine and equipment manufacturing workers, Black 1.2 NS Newspaper pressmen Case-control studieslpopulation-based studies Mack and Los Angeles, Incidence 21 1.3 NA Machinists, White males Paganini-HillZ5 CA (USA) 1981 - a Adapted from NIOSH, Draft Criteria for a Recommended Standard: Occupational Exposures to Metalworking Fluids, 1996.' RR = estimated relative risk. CI = 95% confidence interval. SMR = standardized mortality ratio. e Nested case-control analysis of populations studied in Tolbert et a/43and Eisen et a/4* OR =odds ratio. PMR = proportional mortality ratio. OR = mortality odds ratio. Cancer Causes and Control. Vol 8. 1997 395 RE. Tolbert Table 7. Results for laryngeal cancer from epidemiologic studies of mineral oil-exposed populationsa Author (ref.) Year Location Type of NO. of study1 exposed analysis cases R R ~ (cI)' Study population Cohort-based studies Tolbert et a / 4 3 Michigan SMRd 23 2.0 (1.3-3.0) Autoworkers ever exposed to straight oils, White 1992 (USA) 30 1.4 (1.o-2.0) Autoworkers ever exposed to soluble oils, White 8 1.6 (0.7-3.1) Autoworkers ever exposed to synthetic oils, White 6 1.5 (0.5-3.2) Autoworkers ever exposed to soluble oils, Black RR 11 2.0 (0.9-4.8) Autoworkers exposed to straight oils > 7.5 yrs 6 1.2 (0.4-3.6) Autoworkers exposed to soluble oils > 20 yrs 6 0.6 (0.2-1.8) Autoworkers exposed to synthetic fluids > 8 yrs Eisen et a/42 1992 Delzell et a/57 Michigan (USA) Michigan 8 OR' SMR 28 2.2 (1.3-4.0) Autoworkers exposed to straight oil > 0.5 mg/m3-yrs 38 0.8 (0.6-1. I ) Automotive manufacturing workers, White 1993 Ohio (USA) 17 0.7 (0.4-1. I ) Automotive manufacturing workers, Black Proportional mortality studies Mallin et a/" Illinois (USA) PMR' 1986 7 1.7 NS Diesel engine and equipment manufacturing workers, White Case-control studies/population-based studies Zagraniski et a/26 Connecticut OR 1986 (USA) Wortley et Washington OR 1992 state (USA) Haguenoer et a/ 95 1990 Brown et aIg6 1988 Ahrens et 1981 France OR Texas (USA) OR Germany OR 22 17 NA 10 7 5 NA 2.5 (1.2-5.2) Ever worked as a machinist 2.1 (1.0-4.1) Ever worked as a metal grinder 1.8 (0.5-6.2) Ever employed as grinding, abrading, or buffing operator 1.O (0.5-1.9) Ever employed in precision metal working 1.8 NS Employed in metal work or as mechanic for at least 15 years 0.5 (0.2-1.6) , Ever machinists 2.2 (0.9-5.3) Ever mineral oil exposure a Adapted from NIOSH, Draft Criteria for a Recommended Standard: Occupational Exposures to Metalworking Fluids, 1996.7 b RR = estimated relative risk. CI = 95% confidence interval. SMR = standardized mortality ratio. e OR = odds ratio. PMR = proportional mortality ratio. an inverse trend for both soluble and synthetic fluid expo~ure.A~'nested case-control analysis of lung cancer found that only the inverse association with synthetic fluids persisted in models controlling for multiple expos u r e ~ . ~It' has been hypothesized that endotoxins may inhibit the growth of malignant cells in the lung by stimulating macrophage activity,by inducing interferon release, by mitogen activity,or by inducing tumor necrosis factor and, thus, endotoxins commonly found in synthetic fluids are a possible basis for the observed inverse association with synthetic f l u i d ~ . * ~ . ' ~ . ~ ~ Among other major cohort studies of machining-fluid exposed populations is a study by De~ouflC.A~' standardized mortality analysis was conducted of a cohort of 2,485 White men who had worked for at least five years between 1938 and 1967 in jobs entailing machining fluid exposure at a large US engine manufacturing plant. Among those with 20 years of follow-up, a statistically 396 Cancer Causes and Control. Vol 8. 1997 significant twofold excess of cancers of the stomach and large intestine combined was noted, with 15 deaths cf7.6 expected. (Presumably, results for the two sites were combined due to small numbers of cases of each disease.) Various classes of metalworking fluids were used at the facility studied and no attempt was made to distinguish the different exposures. This study did not support a positive association of machining fluid exposure with lung cancer. Jarvholm et a1 52 reported results of a cancer morbidity and mortality study of a cohort of 792 men with at least five years of machining fluid exposure in a Swedish bearing-ring manufacturing plant. Among grinders, a statistically significant twofold excess of digestive cancers was observed (15 cf 7.8 expected), after 20 years of fol- low-up; six stomach cancers were observed Q 2 . 6 expected. Due to the higher temperatures generated by grinding operations, soluble and synthetic fluids are typi- Oils and cancer ' Table 8. Results for lunghespiratory system cancer from epidemiologic studies of mineral oil-exposed populationsa Author (ref.) Year Location Cohort-based studies Tolbert et a / 4 3 Michigan 1992 (USA) Eisen et Michigan 1992 (USA) Schroeder et a/498' Michigan 1997 (USA) Type of No. of study1 exposed analysis cases RRb SMRd RR SMR OR' 251 1 .o 478 1.1 116 1 .o 35 1.1 64 0.9 91 0.8 115 0.6 30 0.6 60 0.9 88 1 .o 122 0.9 28 0.5 Rotimi et a / 5 6 1993 Acquavella ef i993 Ohio (USA) SMR Iowa (USA) SMR 81 1.2 23 1.4 31 1.8 25 2.5 Jarvholm & ~avenius~~ 1987 Sweden SIRg Park et aI5* 1994 Delzell et a/57 1993 Ohio (USA) SMR Michigan & SMR Ohio (USA) SMR Lynge et a / 6 2 1995 Denmark SIR Leon et a/' 1994 England (UK) SMR Paganini-Hill et a/63 Los Angeles, SMR 1980 CA (USA) Menck & Hendersong7 1976 Los Angeles, SMR CA (USA) Proportional mortality studies Silverstein et Connecticut 1988 (USA) PMR~ Park et a/" 1988 Vena et a/60 1985 Mallin et a/" 1986 Connecticut PMR (USA) MOR' New York PMR (USA) Illinois (USA) PMR 5 0.4 15 1,338 435 26 94 22 10 30 1.3 1.1 1 .o 2.0 1.8 1.5 2.8 1 .o 23 0.9 13 0.6 59 1.2 5 19.3 48 1.3 29 1.4 34 1.3 12 1.3 Lloyd et a/64 1977 USA PMR 41 1 .I (Cl)' Study population (0.9-1.2) Autoworkers ever exposed to straight oils, White (1.o-1.2) Autoworkers ever exposed to soluble oils, White (0.8-1.2) Autoworkers ever exposed to synthetic oils, White (0.7-1.5) Autoworkers ever exposed to straight oils, Black (0.7-1.2) Autoworkers ever exposed to soluble oils, Black (0.6-1.O) Autoworkers exposed to straight oils > 7.5 yrs (0.5-0.9) Autoworkers exposed to soluble oils > 20 yrs (0.4-0.8) Autoworkers exposed to synthetic fluids > 8 yrs (0.7-12 ) Autoworkers, Plant 111, White (0.7-1.3) Autoworkers, exposed to straight oils > 4.88 mg/m3-yrs (0.7-12 ) Autoworkers, exposed to grinding with soluble fluids 7 15.6 mg/m3-yrs (0.3-0.8) Autoworkers, exposed to grinding with synthetic fluids 7 1.94 mg/m3-yrs (1.o-1.5) Engine plant workers, White (0.9-2.0) Engine plant workers, Black (12-2.6) Exposed to cutting oils and metal dusts at metalworking plant (1.6-3.7) First exposed to cutting oils and metal dusts between 1950-1959 (0.1-0.9) Grinders and turners at bearing ring plant (0.7-2.2) Automotive stamping plant (1.o-1.l) Automotive manufacturing workers, White (0.9-1.l) Automotive manufacturing workers, Black (1.3-3.0) Factory workers in newspaper and magazine production (1.4-2.2) Unskilled rotary press operators (0.9-2.2) Newspaper pressmen NS Pressmen NS Newspaper printing (0.8-1.l) NS (1.O-1.6) P = 0.008 NS NS NS NS NS Ball bearing plant workers, Whitemoth primary and secondary lung cancer Employed in grinding 10 or more yrs Ball bearing plant workers, White menlboth primary and secondary lung cancer Females ever employed in grinding Engine plant workers, White Employed in engine plant > 20 yrs Diesel engine and equipment manufacturing workers, White Diesel engine and equipment manufacturing workers, Black Newspaper pressmen Continued Cancer Causes and Control. Vol 8. 1997 397 l? E. Tolbert Table 8. Continued Author (ref.) Year Location Type of No. of study1 exposed analysis cases RRb (Cl)' Study population Case-control studies/population-basedstudies Siemiatycki et a/" Montreal, OR 1987 Canada Jockel et a/" Germany OR 1992 Coggon eta/" 1984 England, UK OR 23 NA 113 26 28 9 Malker & Gemne" Sweden 1987 SIR 149 1.5 9O%CI = Ever exposed to cutting oils/oat cell cancer of (1. 0-2.1) the lung 2.2 (1.1-4.8) Employed 6 months or more as a turner, grinder, driller or cutter 1.4 (1.1-1.8) Ever had an occupation with potential metalworking fluids exposure 1.O (0.6-1.6) Ever had an occupation with potential high metalworking fluids exposure 1.6 (0.9-2.7) Ever had occupation with potential printing ink exposure 2.0 (0.8-5.0) Ever had occupation with potential high printing ink exposure 1.6 (1.3-2.9) Workers in newspaper, journal, bookprinting enterprises a Adapted from NIOSH, Draft Criteria for a Recommended Standard: Occupational Exposures to Metalworking Fluids, 1996.' RR = estimated relative risk. CI = 95% confidence interval. SMR = standardized mortality ratio. e Nested case-control analysis of populations studied in Tolbert et a / 4 3and Eisen et a/.42 OR = odds ratio. SIR = standardized incidence ratio. 'PMR = proportional mortality ratio. MOR = mortality odds ratio. cally used, not straight oils. Grinders also are exposed to abrasives, hypothesized to be co-carcinogens, potentiat- ing the carcinogenic effect of other Extension of follow-up an additional seven years in a later report54did not strengthen the finding of an association of digestive cancer with grinding exposures. Seven scrotal cancers were noted among workers at the plant, whereas less than one was expected." Acquavella et aZ55reported results of a cohort study of 3,630 employees who had worked at least six months at a metal components manufacturing facility. SMRs were generated for subgroups of the workforce categorized according to exposure to specific agents. Among workers exposed to metalworking fluids and metal dusts, the only cancer with sufficient numbers of deaths to be informative was lung cancer, for which an SMR of 1.8 (CI = 1.2-2.6) was observed. The excess of lung cancer was restricted to workers first employed between 1950 and 1959 (SMR = 2.5, CI = 1.6-3.7). No trend in risk with duration of exposure was observed. Rotimi e t a l 56 studied 21,013 foundry and engine plant workers. For the 6,511 men employed only at the engine plant, an excess of stomach cancer was noted among Whites (SMR = 2.5, CI = 1.4-4.2), of pancreatic cancer among Blacks (SMR = 3.0, CI = 1.2-6.2), and of lung can- cer in both races (SMR = 1.2, CI = 1.0-1.5). Nested case-control studies of stomach and lung cancer were undertaken subsequently and results are not yet published. Delzell et al 57rep~rtethde results of the largest stand- ardized mortality study of automotive manufacturing 398 Cancer Causes and Control. Vol 8. 1997 workers (n = 123,232 males) to date. Hourly workers employed at an automotive facility in 1973were followed- up through 1985. The SMR analyses are null for most cancers,with the exception of a slight excess of lung cancer in White males (SMR = 1.1, CI = 1.0-1.1). This analysis was limited to an external comparison with the US popu- lation, and no subgrouping of the workforce according to exposure to metalworking fluids was performed. Nested case-control studies are in progress. Two proportional mortality studies of workers in- volved In ball bearing manufacture in Connecticut provided evidence of an association of gastrointestinal cancers with machining fluid exposure. In one of these studies:' analysesof 702 deaths occurring among employ ~ ees who had worked for at least 10 years at a ball bearing plant yielded statistically significant excesses of stomach cancer (twofold) and rectal cancer (threefold). As in the report by Jarvholm et a1 '' above, the stomach cancer excess appeared to be specific to exposure to water-based fluids in grinding operations. The second included 1,766deaths among workers at another ball bearing plant. Excesses of each major digestive cancer were observed, with proportional mortality ratios for cancer of the esophagus, stomach, and large intestine statistically sig- nificant. Again, the excess of stomach cancer appeared to be specific to grinding work; there was a statistically significant threefold excess of stomach cancer among those with more than 10 years' exposure to grinding. Because an elevation of stomach cancer risk was not ob- served among those exposed to straight oils only, the * 1 1.. Oils and cancer a' ' Table 9. Results for bladder and urinary organ cancer from epidemiologic studies of mineral oil-exposed populationsa Author (ref.) Location ~ y p eof NO. of R R ~ (cI)' Study population fl Year study1 exposed _____ c> Cohort-based studies analysis cases -__ ___ Dec0ufl8~' Michigan SMRd 6 1.2 NS Workers at metal machining plant w/5+yrs . .I * 1978 (USA) exposure to oil mist, White Jarvholm & Sweden SIRe 7 1.0 (0.4-2.2) Grinders and turners at bearing ring plant ~avenius~~ c3 1987 Leon et a/' England, UK SMR 6 1.2 (0.5-2.7) Unskilled rotary press operators 1994 Coggon et a/'' England, UK SMR 5 5.0 (1.O-25.8) Men w/high exposure to printing inks 1984 Delzell et a / Michigan & SMR 63 1.O (0.7-1.2) Automotive manufacturing workers, White 1993 Ohio (USA) 12 0.8 (0.4-1.4) Automotive manufacturing workers, Black Proportional mortality studies Silverstein et Connecticut PMR' 1988 (USA) Vena et New York 1985 Greene et a/" 1979 (USA) USA PMR 2010th et a/'' 1986 USA PMR Case-control studies/population-based studies Silverman et a/3s USA ORg 1989 Siemiatycki et a/22 1987 Claude ef a/1oo 1988 Gonzalez et a/ lo' 1989 Montreal, Canada Germany Spain Steenlandlo2 Ohio (USA) 1987 Vineis & Magnani3' Italy 1985 Schifflers et a/ lo3Belgium 1987 Howe and Lindsay'04 Canada 1980 Coggon et a / * England, UK 1984 OR OR OR OR OR OR OR OR Malker et a/'* 1987 Silverman et a/36 1983 Brownson et a / 9 3 1989 Cart~right~~ 1982 Sweden Detroit, MI (USA) Missouri (USA) England, UK 14 7 17 11 102 51 26 37 47 18 43 21 NA 11 45 16 34 8 NA 52 21 322 137 32 7 18 1.3 (0.8-2.2) Ball bearing plant workers, White 7.3 P < 0.05 Based on US. mortality, White 1.4 (0.9-2.2) Men in general printing 1.1 (0.6-2.0) Men in general printing 1.3 (1.0-1.7) Ever machinist 6 months or more 1.4 (0.9-2.1) Ever drill press operator 6 months or more 1.I (0.6-1.9) Ever metal machinery workers 0.8 (0.5-1.2) Ever printers 1.2 9O%CI= Ever exposed to cutting oils (1.O-1.6) 2.3 (1.O-5.6) Ever turner 0.8 (0.5-1.3) Ever metal worker 0.8 (0.5-1.1) Ever toolmaker 6 months or more 1.9 (1.2-2.8) Ever machinery adjuster, assembler or mechanic 6 months or more 2.0 NS Ever grinding machine operator 0.7 P c 0.05 Ever machinist 1.5 (0.7-3.3) Ever employed in machine tools 6 months or more 2.5 (1.3-4.7) All metal workers 2.6 (0.9-7.2) Turners 2.7 (1.1-7.7) Ever metal machinist 1.3 (0.9-1.9) Ever had an occupation with potential cutting oil exposure 1.5 (0.8-2.8) Ever had an occupation with potential high cutting oil exposure 1.2 P c 0.01 Toolmaker or machinist in 1960 1.1 (0.8-1.5) All metal machinists 1.5 (0.9-2.7) Tool & die workers 3.1 (1.1-8.9) Printing machine operators 3.1 (1.4-6.8) Workers exposed to ink from high-speed presses a Adapted from NIOSH, Draft Criteria for a Recommended Standard: Occupational Exposures to Metalworking Fluids, 1996.' RR = estimated relative risk. CI = 95% confidence interval. SMR = standardized mortality ratio. PMR = proportional mortality ratio. 'e SIR = standardized incidence ratio. OR = odds ratio. Cancer Causes and Control. Vol 8. 1997 399 I!E. Tolbert Table 10. Results for prostate cancer from epidemiologic studies of mineral oil-exposed populationsa Author (ref.) Year Location Type of No. of study/ exposed analysis cases RRb (CI)" Study population - Cohort-based studies Tolbert et a/43 Michigan 1992 (USA) SMRd RR Rotimi et a / 5 6 1993 Decoufle5' 1978 Jarvholm & ~avenius~~ 1987 Delzell ef 1993 Ohio (USA) SMR Michigan (USA) Sweden SMR SIRe Michigan & SMR Ohio (USA) SMR 72 125 26 12 23 40 54 8 8 5 6 6 NA 143 89 1.2 (0.9-1.5) Autoworkers ever exposed to straight oils, White 1.1 (0.9-1.3) Autoworkers ever exposed to soluble oils, White 1.1 (0.7-1.6) Autoworkers ever exposed to synthetic oils, White 1 .o (0.5-1.7) Autoworkers ever exposed to soluble oils, Black 1 .o (0.6-15 ) Autoworkers ever exposed to soluble oils, Black 1.5 (1.O-2.3) Autoworkers exposed to straight oils > 7.5 yrs 1 .o (0.5-1.8) Autoworkers exposed to soluble oils > 20 yrs 1.6 (0.7-3.5) Autoworkers exposed to synthetic fluids > 8 yrs 0.9 (0.4-1.7) Engine plant workers, White 1.3 (0.4-3.0) Engine plant workers, Black 0.6 NS Workers at metal machining plant w/5+yrs exposure to oil mist, White 0.3 (0.1-0.7) Grinders and turners at bearing ring plant 0.4 (0.1-0.8) Grinders and turners, at last 20 years since onset of exposure 1 .o (0.9-1.2) Automotive manufacturing workers, White 1.1 (0.9-1.3) Automotive manufacturing workers, Black Proportional mortality studies Silverstein et Connecticut 1988 Park et also (USA) Connecticut 1988 (USA) Vena et a / 6 0 New York 1985 (USA) PMR' PMR Mallin et a/ 1986 Illinois (USA) SIR 29 1 .o (0.7-1.4) Ball bearing plant workers, White 10 0.9 (0.5-1.7) Ball bearing plant workers, White 7 1.1 NS Engine plant workers, White 7 0.7 NS 10 1.5 NS Diesel engine and equipment manufacturing workers, White Case-control studies/population-based studies Siemiatycki et a/= Montreal, 1987 Canada ORg 47 1.2 90%CI = Ever exposed to cutting oils (1.0-1.6) a Adapted from NIOSH, Draft Criteria for a Recommended Standard: Occupational Exposures to Metalworking Fluids, 1996.' RR = estimated relative risk. CI = 95% confidence interval. ' SMR = standardized mortality ratio. PMR = proportional mortality ratio. e SIR = standardized incidence ratio. OR = odds ratio. excess among grinders was interpreted as implicating exposure to water-based fluids. In addition, an association of exposure to straight oils and pancreatic cancer was observed in this study. For those with 25 years' work in a straight oil-exposed job, there were three- to tenfold excesses of pancreatic cancer. Most of the cohort and proportional mortality studies done to date were too small to detect a moderate excess of bladder cancer, other than the two large cohort studies of auto worker^,"^^^ which had stable numbers of bladder cancer deaths, and in which no association of bladder cancer with machining fluid exposure was observed. One proportional mortality study observed"an excess of blad- 400 Cancer Causes and Control. Vola. 1997 der cancer in engine plant workers, particularly among those employed before 1950, when straight and soluble oils were used at the plant. Printing pressmen Because the purpose of this review is to attempt to discern carcinogenic effects of mineral oil exposure, in describing the epidemiologic literature for the printing professions studies of newspaper press operators will receive greater emphasis since this group has less heterogeneous exposures than other types of printing which have concomitant exposure to pigments containing benzidines, cadmium, and lead. 4 A Oils and mncer Findings from one of the largest cohort studies of printing workers undertaken to date were recently reported.6' The study followed 9,500 men who were members of one of two printing trade unions in the Manchester (UK) area between 1949 and 1963. One of the unions included primarily newspaper production workers, while the other comprised primarily non-newspaper printing professions. There were 3,482 deaths occurring through the end of follow-up (1983). Compared with the population of England and Wales, men in newspaper letterpress printing had elevated rates of lung cancer (SMR = 1.8, CI = 1.4-2.2). When Manchester area rates were used as the referent, the SMR diminished to 1.2 and was of borderline statistical significance (CI = 1.0-13).This finding was followed-up by a case-control study nested within this cohort,' which found that duration of work in newspaper machine rooms was related to lung cancer risk. This finding is unlikely to reflect confounding by smoking because it involved an internal analysis. A cohort of 15,534 men and 3,593 women in Denmark known to be worhng in the printing industry based on a 1979 census was followed-up through 1987.6' Their cancer incidence was compared with that of individuals in the 1970 census employed in other occupations. A lung cancer excess was observed among male factory workers in newspaper and magazine production (SIR = 2.0, CI = 1.3-3.0), while excesses of bladder, renal, and liver cancer were noted among printers involved in other types of printing. The authors point out that many factory workers in newspaper printing in Denmark operated rotary letterpresses and conclude that their findings support the above-described findings of Leon et al' and Leon.61 Paganini-Hill et a1" followed-up through 1978 a cohort of 1,361 newspaper web pressmen who were members of a pressmen's union for a year or more between 1949 and 1965. Mortality was compared with the general US population rates. For most cancers, power was low. Significant excesses of renal cancer and lymphatic and hematopoietic cancers other than leukemia were noted, based on five observed cases for each disease group. The SMR for lung cancer was 1.5 and of borderline significance. Lloyd et a1 conducted a proportional mortality study of 2,604 deaths among members of a printing pressmen's union, categorized by commercial and newspaper printing. For newspaper pressmen, a significantly high proportional mortality ratio (PMR) due to cancer of the buccal cavity and pharynx was observed, using US rates for comparison. Slight, nonsignificant excesses of cancers of the esophagus, stomach, rectum, larynx, and lung, and of leukemia were also evident. Laboratory/mechanistic studies Regarding animal data, the IARC' review concluded that the data are inadequate to evaluate the carcinogenicity to experimental animals of formulated mineral oil products and used products since "the possible carcinogenic activity of individual products is dependent on the severity of processing of the base oils and the nature and concentration of additives and contaminants," and for used products, "the conditions of use." The IARC panel found sufficient evidence for carcinogenicity in experimental animalsof acid-treated oils, mildly hydro-treated oils, and mildly solvent-refined oils. Data on severely hydrotreated oils were determined to be inadequate to permit evaluation of their carcinogenicity to experimental animals and no evidence was found that severely solvent-refined oils are carcinogenic to experimental animals. All of the above types of base oils have been used in mineral oil formulations such as metalworking fluids, but current formulations generally are composed of the more severely refined oils. Known or suspected carcinogens in mineral oil formulations include: PAHs; long-chain aliphatics; nitrosamines; sulfur; chlorinated paraffins, naphthenics and aromatics; formaldehyde; and N-phenyl-2-naphthylamine. PAHs PAHs such as benz[a]pyrene are present to varying degrees in crude oil depending on the geographic source and removed to varying degrees by the refining process. Many PAHs are known carcinogens (e.g., IARC'). They are thought to be largely responsible for the carcinogenicity of mineral oils. Severe solvent refining is very effective in removing PAHs from oil. Whereas acid-refined mineral oils have been found to be highly carcinogenic in animal tests, mildly solvent-refined oils have shown lower carcinogenic activity and severely solvent-refined oils have tested negative in most ~ t u d i e s " "(e~~ception:~~). Generally in the 1970s, severely solvent-refined oils replaced acid-refined oils in metalworking fluids.' In the case of metalworking fluids, PAHs may be reintroduced to varying degrees during the course of use as a result of pyrolytic reaction^."^'^' O n the other hand, print press inks contain carbon black which has been shown to adsorb benz[a]pyrene and other PAHs to its surface. Mean content of benz[a]pyrene in a random sample of newspaper inks manufactured in England was reported to be 8.3 pg/g." The bioavailability of the PAHs adsorbed onto carbon black is unclear, but it has been suggested that the oils in printing inks may elute the PAHs from the carbon black, increasing bioavailability of the carbon black PAHs in ink formulation^.^^ While an IARC expert working group concluded that there was inadequate evidence to evaluate the carcinogenicity to humans of carbon black, the panel found that there was sufficient evidencethat solvent extracts of carbon black are carcinogenic to laboratory animals, Cancer Causes and Control. Vol 8. 1997 401 RE. Tolbert presumably due to PAHs.' Carbon black Long-chain aliphatics Carbon black, added to printing inks, is not suspect itself but has benz[a]pyrene and other PAHs adsorbed to its There is some laboratory evidence that aliphatic hydro- surface (see PAHs, above). carbons 12 to 14 carbons in length, common in mineral oils, are co-car~inogenic.~~~~~ Conclusion Nitrosamines There is clear evidence from studies of metalworkers, The co-occurrence of nitrites or nitrates and amines in cotton spinners, and jute processors that early formula- synthetic and soluble fluids can lead to the formation of tions of straight mineral oils were carcinogenicto the skin. nitrosamines. N-nitrosoethanolamine, detected in the The association is strong, has been observed consistently water-based metalworking fluids, is a potent ~arcinogen.'~ in a large number of studies, and is biologically plausible, In 1976, the National Institute for Occupational Safety in part because the tumors appear on areas that have been and Health in the US issued an alert regarding the possible in direct contact with the fluids. health threat posed by the presence of nitrosamines in Regarding larynx and rectal cancer, several studies have metalworking fluids." The use of nitrites in fluid formu- noted associations of these cancers with metalworking in lations has decreased since that time,' although general, and the largest cohort study that included an nitrosamines continue to be detected at low concentra- tions in formulations without added nitrite^,^**^^ and high levels occasionally are found.80,8' exposure assessment observed a dose-response trend for each specifically with straight oils. An association with larynx cancer has strong biological plausibility due to the Sulfur similarity of the squamous epithelium of this tissue with skin and the fact that this organ is directly exposed to Elemental sulfur and some sulfur compounds, commonly aerosolized droplets. added to metalworking fluids, have been found to accel- Evidence regarding lung cancer is generally negative erate tumorigenesis in laboratory ~ t u d i e s . ~ ~ * ~ * from studies of metalworkers but positive from studies Chlorinated paraffins, naphthenics, and aromatics of printing pressmen. Several cohort studies, including the largest cohort study of newspaper printers, as well as Chlorination of base oils leads to the formation of chlorinated paraffin, naphthenic and aromatic compounds, some of which have been found to be carcinogenic in animal studies (e.g., National Toxicology Program"). During use, these chlorinated compounds may react to form other potentially carcinogenic chlorinated hydrocarbons, such as dioxins. one case-control study, reported a statistically significant associationwith printing ink. A case-control study nested in the largestcohort of printers found a duration-response trend. The primary constituent that distinguishes newspaper printing ink from metalworking fluids is carbon black, which is known t o be contaminated with benz[u]pyrene and other PAHs, plausible etiologic agents for any association of printing ink with lung cancer. Formaldehyde There is suggestive evidence from several studies that Some of the biocides used to control microbial growth in mineral oil formulations act by releasingformaldehyde. `7.Formaldehyde is a suspect carcinogen (e.g., Blair et a1 workers in grinding operations may be at excess risk of cancer of the esophagus, stomach, and pancreas. Exposure-response relationships generally have not been observed in these studies. If there is indeed an association N-phenyl-2-naphthylamine N-phenyl-2-naphthylamine,a suspected bladder carcinogen, is listed as a component of somefluid formulation^.^^^^' of grinding with these cancers, it may not reflect mineral oil exposures, since both soluble and synthetic fluids (which contain no mineral oil) are used in grinding. Possible suspect agents within water-based fluids include Metals nitrosamines, detected in both types of water-based fluids and known to be carcinogenicto the stomach, and PAHs Metal particles are generated in the course of metalwork- which may be higher in fluids used for grinding as a result ing and can dissolve in metalworking fluids and of pyrolysis occurring at the high temperatures generated concentrate with use.s6The type of metal depends, of in grindingoperations.Other exposuresunique to grinding course, on the composition of the metal being worked work, such as abrasives, also must be considered possible and the tool. These may include nickel, chromium, lead, suspect agents. cobalt, or molybdenum, some of which are known car- A number of case-control studies of bladder cancer cinogen~.~' have noted a positive associationwith work as a machinist, 402 Cancer Causes and Control. Vol 8. 1997 ~ Oils and cancer but the cohort studies to date offer little corroboration. This observation is perplexing in light of the fact that the case-control studies are likely to entail more heterogeneous exposure than the cohort studies. There is limited evidence that metalworking fluids are associated with cancer of the colon, prostate, and si- nonasal region. Neither strong internal consistency (e.g., an exposure-response trend) nor consistency across stud- ies has been observed for these cancers. I t should be kept in mind, however, that given the generally unstable numbers of exposed cases in most of the studies to date, and the variability of exposure within and across work set- tings, inconsistency across studies would be likely to occur even if there were indeed a causal association. Data from the US National Occupational Exposure Survey 1981-83 indicate that approximately 1.6 million American workers are exposed occupationally to mineral oils; internationally, the number exposed is many millions.88 Thus, the potential public health impact of any carcinogenic activity of mineral oil formulations is sub- stantial. Whether or not mineral oil formulations in current use are carcinogenic is not known. T h e levels of two of the principal suspect carcinogens in mineral oil formulations -PAHs and nitrosamines -have been reduced over the last 20 years. Severe solvent refining of mineral oil has decreased the content of P A H s in straight and soluble oils, and the reduction in use of nitrite additives has led t o less nitrosamine formation in soluble and synthetic fluids. Data are not yet available to assess possible carcinogenicity of c u r r e n t f o r m u l a t i o n s allowing for sufficient latency. T h e composition of fluids in use 20 to 40 years ago, however, is relevant to cancer risk today. Thus, any health risks associated with the formulations containing higher levels of P A H s and nitrosamines will continue to be evident for some time. References 1. Bingham E, Trosset R, Warshawsky D. Carcinogenic potential of petroleum hydrocarbons: a critical review of the literature. J Environ Pathol Toxic01 1980; 3: 484-563. 2. CONCAWE. Health Aspects of Lubricants. The Hague, The Netherlands: The Oil Companies International Study Group for Conservation of Clean Air and Water - Europe, 1983; Report No. 1/83. 3. International Agency for Research on Cancer. Polynuclear Aromatic Hydrocarbons, Part 11: Carbon Blacks, Mineral Oils (Lubricant-based Oils and Derived Products), and SomeNitroarenes. Lyon,France: IARC, 1984;IARCMonogr Eva1 Carcinog Risks Humans, Vol. 33: 87-168. 4. Hallock MF, Smith TJ, Woskie SR, Hammond SK. Estimation of historical exposures to machining fluids in the automotive industry. A m J Ind Med 1994; 26: 621-34. 5. Kriebel D, Eberiel D, Eisen EA, et al. Field investigations of the acute respiratory effects of machining fluids. Final report to the United Autoworkers - General Motors (UAW-GM) National Joint Committee on Safety and Health, Detroit, Michigan, 1994. 6. Robins T, Seixas N, Franzblau A, Burge H, Abrams L, Minick S. Respiratory Effects of Machining Fluid Aerosols. Final Report to the UAW-GM Occupational Health Advisory Board, Detroit, Michigan, 1994. 7. US National Institute for Occupational Safety and Health. Draft Criteria for a Recommended Standard: Occupational Exposure to Metalworking Fluids. Washington, DC: Dept. Of Health and Human Services (DHHS), 1996. 8. American Automobile Manufacturers Association. Sym- posium Proceedings: The Industrial Metalworking Environment. Detroit, Michigan, March 1996. 9. Leon DA, Thomas P, Hitchings S. Lung cancer among newspaper printers exposed to ink mist: a study of trade union members in Manchester, England. Occup Environ Med 1994; 51: 87-94. 10. Lippmann M, Goldstein DH. Oil-mist studies, environmental evaluation and control. Arch Environ Health 1970; 21: 591-9. 11. Hendricks NV, Linden NJ, Collings GH, Dooley AE, Garrett JT, Rather JB. A review of exposures to oil mist. Arch Environ Health 1962; 4: 139-45. 12. Kinnear J, Rogers J, Finn 0, Mair A. Dermatoses in jute workers. BYJ Znd Med 1955; 12: 36-42. 13. Southam AH, Wilson S. Cancer of the scrotum: the etiology, clinical features, and treatment of the disease. BY Med J 1922;4: 971-3. 14. Henry SA. Occupational cutaneous cancer attributable to certain chemicals in industry. BY Med Bull 1947;4:389-401. 15. Cruickshank CND, Squire JF. Skin cancer in the engineering industry from the use of mineral oil. Br J Ind Med 1950; 7 1-11. 16. Waterhouse JAH. Cutting oils and cancer. Ann Occup Hyg 1971; 14: 171-80. 17. Waldron HA, Waterhouse JAH. Mineral-oil cancers b e t ter]. Lancet 1976; 1: 805. 18. Tourenc ER. Cancer of the scrotum in workers of the screw-cutting industry. A study of 21 cases [in French]. Marseille Chirurgical 1964; 1 6 1-11. 19. Thony C, Thony J. An epidemiologic study on cancer produced by cutting oils [in Japanese]. In: Proceedings of the 16th Congress of the Permanent Commission and International Association on Occupational Health, Tokyo, 1969. Tokyo, Japan: Industrial Safety Association, 1970. Jw,20. Roush GC, Kelly J, Meigs Flannery JT.Scrotal carci- noma in Connecticut metalworkers: sequel to a study of sinonasal cancer. Am J Epidemiol 1982; 116 76-85. 21. Kneller RW, Gao Y, McLaughlin JK, et al. Occupational risk factors for gastric cancer in Shanghai, China. Am J Ind Med 1990; 18: 69-78. 22. Siemiatycki J, Dewar R, Nadon L, Gerin M, Richardson L, Wacholder S. Associations between several sites of cancer and twelve petroleum-derived liquids. Scand J Work Enviion Health 1987; 13: 493-504. 23. Chow W, McLaughlin JK, Malker SR, et al. Occupation and stomach cancer in a cohort of Swedish men. Am J Ind Med 1994; 26: 511-20. 24. Gerhardsson de Verdier M, Plato N, Steineck G, Peters JM. Occupational exposures and cancer of the colon and rectum. Am J Znd Med 1992; 22: 291-303. 25. Mack TM, Paganini-Hill A. Epidemiology of pancreas cancer in Los Angeles. Cancer 1981; 4 7 1474-83. 26. Zagraniski RT, Kelsey JL, Walter SD. Occupational risk factors for larynged carcinoma: Connecticut, 1975-1980. Am J Epidemiol 1986; 124: 67-76. Cancer Causes and Control. Vol 8. 1997 403 PE. Tolbert 27. Ahrens W, Jockel KH, Patzak W, Eisner G. Alcohol, smoking, and occupational factors in cancer of the larynx: a case-control study. Am J Znd Med 1991; 20: 477-93. 28. Coggon D, Pannett B, Acheson ED. Use of job-exposure matrix in an occupational analysis of lung and bladder cancers on the basis of death certificates.JNCZ 1984; 72: 61-5. 29. Jockel K, Ahrens W, Wichmann H, et al. Occupational and environmental hazards associated with lung cancer. Znt J Epidemiol 1992; 21: 202-13. 30. Malker HSR, McLaughlin JK, Silverman DT, et al. Occupational risks for bladder cancer among men in Sweden. Cancer Res 1987; 4 7 6763-6. 31. Dunham LJ, Robson AS, Stewart HL, Frank AS, Young JL. Rates, interview, and pathology study of cancer of the urinary bladder in New Orleans, Louisiana. JNCZ 1968; 41: 683-709. 32. Anthony HM, Thomas GM. Tumors of the urinary bladder: an analysis of the occupations of 1,030 patients in Leeds, England.JNCZ 1970; 45: 879-95. 33. Howe GR, Burch JD, Miller AB, et al. Tobacco use, occupation, coffee, various nutrients and bladder cancer. JNCI 1980; 6 4 701-13. 34. Tola S, Tenho M, Korkala ML, Jarvinen E. Cancer of the urinary bladder in Finland. Znt Arch Occup Environ Health 1980; 46: 43-51. 35. Cartwright R. Occupational bladder cancer and cigarette smoking in West Yorkshire. Scand J Work Environ Health 1982; 8 (Suppl) 1: 79-82. 36. Silverman DT, Hoover RN, Albert S, Graff KM. Occupation and cancer of the lower urinary tract in Detroit. JNCZ 1983; 70: 237-45. 37. Vineis P, Magnani C. Occupation and bladder cancer in males: a case-control study. ZntJ Cancer 1985;35: 599-606. 38. Silverman DT, Levin LI, Hoover RN, Hartge P. Occupational risks of bladder cancer in the United States: I. White men.JNCZ 1989; 81: 1472-80. 39. Dolin PJ, Cook-Mozaffari P. Occupation and bladder cancer: a death-certificate study. B Y ] Cancer 1992; 66: 568-78. 40. Cordier S, Clavel J, Limasset JC, et al. Occupational risks of bladder cancer in France; a multicentre case-controlstudy. Znt J Epidemiol 1993; 22: 403-11. 41. Roush GC, Meigs JW, Kelly J, Flannery JT, Burdo H. Sinonasal cancer and occupation: a case-control study. Am J Epidemiol 1980; 111: 183-93. 42. Eisen EA, Tolbert PE, Monson RR, Smith TJ. Mortality studies of machining fluid exposure in the automobile industry I: a standardized mortality analysis. Am J Znd Med 1992; 22: 809-24. 43. Tolbert PE, Eisen EA, Pothier LJ, Monson RR, Hallock MF, Smith TJ. Mortality studies of machining fluid exposure in the automobile industry. 11. Risks associated with specific fluid types. Scand 1 Work Environ Health 1992; 1 8 351-60. 44. Eisen E, Tolbert E Monson RR, et al. Full cohort analysis of digestiveand respiratory cancer risk among autoworkers. Paper presented at Ninth International Symposium on Epidemiology in Occupational Health, Cincinnati, OH (USA), September 23-25, 1992. 45. Sullivan PA, Eisen EA, Woskie SR, et al. Mortality studies of machining fluid exposure in the automobile industry VI: A case-control study of esophageal cancer. Submitted. 46. Tolbert P, Eisen E, Pothier L, et al. Nested case-control study of rectal cancer in automotive workers exposed to machining fluids. Poster presented at Ninth International Symposium on Epidemiology in Occupational Health, 404 Cancer Causes and Control. Vol8. 1997 Cincinnati, OH (USA), September 23-25, 1992a. 47. Bardin JA, Eisen EA, Hallock MF, et al. Mortality studies of machining fluid exposure in the automobile industry V A case-control study of pancreatic cancer. Am J Znd Med 1997; (in press). 48. Eisen EA, Tolbert PE, Hallock MF, Monson RR, Smith TJ, Woskie SR. Mortality studies of machining fluid exposure in the automobile industry 111: a case-control study of larynx cancer. Am J Znd Med 1994; 26: 185-202. 49. Schroeder JC, Tolbert PE, Eisen EA, et al. Mortality studies of machining fluid exposure in the automobile industry IV A case-control study of lung cancer. Am J Znd Med 1997; 31: 525-33. 50. Enterline PE, Sykora JL, Keleti G, Lange JH. Endotoxins, cotton dust, and cancer. Lancet 1985; ii: 934-5. 51. DecouflC P. Further analysis of cancer mortality patterns among workers exposed to cutting oil mists. JNCZ 1978; 61: 1025-30. 52. Jarvholm B, Lillienburg L, SallstenG, Thiringer G, Axelson 0. Cancer morbidity among men exposed to oil mist in the metal industry. J Occup Med 1981; 23: 333-7. 53. Wang JD, Wegman DH, Smith TJ. Cancer risks in the optical manufacturing industry. BY J Znd Med 1983; 40: 177-81. 54. Jarvholm B, Lavenius B. Mortality and cancer morbidity in workers exposed to cutting fluids. Arch Environ Health 1987; 42: 361-6. 55. Acquavella J, Leet T, Johnson G. Occupational experience and mortality among a cohort of metal components manufacturing workers. Epidemiology 1993; 4: 428-34. 56. Rotimi C, Austin H , Delzell E, Day C, MacalusoM, Honda Y. Retrospective follow-up study of foundry and engine plant workers. Am J Znd Med 1993; 2 4 485-98. 57. Delzell.E, Macaluso M, Honda Y, Austin H. Mortality patterns among men in the motor vehicle manufacturing industry. Am J Ind Med 1993; 24: 471-84. 58. Park R, Krebs J, Mirer F. Mortality at an automotive stamping and assembly plant. Am J Znd Med 1994; 2 6 449-63. 59. Silverstein M, Park R, Marmor M, Maizlish N, Mirer F. Mortality among bearing plant workers exposed to metalworking fluids and abrasives.JOccup Med 1988;30: 706-14. 60. Vena JE, Sultz HA, Fiedler RC, Barnes RE. Mortality of workers in an automobile engine and parts manufacturing complex. BYJ Ind Med 1985; 42: 85-93. 61. Leon DA. Mortality in the British printing industry: a historical cohort study of trade union members in Manchester. Occup Environ Med 1994; 51:79-86. 62. Lynge E, Rix BA, Villadsen E, et al. Cancer in printing workers in Denmark. Occup Environ Med 1995;52: 738-44. 63. Paganini-Hill A, Glazer E, Henderson BE, Ross RK. Cause-specific mortality among newspaper web pressmen. J Occup Med 1980; 22: 542-4. 64. Lloyd JW, Decoufli P, Salvin LG. Unusual mortality experience of printing pressmen.] Occup Med 1977;19:543-50. 65. Gilman JPW, Vesselinovitch SD. Cutting oils and squamous-cell carcinoma. Part 11: an experimental study of the carcinogenicity of two types of cutting oils. B Y ] Ind Med 1955; 12: 244-8. 66. Bingham E, Horton AW, Tye R. The carcinogenic potency of certain oils. Arch Environ Health 1965; 10: 449-51. 67. Gupta KP, Mehrotra NK. Tumor initiation in mouse skin by cutting oils. Environ Res 1989; 49: 225-32. 68. McKee RH, Scala RA, Chauzy C. An evaluation of the epidermal carcinogenic potential of cutting fluids. J Appl Toxic01 1990; 10: 251-6. 69. Jepsen J, Stoyanov S, Unger M, Claussen J, Christensen H. Oils and cancer Cutting fluids and their effects on the skin of mice. Acta Patho Mimobto Scand 1985; Sect A 85: 731-8. 70. Evans MJ, Hooper WB, Ingram AJ, Pullen DL, Aston RH. The chemical, physical and biological properties of a neat cutting oil during prolonged use in a large manufacturing facility. Ann Occup Hyg 1989; 33: 537-53. 71. Apostoli P, Crippa M, Fracasso ME, Cottica D, Alessio L. Increases in polycyclic aromatic hydrocarbon content and mutagenicity in a cutting fluid as a consequence of its use. Int Arch Occup Envtron Health 1993; 64: 473-7. 72. Casey P, Hagger R, Harper P. A collaborative study of `ink mist' in UK newspaper press-rooms. Ann Occup Hyg 1983; 2 7 127-35. 73. Kay K. Toxicological evaluation of chemicals used in the printing and printing inks industries. In: Ayer FA, ed. Environmental aspects of chemical use in printing operations. Conference proceedings, September 1975.Washington DC: Office of Toxic Substances, US Environmental Protection Agency, 1976: 1 1 1-39. 74. Bingham E, Falk HL. Environmental carcinogens: the modifying effect of cocarcinogens on the threshold response. Arch Environ Health 1969; 19: 779-83. 75 Kluwe WM, Abdo KM, Huff J. Chronic kidney disease and organic chemical exposures: Evaluations of causal relationships in humans and experimental animals. Fund Applied Tox 1984; 4: 889-901. 76 Lijinsky W, Reuber MD, Manning WB. Potent carcinogenicity of nitrosodiethanolamine in rats. Nature 1980;288: 589-90. 77. US National Institute of Occupational Safety and Health. Current Intelligence Bulletin No. 15:Nitrosamines in cutting fluids. Cincinnati, OH (USA): NIOSH, 1976. 78. Keefer LK, Goff U, Stevens J, Bennett EO. Persistence of N-nitrodiethanolamine contamination in American metalworking lubricants. Food Chem Toxic01 1990; 28: 531-4. 79. Jarvholm B, Zingmark PA, Osterdahl BG. N-nitroso- diethanolamine in commercial cutting fluids without nitrites. Ann Occup Hyg 1991; 35: 659-63. 80. J b h o l m B, Zingmark PA, Osterdahl BG. High concentration of N-nitrosodiethanolamine in adiluted commercial cutting fluid. Am J Ind Med 1991; 19: 237-9. 81. Monarca S, Sforzolini GS, Spiegelhalder B, Pasquini R, Fatigoni C. Monitoring nitrite and nitrosodiethanolamine in mutagenicity in cutting fluids used in the metal industry. Environ Health Persp 1993; 101: 126-8. 82. Horton AW, Bingham E, Burton MJG, Tye R. Carcino- genesis of the skin. 111.The contribution of elemental sulfur and of organic sulfur compounds. Cancer Res 1965; 25: 1759-63. 83. US National Toxicology Program. NTP Technical Report on the Toxicology and Carcinogenesis of Chlorinated Paraffins. Research Triangle Park, N C (USA): NTP TR 305, USDHHS/PHS/NIH, 1986; N I H Pub. No. 86-2506. 84. Blair A, Stewart PA, Hoover RN. Mortality from lung cancer among workers employed in formaldehyde industries. Am J Ind Med 1990; 1 7 689-99. 85. Moore RM, Woolf BS, Stein HP, Thomas AW, Finklea JF. Metabolic precursors of a known carcinogen. Science 1977; 195: 344. 86. Oxhoj H, Andreasen H, Henius UM. Respiratory symp- toms and ventilatory lung function in machine shop workers exposed to coolant-lubricants. EurJ Resp Dis 1982; 63(Suppl 118): 85-9. 87. Einarsson 0,Kylin B, Lindstedt G. Chromium, cobalt and nickel in used cutting fluids. Contact Dermatitis 1975; 1: 182. 88. US National Institute of Occupational Safety and Health. Unpublished provisional data, as of July 1, 1990, from NOES 1981-83. Cincinnati, OH (USA): NIOSH, 1996. 89. Jarvholm B, Fast K, Lavenius B, Tomsic I? Exposure to cutting oils and its relation to skin tumors and premalignant skin lesions on the hands and forearms. Scand J Work Environ Health 1985; 11: 365-9. 90. Park RM, Wegman DH, Silverstein MA, Maizlish NA, Mirer FE. Causes of death among workers in a bearing manufacturing plant. A m J Ind Med 1988; 42: 85-93. 91. Mallin K, Berkeley L, Young Q. A proportional mortality ratio study of workers in a construction equipment and diesel engine manufacturing plant. A m J Z n d Med 1986; 10: 127-41. 92. Greene MH, Hoover RN, Eck RL, Fraumeni JF Jr. Cancer mortality among printing plant workers. Environ Res 1979; 20: 66-73. 93. Brownson RC, Zahm SH, ChangJC, Blair A. Occupational risk of colon cancer. An analysis by anatomic subsite. A m J Epidemioll989; 130: 675-87. 94. Wortley P, Vaughan TL, Davis S, Morgan MS, Thomas DB. A case-control study of occupational risk factors for laryngeal cancer. BY J Ind Med 1992; 49: 837-44. 95. Haguenoer JM, Cordier S, Morel C, Lefebrve JL, Hemon D. Occupational risk factors for upper respiratory tract and upper digestive tract cancers. BY J Znd Med 1990; 4 7 . 380-3. 96. Brown LM, Mason TJ, Pickle LW, et al. Occupational risk factors for laryngeal canceron the Texas Gulf Coast. Cancer Res 1988; 48: 1960-4. 97. Malker HSR, Gemne GA. A register-epidemiology study on cancer among Swedish printing industry workers. Arch Environ Health 1987; 42: 73-82. 98. Menck HR, Henderson BE. Occupational differences in rates of lung cancer.] Occup Med 1976; 18: 979-801. 99. Zoloth SR, Michaels DM, Villabi JR, Lacher M. Patterns of mortality among commercial pressmen. J N C I 1986; 76: 1047-51. 100. Claude JC, Frentzel-Beyme R, Kunze E. Occupation and risk of cancer of the lower urinary tract among men. A case-control study. Int J Cancer 1988;41: 371-9. 101. Gonzalez CA, Lopez-Abente G, Errezola M, et al. Occupation and bladder cancer in Spain a multi-centre case-control study. Int J Epidemioll989; 18: 569-77. 102. Steenland K, Burnett C, Osorio AM. A case-control study of bladder cancer using city directories as a source of occupational data. Am J Epidemiol 1987; 126: 247-57. 103. Schifflers E, Jamart J, Renard V. Tobacco and occupation as risk factors in bladder cancer: a case-control study in southern Belgium. I n t J Cancer 1987; 39: 287-92. 104. Howe GR, Lindsay JP. A follow-up of a ten-percent sample of the Canadian labor force. I. Cancer mortality in males, 1965-73. I N C I 1983; 70: 37-44. 105. Brownson RC, ChangJC, Davis JR. Occupation, smoking, and alcohol in the epidemiology of bladder cancer. Am J Public Health 1987; 7 7 1298-300. Cancer Causes and Control. Vol 8. 1997 405