Occupational and Environmental Medicine 1995;52: 185-19 1 185 Maintenance of stellite and carbide

tips: respiratory health and exposure-response Occup Environ Med: first published as 10.1136/oem.52.3.185 on 1 March 1995. Downloaded from evaluations

Susan M Kennedy, Moira Chan-Yeung, Steve Marion, John Lea, Kay Teschke

Abstract Hard-metal lung disease occurs in workers Objective-To study exposure to exposed to cobalt in association with tungsten and in saw maintenance carbide (hard-metal)1 2 and in other metal rooms and test respiratory health among working processes that involve exposure to saw filers at lumber mills. Hard-metal cobalt, such as diamond polishing.3 The dis- lung disease is associated with cobalt in ease is an interstitial pneumonitis charac- the manufacture of terised by radiographic infiltrates, reduced tools; recently it has also been reported diffusing capacity and lung volume,4 and a among tool maintenance workers. histology of fibrosing alveolitis with unusual Lumber mills often use tipped with multinucleated giant cells.' Prevalence of tungsten carbide or with a newer alloy, about 1% is generally reported although some stellite (containing more cobalt, as well studies suggest higher rates.46 Fatal cases have as chromium). been reported,7 and the predictors of fatality Methods-A cross sectional study of 118 are not clear. Also, cobalt is known to cause saw filers at eight lumber mills was car- occupational asthma' 8 and asthma like symp- ried out that included a standardized toms.2 questionnaire, spirometry, personal air Most studies of hard-metal lung disease in sampling, and examination of tasks every association with tungsten carbide have been 10 minutes (by observation). Comparison limited to the tool manufacturing industry. data were from a study of bus mechanics We are aware of only one recent report of four tested with similar methods. cases of hard-metal lung disease among work- Results and Conclusion-Cobalt expo- ers who use or maintain these tools.' We sure was associated with tungsten carbide report here results from a study of saw filers, grinding but not with stellite grinding. who maintain saws in lumber mills. Chromium exposure was associated Saw blades in lumber mills are generally mainly with stellite welding. Saw filers made of hardened steel, but the cutting tips had a twofold increase in phlegm and are often composed of specialty compounds wheeze (P < 0.01) and a threefold welded or soldered to the saw body. In the increase in cough, phlegm, and wheeze past two decades, soldered tungsten carbide http://oem.bmj.com/ related to work (P < 0.001), but no tips have become widely used. Tungsten car- increase in breathlessness. Stellite weld- bide is a sintered metal composite with cobalt ing was associated with a significant (from 3-30%) as the binder for the tungsten increase in nasal symptoms and cough carbide and other trace metals.'0 More related to work and a small decrease in recently, saw tips have been welded with a airflow volume in one stellite (Deloro Stellite, Belleville, Occupational and (forced expiratory superalloy,

Environmental Lung second/forced vital capacity (FEV1/ Ontario). In this alloy, cobalt (50-60%), on October 1, 2021 by guest. Protected copyright. Diseases Research FVC%), P < 0.05). Saw filers wet grind- chromium (30-35%), and other trace metals Unit, Respiratory ing with tungsten carbide had significant are heated above their melting points to create Division, Department of Medicine reductions in forced expiratory lung vol- prealloyed powder, which is mixed with an S M Kennedy umes (FEV1 and FVC, P < 0.05) and were organic binder before final sintering.10 Some M Chan-Yeung significantly more likely to have FEV1 and saw filers also sharpen chipper knives that are Department ofHealth FVC values in the abnormal range. predominantly iron (85-90%), but also con- Care and wet and tain some chromium " In lumber Epidemiology Cobalt exposure (in grinding) (8-9%). S M Kennedy duration of work that involved tungsten mills in British Columbia, Canada, neither S Marion carbide grinding were both associated tungsten carbide nor stellite tips are univer- JLea with significant reductions in FEV1 and sally used, although many mills use one or the K Teschke FVC. Average cobalt exposures in this other, and some use both. Occupational Hygiene Programme, study were about 5 pglm3, well below the Depending on the saw type and the University of British currently accepted permissible concen- process, saws in the lumber industry are used Columbia, Vancouver, tration, which suggests that the current for only a few work shifts and then require Canada be too Saw filers a of S M Kennedy workplace limit for cobalt may sharpening. perform variety K Teschke high. tasks including removing saws from produc- Correspondence to: tion lines, transferring them to the filing Dr Susan M Kennedy, (Occup Environ Med 1995;52:185-191) room for maintenance, and reinstalling saws Occupational Hygiene Programme, University of into production. The maintenance work British Columbia, 2206 East involves setting up and operating machinery Mall, Vancouver, British Columia, Canada V6T 1Z3. Keywords: hard-metal lung disease; cobalt; tungsten to solder or weld new tips to the saw body, Accepted 24 November 1994 carbide grinding and sharpening tips, and occasionally 186 Kennedy, Chan-Yeung, Marion, Lea, Teschke

soldering and welding tips manually. which metal concentrations were quantified Grinding and sharpening often requires lubri- by inductively coupled argon plasma atomic

cation with soluble machining fluids. Also, emission spectrophotometry. A total of 278 Occup Environ Med: first published as 10.1136/oem.52.3.185 on 1 March 1995. Downloaded from tasks for quality control and safety assurance air samples were obtained from 112 workers. of the saw bodies are performed. Although Every worker in the saw filing room was workers with different levels of training and involved in the exposure monitoring, regard- seniority perform some of these tasks prefer- less of his job tasks or suspected exposure entially to others, no saw filers are engaged in potential. Coincident with air sampling, every grinding for an entire shift. 10 minutes, an industrial hygiene technician This report describes the results of a survey recorded each worker's specific activity, the of respiratory health among saw filers in eight name of the nearest operating machine, the volunteer lumber mills in coastal British distance between the breathing zone and the Columbia. Our principal objective was to generation point of the contaminant, and determine whether or not these workers were whether or not the worker was smoking. As exposed to excessive concentrations of cobalt well as monitoring airborne exposure, 73 bulk or chromium, and if so, to evaluate the deter- samples of used coolant were obtained from minants of exposure; the respiratory health lubricated grinding machines and analysed for assessment was conducted as a pilot survey metals. only. It should be noted that the study design More than half of the monitored values for (volunteer mills) does not ensure a represen- cobalt and chromium in air were below the tative picture of the entire industry, and that detection limit-that is, severe left censor- the sample size was not expected to be suffi- ing-and the remaining data did not meet cient to detect clinically relevant cases of tests of normality, even with log transforma- hard-metal lung disease. Results from the tion.'2 This presented a challenge for statisti- exposure evaluation are reported separately.'2 cal analysis of the determinants of exposure. Details and rationale for the technique of analysis of the exposure data are presented Methods elsewhere. 12 Briefly, linear regression analyses PARTICIPANTS were conducted with the cobalt or chromium Saw filers were identified from lists provided concentration (original scale) as the depen- by the management at each mill. Each eligible dent variable (with a value of half the detec- participant (all men) was contacted by letter tion limit entered for concentrations below and invited to participate. Participation was the detection limit). Seven activity and 10 entirely voluntary and each participant signed a location variables were considered as the "consent to participate" form before the tests, potential predictors. Both dichotomous (ever after having the aims and objectives of the during the day v never) and continuous study explained to him. The study protocol (number of times observed during the day) was approved by the University of British versions of these variables were considered; Columbia Committee on Ethics in Research however, the dichotomous versions were pre- Involving Human Subjects. ferred in almost all cases. Modelling was done with generalised estimating equations and http://oem.bmj.com/ EXTERNAL COMPARISON POPULATION robust estimation of variance (thus taking into As this was a pilot study, no external compari- account both non-conformance with the son population was included in the protocol. equal variance assumption of ordinary least For some analyses comparisons were made squares regression and the presence of with results obtained from bus mechanics repeated measures from the same worker). studied by us in December 1990 with the This approach has the advantage that each

identical respiratory symptom questionnaire regression coefficient for a dichotomous vari- on October 1, 2021 by guest. Protected copyright. and pulmonary function testing procedures able represents the mean exposure for that and equipment. The technicians and inter- activity or location. viewers for both studies were all trained by the same study coordinator. This group was stud- EVALUATION OF RESPIRATORY HEALTH ied due to their concern about past use of An expanded version of the American asbestos products. Results indicated that Thoracic Society standard respiratory ques- there was little or no impairment of health tionnaire"3 was given by a trained interviewer. among the bus mechanics. Therefore, com- An occupational history, including a record of parison of results from the saw filers with the duration and frequency for each of nine those from the bus mechanics is warranted for job tasks (including type of metal used and some analyses. the use of machining fluids), was also obtained by the interviewer. Routine spiro- EXPOSURE MONITORING AND EVALUATION OF metry was carried out at the worksite by a DETERMINANTS OF EXPOSURE trained technician with a 10 litre dry rolling Details of the exposure assessment and evalu- seal spirometer (S and M Instruments, ation of determinants of exposure are pre- Doylestown, PA, USA), according to stan- sented elsewhere. 12 A brief summary is dard American Thoracic Society proce- provided here. Each mill was visited four dures.'4 Subjects were seated, wearing nose times at least one month apart from June to clips, and were required to perform at least December 1991. On each test day all saw filers three acceptable forced expiratory manoeuv- on the day shift were asked to wear a personal res. The test was ended when a minimum air sampler for a minimum of six hours, from plateau in expired volume of two seconds was Maintenance ofsteiite and tungsten carbide saw tips: respiratory health and exposure-response evaluations 187

reached. The highest value for forced expira- Time spent at these tasks, as well as details tory volume in one second (FEV,) forced vital about the frequency of use of coolants was

capacity (FVC), and the midmaximal flow used to define six categories of potential Occup Environ Med: first published as 10.1136/oem.52.3.185 on 1 March 1995. Downloaded from rate (MMF) from the blow with the largest exposure. Tungsten carbide wet grinding was sum of FEV, and FVC were used for analysis. defined as work in a job that involved grinding All volumes were corrected to body tempera- of tungsten carbide saw tips at least 10% of ture, pressure, and saturation. Prediction the time, for which at least 50% ofthe grinding equations used were those of Crapo and asso- was performed with a coolant. Tungsten car- ciates.'5 Spirometry was performed just before bide dry grinding was similarly defined, but starting work and at the end of the shift on required at least 50% of the grinding to be each of the days that the workers wore the performed without a coolant. Stellite wet and personal air samplers. Thus, each worker dry grinding were defined similarly. Tungsten could have a minimum of one and a maxi- carbide soldering (or stellite welding) was mum of four days of measurements. Subjects defined as a job that involved soldering of were asked to refrain from smoking for at least tungsten carbide saw tips (or welding of one hour before each test. stellite tips) at least 10% of the time. The For analyses of spirometric outcomes in number of years associated with work that which comparisons were made with the exter- involved such tasks was also calculated by nal control population, the average of the val- summing all the years in which a worker's job ues before and after shift on the first day of satisfied one or more of these definitions. testing of the saw filers was used. This was necessary to provide unbiased results that would be directly comparable to those from Results the control population, as the spirometric test DEMOGRAPHICS in the control group was performed only once A total of 131 filing room personnel were (randomly during the day, not before and eligible to participate; 118 took part (90%). after the shift). Reasons for lack of participation were sickness (n = 3), refusal (n = 3), and not being avail- DATA HANDLING AND ANALYSIS able during the test days (n = 7). The mean Analyses were performed with SAS-PC statis- (range) age was 46-7 (31 to 62) years. The tical software (SAS Institute, Cary, NC, duration of employment averaged 21-3 years USA). Respiratory symptoms were defined as a in the lumber mill and 15-5 years in the saw positive answer to the questions: cough, "Do filing trade. Only 19% were current cigarette you usually have a cough?"; phlegm, "Do you smokers; 59% were former smokers. The usually bring up phlegm from your chest?"; external comparison group was slightly wheeze, "Does your chest ever sound wheezy younger (mean 44 1 years), and had about the or whistling occasionally apart from when you same proportion of current smokers (21%) have a cold?"; breathlessness, "Are you trou- but slightly fewer former smokers (42%, bled by shortness ofbreath when you hurry on P < 0 0 1). There were no differences between the level or walk up a slight hill?"; nasal symp- these groups with respect to height, race, toms, "Do you usually have an itchy, runny history of asthma, heart disease, or atopy. http://oem.bmj.com/ nose when you don't have a cold?". A work related symptom was defined as a symptom EXPOSURE EVALUATION that was not present before starting work in Table 1 shows the specific jobs performed by the trade, and that improved when away from these saw filers. Although all 118 workers work on weekends or long holidays. were identified as saw filers, the number To evaluate the relation between tasks per- working at tasks in which exposure may have

formed and respiratory health outcomes, each occurred was more limited. Also, although on October 1, 2021 by guest. Protected copyright. worker's job history was analysed to deter- more workers reported more frequent current mine whether or not the worker's current (or grinding or welding of stellite tips than tung- past) work involved tungsten carbide grind- sten carbide tips, the number of years in ing, tungsten carbide soldering, stellite grind- which tungsten carbide had been handled was ing, stellite welding, and various other tasks. greater than that for stellite. The proportion of the working day spent grinding or welding varied among the workers, but was seldom > 50% of the day. The industrial hygiene assessment showed Table 1 Frequency and duration oftasks within jobs in the saw filing trade (current and detectable cobalt concentrations in 62 of the past) 278 air samples collected. Of the samples Ever Performing Duration * ofjobs above the detection limit, the mean (SD, performing this task in in which this task was this task currentjob was performed maximum) cobalt concentration 9 (20, (n) (n) (y (range)) 106) pg/mi. The permissible eight hour in our local is 100 Tungsten carbide wet grinding (>10%) 18 concentration jurisdiction Tungsten carbide dry grinding (>10%) 43f A 6-9 (0-5-22) ug/in3 with a newly proposed reduction to Stellite wet grinding (>10%) 11 20 Factors that were Stellite dry grinding (>10%) 58f 18 38(0510) pug/M3n. significantly Tungsten carbide soldering (> 10%) 56 19 9-6 (0-5-35) associated with airborne cobalt concentration Stellite welding (>10%) 51 36 3-4 (0-5-8) were: working within five feet of a wet tung- Chipper knife wet grinding 40 15 8-3 (0-3-35) sten carbide grinding machine (coefficient: * Mean duration is shown only for those who have ever performed this task; t distinction 4-7 P < within five feet between wet and dry grinding was not always possible for past jobs (therefore, except for the cur- pg/M3, 0-05), working rent job, these are taken together). of a dry tungsten carbide grinding machine 188 Kennedy, Chan-Yeung, Marion, Lea, Teschke

(coefficient: 4-5 jug/m3, P < 0 05), and having a formed for each symptom, in which age, job title of knife grinder or head filer (both smoking, job (saw filer v bus mechanic), and

coefficients: -0-4 jug/m', P < 0'05). Together, the work tasks (table 1) were included as pre- Occup Environ Med: first published as 10.1136/oem.52.3.185 on 1 March 1995. Downloaded from these factors accounted for 24% of the vari- dictors. The only work task that was a signifi- ance in the cobalt air concentrations. cant predictor for any symptom was stellite Chromium was detected in 105 of the air welding, associated with increased relative samples with a mean (SD, maximum) of 4 (9, odds of 5-0 (95% confidence interval (95% 55) jug/m'. The permissible concentration for CI) 1 9 to 13-1) for nasal symptoms and 3-7 chromium exposure is 50 ,ug/m3 for hexava- (95% CI 1N1 to 12-0) for work related cough. lent chromium and 500 jug/M3 for other forms Inclusion of stellite welding in the regression of the metal. (The analytical method used in model did not reduce the magnitude of the this study did not distinguish the type of increased relative odds associated with saw fil- chromium present.) Work factors significantly ing in general (compared with bus mechan- associated with chromium exposure were: ics), for any of the symptoms shown. working within five feet of a stellite welding operation (coefficient 2-4 pg/M3, P < 0-001), ACUTE PULMONARY FUNCTION EFFECTS working within five feet of a saw steel heating DURING A SHIFT operation (coefficient 1 4 pug/M3, P < 0.001), A total of 35 saw filers had a > 5% drop in and work as a knife grinder (coefficient 12-2 FEVy over the shift on at least one day and a pug/M3, P < 0-001). These factors accounted > 5% average change (over all test days) was for 24% of the variance in this regression seen in 12 workers. There were no significant model. relations found between the change in FEV, The within subject variability in cobalt and over one workshift and any of the following: chromium concentrations was evaluated in measured cobalt or chromium exposure (both the 52 subjects with three or more measure- log transformed and untransformed); number ments. The mean (range) coefficient of of saw changes; cigarette smoking; age; or variation for cobalt was 102-5 (8-4 to 200-0) working in a cedar mill (compared with a and for chromium, 63-3 (0 to 199 9). No job non-cedar mill). Participants with a history of title or mill showed a within subject variation asthma were significantly more likely to expe- significantly different from these mean values. rience an increase in FEVy over the work shift, Cobalt concentrations in used coolant compared with those with no such history samples averaged 664 mg/l (n = 29 samples) (P < 005, data not shown). A detailed case from tungsten carbide grinding machines, 76 by case examination of the respiratory symp- mg/l (n = 23 samples) from stellite grinding toms reported and the changes in lung func- machines, and 0-25 mg/l (n = 18 samples) tion (and corresponding daily cobalt and from chipper knife grinding machines. Con- chromium exposures) showed no obvious or centrations of chromium in these samples suggestive cases of occupational asthma were 0-25 mg/l, 14-8 mg/l, and 1-86 mg/l specific to cobalt or chromium. respectively. PULMONARY FUNCTION RESULTS RESPIRATORY SYMPTOMS No significant differences were found for http://oem.bmj.com/ Compared with the external comparison FEVy or FVC between saw filers as a group group, saw filers reported about twice the rate and the external comparison population, of phlegm production (22% v 11 %) and although saw filers had a lower average wheeze (23% v 10%, both P < 0.01), and FEV1/FVC% (75-6 v 77 0, P < 0 05) and about three times the rate of cough, phlegm, MMF (3-25 v 3*47 U/s, P = 0 06), after con- and wheeze related to work (14,14,9% v trolling for age, height, race, smoking, and

3,4,4% respectively, P < 0 05). No significant amount smoked (pack-years). on October 1, 2021 by guest. Protected copyright. difference was found between the groups for Table 2 shows results from linear regres- breathlessness or nasal symptoms. sion analyses for pulmonary function out- Logistic regression analyses were per- comes in which age, height, race, smoking, pack-years, history of childhood asthma, and the jobs (table 1) were included. Tasks signifi- cantly associated with lung function outcomes Table 2 Effect ofwork task and demographic variables on lungfunction variable were wet grinding of tungsten carbide in the current job (associated with reduced FEV,, FEV, FVC MMF FEV,/FVC FVC, and MMF) and welding of stellite in (7i) (1) (7s) (%/6) the current job (associated with reduced Saw filer v bus mechanic + 0 010 + 0 073 - 0-118 - 0-88 FEV,/FVC). The number of years spent in Tungsten carbide tungsten carbide grinding jobs (either wet or wet grinding - 0.384* - 0.376* - 0-607t - 2-21 dry) was significantly associated with reduced (current job) Stellite welding - 0-016 + 0-158 - 0 334 - 3-27* FVC when current job tasks were not (current job) Age (y) - 0-032** - 0-024** - 0-061** - 0.26** included in the model (not shown), but when Smoking (pack-y): current tungsten carbide wet grinding was Current Former - 0 005* - 0-002 - 0-008* - 0806** taken into account years spent at this task Tungsten carbide wet or dry became only marginally significant (table 2). grinding, duration (y) - 0-007 - 0 023t + 0-023 + 0-22 Table 3 shows odds ratios (ORs) (adjusted tP<0-10; *P<005;**P<0 01. for age, race, history of childhood asthma, t Values shown are coefficients from multiple linear regression models in which height, race, his- and sm tory of childhood asthma were included in the analyses, as well as all of the predictor variables king) for predictors of abnormal shown in the table. FEV1adV and MMF (FEy1 < 80% predicted and Maintenance ofstellite and tungsten carbide saw tips: respiratory health and exposure-response evaluations 189

Table 3 Odds ratio* (95% CI) asjob task predictors ofabnormal levels oflungfunction quantitative exposure-response modelling, FEV, <80% predicted MMF <60% predicted estimated cobalt (and chromium) exposures

were assigned to each saw filer with the pre- Occup Environ Med: first published as 10.1136/oem.52.3.185 on 1 March 1995. Downloaded from Current smoker 2-6 (1-0,7 0) 3-1 (1 5, 6 5) Former smoker 1-7 (0 7,3 9) 1 1 (0-6,2 0) dicted cobalt (and chromium) exposure val- Saw filer ues from the analysis of determinants of (v bus mechanic) 0-5 (0-2, 1-3) 0 9 (0-5, 1-8) Tungsten carbide exposure, according to their current job tasks. wet grinding Only the analysis with respect to cobalt expo- (current job) 6-6 (1-7, 24 8) 1-5 (0-4, 5 4) Stellite welding sure was significant. (current job) 2-9 (0-8, 98) 2-3 (0-8, 6 2) For this analysis, saw filers who did not *From logistic regression models, including age, history of childhood asthma, race, and all of the perform tungsten carbide grinding were variables listed above. assigned the predicted background exposure (077 ug/m3), head filers and knife grinders were assigned the predicted (lower than aver- Table 4 Linear regression models ofexposure response relations (lumber mill population age background) exposure of 0-33 ,ug/m', and only) saw filers who performed tungsten carbide FEV, (7) FVC (7) grinding in their current job were assigned the (coefficient (SEM)) (coefficient (SEM)) cobalt exposure associated with that task (5-6 4ug/m3 for wet grinding, 5-4 pug/m3 for dry Model 1: Cobalt exposure grinding). Regression analyses were per- (ug/m3)t -0 070 (0-025)* - 0-072 (0 029)* formed for pulmonary function outcomes, Model 2: Tungsten carbide which included as predictors this estimated grinding duration (y) -0-020 (0-008)* -0-029 (0-009)** level of cobalt exposure, the difference Model 3: Cobalt exposure between the estimated level of exposure and (ug/m3)4 -0 055 (0-027)* -0-042 (0-031) the average level of exposure for each partici- Tungsten carbide grinding duration (y) -0-012 (0-009) -0-022 (0 011)* pant, and the duration of employment in jobs that involved tungsten carbide grinding. *P <0-05; ** P <001. tValues shown are coefficients (SEM) from multiple linear regression models in which age, Table 4 shows the results for FEVy and FVC. height, race, history ofchildhood asthma, and cigarette smoking and amount were included in the (No significant relations were found between analyses, as well as the predictor variables shown in the table. tCobalt exposure as estimated from regression analyses of determinants of exposure applied to exposure and MMF or FEV,/FVC%.) Both each worker's job history. level of cobalt exposure and duration of tung- sten carbide grinding were significantly asso- ciated with reductions in FEVI and FVC when these variables were included separately MMF < 60% predicted). Wet grinding of in the models (models 1 and 2). When both tungsten carbide in the current job was associ- variables were analysed together (model 3), ated with a sixfold increase in OR for FEV, coefficients for each decreased; FEV1 was < 80% predicted. Stellite welding was associ- slightly more likely to be associated with level ated with a twofold increased OR for both of exposure, whereas FVC was slightly more abnormal FEVI and MMF (although this was likely to be associated with duration of expo- not at the P < 0 05 level). No other sure. significant http://oem.bmj.com/ job task was significantly associated with these When the difference between the estimated outcomes. It was not possible to carry out this exposure and the actual average measured type of analysis for abnormal FVC because exposure was included in these models, the the number of subjects with FVC < 80% pre- only change found was that the coefficients dicted (n = 4, all long term ex-smokers) was for estimated exposure became slightly more too small to allow for stable statistical model- negative and had a slightly higher level of sig- ling. The unadjusted OR for abnormal FVC nificance in every case. The coefficient for the associated with wet tungsten carbide grinding difference between the two exposure values on October 1, 2021 by guest. Protected copyright. in the current job was 5.3 (95% CI 07 to was close to 0 and not significant. These 40-0) and for each decade of wet tungsten results are not shown. carbide grinding 5'0 (95% CI 1-1 to 23-9). Finally, when separate cobalt exposures Several additional work factors were con- were included for wet grinders and all others, sidered in these models but were not found to the wet grinding cobalt exposure term be associated with any of the health outcomes. remained significant and about the same size, These included working in a cedar mill (or whereas the exposure term for other workers duration of time in a cedar mill); average became non-significant and approached 0. number of times each worker changed saws in the mill itself (representing potential for wood dust exposure); and duration of employment Discussion in the saw filing trade, with the current These results suggest that grinding tungsten employer, and in saw milling. carbide tools, especially in the presence of a As personal exposure monitoring in this coolant, may be associated with significant study was limited to data collected on a maxi- reductions in measured lung capacity (FEV, mum of four working days, and the within and FVC). This reduction did not seem to be subject variation was extremely high, the mea- accompanied by breathlessness in workers sured personal exposure values collected on currently employed in the volunteer mills. the study days were unlikely to provide rea- Stellite welding was associated with symptoms sonable quantitative estimates of each of irritation (nasal symptoms and cough) and worker's average exposure over an extended small reductions in airflow rates (specifically period of time. Therefore, to perform limited FEV,/FVC% and MMF). These two 190 Kennedy, Chan-Yeung, Marion, Lea, Teschke

exposures and combination of outcomes are wet grinders were not the group most highly discussed separately. exposed.

It is noteworthy that despite a high cobalt Hard-metal disease is generally charac- Occup Environ Med: first published as 10.1136/oem.52.3.185 on 1 March 1995. Downloaded from concentration in stellite, this was not associ- terised by both reduced FVC and breathless- ated with increased airborne cobalt con- ness. Although a strong exposure-response centrations or in increased cobalt in the relation was seen in this study, the values for coolant fluid from machines on which stellite lung function were generally in the clinically tips were ground. Furthermore, stellite grind- normal range; thus it is not surprising that ing was not associated with decreases in lung there was no obvious increase in breathless- volumes or with increases in respiratory ness. It is not possible to determine, from this symptoms. study, whether or not there will be further Stellite welding was associated with deterioration in these workers to the stage of increased nasal symptoms and slight symptomatic hard-metal lung disease, or decreases in the FEV,/FVC ratio. These whether the results from these volunteer mills effects may be the result of chromium are representative of the industry as a whole. exposure as both nasal symptoms and reduc- Although at the time this study started, we tions in airflow have been associated with air- were not aware of any cases of hard-metal borne chromium exposure in other studies.'6 lung disease among saw filers in lumber mills These effects were not found, however, in in British Columbia, after conclusion of this association with chipper knife grinding, the study, we learned of a recent fatal case of task with the highest measured cobalt expo- hard-metal lung disease in a 50 year old saw sure. One explanation is that the form of filer from a different British Columbian lum- cobalt may differ between these tasks. ber mill (A Churg, personal communication). Exposure to hexavalent chromium has been This confirms the potential for serious pathol- found in association with weld- ogy among workers in this trade. ing'7; thus it is possible that stellite welding In this study, the average cobalt exposure could also produce this form of chromium associated with significant reductions in lung (whereas the chipper knife grinding may not volume was well below the allowable con- generate sufficient heat to oxidise the centrations of most jurisdictions and even chromium). well below the newly recommended 1994 Tungsten carbide grinding in these mills permissible concentration of the American was associated with increased airborne cobalt Conference of Govemmental and Industrial concentrations and with very high concentra- Hygienists. Similar results of significant tions of cobalt in the coolant from tungsten reductions in FEV1 and FVC associated with carbide wet grinding machines. Workers who low cobalt exposure have been reported by performed this task, particularly those who Nemery and colleagues in a study of diamond used coolant most of the time, had significant polishers.20 The average airborne cobalt reductions in both FEVy and FVC compared exposure in the high exposure workshops in with other saw filing tradesmen. that study was 10-2,ug/m3. Also similar to Airborne cobalt exposures were similar for our results was their finding of increased dry and wet grinding, but significant FEV1 symptoms of cough and upper airway http://oem.bmj.com/ and FVC reductions were found only in asso- irritation but no significant increase in ciation with predominantly wet grinding. One breathlessness. might conclude that the coolant itself is the Our results suggest that a permissible con- aetiological agent. Although machining centration for cobalt exposure (at least coolants have been implicated in pulmonary for cobalt in association with tungsten carbide function decrements in other studies,'8 this is grinding) should be set at a value considerably

an unlikely explanation here as wet grinding lower than the currently recommended on October 1, 2021 by guest. Protected copyright. of the other metals-for example, stellite and 20 ,ug/m3. Based on our findings alone, this mild steel-with the same coolants, did not conclusion is most relevant for cobalt associ- result in reductions in lung function. The ated with coolants and is dependent on the grinding fluid may, however, have contributed assumption that the average cobalt con- to a change in the form or toxicity of the centration measured in the eight study mills is cobalt. Cobalt from dry grinding may remain a reasonable reflection of the exposures in elemental form, whereas cobalt from wet that would have been found in these mills grinding may be present as (potentially more over the past several years. As the induction toxic) cobalt ion in solution; cobalt may have period for hard-metal lung disease can vary an enhanced toxic potential when present in from as little as one up to 30 years from first coolant due to an adjuvant effect of the exposure,'4 both recent and past exposures coolant; or additional exposure from ingestion may contribute to the disease process. In the or skin absorption may occur more often in absence of information to evaluate recent and association with wet grinding. Our study can- distant past exposures, it would be prudent to not distinguish between these possible mecha- assume that the exposures measured in this nisms for the difference in physiological effect study are not significantly different from past of wet v dry grinding. exposures in these mills. This, together with Two previous studies have shown wet the previously published evidence of similar grinding of tungsten carbide in tool manu- exposure-response relations in a different facturing to be associated with an increased industry,20 argue in favour of establishing a rate ofhard-metal disease and with reductions permissible concentration for cobalt close to in diffusing capacity,2 19 despite the fact that the average values found in these mills- Maintenance ofstellite and tungsten carbide saw tips: respiratory health and exposure-response evaluations 191

for example, a time weighted average of 9 Barnhart S, Daniell W, Stebbins A, Rosenstock L. Occurrence of hard metal pneumoconiosis at exposure 5 pg/mr. The exposure measurements made levels below the permissible exposure limit. Am Rev

in this study show that cobalt from tungsten RespirDis 1991;143:A263. Occup Environ Med: first published as 10.1136/oem.52.3.185 on 1 March 1995. Downloaded from 10 Bayer AM, Vasco T, Walton LR. Wrought tool steels. In: carbide grinding can be successfully Metals handbook, J(t ed. Vol. 1. Properties and selection: controlled to this air concentration. irons, steels, and high-performance alloys. Ohio: American Society for Metal, 1990:757-63. 11 Material safety data sheet for chipper knife A 8 modified. Delta, British Columbia: Simonds Industries, 1992. 12 Teschke K, Marion SA, van Zuylen MJA, Kennedy SM. Maintenance of stellite and tungsten carbide saw tips: 1 Davison AG, Haslam PL, Corrin B, Coutts II, Dewar A, determinants of exposure to cobalt and chromium. Am Riding WD, et al. Interstitial lung disease and asthma in Ind HygAssoc J 1995 (in press). hard-metal workers: bronchoalveolar lavage, ultrastruc- 13 American Lung Association. Recommended respiratory tural and analytical findings and results of bronchial diseases questionnaire for use with adults and children in provocation tests. Thorax 1983;38:1 19-28. epidemiological research. Am Rev Respir Dis 1978;118:7. 2 Sprince NL, Oliver LC, Eisen EA, Greene RE, 14 American Thoracic Society. Standardization of spirome- Chamberlain RI. Cobalt exposure and lung disease in try-1987 update. Am Rev Respir Dis 1987;136: tungsten carbide production. A cross-sectional study of 1285-98. current workers. Am Rev Respir Dis 1988;138:1220-6. 15 Crapo RO, Morris AH, Gardner RM. Reference spiromet- 3 Demedts M, Gheysens B, Hagels J, Verbeken E, ric values using techniques and equipment that meet Lauweryns J, Van den Eeckhout A, et al. Cobalt lung in American Thoracic Society recommendations. Am Rev diamond polishers. Am Rev Respir Dis 1984;130:130-5. Respir Dis 1981;123:659-64. 4 Fischbein A, Jiin-Chyuan JL, Solomon SJ, Horowitz S, 16 Schneider WD, Grund W, Schmidt G, Wiesnerchromium Hailoo W, Miller A. Clinical findings among hard metal W. Respiratory diseases in welders on stainless steel workers. BrJ Ind Med 1992;49:17-24. and/or aluminium-a cross sectional study. EurJ Respir 5 Ohori NP, Sciurba FC, Owens GR, Hodgson MJ, Yousem Dis 1982;62:179-80. SA. Giant-cell interstitial pneumonia and hard-metal 17 Gray CN. Hexavalent chromium in welding fume and the pneumoconiosis. Am Surg Pathol 1989;13:581-7. role of ozone. Ann Occup Hyg 1987;31:269-70. 6 Meyer-Bisch C, Pham QT, Mur JM, Massin N., Moulin JJ, 18 Kennedy S, Greaves IA, Kriebel D, Eisen EA, Smith TJ, Teculescu D, et al. Respiratory hazards in hard metal Woskie SR Acute pulmonary responses among automo- workers: a cross sectional study. Br Ind Med 1989; bile workers exposed to aerosols of machining fluids. Am 46:302-9. J Ind Med 1989;15:627-41. 7 Nemery B, Nagels J, Verbeken E, Dinsdale D, Demedts 19 Sjogren I, Hillerdal G, Anderson A, Zetterstrom 0. Hard M. Rapidly fatal progression of cobalt lung in a diamond metal lung disease: importance of cobalt in coolants. polisher. Am Rev RespirDis 1990;141:1373-8. Thorax 1980;35:653-9. 8 Gheysens B, Auxerx J, Van den Eeckhout A, Demedts M. 20 Nemery B, Casier P, Roosels D, Lahaye D, Demedts M. Cobalt-induced bronchial asthma in diamond polishers. Survey of cobalt exposure and respiratory health in dia- Chest 1985;88:740-4. mond polishers. Am Rev Respir Dis 1992;145:610-6. http://oem.bmj.com/ on October 1, 2021 by guest. Protected copyright.