Mortality chrome leather workers and chemical exposures in tanneries.

chrome leather tannery workers and chemical exposures in tanneries. Scand J Work Environ Health 13 (1987) 108-117. A retrospectivemortalityanalysis wasconducted in a cohort of 9 365 individualsemployed as of 1940 in two chrome leather tanneries in the United States and followed to the end of 1982. Vital status as of the closing date was determined for over 95 Cllo of the cohort. Potential hazardous workplace exposures varied with department and included nitrosamines, chromate pigments, benzidine-based direct dyestuffs, formaldehyde, leather dust, and aromatic organic solvents. Mortality from all causes combined was lower than expected for each tannery, the standardized mortality ratio being 81 for one and 93 for the other. Deaths from cancer of each site, including the lung, werealso lower than expected compared to thoseof either the population of the United States or of local state rates. A significantexcess of deaths was observed, however, due to accidental causes in one tannery and cirrhosis of the liver, suicide, and alcoholism in the other. These excesses did not appear to be causally associated with occupational ex posures. The findings of this study are consistent with those of the only other mortality investigation of leather tannery employees.

epidermis and sub cutaneous layer of the hide and the subsequent stabilizat io n of the middle portion (derma) . Two main processes fo r leather a re used in th e United States (US ). Vegetable tanning extr ac ts are utili zed to produce firmer , thi cker leathers fo r ba gs, cases, strap leathers, and for sho e soles a nd heel s; tri valent basic chro mic sulfate is used to produce softer , th inner leather s fo r personal leather goods such as handbags, gloves, garments, upholstery , and the upper parts of shoes. A pproxim atel y 85 070 of all leather produced in th e Un ited States is chro me tanned (p ersonal communicati on fro m Dr R Lollar , T echnical Director of th e Leather Industrie s of America); therefo re , em ployees invol ved in producing chro med leather were the foc us of our inves tiga tio n .
A de scription of the processes (21) a nd potential hazardous chemica l exposur es (13,14) for the chrome leather tanneries th at were studied is pre sent ed in ta ble I. No previou s environme nta l analyses ha d ever been conducted a t eith er tannery. The processes in th e two tanneries ha ve rem ained relati vely the sa me since th e end o f the 19th century . Ho wever, some mechanizat ion and che mic al technology have been introduced to reduce the amount of manual operations. For example, rotating drums and hide processors, in whi ch chemicals a re o ften introduced through fixed pip es, ha ve replaced the paddle va ts for soaking , deh airing , and tanning th e hid es. Various chemical changes ha ve included , a mo ng o th ers, the use of d imeth ylamine sulfa te (DMAS) as an accelerat ing agent in th e dehairi ng process, ch lo ro phe nols as a disinfecting agent, and trivalent instead of hexavalent chromium sa lts as th e tanning material. Some of the se chemicals (eg, dim ethylamine sulfa te and the chlo rina ted phenolic

Beamh ouse
Receive, sort , Receive , sort , and t rim the perimeter t rim and sid e areas of th e hid e and split th e hide In half (from tall along the backbone to cente r of neck) Soak and wash Remove blood , di rt , manure, and exce ss salt and rest or e mo ist ure to ope n up the cont racte d fibe rs so th ey are chemically active. Dehair Destroy the hair us ing hydrated Hydr ogen sulfi de 3 0-2.0 1.3 li me contain ing accelerat ing agents N-nit rosod imethlylamine 29 0-11 b 3.5 b (eg, sulfide chemicals and (produced f rom dimethylamine sulfate) dimethylam ine sulfate) Flesh Remove excessive fl esh, fat , and mu sc le Del ime and bate Neu t rali ze the high alka linit y of th e Amm onia 2 9-30 19 biocides), however, have since been replaced in the past decade due to concern about their health consequences.
With new technologies, the increased importation of finished leather goods (mainly from Ital y), and the greater use of man-made materials, the number of workers employed in the US tanning industry has declined from a peak of 33 000 in 1965 (29) to approximately 20000 as of 1982 (38).

Subjects and methods
The study cohort included all production workers employed between I January 1940 and II June 1979 at tannery A (N = 2807) or between I January 1940 and I May 1980 at tannery B (N = 6 558). The cohort members were identified from the per sonnel records, which generally contained name, social security number, sex, race, date of birth, date of hire, date of separation, and specific work history information (ie, department, job assignment and dates of each new work assignment). No information on country of birth (emigration) was available from the employment records.
The vital status of each member of the cohort was determined as of 31 December 1982 from records of the Social Security Administration , Internal Revenue Service, or th e National Death Index (NDI) (34), among others. For each cohort member identified as deceased, a death certificate was obtained from the appropriate state vital statistics office, and underlying and contributory causes of death were coded by a qualified nosologist, according to the revision of the International Classification of Diseases (ICD) in effect at the time of death. Deceased employees for whom no death certificate could be obtained were assumed dead as of the date given by the reporting agenc y, and the cause of death was classified as unknown.
A modified life-table analysis system (41) was utilized to accumulate the person-years at risk for each study member beginning with the first day of employment at either tannery after I January 1940 until either 31 December 1982 or the date of death, whichever came earl ier. The person-years at risk were classified by sex, race, five-year age groups and calendar-time periods, employment periods, and latency periods (time periods since first employment). They were then multiplied by the appropriate sex-, race-, age-, and cause-specific mortality rates of the US population to yield the number of expected deaths . Expected mortality was similarly computed with the use of the state death rates of Minnesota for tannery A and Wisconsin for tannery B in order that po ssible variations in mortality patterns between state and national rates due to population composition and geographic location could be controlled for. For site-specific cancers, state death rates for the years 1950-1979 were utilized; for other causes, state death rates for the years 1962-1979 were used (according to the unpublished data of OM Mar ch, Mortality and Population Data System, University of Pittsburgh).
Duration of employment was used in the analysis as a surrogate for cumulative exposures since past environmental monitoring data from the tanneries were not available. Categories for analysis by duration of employment « I year, 1-9 years, and 2: 10 years) were established from an examination of the distribution of the person-years at risk . For cancers , a minimum laten cy period of 15 year s was selected since most occupationall y related cance rs usually occur after an appropriate induction-latency period (8). Observed and expected numbers of deaths were also stratified according to "ever having been employed" within certain departments of the tannery. The departments were grouped according to process description and chemical exposures as presented in table I. For most employees, potential for exposures depended upon the department (area) in which they worked , although it was realized that exposures between departments could overlap.
A standardized mortality ratio (SMR) was computed for each cau se of death by multiplying the ratio of the observed to the expected number of deaths by 100. A two-sided confidence limit (95 010) for each SMR was then calculated on the assumption of a Poisson distribution for the ob served deaths (31). When the confidence interval did not include 100, the SMR value was considered statistically significa nt. Through 31 December 1982, 95 % of the 9 365 members of the cohort were successfull y traced : 7 316 individuals (78 %) were alive, 1 582 individuals (17 % ) were deceased, and 467 individuals (5 %) were lost to follow-up and considered alive for the purposes of our study (table 2). Death certificates were obtained for 96.8 % of all deaths. Of the 9 365 individuals in the cohort, 5 827 (75.7 %) were male, and 2280 (24.3 %) were female. Approximately 18 % of the cohort was nonwhite.

Cohort description
The mean age of the cohort at hire was 27 years. More than one-half of the work ers (52 %) were hired prior to 1960. Thus the majority of the work force had a minimum of 22 years between date of hire and 1982 (the ending date of the study) . For tannery A, 21 % of the workers were employed for 10 year s or longer. For tannery B, only 8 % were employed for 10 years or longer. These percentages indicate a fairly substantial employment turnover rate at both tanneries, so mewhat typical of the US leather tanning industry.

Total deaths
The I 582 deaths observed among the workers from both tanneries combined were 89 % of the number ex-pected based upon US age-adjusted mortality rates (table 3). Statistically significant deficits of deaths from all causes combined were observed for both tannery A (SMR 81,95 070 CI 74-88) and tannery B (SMR 93, 95 % CI 88-1(0). This reduction in mortality for each tannery was due to decreased risks of death from most causes. Significant elevations in mortality, however, were found for accidents in tannery A and for alcoholism, cirrhosis of the liver, and suicide in tannery B.

Cancers
The number of deaths from all cancers combined was less than expected for both tannery A and tannery B despite whether US rates or state death rates were used for the comparison (table 4). (The discussion of the cancer results has been limited to findings based on state, rather than US, death rates since regional rates .were thought to control better for possible variations in mortality patterns.) The "all cancers" deficit was largely explained by a reduced number of deaths from cancers of the trachea, bronchus, and lung. Analysis by duration of exposure for these sub sites of cancer showed an absence of any positive correlation with excess mortality risk when examined after a time period of 15 years from first exposure, the SMR for 10 years employment being the lowest (table 5). Deficits in mortality from cancers of the trachea, bronchus, and lung were also observed within each department of the tanneries (table 6).
All other primary sites of cancer mortality were also below expectation (table 4). Specifically, for cancers of the buccal cavity and pharynx, eight deaths were observed for both tanneries combined (with 8.5 expected), all resulted after a IS-year latency period (table  • P < 0.05. c 95 % co nfid ence interval (tw o-si ded tes t).

5)
. T hree of these deaths were from cancer of the tongue, two each from can cer of th e lip and pharynx , an d o ne fro m cance r o f th e bucca l cav ity. Ninety-four deaths were due to cancers of the digestive or gan s and perito neum (with 90.2 expected) (table 4). None of the site-specific ca ncers in this gro uping was in excess. Of eight deat hs fro m cancer o f the liver, seven occur red among employee s of tannery B (with 6.9 expec ted) . All seven of th ese deaths occurred after a latency period of 15 years, and four of the seven deaths a fter IO yea rs of emp loyme nt (with 2.7 expected) (ta ble 5). T hree of the deat hs fro m liver cancer were among those employed in th e ta nyard of tannery B, a find ing which was statistica lly significant (SM R 720, 95 % CI 126-1 36 1). However, two of these deaths occ ur red amo ng emp loyees with less than o ne mo nth of em ployment at the tann ery.
Of Mo rtality fro m diseases of the lymph at ic and hemato poietic system was lower than expected for both ta nneries (ta ble 4) . Wi thin this cat egor y, deaths from bot h leu kemia an d aleuke mia and fro m ma ligna nt lymph om as were below tho se expected.
One death resu lted fro m cancer of the nasal ca vity (with 0.4 expected) in an employee with more than 18 year s' experience in the fin ishing department of tannery B.

A lcoholism and cirrhosis of the liver
Significant excesses o f mortality fro m alcoh olism (11 death s, SMR 192,95 % CI 101-344) and cirrhosis of the liver (45 death s, SMR 150, 95 % CI 110-201) were observed amo ng workers in tannery B, but not in tanne ry A (ta ble 3). Th ese statistica lly significa nt excess risks persisted when sta te rates were used as the com parison population . The majority (73 % ) of the deaths from cirrhosis occurre d among tho se employed less tha n one year at the tannery_Nine o f the 45 deaths from cirrhosis occurred among employees of the beamhou se of tan nery B (SM R 138, 95 % C I 62-203). Death cert ificate exa mination indicated tha t five of th ese nine deaths listed chronic alcoholism as a co ntr ibuto ry cause .

A ccidents
Th e excess of accide nta l death s was of borderline statistical signi ficance in tan nery A (SM R 128, 95 070 CI 100-163) but slightly less than expected in tann ery B (SMR 92, 95 070 CI 73-115) (table 3). Of the 65 accid ental deaths among the employees of tannery A , only 23 (with 22.2 expected) resulted from cau ses other th an motor vehicle accidents , and only on e (burned by flash fire) resulted from employment at the tann ery. Most of the accidental deaths occurred among tho se employed in the beamhouse of tannery A (24 cases, SMR 228, 95 % CI 146-394). However, only eight (with 4.7 expected) of these death s resulted from causes other than transportation , and non e occur red at th e tannery. Onl y 6 of the 24 death s occurred within one year of last employment at the tannery .

Su icide
Deaths from suicide were signifi cantly elevated over th ose expected fo r employees of tannery B (42 cases, SMR 171 , 95 % CI 123-232) (table 3). Half of the tanners who died from suicide worked at the tannery fo r less th an one yea r and committ ed suicide, on the avera ge, 8.5 yea rs af ter ter minatio n of employmen t. Onl y 3 of the 42 suicides occ urred within one year of last employment at the tannery.

Discussion
Th e major findin gs of this study suggest, in general, that leather tanners and finishers are no t at an increased risk for those cau ses of death that were of a priori con cern, when their mortality experience was compared to that of either th e US population or the population of the state in which the tanneries were located. The se results seem to corroborate the only other mortalit y investigation focused specifically on emplo yees of the leather tanning and finishing industry (30). Thi s other study, however, was limited by small numbers (N = 833).
The outcomes of a priori concern in our investigation of tanners and finishers included cirrhosis of the liver, cancers of the liver, kidney , lung, bladder, nasal cavity, bu ccal cavity and pharynx, and lar ynx, and lymphoma . The concern fo r the se diseases was ba sed on adverse health effects previously reported by others for the leather and leather manufacturing products indu str y (as cited later) and chemicals known to hav e been used in the tannery environment.
Studie s of empl oyees from the broader category of the leather and leather manufacturing products indus- -P < 0.05. try have shown varying results. Houton et al (18) reported increased risks for cancers of the bladder, buccal cavity and pharynx, and larynx among "operatives in the leather industry" when these workers were compared to clerical workers. In a series of case-referent analyses from the same population, DeCoufie (I I) confirmed Houton's findings and also observed an increased risk for malignant lymphomas compared to noncancer cases. Dubrow & Wegman (12) and Cole & Goldman (8) found statistically significant increased risks for bladder cancer and cirrhosis of the liver among "leather workers," who included tanners, as well as among employees involved in the cutting, assembling and buffing of leather. Increased risk of kidney cancer has been reported among "leather workers" (2,6,26), as has risks from lung cancer (17,22) and nasal cancer (10,30). Of particular interest were mortality risks associated with exposure to N-nitrosodimethylamine (NOMA). This chemical is an extremely potent carcinogen (II) that can be formed by the interaction of oxides of nitrogen and dimethylamine sulfate, a chemical used to accelerate dehairing in the tannery process. N-nitrosodimethylamine has been shown to cause cancer in a variety of organs, especially in the liver and kidney 114 (4,16), in virtually every animal species in which it has been tested (28), although the doses used far exceeded the maximum exposures tolerated by man (24,25). The carcinogenic risk to humans has yet to be established. Barnes & Magee (4) observed two cases of cirrhosis of the liver and liver damage among three men working in a research laboratory of a large industrial complex where N-nitrosodimethylamine had been used as a solvent.
In our study an elevated mortality risk was observed from cirrhosis of the liver (9 deaths, SMR 138,95 Olo CI 62-203) and kidney cancer (2 deaths, SMR 222, 95 % CI  among the beam house employees of tannery B, where dimethylamine sulfate had been used to accelerate the dehairing process. A significantly increased risk for liver cancer (3 deaths, SMR 720, 95 % CI 126-1 361) was also observed in an adjacent (tanyard) department. The airborne levels of nitrosamines found ranged from 0.1-11 jlg/m" which, while very low, appear to be typical of a tannery that uses dimethylamine sulfate. However, higher levels have been observed in another tannery (32).
Of the nine deaths from cirrhosis of the liver in our study, five of the death certificates listed chronic alcoholism as a contributing cause, and, of the three deaths due to liver cancer, two of the persons had been employed for less than one month at the tannery. Similarly, one of the two persons who died of kidney cancer had been employed for less than one month in the tannery. In addition, none of the deaths from nonmalignant diseases of the genitourinary system, including nephritis and nephrosis and kidney disease, were elevated. Therefore, the increased risks cited must be interpreted with caution.
In contradistinction to previous studies of workers in the leather and leather manufacturing products industry (10,17,22), this analysis did not find increased mortality from lung cancer or other nonmalignant respiratory diseases. In the earlier years (prior to World War II), US tanneries which tanned skins often used the two-bath tanning method in which hides were saturated with hexavalent chromium salts (potassium or sodium dichromate) and sulfuric acid and then removed manually and placed in a bath that reduced the dichromate to trivalent chromium sulfate. Exposures to hexavalent chromium compounds, some of which have been shown to be respiratory carcinogens in both animals and humans (20,27), were probable. Several explanations may account for the apparent discordance between our results and those of the previous studies. First, since the early I940s, most chrome tanneries in the United States have switched to the one-bath tanning method in which hexavalent chromium had already been reduced to trivalent chromium. Since fewer than 8 070 of our cohort had been employed prior to 1940, worker exposure to hexavalent chromium was minimal. Second, the three previous studies referenced (10,17,22) included employees from the broader occupational category of the leather and leather manufacturing products industry. These employees may have been exposed to respiratory carcinogens not normally present in the tannery environment. Third, although the tobacco smoking habits of our cohort were unknown, the two tanneries under study had strict enforcement of antismoking rules and, therefore, employees probably did not smoke more than the comparison population. The only study of leather workers which has taken smoking status into account (40) also observed a lower than expected risk from lung cancer and other respiratory diseases among tannery workers.
Cancer of the bladder was of concern because of the use of some dyestuffs derived from benzidine and betanaphthylamine in the dyeing of the chrome leather. Benzidine and beta-naphthylamine are dye intermediates which have been shown to be human bladder carcinogens (7,19,33). One of the two tanneries in our study, tannery B, showed detectable concentrations of benzidine in bulk dyes in the dye room. Of the four bladder cancer deaths that occurred in this tannery, two of the employees that died (with 1.0 expected) had once worked in the dye room. However, neither employee had worked for longer than two months at the tannery and, due to small numbers, the results could have been due to chance.
In the finishing department of tannery B, there was one death due to squamous cell carcinoma of the nasal cavity. Cancer of the nasal cavity is rare. The annual incidence for white males is approximately eight in one million in the United States, about one-half of these cases being fatal within five years. Sixty percent of the fatal cases of nasal cancer are squamous cell carcinomas. Both formaldehyde exposure and leather dust have been linked with squamous cell carcinoma of the nasal cavity in rodents (3,36) although the findings in human epidemiologic studies are still unclear (42). In the finishing department of tannery B, formaldehyde was used as a leather preservative and as a protein fixer on glazed leather, and leather dust originated from the buffing operation. The person with the single fatal case of squamous cell carcinoma of the nasal cavity in our study had worked in the finishing department for more than 18 years and died 55 years after initial employment, an induction-latency period consistent with occupational nasal carcinoma. The observation of nasal cancer among leather tanners was also recently reported in England (1 observed versus 0.21 expected) (30).
Deaths from accidental causes were found to be significantly elevated among employees of tannery A (SMR 128, 95 % CI 100-163), especially among those who had worked in the beamhouse. This finding was not unexpected since tanneries traditionally rank high in accidental injuries among all industries surveyed (39). However, only one accidental death occurred in this tannery. Almost all of the other accidental deaths resulted from fatal car or motorcycle accidents, accidental falls at home, or drownings. This result is consistent with data from the Bureau of Labor Statistics, which showed that, of 10 309 employment compensation claims filed during the years 1976-1983 for injuries in SIC 31I, only four were for fatalities (37).
Our study had several limitations which should be noted. First, although our environmental surveys, conducted during 1979 and 1980, at the two tanneries had found all exposure levels to be below OSHA standards, except for formaldehyde, historic environmental measurements had never been conducted at either tannery. Because detailed information concerning past environmental exposures was not available, we defined exposures by classifying workers according to (i) the duration of their employment, (ii) the length of time interval since their first employment (latency), and (iii) their assignments to various departments. These are crude surrogate measures of actual exposure. Had we been able to identify more precisely those members of the cohort with the heaviest cumulative exposures, we might have been able to define more precisely the effects of such exposures on mortality. Second, mortality is not always an adequate indicator of potential health risks associated with employment, 115 particularly for those diseases such as bladder cancer which are readily treatable (35). Morbidity, which may be a better measurement of potential health risk in some cases, was beyond the scope of the present study.
Third, ascertainment of vital status was only 95 070 complete. While this percentage of follow-up is well within the range generally considered acceptable for cohort mortality studies, the 5 % deficit may have had the effect of inflating person-years at risk and thus lowering the SMR values. In addition, death certificates were not recorded for 50 (3.2 %) of the I 582 known deaths. This occurrence had the effect of lowering the cause-specific SMR values by an additional 3.2 !t/ o, on the average. Fourth, although there is no particular reason to assume so, there is always the possibility that the personnel records may have been incomplete. Finally, the statistical power of this study to detect increases in mortality for diseases of a priori interest varied widely according to the background frequency of each condition (5), as shown by the confidence intervals.
In summary, although this study revealed some elevated risks for certain causes of death among tannery workers, no significantly increased risks were noted for any cause of death thought a priori to be occupationally related. Several limitations of this study, however, are discussed which may have accounted for our negative findings. Considering the limitations of this study, it would be improper to conclude that employment in the leather, tanning, and finishing industry presents no occupational health risks.