1.1. Historical overview

Dry cleaning is believed to have originated in France in 1825: The French Federation of Dyeing and Cleaning gave an account of a servant in the household of Jean-Baptiste Jolly in Paris in 1825 who spilled the contents of a lamp on a soiled tablecloth. When the tablecloth dried, the spots had disappeared. The liquid in the lamp was probably camphene, produced from pinene, which is the main ingredient of turpentine. Camphene was initially used as a spotting agent, but it was later used for complete cleaning by hand. Use of volatile organic liquids to clean cloths became known as ‘French cleaning’ in Scotland and ‘chemical cleaning’ in Germany (Michelsen, 1957). Eventually, such processes were called ‘dry cleaning’ because they do not involve the use of water.

The common cleaning practice of early dyers and cleaners was to scour, using soap and water, and to add various acids and alkalis to the rinse. In the 1840s, garments were sometimes sponged inside and out with water and ox gall; dark clothes were sponged with water and logwood. A formulation for removing grease spots comprised fuller's earth, French chalk, yellow soap, pearlash and turpentine. When white fabrics became too soiled to be cleaned, they were dyed a darker colour. The early dry cleaning operation consisted of four steps: dusting (with rattan whips and brushes), hand brushing, solvent washing and drying (Michelsen, 1957).

Camphene was later replaced by benzene soap (1 lb soap for 2 gallons of benzene [60 g/L]), benzene (see IARC, 1987a), benzine (light petroleum naphtha), benzoline (a low-boiling subfraction of benzine), naphtha (see IARC, 1989a) and gasoline (see IARC, 1989b). The agents used early on for removing persistent spots on fabric (spotting agents) were benzine, turpentine, water, alcohol, cream of tartar, diethyl ether, chloroform (see IARC, 1987b), ox gall, egg yolks, boiling milk, honey, fuller's earth and spirits of wine. Attempts were made to provide mechanical agitation by adding sawdust, crushed marble and bran; however, use of these substances was soon discontinued because they were difficult to remove from the folds of the fabric (Michelsen, 1957).

The main hazard associated with the increasing use of volatile organic solvents in dry cleaning was fire. In 1928, W.J. Stoddard, president of the United States National Institute of Drycleaning (now the International Fabricare Institute) introduced Stoddard solvent, a petroleum-based solvent (see IARC, 1989c; Reich & Cormany, 1979) that was less inflammable and odour-free, which rapidly gained acceptance in the industry. Around 1917, many companies used ‘cleaners’ naphtha', which was similar to the light naphtha used in the paint industry. Subsequently, other petroleum solvents were marketed, including mineral turpentine, mineral spirits and oleum spirits. Use of these materials in the war effort in the early 1940s greatly curtailed their availability for dry cleaning (Michelsen, 1957).

Carbon tetrachloride (see IARC, 1987c) was the first chlorinated hydrocarbon solvent used for dry cleaning. It was introduced in the United States of America and Europe as an alternative to the expensive petroleum-based solvents, and by 1940 it was estimated that 20 thousand tonnes were used annually for dry cleaning in the United States. Widespread use of carbon tetrachloride in this context was discontinued in the 1950s because of its toxicity and corrosiveness, and tetrachloroethylene (see monograph, this volume) became the most widely used solvent. Trichloroethylene (see monograph, this volume) was also used, but on only a limited basis because it caused bleeding of many acetate dyes (Reich & Cormany, 1979).

Figure 1 shows the evolution of solvent use in dry cleaning over time.

Fig. 1.

Use of solvents in dry cleaning over time

1.2. Description of the industry

Cleaning is the process of removing dirt from fabric through the use of aqueous and nonaqueous solvents. Aqueous solvents swell the hydrophilic textile fibres, so that the dimensions of the fibre change, which can cause wrinkles or shrinkage of the fabric. Nonaqueous dry cleaning solvents do not swell the fibres and can remove oily stains at low temperatures, whereas a high-temperature, colloidal suspension is needed in aqueous (laundering) processes (Reich & Cormany, 1979).

Dry cleaning is a batch process that involves washing fabrics in a solvent solution, extracting the solvent from the fabrics during a spin cycle and tumble-drying the damp fabrics with warm air. In many commercial facilities where tetrachloroethylene is used, washing and extraction are done in a single machine. The damp fabric is then transferred manually to a tumbler for drying. This is known as the ‘transfer’ process (Stricoff, 1983).

A more recent process, in which all three steps are carried out in one machine, is referred to as the ‘dry-to-dry’ process. After the wash cycle, the solvent is drained, and residual solvent is removed by centrifugal force. This method has eliminated the handling of solvent-laden fabrics. All coin-operated dry cleaning equipment is dry-to-dry, while industrial processes may be either transfer or dry-to-dry (Stricoff, 1983).

Dry cleaning involves three basic steps: washing, solvent extraction and drying. The common tetrachloroethylene-based process is described below (Reich & Cormany, 1979).

1.3. Exposures in the workplace

1.3.1. Current practices

Today, commercial dry cleaners make up the largest segment of the industry (by number of establishments and quantity of garments cleaned). These firms handle primarily clothing and household products. Coin-operated dry cleaning machines are used directly by consumers for cleaning clothes and household furnishings. Industrial dry cleaners supply businesses, industrial plants and institutions with high-volume cleaning of uniforms, towels, linen and cleaning cloths, which are sometimes heavily soiled with industrial chemicals, oil and other materials.

In Europe, tetrachloroethylene comprises about 90% of the solvents used in dry cleaning; the main chlorofluorocarbon used is 1,1,2-trichloro-1,2,2-trifluoroethane (CFC-113) (Vieths et al., 1988). There are an estimated 6500 dry cleaning establishments in the United Kingdom (Edmondson & Palin, 1993). In 1985, there were an estimated 2000 dry cleaners in Finland (Rantala et al., 1992).

In 1970, there were 2886 laundries and dry cleaning shops in Denmark. White spirit (a petroleum-based solvent) was the most commonly used solvent in these establishments after the Second World War, but consumption decreased when automatic machines were introduced in the late 1950s. Tetrachloroethylene then became the most prevalent solvent, supplemented by trichloroethylene for the cleaning of work clothes and the fluorocarbons trichlorofluoromethane (CFC-11) and CFC-113 for cleaning fur coats (Lynge & Thygesen, 1990).

In Switzerland in 1991, 25% of dry cleaning shops used CFC-113 (Grob et al., 1991). In France, 71 spotting agents used in dry cleaning shops were analysed. Benzene was not found; toluene (see IARC, 1989d) was found in 7% of the shops, xylenes (see IARC, 1989e) in 34%, methanol in 3% and chloroform in 1%. Chlorofluorocarbons were used in four of 30 shops studied and tetrachloroethylene in 26 (Davezies et al., 1983).

In Japan, both dry cleaning and laundering are done in dry cleaning shops. Of the dry cleaning machines used, 70% (35 500) use petroleum-based solvents, 20% (10 700) use tetrachloroethylene, 5% (1900) use CFC-113 and 5% (1800) use 1,1,1-trichloroethane. In 1989, the industry consumed 7700 tonnes of petroleum solvent, 18 000 tonnes of tetrachloroethylene, 2700 tonnes of 1,1,1-trichloroethane and 4100 tonnes of CFC-113 (Kubota, 1992), In Nagoya, Japan, 50% of dry cleaning shops were reported to use petroleum solvents, 46% tetrachloroethylene, 4% CFC-113 and 2% a combination of tetrachloroethylene and petroleum solvents (Takeuchi et al., 1981). Almost all of the machines involving use of petroleum-based solvents are transfer types, whereas in those that use the other three types of solvents the equipment is dry-to-dry. There are relatively few coin-operated machines in operation (Kubota, 1992).

By 1991, tetrachloroethylene was used in an estimated 28 121 dry cleaning plants in the United States, including 24 947 commercial plants and 130 industrial and 3044 coin-operated establishments (United States Environmental Protection Agency, 1991a). About two-thirds had dry-to-dry machines and one-third, transfer machines. Of commercial dry cleaning establishments in the United States in 1991, 82% used tetrachloroethylene, 14% used Stoddard solvent, 3% used CFC-113 and less than 1% used 1,1,1-trichloroethane (United States Environmental Protection Agency, 1991b; Earnest et al., 1994). Of industrial cleaners, about 50% used tetrachloroethylene and 50% used Stoddard solvent. Coin-operated facilities used only tetrachloroethylene (Linak et al., 1992). In Oklahoma, about 50% of dry cleaning involves use of Stoddard solvent (Duh & Asal, 1984). In Canada, 90% of commercial dry cleaning is based on tetrachloroethylene (Rice & Weinberg, 1994).

In December 1991, the United States Environmental Protection Agency (1991c) began regulating tetrachloroethylene as a hazardous air pollutant under Section 112 of the Clean Air Act. Regulatory action with regard to tetrachloroethylene is also being taken on an international level (Linak et al., 1992). The international dry cleaning industry has responded with increased research into alternative solvents (Kirschner, 1994) and cleaning methods: a shift from transfer machines to closed-loop dry-to-dry machines and innovations in vapour recovery equipment and other devices to reduce occupational exposures and environmental emissions. Many of the exposure problems identified during studies in the late 1970s and early 1980s still exist, however, because transfer machines continue to be used, many of the controls that have been developed are prohibitively expensive and some work practices are inadequate.

The several distinct job titles and duties involved in dry cleaning are summarized in Table 1. It is estimated that several million people are employed in dry cleaning worldwide.

Table 1.

Job titles and duties in the dry cleaning industry.

1.3.3. Measurements of exposure

Occupational exposure to tetrachloroethylene has been measured in the dry cleaning industry throughout the world (Table 2). Because tetrachloroethylene has been the most commonly used solvent during the last two or three decades, when most of the measurements were made, only data on that compound are included in Table 2. Data on other exposures to tetrachloroethylene, for example, in the vicinity of dry cleaning establishments and in dry cleaned clothes carried home, are included in the monograph on tetrachloroethylene (this volume).

Table 2.

Exposures to tetrachloroethylene in dry cleaning shops.

Differences in personal exposures to tetrachloroethylene between plants and shops are often many times larger than the differences between machine operators and other staff within dry cleaning premises. Such differences between staff have decreased in time from a factor of roughly 6 in the early 1980s to a factor of 2 in the early 1990s. As can be seen from Table 2, this coincides with decreasing average exposure levels, from 50–100 ppm [339–678 mg/m³] in the 1970s to 10–50 ppm [67.8–339 mg/m³] in the 1980s. Applying fixed multipliers for job titles in the calculation of cumulative exposure to tetrachloroethylene will introduce severe misclassification when factors that modify exposures in specific plants and shops are not taken into account.

In a study of dry cleaners and their families in Italy, a significant correlation was observed between the concentration of tetrachloroethylene in alveolar air samples collected at the end of the work day and 8-h time-weighted average (TWA) levels in ambient air (r = 0.758; p < 0.001; n = 49). A similar relationship was found between tetrachloroethylene concentrations in alveolar air samples collected from dry cleaners at home (n = 33), about 2 h after the end of the exposure, and the 8-h TWA concentration during the work day (r = 0.665; p < 0.001). Alveolar samples collected the next day still showed a significant correlation with the 8-h TWA value of the day before (r = 0.549; p < 0.001). The concentrations of tetrachloroethylene in alveolar air in this population were 0.49–353 mg/m³, with a geometric mean of 15.4 mg/m³ (Aggazzotti et al., 1994).

In a study in Japan, the personal TWA levels of tetrachloroethylene to which dry cleaning workers were exposed were 0.6–100.8 ppm [4.07–683 mg/m³]. The concentrations were 0.3–87 ppm [2.03–590 mg/m³] in expired air and 0.01–0.73 mg/L in blood; the level of total trichlorinated compounds in urine was 0.6–19.2 mg/L (Hayashi et al., 1990). In another study in Japan, the average annual consumption of solvents in dry cleaning shops was estimated to be 1280 kg of petroleum solvent, 1450 kg of tetrachloroethylene and 275 kg of CFC-113. The concentration of total trichloro compounds in urine samples was 3.2 ± 2.0 mg/L in the morning and 2.6 ± 1.9 mg/L in the afternoon (Takeuchi et al., 1981).

In one dry cleaning shop in Switzerland, the concentrations of tetrachloroethylene and CFC-113 were 10 mg/m³ and 150 mg/m³ inside the shop and 25 mg/m³ and 1000 mg/m³ in the machine room (Grob et al., 1991).

Analysis of 138 blood samples taken from workers in 55 dry cleaning shops in Finland in 1987 showed a geometric mean tetrachloroethylene concentration of 0.75 µmol/L (range, < 0.1–14.0 µmol/L) [0.12 (0.02–2.3) mg/L]; the arithmetic mean was 1.25 µmol/L [0.21 mg/L] (Rantala et al., 1992).

In a study conducted by the United States National Institute for Occupational Safety and Health in 1988, occupational exposures to tetrachloroethylene and trichloroethylene in dry cleaning and exposure to 2-butoxyethanol, hexylene glycol, n-butyl acetate and methyl isobutyl ketone during spotting were evaluated. The concentrations of spotting agents in air were low: 2-Butoxyethanol was not detectable (n = 10), and the mean concentrations were 1.2 ppm [5.8 mg/m³] (range, 0–6.9 ppm [0–33 mg/m³]; n = 17) for hexylene glycol, 0.32 ppm [1.3 mg/m³] (range, 0–4.2 ppm [0–17.2 mg/m³]; n = 57) for methyl isobutyl ketone and 1.2 ppm [5.7 mg/m³] (range, 0–5.69 ppm [0–27 mg/m³]; n = 57) for n-butyl acetate. Spotters were exposed to mean concentrations of 4.3 ppm [29 mg/m³] (range, 0–38.7 ppm [0–262 mg/m³]; n = 115) tetrachloroethylene and 1.4 ppm [7.5 mg/m³] (range, 0–9.16 ppm [0–49 mg/m³]; n =115) trichloroethylene in air (Earnest, 1994).

In Japan, dry cleaned test samples of fabric were found to contain up to 13.6 mg/g tetrachloroethylene. In 53 air samples from 11 establishments where dry cleaned clothes were handled, the mean concentration was [434 µg/m³] (range, 1–4813 µg/m³). The mean level in alveolar air from 31 dry cleaners was [37.8 µg/m³] (average, 36.9; range, 0–168 µg/m³) (Kawauchi & Nishiyama, 1989).

The average concentration of tetrachloroethylene to which 22 workers were exposed in six dry cleaning shops in Belgium was 21 ppm [142 mg/m³], with a range of 9–38 ppm [61.0–258 mg/m³]. The mean concentration of tetrachloroethylene in alveolar air was 1.9 ppm [13 mg/m³] before work and 5.1 ppm [35 mg/m³] 30 min after exposure. The mean concentrations in the blood at the same time intervals were 0.4 and 1.2 mg/L, respectively. Trichloroacetic acid was not detected in any urine samples. The authors concluded that a blood concentration of < 1 mg/L tetrachloroethylene 16 h after exposure probably corresponded to an 8-h time-weighted average exposure of < 50 ppm [339 mg/m³] (Lauwerys et al., 1983).

In a study of 195 individuals in one industrial and 12 commercial dry cleaning facilities in Detroit, MI, United States, in 1986–87, the concentrations of tetrachloroethylene in 181 samples of exhaled breath and 54 personal samples were measured during use of dry-to-dry and transfer machinery. In facilities where transfer machines were used, the mean concentration in breath samples was 7.72 ppm [52.3 mg/m³] (range, 2–14 ppm [12–95 mg/m³]), which was over five times higher than that measured in facilities where dry-to-dry machinery was used (1.47 ppm [9.97 mg/m³]). The mean concentration in air samples from facilities using transfer machines (29.5 ppm [200 mg/m³]; range, 1–58 ppm [7–100 mg/m³]) exceeded that for dry-to-dry machines (7.09 ppm [48.1 mg/m³]) by more than four times. The mean concentrations of tetrachloroethylene in breath and air samples (ppm [mg/m³]) for various job titles were, respectively: operator, 12.5 [84.8] and 46.5 [315]; presser, 5.88 [39.9] and 14.8 [100]; clerk, 4.52 [30.6] and 11.7 [79.3]; seamstress, 6.50 [44.1] and 19.0 [129]; managerial., 5.50 [37.3] and not reported; other, 2.55 [17.3] and 5.00 [33.9]. Of 12 air samples taken from the breathing zone of operators, eight contained > 25 ppm [170 mg/m³], 6 contained > 50 ppm [339 mg/m³] and one contained > 100 ppm [678 mg/m³]. None of the air samples from the four shops where only dry-to-dry machinery was used had concentrations > 25 ppm [170 mg/m³] (Solet et al., 1990). In the same population of dry cleaners, the ratios of total urinary protein, albumin and n-acetylglucosaminidase to creatinine in 192 urine samples were studied in relation to tetrachloroethylene levels in breath and estimates of chronic exposure; no consistent relationship was seen (Solet & Robins, 1991).

In a study in Croatia of 19 dry cleaners exposed to tetrachloroethylene for a mean duration of 16.6 years (range, 2–30), the concentrations of tetrachloroethylene and trichloroacetic acid in blood on a Monday before work were 2.96 (1.63–32) and 6.5 (1.65–17.1) µmol/L [0.5 (0.3–53) and 1.1 (0.3–2.8) mg/L], and those on a Thursday after work were 8.44 (4.8–80) and 8.57 (1.7–20.9) µmol/L [1.4 (0.8–13.3) and 1.4 (0.3–3.4) mg/L], respectively. The concentrations of trichloroethanol and trichloroacetic acid in urine were 0.03 (0–1.7) and 2.3 (0.5–15.8) µmol/mmol [0.04 (0–2.3) and 3 (0.7–23) mg/g] creatinine on a Monday before work, and 0.13 (0–0.95) and 2.3 (0.8–10.8) µmol/mmol [0.17 (0–1.3) and 3.3 (1.2–15.6) mg/g] creatinine on a Thursday after work (Skender et al., 1986).

In a study in the United States (Petreas et al., 1992), the mixed exhaled air of dry cleaning machine operators contained a mean concentration of 3.4 ppm [23.1 mg/m³] tetrachloroethylene at the end of the work shift and 2.2 ppm [14.9 mg/m³] the next morning; the mean concentrations for a presser were 3.3 and 1.4 ppm [22.4 and 9.49 mg/m³], respectively. Exhaled air of operators using dry-to-dry machines contained 2.4 ppm [16.3 mg/m³] at the end of the work shift and 1.5 ppm [10.2 mg/m³] the next morning; the exhaled air of transfer machine operators contained 5.4 and 3.1 ppm [36.6 and 21.0 mg/m³], respectively. The levels of tetrachloroethylene in 12 blood samples were 17–1154 µg/L, with a mean of [283 µg/L]. The mean blood:breath concentration ratio was thus 21.6, similar to that of 23 derived by Monster et al., (1979). Petreas et al., (1992) found strong correlations between tetrachloroethylene concentrations in mixed exhaled air and in blood (r = 0.94) and in mixed exhaled air and personal air samples collected over the entire shift (r = 0.89).

2.1. Case reports

A man and four of his five children, all of whom worked in dry cleaning in the United States, developed chronic lymphocytic leukaemia (Blattner et al., 1976).

Jalihal and Barlow (1984) reported a case of myeloid leukaemia diagnosed in a 60-year-old dry cleaner in the United Kingdom. He had had heavy exposure for many years, first to trichloroethylene and later to tetrachloroethylene.

2.2. Descriptive studies

Guralnick (1963) used death certificates to evaluate the cause-specific mortality by occupation for all men aged between 20 and 64 who had died in the United States during 1950. There were 1193 deaths among laundry or dry cleaning operatives. Nonsignificant excesses of cancer mortality were seen, both overall (standardized mortality ratio [SMR], 1.4) and for cancers at specific sites, including the stomach (1.5), intestine and rectum (1.6) and trachea, bronchus and lung (1.5).

Gallagher et al., (1989) used death certificates to study cause-specific mortality by occupation for all men aged 20–65 in British Columbia (Canada) for the years 1950–84. There were 165 deaths among launderers and dry cleaners, and significantly elevated proportionate mortality ratios (PMRs) were observed for cancer of the lip (PMR, 53; 95% confidence interval [CI], 6.4–100, based on two deaths) and for all tumours of the digestive organs (2.0; 1.3 –3.0; 23 deaths), especially the stomach (2.5; 1.1 –4.9; eight deaths).

Milham (1992) analysed information on occupation and cause of death from the death certificates of 1652 male laundry and dry cleaning operatives who died in Washington State, United States, between 1950 and 1989. Elevated PMRs were seen for cancers of the buccal cavity and pharynx and connective tissues.

Data from death certificates and censuses for Great Britain for the years 1979–80 and 1982–83 showed that mortality among men aged 20–64 and single women aged 20–59 [data for married women were not reported] who were launderers, dry cleaners or pressers was higher than that in the general population (males: SMR, 1.3 [95% CI, 1.2–1.5]; single females: 1.4 [1.1–1.7]) (Office of Population Censuses and Surveys, 1986). There was a significant overall excess of malignant neoplasms in males (1.4 [1.1–1.7]), and excesses were reported for cancers of the oesophagus (1.7 [0.5–4.5], four deaths), stomach (2.0 [0.9–3.6], 10 deaths), trachea, bronchus and lung (1.7 [1.2–2.3], 41 deaths) and urinary bladder (2.8 [0.9–6.5], five deaths). In single women, there was no overall excess of malignant neoplasms (SMR, 1.0 [0.7–1.5]), but some excess was reported for cancers of the digestive organs (1.4 [0.8–2.9], seven deaths), rectum (3.3 [0.7–9.7], three deaths) and bladder (7.3 [ 1.2–36], two deaths).

A proportionate mortality study in Wisconsin, United States, was based on the death certificates of 671 female laundry and dry cleaning workers in the period 1963–77 (Katz & Jowett, 1981). Comparisons were made with all occupations as well as with lower-wage occupations obtained from a data file of 66 230 deaths in the State among white women who had ever been employed. The PMRs given below are those obtained from comparisons with lower-wage occupations. Elevated risks were found for some cancers and for diabetes mellitus. For cancer of the kidney, the PMR was 2.5 [95% CI, 1.0–5.2], based on seven deaths; types of cancer associated with nonsignificantly increased PMRs included cancer of the skin (2.6 [0.73–6.8], four deaths), lymphosarcoma (1.8 [0.65–3.8], six deaths), urinary bladder cancer (1.9 [0.62–4.5], five deaths), cervical cancer (1.4 [0.68–2.6], 10 deaths) and rectal cancer (1.3 [0.45–2.7], six deaths). There was no overall excess of neoplasms.

The causes of death of 440 laundry and dry cleaning workers in Oklahoma, United States, were investigated in a study which covered the years 1975–81 (Duh & Asal, 1984). Petroleum solvents accounted for > 50% of the solvents used in the State. The distribution of deaths in the United States in 1978 by age, sex, race and cause was used as the standard of comparison. Some SMRs were decreased, most notably that for ischaemic heart disease (0.8; 95% CI, 0.7–1.0), based on 134 deaths. For cancers at all sites, the standardized mortality odds ratio was 0.9 (0.7–1.2), based on 97 deaths. Cancer of the respiratory system was the cause of death for 39 persons (standardized mortality odds ratio, 1.8; 95% CI, 1.3–2.5), of whom 37 died of lung cancer (1.7; 1.2–2.5). For renal cancer, the standardized mortality odds ratio was 3.8 (1.9–7.6), based on seven deaths. There were two deaths from cervical cancer (1.3; 0.3–5.3), five from cancer of the ovary (1.5; 0.6–3.6), one from urinary bladder cancer (0.4) and one from breast cancer (0.1).

A further report on the Oklahoma dry cleaners was published as an abstract (Petrone, 1988). PMRs were based on the mortality of 474 white male dry cleaners whose exposure histories were obtained from plant records. The men had been exposed to petroleum solvents for an average of 10.5 years and had had no known exposure to ‘synthetic solvents’. For all malignant neoplasms, the PMR was 1.0 (95% CI, 0.83–1.3), based on 84 deaths. For respiratory cancer, the PMR was 1.5 (1.0–2.1), based on 35 deaths; no relationship was seen with latency or degree of exposure. A PMR of 2.0 (0.89–3.7) was obtained for pancreatic cancer. The PMR for renal cancer was 2.0 (0.66–4.7), based on five deaths; when only men exposed to petroleum solvents were analysed, the PMR was 1.1 (0.13–3.9), based on two deaths. [The degree of overlap between this study and that of Duh & Asal (1984) is unknown.]

Nakamura (1985) conducted a proportionate mortality study of members of the All-Japan Laundry and Dry-cleaning Association. Trichloroethylene was reported to be the solvent used most commonly for dry cleaning in Japan between 1903 and 1940; more recently, 30% of dry cleaning was stated to be done using tetrachloroethylene and 65% with petroleum solvents. Death certificates were obtained for 1711 members who had died between 1971 and 1980, and expected numbers of deaths were calculated from mortality rates for Japanese males and females in 1975. When the results for men and women were combined, there was an increased risk for cancer of the bone ([PMR, 2.9; 95% CI, 0.9–6.9]; five deaths).

2.3. Cohort studies

A union of dry cleaning workers in Missouri, United States, had 11 062 members between 1945 and 1978, of whom 5790 had held membership for one year or more. After exclusion of 425 members for whom information on race, sex or date of birth was unavailable, the analysis was restricted to 5365 members (Blair et al., 1986, abstract; Blair et al., 1990). The members were followed-up from entry to the Union or 1 January 1948 (whichever came later) until 1 January 1979: follow-up was 88% successful, with slight variations for men and women and for race. The cohort contributed 98 818 person-years, and the expected numbers of deaths were calculated from national rates. The mortality rate was slightly lower than expected for all causes combined (SMR, 0.9; 95% CI, 0.9–1.0; 1129 deaths) but slightly raised for cancer (1.2; 1.0–1.3; 294 deaths). Excesses were found for cancers of the oesophagus (SMR, 2.1; 1.1–3.6; 13 deaths), larynx (1.6; 0.3–4.7; three deaths), lung (1.3; 0.9–1.7; 47 deaths), cervix uteri (1.7; 1.0–2.0; 21 deaths), bladder (1.7; 0.7–3.3; eight deaths) and thyroid (3.3; 0.7–9.8; three deaths) and for lymphosarcoma and reticulosarcoma (1.7; 0.7–3.4; seven deaths) and Hodgkin's disease (2.1; 0.6–5.3; four deaths). There was no excess of cancer of the liver or of the kidney. The excess risk for cancer of the oesophagus was restricted to black men (SMR, 3.5; 11 deaths). The relative risk for cancer at this site for all cohort members was related to estimated cumulative exposure to dry cleaning solvents and was 2.8 for those in the highest category of exposure. For deaths from causes other than cancer, an SMR of 2.0 (95% CI, 1.1–3.4) for emphysema was the most relevant finding. With the highest level of exposure to dry cleaning solvents, there were five deaths from lymphatic and haematopoietic malignancies, whereas 1.3 were expected (SMR, 4.0 [1.2–90]). The authors stated that the relative risks were similar for workers first employed before 1960 and those employed after that date, when use of tetrachloroethylene became predominant, but no details were provided.

Proportionate mortality among the same workers, reported in a previous study (Blair et al., 1979), showed excess numbers of deaths from cancers of the liver, cervix and skin.

An updating of an earlier cohort mortality study in the United States (Kaplan, 1980; Brown & Kaplan, 1987) from California, Illinois, Michigan and New York, included 1701 dry cleaning workers in four labour unions (Ruder et al., 1994). The inclusion criteria were employment for at least one year before 1960 in a shop where tetrachloroethylene was the primary solvent used and there was no known exposure to carbon tetrachloride. A survey in 1977–79 (Brown & Kaplan, 1987) showed geometric mean, time-weighted average concentrations of tetrachloroethylene in the range of 3–22 ppm [20.3–149 mg/m³]; other solvents used for spot cleaning were not detected in the samples. In the analyses, two subcohorts were defined: people employed only in shops where tetrachloroethylene was the primary solvent, and people whose work also involved exposure to other solvents. The mean duration of employment in dry cleaning through 31 December 1990 was 6.4 years for those exposed only to tetrachloroethylene and 11.4 years for those also exposed to other solvents; the latter group had a mean duration of 6.0 years' exposure to tetrachloroethylene. Expected numbers of deaths were calculated from the national death rates. The follow-up reached 94% of the women and 96% of the men, contributing 47 273 person-years. There were 769 deaths (SMR, 1.0; 95% CI, 0.94–1.1) and 209 cancer deaths (1.2; 1.1–1.4). The SMR for urinary bladder cancer was significantly increased (2.5; 1.2–4.8; nine deaths), as were those for oesophageal cancer (2.1; 1.0–3.9; 10 deaths) and cancers of the colon and small intestine (1.6; 1.0–2.3; 26 deaths). Some excess of pancreatic cancer was noted (1.7; 0.93–2.8; 15 deaths). When the analysis was restricted to workers with a 20-year latency since first employment and with a length of employment ≥ 5 years, the SMRs were increased for cancers at all sites combined (1.5; 1.2–1.9; 83 deaths) and, notably, for cancers of the oesophagus (5.4; 2.3–11; eight deaths) and the urinary bladder (6.5; 2.8–13; eight deaths). Some overall excess was also seen for cervical cancer (1.8; 0.86–3.3; 10 deaths) and for renal cancer (1.5; 0.40–3.7; four deaths), but there was no association with time since first employment or length of employment. There were also three cancers of the tongue (3.5; 0.73–10). Only one death from cancer of the liver and biliary tract was found [with 4.8 expected].

A Danish cohort of 8567 women and 2033 men aged 20–64 years and employed in laundry and dry cleaning in 1970 was followed-up for 10 years (Lynge & Thygesen, 1990). One-fourth of the cohort was estimated to work in dry cleaning, but at the individual level there was no way of distinguishing those who worked in dry cleaning from those who worked in laundries. The dry cleaners in the cohort were mainly exposed to tetrachloroethylene after the late 1950s, but had also been exposed to trichloroethylene and chlorofluorocarbons CFC-11 and CFC-113. In measurements taken in 1979–80, the concentrations of tetrachloroethylene in clothes being taken out of the machines were 1000–7000 ppm [mg/kg] in some samples but were usually less than 100 ppm. The expected numbers of cases were calculated on the basis of the rates for the economically active proportion of the general population. A total of 510 cancer cases were observed [standardized incidence ratio (SIR), 1.0; 95% CI, 0.9–1.1]. For colon cancer, the SIR was [1.0] ([0.70–1.4]; 35 observed); for lung cancer, the SIR was [1.2] ([0.9–1.6]; 60 observed). Seven cases of liver cancer were observed [SIR, 2.2; 95% CI, 0.9–4.5], and there were 22 cases of pancreatic cancer (1.7; 1.1–2.6).

When the study was updated for the period 1981–87, a total of 10 additional cases of liver cancer were observed (4.5 expected) (Lynge, 1994).

The Swedish population at the time of the 1960 census was followed-up for cancer incidence, first for 1961–73 (Malker & Weiner, 1984) and later for 1961–79 (McLaughlin et al., 1987). In the first follow-up, data on the following cancers were reported for laundry and dry cleaning workers: all cancers (SIR, 1.0 [95% CI, 0.9–1.1]; 646 observed) and cancers of the buccal cavity and nasopharynx ([2.9; 1.7–4.7]; 16 observed), liver and biliary passages (1.2 [0.7–1.8]; 17 observed), lung (1.2 [0.9–1.7]; 34 observed) and breast (0.8 [0.7–1.0]; 113 observed). In the second follow-up, the only site for which data were reported for this occupational group was the kidney: the SIR was 0.99 ([0.6–1.6] 18 observed) in men and 0.86 ([0.6–1.3] 25 observed) in women.

Table 3 gives the details of these cohort studies. Results are shown for all cancer sites for which at least one of the cohort studies showed a significantly increased risk. In addition, results are shown for cancer sites on which a relevant case–control study was conducted (see Table 4).

Table 3.

Results of cohort studies of dry cleaners.

Table 4.

Risks for cancer associated with dry cleaning work in case–control studies of people with a variety of exposures and occupations.

2.4. Case–control studies

Dry cleaning (both in association with and independently of work in laundries) has been evaluated for possible associations with several types of cancer in case-control studies. Risks were investigated in relation to a variety of exposures and occupations and were not specifically concerned with dry cleaners. The studies addressed cancers of the colon (Fredriksson et al., 1989), liver (Stemhagen et al., 1983; Austin et al., 1987; Suarez et al., 1989), lung (Brownson et al., 1993), kidney (Asal et al., 1988; McCredie & Stewart, 1993; Mellemgaard et al., 1994), pancreas (Mack et al., 1985), oral cavity and pharynx (Huebner et al., 1992) and non-Hodgkin's lymphoma (Blair et al., 1992, 1993). Data from the National Bladder Cancer study of the United States National Cancer Institute conducted in 1978, in relation to dry cleaners, were used in four studies (Silverman et al., 1983; Schoenberg et al., 1984; Smith et al., 1985; Silverman et al., 1989). A study of multiple sites was reported by Siemiatycki (1991).

A record-linkage study from Denmark on parental employment at the time of conception and the risk of cancer in offspring mentions an increased risk for cancers at all sites combined among children of mothers who owned a laundry or dry cleaning establishment (odds ratio, 3.7; p < 0.01; 6 observed) (Olsen et al., 1991).

Table 4 summarizes the results of the case-control studies. [The Working Group noted that positive results may have been reported preferentially.]

3.1. Absorption, distribution, metabolism and excretion in humans

Studies of the toxicokinetics of tetrachloroethylene are described in the relevant monograph (this volume).

3.2. Toxic effects in humans

Slight renal changes have been reported in workers exposed to tetrachloroethylene in dry cleaning shops in several studies. In a multi-centre cross-sectional investigation in several European countries, of nine men and 41 women (mean age, 41 years; range, 17–65) who had an average of 10 years of exposure to tetrachloroethylene (median air concentration, 15 ppm [102 mg/m³]), high-relative-molecular-mass proteinuria and increased release of laminin fragments, fibronectin and glycosaminoglycans indicated slight nephrotoxicity. In addition, shedding of epithelial membrane components from tubular cells into the urine was observed. No such renal changes were observed in 50 blood donors matched by sex, age, smoking habits, alcohol consumption and use of medication. The presence of several other possible markers of renal damage, including urinary retinol-binding protein and β₂-microglobulin, was not increased (Mutti et al., 1992). [The Working Group noted that the period of employment was not taken into consideration.]

Biochemical markers of kidney damage were studied in 16 female dry cleaning workers in the former Czechoslovakia who were chronically exposed to tetrachloroethylene and compared with those in 13 women with only administrative functions and no known exposure to organic solvents. The concentration of tetrachloroethylene in the breathing zone of the dry cleaning workers was 157 mg/m³ (time-weighted average exposure; range, 9–799 mg/m³). Urinary excretion of lysozyme was 0.15 mg/g creatinine (range, 0–1.49) in the controls and 0.56 rng/g creatinine (range, 0–4.16) in the 16 exposed subjects, whereas there were no differences in the urinary excretion of albumin, (β-microglobulin, lactate dehydrogenase, total protein or glucose. Moreover, no correlation was observed between the level of tetrachloroethylene in air and bio-chemical parameters of renal function (Vyskocil et al., 1990).

In contrast, several other studies found no relationship between exposure to tetrachloroethylene in dry cleaning and renal damage. Solet and Robins (1991) obtained histories of exposure and medical conditions and samples of urine and breath from 192 dry cleaning workers in the United States. No association was found between exposure and parameters indicative of renal damage. In a study of 22 subjects exposed to tetrachloroethylene in six dry cleaning shops in Belgium (Lauwerys et al., 1983), exposure was assessed by monitoring air in the breathing zone, by urinary analysis for trichloroacetic acid and by analysis of expired air and venous blood for tetrachloroethylene. The results were compared with those for 33 subjects who were not occupationally exposed to organic solvents. The time-weighted average exposure to tetrachloroethylene was 21 ppm [142 mg/m³] (range, 9–38 ppm [61.0–258 mg/m³]). The urinary concentrations of albumin, (β₂-microglobulin and retinol-binding protein were similar in the exposed and control workers.

3.3. Reproductive and prenatal effects in humans

3.4. Genetic and related effects in humans

4.1. BACKGROUND

The process of cleaning fabrics with nonaqueous liquids is believed to have begun in France in 1825. The process has evolved into an industry called ‘dry cleaning’. ‘Camphene’ (turpentine) was used initially; in the late 1800s, benzene, benzene soap, naphtha and gasoline began to be used. In the 1920s, Stoddard solvent (mineral spirits or white spirits) was introduced in the United States in order to minimize the fire hazards associated with use of the more volatile hydrocarbon-based solvents. Carbon tetrachloride, the first chlorinated solvent used for dry cleaning, was introduced because of the high cost of petroleum solvents and was widely used until the 1950s. Its use was discontinued because of its toxicity and corrosiveness. Trichloroethylene was introduced in the 1930s. It is still used to a limited extent in Europe and in industrial cleaning plants throughout the world, but it has had a limited market in dry cleaning in the United States because of its incompatibility with acetate dyes.

Use of tetrachloroethylene began to increase in the 1940s, and by the late 1950s it had virtually replaced carbon tetrachloride and trichloroethylene in commercial dry cleaning. Tetrachloroethylene is currently the solvent of choice in most of the world, except in regions, such as Japan, where petroleum-based solvents have remained important in the dry cleaning industry. In 1990, about 53% of the world demand for tetrachloroethylene was for dry cleaning, and about 75% of all dry cleaners used it to clean garments. Chlorofluorocarbon solvents (especially CFC-113) were introduced for use in dry cleaning in the 1970s; however, because of environmental concerns, their use is declining rapidly.

It is estimated that several million people are employed in dry cleaning worldwide. The predominant route of exposure to the solvents used in dry cleaning is by inhalation, although skin absorption and ingestion may also occur. In addition, a wide range of chemicals are used in ‘spotting’ (treatment of spots); they include chlorinated solvents, amyl acetate, bleaching agents, acetic acid, aqueous ammonia, oxalic acid, hydrogen peroxide and dilute hydrogen fluoride solutions.

Improvements in equipment, solvent reclamation and engineering controls in the dry cleaning industry are resulting in decreasing occupational exposures to chlorinated solvents. The trend to use of ‘dry-to-dry’ machines, as opposed to the transfer process, has also resulted in reduced emissions and exposures. Typical average exposures to tetrachloroethylene in dry cleaning declined from about 350–700 mg/m³ in the 1970s to 70–350 mg/m³ in the late 1980s. The differences in airborne concentrations between dry cleaning shops are often many times greater than the differences in the exposures of machine operators and other staff within a shop.

4.2. Human carcinogenicity data

The relationship between employment in dry cleaning and the occurrence of cancer has been assessed in proportionate mortality studies, case-control studies and four cohort studies. Two cohort studies restricted to dry-cleaning workers in the United States were given greater weight in the evaluation than were the results of cohort studies of laundry and dry-cleaning workers (from Denmark and Sweden).

The relative risks for mortality from urinary bladder cancer were elevated in both United States cohorts (relative risks of 1.7 and 2.5, total of 17 deaths), with evidence in one of the studies of an increasing risk with increasing duration of employment. These results are consistent with those from case-control studies in the United States and in Canada and with the findings of a proportionate mortality study in the United States (although the Danish cohort study found no elevated incidence of bladder cancer) and do not appear to be due to confounding by cigarette smoking.

The relative risk for mortality from oesophageal cancer was elevated by a factor of two in both United States cohorts (23 observed deaths in the two studies combined) and increased with increasing duration and/or intensity of employment. This cancer also occurred in slight excess in a proportionate mortality study in the United Kingdom with respect to launderers, dry cleaners and pressers. Risk estimates for oesophageal cancer were not provided in either of the two Nordic studies of laundry and dry cleaning workers. While in a case–control study of oesophageal cancer in Montréal, Canada, none of the case subjects had worked in dry cleaning, the study was relatively small. The relative incidence of oesophageal cancer is increased by consumption of alcohol drinking and cigarette smoking, but potential confounding by these exposures could not be explored directly in these studies.

The relative risk for mortality from cancer of the pancreas was modestly increased in both United States cohort studies; however, this result was not confirmed in two North American case-control studies.

The occurrence of lung cancer was increased slightly in each of the four cohort studies. The mortality rate from lung cancer in the subgroup with long duration of employment and a long interval since first employment (in the one study that evaluated these characteristics) was not elevated. Two case-control studies in North America gave conflicting results.

The relative risk for mortality from cervical cancer was increased by 70–80% in the two United States cohort studies but not at all among Danish dry cleaning and laundry workers. Socioeconomic characteristics were not adjusted for in these studies.

While in a Swedish case-control study and in one of the United States cohort studies moderate increases were found in the relative risk for cancer of the colon in association with employment in dry cleaning, there was no suggestion of an increase in the risk for this form of cancer in the other relevant studies. Furthermore, in the United States cohort study, there was no particular accentuation of the increased risk in relation to increasing duration of employment.

The Danish cohort study of dry cleaning and laundry workers showed some increase in the incidence of cancers of both the liver and the gall-bladder; however, this result was not confirmed in the two United States cohort studies. The results of the case-control studies of liver cancer in the United States are conflicting.

There was a suggestion of an increased risk for non-Hodgkin's lymphoma in one of the two United States cohort studies (relative risk, 1.7, based on seven deaths) and in a large case-control study in the United States; however, no increase in risk for non-Hodgkin's lymphoma was observed in the other United States cohort study, in the Danish cohort study or in a case–control study from Montréal, Canada.

The results of the four cohort studies do not suggest an increase in the risk for cancer of the kidney, while the results of proportionate mortality studies in Wisconsin and Oklahoma and of the case-control studies from Canada, Denmark and the United States indicate an increase in risk associated with a history of work as a dry cleaner. It may be noteworthy that the petroleum solvents used for dry cleaning in Oklahoma are not typical of those used in much of the rest of the world.

Variation within individual studies of dry cleaners may depend on the nature and level of exposure, which varies from shop to shop and across studies of dry cleaning workers. There is also variation in the types of solvents used over time and across geographic regions. These limitations notwithstanding, the epidemiological studies on dry cleaning indicate that the risks for cancers at two sites, urinary bladder and oesophagus, may be increased by employment in dry cleaning.

4.3. Other relevant data

Inconsistent evidence of slight renal damage among workers exposed to tetrachloroethylene in dry cleaning shops was found in two studies, whereas two other studies in which exposure to tetrachloroethylene was at least as high did not find such an association.

Disturbances of sperm quality and fertility have been observed among dry cleaning workers in a few studies of limited size. Several studies performed in Nordic countries have shown a consistent increase in the risk for spontaneous abortion among dry cleaners, but the studies are not entirely independent of each other. No effect has been observed on other reproductive outcomes, such as stillbirth, congenital malformation or low birth weight, but the power of the studies was limited.

In single studies, lymphocytes from dry cleaning workers showed no increase in the frequency of alkaline-labile sites/DNA single-strand breaks. There was inadequate information to evaluate the genetic effects in humans of exposures in dry cleaning.

4.4. Evaluation¹

There is limited evidence in humans for the carcinogenicity of occupational exposures in dry cleaning.