National Industrial Chemicals Notification and Assessment Scheme

11.7.1Cohort studies

In the cohort investigated by Axelson et al (1994) no significant increases in cancer mortality or morbidity were observed. Mortality was analysed in worker subgroups categorised according to urinary trichloroacetic acid levels (<49, 50 to 99 or >100 mg/L) and exposure time (< or > 2 years). The two lowest exposure groups had low cancer mortality.

However in the > 100 mg/L group, the standardised mortality ratio (SMR) for cancer was slightly increased. In males, some excess of cancers of the liver, larynx, prostate and of non-Hodgkins lymphoma were observed but the excess of liver and prostate cancers and lymphomas were in the low exposure group. A statistically significant increase was seen for malignant skin tumours in men (standardised incidence ratio SIR 2.36, 95% confidence interval 1.02 - 4.65). This increase, however, was in the low exposure group. The overall female cancer morbidity was slightly higher than expected (SIR 1.32, CI 0.53 - 3.79) among women with < 2 years exposure. Half of these cases had tumours of the breast or genital organs and the other cancers were in the gastrointestinal tract. There were no cases of skin or liver cancers, lymphoma or leukaemia among women.

An epidemiological study was conducted in a group of workers at an aircraft maintenance facility in Utah (Spirtas et al., 1991). A number of chemicals were used at the facility, such as chlorinated hydrocarbons (including trichloroethylene), aromatic hydrocarbons (such as toluene and xylene) and some alcohols. Exposure indices were calculated based on job, frequency of exposure, frequency of peak exposure and duration of exposure. Mortality from all causes was significantly less than expected for both men and women. There were no statistically significant excesses for cancer deaths in general or for specific kinds of cancer.

Blair et al (1998) followed up the cohort of 14457 aircraft maintenance workers previously reported by Spirtas et al (1991). Workers exposed to trichloroethylene showed non-significant excesses for Hodgkin’s lymphoma (RR 2.0), cancers of the oesophagus (RR 5.6), colon (RR 1.4), primary liver (RR 1.7), breast (RR 1.8), cervix (RR 1.8), kidney (RR 1.6) and bone (RR 2.1). The findings in this study did not show a strong association between trichloroethylene exposure and any cancers. The associations were not significant or dose related and not consistent between men and women. This study included a large cohort of workers with a follow up period of about 40 years enabling detection of cancers with a long latency period. However workers were exposed to a number of chemicals and it was not possible to evaluate risks from individual chemicals.

Antilla et al (1995) divided the total cohort into sub-groups, on the basis of the observation period, into 0-9, 10-19 or more than 20 years. The average urinary trichloroacetic acid levels were 8.3 mg/L in women and 6.3 mg/L in men. The mean latency period was 18 years. Risk of cervical cancer was significantly increased for the study cohort , with higher numbers in the shortest follow-up group (0-9 yrs). There was a significantly increased incidence of cancer in general (SIR 2.98 95% CI 1.20 - 6.13 ) in the group with the follow-up of > 20 yrs. There was a significant increase in the incidence of tumours of the liver (SIR 6.1; 95% CI 1.3 - 18), prostate (SIR 3.56; 95% CI 1.5 - 7.0), stomach (SIR 3.0; 95% CI 1.2 - 6.1) and lymphohaematopoietic system (SIR 3.0; 95% CI 1.2 - 6.1) in the group with a follow up of > 20 yrs.

Henschler et al (1995) reported a study of renal cancer in workers exposed to trichloroethylene at a cardboard manufacturing factory. Physical examination of the workers included abdominal sonography and causes of death were obtained from hospital records. Tumour diagnosis date was the date of surgery and renal tumours were verified by histopathological examination. Air concentrations of trichloroethylene or metabolites in urine were not available. Information indicates that the workers in the cohort were exposed to high concentrations of trichloroethylene over long periods of time. The average period of exposure was 18 years and the average observation period was 30 years. The incidence of renal cancer in the cohort was compared directly with the incidence in the control group and with data of the cancer registries of Denmark and German Democratic Republic.

Five cases of renal cancer were diagnosed by the close of the study and two additional cases were diagnosed later giving a total of seven cases in the cohort. No renal cancer was observed in the control group. A statistically significant increase in incidence of renal cancer was obtained compared with cancer registry of Denmark (SIR of 7.97; 95% CI 2.6 - 19) and German Democratic Republic (SIR of 9.66; 95% CI 3.1 - 23). This study has been criticised for a number of reasons. IARC (1995) have noted that the study may have been initiated after observing a cluster of cancer cases. Others have also noted that the study was a cluster study and that physician and hospital records should not be compared with general population mortality rates (Bloemen & Tomenson, 1995; Swaen, 1995). Though this study appears to be a cluster investigation rather than a retrospective cohort the findings of this study raise concern of an association between high trichloroethylene exposure and renal cancer.

Mortality at a plant in the US using trichloroethylene as a degreasing agent was investigated in a study by Shindell and Ulrich (1985). Persons working for more than three months from 1957 to 1983 were included in the study. No data on exposure levels were available. Overall mortality (SMR for white males 0.79) and cancer mortality (SMR for white males 0.62) were found to be less than expected.

11.7.2Case-control studies

A case-control study including 59 nephrectomised patients has been described by Vamvakas et al (cited in Deutsche Forschungsgemeinschaft, 1996). The study included all patients who had been diagnosed with renal cell tumours at histopathology after nephrectomy between December 1987 and May 1992. The control group included 84 traffic accident patients treated in the same clinic. Abdominal ultrasonography was used to exclude renal tumours in the control group. Exposure evaluation was carried out by questionnaires and personal interviews. Information was also obtained from physicians and occupational hygienists. Of the nephrectomy cases, 20 had been exposed to trichloroethylene and none to tetrachloroethylene. Five from the control group had been exposed to trichloroethylene and 2 to tetrachloroethylene. The average exposure period for the cases was 19 years. A highly significant odds ratio of 13.42 (95% CI 3.50 - 51.39) was obtained for the combined exposure to trichloroethylene and tetrachloroethylene. However, no exposure to tetrachloroethylene had been reported in the nephrectomised patients. Factors such as age, sex, smoking habits, blood pressure and consumption of diuretics were allowed for by logistic regression. The significant odds ratio is suggestive of an association between trichloroethylene exposure and renal cell carcinomas. The nephrectomised patients were classified into high, medium or low level exposures on the basis of duration, frequency of exposure and the workplace description. Eight patients with renal tumours were in the high exposure group, 10 in the medium and 2 in the low. Of the controls, 2 were in the high exposure group, 3 in the medium and 2 in the low exposure group. Only a summary of this study was available during the Priority Existing Chemical assessment. The data are therefore limited and do not allow an in depth assessment of the quality of the study.

Several other case-control studies have investigated the carcinogenic effects of trichloroethylene. Trichloroethylene exposure was not found to be a risk factor for astrocytic brain tumours (Heineman & et al., 1994). The incidence of liver cancer was investigated in people exposed to trichloroethylene in separate studies (Novotna et al., 1979; Paddle, 1983). None of the liver cancer cases identified were found to be occupationally exposed to trichloroethylene. These two studies only looked at one type of cancer and included limited numbers. A high odds ratio (7.4) was found for dry cleaners exposed to trichloroethylene in a case-control study investigating risk factors for colon cancer (Fredrickson et al., 1989). The odds ratios were not significantly increased for all dry cleaners or for all exposed to trichloroethylene. This study does not provide sufficient evidence of a causal association between trichloroethylene and colon cancer in dry cleaners because of the small number of exposed people.