Trichloroethylene and Cancer

Trichloroethylene (TCE) was for many decades the solvent of choice for cleaning and vapor degreasing of metal parts, and has been commonly used in several other industrial applications. Since the early 1970s, however, concern over potential risks of cancer and other health effects among TCE-exposed workers, coupled with its documented environmental release through air and wastewater emissions, has led to a decline in its use. Strong evidence of carcinogenicity has come from experimental research, with TCE shown to induce tumors in kidney, liver, lung, testes, and hematopoietic tissue in rodents (1). The evidence from occupational epidemiologic studies has been less clear, although a recent meta-analysis of published cohort and case-control findings supports an association with kidney cancer and more limited evidence of association with liver cancer and non-Hodgkin lymphoma (NHL) (2). In October 2012, an expert review panel at the International Agency for Research on Cancer (IARC) upgraded the classification of TCE to known human carcinogen (group 1) (3), largely based on findings related to kidney cancer.

In this issue of the Journal, Hansen et al. report new findings from an updated analysis of three cohorts of TCE-exposed workers, with additional follow-up of 10 to 15 years of cancer incidence for each cohort (4). A unique strength of these cohorts, located in Denmark, Sweden, and Finland, is that exposure to TCE among workers was confirmed from past measurements of urinary trichloroacetic acid (U-TCA), a major metabolite of TCE and a biomarker of recent exposure. In a combined analysis of all 5553 exposed workers, the investigators observed statistically significant elevated standardized incidence ratios (SIRs) for cancers of the liver (1.93; 95% confidence interval [CI] = 1.19 to 2.95) and cervix (2.31; 95% CI = 1.32 to 3.75), but not for kidney cancer (1.01; 95% CI = 0.70 to 1.42). A statistically significant association with NHL was observed for men (SIR = 1.55; 95% CI = 1.06 to 2.20) but not women (SIR = 0.63; 95% CI = 0.23 to 1.37), with inconclusive results overall (SIR = 1.26; 95% CI = 0.89 to 1.73). In internal analyses, evidence of an exposure-response relationship with mean U-TCA was found only for cervical cancer. However, given the short half-life of U-TCA of 35–70 hours (5), the large proportion of cohort members with only one measurement (Denmark, 29%; Finland, 37%; Sweden, unknown), and the absence of information on exposure duration, it is unclear to what extent mean U-TCA is representative of cumulative TCE exposure. The analyses were also limited by the small numbers of accrued patients (cervix, 16; kidney, 32; liver, 36; NHL, 38). The investigators were unable to adjust for smoking or alcohol consumption; however, TCE was not found to be associated with all smoking-related cancers combined, nor for alcohol-related cancers, arguing against confounding from these exposures as an explanation for these findings.

Given the timing of this article so soon after the IARC evaluation, it is natural to wonder whether these new results would have affected the panel’s decision to upgrade TCE to group 1 status. The findings of Hansen et al. for liver cancer and kidney cancer are consistent with the earlier findings from these cohorts, whereas the new results for NHL are weaker than those previously observed (SIR = 3.1, 1.52, and 1.81 for Denmark, Sweden, and Finland, respectively) (6–8). However, given the small sizes of the three cohorts, it is unlikely that the new findings from updated follow-up would have meaningfully affected the summary results of the recent meta-analysis. The pooled results for kidney cancer observed by Hansen et al. arguably do not constitute an “informative null” in that, as acknowledged by the authors, their analysis had limited power to detect an association comparable in size to the meta-analysis summary relative risk of 1.27 (2). Moreover, a statistically nonsignificant elevated hazard ratio for kidney cancer was observed for workers with a mean U-TCA of 50mg/L and greater relative to those with U-TCA less than 5mg/L (2.04; 95% CI = 0.81 to 5.17), suggestive of a TCE effect at high exposure levels. Given these observations, it is unlikely that these findings from Hansen et al. would have changed the conclusion of the IARC panel.

Although TCE is now classified as a known carcinogen, there remain several important research questions requiring further investigation. Better identification of the exposure levels that are capable of inducing relevant biologic effects (eg, alterations in immune markers (9) and biomarkers of kidney or liver injury) in TCE-exposed workers is needed to inform the regulation of safe exposure limits. Such findings may also provide further insight into the underlying carcinogenic mechanisms of TCE at different cancer sites. Additional questions for epidemiologic investigation include whether TCE effects on cancer risk are modified by inherited metabolic gene variants, as was suggested in a recent study of kidney cancer (10), or differ across disease subtypes. There is also a need to resolve whether TCE increases risk of cervical cancer and other malignancies with limited or inconsistent epidemiologic evidence. Opportunities for new epidemiologic studies may be most promising in developing countries, where TCE use is actually increasing (11).

Employers can adopt a variety of available approaches to reduce workers’ exposure to TCE. Where possible, TCE should be substituted by safer alternative chemicals and/or emissions should be reduced. Conversion from conventional vapor degreasers to new low-emission equipment such as enclosed vapor degreasing systems can greatly reduce solvent exposures in the workplace, and aqueous cleaning systems may also be feasible alternatives in certain applications.

Funding

Supported by the Intramural Research Program of the National Institutes of Health and the National Cancer Institute. The study sponsor had no role in the writing of the editorial or the decision to submit the editorial for publication.

The author has no conflicts of interest to declare.