Risk of pancreatic cancer in workers exposed to chlorinated hydrocarbon solvents and related compounds: a meta-analysis.

This is a meta-analysis of occupational exposures to chlorinated hydrocarbon (CHC) solvents and pancreatic cancer, based primarily on studies that addressed exposure directly (agent studies) and secondarily on studies that reported data without verification of individual CHC exposures (job title studies), all of which were listed in databases for the period January 1969 to May 1998. Standardized extraction of data and double-checking of consistency of data extraction by five extractors were done. Simple random models estimated meta-relative risks. Suggestive weak excesses were found for trichloroethylene (meta-relative risk (MRR) = 1.24, 95% confidence interval (CI): 0.79, 1.97), polychlorinated biphenyls (MRR = 1.37, 95% CI: 0.56, 3.31), methylene chloride (MRR = 1.42, 95% CI: 0.80, 2.53), and vinyl chloride (MRR = 1.17, 95% CI: 0.71, 1.91) but not for carbon tetrachloride. One study addressed tetrachloroethylene (MRR = 3.08, 95% CI: 0.63, 8.99); another investigated chlorohydrin production (MRR = 4.92, 95% CI: 1.58, 11.4). Exposure-response meta-analyses for trichloroethylene and methylene chloride failed to reveal trends. Job title studies on metal degreasing and dry cleaning revealed significant MRRs (2.0 and 1.4, respectively). Publication bias was unlikely. Confounding may have remained insufficiently controlled. Unless the results are seriously biased by exposure or endpoint misclassification or by confounding, strong causal associations between CHC compounds and pancreatic cancer can be judged unlikely. Interactions between environmental and occupational agents, lifestyle factors, and genetic susceptibility remain a possibility, but the data for this meta-analysis did not address interactions.