CNS tumors and exposure to acrylonitrile: Inconsistency between experimental and epidemiology studies

Abstract

Acrylonitrile is a potent CNS tumorigen in rats leading to concern that it may be a tumorigen in humans. There have been 12 epidemiology studies of 37,352 workers exposed to acrylonitrile which evaluate CNS cancers. We summarize and evaluate these epidemiology studies for CNS cancers using the methods of meta-analysis. Our analyses indicate that workers with acrylonitrile exposure have null findings for CNS cancer (relative risk = 1.1, 95% confidence interval 0.8-1.5), which are in stark contrast to the projected risk to humans using the rat findings (relative risk = 3.5, 95% confidence interval 3.0-4.0). We discuss several explanations for the inconsistency between animal and human findings, including the possibility that the acrylonitrile-induced rat CNS tumors may not be relevant to humans. Given the rarity of CNS tumors in humans and a lack of understanding of the causal mechanisms of these tumors in rats, however, a more definitive conclusion will have to await additional experimental and observational data. Nevertheless, the epidemiology evidence indicates that acrylonitrile is not a potent CNS tumorigen.

Using Epidemiological Studies to Check the Consistency of the Cancer Risks Predicted by High‐Dose Animal Experiments: A Methodological Review

“…Aggregate meta-analysis generates a single summary estimate that has a considerably narrower confi- (19) reported a significantly lower acrylonitrile risk for a meta-analysis of more than 10 mortality studies (relative risk [RR] Ϸ 1.0) compared to a rodent-based prediction (RR Ϸ 2.5). This precise null summary estimate obscures the fact that five individual epidemiological studies reported a nonsignificant RR of 2 to 4 and that the meta-RR was largely driven by the results of a single, large, U.S. study with approximately twice as many subjects as the other studies combined.…”

“…When chemical specific cancers have not been clearly identified, investigators have generally compared the tumor type that is statistically increased in rodent studies with the same or similar type in people. (2)(3)(4)(5)(6)11,18,19) This seems reasonable if primarily one tumor type is increased in rodents and if this tumor has a counterpart in people (e.g., brain, bladder, and so forth). This does not, however, guarantee that an appropriate animal-to-human comparison will necessarily result.…”

“…Meijers et al (23) generated human risk for vinyl chloride using a 1986 Dutch Health Council report, which is an obscure reference written in Dutch. Collins and Strother (19) presented animal predictions for acrylonitrile that were substantiated by a reference to the methods used by Ward and Starr, (18) without a description of those methods. Although we are not suggesting that these data sources are not useful, this lack of transparency detracts from reader understanding and confidence in the results.…”

“…Most of the investigators that made statistical comparisons provided either a 95% CI around the observed number of deaths, (2,3) a 95% CI around the predicted SMR, (4) or a p value for the difference between the observed and predicted. (2,4,19) However, two studies presented arguments based on statistical power. Tollefson et al (11) suggested that the human and animal evidence for MC were consistent, because the Kodak cohort study did not have sufficient power to detect a predicted SMR of 1.01 to 1.4.…”

Acrylonitrile-induced extracellular signal-regulated kinase (ERK) activation via protein kinase C (PKC) in SK-N-SH neuroblastoma cells

“…the brain, forestomach and lung (Ghanayem et al, 1997;Ghanayem et al, 2002;Quast, 2002). However, epidemiological evidence associating ACN exposure and cancer are inconclusive (Collins and Strother, 1999;Marsh et al, 2001;Starr et al, 2004). Therefore, it remains to be determined whether ACN causes tumors in humans.…”