that remove these compounds from the final product (Source:

Adverse Health Effects An extensive experimental literature on rodent and nonhuman primate models demonstrates that prenatal exposure to PCBs can interfere with neurodevelopment (Rice, 2000; Faroon et al., 2001; Bowers et al., 2004; Nguon et al., 2005). This literature is complemented by numerous prospective epidemiological studies of children conducted in Michigan; North Carolina; Oswego, NY; Germany; Faroe Islands; and the Netherlands (Schantz et al., 2003). The cohorts were often chosen to include children born to women who consumed fish from waters known to be contaminated with PCBs. The results of these epidemiological studies are generally congruent with those using animal models, although, as in most areas of observational research in humans, results are not always consistent across studies or consistent over time in a particular study.

Higher prenatal exposures have been associated with deficits in various functional domains including intelligence, attention, response inhibition, activity, and play behaviors (Jacobson and Jacobson, 1996; Patandin et al., 1999; Walkowiak et al., 2001; Vreugdenhil et al., 2002a,b; 2004; Jacobson and Jacobson, 2003; Stewart et al., 2003). However, there are some uncertainties about many key issues. One issue is the shape of the dose-effect relationship curve and, specifically, whether a threshold exists. A second is whether PCB exposure leading to adverse effects occurs prenatally or postnatally. Although most of the focus has been on prenatal exposures, some recent studies suggest that early postnatal exposures are also associated with neurotoxicities (Huisman et al., 1995; Walkowiak et al., 2001; Winneke et al., 2002). A third issue is the relative potency of the different congeners. For some neurodevelopmental outcomes, it is exposure to the dioxin-like congeners that is most strongly associated with deficits. A fourth issue is the impact of synergism between PCBs and other toxicants. Some studies suggest that adverse effects arise only when PCB exposure occurs in the presence of methylmercury or in environments in which individuals may be exposed to increased levels or multiple exposures (Grandjean et al., 2001; Roegge et al., 2004; Roegge and Schantz, 2006).

The PCB exposures identified in these study samples were considerably greater than those of the general US population. The median concentration of PCB 153 in the 10 studies, the only basis for direct comparison, ranged from 30 to 450 ng/g serum lipid, and the median of the 10 medians was 110 ng/g. The exposure levels in the two recent US studies were about one-third of those in the four earlier US studies or recent Dutch, German, and northern Québec studies (Longneker et al., 2003), consistent with exposure surveys indicating that PCB levels in human tissues in the United States have declined in recent decades (Sjodin et al., 2004a). In the most recent

The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement