Background: Polychlorinated biphenyl (PCB) exposures are encountered by the general public by eating contaminated food or living near a previously operating PCB factory or hazardous waste site. PCBs affect the immune, reproductive, nervous, and endocrine systems and are carcinogens. PCBs were banned in the United States in 1977. For public health, it is important to be able to estimate individual risk, especially for vulnerable populations, to monitor the decline in risk over time and to alert the public health community if spikes occur in PCB exposures, by measuring serum PCB levels. The historical decline in PCB exposures cannot be documented within a repeatedly tested general population, since there is no such population. Therefore, our aim was to model serum PCB levels in the US general population over time using published data.
Methods: Models were developed based on 45 publications providing 16,914 background PCB levels in sera collected 1963-2003. Multiple linear regression and exponential decay were used to model the summary PCB levels.
Results: Background levels of higher-chlorinated PCBs (five or more chlorines) in sera increased before 1979 and decreased after 1979; a quadratic model was the best fit. However, the exponential decay model explained better the low PCB serum levels still seen in the general population. For lower-chlorinated serum PCBs, no increase or decrease was shown (1.7ppb for all years).
Conclusions: Limitations for both models were lack of repeated measures, non-randomly selected study participants, selected years, concentration on geographic areas centered on PCB waste sites, lack of adjustment for BMI or for laboratory methods. Despite the limitations, this analysis shows that background PCB levels in the general population are still of concern. Future work should focus on uncertainties governing how to interpret the levels with respect to possible long term health effects.