An investigation of the co-variation in circulating levels of a large number of environmental contaminants

We are daily exposed to many different environmental contaminants. Mixtures of these contaminants could act together to induce more pronounced effects than the sum of the individual contaminants. To evaluate the effects of such mixtures, it is of importance to assess the co-variance amongst the contaminants. Thirty-seven environmental contaminants representing different classes were measured in blood samples from 1016 individuals aged 70 years. Hierarchical cluster analysis and principal component analysis were used to assess the co-variation among the contaminants. Within each identified cluster, possible marker contaminants were sought for. We validated our findings using data from the National Health and Nutrition Examination Survey (NHANES) 2003--2004 study. Two large clusters could be identified, one representing low/medium chlorinated polychlorinated biphenyls (PCBs) (<= 6 chlorine atoms), as well as two pesticides and one representing medium/high chlorinated PCBs (>= 6 chlorine atoms). PCBs 118 and 153 could be used as markers for the low/medium chlorinated cluster and PCBs 170 and 209 could be used as markers for the medium/high chlorinated cluster. This pattern was similar to data from the NHANES study. Apart from the PCBs, little co-variation was seen among the contaminants. Thus, a large number of chemicals have to be measured to adequately identify mixtures of environmental contaminants.