Classification of NDL-PCB congeners based on extensive in vitro screening and multivariate statistics

Abstract

Principal component analysis (PCA) was applied to a data set containing the levels of indicator polychlorinated biphenyls (PCBs) in human milk of mothers living in the Wielkopolska region, Poland, in order to investigate the information captured in the PCB patterns and to elucidate the relationship between PCB concentrations in milk and donor characteristics. According to the obtained PCA results milk fat content was the most influential factor affecting the PCB levels in milk of the Wielkopolska cohort. The lifestyle data collected from the questionnaire completed by the donors appeared to have no influence on PCB concentrations in breast milk. The score plots revealed the PCB contents of milk were quite low and uniform with a few outliers, without discrimination observed either between the primipareous and secundipareous females or between donors from the urban and rural areas. Comparison of the PCB levels and profiles of human milk from the Wielkopolska region and from various European and Asian locations made by PCA reflected a generally low background exposure and indicated the possible reasons for the outlying of some samples. In order to enhance the chances of observing the relationship between donor habits and PCB levels in breast milk it was suggested that the questionnaire be redesigned to gather information about vegetable product consumption and indoor air exposure.