Serum and adipose tissue as matrices for assessment of exposure to persistent organic pollutants in breast cancer patients.

The aim of this study was to assess differences between two biological matrices (serum and breast adipose tissue) in the evaluation of persistent organic pollutant (POP) exposure in breast cancer patients. The study population consisted of 103 women undergoing surgery for newly diagnosed breast carcinoma in a public hospital in Granada, Southern Spain. Independent variables were gathered from questionnaires and clinical records. POP concentrations were quantified in breast adipose tissue and serum samples. Spearman correlation tests were performed between pairs of POP concentrations and stepwise multivariable linear regression analyses were conducted to assess predictors of concentrations in the two matrices. p,p'- Dichlorodiphenyldichloroethylene (p,p'-DDE) showed the the highest median concentration in both matrices (194.34 and 173.84 ng/g lipid in adipose tissue and serum, respectively). Median wet-basis adipose tissue:serum ratios ranged from 109.34 to 651.62, while lipid-basis ratios ranged from 0.88 to 4.34. In general, we found significant positive correlation coefficients between pairs of POPs in adipose tissue and in serum, which were always higher in adipose tissue. We found positive and statistically significant correlations between serum and adipose tissue concentrations of p,p'-DDE and hexachlorobenzene (HCB) but not of polychlorinated biphenyls (PCBs). Age was positively associated with most POPs in adipose tissue and serum, while the body mass index was positively associated with adipose tissue HCB concentrations and negatively associated with serum PCB-153 and PCB-138 concentrations. Recent weight loss was inversely associated with POP residues in adipose tissue and positively associated with POP residues in serum. Serum HCB and PCB-180 concentrations were lower in patients who had received preoperative chemotherapy. According to our results, serum and adipose tissue POP concentrations in breast cancer patients may be differentially affected by external predictors. Taken together, these findings indicate the need to take account of the individual POP(s) under study and the biological matrix used when relating internal POP exposure to breast cancer disease and to make a careful selection of covariates for adjusting the model.