Associations of perfluoroalkyl substances (PFAS) with terminal ductal lobular unit involution of the normal breast.

BACKGROUND

Per- and polyfluoroalkyl substances (PFAS) may be carcinogenic, and animal studies demonstrate their harmful effects on mammary gland development. Terminal ductal lobular units (TDLUs) are the structures that produce milk following childbirth, and involution of TDLUs normally occurs with aging. Most breast cancers arise from TDLUs, and a greater degree of TDLU involution may be associated with lower breast cancer risk. We estimated associations between PFAS concentrations and TDLU involution in normal breast tissue samples.

METHODS

Concentrations of seven PFAS were measured in serum provided by a subset of 263 cancer-free volunteer participants from the Susan G. Komen for the Cure Tissue Bank (KTB) who were postmenopausal, not currently using hormone therapy, and had available TDLU measurements made by a trained pathologist examining H&E stained section of a core biopsy sample of tissue from the outer upper quadrant of a single breast. Bayesian kernel machine regression and quantile-G computation were used to estimate covariate-adjusted associations between the PFAS mixture and measures of TDLU involution (presence of TDLUs, number of observed TDLUs, and median TDLU span) within this population and with stratification on parity and breastfeeding history.

RESULTS

TDLUs were observed in breast tissue samples of 40.3% (N = 106) of the study population, with similar PFAS concentrations between participants with and without observed TDLUs. No strong, statistically significant associations were observed between individual PFAS and presence of observed TDLUs. The overall effect of the PFAS mixture suggested an inverted U-shaped association with odds of observed TDLUs, although this was not statistically significant (β = 0.03 95% CI -2.75, 2.81; p = 0.98). Among the subgroup of parous women, stratified analyses suggested a positive association between the PFAS mixture and observed TDLUs among those who had ever breastfed, but a slightly negative association among those who had never breastfed.

CONCLUSIONS

Overall, our analysis does not support meaningful effects of PFAS on TDLU involution, although we note that these findings are not applicable to premenopausal women or to postmenopausal women using hormone therapy.