The identification of host factors that are predictors of changes in serum polyhalogenated biphenyl contaminants over time has been a difficult challenge in epidemiologic studies of exposed individuals. Of particular concern are age at exposure, reproductive and lactational histories, and changes in body mass index. Using both cross-sectional and longitudinal approaches, this study examined factors related to high initial serum PBB and PCB levels and changes in these levels over time among women of varying ages at exposure (n=1772; age range<1 to 45 years). In 1973, PBB exposure occurred through consumption of farm products contaminated with PBB added to cattle feed. Exposures to PCBs began in 1941 through PCB-contaminated silo sealant deteriorating into animal feed. The Michigan Department of Public Health began enrolling participants in 1977 and has continued to follow them through annual updates. At enrollment, questionnaires were administered to obtain demographic, lifestyle, and anthropometric measurements, medical/reproductive and occupational histories, and contaminated food consumption patterns. Blood samples were collected for PBB and PCB analysis at enrollment for all participants; additional serum tests were done on a subset of the population during follow-up. Median serum levels at enrollment were 2.0 ppb PBB and 5.0 ppb PCB. A decline in serum PBB level over an interval that ranged from 1 to 146 months (median=31) was observed for 44.6% of the women (median=1.0 ppb), while 12.2% showed an increase (median=1.0 ppb). PCB levels declined in 50.3% of the women (median=3.0 ppb) while 12.2% increased (median=2.0 ppb). Relative to women whose contaminant levels were stable, higher initial serum level was a predictor of decline for both PBB and PCB (OR=1.66, 95% CI 1.52–1.82; OR=3.26, 95% CI 2.58–4.12, respectively); a yearly increase in interval between tests was related to declining PCBs (OR=1.65, 95% CI 1.46–1.87). In addition, age≤10 years at exposure (OR=1.72, 95% CI 1.03–2.86) and residence on a quarantined farm (OR=1.40, 95− CI 1.03–1.90) were predictors of a decrease in PBBs. Factors related to an increase in PBB levels were age≤10 years at exposure (OR=0.30, 95% CI 0.10–0.96) and initial PBB level (OR=1.24, 95% CI 1.15–1.33); and for PCBs, high initial level (OR=1.34, 95% CI 1.17–1.53) and body mass index (OR=1.07, 95% CI 1.01–1.13). One or more live births during the interval between tests were not related to changing levels of either contaminant; breastfeeding data were not available for examination. Early age at exposure appears to be an important predictor of changes in serum PBB levels OVER TIME.