PCBs and OH-PCBs in serum from children and mothers in urban and rural U.S. communities

Abstract

East Chicago, Indiana is a heavily industrialized community bisected by the Indiana Harbor and Ship Canal, which volatilizes ~7.5 kg/yr polychlorinated biphenyls (PCBs). In contrast, the rural Columbus Junction, Iowa area has no known current or past PCB industrial sources. Blood from children and their mothers from these communities were collected April 2008 to January 2009 (n = 177). Sera were analyzed for all 209 PCBs and 4 hydroxylated PCBs (OH-PCBs). Sum PCBs ranged from nondetect to 658 ng/g lw (median = 33.5 ng/g lw). Sum OH-PCBs ranged from nondetect to 1.2 ng/g fw (median = 0.07 ng/g fw). These concentrations are similar to those reported in other populations without high dietary PCB intake. Differences between the two communities were subtle. PCBs were detected in more East Chicago mothers and children than Columbus Junction mothers and children, and children from East Chicago were enriched in lower-molecular weight PCBs. East Chicago and Columbus Junction residents had similar levels of total and individual PCBs and OH-PCBs in their blood. Concentrations of parent PCBs correlated with concentrations of OH-PCBs. This is the first temporally and methodologically consistent study to evaluate all 209 PCBs and major metabolites in two generations of people living in urban and rural areas of the United States.

Figure 1

Detection frequency of each PCB and OH-PCB congener in our sample set of East Chicago and Columbus Junction residents (n=177). See Tables S6 and S7 for congener-specific data.

Figure 2

Fraction of low-molecular weight PCBs, defined as mono-penta CBs divided by total PCBs in mothers and children from East Chicago (left) and Columbus Junction (right). Data are plotted as box plots with the median indicated by the bold horizontal line, the two middle quartiles shown as polyhedrons above and below the median and the 95th percentiles shown as the horizontal lines connected by the dashed line. Outlier points are indicated by open circles.

Figure 3

Fresh weight concentrations of sum OH-PCBs and sum of their precursor PCBs display a linear trend (R² = 0.62, p < 0.0001). Each data point represents one participant. One leverage point, a mother with much higher concentrations than the other participants, was excluded. Participants with values <LOQ were excluded.

Figure 4

Best fitting linear regressions between specific PCB parents and their OH-PCB metabolites in mothers and children: (a) 4-OH-PCB 107 with PCB 105 (R² = 0.02) and PCB 118 (R² = 0.11, p = 0.024); (b) 3′-OH-PCB 138 with PCB 138 (R² = 0.13), and PCB 157 (R² = 0.18); (c) 4-OH-PCB 146 with PCB 138 (R² = 0.26, p < 0.0001), PCB 146 (R² = 0.75, p = 0.0025), and PCB 153 (R² = 0.39, p < 0.0001); (d) 4-OH-PCB 187 with PCB 183 (R² = 0.27, p = 0.070) and PCB 187 (R² = 0.47, p < 0.0001). One leverage point, a mother with much higher concentrations than the other participants, was removed from all 4 graphs. An outlier, a mother with a very high concentration of 4-OH-PCB 146 compared to the other metabolites was also removed from graph (c). Participants with values <LOQ were excluded; it was therefore not possible to determine correlations between PCB 107 and PCB 130 and their respective metabolites.

Figure 5

Comparison of PCBs 28, PCB 105, and PCB 153 levels in units of nanogram per gram lipid weight in populations around the world, including this study. Population demographics and sample collection years are indicated in the figure. The published reports did not all use consistent measures of central tendency or range. These differences are indicated a–c, where a = min, median, max; b = mean & standard deviation; c = median. (ref 1), (ref 2), (ref 3). These references were chosen for comparison because they provided lipid weight concentration data for all three congeners and represented a variety of target populations.

Figure 6

Comparison of OH-PCB levels in units of nanogram per gram fresh weight in populations around the world, including this study. Population demographics and sample collection years are indicated in the figure. The published reports did not all use consistent measures of central tendency or range. These differences are indicated a–c, where a = median; b = min, median, max; c = 5%, median, 95%. (ref 1), (ref 2), (ref 3), (ref 4), (ref 5). These references were chosen for comparison because they provided data for all four congeners and represented a variety of target populations.