Peripubertal serum levels of dioxins, furans and PCBs in a cohort of Russian boys: can empirical grouping methods yield meaningful exposure variables?

BACKGROUND

Dioxins, furans and polychlorinated biphenyls (PCBs) are endocrine disruptors with half-lives of months to years in humans. Peripubertal exposure to these chemicals may disrupt pubertal timing. Biomarker exposure metrics like the sum of non-dioxin-like PCBs in serum (∑NDL-PCBs) involve chemicals with different hypothetical effects. Empirical exposure metrics represent an alternative that requires no knowledge of biological mechanisms.

METHODS

From 2003 to 2005, 516 Russian boys aged 8-9 residing near a plant that historically produced organochlorine pesticides were enrolled in the Russian Children's Study. At enrollment, blood was collected for measurement of organochlorine chemicals by the Centers for Disease Control and Prevention (CDC). Variable cluster analysis was used to empirically group serum levels of 45 dioxins, furans and PCBs into eight cluster scores. These scores were compared to the sum of dioxin-like toxic equivalents (∑TEQs) and ∑NDL-PCBs using Spearman correlation coefficients.

RESULTS

Clustering appeared to reflect number and position of chlorine atoms. Clusters 4, 7 and 1 contained, respectively, tri/tetra-, tetra/penta-, and hexa/hepta-chlorinated PCBs with chlorines at the 4,4' positions. Clusters 2, 8 and 5 contained, respectively, tetra-to hexa-, hexa/hepta-, and hepta/octa-chlorinated PCBs with chlorines at the 2,2' positions. ∑NDL-PCBs were highly correlated with 4,4'-chlorinated clusters 1 and 7 (Spearman's ρ = 0.8) and less so with all other clusters (Spearman's ρ = 0.4-0.5).

CONCLUSIONS

In this cohort of Russian boys, baseline serum dioxins, furans and PCBs seemed to cluster by chlorination pattern, possibly reflecting shared persistence, metabolism or source of exposure. The widely used measure ∑NDL-PCB seemed most representative of 4,4'-chlorinated PCBs.