Control Priorities for Persistent Organic Pollutant Mixture Based on Neurodevelopmental Disorders

Long-term population-level exposure data of persistent organic pollutants (POPs) are fragmented, and chemical control strategies are often based on exposure level, with limited consideration of biological relevance. Here, we reconstructed long-term exposure trajectories of 91 POP mixtures using NHANES biomonitoring data by integrating low-rank matrix completion and multitask temporal forecasting (R2 = 0.93), covering DIOXINs, organochlorine pesticides (OCPs), perfluoroalkyl and polyfluoroalkyl substances (PFASs), polybrominated diphenyl ethers (PBDEs), polychlorinated biphenyls (PCBs). Reconstructed exposure profiles were further embedded into multilayer biological networks of "exposure-biology-disease" framework focusing on neurodevelopmental disorders (NDDs, 349 subtypes) as an application scenario. Across POP categories, reconstructed population-level concentrations declined substantially (>80-95% projected by 2030-2040), which increasingly concentrated among a limited number of persistent and biologically connected chemicals. Chemical risk rankings exhibited pronounced temporal reorganization, with early period relevance largely driven by PFASs, while under low-exposure futures, disease relevance shifted toward other POP categories, particularly in protein-centric interaction networks. By integrating long-term exposure reconstruction with network-informed NDDs relevance, our study provides a dynamic framework for prioritizing control of POP mixtures. In the early postregulation period (2004-2009), PFAS dominated NDD relevance in both frameworks, accounting for more than half of the total score. By 2025, however, the relative contribution of PFAS decreased, while PBDEs and PCBs increased markedly. In sum, we proposed that given the dramatic decline of conventional POPs following the implementation of the Stockholm Convention, increasing attention should now be directed toward other emerging pollutants.