Urban areas contribute substantially to xenobiotic contaminant loads in rivers, but their effects have been investigated more for individual organisms and sensitive taxa, rather than through the emergent properties of communities. Here, we use replicated, catchment-scale sampling of benthic invertebrates and novel multivariate techniques to assess whether urban wastewater contaminants affected the structure and function of river food webs. We postulated that the continued occurrence of selected contaminants in river systems might explain the incomplete recovery of urban rivers from legacy gross pollution. Benthic invertebrate communities were sampled monthly over a year (2016–2017) at 18 sites across 3 river systems in South Wales (United Kingdom). Contaminant sources were characterised using remote sensing, water quality data from routine monitoring and measured concentrations of selected persistent xenobiotic pollutants (polychlorinated biphenyls and polybrominated diphenyl ethers). Urban wastewater discharges had relatively limited effects on river water quality, with small increases in nitrate, phosphate, temperature, conductivity and total dissolved solids in urban systems. Concentrations of polychlorinated biphenyls and polybrominated diphenyl ethers in invertebrates, however, were significantly higher under greater urban land cover and wastewater discharge. Food webs at the most highly contaminated urban sites were characterised by: (i) reduced taxonomic and functional diversity; (ii) simplified food web structure with reduced network connectance; and (iii) reductions in the abundance of prey important for apex predators such as the Eurasian dipper (Cinclus cinclus). Although correlative and partially confounded by other effects, these data provide support for the hypothesis that impairment to food webs resulting from urban pollutants might explain population, community and ecosystem-level effects in urban river systems, and hence incomplete recovery from past pollution.