The in situ feeding bioassay in Gammarus fossarum is recognised as a reliable tool for monitoring the toxicity of freshwater contamination. However, whether recorded feeding inhibitions can potentially provoke population-level adverse outcomes remains an open question. Here we present an experimental study in G. fossarum, which contributes to the quantitative description of the links between feeding inhibitions and impacts on female reproductive performance. We study the impacts of food deprivation on reproductive endpoints (i.e. fecundity, fertility, moult cycle) during two successive moulting cycles. Among the main results, we found that food deprivation triggers a slowdown of the moulting process and a reduction in fertility, but no alteration on embryonic development. These reproductive impairments appear for feeding inhibition values usually recorded in monitoring programs of environmental pollution. Using a population model translating Gammarus life-history, we predict that these reproductive alterations are predictive of a strong degradation of population dynamics. This study underlines the importance of feeding inhibition in population-level risk assessment. With our example, we discuss how establishing upscaling schemes based on quantitative mechanistic links between impacts at different levels of biological organization can be applied in environmental monitoring, in order to propose an ecotoxicological assessment of water quality, which would be sensitive, specific and ecologically relevant.