The effects of complex metal oxide mixtures on three soil invertebrates with contrasting biological traits

For regulatory purposes, the concentration addition model is the default first tier for assessing joint-action toxicity of metal mixtures. Although many researchers have evaluated binary and ternary mixtures, fewer have investigated joint-action toxicity in more complex mixtures, where deviations from additivity are more likely due to the greater number of potential interactions. In this study, we tested fixed ratios of five metals (lead, copper, nickel, zinc, cobalt) as metal oxide mixtures on three soil invertebrate species (Enchytraeus crypticus, Folsomia candida, Oppia nitens) at different dose effect levels (EC10-EC90) in an acid sandy forest and a loamy soil. Total metal concentrations for mixture ratios in soil did not explain or correlate with species responses. For F. candida, toxicity was linked to metal solubility, while for O. nitens and E. crypticus, toxicity did not correlate with total or extractable metals. In O. nitens and E. crypticus, however, soil ingestion could be an important route of exposure. Analysing the joint effect of metal mixtures, F. candida response was globally additive, while E. crypticus and O. nitens both presented synergistic effects at low-dose effect levels. Estimations at the EC50 level underestimated the deviations from additivity which were larger at higher and especially lower effect levels. Testing across different effect concentrations (EC10-EC90) was an important tool allowing the identification of these larger deviations from additivity outside the EC50 threshold. Considering most protection thresholds are set below the EC50 level, and it was in this low effect range where the highest synergisms were observed, risk assessment schemes should test additivity at the target protection level using representative test organisms. [Display omitted] •Total metal concentrations did not correlate with species responses to mixtures.•Joint action only at the EC50 level would underestimate deviations from additivity.•Highest synergisms were detected below the EC50 level.•Risk assessment schemes should test additivity at the target protection level.