Much of our understanding of anticoagulant rodenticide toxicity to non-target wildlife has been derived from molecular through whole animal research and registration studies in domesticated birds and mammals, and to a lesser degree from trials with captive wildlife. Using these data, an adverse outcome pathway identifying molecular initiating and anchoring events (inhibition of vitamin K epoxide reductase, failure to activate clotting factors), and established and plausible linkages (coagulopathy, hemorrhage, anemia, reduced fitness) associated with toxicity, is presented. Controlled exposure studies have demonstrated that second-generation anticoagulant rodenticides (e.g., brodifacoum) are more toxic than first- and intermediate-generation compounds (e.g., warfarin, diphacinone), however the difference in potency is diminished when first- and intermediate-generation compounds are administered on multiple days. Differences in species sensitivity are inconsistent among compounds. Numerous studies have compared mortality rate of predators fed prey or tissue containing anticoagulant rodenticides. In secondary exposure studies in birds, brodifacoum appears to pose the greatest risk, with bromadiolone, difenacoum, flocoumafen and difethialone being less hazardous than brodifacoum, and warfarin, coumatetralyl, coumafuryl, chlorophacinone and diphacinone being even less hazardous. In contrast, substantial mortality was noted in secondary exposure studies in mammals ingesting prey or tissue diets containing either second- or intermediate-generation compounds. Sublethal responses (e.g., prolonged clotting time, reduced hematocrit and anemia) have been used to study the sequelae of anticoagulant intoxication, and to some degree in the establishment of toxicity thresholds or toxicity reference values. Surprisingly few studies have undertaken histopathological evaluations to identify cellular lesions and hemorrhage associated with anticoagulant rodenticide exposure in non-target wildlife. Ecological risk assessments of anticoagulant rodenticides would be improved with additional data on (i) interspecific differences in sensitivity, particularly for understudied taxa, (ii) sublethal effects unrelated to coagulopathy, (iii) responses to mixtures and sequential exposures, and (iv) the role of vitamin K status on toxicity, and significance of inclusion of supplemental vitamin K or menadione (provitamin) in the diet of test organisms. A more complete understanding of the toxicity of anticoagulant rodenticides in non-target wildlife would enable regulators and natural resource managers to better predict and even mitigate risk.