From 1930 to 1995, the Upper Columbia River (UCR) of northeast Washington State received approximately 12 million metric tons of smelter slag and associated effluents from a large smelter facility located in Trail, British Columbia, approximately 10 km north of the United States–Canadian border. Studies conducted during the past two decades have demonstrated the presence of toxic concentrations of heavy metals in slag-based sandy sediments, including cadmium, copper, zinc, and lead in the UCR area as well as the downstream reservoir portion of Lake Roosevelt. We conducted standardized whole-sediment toxicity tests with the amphipod Hyalella azteca (28-day) and the midge Chironomus dilutus (10-day) on 11 samples, including both UCR and study-specific reference sediments. Metal concentrations in sediments were modeled for potential toxicity using three approaches: (1) probable effects quotients (PEQs) based on total recoverable metals (TRMs) and simultaneously extracted metals (SEMs); (2) SEMs corrected for acid-volatile sulfides (AVS; i.e., ∑SEM − AVS); and (3) ∑SEM − AVS normalized to the fractional organic carbon (f_oc) (i.e., ∑SEM − AVS/f_oc). The most highly metal-contaminated sample (∑PEQ_TRM = 132; ∑PEQ_SEM = 54; ∑SEM − AVS = 323; and ∑SEM − AVS/_foc = 64,600 umol/g) from the UCR was dominated by weathered slag sediment particles and resulted in 80% mortality and 94% decrease in biomass of amphipods; in addition, this sample significantly decreased growth of midge by 10%. The traditional ∑AVS – SEM, uncorrected for organic carbon, was the most accurate approach for estimating the effects of metals in the UCR. Treatment of the toxic slag sediment with 20% Resinex SIR-300 metal-chelating resin significantly decreased the toxicity of the sample. Samples ∑SEM − AVS > 244 was not toxic to amphipods or midge in laboratory testing, indicating that this value may be an approximate threshold for effects in the UCR. In situ benthic invertebrate colonization studies in an experimental pond (8-week duration) indicated that two of the most metal-contaminated UCR sediments (dominated by high levels of sand-sized slag particles) exhibited decreased invertebrate colonization compared with sand-based reference sediments. Field-exposed SIR-300 resin samples also exhibited decreased invertebrate colonization numbers compared with reference materials, which may indicate behavioral avoidance of this material under field conditions. Multiple lines of evidence (analytical chemistry, laboratory toxicity, and field colonization results), along with findings from previous studies, indicate that high metal concentrations associated with slag-enriched sediments in the UCR are likely to adversely impact the growth and survival of native benthic invertebrate communities. Additional laboratory toxicity testing, refinement of the applications of sediment benchmarks for metal toxicity, and in situ benthic invertebrate studies will assist in better defining the spatial extent, temporal variations, and ecological impacts of metal-contaminated sediments in the UCR system.