Western corn rootworm, Diabrotica virgifera virgifera (LeConte) (Coleoptera: Chrysomelidae), is a major pest of maize in the United States. Transgenic maize producing insecticidal toxins from the bacterium Bacillus thuringiensis (Bt) have been used to manage this pest since 2003. Refuges of non-Bt maize have been used to delay resistance to Bt maize by western corn rootworm, and are planted in conjunction with maize producing single or multiple (i.e., pyramids) Bt toxins. Two Bt toxins, Cry3Bb1 and Gpp34/Tpp35Ab1, were used individually before being combined as a pyramid, at which point resistance had already evolved to Cry3Bb1. Pyramids targeting western corn rootworm therefore contained at least one toxin to which resistance had evolved. Western corn rootworm has now evolved resistance to all four commercially available Bt toxins used for rootworm management. We used laboratory and field-generated data to parameterize a deterministic model to simulate the effectiveness of refuges and Bt pyramids to delay resistance to Bt maize in western corn rootworm. Resistance to the pyramid of Cry3Bb1 with Gpp34/Tpp35Ab1 evolved more rapidly when resistance to Cry3Bb1 was already present. This effect arose when model conditions affecting initial resistance allele frequency, inheritance of resistance, and fitness costs were varied. Generally, resistance evolved faster when initial resistance allele frequencies were higher, inheritance of resistance was nonrecessive, and fitness costs were absent, which is consistent with previous models that simulated resistance evolution. We conclude that new transgenic pyramids should pair novel, independently acting toxins with abundant refuges to minimize the risk of rapid resistance evolution.