Dutch elm disease (DED) is a vascular wilt disease caused by the pathogens Ophiostoma ulmi and Ophiostoma novo-ulmi with multiple ecological phases including pathogenic (xylem), saprotrophic (bark) and vector (beetle flight and beetle feeding wound) phases. Due to the two DED pandemics during the twentieth century the use of elms in landscape and forest restoration has declined significantly. However new initiatives for elm breeding and restoration are now underway in Europe and North America. Here we discuss complexities in the DED 'system' that can lead to unintended consequences during elm breeding and some of the wider options for obtaining durability or 'field resistance' in released material, including (1) the phenotypic plasticity of disease levels in resistant cultivars infected by O. novo-ulmi; (2) shortcomings in test methods when selecting for resistance; (3) the implications of rapid evolutionary changes in current O. novo-ulmi populations for the choice of pathogen inoculum when screening; (4) the possibility of using active resistance to the pathogen in the beetle feeding wound, and low attractiveness of elm cultivars to feeding beetles, in addition to resistance in the xylem; (5) the risk that genes from susceptible and exotic elms be introgressed into resistant cultivars; (6) risks posed by unintentional changes in the host microbiome; and (7) the biosecurity risks posed by resistant elm deployment. In addition, attention needs to be paid to the disease pressures within which resistant elms will be released. In the future, biotechnology may further enhance our understanding of the various resistance processes in elms and our potential to deploy trees with highly durable resistance in elm restoration. Hopefully the different elm resistance processes will prove to be largely under durable, additive, multigenic control. Elm breeding programmes cannot afford to get into the host-pathogen arms races that characterise some agricultural host-pathogen systems.