Objective This study tested the hypothesis that the use of ethylenediamine tetraacetic acid (EDTA) as a final irrigant causes canal wall erosion only after prolonged use of 5.25% sodium hypochlorite (NaOCl) as the initial irrigant. Methods Two irrigation protocols were employed: (1) 5.25% NaOCl as the initial irrigant for 10 min and 17% EDTA as the final irrigant for 2 min; and (2) 5.25% NaOCl for 60 min and 17% EDTA for 2 min. Flexural strengths of dentine beams prepared from human mid-coronal dentine were evaluated using a miniature three-point bending device after they were irrigated with either protocol. Transmission electron microscopy was used to examine the conditions of instrumented canal walls irrigated with either protocol after they were bonded with methacrylate resin-based sealers. Results Immersion of dentine beams in NaOCl for 60 min instead of 10 min prior to the use of EDTA resulted in a significant ( p < 0.001) decline in flexural strength. For both NaOCl initial rinsing times, a 2-min final irrigation with EDTA created similar 0.5 μm thick demineralisation fronts. However, extensive surface and subsurface erosion of radicular dentine was seen only when the dentine was immersed in NaOCl for 60 min. Conclusion The apparent aggressiveness of EDTA in causing canal wall erosion is attributed to the prolonged use of NaOCl. The associated decline in dentine flexural strength has potential clinical relevance when thin pulp chamber dentine is immersed in NaOCl for lengthy periods during canal instrumentation. This may render root-treated teeth more prone to vertical fracture.