The aim of this study was to assess the amount of apically extruded debris during filling removal with WaveOne Gold (WOG), ProTaper Universal Retreatment (PTR), D-RaCe Retreatment (DRR) or hand files (HF), to compare the working time during filling removal, and to describe failures of NiTi instruments. Forty mesiobuccal roots of maxillary first molars were prepared with WOG Primary, obturated and divided into 4 groups (n=10), according to the instruments used: WOG, PTR, DRR or HF. Distilled water was used as irrigant and the extruded debris were collected in Eppendorf tubes and dried. The amount of extruded debris was determined by subtracting the final from the initial weight. The time of filling removal for each canal was recorded and the instruments used were analyzed pre and post-operatively by SEM. Kruskal-Wallis and Dunn's test analyzed extruded debris data and ANOVA, followed by Tukey's test, compared the working time data (α=0.05). Instrument deformation and fracture were described. WOG produced significantly less debris compared with HF and DRR (p<0.05), and similar to PTR (p>0.05). HF, PTR and DRR showed no significant difference (p>0.05). Working time in HF group was significantly higher than others (p<0.05). SEM analyses showed, from the 18 instruments evaluated, 3 fractures and 10 deformations. All instruments tested caused debris extrusion. WOG was associated with less extrusion than DRR and HF. Filling removal with HF was slower than with the other instruments. All NiTi systems presented fracture and deformation.