Objectives: This study investigated the long-term survival and the prognostic factors of endodontic treatments performed in a dental teaching hospital. The aim was to calculate the probabilities of success or failure according to the follow-up extent and to assess the time allowed for a complete periapical healing. Materials and methods. A cohort of 185 teeth were re-examined 1–4 years after treatment. The outcome was assessed on the basis of radiographic and clinical criteria as success, uncertain or failure. A survival analysis using the Cox model was used (i) to explore tooth survival and periapical healing over the time and (ii) to highlight the predictive factors of treatment outcome. Results. After 2 years, the appearance of an apical periodontitis remained lower than 3.5%, whereas only 22.8% of periapical healed cases were noti fied. The prognosis factors are: (i) for teeth with initial healthy periapical conditions, coronal leakage (p = 0.002) with the higher risk of failure (RR = 19.77), absence of correspondence filling length/shaping = 0.026), type of teeth (p = 0.041) and (ii) for teeth with apical periodontitis, number of root canals (p = 0.000,91), correspondence filling length/shaping length ( p = 0.017) and over-filling (p = 0.09). New periapical lesions or tooth loss were recorded after 2 years. Half of the successful cases of periapical healing were observed during the follow-up from 2 – 4 years. Conclusions. This longitudinal study shows that coronal leakage is responsible for late failure and that periapical healing is long to achieve. Therefore, endodontic treatments may require a follow-up of over 2 years