The primary aim of the study is to compare the outcome of one vs two visits endodontics (root canal treatments) performed in general dental practice of infected (necrotic) teeth, one and four years after treatment. Secondary aims are to compare the root filling quality and the frequency of complications for both treatment modalities. A pilot study will identify the shortcomings of the protocol, which will be revised accordingly. A multi-center study in several clinics from five counties throughout Sweden is planned. Consecutive eligible patients (see inclusion and exclusion criteria) with one tooth each will be included. The study sample size calculation was based on a judgement that a minimal worthwhile effect of less than 10% was not clinically relevant. The presumption was that successful outcomes would be of 75% and 65% for the different treatment arms, respectively. Three hundred and fifty six patients per group are thus needed to detect that difference as statistical significant with a power of 80% and an alfa-error of 5%. To compensate for around 30% drop-outs a number of 1000 patients will be recruited. Patients will be randomly enrolled in two arms: one in which the endodontic treatment (root canal treatment) is performed in one visit, and in the other arm in two visits. The allocation sequences for one vs multiple visit endodontics will be computer generated followed by a dark colored envelope concealed allocation. The patients will be contacted by telephone 5-7 days post-operatively, and a control visit at the clinic will be scheduled one month post-operatively to assess complications related to the treatment. A clinical control will be performed at the clinic one and four years after treatment. Intra-oral radiographs will be taken. The radiographs will be examined by two blinded and independent observers. A successful outcome is defined by clinical and radiographic normalcy. The statistical methods selected to analyse the primary outcome will be the Chi square test of 2x2 tables, or Fisher´s exact test when proportions are small and the expected value \<5. The calculations will be two-tailed. Results will be considered statistically significant at p\<0.05.