Pulpotomy vs Pulpectomy Outcome.

Epidemiological data show a significant rate of failure of endodontic treatment of 20 to 50% worldwide, probably directly related to the difficulty of the procedure. A successful pulpotomy allows the preservation of a vital radicular pulp into the root canal. The presence of a biological tissue into the root canal is definitely more efficient than a \"complete\" filling with an inert material. It has been shown on animal and human studies that this pulp stump in contact with the biomaterial is able to regenerate a dentinal plug, with the same architecture as physiologic dentin.

Primary objective : To compare the success rates of root canal treatment (reference) and conservative treatment (pulpotomy) for treating inflamed dental pulp on permanent mature teeth.

Secondary objectives : (1) To describe the clinical and biological prognostic factors of these two treatments; (2) to assess the additional value of biomarkers expression levels as a prognostic tool for clinical decision making (radical vs. conservative treatment); (3) to assess the impact of treatment on post-procedural pain.

This trial aims to demonstrate the non-inferiority of conservative pulpal treatment over endodontic treatment.

Patients consulting in one of the seven study centers, presenting one of the indications retained for the trial and giving written informed consent will receive the treatment (endodontic treatment or conservative treatment) allocated by randomization (stratified over the clinical diagnosis of the pulp status).

The follow up of patients include, a phone call at D15, and visit at 1, 6, 12 and 14 post operative months. Clinical examination and Xrays at 6, 12 and 24 months) will be used to evaluate the success or failure of the treatment.

During the treatment, a sample of pulp tissue will be withdrawn and transferred to a molecular biology laboratory for analysis of inflammation biomarkers. The aim of this part of the sudy is to assess a putative relationship (1) of regulation of biomarkers expression and clinical diagnosis, and (2) of regulation of these biomarkers and success rate of pulpotomy.

Success/failure evaluation:

The primary endpoint is the time to necessity of endodontic reintervention (analysed as a time to failure). This study will use an Intention To Treat analysis as its main assessment ; a secondary assessment accounting for peroperative conversions will assess the practical impact of these conversions. We will distinguish

* Direct failure (means that the failure is directly correlated to the treatment) : the reintervention need is due to the evolution of the treated tooth. This includes delayed onset of desmodontitis, periodontal space enlargement and/or periapical/periradicular radiolucency (PAI\>2) demonstrating an infection of the root canal system (filled by either pulpal stump or filling material).
* Indirect failure (means that the cause of the failure is not directly related to the endodontic treatment choice) : any event leading to endodontic reintervention indication NOT caused by radicular infection or restoration failure attributable to inadequate restoration. For example : new need of post-placement for treatment of loss of another tooth, unexpected progression of periodontal disease.

Both these failure modes are of interest for analysis : the direct failure time is an indication if the intrinsic value of a therapy, whereas the gross (direct+indirect) failure time is an indication of its clinical relevance (a good therapy applicable in rare cases may be less interesting than a mediocre but widely applicable one).

Statistical analysis:

The classical methods of descriptive analysis will be used to describe the raw results.

In order to make inferences directly on possible clinical results, this study will be analyzed in a Bayesian framework.

This study has been designed in reference with a frequentist demonstration of non-inferiority.

A non-inferiority trial with first and second species error rates α and β has the same operational characteristics as a superiority (unilateral trial) of error rates alpha and beta, which in turn needs the same study size as a comparison (=bilateral) trial with error rates α and 2β.

The final planned size of the trial is established as follows :

* Ideal plan : a nonparametric comparison (logrank test) fulfilling these goals according to this plan needs 158 patients overall under \"perfect information\" assumptions (no loss to follow-up, single analysis)
* Loss to follow-up : the expected loss to follow-up will cause about 22% of included patients to drop out of the study before final analysis this leads to include 194 patients overall.

Sequential analysis : since we wish to be able to follow the progress of the study, and to interrupt it if the main goal is reached, we choose to use a sequential analysis. A Pocock scheme needs to increase the sample size by 16% , leading to plan to recruit 226 patients overall.