This manuscript was written in accordance with the STrengthening the Reporting of OBservational Studies in Epidemiology (STROBE) guidelines.

All patients referred for treatment of non-vital anterior permanent teeth to the Department of Pediatric Dentistry, Behavioral Science and Forensic Dentistry at University of Oslo from 2008 to 2021 were included in the study. The inclusion criteria were permanent anterior teeth diagnosed with pulp necrosis, apexification with a bioceramic apical plug and age under 18 years. In total, 48 patients met the inclusion criteria. Two patients were lost to follow-up and the study group comprised 46 patients (34 boys and 12 girls) with 51 teeth.

Postgraduate dentists under supervision of a specialist in paediatric dentistry performed treatments and follow-ups. Data were extracted from dental records. Preoperative data included gender, tooth, child age and stage of root development at pulp necrosis, aetiology of necrosis, clinical signs of infection, radiographic evidence of pulp necrosis and root resorption.

The stage of root development prior to treatment was classified according to Moorrees classification (Moorrees et al. 1963) and divided into three groups of increasing tooth maturity: Very immature (ranged from initial root formation to root length completed with parallel ends), immature (full root length and apical foramen half closed) or mature (full root length and closed apex). Five teeth had sustained external apical root resorptions (three teeth) or root fractures (two teeth) which left the teeth with open apices and short roots prior to apexification, and these were classified as very immature. These five teeth were excluded from the analyses of spontaneous root fracture. Root development was categorised as very immature or immature in the analyses as no teeth had mature root development.

Aetiology of pulp necrosis was assessed as dental trauma or developmental defects such as invagination.

Clinical signs of infection included sensitivity to percussion and palpation, sinus tract or abscess, and was classified as present or absent in the analyses. Pulp necrosis was verified using an electric pulp tester (Dahlin Electronic Pulp Tester) and cold pulp sensibility test (Endo Ice; Roeko, 71 Langenau, Germany; Endo Frost, Coltène Whaledent Roeko, Germany).

Radiographic evidence of pulp necrosis was reported as periapical lesion or arrested root development. Root resorption included infection related root resorption and replacement resorption and were classified as present or absent.

Periapical status at apexification and at follow-ups was scored on digital radiographs using the periapical index (PAI) (Ørstavik et al. 1986). PAI classifies radiographically the tooth’s periapical status on a grade from 1 (healthy) to 5 (severe apical periodontitis). Before scoring the radiographs, the examiner (FSA) was calibrated using the PAI kit where 100 reference radiographs, not related to this study, were evaluated and graded on a scale from 1 to 5. Two score sets were produced 10 days apart and was compared to a standard established by Ørstavik et al. to determine agreement (Ørstavik et al. 1986). Kappa values showed substantial agreement for inter-examiner (κ = 0.78) and intra-examiner (κ = 0.79) reproducibility against the standard. All periapical radiographs were presented on a computer screen and blindly evaluated in a dark room by the first author (FSA). The periapical area was graded on a scale from 1 to 5 and compared with five reference radiographs. Any uncertainty regarding PAI-score for a particular tooth was discussed with the other authors (TW and IJB) and resulted in joint assessment.

Treatment was performed according to a standard protocol of root canal treatment used at The Department of Pediatric Dentistry, Behavioral Science and Forensic Dentistry at University of Oslo, based on the recommendations from European Society of Endodontology (European Society of Endodontology 2006). Teeth were isolated with rubber dam after injection of local anaesthesia (Septocaine (Articaine) 40 mg/ml + 5 µg/ml, Septodont Inc, New Castle, DE). Chemo-mechanical debridement was achieved using endodontic files and irrigation with 1% sodium hypochlorite (NaOCL) and 17% ethylenediaminetetraacetic acid (EDTA). Calcium hydroxide paste was used as an inter-appointment medicament. At the time of obturation, Biodentine™ (Septodont, Saint Maur des Faussés, France), white MTA ProRoot® (Dentsply, Tulsa Dental, Tulsa, OK, USA) or white MTA Angelus® (Angelus Dental Solutions, Londrina, Parana, Brazil) was placed in the apical portion of the root using either a Micro-Apical Placement System (Dentsply Maillefer, Vevey Switzerland) or an endodontic plugger. A periapical radiograph was taken to confirm the correct position of the bioceramic plug. The root canal was obturated with thermo-plasticized GP (Obtura Spartan, Fenton, MO, USA) or with GP placed with cold lateral condensation technique (one tooth) with AH Plus® sealer (Dentsply Sirona Endodontics) and IRM® (Dentsply) in the coronal portion of the root. Obturation with GP was conducted in the same appointment as apexification when Biodentine was used, whilst in a second appointment when MTA was used for apexification. Composite resin material was used to seal the access cavity.

Follow-up data included the presence or absence of clinical symptoms and periapical status. Any pain or soft tissue lesion associated with the tooth was defined as a clinical symptom. Tenderness to percussion or palpation was considered a subjective symptom and accepted if unaccompanied by any other clinical signs or symptoms. Outcomes were recorded at follow-ups and were based on radiographic outcome (PAI-score) and by criteria from Holden et al. (Holden et al. 2008). The radiographic outcome was considered healed when PAI-score was 1 or 2, or non-healed when PAI-score was 3, 4 or 5. The outcome according to Holden’s criteria included clinical symptoms in addition to radiographic measures as shown in Table 1. In the analyses, the radiographic outcome and outcome according to Holden were dichotomized as healed (healed or healing) or non-healed.

The position of the bioceramic plug was scored as ideal (plug was positioned in the apical portion of the root) or non-ideal (overfilled or underfilled) at apexification. The radiopacity of the bioceramic material was assessed on radiographs at apexification and at last follow-up and categorised as no change of radiopacity or change of radiopacity. A spontaneous root fracture was recorded when it was seen and not caused by a second injury. The five teeth that had undergone root fracture or had external root resorption before treatment were excluded from these analyses.

The statistical analyses were conducted using IBM SPSS Statistics for Windows, version 28 (IBM Corp., Armonk, N.Y., USA). Data were cross-tabulated and tested with chi-square statistics, and presented as frequency, mean and standard deviation. The level of significance was set at 5%.