Cepa 2017.

Methods	Trial design: 3‐year follow‐up, randomised parallel group trial Parallel group randomised clinical trial in which the individuals were randomly assigned to the 2 trial groups consisting of ball and telescopic attachments. Randomisation was carried out by a statistician not involved in the trial using sealed envelopes. The envelope containing information on the abutments to be used was opened by a dental nurse
Participants	Inclusion criteria: participants recruited for the trial were dissatisfied with their conventional mandibular overdenture Exclusion criteria: general contraindications for oral surgical procedures such as infectious or metabolic diseases, cardiovascular disease and pregnancy; local contraindications (e.g. tumours, ulcers); alcohol abuse or smoking; psychological disease; non‐compliance and intolerance of pre‐ or post‐operative medication in the presence of chronic disease Age at baseline (years): mean 65.2 Gender (M/F): not specified Number of mandibular overdentures randomised: 25 (mandibular 2‐implant overdentures: 12 ball; 13 telescopic attachment) Number of mandibular overdentures evaluated at 1 year: 23 (12 ball; 11 telescopic attachment) Number of mandibular overdentures evaluated at 2 years: 21 (11 ball; 10 telescopic attachment) Number of mandibular overdentures evaluated at 3 years: 16 (11 ball; 5 telescopic attachment)
Interventions	Mandibular 2‐implant overdentures ‐ unsplinted prosthodontic design Test group 1 (n = 12): Ankylos® (Dentsply Implants, Mannheim, Germany) ball attachments with connecting matrices were polymerized into spacers of the metal framework using a self‐curing resin (Pattern Resin LS, GC Europe, Leuven, Belgium) Test group 2 (n = 13): prefabricated telescopic (conus) attachments with connecting matrices were polymerized into spacers of the metal framework using a self‐curing resin (Pattern Resin LS, GC Europe, Leuven, Belgium)
Outcomes	Prosthodontic maintenance by general categorization Patient satisfaction (trial did not use validated criteria for this outcome) Data collection was performed on the day of overdenture delivery and after 1, 2 and 3 years
Notes	Location: Germany Funding source: this investigation was supported by a grant from Dentsply Implants, Mannheim, Germany Patient satisfaction data presented did not use validated criteria Authors contacted by email (7 February 2017): requested full prosthodontic maintenance categorization to facilitate data analysis ‐ no reply received Details of ethical approval: Freiburg Regional Council and the Ethics Committee of the Medical‐Center, Freiburg, Germany (Votum number: 105/04)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "The patients were randomly assigned to the two trial groups consisting of ball and telescopic attachments" Comment: the trial does not include any description of how the sequence was generated
Allocation concealment (selection bias)	Low risk	Quote: "Randomization was carried out by a statistician not involved in the study using sealed envelopes. The envelope containing information on the abutments to be used was opened by a dental nurse"
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: "As the prosthetic reconstruction was visible, blinding of the investigators was not possible" Comment: although participants and personnel were not blinded, we judge the risk of bias as low since it is an operative procedure and the outcomes are not likely to be influenced by lack of blinding of participants and personnel
Blinding of outcome assessment (detection bias) Objective outcomes	Low risk	The trial did not report blinding, however blinding is not possible due to the nature of the intervention, but we judge the risk of bias as low regarding prosthodontic maintenance due to being objective and is not likely to be influenced by lack of blinding
Blinding of outcome assessment (detection bias) Patient‐reported outcomes	High risk	The trial did not report blinding, however blinding is not possible due to the nature of the intervention, but we judge the risk of bias as high regarding patient satisfaction due to being subjective and is likely to be influenced by lack of blinding
Incomplete outcome data (attrition bias) All outcomes	High risk	Total losses: 9/25 (36%) Ball group losses: 1/12 Telescopic group losses: 8/13 Ball attachment group: 12 participants received IODs. 12 participants were followed up after 1 year, whereas 11 participants were followed up after 2 and 3 years. 1 participant was deceased within the second year of observation Prefabricated telescopic group: 13 participants received IODs. 11 participants were followed up after 1 year. 1 participant deceased, and 1 refused to further participate in the investigation. 10 participants were followed up after 2 years, as another participant refused to further participate in the investigation. Only 5 participants were followed up after 3 years of observation because further 5 dropouts occurred. Another participant deceased, and 4 more participants refused to further participate in the follow‐up appointments. 4 dropouts of the telescopic group were changed to ball attachments during their ongoing aftercare appointments in the department In addition, 'per protocol' analysis was performed when analysing the patient satisfaction outcome
Selective reporting (reporting bias)	Low risk	The trial protocol is not available, but we judge the risk of bias as low due to reporting key outcomes that are expected to be reported for such a trial
Other bias	Low risk	None noted