Bolla 2012.

Methods	2‐arm parallel randomised controlled trial
Participants	Numbers recruited: 85 (56 females and 29 males) Mean age: 21.9 years Inclusion criteria: people who had been treated with fixed appliances, with: good oral hygiene mild crowding in upper labial segments treated non‐extraction post‐treatment Class I relationship with an overjet and overbite of 1‐3 mm correct anterior Bolton's index Exclusion criteria: dental agenesis or anomalies enamel decalcification or white spots on lingual surface of anterior teeth periodontal disease considerable dental abrasion parafunctional habits such as nocturnal bruxism Setting: university clinic in Italy and participants had to pay for treatment (information provided by authors)
Interventions	Comparison: polyethylene ribbon retainer vs. multistrand retainer Group 1: polyethylene ribbon retainer were silanised, 14 µm, unidirectional, glass fibre devices embedded in a resin matrix Group 2: multistrand retainer were 0.0175" dead‐soft passive multistrand wire Both types were bonded lateral incisor to lateral incisor in upper, and canine to canine in the lower All were placed under rubber dam using a standardised bonding process
Outcomes	Outcome relevant to review: survival of retainers Number of detachments or bond failures recorded after 6 years
Notes	No measures of patient satisfaction, effect on health or stability were recorded
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Contacted authors who confirmed, "the operator tossed a coin for every patient" after agreeing to be in the study
Allocation concealment (selection bias)	Low risk	Contacted authors who confirmed, "the operator tossed a coin for every patient" after agreeing to be in the study
Blinding of outcome assessment (detection bias) All outcomes	High risk	Not feasible as the outcome assessor could not be blinded to the presence or absence of the wire
Incomplete outcome data (attrition bias) All outcomes	High risk	Paper recorded no drop‐outs, but on contacting the authors reported that, "we described only the patients that completed the trial, we estimate, now, we had a drop out of 15‐20 per group"
Selective reporting (reporting bias)	High risk	Contacted authors who reported, "we started to check the periodontium, but it was making the visit too long for the administration therefore we had to terminate that evaluation" They also started "evaluating the bacterial flora taking a sample every 3 months, unfortunately we ran out of funds"
Other bias	Low risk	No other areas of risk of bias identified