Edman Tynelius 2010.

Study characteristics
Methods	3‐arm parallel RCT
Participants	Numbers recruited: 75 participants, 25 per group Sex (M:F): 30:45 Age (years): group 1 mean 14.1 (SD 1.3), group 2 mean 14.7 (SD 1.8), group 3 mean 14.3 (SD 1.5) Inclusion criteria: people who had completed upper and lower fixed appliances, with: no other previous experience of orthodontic treatment; permanent dentition; crowding in both arches; normal skeletal and dentoalveolar sagittal, vertical and transverse relationships; Class I molar relationship or up to 3 mm anterior or posterior deviation; treatment plan involving extraction of 4 premolars Exclusion criteria: no additional criteria specified Setting: Swedish National Health Service clinic
Interventions	Two types of removable retainer and a positioner  Group 1: upper clear plastic retainer and a ⌀ 0.7 mm stainless steel‐bonded retainer bonded to the lower canines only. Clear plastic retainers were worn 22‐24 hours for the first 2 days and then nights only after this. The bonded retainer was 0.7 mm stainless steel and bonded to the lingual of the canines only Group 2: upper clear plastic retainer with interproximal stripping of the lower anterior teeth, but no mandibular retainer. Clear plastic retainers were worn 22‐24 hours for the first 2 days and then nights only after this. Interproximal stripping was undertaken 5‐6 weeks before debond and at debond. Group 3: prefabricated positioner covering all the maxillary and mandibular teeth. Positioner was worn 30 min during the day (and participant was encouraged to bite into it) and also during sleep. It was a soft plastic device covering all teeth In the second year of retention, participants wore their removable retainers every other night. After 2 years, all retainers were stopped and bonded retainers removed.
Outcomes	Stability: based on LII in the upper and lower labial segments, intercanine width, intermolar width, arch length, overjet and overbite. Stability data collected after 1 year Time points: 1, 2 and 5 years post debond. Change data available for years 1 and 2. Final LII data only in both arches were available for 5 years post‐debond
Notes	No measurement of survival of retainers, adverse effects on health or patient satisfaction It was reported that 2 participants did not comply with the positioner. On contact with the study author, the failure rates for the bonded retainers after 1 year were: 18 participants had no failures, 5 had 1 failure, 1 had 2 failures and 1 had 4 failures. Cost‐effectiveness also reported, but not an outcome in this review.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"The generation of randomisation sequence was performed in blocks of five to ensure that equal numbers of patients were allocated to each of the three retention groups". The study author confirmed appropriate random number generation.
Allocation concealment (selection bias)	Low risk	Direct communication with the study author confirmed that they used adequate concealment, and neither the participant nor clinician could have known of the intervention to be used in advance. This was done by placing ballots in a hat by an independent observer, so that the clinician could not know what intervention was going to be drawn out.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	The paper reported that the outcome assessor was blinded to the retention protocol.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Dropouts were clearly reported and not included in the final analysis.
Selective reporting (reporting bias)	Low risk	All the data were reported over the 5‐year study period.
Other bias	Low risk	No other areas of bias identified