Millett 2007.

Methods	2‐arm parallel randomised controlled trial
Participants	Numbers recruited: 85 (46 females and 39 males) Mean age: not reported Inclusion criteria: all participants had undergone fixed appliance treatment in at least the lower arch with satisfactory alignment and: were in good health brushed teeth at least twice a day had full complement of teeth in lower labial segment with these teeth being of normal size and shape were willing to comply with trial protocol included extraction and non‐extraction cases Exclusion criteria: general medical health problems such as diabetes mellitus, epilepsy, physical or mental disability poor periodontal health uncontrolled caries absent or unusually shaped lower incisors cleft palate or other severe facial deformities Setting: hospital department in Cork, Republic of Ireland
Interventions	Comparison: lower bonded retainer vs. lower vacuum‐formed retainer Group 1: bonded retainer was 0.018" stainless steel multistrand wire bonded to lingual of all lower incisors and canines Group 2: vacuum‐formed retainer was full‐occlusal coverage worn on a nights only basis from the day it was fitted In the upper arch, the clinician chose the upper retainer (Hawley or vacuum‐formed)
Outcomes	Outcomes relevant to review: stability, survival of retainers, adverse effects on health and participant satisfaction Stability assessed by Little's Irregularity Index. The initial protocol discussed the use of PAR index, but these data were not collected Survival of retainers ‐ vacuum‐formed retainers were assessed as failed if they fractured or were lost (but not worn); bonded retainers were assessed as failed if debonded (clarified this with the authors) Adverse effects on health ‐ caries (evidence of decalcification) and periodontal health (gingival bleeding and periodontal probing > 3 mm) was assessed for lower anterior teeth Participant satisfaction ‐ participant questionnaire assessing retainer wear and acceptability recorded. Data for both retainers available for the following questions: "How acceptable did you find the retainer brace to wear?" "How easy was it to keep your retainer brace clean?" "How happy are you with the way your teeth look now?" (at the end of the retention period) Outcomes collected 1 year after debond
Notes	Additional outcome was operator perception of each retainer type ‐ this will not be reported here, as this is not part of this review.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Authors confirmed computer‐generated random sequence
Allocation concealment (selection bias)	Low risk	Sealed consecutive opaque envelopes in a separate co‐ordinating centre
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Contact with the author confirmed the assessor was blinded for the intervention type
Incomplete outcome data (attrition bias) All outcomes	Low risk	Drop‐outs were reported. Authors confirmed the data was analysed on an intention‐to‐treat basis
Selective reporting (reporting bias)	High risk	Not all outcomes were reported in the abstracts that had been described in the protocol (PAR index for occlusal change was not collected). Further data will be reported after 2 years of wear
Other bias	Unclear risk	As this study was only available as research abstracts, it is unclear as to whether other bias exists