Retention procedures for stabilising tooth position after treatment with orthodontic braces

Summary of findings for the main comparison. Removable retainer versus fixed retainer to stabilise tooth position

Removable retainer versus fixed retainer to stabilise tooth position
Patient or population: people who had received fixed appliance treatment Setting: specialist orthodontic practice and hospital orthodontic department Intervention: removable retainers Comparison: fixed retainers
Outcomes	*Anticipated absolute effects (95% CI)**		Relative effect (95% CI)	No of participants (studies)	Quality of the evidence (GRADE)	Comments
	Risk with fixed retainers	Risk with removable retainers
Stability ‐ Little's Irregularity Index in lower arch Ideally 0 mm 12 months	Mean = 0.43 mm	0.6 mm more (0.17 more to 1.03 more)	‐	84¹ (1 RCT)	⊕⊕⊝⊝ low²	1 study with low risk of bias suggested that there is a statistically significant reduction in relapse with the fixed retainer, but this is unlikely to be clinically significant
Failure of retainers (lower)	See comment		‐	133 (2 RCTs)	⊕⊝⊝⊝ very low³	2 studies at high risk of bias making multiple comparisons of different types of fixed and removable retainers showed inconsistent results
Adverse effects on health Evidence of gingival bleeding 12 months	617⁴ per 1000	327 per 1000 (191 to 543)	RR 0.53 (0.31 to 0.88)	84¹ (1 RCT)	⊕⊕⊝⊝ low²	1 study also looked at adverse effects on dental health: caries and periodontal pocketing. There was low quality evidence to suggest that there was additional periodontal pocketing with the fixed retainer
Patient satisfaction How acceptable was the retainer to wear? 0 to 100 VAS (100 is very satisfied) 12 months	Mean = 91.62	12.84 lower (18.6 lower to 7.09 lower)	‐	81 (1 RCT)	⊕⊕⊝⊝ low²	1 RCT found that the participants thought the fixed retainer was easier to keep clean but found no difference in the appearance of their teeth after 1 year of retention
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; RCT: randomised controlled trial; RR: risk ratio; VAS: visual analogue scale
GRADE Working Group grades of evidence High quality: We are very confident that the true effect lies close to that of the estimate of the effect Moderate quality: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low quality: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low quality: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect
¹ Removable thermoplastic versus fixed multistrand ² Downgraded twice: once for high risk of bias and once for imprecision (only one study) ³ Downgraded three times: once for high risk of bias, once for imprecision and once for inconsistency ⁴ From control group (fixed) of study

Summary of findings 2. Fixed retainer versus fixed retainer to stabilise tooth position

Fixed retainer versus fixed retainer to stabilise tooth position
Patient or population: people who have received fixed appliance treatment Setting: specialist practice or university orthodontic clinic Intervention: one type of fixed retainer Comparison: another type of fixed retainer
Outcomes	*Anticipated absolute effects (95% CI)**		Relative effect (95% CI)	No of participants (studies)	Quality of the evidence (GRADE)	Comments
Outcomes	Risk with fixed retainers	Risk with fixed retainers	Relative effect (95% CI)	No of participants (studies)	Quality of the evidence (GRADE)	Comments
Stability ‐ Little's Irregularity Index	‐	‐	‐	0 (0)	‐	Not reported
Failure of retainers (lower arch) 18 months to 6 years	336¹ per 1000	370 per 1000 (259 to 528)	RR 1.10 (0.77 to 1.57)	228 (3 RCTs)²	⊕⊕⊝⊝ low³	2 of these RCTs also compared failure rates in the upper arch and found no statistical difference in failure rate (RR 1.25, 95% CI 0.87 to 1.78). 1 study with a high risk of bias showed no evidence of a difference in failure rates between 3 types of metal fixed retainers
Adverse effects on health	‐	‐	‐	0 (0)	‐	Not reported
Patient satisfaction	‐	‐	‐	0 (0)	‐	Not reported
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; RCT: randomised controlled trial; RR: risk ratio
GRADE Working Group grades of evidence High quality: We are very confident that the true effect lies close to that of the estimate of the effect Moderate quality: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low quality: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low quality: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect
¹ From control group ² Mandibular polyethylene ribbon bonded retainer versus mandibular multistrand retainer ³ Downgraded twice: once for high risk of bias and once for imprecision

See: Summary of findings for the main comparison Removable retainer versus fixed retainer to stabilise tooth position; Summary of findings 2 Fixed retainer versus fixed retainer to stabilise tooth position; Summary of findings 3 Removable retainer versus removable retainer to stabilise tooth position

Summary of findings 3. Removable retainer versus removable retainer to stabilise tooth position

Removable retainer versus removable retainer to stabilise tooth position
Patient or population: people who have received fixed appliance treatment Setting: specialist practice, university or hospital orthodontic department Intervention: one type of removable retainer or wear time Comparison: another type of removable retainer or wear time
Outcomes	*Anticipated absolute effects (95% CI)**		Relative effect (95% CI)	No of participants (studies)	Quality of the evidence (GRADE)	Comments
Outcomes	Risk with Removable Retainers	Risk with Removable	Relative effect (95% CI)	No of participants (studies)	Quality of the evidence (GRADE)	Comments
Stability ‐ Little's Irregularity Index in upper arch Ideally 0 mm 12 months	‐	‐	‐	419 (3 RCT)	⊕⊕⊝⊝ low to ⊕⊕⊕⊝ moderate	Data used from 3 studies comparing different removable retainers used for different time periods 1 study at low risk of bias comparing part‐time thermoplastic vs. full‐time thermoplastic retainer showed no evidence to suggest a difference (graded moderate quality evidence: downgraded once due to imprecision). There was also data from additional study with a low risk of bias that could not be analysed using Review Manager 5 that suggested no evidence of difference in relapse when thermoplastic retainers were worn full‐time or part‐time 1 study comparing part‐time and full‐time Hawley retainers showed no evidence of a difference (graded low quality of evidence: downgraded twice due to high risk of bias and imprecision) 1 study at high risk of bias comparing upper Hawley retainers and upper thermoplastic retainers suggested there was statistical, but not clinically significant, reductions in relapse with thermoplastic retainers (graded low quality: downgraded twice due to high risk of bias and imprecision)
Stability ‐ Little's Irregularity Index in lower arch Ideally 0 mm 6‐12 months	‐	‐	‐	643 (4 RCT)	⊕⊕⊝⊝ low to ⊕⊕⊕⊝ moderate	Data used from 4 studies comparing different removable retainers used for different time periods. 1 study at low risk of bias comparing part‐time thermoplastic vs. full‐time thermoplastic retainer showed no evidence to suggest a difference (graded moderate quality evidence, downgraded once due to imprecision) 2 studies at high risk of bias suggested that stability was better for thermoplastic retainers vs. Hawley retainers, and for thermoplastic full‐time retainers vs. Begg (full‐time) retainers (both low quality evidence; downgraded twice due to high risk of bias and imprecision) 1 study with a high risk of bias comparing part‐time and full‐time wear of lower Hawley retainers found no evidence of any difference in relapse (low quality evidence: downgraded twice due to high risk of bias and imprecision). Data from an additional study with a low risk of bias that could not be analysed using Review Manager to support this evidence
Failure of retainers How many broke in total? 12 months	‐	‐	‐	457 (2 RCTs)	⊕⊕⊝⊝ low	2 studies provided data 1 study comparing Hawley retainers vs. part‐time thermoplastic retainers showed no difference in the number of retainers that were lost (low quality evidence, downgraded twice due to high risk of bias) 1 study with a high risk of bias comparing full‐time wear of Hawley retainers vs. full‐time thermoplastic retainers found greater failure rates in the lower thermoplastic retainers (low quality evidence, downgraded twice due to high risk of bias, and imprecision)
Adverse effects on health	‐	‐	‐	0 (0)	‐	Not reported
Patient satisfactionEmbarrassed to wear retainer? 12 months	72¹ per 1000	174 per 1000 (93 to 322)	RR 2.42 (1.30 to 4.49)	348 (1 RCT)²	⊕⊕⊝⊝ low	1 study comparing Hawley and thermoplastic retainers found participants who wore the Hawley retainers found them harder to wear as instructed, and judged them more negatively when comparing them to the fixed appliances they had worn. There was no evidence of a difference between the Hawley and thermoplastic retainer groups in terms of discomfort or ability to wear them away from home (low quality evidence, downgraded twice as one study with high risk of bias and imprecision)
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; RCT: randomised controlled trial; RR: risk ratio
GRADE Working Group grades of evidence High quality: We are very confident that the true effect lies close to that of the estimate of the effect Moderate quality: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low quality: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low quality: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect
¹ From study control group ² Hawley versus thermoplastic retainers

Background

Description of the condition

Retention is the phase of orthodontic treatment that attempts to keep teeth in the corrected positions after correction with orthodontic (dental) braces. Without retention there is a tendency for the teeth to return to their initial position. This unfavourable change from the corrected position is known as relapse. The causes of relapse are not fully understood, but are felt to relate to recoil of the fibres that hold the teeth in the jaw bone; pressures from the lips, cheeks and tongue; further growth and the way the teeth meet together (Melrose 1998). To minimise relapse almost every person who has orthodontic treatment will require some type of retention.

Description of the intervention

Retention can be achieved by placing appliances, called retainers, on the teeth, or by undertaking additional or 'adjunctive' procedures to the teeth or the surrounding structures.

The retainers can either be removable, so that the person can take them out to clean, or they can be fixed to the teeth (Atack 2007). Bonded retainers are usually glued or 'bonded' on the back (inside) of the front teeth. These bonded retainers are sometimes referred to as 'fixed retainers'. Some clinicians use an appliance called a positioner after treatment is completed: this helps to 'fine‐tune' the result and can then be worn part‐time to help reduce relapse. Appendix 1 contains a glossary of terms to help describe some of the common type of retainers and adjunctive procedures that are described in this review.

There is no recognised duration for the time that retainers need to be worn. It has been shown that if people stop wearing retainers after one to two years there is a risk of long‐term relapse of the teeth (Little 1981; Little 1988). Therefore, some clinicians prefer to retain for longer periods, sometimes indefinitely. It is also not clear how many hours a day removable retainers need to be worn: some people are asked to wear their retainer full‐time (24 hours a day), while other people are only asked to wear them part‐time (less than 24 hours a day).

Clinicians may also try to reduce relapse by using 'adjunctive' procedures to the teeth (hard tissues) or the surrounding gum (soft tissues). This can involve reshaping of the contact points between teeth known as interproximal reduction (Aasen 2005), or cutting the fibres around the neck of the tooth, that hold the tooth in the jaw bone, known as pericision (Edwards 1988). The glossary of terms in Appendix 1 provides more information on this.

How the intervention might work

Retainers or adjunctive procedures aim to maintain the teeth in the position they were in at the end of orthodontic treatment. Retainers fit over or around the teeth, and prevent them moving away from their final position. Adjunctive procedures may work by improving the contacts between teeth (interproximal reduction) or by cutting fibres that connect the teeth to the gum around the neck of the teeth. These fibres may pull the teeth back towards their original position.

In order for retainers to work, they must keep the teeth in position without doing any harm and be comfortable and acceptable for people to wear. Potentially they could cause damage to the teeth by collecting plaque and calculus, and by making it difficult for people to keep their teeth clean. This build‐up of plaque may cause decay (caries) or inflammation and damage to the surrounding gum. This damage to the gum can cause gingivitis (inflammation of the gums) or periodontal disease (loss of attachment of the tooth to its surrounding gum and bony socket).

Why it is important to do this review

The Cochrane Oral Health Group undertook an extensive prioritisation exercise in 2014 to identify a core portfolio of titles that were the most clinically important ones to maintain on The Cochrane Library (Worthington 2015). Consequently, the orthodontic expert panel identified this review as a priority title (Cochrane OHG priority review portfolio).

Retention is a key part of orthodontic treatment. Unless we can maintain the teeth in position after orthodontic treatment, the beneficial effects of the treatment can disappear. There are currently many different types of removable and fixed retainers and it is unclear which retainers are the best, and how long they should be used for. There is also little known about the possible benefits and risks of adjunctive procedures aimed at reducing relapse.

This review investigated the effectiveness of different retention strategies used to stabilise tooth position after treatment with orthodontic braces. It did not attempt to identify the causes of relapse. This review looked at the effects of retainers while in place, not the long‐term effects after they are no longer in use.

Objectives

To evaluate the effects of different retention strategies used to stabilise tooth position after orthodontic braces.

Methods

Criteria for considering studies for this review

Types of studies

Randomised controlled trials.

Types of participants

Children and adults who have had retainers fitted or adjunctive procedures undertaken following treatment with orthodontic braces. There was no restriction for the presenting malocclusion or type of active orthodontic treatment undertaken. The participants had to be followed up at least three months after completing their orthodontic treatment.

We excluded:

people who had surgical correction of the jaws;
people with a cleft lip or palate, or both, or other craniofacial syndrome;
people who had orthodontic treatment based on extractions alone or the fitting of a passive space maintainer, or both.

Types of interventions

Retainers or adjunctive techniques, or both, after treatment with orthodontic braces. We included only studies where the full course of definitive orthodontic treatment was completed ‐ therefore, we excluded data on retention strategies at the end of an initial phase of removable or functional appliance treatment.

Types of outcome measures

Primary outcome

Stability.

This could be assessed by an index of tooth irregularity, for example, Little's Irregularity Index (Little 1981), which measures how crooked anterior teeth are or by crowding (see glossary of terms in Appendix 1). It can also be assessed by a change in the shape or size of each arch: this can be measured by intercanine width, intermolar width or arch length. The way the teeth meet together (occlusion) can also be assessed using measurements such as overjet, overbite and assessing the quality of the final results using an index such as the Peer Assessment Rating (PAR index) (Richmond 1992) (see glossary of terms in Appendix 1).

The assessment of stability had to be made at least three months after the fitting of the retainer or after the adjunctive procedure was carried out, or after both.

Secondary outcomes

Failure of the retainers.

This assessed how long retainers lasted without breaking (in months) or how many times they needed to be replaced or repaired during wear. If retainers were lost, we reported this as a failure, since the retainers could not fulfil their role. It is usually recorded how many retainers fail over the observation period.

Adverse effects on the oral health of the:

- teeth (in terms of decay) ‐ assessed using indices of demineralisation or identifying the presence of caries;
- surrounding structures (gums and other supporting structures) ‐ assessed using periodontal indices or other markers of periodontal disease.
Patient satisfaction.

This often takes the form of a questionnaire.

Search methods for identification of studies

We developed detailed search strategies for for each database searched for the identification of studies for this review. These were based on the search strategy developed for MEDLINE but revised appropriately for each database to take account of differences in controlled vocabulary and syntax rules. The MEDLINE search strategy combined the subject search with the Cochrane Highly Sensitive Search Strategy (CHSSS) for identifying randomised trials in MEDLINE: sensitivity maximising version (2009 revision) as referenced in Chapter 6.4.11.1 and detailed in box 6.4.c of The Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011).

The MEDLINE subject search used a combination of controlled vocabulary and free‐text terms and is published in Appendix 2.

Electronic searches

We searched the following databases:

Cochrane Oral Health Group's Trials Register (to 26 January 2016) (see Appendix 3);
Cochrane Central Register of Controlled Trials (CENTRAL) (2015, Issue 12) (see Appendix 4);
MEDLINE via Ovid (1946 to 26 January 2016) (see Appendix 2);
EMBASE via Ovid (1980 to 26 January 2016) (see Appendix 5).

There were no restrictions on language or date of publication in the searches of the electronic databases.

Searching other resources

Ongoing trials

We searched the following databases for ongoing trials (see Appendix 6 for details of the search):

US National Institutes of Health Trials Registry (clinicaltrials.gov) (to 26 January 2016);
The World Health Organization (WHO) Clinical Trials Registry Platform (apps.who.int/trialsearch/default.aspx) (to 26 January 2016).

Handsearching

We handsearched conference proceedings and abstracts from the British Orthodontic Conference, the European Orthodontic Conference and the International Association for Dental Research (IADR) from 2011 to 2015.

Checking reference lists

We checked the bibliographies of papers and review articles that we identified.

Personal communication

We contacted the first named authors of randomised trials identified where we required additional information. We requested further information relevant to the review that was not apparent in the published work. We also asked if they knew of any other published or unpublished studies relevant to the review that were not included in the list. In addition, the authors of this review act as referees for many orthodontic publications. If any study submitted for review appeared to fulfil the criteria, then we contacted the authors for further details.

Data collection and analysis

Selection of studies

Two review authors independently carried out the selection of papers, decision about eligibility, quality assessment, risk of bias, and data extraction. We resolved any disagreements by discussion with one of the other two review authors in the team. If additional information was required, we contacted the author directly and categorised the study as 'awaiting assessment'.

Data extraction and management

We extracted data and entered them on a customised data collection form. We recorded the following:

citation details of publication, if appropriate;
summary of the study design;
participants (sample size, age, inclusion and exclusion criteria, setting of study, costs involved to participants);
interventions:

- type of retainer;
- type of adjunctive procedure;
- prescribed and actual duration of retention;

quality assessment of paper (see Assessment of risk of bias in included studies);
outcomes:

- primary outcome ‐ assessment of stability;
- secondary outcomes ‐ failure of retainers, adverse effects on health, patient satisfaction assessment;

Any additional information that may affect the assessment of the study.

Assessment of risk of bias in included studies

We used the Cochrane 'Risk of bias' tool to assess the risk of bias in the studies (see Section 8.5, Higgins 2011). We assessed the following domains:

random sequence generation;
allocation concealment;
blinding of outcome assessors;
incomplete outcome data reporting;
any selective outcome reporting;
any other sources of bias.

We assessed a study as low risk of bias overall if all the domains were low, unclear if at least one domain was unclear and high if at least one domain was high risk.

Measures of treatment effect

For studies considered eligible for this review, we used the following analyses in line with Cochrane guidance. For dichotomous outcomes, we expressed the estimate of effect of an intervention as risk ratios (RR) together with 95% confidence intervals (CIs). When analysing data for stability, we used the change in irregularity from end of active treatment to final position. If change data were not available, then we used the final irregularity score. For continuous outcomes, we used mean differences (MD) and standard deviations to summarise the data for each group using MD and 95% CIs.

Unit of analysis issues

These were parallel group studies and the statistical unit was the participant.

Dealing with missing data

We contacted trial authors to obtain missing data. We did not exclude a study from the review because of missing summary data; however, we would have discussed the potential implications of its absence from any meta‐analysis if applicable.

Assessment of heterogeneity

We assessed the significance of any discrepancies in the estimates of the treatment effects from the different trials by means of Cochrane's test for heterogeneity and considered heterogeneity significant if P value < 0.10. We used the I² statistic, which describes the percentage total variation across studies that is due to heterogeneity rather than chance, to quantify heterogeneity with I² over 50% being considered moderate to high heterogeneity.

Assessment of reporting biases

Only a proportion of research projects are ultimately published in an indexed journal and become easily identifiable for inclusion in a systematic review. We investigated and attempted to minimise potential reporting biases including publication bias, duplicate publication bias and language bias in this review.

If there had been sufficient numbers of trials (more than 10) in any meta‐analysis, we would have assessed publication bias according to the recommendations on testing for funnel plot asymmetry described in Higgins 2011. If we had identified asymmetry, we would have examined possible causes.

Data synthesis

If there were studies of similar comparisons reporting the same outcome measures, we combined them in a meta‐analysis. We combined RRs for dichotomous data, and would have used MDs for continuous data. We used the fixed‐effect model where there were two or three studies combined and would have used random‐effects models where there were more than three studies in the meta‐analysis.

Subgroup analysis and investigation of heterogeneity

We assessed clinical heterogeneity by examining the types of participants and interventions for all outcomes in each study. We did not formulate any hypotheses to be investigated for subgroup analyses.

Sensitivity analysis

We planned to undertake sensitivity analyses to examine the effect of the risk of bias on the overall estimates of effect, by removing studies that were unclear or high risk of bias from the analysis. In addition, we also planned to examine the effect of including unpublished literature on the review's findings. However, there were too few trials to undertake these analyses.

Presentation of main results

We produced a 'Summary of findings' table for each comparison using GRADEpro software. We assessed the quality of the body of evidence with reference to the overall risk of bias of the included studies, the directness of the evidence, the inconsistency of the results, the precision of the estimates, the risk of publication bias and the magnitude of the effect. We categorised the quality of the body of the evidence for each of the outcomes as high, moderate, low or very low and produced 'Summary of findings' tables for the main outcomes in this review. Where there were studies reporting different types of retainer under different comparison headings, we chose to report the findings of the study with the lowest risk of bias in the 'Summary of findings' table, and then included a narrative summary of the other studies in the comments box.

Results

Description of studies

Results of the search

See table of Characteristics of included studies and Characteristics of excluded studies.

The database search in January 2015 identified 441 articles and four additional articles were identified from additional sources (authors of this review). Of these, 11 were duplicates. Of the remaining 434, screening the titles and abstracts discarded 385. Of the remaining 49 articles, for which the full text was examined, we excluded 29. The remaining 20 articles reported on 15 studies. See Figure 1.

Figure 1

Study flow diagram

On 26 January 2016, we re‐ran the searches and identified seven potentially relevant trials from 53 records. Three of these are ongoing and are described in Characteristics of ongoing studies. Of the four completed and published studies, only one assessed our primary outcome 'stability' but the full text was not available and further information is being sought from the authors. The remaining three studies only assessed secondary outcomes; two of them require translation into English. See the Characteristics of studies awaiting classification section.

Included studies

Types of included studies

We included 15 studies (1722 participants) in this review (Årtun 1997; Aslan 2013; Bolla 2012; Edman Tynelius 2015; Gill 2007; Kumar 2011; Millett 2007; O'Rourke [pers comm]; Rohaya 2006; Rose 2002; Rowland 2007; Salehi 2013; Shawesh 2010; Sun 2011; Thickett 2010). All studies were parallel randomised controlled trials. Thirteen studies were two‐arm trials, one was a three‐arm trial (Edman Tynelius 2015), and one was a four‐arm trial (Årtun 1997).

Data from one study were collected from two articles, publishing different outcomes from the same study (Rowland 2007). Data from Millett 2007 were gathered from three published research abstracts and data from Edman Tynelius 2015 were collected from three articles reporting outcomes at different time intervals. Data and participants from a previous smaller trial were reported in the primary reference for Årtun 1997.

This version of the review did not include three studies that had been included in previous reviews (Edwards 1988; Sauget 1997; Taner 2000). This is because we decided to remove quasi‐randomised trials from the protocol, due to the high risk of bias.

Characteristics of the trial settings and investigators

Nine studies were undertaken in a hospital or university setting (Aslan 2013; Gill 2007; Kumar 2011; Millett 2007; O'Rourke [pers comm]; Rohaya 2006; Rose 2002; Sun 2011; Thickett 2010), five in a specialist practice (Årtun 1997; Bolla 2012; Rowland 2007; Salehi 2013; Shawesh 2010), and one in a National Health Service clinic (Edman Tynelius 2015).

There were six trials from the UK (Gill 2007; O'Rourke [pers comm]; Rohaya 2006; Rowland 2007; Shawesh 2010; Thickett 2010), one from the USA (Årtun 1997), one from Turkey (Aslan 2013), one from Italy (Bolla 2012), one from Sweden (Edman Tynelius 2015), one from India (Kumar 2011), one from the Republic of Ireland (Millett 2007), one from Germany (Rose 2002), one from Iran (Salehi 2013), and one from China (Sun 2011).

Orthodontists provided the care for all the participants in the trials. Four studies stated there was one operator (Bolla 2012; Edman Tynelius 2015; Millett 2007; Rowland 2007), one study stated they had two operators (Årtun 1997), and the remainder did not disclose the number of operators.

Three studies declared external funding sources (Edman Tynelius 2015; Gill 2007; Sun 2011).

Characteristics of the participants

The studies were undertaken on both children and adults. In eight studies, the mean age of participants was under 18 years (Aslan 2013; Edman Tynelius 2015; Gill 2007; O'Rourke [pers comm]; Rowland 2007; Shawesh 2010; Sun 2011; Thickett 2010), in three studies the mean age was over 18 years (Bolla 2012; Rose 2002; Salehi 2013), and four studies did not state the mean age (Årtun 1997; Kumar 2011; Millett 2007; Rohaya 2006).

There were between 20 (Rose 2002) and 397 (Rowland 2007) participants in the 15 studies.

In 10 studies, there were more females than males (Aslan 2013; Bolla 2012; Edman Tynelius 2015; Gill 2007; Millett 2007; O'Rourke [pers comm]; Rowland 2007; Salehi 2013; Shawesh 2010; Thickett 2010), one study had more males than females (Rose 2002), one study had equal numbers of males and females (Sun 2011), and three studies did not report this (Årtun 1997; Kumar 2011; Rohaya 2006).

Characteristics of the interventions

Eight studies compared different types of removable retainers (Aslan 2013; Gill 2007; Kumar 2011; Rohaya 2006; Rowland 2007; Shawesh 2010; Sun 2011; Thickett 2010).

Four studies compared different types of fixed retainers (Årtun 1997; Bolla 2012; Rose 2002; Salehi 2013).

Three studies compared removable retainers with fixed retainers (Årtun 1997; Millett 2007; O'Rourke [pers comm]).

One study compared a combination of removable and fixed retainers, use of an adjunctive procedure and a positioner (Edman Tynelius 2015).

Of the studies comparing removable retainers, the following comparisons were made:

modified thermoplastic retainer versus full coverage thermoplastic retainer (Aslan 2013);
part‐time versus full‐time wear of thermoplastic retainers (Gill 2007; Thickett 2010);
Begg retainers versus thermoplastic retainers, with bonded retainer in lower arch in both groups (Kumar 2011);
Hawley retainers versus thermoplastic retainers (Rohaya 2006; Rowland 2007; Sun 2011);
part‐time versus full‐time wear of Hawley retainers (Shawesh 2010).

Of the studies comparing fixed retainers, the following comparisons were made:

polyethylene ribbon retainer versus multistrand stainless steel;
thick plane stainless steel versus thick spiral stainless steel versus thin flexible stainless steel (Årtun 1997).

Of the studies comparing removable retainers and fixed retainers, the following comparisons were made:

lower multistrand stainless steel versus lower thermoplastic (Millett 2007; O'Rourke [pers comm]);
thick plane stainless steel, thick spiral stainless steel and thin flexible stainless steel versus removable retainer (Årtun 1997).

Characteristics of the outcomes

With the data available, it was possible to analyse the following outcomes:

assessment of stability (Aslan 2013; Gill 2007; Kumar 2011; Millett 2007; O'Rourke [pers comm]; Rohaya 2006; Rowland 2007; Salehi 2013; Shawesh 2010; Thickett 2010);
survival of retainers (Årtun 1997; Bolla 2012; Millett 2007; Rose 2002; Sun 2011);
participant satisfaction (Millett 2007; Rowland 2007);
adverse effects on oral health (Millett 2007).

Assessment of stability was assessed using different measurements:

Little's Irregularity Index (Edman Tynelius 2015; Gill 2007; Millett 2007; O'Rourke [pers comm]; Rowland 2007; Shawesh 2010; Thickett 2010);
Crowding (Shawesh 2010);
Intercanine and intermolar width (Edman Tynelius 2015; Gill 2007; O'Rourke [pers comm]; Rowland 2007; Thickett 2010);
Arch length (Edman Tynelius 2015; O'Rourke [pers comm]; Thickett 2010);
Overjet and overbite (Edman Tynelius 2015; Gill 2007; Thickett 2010);
PAR (Kumar 2011; Thickett 2010);
Settling of the occlusion (Aslan 2013).

This review gave priority to reporting the Little's Irregularity Index (in the 'Summary of findings' tables and conclusions). There is no agreed Core Outcome Measures in Effectiveness Trials (COMET) for stability trials in orthodontics at the present time, so the review authors decided to use Little's Irregularity Index for the preferred outcome measure in this review.

Assessment of survival of retainers were assessed by measuring the number of retainers that failed:

bonded retainer partially or completely detached from the teeth (Årtun 1997; Bolla 2012; Millett 2007; Rose 2002; Salehi 2013; Sun 2011);
bonded retainer fractured (Årtun 1997; Bolla 2012; Millett 2007; Rose 2002; Salehi 2013; Sun 2011);
removable retainer was lost (Årtun 1997; Millett 2007; Sun 2011);
removable retainer no longer fitted (Sun 2011);
removable retainer showed excessive wear (Sun 2011).

Assessment of patient satisfaction was assessed using questionnaires. The following questions were asked:

Was retainer acceptable to wear? Was retainer easy to keep clean? Were you happy with the appearance? (Millett 2007);
Able to wear retainer as instructed? Able to wear retainer away from home? Embarrassed to wear retainer? Discomfort wearing retainer? How did it compare to wearing fixed appliances? (Rowland 2007).

Outcomes were reported at the following intervals:

three months (Kumar 2011);
six months (Aslan 2013; Gill 2007; Kumar 2011; Rowland 2007);
nine months (Aslan 2013);
one year (Edman Tynelius 2015; Millett 2007; Rohaya 2006; Shawesh 2010; Sun 2011; Thickett 2010);
18 months (O'Rourke [pers comm]; Salehi 2013);
two years (Edman Tynelius 2015; Rose 2002);
three years (Årtun 1997);
five years (Edman Tynelius 2015);
six years (Bolla 2012).

Some of the results were provided without standard deviations, or with data that were not amenable to meta‐analysis.

Excluded studies

Of the 29 excluded studies:

16 were not RCTs (Al‐Nimri 2009; Arora 2014; Atack 2007; Axelsson 1993; Barlin 2011; Basciftici 2007; Bauer 2010; Bock 2008; Edwards 1988; Jäderberg 2012; Lindauer 1998; Pandis 2007; Sauget 1997; Tacken 2010; Taner 2000; Vecere 1983);
four investigated methods of placing or supervising retainers rather than comparing different retainers (Bazargani 2012; Conway 2011; ISRCTN56295329; Pandis 2013);
three recruited insufficient participants to complete study (ACTRN12612000670875; ISRCTN22535947; ISRCTN26364810);
three had less than three months' follow‐up (Ahrens 1981; Carvalho 2006; Horton 2009);
three had insufficient data to analyse (Haydar 1992; Larsson 1983; Stormann 2002).

Risk of bias in included studies

See table of Characteristics of included studies and Figure 2 and Figure 3.

Figure 2

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies

Figure 3

Risk of bias summary: review authors' judgements about each risk of bias item for each included study

Allocation

Sequence generation

Twelve studies had adequate sequence generation (Bolla 2012; Edman Tynelius 2015; Gill 2007; Millett 2007; O'Rourke [pers comm]; Rohaya 2006; Rose 2002; Rowland 2007; Salehi 2013; Shawesh 2010; Sun 2011; Thickett 2010). One study used an approach with a high risk of bias sequence (Aslan 2013). For two studies, it is unclear how the sequence generation was performed (Årtun 1997; Kumar 2011).

Allocation concealment

Allocation concealment was adequate for 11 studies (Bolla 2012; Edman Tynelius 2015; Gill 2007; Millett 2007; O'Rourke [pers comm]; Rohaya 2006; Rose 2002; Rowland 2007; Shawesh 2010; Sun 2011; Thickett 2010). There was a high risk of bias in one study (Aslan 2013), and it was unclear in three studies (Årtun 1997; Kumar 2011; Salehi 2013).

Blinding

Blinding of participants and clinicians is usually not possible in retention research, due to the nature of the intervention. This is because it is not possible to hide either the retainer type or adjunctive procedure. Therefore, the blinding of participants and clinicians was not included in the 'Risk of bias' tables for each study. The lack of ability to blind the participants and clinicians could potentially bias this type of research and should be noted when interpreting the findings.

However, blinding of the assessors is often possible, particularly when assessing stability on study models (except where bonded retainers are still in place). Blinding of the assessors is also possible when assessing the results of participant satisfaction questionnaires. Nine studies used blinding of outcome assessors where possible (Edman Tynelius 2015; Gill 2007; Millett 2007; O'Rourke [pers comm]; Rohaya 2006; Rowland 2007; Salehi 2013; Shawesh 2010; Thickett 2010). For three studies, it was not possible to provide blinding (Bolla 2012; Rose 2002; Sun 2011), and in one study, the outcome assessor was not blinded (Aslan 2013). In two of the studies, it was unclear if blinding of assessors was used (Årtun 1997; Kumar 2011).

Incomplete outcome data

Ten studies reported and explained drop‐outs and withdrawals (Edman Tynelius 2015; Gill 2007; Millett 2007; O'Rourke [pers comm]; Rose 2002; Rowland 2007; Salehi 2013; Shawesh 2010; Sun 2011; Thickett 2010). In three studies, the drop‐outs were not fully reported on (Aslan 2013; Bolla 2012; Rohaya 2006), and in two studies, it was unclear (Årtun 1997; Kumar 2011).

Selective reporting

Seven studies showed no evidence of selective reporting (Aslan 2013; Edman Tynelius 2015; Gill 2007; O'Rourke [pers comm]; Salehi 2013; Shawesh 2010; Thickett 2010). Five studies had a high risk of bias due to reporting bias. The Rowland 2007 study measured, but did not report, overjet and overbite changes, and also selectively reported the participant satisfaction data. One study mentioned use of the PAR index as a measure of stability in the methodology, but the outcome was not reported (Millett 2007), and one study described some aspects of participant satisfaction, without giving data to support this (Rose 2002). In another study, there were missing stability and adverse effects data (Årtun 1997). The author of one study reported that there were more outcomes to be reported (Rohaya 2006). In one study, the lower irregularity was reported, but not the upper (Kumar 2011). We contacted the author of one trial and they confirmed they started collecting data for periodontal health, but abandoned this part way through the study due to time constraints (Bolla 2012). For one study, there were outcomes mentioned on a clinical trial register that have not yet been reported (Sun 2011).

Other potential sources of bias

Eleven studies appeared to be free of other bias (Årtun 1997; Aslan 2013; Bolla 2012; Edman Tynelius 2015; Gill 2007; O'Rourke [pers comm]; Rohaya 2006; Rose 2002; Rowland 2007; Salehi 2013; Thickett 2010).

Two studies showed a high risk of other bias. In one study, the authors tried to qualify the quality of their occlusal result by comparing it to a so‐called "ideal" group, but this group may not have been ideal (Aslan 2013). One study identified that despite adequate randomisation, there was not pre‐treatment equivalence between the two intervention groups (Shawesh 2010). As a result, the authors warned that the results should be interpreted with caution.

For two studies, it was not possible to make a decision as to whether there was other risk of bias due to inadequate information in the articles (Kumar 2011; Millett 2007).

Overall risk of bias

Four studies showed overall low risk of bias (Edman Tynelius 2015; Gill 2007; O'Rourke [pers comm]; Thickett 2010).

Ten studies had a high risk of bias (Årtun 1997; Aslan 2013; Bolla 2012; Kumar 2011; Millett 2007; Rohaya 2006; Rose 2002; Rowland 2007; Shawesh 2010; Sun 2011).

One study was at unclear risk of bias (Salehi 2013).

See Figure 2; Figure 3.

Effects of interventions

Comparison 1: removable retainers versus fixed retainers

There were three trials that investigated removable retainers versus fixed retainers (Årtun 1997; Millett 2007; O'Rourke [pers comm]). In all three studies, the comparison was in the lower arch only.

The Årtun 1997 study compared a removable acrylic and wire retainer with types of fixed retainers.
The Millett 2007 study compared a lower thermoplastic (vacuum‐formed) retainer with a multistrand stainless steel wire bonded to the canines and incisors.
The O'Rourke [pers comm] study compared a lower thermoplastic (vacuum‐formed) retainer with a multistrand stainless steel wire bonded to the canines and incisors. The data from this study were non‐parametric and, therefore, reported in medians, so could not be entered into Review Manager 5 (RevMan 2014). Table 1 shows the data and analysis from the original study.

Table 1. O'Rourke study results: changes in stability measurements at 6, 12 and 18 months

Measure

Stability assessments at 6, 12 and 18 months

Bonded retainer

42 participants

(interquartiles)

Vacuum‐formed retainer

40 participants

(interquartiles)

Mann Whitney P value

Little's Irregularity Index

Change between T0 and T1

Change between T1 and T2

Change between T2 and T3

0.03 (0.00‐0.82)

0.04 (0.02‐0.18)

0.05 (0.02‐0.20)

0.08 (0.01‐0.40)

0.10 (0.06‐0.32)

0.07 (0.00‐0.64)

0.003 (P value < 0.05)

0.03 (P value < 0.05)

0.04 (P value < 0.05)

Intercanine width

Change between T0 and T1

Change between T1 and T2

Change between T2 and T3

0.11 (0.05‐0.39)

0.22 (0.11‐0.60)

0.29 (0.12‐0.57)

0.25 (0.10‐0.50)

0.25 (0.09‐0.58)

0.28 ( 0.12‐0.57)

0.56

0.43

0.32

Intermolar width

Change between T0 and T1

Change between T1 and T2

Change between T2 and T3

0.29 (0.11‐0.67)

0.47 (0.15‐1.04)

0.38 (0.12‐0.98)

0.18 (0.66‐0.52)

0.38 (0.14‐0.90)

0.50 (0.22‐1.05)

0.92

0.36

0.24

Arch length

Change between T0 and T1

Change between T1 and T2

Change between T2 and T3

0.24 (0.08‐0.55)

0.39 (0.13‐0.68)

0.49 (0.09‐0.95)

0.31 (0.07‐1.18)

0.26 (0.17‐2.01)

0.22 (0.12‐1.79)

0.17

0.20

0.44

Extraction site space

Change between T0 and T1

Change between T1 and T2

Change between T2 and T3

0.00 (0.00‐0.01)

0.00 (0.00‐0.17)

0.03 (0.00‐0.10)

0.00 (0.00‐0.03)

0.00 (0.00‐0.11)

0.14 (0.00‐0.17)

0.37

0.47

0.01 (P value < 0.05)

Stability

Two trials reported stability using Little's Irregularity Index, both comparing lower removable thermoplastic retainers with lower bonded retainers (Millett 2007; O'Rourke [pers comm]).

Lower removable thermoplastic versus lower bonded retainer

Millett 2007 found a statistically significant difference of 0.6 mm more relapse with a lower removable thermoplastic retainer compared with a lower bonded multistrand retainer (95% CI 0.17 to 1.03; P value = 0.0061). While this was statistically significant, a difference of 0.6 mm would only be clinically significant if restricted to the area of one tooth contact point (Analysis 1.1).

O'Rourke [pers comm] found a statistically significant difference of 0.05 mm more relapse with a lower removable thermoplastic retainer compared with a lower bonded multistrand retainer after six months (P value = 0.003). There was a similar small, but statistically significant, increase in relapse in the removable retainer group at 12 and 18 months. Once again, while these differences were statistically significant, they were not clinically significant. There was no difference reported in intercanine width, intermolar width, arch length and extraction site space opening.

It was not possible to pool the results as the O'Rourke [pers comm] data could not be put into Review Manager 5 (RevMan 2014); they are presented in Table 1.

Failure of retainers

Two studies reported failure of retainers (Årtun 1997; Millett 2007). One study compared three types of bonded retainers and a removable retainer (Årtun 1997), while the other study compared lower thermoplastic retainer with lower bonded retainer.

Comparing three types of lower bonded retainers and removable retainer

In the Årtun 1997 study, there was no difference between the three types of fixed retainers and the removable retainer used (Analysis 1.2).

Lower removable thermoplastic versus lower bonded retainer

In Millett 2007, there was a statistically significant difference of more failures in the removable retainer group than the bonded retainer group (RR 0.66, 95% CI 0.49 to 0.90; P value = 0.007). This was due to 10 retainers being fractured and five being lost in the removable retainer group, compared with two with some composite loss and five completely debonding in the fixed retainer group. This increase of eight more removable retainer failures was clinically significant (Analysis 1.2).

We did not pool the results due to comparison of markedly different types of retainers. This may explain the different findings between the two studies.

Adverse effects on health

One study investigated adverse effects on health, when comparing lower thermoplastic retainers with lower bonded retainers (Millett 2007).

Lower removable thermoplastic versus lower bonded retainer

The Millett 2007 study reported adverse effects on dental health in terms of evidence of caries, gingival bleeding and periodontal pocketing. There was no evidence of caries in either group (Analysis 1.3). However, there was significantly more gingival bleeding in the bonded retainer group than the removable retainer group (RR 0.53, 95% CI 0.31 to 0.88; P value = 0.0015), which is clinically significant (Analysis 1.4). There was more periodontal pocketing in the bonded retainer group (RR 0.32, 95% CI 0.12 to 0.87; P value = 0.026). However, this periodontal pocketing result should be viewed with caution as the two groups were not equal at debond, with four times as many participants with pocketing in the bonded retainer group (Analysis 1.5).

Patient satisfaction

One study investigated patient satisfaction when comparing lower thermoplastic retainers with lower bonded retainers.

Lower removable thermoplastic versus lower bonded retainer

The Millett 2007 study assessed patient satisfaction using the response to three questions.

There was a statistically significant difference in two questions related to patient satisfaction, with the participants finding the fixed retainer more acceptable to wear (MD ‐12.84, 95% CI ‐18.10 to ‐7.09; P value = 0.000012) (Analysis 1.6) and they also perceived it easier to clean (MD ‐10.31, 95% CI ‐20.05 to ‐0.58; P value = 0.038), although the results for ease of cleaning were unlikely to be clinically significant (Analysis 1.7). The participants were equally happy with the appearance of their teeth with both retainers after one year of retention (Analysis 1.8).

Comparison 2: fixed retainers versus fixed retainers

Four trials compared different types of fixed retainers (Årtun 1997; Bolla 2012; Rose 2002; Salehi 2013).

The Årtun 1997 study compared three types of fixed retainers: thick (0.032" (0.08 cm)) plain stainless steel wire bonded only to the canines, thick (0.032") spiral stainless steel wire bonded only to the canines and a thin (0.0205") spiral wire bonded to the canines and incisors.

Three other studies compared polyethylene ribbon bonded retainers with multistrand bonded retainers (Bolla 2012; Rose 2002; Salehi 2013). Two of these studies compared upper and lower retainers (Bolla 2012; Salehi 2013), and one study compared lower retainers only (Rose 2002).

Stability

Stability was not reported.

Failure of retainers

Four studies comparing bonded retainers reported failure of retainers. One studies comparing three types of bonded metal bonded retainers, and three studies comparing polyethylene ribbon bonded retainers with multistrand retainers (Analysis 2.1).

Three types of metal bonded retainers

There was no difference between the failure rates of the three types of bonded retainers in the Årtun 1997 study (Analysis 2.1).

Polyethylene ribbon bonded versus multistrand bonded retainers

Two studies compared failure rates of polyethylene ribbon bonded retainers with multistrand bonded retainers in the maxilla (Bolla 2012; Salehi 2013), and it was possible to undertake a meta‐analysis of these data. The pooled estimate showed an RR of 1.25 (95% CI 0.87 to 1.78; P value = 0.22) indicating no difference in the failure rates. There was low heterogeneity between the studies (P value = 0.26; I² = 22), with very similar methodology and materials used in both trials (Analysis 2.1).

Three studies compared failure rates of polyethylene ribbon bonded retainers with multistrand bonded retainers in the mandible (Bolla 2012; Rose 2002; Salehi 2013), and it was possible to undertake a meta‐analysis of these data. The pooled estimate showed an RR of 1.10 (95% CI 0.77 to 1.57; P value = 0.59) indicating no difference in the failure rates. There was moderate heterogeneity between the studies (P value = 0.15; I² = 47%), with similar methodology and materials used in both trials (Analysis 2.1).

Adverse effects on health

Adverse effects on health were not reported.

Patient satisfaction

Patient satisfaction was not reported.

Comparison 3: removable retainers versus removable retainers

Eight studies compared different types of removable retainers (Aslan 2013; Gill 2007; Kumar 2011; Rohaya 2006; Rowland 2007; Shawesh 2010; Sun 2011; Thickett 2010).

Three studies compared Hawley retainers with thermoplastic retainers, but they were worn for different lengths time per day (Rohaya 2006; Rowland 2007; Sun 2011):

upper Hawley retainers full‐time for three months then six months nights only versus upper and lower thermoplastic retainers worn full‐time for one week then part‐time (nights only) (Rohaya 2006);
upper and lower Hawley retainers worn full‐time for three months then nights only (12 hours per day) versus upper and lower thermoplastic retainers worn full‐time for one week then part‐time (12 hours per day) (Rowland 2007);
upper and lower Hawley retainers versus upper and lower thermoplastic retainers worn full‐time (Sun 2011).

One study compared Begg retainers with thermoplastic retainers (Kumar 2011). The participants were instructed to wear their retainers full‐time for six months then part‐time for six months (12 hours per day).

One study compared modified thermoplastic retainers and full‐coverage thermoplastic retainers (Aslan 2013). Participants were asked to wear their retainers full‐time for six months then nights only for three months.

Two studies compared part‐time wear of thermoplastic retainers with full‐time wear of thermoplastic retainers (Gill 2007; Thickett 2010). Gill 2007 defined part‐time as eight hours per day, and Thickett 2010 defined part‐time as 10 hours per day.

One study compared part‐time wear of Hawley retainers (nights only) with full‐time wear of Hawley retainers (Shawesh 2010).

All eight studies reported stability but they assessed it in a variety of ways:

Little's Irregularity Index for each arch (Gill 2007; Rowland 2007; Shawesh 2010; Thickett 2010);
mean Little's Irregularity Index for both arches (Kumar 2011);
crowding (Shawesh 2010);
intercanine and intermolar width (Gill 2007; Rowland 2007);
overjet and overbite (Gill 2007);
settling assessed by increase in occlusal contacts (Aslan 2013);
stability of corrected rotations (Rohaya 2006);
quality of finish measured by PAR index (Kumar 2011).

Two studies investigated failure of retainers (Rowland 2007; Sun 2011).

One study investigated participant satisfaction (Rowland 2007).

Although there were eight studies comparing removable retainers with other removable retainers, only two studies compared similar interventions and similar outcomes (Gill 2007; Thickett 2010). However, it was not possible to pool the results in a meta‐analysis, as the data from the Thickett 2010 study were presented as medians and interquartile ranges, so could not be put into Review Manager 5 (RevMan 2014), but are presented in Table 2.

summary of findings Table for the main comparison.

Table 2. Thickett table of results: medians for full‐time (group 1) and part‐time (group 2) at debond and after one year

Outcome	Full‐time at debond	Part‐time at debond	P value at debond	Full‐time after 1 year	Part‐time after 1 year	P value after 1 year
LII	0.04	0.14	0.46	0.71	0.89	0.5
UII	0.21	0.14	0.14	1.08	1.09	0.8
Lower intercanine width	27.24	27.07	0.22	27.07	26.47	0.65
Lower intermolar width	34.09	33.32	0.52	34.35	33.6	0.61
Lower arch length	21.1	21.02	0.44	21.28	20.14	0.06
Upper intercanine width	35.49	34.79	0.08	34.93	34.56	0.52
Upper intermolar width	40.23	39.57	0.35	40.34	39.39	0.68
Upper arch length	24.98	24.19	0.22	25.07	25.15	0.97
Overjet	2.54	2.36	0.6	2.76	2.39	0.37
Overbite	2.86	3.31	0.14	3.14	3.74	0.05 (P value < 0.05)

LII: lower Little's Irregularity Index in mm; UII: upper Little's Irregularity Index in mm.

Stability

Seven studies investigated stability when comparing removable retainers.

Modified versus full coverage thermoplastic retainers

Aslan 2013 assessed stability by reporting "settling". This describes the favourable changes in occlusion that happens after treatment is completed, and may be affected by the nature and duration of wear of retainers (see glossary of terms in Appendix 1). There was no difference in the number of total occlusal contacts after six months of full‐time wear. After three further months of nights‐only wear, there was a statistically significant increase in posterior occlusal contacts in the modified thermoplastic group compared with the full coverage thermoplastic retainers (MD 0.95, 95% CI 0.05 to 1.85; P value = 0.038) (Analysis 3.11). However, it is doubtful that an increase in total occlusal contacts from 23 to 35 is clinically significant. There was a statistically significant increase in total anterior occlusal contacts in the full coverage thermoplastic retainers; however, once again the difference of one extra occlusal contact would not be regarded as clinically significant (Analysis 3.12).

Full‐time versus part‐time wear of thermoplastic retainers

Gill 2007 and Thickett 2010 both compared full‐time wear of thermoplastic retainers versus part‐time wear of thermoplastic retainers. The results from the Gill 2007 study showed no statistical difference in terms of Little's Irregularity Index, intercanine width and intermolar width in both upper and lower arches, and no difference in overjet or overbite (Analysis 3.5; Analysis 3.6; Analysis 3.7; Analysis 3.8; Analysis 3.9; Analysis 3.10). Table 2 presents the results from Thickett 2010. They found no statistical difference in Little's Irregularity Index, arch length, intercanine width and intermolar width in both upper and lower arches, and no difference in overjet. They did find a statistical difference in overbite, but as the difference was 0.6 mm it is unlikely to be clinically significant. While the results could not be combined in a meta‐analysis, the finding of both studies was that there was no evidence of a difference in stability when thermoplastic retainers were worn full‐time or part‐time.

Full‐time versus part‐time wear of Hawley retainers

Shawesh 2010 compared full‐time wear of Hawley retainers with part‐time wear of Hawley retainers. They showed no statistical difference in Little's Irregularity Index or crowding in both arches between the two wear regimens (Analysis 3.1; Analysis 3.2; Analysis 3.3; Analysis 3.4).

Hawley retainers versus thermoplastic retainers

Rowland 2007 compared Hawley retainers with thermoplastic retainers. They found no statistically significant difference in intercanine and intermolar widths in both arches (Analysis 3.5; Analysis 3.6; Analysis 3.7; Analysis 3.8). However, they did find a statistically significant difference in Little's Irregularity Index in the upper arch, with 0.25 mm (95% CI 0.08 to 0.42; P value = 0.004) more relapse in Hawley retainer group (Analysis 3.1), and in the lower arch of 0.42 mm (95% CI 0.23 to 0.61; P value < 0.0001) more relapse in the thermoplastic retainer group (Analysis 3.2). These differences were very small. Certainly in the upper arch this difference would not be clinically significant. The difference in the lower arch may be clinically significant if the irregularity was restricted to one tooth contact.

The Rohaya 2006 study investigated the ability to maintain previously rotated teeth in a stable position, comparing Hawley retainers and thermoplastic retainers. They found that people wearing thermoplastic retainers were 4.88 times more likely to hold derotated teeth stable than people who wore Hawley retainers (95% CI 1.13 to 21.07; P value = 0.03) (Analysis 3.13). This is likely to be clinically significant.

Begg retainers versus thermoplastic retainers

Kumar 2011 compared Begg retainers with thermoplastic retainers (both groups had a bonded retainer in addition in the lower arch). There was a statistically significant difference of 0.25 mm (95% CI 0.19 to 0.31; P value < 0.000001) more irregularity in the lower arch with the Begg retainers. This is unlikely to be clinically significant. There was also a statistically significant reduction in the quality of the result of 1.71 points (95% CI 1.44 to 1.98; P value < 0.0001) on the PAR index in the Begg retainer group, indicating more relapse in this group (Analysis 3.14). Although small, this may be clinically significant.

Failure of retainers

Hawley retainers versus thermoplastic retainers

Rowland 2007 compared failure rates of Hawley retainers with thermoplastic retainers. They assessed failure rates by recording how many retainers broke and how many were lost. Hawley retainers were more likely to break (32 Hawley retainers broke compared with 12 thermoplastic retainers) giving an RR of 2.96 (95% CI 1.58 to 5.55; P value = 0.000072), which is clinically significant (Analysis 3.15). There was no difference in the number of retainers that were lost (Analysis 3.16).

Sun 2011 also compared the failure rate of Hawley retainers and thermoplastic retainers. A retainer was assessed as failed if it fractured, no longer fit, showed local serious abrasion causing holes in the retainer, or loss of retainers. Failure rate was assessed separately for upper and lower retainers. Sun 2011 showed no difference in failure rates in the upper arch (Analysis 3.17). However, in the lower arch there was a statistically significant chance of thermoplastic retainers failing compared with Hawley retainers (25 Hawley compared with 41 thermoplastic). The RR was 0.60 (95% CI 0.43 to 0.83; P value = 0.0023), which is clinically significant (Analysis 3.18).

Since failure was measured in different ways, and one study measured failure rates in total while the other measured them per arch, we could not pool the results of Rowland 2007 and Sun 2011.

Adverse effects on health

Adverse effects on health were not reported.

Patient satisfaction

Hawley retainers versus thermoplastic retainers

One study assessed patient satisfaction by measuring response to five questions (Rowland 2007). There was a statistically significant response in three of the questions: the participants in the thermoplastic group were more likely to report being able to wear the retainer than participants in the Hawley group, with an RR of 0.89 (95% CI 0.83 to 0.96; P value = 0.0023) (Analysis 3.19); the participants wearing the thermoplastic retainers were less likely to report feeling embarrassed by the retainers, with an RR of 2.42 (95% CI 1.30 to 4.49; P value = 0.0052) (Analysis 3.21); more participants reported that it was more difficult wear Hawley retainers than thermoplastic retainers, when compared with fixed appliances, with an RR of 9.37 (95% CI 3.80 to 23.10; P value < 0.00001) (Analysis 3.23). All of these findings were clinically significant. There was no difference between the two retainers in the amount of discomfort experienced (Analysis 3.22), or whether people were able to wear the retainers away from home (Analysis 3.20).

Comparison 4: upper removable and lower fixed retainer versus upper removable retainer and lower adjunctive procedure

Upper removable and lower fixed retainer versus upper removable retainer and lower adjunctive procedure

The Edman Tynelius 2015 study compared an upper thermoplastic retainer and lower rigid stainless steel retainer from lower canine to canine with an upper thermoplastic retainer and the adjunctive procedure of interproximal reduction in the lower arch, with no lower retainer. They measured stability at one, two and five years. Change data were reported for Little's Irregularity Index in both arches, upper and lower intercanine and intermolar width, arch length, overjet and overbite at one and two years. The final Little's Irregularity Index (change data were not available) was reported for both arches at five years.

After one year, there was a statistically significant reduction of 1 mm in the lower intercanine width in the interproximal stripping without lower retainer group (95% CI ‐1.22 to ‐0.78; P value < 0.00001), which may be clinically significant (Analysis 4.6). There was also a statistically significant increase of 0.7 mm in the lower arch length in the lower fixed retainer group (95% CI 0.2 to 1.2; P value = 0.0063), which we would not perceive as clinically significant (Analysis 4.8). There was no difference in the Little's Irregularity Index in either arch, intercanine width in the upper arch, intermolar width in either arch, arch length in the upper arch, overjet or overbite (Analysis 4.1; Analysis 4.2; Analysis 4.3; Analysis 4.4; Analysis 4.5; Analysis 4.7; Analysis 4.9; Analysis 4.10).

After two years, there was a statistically significant reduction of 0.8 mm in the lower intercanine width in the interproximal stripping without lower retainer group (95% CI ‐1.26 to ‐0.34; P value = 0.0057), which is unlikely to be clinically significant (Analysis 4.6). There was no difference in the Little's Irregularity Index in either arch, intercanine width in the upper arch, intermolar width in either arch, arch length in either arch, overjet or overbite (Analysis 4.1; Analysis 4.2; Analysis 4.3; Analysis 4.4; Analysis 4.5; Analysis 4.7; Analysis 4.8; Analysis 4.9; Analysis 4.10).

After five years, there was no difference in the final Little's Irregularity Index in either the upper arch or the lower arch (Analysis 4.11; Analysis 4.12).

Comparison 5: upper removable and lower fixed retainer versus positioner

Upper removable and lower fixed retainer versus positioner

The Edman Tynelius 2015 study compared an upper thermoplastic retainer and lower rigid stainless steel retainer from lower canine to canine with a positioner. They measured stability at one, two and five years. Change data were reported for Little's Irregularity Index in both arches, upper and lower intercanine and intermolar width, arch length, overjet and overbite at one and two years. The final Little's Irregularity Index (change data were not available) was reported for both arches at five years.

After one year, there was a statistically significant reduction of 0.7 mm in the lower intercanine width in the interproximal stripping without lower retainer group (95% CI ‐1.14 to ‐0.26; P value = 0.0017), but we would not perceive this to be clinically significant (Analysis 5.6). There was no difference in the Little's Irregularity Index in either arch, intercanine width in the upper arch, intermolar width in either arch, arch length in either arch, overjet or overbite (Analysis 5.1; Analysis 5.2; Analysis 5.3; Analysis 5.4; Analysis 5.5; Analysis 5.7; Analysis 5.8; Analysis 5.9; Analysis 5.10).

After two years, there was a statistically significant difference in the upper intercanine width, lower Little's Irregularity Index, lower intercanine width and lower intermolar width. The upper intercanine width reduced more in the positioner group by 0.8 mm (95% CI ‐1.50 to ‐0.10; P value = 0.026), but this is unlikely to be clinically significant (Analysis 5.2). The Little's Irregularity Index in the lower arch increased by 1 mm in the positioner group (95% CI ‐1.65 to ‐0.35; P value = 0.0025), and this is clinically significant (Analysis 5.5). The lower intercanine width decreased by 0.8 mm in the positioner group (95% ‐1.34 to ‐0.26; P value = 0.0037), which we would not regard as clinically significant (Analysis 5.6). The intermolar width decreased by 1 mm in the positioner group (95% CI ‐1.93 to ‐0.07; P value = 0.035), which may be clinically significant (Analysis 5.7). There was no difference in the Little's Irregularity Index in the upper arch, upper intermolar width, upper arch length, lower arch length, overjet and overbite (Analysis 5.1; Analysis 5.3; Analysis 5.4; Analysis 5.7; Analysis 5.8; Analysis 5.9; Analysis 5.10).

After five years, there was significantly more relapse of 1.30 mm change in Little's Irregularity Index in the lower arch with the positioner (95% CI ‐2.42 to ‐0.18; P value = 0.023), which is clinically significant (Analysis 5.12). There was no difference in the upper arch Little's Irregularity Index (Analysis 5.11).

Comparison 6: upper removable retainer and lower adjunctive procedure versus positioner

Upper removable retainer and lower adjunctive procedure versus positioner

The Edman Tynelius 2015 study compared an upper thermoplastic retainer and lower adjunctive procedure of interproximal stripping with a positioner. They measured stability at one, two and five years. Change data were reported for Little's Irregularity Index in both arches, upper and lower intercanine and intermolar width, arch length, overjet and overbite at one and two years. The final Little's Irregularity Index (change data were not available) was reported for both arches at five years.

After one year, there was a statistically significant increase of 0.5 mm in the lower arch length with the positioner (95% CI ‐0.95 to ‐0.05; P value = 0.031), which was not clinically significant (Analysis 6.8). There was no difference in the Little's Irregularity Index in both arches, intercanine or intermolar widths in both arches, upper arch length, overjet and overbite (Analysis 6.1; Analysis 6.2; Analysis 6.3; Analysis 6.4; Analysis 6.5; Analysis 6.6; Analysis 6.7; Analysis 6.9; Analysis 6.10).

After two years, there was a statistically significant difference reduction of 0.90 mm in upper intercanine width with the positioner (95% CI ‐1.63 to ‐0.17; P value = 0.0015), which is not clinically significant (Analysis 6.2). There was also an increase of 0.70 mm in the lower Little's Irregularity Index (95% CI ‐1.37 to ‐0.03; P value = 0.0041), which was also not clinically significant (Analysis 6.5). There was no difference in Little's Irregularity Index in the upper arch, intercanine width in the lower arch, intermolar width in both arches, arch length in both arches, overjet and overbite (Analysis 6.1; Analysis 6.3; Analysis 6.4; Analysis 6.6; Analysis 6.7; Analysis 6.8; Analysis 6.9; Analysis 6.10).

After five years, there was a statistically significant 1.40 mm more relapse in the positioner group in the final lower arch Little's Irregularity Index (95% CI ‐2.77 to ‐0.03; P value = 0.045), which was clinically significant (Analysis 6.12). There was no difference in the final upper Little's Irregularity Index (Analysis 6.11).

Discussion

Summary of main results

Removable retainers versus fixed retainers

Lower multistrand retainer versus lower thermoplastic retainer

We found limited evidence to suggest that multistrand bonded retainers were slightly better at reducing relapse in the lower arch than thermoplastic retainers (Millett 2007; O'Rourke [pers comm]). The bonded retainers were better at reducing irregularity, but there was no difference in maintaining intercanine width, intermolar width and arch length of extraction space opening. There was conflicting evidence about failure rates of these types of retainers (Millett 2007; O'Rourke [pers comm]). There was limited evidence to suggest that bonded retainers lead to increased gingival bleeding compared with thermoplastic retainers, but neither retainer appeared to predispose to caries (Millett 2007). There was some evidence to show that patient satisfaction was higher with bonded retainers, with participants reporting that they were more acceptable to wear than thermoplastic retainers (Millett 2007).

Fixed retainers versus fixed retainers

Polyethylene ribbon fixed retainer versus multistrand fixed retainer

We found weak evidence to suggest that there is no difference in failure rates of polyethylene ribbon fixed retainers and multistrand fixed retainers (Bolla 2012; Rose 2002; Salehi 2013). We were able to pool the data for these studies as the materials and methodology were similar.

Removable retainers versus removable retainers

Part‐time wear versus full‐time wear of thermoplastic retainers

Two studies failed to find a difference in stability between part‐time and full‐time wear (Gill 2007; Thickett 2010). The data from the studies could not be pooled, as the data from Thickett 2010 could not be entered into Review Manager 5 (RevMan 2014); however, both studies reached the same conclusions and we assessed this evidence as of moderate quality.

Hawley retainers versus thermoplastic retainers

There was limited evidence to suggest that thermoplastic retainers offer better retention in the lower arch than Hawley retainers. Rowland 2007 showed a small increase in irregularity in the lower arch with Hawley retainers, although this was unlikely to be significant unless the irregularity was restricted to one tooth contact. One study showed that thermoplastic retainers prevented relapse of previously rotated teeth in the upper arch better than Hawley retainers (Rohaya 2006).

We found limited evidence from two studies looking at failure rates of these two types of retainers (Rowland 2007; Sun 2011). The studies could not be combined due to the marked difference in materials and methodology, in particular that in one study the retainers were worn for long periods on a part‐time basis (Rowland 2007), and in the other study they were worn full‐time except at meal times (Sun 2011). When Hawley retainers were worn three months full‐time, then nine months part‐time, and the vacuum‐formed retainers were worn part‐time, there were more breakages in the Hawley retainers (Rowland 2007). When both sets of retainers were worn full‐time except at meal times, there were more failures in the thermoplastic retainer group (Sun 2011).

There was limited evidence of a greater level of patient satisfaction with thermoplastic retainers than Hawley retainers (Rowland 2007). Participants wearing thermoplastic retainers reported less embarrassment at wearing them than reported by participants wearing Hawley retainers. The participants who wore thermoplastic retainers also found them easier to wear than participants with Hawley retainers, when comparing the retainers to their experience with fixed appliances.

Begg retainers versus thermoplastic retainers

There was limited evidence that thermoplastic retainers may retain the quality of the result better than Begg retainers (Kumar 2011).

Part‐time wear versus full‐time wear of Hawley retainers

There was limited evidence that part‐time wear of Hawley retainers was as effective for preventing relapse of irregular and crowded teeth as full‐time wear (Shawesh 2010).

Full‐coverage versus modified thermoplastic retainers

There was limited evidence that there was no difference in the amount of settling between full‐coverage retainers and modified thermoplastic retainers after six months of full‐time wear (Aslan 2013). However, after moving to a period of part‐time wear, there was more settling (increased occlusal contacts) with the modified thermoplastic retainers.

Upper removable and fixed retainer versus upper removable and lower adjunctive procedure versus positioner

Based on the results of one study, all three approaches maintained the teeth in a reasonably stable position for the first year (Edman Tynelius 2015). However, after two years, there was greater irregularity in the lower arch with the positioner, which could be clinically significant. The fact that adequate stability was maintained in the group with adjunctive procedure but no lower retainer is interesting, because in recent decades it has been presumed that all participants had to wear some sort of retainer to maintain stability. It is important to emphasise that the participants initially presented with normal jaw relationships in all three dimensions (anteroposterior, vertical and transverse proportions), and so were essentially Class I crowding cases. Typically, teeth do not need to be moved as far in these cases. When trying to straighten teeth and obtain a good bite when the jaws do not meet together well, teeth need to be moved further to compensate for this. As a result, it may not be possible to extrapolate these results to most people requiring orthodontic treatment.

Overall completeness and applicability of evidence

Overall, we included 15 studies investigating different approaches to retention and with multiple outcomes in this review. Three studies compared different removable retainers with fixed retainers. Four studies compared different types of fixed retainers. Eight studies compared different types of removable retainers.

One further study, compared a combination of upper thermoplastic and lower bonded versus upper thermoplastic with lower adjunctive procedures versus positioner (Edman Tynelius 2015). It is important to note that there were very strict inclusion criteria for this study and, while the study was well conducted with a low risk of bias, it is important to recognise that the findings may only be applicable to people with similar malocclusions (crowded cases with normal positions of the jaws in all three dimensions).

Only three of the 15 studies reported outcomes beyond two years. This is important as it is felt that retainers may be needed indefinitely to reduce the chances of relapse, so the findings of this review predominantly will inform clinicians of the effects of different approaches to retention in the short term.

When interpreting the results of this review, it is important to remember that the results of the studies may be affected by the age of the participants, initial malocclusions, treatment procedure, amount and type of tooth movement, along with other possible factors that may affect relapse. The impact of these areas is still not fully understood, and is beyond the scope of this review. However, the broader context should always be considered when interpreting studies reporting the stabilising effects of retainers and adjunctive techniques.

There is no agreed Core Outcome Measure in Effectiveness Trials (COMET) for relapse studies in orthodontics at the present time. Therefore, the review authors have prioritised reporting Little's Irregularity Index as the preferred outcome for this review.

We identified four ongoing studies and four studies that are currently awaiting classification and may be included in the next update of this review (see Characteristics of studies awaiting classification and Characteristics of ongoing studies).

Quality of the evidence

The overall quality of evidence is reported in the 'Summary of findings' tables (summary of findings Table for the main comparison; summary of findings Table 2; summary of findings Table 3).

Removable retainer versus fixed retainer

There was low quality evidence to assess stability, in the lower arch, adverse effects on health and patient satisfaction. We downgraded the evidence due to high risk of bias and imprecision. There was very low evidence for failure of retainers, with the quality of evidence downgraded three times due to high risk of bias, imprecision and inconsistency. This low quality, or very low quality, of evidence means that future research is very likely to have an impact on our confidence in the estimate of effect and likely to change the estimate.

Fixed retainer versus fixed retainer

summary of findings Table 2.

There was low quality evidence on failure of retainers, downgraded twice due to high risk of bias and imprecision. This low quality evidence means that future research is very likely to have an impact on our confidence in the estimate of effect and likely to change the estimate.

Removable retainer versus removable retainer

summary of findings Table 3.

There was moderate quality evidence comparing full‐time wear with part‐time wear of thermoplastic retainers. The results were based on only one study and we downgraded this evidence once due to imprecision. Moderate quality evidence means that further research in this area is likely to have an important impact on our confidence in the estimate of effect, and may change the estimate.

There was low quality evidence in further studies comparing different types of removable retainers: part‐time versus full‐time wear of Hawley retainers; Hawley versus thermoplastic retainers and Begg versus thermoplastic retainers. We downgraded all these low quality studies twice due to high risk of bias and imprecision. Low quality evidence means that future research in these areas is very likely to have an impact on our confidence in the estimate of effect and likely to change the estimate.

Potential biases in the review process

We have used a broad sensitive search strategy of multiple databases, and also searched for unpublished studies and data, without language restrictions. Therefore, we hope that potential bias in the review process has been minimised. We could not analyse two of the studies in Review Manager 5 (RevMan 2014), as the data were reported as medians (O'Rourke [pers comm]; Thickett 2010). Therefore, we can only report the findings of the authors and so we have not included their findings in our final conclusions.

Agreements and disagreements with other studies or reviews

Mai 2014 is a systematic review that investigated Hawley retainers compared to thermoplastic retainers. They identified seven studies to include, but this was because, in addition to randomised controlled clinical trials, they included quasi‐randomised and controlled clinical trials. They reported that there was some evidence to suggest "that there are no differences with respect to changes in intercanine and intermolar widths between Hawley retainers and vacuum‐formed retainers after orthodontic retention". This finding is different to our conclusions, but this was due to different inclusion criteria to our review.

Figure 1

Study flow diagram

Figure 2

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies

Figure 3

Risk of bias summary: review authors' judgements about each risk of bias item for each included study

Analysis 1.1

Comparison 1 Removable versus fixed retainers, Outcome 1 Stability ‐ Little's Irregularity Index.

Analysis 1.2

Comparison 1 Removable versus fixed retainers, Outcome 2 Failure of retainers in lower.

Analysis 1.3

Comparison 1 Removable versus fixed retainers, Outcome 3 Adverse effects on health: evidence of caries.

Analysis 1.4

Comparison 1 Removable versus fixed retainers, Outcome 4 Adverse effects on health: evidence of gingival bleeding.

Analysis 1.5

Comparison 1 Removable versus fixed retainers, Outcome 5 Adverse effects on health: evidence of periodontal pocketing.

Analysis 1.6

Comparison 1 Removable versus fixed retainers, Outcome 6 Patient satisfaction: how acceptable was the retainer to wear?.

Analysis 1.7

Comparison 1 Removable versus fixed retainers, Outcome 7 Patient satisfaction: how easy was retainer to keep clean?.

Analysis 1.8

Comparison 1 Removable versus fixed retainers, Outcome 8 Patient satisfaction: how happy are you with appearance of teeth after 1 year of retention.

Analysis 2.1

Comparison 2 Fixed versus fixed retainers, Outcome 1 Failure of retainers.

Analysis 3.1

Comparison 3 Removable versus removable retainers, Outcome 1 Stability ‐ Little's Irregularity Index ‐ upper labial segment ‐ 1 year.

Analysis 3.2

Comparison 3 Removable versus removable retainers, Outcome 2 Stability ‐ Little's Irregularity Index ‐ lower labial segment ‐ 1 year.

Analysis 3.3

Comparison 3 Removable versus removable retainers, Outcome 3 Stability ‐ crowding upper labial segment ‐ 1 year.

Analysis 3.4

Comparison 3 Removable versus removable retainers, Outcome 4 Stability ‐ crowding lower labial segment ‐ 1 year.

Analysis 3.5

Comparison 3 Removable versus removable retainers, Outcome 5 Stability: lower intercanine width.

Analysis 3.6

Comparison 3 Removable versus removable retainers, Outcome 6 Stability: lower intermolar width.

Analysis 3.7

Comparison 3 Removable versus removable retainers, Outcome 7 Stability: upper intercanine width.

Analysis 3.8

Comparison 3 Removable versus removable retainers, Outcome 8 Stability: upper intermolar width.

Analysis 3.9

Comparison 3 Removable versus removable retainers, Outcome 9 Stability: overjet.

Analysis 3.10

Comparison 3 Removable versus removable retainers, Outcome 10 Stability: overbite.

Analysis 3.11

Comparison 3 Removable versus removable retainers, Outcome 11 Stability: settling with increased posterior contacts.

Analysis 3.12

Comparison 3 Removable versus removable retainers, Outcome 12 Stability: settling with increased anterior contacts.

Analysis 3.13

Comparison 3 Removable versus removable retainers, Outcome 13 Stability: maintaining corrected rotations in the upper.

Analysis 3.14

Comparison 3 Removable versus removable retainers, Outcome 14 Stability: maintaining quality of finish (PAR).

Analysis 3.15

Comparison 3 Removable versus removable retainers, Outcome 15 Survival of retainers: how many broke in total.

Analysis 3.16

Comparison 3 Removable versus removable retainers, Outcome 16 Survival of retainers: how many were lost in total.

Analysis 3.17

Comparison 3 Removable versus removable retainers, Outcome 17 Survival of retainers: upper.

Analysis 3.18

Comparison 3 Removable versus removable retainers, Outcome 18 Survival of retainers: lower.

Analysis 3.19

Comparison 3 Removable versus removable retainers, Outcome 19 Patient satisfaction: able to wear retainer as instructed?.

Analysis 3.20

Comparison 3 Removable versus removable retainers, Outcome 20 Patient satisfaction: able to wear retainers away from home?.

Analysis 3.21

Comparison 3 Removable versus removable retainers, Outcome 21 Patient satisfaction: embarrassed to wear retainer?.

Analysis 3.22

Comparison 3 Removable versus removable retainers, Outcome 22 Patient satisfaction: amount of discomfort ‐ never or only on occasion.

Analysis 3.23

Comparison 3 Removable versus removable retainers, Outcome 23 Patient satisfaction: worse or much worse than wearing fixed appliances?.

Analysis 4.1

Comparison 4 Upper removable and lower fixed retainer versus upper removable and lower adjunctive procedure, Outcome 1 Irregularity change of upper labial segment (V‐CTC vs. V‐S).

Analysis 4.2

Comparison 4 Upper removable and lower fixed retainer versus upper removable and lower adjunctive procedure, Outcome 2 Intercanine change in upper (V‐CTC vs. V‐S).

Analysis 4.3

Comparison 4 Upper removable and lower fixed retainer versus upper removable and lower adjunctive procedure, Outcome 3 Intermolar width change in upper (V‐CTC vs. V‐S).

Analysis 4.4

Comparison 4 Upper removable and lower fixed retainer versus upper removable and lower adjunctive procedure, Outcome 4 Arch length change in upper (V‐CTC vs. V‐S).

Analysis 4.5

Comparison 4 Upper removable and lower fixed retainer versus upper removable and lower adjunctive procedure, Outcome 5 Irregularity change in lower labial segment (V‐CTC vs. V‐S).

Analysis 4.6

Comparison 4 Upper removable and lower fixed retainer versus upper removable and lower adjunctive procedure, Outcome 6 Intercanine width change in lower (V‐CTC vs. V‐S).

Analysis 4.7

Comparison 4 Upper removable and lower fixed retainer versus upper removable and lower adjunctive procedure, Outcome 7 Intermolar width change in the lower (V‐CTC vs. V‐S).

Analysis 4.8

Comparison 4 Upper removable and lower fixed retainer versus upper removable and lower adjunctive procedure, Outcome 8 Arch length change in lower (V‐CTC vs. V‐S).

Analysis 4.9

Comparison 4 Upper removable and lower fixed retainer versus upper removable and lower adjunctive procedure, Outcome 9 Overjet change (V‐CTC vs. V‐S).

Analysis 4.10

Comparison 4 Upper removable and lower fixed retainer versus upper removable and lower adjunctive procedure, Outcome 10 Overbite change (V‐CTC vs. V‐S).

Analysis 4.11

Comparison 4 Upper removable and lower fixed retainer versus upper removable and lower adjunctive procedure, Outcome 11 Final irregularity in upper after more than 5 years (V‐CTC vs. V‐S).

Analysis 4.12

Comparison 4 Upper removable and lower fixed retainer versus upper removable and lower adjunctive procedure, Outcome 12 Final irregularity in lower after more than 5 years (V‐CTC vs. V‐S).

Analysis 5.1

Comparison 5 Upper removable retainer and lower fixed retainer versus positioner, Outcome 1 Irregularity change in upper labial segment (V‐CTC vs. P).

Analysis 5.2

Comparison 5 Upper removable retainer and lower fixed retainer versus positioner, Outcome 2 Intercanine width change in upper (V‐CTC vs. P).

Analysis 5.3

Comparison 5 Upper removable retainer and lower fixed retainer versus positioner, Outcome 3 Intermolar width change in upper (V‐CTC vs. P).

Analysis 5.4

Comparison 5 Upper removable retainer and lower fixed retainer versus positioner, Outcome 4 Arch length change in upper (V‐CTC vs. P).

Analysis 5.5

Comparison 5 Upper removable retainer and lower fixed retainer versus positioner, Outcome 5 Irregularity change in lower labial segment (V‐CTC vs. P).

Analysis 5.6

Comparison 5 Upper removable retainer and lower fixed retainer versus positioner, Outcome 6 Intercanine width change in lower (V‐CTC vs. P).

Analysis 5.7

Comparison 5 Upper removable retainer and lower fixed retainer versus positioner, Outcome 7 Intermolar width change in lower (V‐CTC vs. P).

Analysis 5.8

Comparison 5 Upper removable retainer and lower fixed retainer versus positioner, Outcome 8 Arch length change in lower (V‐CTC vs. P).

Analysis 5.9

Comparison 5 Upper removable retainer and lower fixed retainer versus positioner, Outcome 9 Overjet change (V‐CTC vs. P).

Analysis 5.10

Comparison 5 Upper removable retainer and lower fixed retainer versus positioner, Outcome 10 Overbite change (V‐CTC vs. P).

Analysis 5.11

Comparison 5 Upper removable retainer and lower fixed retainer versus positioner, Outcome 11 Final irregularity in upper after more than 5 years (V‐CTC vs. P).

Analysis 5.12

Comparison 5 Upper removable retainer and lower fixed retainer versus positioner, Outcome 12 Final irregularity in lower after more than 5 years (V‐CTC vs. P).

Analysis 6.1

Comparison 6 Upper removable retainer and lower adjunctive procedure versus positioner (P), Outcome 1 Irregularity change in upper labial segment (V‐S vs. P).

Analysis 6.2

Comparison 6 Upper removable retainer and lower adjunctive procedure versus positioner (P), Outcome 2 Intercanine width change in upper (V‐S vs. P).

Analysis 6.3

Comparison 6 Upper removable retainer and lower adjunctive procedure versus positioner (P), Outcome 3 Intermolar width change in upper (V‐S vs. P).

Analysis 6.4

Comparison 6 Upper removable retainer and lower adjunctive procedure versus positioner (P), Outcome 4 Arch length change in upper (V‐S vs. P).

Analysis 6.5

Comparison 6 Upper removable retainer and lower adjunctive procedure versus positioner (P), Outcome 5 Irregularity change in lower labial segment (V‐S vs. P).

Analysis 6.6

Comparison 6 Upper removable retainer and lower adjunctive procedure versus positioner (P), Outcome 6 Intercanine width change in lower (V‐S vs. P).

Analysis 6.7

Comparison 6 Upper removable retainer and lower adjunctive procedure versus positioner (P), Outcome 7 Intermolar width change in lower (V‐S vs. P).

Analysis 6.8

Comparison 6 Upper removable retainer and lower adjunctive procedure versus positioner (P), Outcome 8 Arch length change in lower (V‐S vs. P).

Analysis 6.9

Comparison 6 Upper removable retainer and lower adjunctive procedure versus positioner (P), Outcome 9 Overjet change (V‐S vs. P).

Analysis 6.10

Comparison 6 Upper removable retainer and lower adjunctive procedure versus positioner (P), Outcome 10 Overbite change (V‐S vs. P).

Analysis 6.11

Comparison 6 Upper removable retainer and lower adjunctive procedure versus positioner (P), Outcome 11 Final irregularity in upper after more than 5 years (V‐S vs. P).

Analysis 6.12

Comparison 6 Upper removable retainer and lower adjunctive procedure versus positioner (P), Outcome 12 Final irregularity in lower after more than 5 years (V‐S vs. P).

Summary of findings for the main comparison. Removable retainer versus fixed retainer to stabilise tooth position

Removable retainer versus fixed retainer to stabilise tooth position
Patient or population: people who had received fixed appliance treatment Setting: specialist orthodontic practice and hospital orthodontic department Intervention: removable retainers Comparison: fixed retainers
Outcomes	*Anticipated absolute effects (95% CI)**		Relative effect (95% CI)	No of participants (studies)	Quality of the evidence (GRADE)	Comments
	Risk with fixed retainers	Risk with removable retainers
Stability ‐ Little's Irregularity Index in lower arch Ideally 0 mm 12 months	Mean = 0.43 mm	0.6 mm more (0.17 more to 1.03 more)	‐	84¹ (1 RCT)	⊕⊕⊝⊝ low²	1 study with low risk of bias suggested that there is a statistically significant reduction in relapse with the fixed retainer, but this is unlikely to be clinically significant
Failure of retainers (lower)	See comment		‐	133 (2 RCTs)	⊕⊝⊝⊝ very low³	2 studies at high risk of bias making multiple comparisons of different types of fixed and removable retainers showed inconsistent results
Adverse effects on health Evidence of gingival bleeding 12 months	617⁴ per 1000	327 per 1000 (191 to 543)	RR 0.53 (0.31 to 0.88)	84¹ (1 RCT)	⊕⊕⊝⊝ low²	1 study also looked at adverse effects on dental health: caries and periodontal pocketing. There was low quality evidence to suggest that there was additional periodontal pocketing with the fixed retainer
Patient satisfaction How acceptable was the retainer to wear? 0 to 100 VAS (100 is very satisfied) 12 months	Mean = 91.62	12.84 lower (18.6 lower to 7.09 lower)	‐	81 (1 RCT)	⊕⊕⊝⊝ low²	1 RCT found that the participants thought the fixed retainer was easier to keep clean but found no difference in the appearance of their teeth after 1 year of retention
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; RCT: randomised controlled trial; RR: risk ratio; VAS: visual analogue scale
GRADE Working Group grades of evidence High quality: We are very confident that the true effect lies close to that of the estimate of the effect Moderate quality: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low quality: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low quality: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect
¹ Removable thermoplastic versus fixed multistrand ² Downgraded twice: once for high risk of bias and once for imprecision (only one study) ³ Downgraded three times: once for high risk of bias, once for imprecision and once for inconsistency ⁴ From control group (fixed) of study

Summary of findings for the main comparison. Removable retainer versus fixed retainer to stabilise tooth position

Summary of findings 2. Fixed retainer versus fixed retainer to stabilise tooth position

Fixed retainer versus fixed retainer to stabilise tooth position
Patient or population: people who have received fixed appliance treatment Setting: specialist practice or university orthodontic clinic Intervention: one type of fixed retainer Comparison: another type of fixed retainer
Outcomes	*Anticipated absolute effects (95% CI)**		Relative effect (95% CI)	No of participants (studies)	Quality of the evidence (GRADE)	Comments
Outcomes	Risk with fixed retainers	Risk with fixed retainers	Relative effect (95% CI)	No of participants (studies)	Quality of the evidence (GRADE)	Comments
Stability ‐ Little's Irregularity Index	‐	‐	‐	0 (0)	‐	Not reported
Failure of retainers (lower arch) 18 months to 6 years	336¹ per 1000	370 per 1000 (259 to 528)	RR 1.10 (0.77 to 1.57)	228 (3 RCTs)²	⊕⊕⊝⊝ low³	2 of these RCTs also compared failure rates in the upper arch and found no statistical difference in failure rate (RR 1.25, 95% CI 0.87 to 1.78). 1 study with a high risk of bias showed no evidence of a difference in failure rates between 3 types of metal fixed retainers
Adverse effects on health	‐	‐	‐	0 (0)	‐	Not reported
Patient satisfaction	‐	‐	‐	0 (0)	‐	Not reported
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; RCT: randomised controlled trial; RR: risk ratio
GRADE Working Group grades of evidence High quality: We are very confident that the true effect lies close to that of the estimate of the effect Moderate quality: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low quality: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low quality: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect
¹ From control group ² Mandibular polyethylene ribbon bonded retainer versus mandibular multistrand retainer ³ Downgraded twice: once for high risk of bias and once for imprecision

Summary of findings 2. Fixed retainer versus fixed retainer to stabilise tooth position

Summary of findings 3. Removable retainer versus removable retainer to stabilise tooth position

Removable retainer versus removable retainer to stabilise tooth position
Patient or population: people who have received fixed appliance treatment Setting: specialist practice, university or hospital orthodontic department Intervention: one type of removable retainer or wear time Comparison: another type of removable retainer or wear time
Outcomes	*Anticipated absolute effects (95% CI)**		Relative effect (95% CI)	No of participants (studies)	Quality of the evidence (GRADE)	Comments
Outcomes	Risk with Removable Retainers	Risk with Removable	Relative effect (95% CI)	No of participants (studies)	Quality of the evidence (GRADE)	Comments
Stability ‐ Little's Irregularity Index in upper arch Ideally 0 mm 12 months	‐	‐	‐	419 (3 RCT)	⊕⊕⊝⊝ low to ⊕⊕⊕⊝ moderate	Data used from 3 studies comparing different removable retainers used for different time periods 1 study at low risk of bias comparing part‐time thermoplastic vs. full‐time thermoplastic retainer showed no evidence to suggest a difference (graded moderate quality evidence: downgraded once due to imprecision). There was also data from additional study with a low risk of bias that could not be analysed using Review Manager 5 that suggested no evidence of difference in relapse when thermoplastic retainers were worn full‐time or part‐time 1 study comparing part‐time and full‐time Hawley retainers showed no evidence of a difference (graded low quality of evidence: downgraded twice due to high risk of bias and imprecision) 1 study at high risk of bias comparing upper Hawley retainers and upper thermoplastic retainers suggested there was statistical, but not clinically significant, reductions in relapse with thermoplastic retainers (graded low quality: downgraded twice due to high risk of bias and imprecision)
Stability ‐ Little's Irregularity Index in lower arch Ideally 0 mm 6‐12 months	‐	‐	‐	643 (4 RCT)	⊕⊕⊝⊝ low to ⊕⊕⊕⊝ moderate	Data used from 4 studies comparing different removable retainers used for different time periods. 1 study at low risk of bias comparing part‐time thermoplastic vs. full‐time thermoplastic retainer showed no evidence to suggest a difference (graded moderate quality evidence, downgraded once due to imprecision) 2 studies at high risk of bias suggested that stability was better for thermoplastic retainers vs. Hawley retainers, and for thermoplastic full‐time retainers vs. Begg (full‐time) retainers (both low quality evidence; downgraded twice due to high risk of bias and imprecision) 1 study with a high risk of bias comparing part‐time and full‐time wear of lower Hawley retainers found no evidence of any difference in relapse (low quality evidence: downgraded twice due to high risk of bias and imprecision). Data from an additional study with a low risk of bias that could not be analysed using Review Manager to support this evidence
Failure of retainers How many broke in total? 12 months	‐	‐	‐	457 (2 RCTs)	⊕⊕⊝⊝ low	2 studies provided data 1 study comparing Hawley retainers vs. part‐time thermoplastic retainers showed no difference in the number of retainers that were lost (low quality evidence, downgraded twice due to high risk of bias) 1 study with a high risk of bias comparing full‐time wear of Hawley retainers vs. full‐time thermoplastic retainers found greater failure rates in the lower thermoplastic retainers (low quality evidence, downgraded twice due to high risk of bias, and imprecision)
Adverse effects on health	‐	‐	‐	0 (0)	‐	Not reported
Patient satisfactionEmbarrassed to wear retainer? 12 months	72¹ per 1000	174 per 1000 (93 to 322)	RR 2.42 (1.30 to 4.49)	348 (1 RCT)²	⊕⊕⊝⊝ low	1 study comparing Hawley and thermoplastic retainers found participants who wore the Hawley retainers found them harder to wear as instructed, and judged them more negatively when comparing them to the fixed appliances they had worn. There was no evidence of a difference between the Hawley and thermoplastic retainer groups in terms of discomfort or ability to wear them away from home (low quality evidence, downgraded twice as one study with high risk of bias and imprecision)
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; RCT: randomised controlled trial; RR: risk ratio
GRADE Working Group grades of evidence High quality: We are very confident that the true effect lies close to that of the estimate of the effect Moderate quality: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low quality: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low quality: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect
¹ From study control group ² Hawley versus thermoplastic retainers

Summary of findings 3. Removable retainer versus removable retainer to stabilise tooth position

Table 1. O'Rourke study results: changes in stability measurements at 6, 12 and 18 months

Measure

Stability assessments at 6, 12 and 18 months

Bonded retainer

42 participants

(interquartiles)

Vacuum‐formed retainer

40 participants

(interquartiles)

Mann Whitney P value

Little's Irregularity Index

Change between T0 and T1

Change between T1 and T2

Change between T2 and T3

0.03 (0.00‐0.82)

0.04 (0.02‐0.18)

0.05 (0.02‐0.20)

0.08 (0.01‐0.40)

0.10 (0.06‐0.32)

0.07 (0.00‐0.64)

0.003 (P value < 0.05)

0.03 (P value < 0.05)

0.04 (P value < 0.05)

Intercanine width

Change between T0 and T1

Change between T1 and T2

Change between T2 and T3

0.11 (0.05‐0.39)

0.22 (0.11‐0.60)

0.29 (0.12‐0.57)

0.25 (0.10‐0.50)

0.25 (0.09‐0.58)

0.28 ( 0.12‐0.57)

0.56

0.43

0.32

Intermolar width

Change between T0 and T1

Change between T1 and T2

Change between T2 and T3

0.29 (0.11‐0.67)

0.47 (0.15‐1.04)

0.38 (0.12‐0.98)

0.18 (0.66‐0.52)

0.38 (0.14‐0.90)

0.50 (0.22‐1.05)

0.92

0.36

0.24

Arch length

Change between T0 and T1

Change between T1 and T2

Change between T2 and T3

0.24 (0.08‐0.55)

0.39 (0.13‐0.68)

0.49 (0.09‐0.95)

0.31 (0.07‐1.18)

0.26 (0.17‐2.01)

0.22 (0.12‐1.79)

0.17

0.20

0.44

Extraction site space

Change between T0 and T1

Change between T1 and T2

Change between T2 and T3

0.00 (0.00‐0.01)

0.00 (0.00‐0.17)

0.03 (0.00‐0.10)

0.00 (0.00‐0.03)

0.00 (0.00‐0.11)

0.14 (0.00‐0.17)

0.37

0.47

0.01 (P value < 0.05)

Table 1. O'Rourke study results: changes in stability measurements at 6, 12 and 18 months

Table 2. Thickett table of results: medians for full‐time (group 1) and part‐time (group 2) at debond and after one year

Outcome	Full‐time at debond	Part‐time at debond	P value at debond	Full‐time after 1 year	Part‐time after 1 year	P value after 1 year
LII	0.04	0.14	0.46	0.71	0.89	0.5
UII	0.21	0.14	0.14	1.08	1.09	0.8
Lower intercanine width	27.24	27.07	0.22	27.07	26.47	0.65
Lower intermolar width	34.09	33.32	0.52	34.35	33.6	0.61
Lower arch length	21.1	21.02	0.44	21.28	20.14	0.06
Upper intercanine width	35.49	34.79	0.08	34.93	34.56	0.52
Upper intermolar width	40.23	39.57	0.35	40.34	39.39	0.68
Upper arch length	24.98	24.19	0.22	25.07	25.15	0.97
Overjet	2.54	2.36	0.6	2.76	2.39	0.37
Overbite	2.86	3.31	0.14	3.14	3.74	0.05 (P value < 0.05)
LII: lower Little's Irregularity Index in mm; UII: upper Little's Irregularity Index in mm.

Table 2. Thickett table of results: medians for full‐time (group 1) and part‐time (group 2) at debond and after one year

Comparison 1. Removable versus fixed retainers

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Stability ‐ Little's Irregularity Index Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

1.1 Removable thermoplastic vs. fixed multistrand (lower)	1	84	Mean Difference (IV, Fixed, 95% CI)	0.60 [0.17, 1.03]
2 Failure of retainers in lower Show forest plot	2		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

2.1 Removable vs. thick plain canine to canine	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.2 Removable vs. thick spiral canine to canine	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.3 Removable vs. thin spiral wire (incisors and canines)	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.4 Removable thermoplastic vs. fixed multistrand (lower)	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
3 Adverse effects on health: evidence of caries Show forest plot	1		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only

3.1 Removable thermoplastic vs. fixed multistrand (lower)	1	84	Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
4 Adverse effects on health: evidence of gingival bleeding Show forest plot	1		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only

4.1 Removable thermoplastic vs. fixed multistrand (lower)	1	84	Risk Ratio (M‐H, Fixed, 95% CI)	0.53 [0.31, 0.88]
5 Adverse effects on health: evidence of periodontal pocketing Show forest plot	1		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only

5.1 Removable thermoplastic vs. fixed multistrand (lower)	1	84	Risk Ratio (M‐H, Fixed, 95% CI)	0.32 [0.12, 0.87]
6 Patient satisfaction: how acceptable was the retainer to wear? Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

6.1 Removable thermoplastic vs. fixed multistrand (lower)	1	81	Mean Difference (IV, Fixed, 95% CI)	‐12.84 [‐18.60, ‐7.09]
7 Patient satisfaction: how easy was retainer to keep clean? Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

7.1 Removable thermoplastic vs. fixed multistrand (lower)	1	81	Mean Difference (IV, Fixed, 95% CI)	‐10.31 [‐20.05, ‐0.58]
8 Patient satisfaction: how happy are you with appearance of teeth after 1 year of retention Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

8.1 Removable thermoplastic vs. fixed multistrand (lower)	1	81	Mean Difference (IV, Fixed, 95% CI)	1.5 [‐3.50, 6.50]

Comparison 1. Removable versus fixed retainers

Comparison 2. Fixed versus fixed retainers

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Failure of retainers Show forest plot	4		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only

1.1 Thick plain wire canine to canine vs. thick spiral wire canine to canine in lower	1	24	Risk Ratio (M‐H, Fixed, 95% CI)	0.30 [0.04, 2.27]
1.2 Thick plain wire canine to canine vs. thin spiral wire (incisors and canines) in lower	1	22	Risk Ratio (M‐H, Fixed, 95% CI)	0.33 [0.04, 2.73]
1.3 Thick spiral wire canine to canine vs. thin spiral wire (incisors and canines) in lower	1	24	Risk Ratio (M‐H, Fixed, 95% CI)	1.13 [0.32, 3.99]
1.4 Maxillary polyethylene ribbon bonded retainer vs. maxillary multistrand bonded retainer	2	174	Risk Ratio (M‐H, Fixed, 95% CI)	1.25 [0.87, 1.78]
1.5 Mandibular polyethylene ribbon bonded retainer vs. mandibular multistrand retainer	3	228	Risk Ratio (M‐H, Fixed, 95% CI)	1.10 [0.77, 1.57]

Comparison 2. Fixed versus fixed retainers

Comparison 3. Removable versus removable retainers

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Stability ‐ Little's Irregularity Index ‐ upper labial segment ‐ 1 year Show forest plot	3		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

1.1 Night wear Hawley 1 year vs. 24/7 for 6 months + night wear 6 months Hawley	1	52	Mean Difference (IV, Fixed, 95% CI)	0.0 [‐0.41, 0.41]
1.2 Hawley 24/7 for 3 months then 12/7 vs. thermoplastic 12/7 for 1 week then night wear	1	310	Mean Difference (IV, Fixed, 95% CI)	0.25 [0.08, 0.42]
1.3 Part‐time 8/7 thermoplastic vs. full‐time 24/7 thermoplastic	1	57	Mean Difference (IV, Fixed, 95% CI)	0.03 [‐0.39, 0.45]
2 Stability ‐ Little's Irregularity Index ‐ lower labial segment ‐ 1 year Show forest plot	4		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

2.1 Night wear Hawley 1 year vs. 24/7 for 6 months + night wear 6 months	1	52	Mean Difference (IV, Fixed, 95% CI)	0.20 [‐0.28, 0.68]
2.2 Hawley 24/7 for 3 months then 12/7 vs. thermoplastic 12/7 for 1 week then night wear	1	310	Mean Difference (IV, Fixed, 95% CI)	0.42 [0.23, 0.61]
2.3 Upper and lower (mean) Begg full time vs. upper and lower (mean) thermoplastic retainers full time	1	224	Mean Difference (IV, Fixed, 95% CI)	0.25 [0.19, 0.31]
2.4 Part‐time 8/7 thermoplastic vs. full‐time 24/7 thermoplastic	1	57	Mean Difference (IV, Fixed, 95% CI)	0.02 [‐0.34, 0.38]
3 Stability ‐ crowding upper labial segment ‐ 1 year Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected

3.1 Night wear Hawley 1 year vs. 24/7 for 6 months + night wear 6 months	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
4 Stability ‐ crowding lower labial segment ‐ 1 year Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected

4.1 Night wear Hawley 1 year vs. 24/7 for 6 months + night wear 6 months	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
5 Stability: lower intercanine width Show forest plot	2		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

5.1 Hawley 24/7 for 3 months then 12/7 vs. thermoplastic 12/7 for 1 month	1	310	Mean Difference (IV, Fixed, 95% CI)	0.08 [0.01, 0.15]
5.2 Part‐time 8/7 thermoplastic vs. full‐time 24/7 thermoplastic	1	57	Mean Difference (IV, Fixed, 95% CI)	0.05 [‐0.22, 0.32]
6 Stability: lower intermolar width Show forest plot	2		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

6.1 Hawley 24/7 for 3 months then 12/7 vs. thermoplastic 12/7 for 1 month	1	310	Mean Difference (IV, Fixed, 95% CI)	0.08 [‐0.01, 0.17]
6.2 Part‐time 8/7 thermoplastic vs. full‐time 24/7 thermoplastic	1	56	Mean Difference (IV, Fixed, 95% CI)	‐0.11 [‐0.59, 0.38]
7 Stability: upper intercanine width Show forest plot	2		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

7.1 Hawley 24/7 for 3 months then 12/7 vs. thermoplastic 12/7 for 1 month	1	310	Mean Difference (IV, Fixed, 95% CI)	0.05 [‐0.04, 0.14]
7.2 Part‐time 8/7 thermoplastic vs. full‐time 24/7 thermoplastic	1	57	Mean Difference (IV, Fixed, 95% CI)	‐0.00 [‐0.31, 0.30]
8 Stability: upper intermolar width Show forest plot	2		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

8.1 Hawley 24/7 for 3 months then 12/7 vs. thermoplastic 12/7 for 1 month	1	310	Mean Difference (IV, Fixed, 95% CI)	‐0.02 [‐0.12, 0.08]
8.2 Part‐time 8/7 thermoplastic vs. full‐time 24/7 thermoplastic	1	57	Mean Difference (IV, Fixed, 95% CI)	0.12 [‐0.67, 0.92]
9 Stability: overjet Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

9.1 Part‐time 8/7 thermoplastic vs. full‐time 24/7 thermoplastic	1	57	Mean Difference (IV, Fixed, 95% CI)	0.06 [‐0.35, 0.48]
10 Stability: overbite Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

10.1 Part‐time 8/7 thermoplastic vs. full‐time 24/7 thermoplastic	1	57	Mean Difference (IV, Fixed, 95% CI)	0.34 [‐0.01, 0.69]
11 Stability: settling with increased posterior contacts Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

11.1 Modified thermoplastic retainer vs. full coverage thermoplastic retainer (both 6/12 full‐time)	1	36	Mean Difference (IV, Fixed, 95% CI)	‐0.34 [‐1.40, 0.72]
11.2 Modified thermoplastic retainer vs. full coverage thermoplastic retainer (both 6/12 full‐time then 3/12 part‐time)	1	36	Mean Difference (IV, Fixed, 95% CI)	0.95 [0.05, 1.85]
12 Stability: settling with increased anterior contacts Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

12.1 Modified thermoplastic retainer vs. full coverage thermoplastic retainer (both 6/12 full‐time)	1	36	Mean Difference (IV, Fixed, 95% CI)	‐1.89 [‐2.32, ‐1.46]
12.2 Modified thermoplastic retainer vs. full coverage thermoplastic retainer (both 6/12 full‐time then 3/12 nights only)	1	36	Mean Difference (IV, Fixed, 95% CI)	‐1.5 [‐1.94, ‐1.06]
13 Stability: maintaining corrected rotations in the upper Show forest plot	1		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only

13.1 Hawley full‐time for 3/12 then nights only vs. thermoplastic part‐time	1	75	Risk Ratio (M‐H, Fixed, 95% CI)	4.88 [1.13, 21.07]
14 Stability: maintaining quality of finish (PAR) Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

15 Survival of retainers: how many broke in total Show forest plot	1	346	Risk Ratio (M‐H, Fixed, 95% CI)	2.96 [1.58, 5.55]

15.1 Hawley 24/7 for 3 months then 12/7 vs. thermoplastic 12/7 for 1 week then night wear	1	346	Risk Ratio (M‐H, Fixed, 95% CI)	2.96 [1.58, 5.55]
16 Survival of retainers: how many were lost in total Show forest plot	1		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only

16.1 Hawley 24/7 for 3 months then 12/7 vs. thermoplastic 12/7 for 1 week then night wear	1	346	Risk Ratio (M‐H, Fixed, 95% CI)	1.20 [0.56, 2.56]
17 Survival of retainers: upper Show forest plot	1		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only

17.1 Hawley 24/7 for 1 year vs. vacuum formed 24/7 for 1 year	1	111	Risk Ratio (M‐H, Fixed, 95% CI)	0.87 [0.63, 1.19]
18 Survival of retainers: lower Show forest plot	1		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only

18.1 Hawley 24/7 for 1 year vs. thermoplastic 24/7 for 1 year	1	111	Risk Ratio (M‐H, Fixed, 95% CI)	0.60 [0.43, 0.83]
19 Patient satisfaction: able to wear retainer as instructed? Show forest plot	1	347	Risk Ratio (M‐H, Fixed, 95% CI)	0.89 [0.83, 0.96]

19.1 Hawley 24/7 for 3 months then 12/7 vs. thermoplastic 12/7 for 1 week then night wear	1	347	Risk Ratio (M‐H, Fixed, 95% CI)	0.89 [0.83, 0.96]
20 Patient satisfaction: able to wear retainers away from home? Show forest plot	1		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only

20.1 Hawley 24/7 for 3 months then 12/7 vs. thermoplastic 12/7 for 1 week	1	344	Risk Ratio (M‐H, Fixed, 95% CI)	0.92 [0.81, 1.06]
21 Patient satisfaction: embarrassed to wear retainer? Show forest plot	1		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only

21.1 Hawley 24/7 for 3 months then 12/7 vs. thermoplastic 12/7 for 1 week	1	348	Risk Ratio (M‐H, Fixed, 95% CI)	2.42 [1.30, 4.49]
22 Patient satisfaction: amount of discomfort ‐ never or only on occasion Show forest plot	1		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only

22.1 Hawley 24/7 for 3 months then 12/7 vs. thermoplastic 12/7 for 1 week then night wear	1	349	Risk Ratio (M‐H, Fixed, 95% CI)	0.96 [0.89, 1.03]
23 Patient satisfaction: worse or much worse than wearing fixed appliances? Show forest plot	1		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only

23.1 Hawley 24/7 for 3 months then 12/7 vs. thermoplastic 12/7 for 1 week	1	349	Risk Ratio (M‐H, Fixed, 95% CI)	9.37 [3.80, 23.10]

Comparison 3. Removable versus removable retainers

Comparison 4. Upper removable and lower fixed retainer versus upper removable and lower adjunctive procedure

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Irregularity change of upper labial segment (V‐CTC vs. V‐S) Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

1.1 Irregularity change in upper labial segment after 1 year	1	50	Mean Difference (IV, Fixed, 95% CI)	0.10 [‐0.45, 0.65]
1.2 Irregularity change of upper labial segment after 2 years	1	47	Mean Difference (IV, Fixed, 95% CI)	‐0.4 [‐0.95, 0.15]
2 Intercanine change in upper (V‐CTC vs. V‐S) Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

2.1 Intercanine width in the upper after 1 year	1	50	Mean Difference (IV, Fixed, 95% CI)	0.10 [‐0.26, 0.46]
2.2 Intercanine width change in upper after 2 years	1	47	Mean Difference (IV, Fixed, 95% CI)	0.10 [‐0.39, 0.59]
3 Intermolar width change in upper (V‐CTC vs. V‐S) Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

3.1 Intermolar change in upper after 1 year	1	50	Mean Difference (IV, Fixed, 95% CI)	0.0 [‐0.77, 0.77]
3.2 Intermolar width change after 2 years	1	47	Mean Difference (IV, Fixed, 95% CI)	‐0.10 [‐0.90, 0.70]
4 Arch length change in upper (V‐CTC vs. V‐S) Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

4.1 Arch length change after 1 year	1	50	Mean Difference (IV, Fixed, 95% CI)	0.20 [‐0.26, 0.66]
4.2 Arch length change after 2 years	1	47	Mean Difference (IV, Fixed, 95% CI)	0.2 [‐0.29, 0.69]
5 Irregularity change in lower labial segment (V‐CTC vs. V‐S) Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

5.1 Irregularity change in lower after 1 year	1	50	Mean Difference (IV, Fixed, 95% CI)	‐0.20 [‐0.62, 0.22]
5.2 Irregularity in lower after 2 years	1	47	Mean Difference (IV, Fixed, 95% CI)	‐0.30 [‐0.73, 0.13]
6 Intercanine width change in lower (V‐CTC vs. V‐S) Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

6.1 Intercanine width change in lower after 1 year	1	50	Mean Difference (IV, Fixed, 95% CI)	‐1.0 [‐1.22, ‐0.78]
6.2 Intercanine width change in lower after 2 year	1	47	Mean Difference (IV, Fixed, 95% CI)	‐0.8 [‐1.26, ‐0.34]
7 Intermolar width change in the lower (V‐CTC vs. V‐S) Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

7.1 Intermolar width change in lower after 1 year	1	50	Mean Difference (IV, Fixed, 95% CI)	‐0.6 [‐1.43, 0.23]
7.2 Intermolar width change in lower after 2 years	1	47	Mean Difference (IV, Fixed, 95% CI)	‐0.7 [‐1.68, 0.28]
8 Arch length change in lower (V‐CTC vs. V‐S) Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

8.1 Arch length change in lower after 1 year	1	50	Mean Difference (IV, Fixed, 95% CI)	0.7 [0.20, 1.20]
8.2 Arch length change in lower after 2 years	1	47	Mean Difference (IV, Fixed, 95% CI)	0.30 [‐0.27, 0.87]
9 Overjet change (V‐CTC vs. V‐S) Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

9.1 Overjet change after 1 year	1	50	Mean Difference (IV, Fixed, 95% CI)	‐0.3 [‐0.83, 0.23]
9.2 Overjet change after 2 years	1	47	Mean Difference (IV, Fixed, 95% CI)	‐0.2 [‐0.83, 0.43]
10 Overbite change (V‐CTC vs. V‐S) Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

10.1 Overbite change after 1 year	1	47	Mean Difference (IV, Fixed, 95% CI)	0.0 [‐0.57, 0.57]
10.2 Overbite change after 2 years	1	47	Mean Difference (IV, Fixed, 95% CI)	0.0 [‐0.57, 0.57]
11 Final irregularity in upper after more than 5 years (V‐CTC vs. V‐S) Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

12 Final irregularity in lower after more than 5 years (V‐CTC vs. V‐S) Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

Comparison 4. Upper removable and lower fixed retainer versus upper removable and lower adjunctive procedure

Comparison 5. Upper removable retainer and lower fixed retainer versus positioner

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Irregularity change in upper labial segment (V‐CTC vs. P) Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

1.1 Irregularity change in upper labial segment after 1 year	1	48	Mean Difference (IV, Fixed, 95% CI)	‐0.20 [‐0.83, 0.43]
1.2 Irregularity change in upper labial segment after 2 years	1	46	Mean Difference (IV, Fixed, 95% CI)	‐0.60 [‐1.27, 0.07]
2 Intercanine width change in upper (V‐CTC vs. P) Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

2.1 Intercanine width change in upper after 1 year	1	48	Mean Difference (IV, Fixed, 95% CI)	‐0.90 [‐1.92, 0.12]
2.2 Intercanine width change in upper after 2 years	1	46	Mean Difference (IV, Fixed, 95% CI)	‐0.8 [‐1.50, ‐0.10]
3 Intermolar width change in upper (V‐CTC vs. P) Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

3.1 Intermolar width change in upper after 1 year	1	48	Mean Difference (IV, Fixed, 95% CI)	‐0.20 [‐1.00, 0.60]
3.2 Intermolar width change in upper after 2 years	1	46	Mean Difference (IV, Fixed, 95% CI)	‐0.30 [‐1.08, 0.48]
4 Arch length change in upper (V‐CTC vs. P) Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

4.1 Arch length change in upper after 1 year	1	48	Mean Difference (IV, Fixed, 95% CI)	0.0 [‐1.11, 1.11]
4.2 Arch length change in upper after 2 years	1	46	Mean Difference (IV, Fixed, 95% CI)	‐0.1 [‐1.30, 1.10]
5 Irregularity change in lower labial segment (V‐CTC vs. P) Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

5.1 Irregularity change in lower after 1 year	1	48	Mean Difference (IV, Fixed, 95% CI)	‐0.5 [‐1.01, 0.01]
5.2 Irregularity change in lower after 2 years	1	46	Mean Difference (IV, Fixed, 95% CI)	‐1.0 [‐1.65, ‐0.35]
6 Intercanine width change in lower (V‐CTC vs. P) Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

6.1 Intercanine width change in lower after 1 year	1	48	Mean Difference (IV, Fixed, 95% CI)	‐0.7 [‐1.14, ‐0.26]
6.2 Intercanine width change in lower after 2 years	1	46	Mean Difference (IV, Fixed, 95% CI)	‐0.8 [‐1.34, ‐0.26]
7 Intermolar width change in lower (V‐CTC vs. P) Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

7.1 Intermolar width change in lower after 1 year	1	48	Mean Difference (IV, Fixed, 95% CI)	‐0.8 [‐1.65, 0.05]
7.2 Intermolar width change in lower after 2 years	1	46	Mean Difference (IV, Fixed, 95% CI)	‐1.0 [‐1.93, ‐0.07]
8 Arch length change in lower (V‐CTC vs. P) Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

8.1 Arch length change in lower after 1 year	1	48	Mean Difference (IV, Fixed, 95% CI)	0.20 [‐0.31, 0.71]
8.2 Arch length change in lower after 2 years	1	46	Mean Difference (IV, Fixed, 95% CI)	0.4 [‐0.15, 0.95]
9 Overjet change (V‐CTC vs. P) Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

9.1 Overjet change after 1 year	1	48	Mean Difference (IV, Fixed, 95% CI)	‐0.3 [‐0.81, 0.21]
9.2 Overjet change after 2 years	1	46	Mean Difference (IV, Fixed, 95% CI)	‐0.10 [‐0.71, 0.51]
10 Overbite change (V‐CTC vs. P) Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

10.1 Overbite change after 1 year	1	48	Mean Difference (IV, Fixed, 95% CI)	0.2 [‐0.32, 0.72]
10.2 Overbite change after 2 years	1	46	Mean Difference (IV, Fixed, 95% CI)	0.2 [‐0.50, 0.90]
11 Final irregularity in upper after more than 5 years (V‐CTC vs. P) Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

12 Final irregularity in lower after more than 5 years (V‐CTC vs. P) Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

Comparison 5. Upper removable retainer and lower fixed retainer versus positioner

Comparison 6. Upper removable retainer and lower adjunctive procedure versus positioner (P)

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Irregularity change in upper labial segment (V‐S vs. P) Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

1.1 Irregularity change in upper labial segment after 1 year	1	48	Mean Difference (IV, Fixed, 95% CI)	‐0.30 [‐0.93, 0.33]
1.2 Irregularity change in upper labial segment after 2 years	1	45	Mean Difference (IV, Fixed, 95% CI)	‐0.20 [‐0.94, 0.54]
2 Intercanine width change in upper (V‐S vs. P) Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

2.1 Intercanine width change in upper after 1 year	1	48	Mean Difference (IV, Fixed, 95% CI)	0.0 [‐1.01, 1.01]
2.2 Intercanine width change in upper after 2 years	1	45	Mean Difference (IV, Fixed, 95% CI)	‐0.9 [‐1.63, ‐0.17]
3 Intermolar width change in upper (V‐S vs. P) Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

3.1 Intermolar width change in upper after 1 year	1	48	Mean Difference (IV, Fixed, 95% CI)	‐0.20 [‐0.80, 0.40]
3.2 Intermolar width change in upper after 2 years	1	45	Mean Difference (IV, Fixed, 95% CI)	‐0.20 [‐0.93, 0.53]
4 Arch length change in upper (V‐S vs. P) Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

4.1 Arch length change in upper after 1 year	1	48	Mean Difference (IV, Fixed, 95% CI)	‐0.20 [‐1.23, 0.83]
4.2 Arch length change in upper after 2 years	1	45	Mean Difference (IV, Fixed, 95% CI)	0.1 [‐1.06, 1.26]
5 Irregularity change in lower labial segment (V‐S vs. P) Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

5.1 Irregularity change in lower labial segment after 1 year	1	48	Mean Difference (IV, Fixed, 95% CI)	‐0.30 [‐0.84, 0.24]
5.2 Irregularity change in lower labial segment after 2 years	1	45	Mean Difference (IV, Fixed, 95% CI)	‐0.70 [‐1.37, ‐0.03]
6 Intercanine width change in lower (V‐S vs. P) Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

6.1 Intercanine width change in lower after 1 year	1	48	Mean Difference (IV, Fixed, 95% CI)	‐0.7 [‐1.14, ‐0.26]
6.2 Intercanine width change in lower after 2 years	1	45	Mean Difference (IV, Fixed, 95% CI)	‐0.10 [‐0.75, 0.55]
7 Intermolar width change in lower (V‐S vs. P) Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

7.1 Intermolar width change in lower after 1 year	1	48	Mean Difference (IV, Fixed, 95% CI)	‐0.20 [‐1.22, 0.82]
7.2 Intermolar width change in lower after 2 years	1	45	Mean Difference (IV, Fixed, 95% CI)	0.1 [‐0.40, 0.60]
8 Arch length change in lower (V‐S vs. P) Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

8.1 Arch length change in lower after 1 year	1	48	Mean Difference (IV, Fixed, 95% CI)	‐0.5 [‐0.95, ‐0.05]
8.2 Arch length change in lower after 2 years	1	45	Mean Difference (IV, Fixed, 95% CI)	0.1 [‐0.40, 0.60]
9 Overjet change (V‐S vs. P) Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

9.1 Overjet change after 1 year	1	48	Mean Difference (IV, Fixed, 95% CI)	0.0 [‐0.54, 0.54]
9.2 Overjet change after 2 years	1	45	Mean Difference (IV, Fixed, 95% CI)	0.10 [‐0.51, 0.71]
10 Overbite change (V‐S vs. P) Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

10.1 Overbite change after 1 year	1	48	Mean Difference (IV, Fixed, 95% CI)	0.0 [‐0.49, 0.49]
10.2 Overbite change after 2 years	1	45	Mean Difference (IV, Fixed, 95% CI)	0.2 [‐0.46, 0.86]
11 Final irregularity in upper after more than 5 years (V‐S vs. P) Show forest plot	1	33	Mean Difference (IV, Fixed, 95% CI)	0.30 [‐0.87, 1.47]

12 Final irregularity in lower after more than 5 years (V‐S vs. P) Show forest plot	1	33	Mean Difference (IV, Fixed, 95% CI)	‐1.4 [‐2.77, ‐0.03]

Comparison 6. Upper removable retainer and lower adjunctive procedure versus positioner (P)