The effects of fixed and removable orthodontic retainers: a systematic review

In the view of the widespread acceptance of indefinite retention, it is important to determine the effects of fixed and removable orthodontic retainers on periodontal health, survival and failure rates of retainers, cost-effectiveness, and impact of orthodontic retainers on patient-reported outcomes. A comprehensive literature search was undertaken based on a defined electronic and gray literature search strategy (PROSPERO: CRD42015029169). The following databases were searched (up to October 2015); MEDLINE via OVID, PubMed, the Cochrane Central Register of Controlled Trials, LILACS, BBO, ClinicalTrials.gov, the National Research Register, and ProQuest Dissertation and Thesis database. Randomized and non-randomized controlled clinical trials, prospective cohort studies, and case series (minimum sample size of 20) with minimum follow-up periods of 6 months reporting periodontal health, survival and failure rates of retainers, cost-effectiveness, and impact of orthodontic retainers on patient-reported outcomes were identified. The Cochrane Collaboration’s Risk of Bias tool and Newcastle-Ottawa Scale were used to assess the quality of included trials. Twenty-four studies were identified, 18 randomized controlled trials and 6 prospective cohort studies. Of these, only 16 were deemed to be of high quality. Meta-analysis was unfeasible due to considerable clinical heterogeneity and variations in outcome measures. The mean failure risk for mandibular stainless steel fixed retainers bonded from canine to canine was 0.29 (95 % confidence interval [CI] 0.26, 0.33) and for those bonded to canines only was 0.25 (95 % CI: 0.16, 0.33). A meta-regression suggested that failure of fixed stainless steel mandibular retainers was not directly related to the period elapsed since placement (P = 0.938). Further well-designed prospective studies are needed to elucidate the benefits and potential harms associated with orthodontic retainers.


Introduction
Retention procedures are considered necessary to maintain the corrected position of teeth following orthodontic treatment and to mitigate against characteristic agerelated changes which, if unchecked, are known to culminate in mandibular anterior crowding [1]. Retention procedures are continually being refined with a recognition that existing protocols are infallible [2]. Nevertheless, both fixed and removable retainers continue to be in vogue, although adjunctive procedures including interproximal enamel reduction and minor oral surgical procedures have also been advocated.
A recent Cochrane review exposed a lack of highquality evidence to favor one method of retention over another in terms of stability [3]. Given this absence of definitive evidence, retainer selection is often based on individual preference. This is evidenced by marked geographical variation with maxillary Hawley or vacuumformed retainers and mandibular fixed lingual retainers with full-time wear of removable retainers most popular in the USA [4,5]. In Australia and New Zealand, mandibular fixed and maxillary vacuum-formed retainers are shown to be the most prevalent combination [6], while a preference for the use of fixed retainers in both arches has been shown in the Netherlands [7].
The duration of wear of orthodontic retainers has long been a dilemma in orthodontics. However, there is now widespread acceptance of the necessity for indefinite retention to minimize both relapse and maturational changes [5,8]. Prolonged retention may pose increased risk to the periodontium and dental hard tissues; it is therefore important to investigate the implications of the long-term use of fixed and removable retainers on the supporting tissues [3,9,10].
A further consideration is patient experiences of retention and compliance with prolonged retention regimes; it is intuitive to expect that co-operation with retention regimes would decline over time. Moreover, both fixed and removable retainers are prone to breakage, loss, and degradation [2,11]. Repeated breakage and requirement for replacement may have implications for the costeffectiveness of both fixed and removable approaches. There is however limited evidence concerning the costeffectiveness of either approach [12,13].
The primary aim of this systematic review was to determine the influence of fixed and removable orthodontic retainers on periodontal health in patients who have completed orthodontic treatment with fixed appliances. A secondary aim was to evaluate survival and failure rates, impact of orthodontic retainers on patient-reported outcomes, and cost-effectiveness.

Materials and methods
This protocol for this systematic review was registered on PROSPERO (www.crd.york.ac.uk/prospero; CRD42015 029169). The following selection criteria were applied: Study design: randomized and non-randomized controlled clinical trials, prospective cohort studies, and case series (with a minimum sample size of 20 patients) with minimum follow-up periods of 6 months Participants: patients having had orthodontic treatment with fixed or removable appliances followed by orthodontic retention Interventions: fixed retainers, removable retainers, and interproximal reduction Outcome measures: periodontal outcomes, survival and failure rates (including detachment of fixed retainers, breakages, retainer loss, or the need for replacement), patient-reported outcomes, and cost-effectiveness measures Search strategy for identification of studies  (6) outcomes (primary and secondary).

Risk of bias (quality) assessment
For randomized controlled trials sequence generation, allocation concealment, blinding of outcome assessors, incomplete outcome data, selective reporting, and other biases were assessed using the Cochrane Collaboration's Risk of Bias tool. Any disagreement was resolved by joint discussion (DA, PSF). Only studies at low or unclear risk of bias overall were to be included in the meta-analysis. The methodological quality of the included nonrandomized studies was assessed using the Newcastle-Ottawa Scale. Studies adjudged to be of moderate or high methodological quality overall (more than five stars) were to be included in the meta-analysis. The authors of the included studies were contacted for clarification if required.

Strategy for data synthesis
Clinical heterogeneity was assessed according to the treatment interventions, wear regimen for removable retainers, measurement approach, and location of the retainers. For periodontal outcomes, the index used and surfaces examined were considered. Statistical heterogeneity was to be assessed by inspecting a graphic display of the estimated treatment effects from individual trials with associated 95 % confidence intervals. Heterogeneity would be quantified using I-squared with values above 50 % indicative of moderate to high heterogeneity which might preclude meta-analysis. A weighted treatment effect was to be calculated, and the results for retainer failure were expressed as odds ratios. All statistical analyses were undertaken using the Stata statistical software package (version 12.1; StataCorp, College Station, Tex).

Description of the included studies
Sixty-four were considered potentially relevant to the review. Following retrieval of the full-text articles, 36 studies were excluded. Overall, 24 studies met the inclusion criteria ( Fig. 1). Reasons for exclusion at the final stage are presented (Appendix 2). The study design, characteristics of participants, comparison groups, follow-up period, and the outcomes of the included studies are presented in Table 1.

Periodontal outcomes
Of the included trials, only seven trials assessed periodontal outcomes (Tables 2 and 3) [14,16,23,25,28,30,31]. Four of these were randomized controlled trials [14,16,23,31], and the other three were prospective cohort studies [25,28,30]. Two trials did not report baseline scores [14,25], and another two studies reported the periodontal outcome with no distinction made between maxillary and mandibular measurements [30,31]. No significant difference was found between mandibular stainless steel fixed retainers bonded to the anterior teeth and canines only in terms of periodontal outcomes, at 12-month and 3-year follow-ups in two studies [25,28]. With regard to periodontal outcomes of mandibular Hawley retainers in comparison to mandibular stainless steel fixed retainers, no significant difference was found at 3-year follow-up [28]. When mandibular fiber-reinforced composite was compared to mandibular stainless steel fixed retainers, no significant difference in probing depths, bleeding on probing, and calculus scores Fig. 2 Risk of bias for included randomized controlled trials. Low risk of bias (green). Unclear risk of bias (yellow). High risk of bias (red) at 6-month follow-up was found [16,23]. Probing depths and bleeding on probing were further measured at 12month follow-up and showed no significant difference between the two groups [23]. However, gingival and plaque indices scores were found to be higher in maxillary and mandibular fiber-reinforced composite compared to stainless steel fixed retainers at 6-month follow-up [16]. Very few overlapping studies were identified, however. Metaanalysis was therefore not possible in view of heterogeneity.
In terms of the natural history of periodontal changes related to stainless steel fixed retainers, plaque and gingival indices scores on the lingual surfaces of mandibular anterior teeth increased from baseline to 6 months followup; however, this was not statistically significant [16]. At 3-year follow-up, plaque and gingival indices scores remained low [28]. No significant changes in Calculus Index scores at 6-month [16] and 3-year follow-ups [28] were observed in two studies. Bleeding on probing scores for stainless steel fixed retainer increased at both 6 months [16,23] and 12 months [23] from baseline, although only one study found this to be statistically significant [23]. Similar patterns were observed for fixed fiber-reinforced composite retainers [16,23]. Conversely, plaque, calculus, and gingival indices scores reduced at 3-year follow-up in relation to the lingual of the mandibular anterior teeth with Hawley retainers [28]. However, Gingival Index scores were shown to increase on the buccal surfaces of maxillary and mandibular anterior teeth in one study at 6-month follow-up [30].

Survival and failure rates of retainers
The survival rate of fixed retainers was reported over 12 to 24 months [18,24,26]. In terms of retainer material, one study found fiber-reinforced thermoplastic fixed retainer with polyethylene terephthalate glycol matrix resin survived significantly less than fiber-reinforced thermoplastic fixed retainer with polycarbonate matrix resin [26]. Two other studies found no significant difference in the survival rate of multistrand stainless steel fixed and esthetic retainers made of polyethylene woven ribbon or polyethelene fiber-reinforced resin composite [18,24]. No statistical difference was found in the survival rate between maxillary and mandibular fixed retainers [18,26]. Interestingly, in one study, the survival rate of fiber-reinforced thermoplastic fixed retainers was directly related to the thickness of the wire and the number of teeth bonded [26].
All the studies that involved mandibular stainless steel retainers reported failures per patient [13, 14, 17-20, 22-25, 27, 28], except for two studies in which the failure was reported per tooth [17,24] (Table 4). The mean failure risk for mandibular stainless steel fixed retainers bonded to canine to canine was 0.29 (95 % confidence interval [CI], 0.26, 0.33) based on nine studies (n = 555) (Fig. 4). The follow-up period ranged from 6 to 36 months. Similarly, the failure risk for mandibular stainless steel fixed retainers bonded to canines was 0.25 (95 % CI, 0.16, 0.33) based on three studies [13,25,28] (n = 79) over a follow-up period of 12 to 36 months (Fig. 5). Considerable statistical heterogeneity was noted in both analyses (I-squared = 89 %) reflecting high levels of inconsistency and limited numbers of events. A metaregression shows that follow-up period was not a predictor of failure rate for mandibular stainless steel fixed retainers (P = 0.938).
One study reporting failure rates of mandibular Hawley retainers was unclear regarding the stipulated duration of wear [28]. However, two studies found around 12 % failure over a period of 6 months and 14 % at 3-year follow-up [12,28]. Similarly, the failure rate for maxillary vacuumformed retainers was found to be 10 % over 2 years [13], while a further study reported a higher rate of 17 % over 6 months [12].

Patient-reported outcomes and cost-effectiveness
Patient-reported outcomes were reported in two studies [12,24] (Table 5). Removable retainers were found to be associated with discomfort, with those in the Hawley          retainer group reporting higher levels of embarrassment in terms of speech and esthetics [12].
In terms of cost-effectiveness (Table 5), vacuumformed retainers were found to be significantly more cost-effective than Hawley retainers within the National Health Service over a 6-month retention period [12]. One study, over 2 years, found interproximal reduction as a retention method and positioners to be more costeffective than mandibular stainless steel fixed retainers bonded to canines [13].

Discussion
This systematic review found a lack of evidence to endorse the use of one type of orthodontic retainer based on their effect on periodontal health, survival and failure rates, patient-reported outcomes, and cost-effectiveness. Largely, this finding can be attributed to a lack of highquality, relevant research. In this respect, the results of the present systematic review are in line with previous systematic reviews [3,9,10]. Interestingly, it was found that failure of fixed stainless steel mandibular retainers was not directly related to the duration of follow-up. This suggests that other factors including the influence of operator technique and experience might override the effects of retainer design or materials, although followup did not extend beyond 3 years in the present review.
Generally, relatively minor changes in periodontal parameters were reported; however, given that most studies did not incorporate an untreated control, or indeed a control group without retention, it was unclear whether these changes were attributable to the intervention or temporal changes, in isolation. As such additional research including prospective cohort studies with matched controls incorporating baseline assessment would be helpful in providing more conclusive information. It is worthy of mention that the current standard of care is to recommend bonded retention to preserve orthodontic correction in those with a history of periodontal disease as these patients are known to be particularly susceptible to post-treatment changes [32,33]. It is therefore important that there is greater clarity in relation to the compatibility of fixed retention with periodontal health and indeed on variations that may facilitate maintenance of optimal hygiene.
A minimum follow-up period of 6 months was set to distinguish between gingival inflammation associated with fixed orthodontic treatment and periodontal sideeffects related to the orthodontic retainers [34]. Previous reviews have stipulated a minimum observation period of 3 months [3,9] to 2 years [10]. However, a 3-month period might be insufficient to allow for the resolution of inflammatory changes related to the presence of active appliances. Using a minimum of 2-year observation period risks omission of a considerable amount of relevant research. Moreover, in this review, just one study focusing on periodontal outcomes involved follow-up in excess of 2 years [28]. It is therefore clear that the prolonged effect of orthodontic retention on periodontal health has not been adequately addressed in prospective research.
Intuitively, a significant difference in patient-reported outcomes and experiences could be expected with fixed or removable retainers in view of differences in appearance, size, and requirement for compliance. Notwithstanding this, only two studies reported on satisfaction with the appearance of retainers or on levels of associated embarrassment or discomfort [12,24]. This tendency for researchers to concentrate on objective, often cliniciancentered outcomes has recently been documented both within orthodontics and general dental research more broadly [35,36]. Further studies incorporating patientreported outcomes are therefore necessary to provide a more holistic assessment of benefits, harms, and experiences associated with orthodontic retainers.
While the primary focus of this review was to compare the effectiveness of retainer types, it was also possible to generate epidemiological data on the risk of failure of fixed retainers based on the primary studies. Failure risk of 0.29 was found for fixed wires bonded to the six anterior teeth and approximately one-quarter of retainers bonded to mandibular canines only, based on observation periods of 6 months to 3 years. This data highlights that the risk of failure is considerable and that fixed retention does not guarantee prolonged stability. Similar findings have been observed in observational studies [2]. The onus on realistic treatment planning with due consideration for placement of teeth into a zone of relative stability therefore remains [37].
Attempts were made to identify all trials meeting the inclusion criteria in the present review with no restrictions based on either publication date or language. Furthermore, we planned to include both prospective cohort studies and randomized controlled trials. Cohort studies were included, in particular, to permit assessment of periodontal outcomes as they are more likely to involve more prolonged periods of follow-up, which may be necessary to reveal the extent of prolonged periodontal  effects. Meta-analysis was not undertaken in view of the clinical heterogeneity between the limited number of included studies, which made statistical pooling inappropriate in relation to periodontal health, survival and failure rates, patient-reported outcomes, and cost-effectiveness. This inability to undertake meta-analysis is common to many orthodontic systematic reviews with meta-analysis found in just 27 % of 157 reviews over a 14-year period with a median of just 4 studies for those that did incorporate meta-analysis [38]. The onus on producing highquality primary research studies in orthodontics remains.

Conclusions
There is a lack of high-quality evidence to endorse the use of one type of orthodontic retainer based on their effect on periodontal health, risk of failure, patientreported outcomes, and cost-effectiveness. Further welldesigned prospective studies are therefore required to provide further definitive information in relation to the benefits and potential harms of prolonged retention. Table 6 Excluded studies with reasons for exclusion (n = 36)