RANDOMIZED CONTROLLED TRIAL

Effectiveness of bonded and vacuum-formed

retainers: A prospective randomized controlled
clinical trial
Niamh O'Rourke,a Hussein Albeedh,a Pratik Sharma,b and Ama Johalc
London, United Kingdom

Introduction: The objective of this prospective trial was to compare the clinical effectiveness of bonded retainers
with vacuum-formed retainers, in terms of maintaining the results of orthodontic treatment in the lower arch up to
18 months post debond. Methods: This was a hospital-based, prospective randomized controlled clinical trial in
which a total of 82 subjects were randomly allocated using a computer-generated number sequence to 1 of 2
groups, receiving either a vacuum-formed retainer (Essix Ace plastic (120 mm; DENTSPLY Raintree Essix,
Sarasota, Fla) or a bonded retainer (0.0175 coaxial archwire (Ortho-Care, UK, Shipley, United Kingdom)
bonded in place with Transbond LR (3M United Kingdom, Brachnell, United Kingdom) for the mandibular
arch. Each number was placed in an opaque, concealed envelope and chosen randomly by the study
subject; this determined the allocation group. Eligibility criteria included patients nearing debond after
treatment with 0.022 3 0.028-in slot size preadjusted edgewise fixed orthodontic appliances whose
pretreatment records and study models were available to confirm pretreatment labial segment crowding or
spacing and who had clinically acceptable alignment at the end of treatment. The main outcome was to
investigate the clinical effectiveness of the 2 types of retainers in terms of changes in incisor irregularity at
6 months of retention. The following measurements were recorded at each time point (6, 12, and 18 months)
with a digital caliper: Little's irregularity index, intercanine width, intermolar width, arch length, and extraction
site opening. Blinding was applicable only at debond because of the permanence of 1 intervention. Results:
The 2 groups were well matched with respect to age, sex, clinical characteristics, and treatment plans. There
was a statistically significant difference between the groups for changes in Little's irregularity index at 6 months,
with the vacuum-formed retainer group showing greater changes than the bonded retainer group (P 5 0.008).
There was no statistically significant difference between the groups for changes in Little's irregularity index at
12 and 18 months.There were also no statistically significant changes at any time for intercanine width,
intermolar width, arch length, or extraction site opening. Conclusions: Some relapse is likely after fixed appli-
ance therapy irrespective of retainer choice, and this is minimal in most patients at 6 months after debond.
Bonded retainers have a better ability to hold the mandibular incisor alignment in the first 6 months after treat-
ment than do vacuum-formed retainers. Registration: Not applicable. Protocol: The protocol was not published
before trial commencement. Funding: There is no funding or conflict of interest to be declared. (Am J Orthod
Dentofacial Orthop 2016;150:406-15)

a
Specialty registrar in orthodontics, Centre for Oral Growth and Development,
Barts and the London School of Medicine and Dentistry, London, United
Kingdom; postgraduate student in orthodontics, Institute of Dentistry, Queen
Mary University, London, United Kingdom.
b
Senior clinical lecturer and consultant orthodontist, Centre for Oral Growth and
Development, Barts and the London School of Medicine and Dentistry, London,
United Kingdom.
c
Reader, honorary consultant orthodontist, and academic lead in orthodontics,
Institute of Dentistry, Queen Mary University, London, United Kingdom.
All authors have completed and submitted the ICMJE Form for Disclosure of
Potential Conflicts of Interest, and none were reported.
Address correspondence to: Ama Johal, Oral Growth & Development, Fourth
Floor, Institute of Dentistry, Turner Street, Whitechapel London E1 2AD;
e-mail, a.s.johal@qmul.ac.uk.
Submitted, March 2015; revised and accepted, March 2016.
0889-5406/$36.00
Ó 2016 by the American Association of Orthodontists. All rights reserved.
http://dx.doi.org/10.1016/j.ajodo.2016.03.020
406
T
he goal of orthodontic treatment is to produce an
ideal occlusion that is morphologically stable,
esthetic, and functional.1 Despite proper diagnosis
and carefully rendered treatment mechanics, the results
achieved at the end of active treatment are not neces-
sarily stable over the long term.
Posttreatment relapse is perhaps the most common
risk of orthodontic treatment, and planning for
postretention stability should be part of the initial
treatment plan and discussed with the patient during
the informed consent process before treatment, so that
any relapse is not a disappointment for either the
clinician or patient.
O'Rourke et al
Stability and relapse, in both treated and untreated
malocclusions, have been studied intently over many
years,2-10 and the long-term results have been similar
and not hugely optimistic. Sadowsky and Sakols6 fol-
lowed patients on average for 20 years postretention
and found that 9% had an increase in mandibular
crowding when compared with pretreatment, and 73%
had dental relationships “outside the norm.” Similarly,
Little et al10 noted that only 10% of patients had main-
tained satisfactory mandibular incisor alignment at
20 years postretention.
This previous research demonstrates that the only
apparent guarantee of long-term stability is long-term
retention. This is due to the variety of factors that are
reported to affect tooth positions in both treated and un-
treated malocclusions. These include skeletal and soft tissue
growth11-13; dental factors14-16; treatment mechanics such
as changes in arch form,17 length,18 width,19 and treatment
plan20-26; final interdigitation27,28; and functional
occlusion,29 as well as elements of the pretreatment
malocclusion.30
Retention is necessary to allow reorganization of the
gingival and periodontal tissues affected by orthodontic
tooth movement, to prevent unwanted movement as a
result from growth changes, and to prevent the relapse
tendency of teeth that have been moved to an inherently
unstable position.31
In the United Kingdom, the most common types of
retention appliances are vacuum-formed retainers,
Hawley retainers, and bonded retainers, with the latter
the most frequently used by private practitioners; the
former are more commonly prescribed by the National
Health Service.32 A similar study in the United States
found that a maxillary Hawley and a mandibular bonded
retainer were the most popular.33 In a trial carried out in
a specialist practice in the National Health Service, Hi-
chens et al34 reported that a vacuum-formed retainer
was preferred by the patients over Hawley retainers.
Cerny eta al35 identified a patient preference for bonded
retainers in private practice. More recently, social per-
ceptions of intellectual ability and attractiveness have
also been found to be influenced by retainer design
and appearance.36
Previous prospective research evaluating the clinical
effectiveness of removable retention is limited. In a trial
reported by Rowland et al,37 a statistically significant
difference was found between the clinical effectiveness
of vacuum-formed retainers and Hawley retainers, with
the vacuum-formed group more successful in maintain-
ing posttreatment alignment of the anterior teeth after
6 months.
Previous research involving bonded retention has
been mainly retrospective,38-41 and the few available
American Journal of Orthodontics and Dentofacial Orthopedics
407
prospective studies have investigated failure rates and
dental health associated with fixed retainer types as
opposed to their clinical effectiveness.42-46 It was
noted in 1 study that thin multistranded wires were
superior for maintaining mandibular incisor positions
compared with a thicker wire and a prefabricated wire.44
There is 1 prospectively designed trial comparing
bonded and vacuum-formed retainers up to 24 months
after debond.47,48 In these studies, it was reported that a
prefabricated positioner used as a retainer showed a
statistically significant difference in its inability to
maintain incisor positions after treatment (measured
with Little's irregularity index) compared with a
vacuum-formed retainer or a bonded retainer after
6 months,47 but no statistically significant difference
was found after 2 years.48
Retention type and duration of wear are also ongoing
contentious issues in the specialty.49-52 Two Cochrane
reviews have been published on relapse; the latest
reviewed the management of relapse and found no
study to include in the review.53,54 The former review
identified limitations to previous research on retention
type including short follow-up periods, inappropriate
or no controls, retrospective designs, and insufficient
or irrelevant data. Thus, both highlighted the need for
randomized controlled trails in this area to aid in deter-
mining the most effective and safe method for manag-
ing the relapse of alignment of the mandibular front
teeth.
The purposes of this study were to quantify and
compare the changes in a number of intra-arch variables
with vacuum-formed retainers and bonded retainers
from debond to 6, 12, and 18 months and to determine
whether 1 type of retainer is superior to the other in
terms of maintaining the orthodontic results. These
particular retainers have to date not been directly
compared in a randomized controlled trial.
SPECIFIC OBJECTIVES AND HYPOTHESES
The main aim of this randomized controlled trial was
to compare the clinical effectiveness of 2 types of ortho-
dontic retainers in the mandibular arch in terms of reten-
tion of the treated results at 6 months after debond.
More specifically, our aim was to determine whether
there are any differences in the clinical effectiveness of
vacuum-formed retainers compared with bonded re-
tainers in maintaining alignment in the mandibular
labial segment (Little's irregularity index) at 6 months
after debond. Also, we aimed to investigate whether
there are any differences in the clinical effectiveness of
vacuum-formed retainers compared with bonded re-
tainers in maintaining arch width (intercanine width)
September 2016
Vol 150
Issue 3
408
and intermolar width, arch length, and extraction site
closure. In addition to the 6-month time point, 12 and
18 months were also considered because previous retro-
spective studies have shown that most relapse after or-
thodontic treatment occurs within the first 12 months
after debonding of the appliances.47,48
The null hypothesis tested in this study was that there
is no difference in the ability of both types of retainers to
retain the treated result in the short and medium terms.
MATERIAL AND METHODS
Trial design
This was a hospital-based parallel design randomized
controlled clinical trial.
Participants, eligibility criteria, and setting
Ethical approval was granted by the National
Research Ethics Services Committee (reference number,
10/H0713/57).
The study population was drawn from patients near-
ing debond after treatment with 0.022 3 0.028-in slot
size preadjusted edgewise fixed orthodontic appliances
who fulfilled the following inclusion criteria: (1) patients
nearing debond of fixed orthodontic appliances, (2) rec-
ords and study models available to confirm the pretreat-
ment labial segment crowding or spacing, and (3)
clinically acceptable alignment at the end of treatment
(no clinically detectable contact point displacement).
The exclusion criteria were the following: (1) patients
who completed treatment early or had repeated break-
ages during treatment, (2) patients with poor oral hygiene
during treatment, (3) patients with prosthodontic needs
in the mandibular arch after treatment, (4) patients
with a history of periodontal disease, and (5) patients
with a learning difficulty.
Subjects who fulfilled the selection criteria were iden-
tified and invited to take part in this study. Informed
consent and assent were obtained on the day of debond.
Enrollment started in January 2011 and was
completed by February 2012. Eighty-five subjects were
entered into the trial, but during recruitment 1 declined
to take part and 2 more subjects did not meet the inclu-
sion criteria. Therefore, 82 subjects were recruited in to-
tal. At the debond appointment after appliance removal
(T0), a set of alginate impressions was taken for study
models, and an additional polysilicone impression of
the mandibular arch was taken to obtain measurements
for this study. The polysilicone impression was then cast
in hard stone (type III) in the laboratory on the same day.
All subjects were then randomized into the intervention
groups.
September 2016
Vol 150
Issue 3 American Journal of Orthodontics and Dentofacial Orthopedics
O'Rourke et al
Interventions
The vacuum-formed retainer was constructed from
Essix Ace plastic (120 mm in diameter, 0.03 in thick;
DENTSPLY Raintree Essix, Sarasota, Fla). This was fabri-
cated by a qualified orthodontic technician under stan-
dardized conditions, using the same technique and the
same machine for each retainer made for this study.
This was fitted within 7 days with instructions for full-
time wear for the first 6 months followed by 6 months
of nighttime wear and a further 6 months of alternate-
night wear, which was the department protocol after
delivery of retainers.
For the bonded retainer group, the teeth were pol-
ished with pumice, and scaling (Cavitron; DENTSPLY In-
ternational, York, Pa) was used if deemed necessary. A
0.0175-in stainless steel coaxial archwire (Ortho-Care
UK, Shipley, United Kingdom) was formed at chairside
to fit passively against the mandibular labial segment
from canine to canine, using a 37% phosphoric acid
etch followed by copious washing, drying, and applica-
tion of an adhesive primer (OrthoSolo bonding agent;
Ormco, Orange, Calif), which was subsequently cured.
The wire was then bonded with Transbond LR composite
material (3M United Kingdom, Bracknell, United
Kingdom).
All subjects were reviewed by a member of the
research team (N.O'R. or H.A.), and a mandibular arch
dental impression was taken with polysilicone at
6 months (T1), 12 months (T2), and 18 months (T3) to
obtain stone models for measurement purposes.
When patients lost their retainers, new impressions
and retainers were provided. When there was appliance
breakage or loss, the patients were advised to attend
the daily emergency clinic, where a new appliance was
made; bonded retainers were repaired.
Outcomes (primary and secondary and any
changes after trial commencement)
All measurements were made manually with a digital
caliper (150 mm DIN 862, ABSOLUTE Digimatic caliper,
model 500-191U; Mitutoyo, Andover, Hampshire,
United Kingdom) with a resolution of 60.01 mm. The
arch was viewed from above, the caliper was held parallel
to the occlusal plane, and the distance between the con-
tact points of the caliper was recorded.
Irregularities of the mandibular incisors were
measured on the study models at T0, T1, T2, and T3 us-
ing the method described by Little55 (Fig 1).
Intra-arch measurements included intercanine width,
which was measured as the distance between the 2
canine cusp tips; in case of cuspal wear, an estimation
of the middle of the surface was made (Fig 2). Intermolar
O'Rourke et al
Fig 1. Little's irregularity index: sum of the distances
between the contact points of the mandibular canines.55
Fig 2. Intercanine and intermolar widths.
width was measured as the distance between the mesio-
buccal cusp tips of both mandibular first molars; simi-
larly, if the cusps tips were worn, an estimation of the
middle of the surface was made (Fig 2). Arch length
was measured at a point midway between the incisal
edges of the central incisors, bisecting the line connect-
ing the mesial marginal ridges of the left and right per-
manent molars (Fig 3). The extraction space opening was
measured as the sum of contact point displacements in
the anteroposterior plane where the extractions were
carried out.
All model measurements were collected by 1 investi-
gator (N.O'R.). Random errors and intraoperator reli-
ability were assessed by reproducing 2 sets of
measurements on 25 randomly selected study models
2 weeks later.56
No harm was observed as a result of either interven-
tion. Unlike previous data in the literature, there was a
low failure rate associated with the bonded retainer group
over 18 months, with only 3 subjects returning with 1 pad
that had debonded.38,40,43 This may have been because
American Journal of Orthodontics and Dentofacial Orthopedics
409
Fig 3. Arch depth.
most (39 of 42) of the bonded retainers were placed by
1 operator (N.O'R.) under a standardized protocol.
Sample size calculation
The main outcome measure in this study was change
in incisor irregularity in the groups after 6 months of
retention. Sample size estimation, using PASS 11,57
showed that a total of 72 participants, 36 in each treat-
ment arm, were required to demonstrate a clinically sig-
nificant difference in the primary outcome of 0.5 mm
between the groups, with 90% power, a standard devia-
tion of 0.79, and an alpha of 0.05, using the Mann-
Whitney U test, and with the null hypothesis of equal
effect. This calculation was based on a previous random-
ized controlled trial that identified a treatment effect of
0.5 mm at the 6-month follow-up.58 In addition, to
allow for a potential 15% dropout rate, the sample
size was increased to 82 subjects.
Interim analyses and stopping guidelines
No interim analysis was performed during the study.
During the protocol stage, we decided that the trial
would be stopped if it became obvious that 1 interven-
tion was clinically superior to the other or if any harm
came as a result of the interventions.
Randomization
An electronic randomization program was used,
whereby a subject allocation sequence was generated
using an online computer program (www.random.org).
The numbers were assigned by the computer to 2 groups
that corresponded to 1 of the 2 treatment options to be
studied. The random numbers were then placed in opa-
que envelopes that were identical, tamper-proof, and
prepared in advance. One envelope was then selected
randomly by the patient and opened. Forty subjects
September 2016
Vol 150
Issue 3
410
were randomized to the vacuum-formed retainer group
and 42 subjects to the bonded retainer group, and all
subjects received their allocated intervention.
Blinding
It was not possible to be blinded to the randomiza-
tion allocation after T0 because of the permanence of
a bonded retainer. However, measuring the models in
a random order and blinding during data analysis were
aimed to minimize bias.
Statistical analysis
Mann-Whitney U tests were performed to allow com-
parison of the vacuum-formed retainer and bonded
retainer groups in relation to the different variables at
all time periods. A P value of 0.05 was taken to be sta-
tistically significant. Statistical analyses were performed
on the available data set. No missing data imputation
was carried out.
Intraobserver reliability was assessed by performing
measurements for Little's irregulatiry index at a
2-week interval on 25 casts with intraclass correlation
coefficients, which confirmed excellent agreement
between the measurements, with 95% of the sample
demonstrating an intraclass correlation coefficient
between 0.92 and 0.98.56
RESULTS
Participant flow
The recruitment and follow-up of all patients is
shown in the CONSORT flow diagram (Fig 4). Eighty-
two subjects (mean age, 17.73 years; SD, 3.52 years)
were randomized to either vacuum-formed or bonded
retainers. At T1, 8 subjects failed to attend their appoint-
ments at T2; at T3, a further loss to follow-up was noted.
There was a substantial loss to follow-up overall, with
fewer subjects returning for their T3 follow-ups in the
vacuum-formed group (60%) than in the bonded
retainer group (71.4%).
Baseline data
Baseline demographics including age, sex, features of
the original malocclusion, and treatment plans were
similar in both groups. Table I shows the mean ages,
sex distributions, and clinical characteristics of the sub-
jects in their randomized groups.
The treatment details of each group are given in
Table II. In both groups, there were similar numbers of
extraction and nonextraction treatments. Second pre-
molars were overall the most popular choices for extrac-
tion. The amounts of crowding were similar in the
September 2016
Vol 150
Issue 3 American Journal of Orthodontics and Dentofacial Orthopedics
O'Rourke et al
groups (3.58 mm [SD, 3.3 mm] in bonded retainer group
and 3.69 mm [SD, 3.94 mm] in the vacuum-formed
retainer group).
Numbers analyzed for each outcome, estimation
and precision, subgroup analyses
For the vacuum-formed retainer group, data from
40, 37, 29, and 21 subjects were analyzed at T0, T1,
T2, and T3, respectively. For the bonded retainer group,
data from 42, 37, 30, and 27 subjects were analysed at
T0, T1, T2, and T3, respectively.
There was a statistically significant difference be-
tween the changes observed between the groups at
6 months after debond for Little's irregularity index
(P 5 0.008) (Table III).
The median changes in Little's irregularity index, in-
tercanine width, and arch length in the vacuum-formed
retainer group were greater in this time period than in
the bonded retainer group. The differences in the
amount of change observed in the latter 2 outcome
measures did not reach statistical significance. A greater
change in intermolar width was observed in the bonded
retainer group. There was no significant difference be-
tween the groups with respect to changes in extraction
site space opening during the first 6 months.
At 12 months, there was no statistically significant
difference in the amount of change between the groups
for any outcome. Extraction site space opening re-
mained minimal in both groups (Table III).
At 18 months, there was no statistically significant
difference in the amounts of change between the groups
for any outcome (Table III).
Harms
No significant harms were observed in either group
during the trial.
DISCUSSION
Main findings in the context of the existing
evidence, interpretation
This randomized clinical trial was designed to inves-
tigate the clinical effectiveness of 2 retainer types at
maintaining incisor regularity in the mandibular arch,
up to 6 months after debond. From an esthetic stand-
point, relapse of the anterior teeth weighs heavily in
any assessment of stability of the results because the pa-
tient tends to focus almost exclusively on alignment of
the incisors and canines. However, effectiveness could
not be based on this alone; therefore, additional
outcome measures were selected because they indicate
the stability of the treatment results and thus the
O'Rourke et al 411

Patients assessed for

eligibility

n=85
Excluded (n=3)

Not meeƟng inclusion

criteria (n=2)
Randomization
Refused to parƟcipate
(n=1)

Allocated to bonded retainer group Allocated to VFR group

n=42 Allocation n=40
Received allocated intervention Received allocated intervention
n=42 n=40

Analyzed at T0: debond Analyzed at T0: debond

n=42 Analysis n=40

Lost to follow-up due to Lost to follow-up due to

failure to attend T1 visit Follow up failure to attend T1 visit
n=5 n=3

Analyzed T0-T1 change: after 6 Analyzed T0-T1 change: after 6

months of retention Analysis months of retention
n=37 n=37

Lost to follow-up due to Lost to follow-up due to

Follow up failure to attend T2 visit
failure to attend T2 visit
n=8* n=9*

Analyzed T1-T2 change: after 12 Analyzed T1-T2 change: after 12

Analysis
months of retention months of retention
n=30 n=29

Lost to follow-up due to Follow up Lost to follow-up due to

failure to attend T3 visit failure to attend T3 visit
n=12† n=16†

Analyzed T2-T3 change, after 18 Analyzed T2-T3 change, after 18

months of retention Analysis
months of retention
n=27 n=21

Fig 4. CONSORT diagram of subjects' flow through the trial. VFR, Vacuum-formed retainer. *There
were 1 and 5 additional patients in the bonded and VFR groups, respectively, who attended their T2
follow-up visit. Since they were lost to follow-up at the T1 visit, their T1-T2 data were missing and
not analyzed. yThere were 3 additional patients in each group who attended their T3 follow-up visit.
Since they were lost to follow-up at the T2 visit, their T2-T3 data were missing and not analyzed.

American Journal of Orthodontics and Dentofacial Orthopedics September 2016
Vol 150
Issue 3
412 O'Rourke et al

Table I. Demographic and clinical characteristics of the sample at debond

Bonded retainer Vacuum-formed
Overall sample sample retainer sample
(n 5 82) (n 5 42) (n 5 40)
Age (y) 17.73 (3.52) 18.47 (4.41) 16.95 (2.02)
Sex (n)
Male 23 (28.1%) 9 (21.4%) 14 (35%)
Female 59 (71.9%) 33 (78.6%) 26 (65%)
Incisor classification59 (n)
Class I 11 (13.5%) 5 (11.9%) 6 (15%)
Class II Division 1 40 (48.8%) 20 (47.6%) 20 (50%)
Class II Division 2 10 (12.2%) 4 (9.5%) 6 (15%)
Class III 21 (25.5%) 13 (31%) 8 (20%)
Skeletal pattern (n)
Skeletal I 23 (28%) 11 (26.2%) 12 (30%)
Skeletal II 39 (47.6%) 17 (40.5%) 22 (55%)
Skeletal III 20 (24.4%) 14 (33.3%) 6 (15%)
Crowding/spacing (mm)
Spacing 10 (12.2%) 5 (11.8%) 5 (12.5%)
No crowding or spacing 2 (2.4%) 1 (2.4%) 1 (2.5%)
Mild 24 (29.3%) 13 (31%) 11 (27.5%)
Moderate 30 (36.6%) 16 (38.1%) 14 (35%)
Severe 16 (19.5%) 7 (16.7%) 9 (22.5%)
Amount of crowding (mm) 3.629 (3.62) 3.64 (3.25) 3.675 (3.8)
Values are mean (SD) or n (%).

because of 2 older subjects (28 years of age). There were

Table II. Treatment summary of the sample
more female subjects in the trial; this is a common
Vacuum- occurrence in orthodontic studies.26-39,41,47 This may
Bonded formed be because female patients are more self-aware and con-
Overall retainer retainer
sample sample sample cerned with dental health than are males58 and are
(n 5 82) (n 5 42) (n 5 40) perhaps more likely to seek treatment, but also because
Extraction summary female patients have a greater desire for orthodontic
Extraction 38 (46.3%) 20 (47.6%) 18 (45%) treatment than do males.60 There was also a higher pro-
Nonextraction 44 (53.7%) 22 (52.4%) 22 (55%)
portion of Class II Division 1 malocclusions, and this can
Extraction pattern
First premolars 14 (36.8%) 7 (35%) 7 (38.9%) be explained by the greater prevalence of this malocclu-
Second premolars 16 (42.1%) 8 (40%) 8 (44.4%) sion in society,61 as well as in a referred population
Asymmetric premolars 3 (7.9%) 2 (10%) 1 (5.6%) because this is a common malocclusion for which people
Other 5 (13.2%) 3 (15%) 2 (11.1%) seek treatment.62 The tooth choice for extraction
Values are mean (SD) or n (%). may reflect that mild and moderate crowding are the
most prevalent patterns of crowding, and both can be
effectiveness of a retention method. These measure- successfully relieved with extraction of the second
ments were intercanine width, intermolar width, arch premolars.
length, and extraction site space opening; these have Unlike the findings of previous studies, not all sub-
been routinely used in many previous studies on jects exhibited changes in Little's irregularity index
stability.5,10,23,30,37,43,45,47,50,51 over the 18 months after debond.39 However, there
One advantage of the randomization process is that it was a median increase in Little's irregularity index in
attempts to ensure that confounding variables such as both treatment groups over the first 6 months after de-
pretreatment malocclusion, irregularity, treatment bond, and the increase reported here was similar to that
choice, and mechanics are equally divided among the reported previously regardless of retention regimen and
groups so that the groups are equal in all respects except type.37,42,44-52 No patient had a Little's irregularity index
for intervention type. score after 18 months of more than 3.5 mm. A score less
In this study, the randomization worked well, and the than 3.5 mm has been deemed to be clinically acceptable
2 groups were well matched. The age range of the pa- in other studies.39,47,48,50-52 Our findings for the
tients in the bonded retention group was slightly higher vacuum-formed retainer group were similar to those of

September 2016
Vol 150
Issue 3 American Journal of Orthodontics and Dentofacial Orthopedics
O'Rourke et al 413

Table III. Changes at 6 (T1), 12 (T2), and 18 (T3) months in the groups
Bonded retainers Vacuum-formed Mann-Whitney
(n 5 42) retainers (n 5 40) P value
Little's irregularity index
Change, T0-T1 0.03 (0.00-0.07) 0.08 (0.01-0.31) 0.008
Change, T1-T2 0.03 (0.00-0.06) 0.05 (0.01-0.20) 0.195
Change, T2-T3 0.03 (0.00-0.10) 0.05 (0.02-0.18) 0.300
Intercanine width
Change, T0-T1 0.11 (0.04-0.39) 0.23 (0.10-0.41) 0.214
Change, T1-T2 0.17 (0.09-0.42) 0.20 (0.08-0.37) 0.720
Change, T2-T3 0.17 (0.10-0.32) 0.26 (0.14-0.33) 0.306
Intermolar width
Change, T0-T1 0.26 (0.10-0.54) 0.16 (0.06-0.33) 0.169
Change, T1-T2 0.38 (0.75-0.74) 0.25 (0.88-0.43) 0.565
Change, T2-T3 0.18 (0.70-0.41) 0.25 (0.13-0.41) 0.439
Arch length
Change, T0-T1 0.19 (0.05-0.49) 0.23 (0.06-0.68) 0.512
Change, T1-T2 0.20 (0.10-0.57) 0.19 (0.12-0.66) 0.515
Change, T2-T3 0.18 (0.07-0.56) 0.19 (0.08-0.32) 0.779
Extraction site space
Change, T0-T1 0.00 (0.00-0.00) 0.00 (0.00-0.00) 0.881
Change, T1-T2 0.00 (0.00-0.00) 0.00 (0.00-0.00) 0.799
Change, T2-T3 0.00 (0.00-0.00) 0.00 (0.00-0.02) 0.831
Values are n (interquartile).
T0, Debond.

Rowland et al37 and Shawesh et al.51 Our figures are
more favorable than those reported by Lindauer and
Shoff.49 Our findings for the bonded retainer group are
more favorable than those of Atack et al39 and Edman
Tynelius et al.47 The differences may be related to
different measuring techniques used in the latter studies.
Changes of similar magnitudes were documented in
longer retrospective studies.38,48
Our findings also suggest that for at least 6 months
after debond the bonded retainer is superior to a
vacuum-formed retainer in maintaining alignment of
the mandibular incisors. Changes in Little's irregularity
index over time may be related to remaining growth
but could also be due to failure to comply with the reten-
tion regimen, particularly because the vacuum-formed
retainer is removable. Other reasons may be inadequate
fit of the vacuum-formed retainer,63 failure of the
bonded retainer to remain in situ,46 and thus failure of
the retainer to prevent relapse.
There was no statistically significant difference in the
amount of change in intercanine width over the duration
of observation between the 2 groups, but this study
showed that the bonded retainer has less change associ-
ated with it during the first 12 months. The size of the
change was small, and it is not likely to be noted clini-
cally. This is similar to the findings of Renkema et al,38
who reported that intercanine width was well main-
tained with bonded retainers. Edman Tynelius et al48
also reported a minimal change in intercanine width in
the bonded retainer group. In keeping with the findings
of previous studies, the intercanine width in the
vacuum-formed retainer group showed a small increase
over the study period.38,47,48 In contrast to this, Thickett
and Power50 reported that intercanine width was well
maintained by a vacuum-formed retainer for 6 months
and 1 year.
In the first 12 months posttreatment, there was as
slightly greater increase in intermolar width associated
with the bonded retainer group. This may have been
because the retainer does not extend to the molar re-
gion; thus, stability in this area entirely depends on
excellent interdigitation of the buccal segments. The
changes were slight in both groups, and we concluded
that the intermolar width was also generally well main-
tained in both retainer types; this agrees with the find-
ings of other similar studies.39,47,48
Arch length changes were equally small. Although
the change was greater in the vacuum-formed retainer
group, it was not statistically significant; an explanation
for the difference could be related to adherence to the
recommended retention regimen.
Extraction site space opening has not been well re-
ported in the literature; therefore, direct comparisons
with our findings are not possible. Extraction site re-
opening was seen in a few patients in both groups (7
in the bonded retainer group; 8 in the vadcuum-
formed retainer group), and the increases in space open-
ing were similar for both groups.

American Journal of Orthodontics and Dentofacial Orthopedics September 2016
Vol 150
Issue 3
414
In the bonded retainer group, in 1 subject, when
space opened by 3.42 mm, this was likely to be clinically
significant. One would expect that good buccal interdig-
itation would enhance the treatment stability27,28;
however, because the periodontal ligament can take
up to 232 days to rearrange, it is likely that because
there was no retainer posterior to the canines to help
ensure that the periodontal fibers reorganized in the
approximated tooth positions, the result was
extraction-site reopening.31 In the vacuum-formed
group, the only true explanation for extraction site space
opening was the lack of compliance with retainer wear.
Limitations
There is a difficulty in directly comparing results of this
study with others because, although similar research ques-
tions are asked in other studies,37,39 they are not all similar
in their methods,37,44 outcome measures,42-46,50,51
measurement techniques,37,39,41 and durations of
recall.5,23,38,39 In addition to these issues, other authors
did not publish their full results or displayed them
differently.41,45,47,48,54
Adherence with retainer wear was not assessed in this
study; therefore, it cannot be confirmed that some of the
observed relapse might be the result of nonadherence to
the retention regimen rather than a failure in the retainer
itself. Previous studies that focused on adherence with
different regimens in orthodontics showed various out-
comes and highlight the difficulties associated with
recording compliance.64,65 In 1 study involving timers
to measure wear of an appliance, the patients adhered
to their requested regimen for slightly over 50% of the
expected duration of wear.65 Attempts have been
made to record compliance with retainer wear, but this
is usually a subjective assessment of the condition of
the retainer at the review period or a self-assessment
by the patients, both which can be inaccurate because
of the Hawthorne effect, or patient or operator judgment
bias.37,47,48,64
Generalizability
The generalizability of these results might be limited
because this research was undertaken in a single-center
hospital mostly by 1 clinician (N.O'R.). The methods and
materials are described in detail to aid in similar studies
in the future.
CONCLUSIONS
There was a significant difference in incisor irregular-
ity changes at 6 months after debond between the
groups. Thus, the null hypothesis can be rejected.
Bonded retainers are more effective in their ability to
September 2016
Vol 150
Issue 3 American Journal of Orthodontics and Dentofacial Orthopedics
O'Rourke et al
maintain incisor alignment in the mandibular arch in
the first 6 months after debond of fixed appliances
when compared with vacuum-formed retainers.
All changes during the first 18 months after debond
were minimal.
REFERENCES
1. Angle EH. Malocclusion of the teeth. 7th ed. Philadelphia: SS
White Dental Manufacturing; 1907.
2. Riedel RA. A review of the retention problem. Angle Orthod 1960;
30:179-99.
3. Horowitz SL, Hixon EH. Physiologic recovery following orthodon-
tic treatment. Am J Orthod 1969;58:1-4.
4. Gardner SD, Chaconas SJ. Posttreatment and postretention
changes following orthodontic therapy. Angle Orthod 1976;46:
151-61.
5. Little RM, Wallen TR, Riedel RA. Stability and relapse of mandibular
anterior alignment—first premolar extraction cases treated by
traditional edgewise orthodontics. Am J Orthod 1981;80:349-65.
6. Sadowsky C, Sakols EI. Long-term assessment of orthodontic
relapse. Am J Orthod 1982;82:456-63.
7. Sinclair P, Little R. Maturation of untreated normal occlusions. Am
J Orthod 1983;83:114-23.
8. Uhde MD, Sadowsky C, BeGole EA. Long-term stability of dental
relationships after orthodontic treatment. Angle Orthod 1983;
53:240-52.
9. Gilmore CA, Little RM. Mandibular incisor dimensions and crowd-
ing. Am J Orthod 1984;86:493-502.
10. Little RM, Riedel RA, Artun J. An evaluation of changes in mandib-
ular anterior alignment from 10 to 20 years postretention. Am J
Orthod Dentofacial Orthop 1988;93:423-8.
11. Behrents RG. Growth in the aging craniofacial skeleton. Craniofa-
cial Growth Series. Ann Arbor: Center for Human Growth and
Development; University of Michigan; 1985. p. 1-145.
12. Mamandras AH. Linear changes of the maxillary and mandibular
lips. Am J Orthod Dentofacial Orthop 1998;94:405-10.
13. Nanda RS, Nanda SK. Considerations of dentofacial growth in
long-term retention and stability: is active retention needed? Am
J Orthod Dentofacial Orthop 1992;101:297-303.
14. Richardson ME. Late lower arch crowding in relation to primary
crowding. Angle Orthod 1982;44:56-61.
15. Richardson ME. Late lower arch crowding in relation to the direc-
tion of eruption. Eur J Orthod 1996;18:341-7.
16. Peck S, Peck H. Crown dimensions and mandibular incisor align-
ment. Angle Orthod 1972;42:148-53.
17. De La Cruz AR, Sampson P, Little RM,  Artun J, Shapiro PA. Long-
term changes in arch form after orthodontic treatment and reten-
tion. Am J Orthod Dentofacial Orthop 1995;107:518-30.
18. Bondemark L, Holm AK, Hansen K, Axelsson S, Mohlin B,
Brattstrom V, et al. Long term stability of orthodontic treatment
and patient satisfaction: a systematic review. Angle Orthod
2007;77:181-91.
19. Burke SP, Silveira AM, Goldsmith LJ, Yancey JM, Van Stewart A,
Scarfe WC. A meta-analysis of mandibular intercanine width in
treatment and postretention. Angle Orthod 1998;68:53-60.
20. McReynolds DC, Little RM. Mandibular second premolar extrac-
tion—postretention evaluation of stability and relapse. Angle Or-
thod 1991;61:133-44.
21. Haruki T, Little RM. Early versus late treatment of crowded first
premolar extraction cases: postretention evaluation of stability
and relapse. Angle Orthod 1988;68:61-8.
O'Rourke et al
22. Riedel RA, Little RM, Bui TD. Mandibular incisor extraction—post-
retention evaluation of stability and relapse. Angle Orthod 1992;
62:103-16.
23. Shields TE, Little RM, Chapko MK. Stability and relapse of mandib-
ular anterior alignment: a cephalometric appraisal of first premolar
extraction cases treated by traditional edgewise orthodontics. Am J
Orthod 1985;87:27-38.
24. Mills JR. The stability of the lower labial segment. A cephalometric
survey. Dent Pract Dent Rec 1986;18:293-306.
25. Paquette DE, Beattie JR, Johnston LE Jr. A long-term comparison
of nonextraction and premolar extraction edgewise therapy in
“borderline” Class II patients. Am J Orthod Dentofacial Orthop
1992;102:1-14.
26. Erdinc AE, Nanda RS, Isiksal E. Relapse of anterior crowding in pa-
tients treated with extraction and nonextraction of premolars. Am
J Orthod Dentofacial Orthop 2006;129:775-84.
27. Andrews LF. The six keys to normal occlusion. Am J Orthod 1972;
72:296-309.
28. Harris EF, Behrents RG. The intrinsic stability of Class I molar rela-
tionship: a longitudinal study of untreated cases. Am J Orthod
Dentofacial Orthop 1988;94:63-7.
29. Weiland F. The role of occlusal discrepancies in the long-term sta-
bility of the mandibular arch. Eur J Orthod 1994;16:521-9.
30. Little RM, Riedel RA. Postretention evaluation of stability and
relapse—mandibular arches with generalized spacing. Am J Orthod
Dentofacial Orthop 1989;95:37-41.
31. Reitan K. Clinical and histological observations on tooth move-
ment during and after orthodontic treatment. Am J Orthod
1967;53:721-45.
32. Singh P, Grammati S, Kirschen R. Orthodontic retention patterns in
the United Kingdom. J Orthod 2009;36:115-21.
33. Valiathan M, Hughes E. Results of a survey-based study to identify
common retention practices in the United States. Am J Orthod
Dentofacial Orthop 2010;137:170-7.
34. Hichens L, Rowland H, Williams A, Hollinghurst S, Ewings P,
Clark S, et al. Cost-effectiveness and patient satisfaction: Hawley
and vacuum-formed retainers. Eur J Orthod 2007;29:372-8.
35. Cerny R, Cockrell D, Lloyd D. A survey of patient opinions on fixed
vs. removable retainers. J Clin Orthod 2009;43:784-7.
36. Meade M, Millett D, Cronin M. Social perceptions of orthodontic
retainer wear. Eur J Orthod 2014;36:649-56.
37. Rowland H, Hichens L, Williams A, Hills D, Killingback N, Ewings P,
et al. The effectiveness of Hawley and vacuum-formed retainers: a
single-center randomized controlled trial. Am J Orthod Dentofa-
cial Orthop 2007;132:730-7.
38. Renkema AM, Renkema A, Bronkhorst E, Katsaros C. Long-term
effectiveness of canine-to-canine bonded flexible spiral wire lingual
retainers. Am J Orthod Dentofacial Orthop 2011;139:614-21.
39. Atack N, Harradine N, Sandy J, Ireland A. Which way forward?
Fixed or removable lower retainers. Angle Orthod 2007;77:
594-9.
40. Zachrisson BU. Long-term experience with direct-bonded re-
tainers: update and clinical advice. J Clin Orthod 2007;41:728-37.
41. Barlin S, Smith R, Reed R, Sandy J, Ireland AJ. A retrospective
randomized double-blind comparison study of the effectiveness
of Hawley vs vacuum-formed retainers. Angle Orthod 2011;81:
404-9.
42. Rose E, Frucht S, Jonas IE. Clinical comparison of a multistranded
wire and a direct-bonded polyethylene ribbon-reinforced resin
composite used for lingual retention. Quintessence Int 2002;33:
579-83.
American Journal of Orthodontics and Dentofacial Orthopedics
415
43. 
Artun J, Spadafora AT, Shapiro PA. A 3-year follow-up study of
various types of orthodontic canine to canine retainers. Eur J Or-
thod 1997;19:501-9.
44. Stormann I, Ehmer U. A prospective randomized study of different
retainer types. J Orofac Orthop 2002;63:42-50.
45. Tacken MP, Cosyn J, De Wilde P, Aerts J, Govaerts E, Vannet BV.
Glass fibre reinforced versus multistranded bonded orthodontic re-
tainers: a 2 year prospective multi-centre study. Eur J Orthod
2010;32:117-23.
46. Taner T, Aksu M. A prospective clinical evaluation of mandibular
lingual retainer survival. Eur J Orthod 2011;34:470-4.
47. Edman Tynelius G, Bondemark L, Lilja-Karlander E. Evaluation of
orthodontic treatment after 1 year of retention—a randomized
controlled trial. Eur J Orthod 2010;32:542-7.
48. Edman Tynelius G, Bondemark L, Lilja-Karlander E. A randomized
controlled trial of three orthodontic retention methods in Class I
four premolar extraction cases—stability after 2 years in retention.
Orthod Craniofac Res 2013;16:105-15.
49. Lindauer SJ, Shoff RC. Comparison of Essix and Hawley retainers. J
Clin Orthod 1998;32:95-7.
50. Thickett E, Power S. A randomized clinical trial of thermoplastic
retainer wear. Eur J Orthod 2010;32:1-5.
51. Shawesh M, Bhatti B, Usmani T, Mandall N. Hawley retainers full- or
part-time? A randomized clinical trial. Eur J Orthod 2010;32:165-70.
52. Gill DS, Naini FB, Jones A, Tredwin CJ. Part-time versus full-time
retainer wear following fixed appliance therapy: a randomized
prospective controlled trial. World J Orthod 2006;8:300-6.
53. Littlewood SJ, Millett DT, Doubleday B, Bearn DR,
Worthington HV. Retention procedures for stabilising tooth posi-
tion after treatment with orthodontic braces. Cochrane Database
Syst Rev 2006;(1):CD002283.
54. Yu Y, Sun J, Lai W, Wu T, Koshy S, Shi Z. Interventions for manag-
ing relapse of the lower front teeth after orthodontic treatment.
Cochrane Database Syst Rev 2013;(9):CD008734.
55. Little RM. The irregularity index: a quantitative score of mandib-
ular anterior alignment. Am J Orthod 1975;68:554-63.
56. Houston WJ. The analysis of errors in orthodontic measurements.
Am J Orthod 1983;83:382-90.
57. Hintze J. PASS 11. Kaysville, Utah: NCSS; 2011. Available at: www.
ncss.com. Accessed May 1, 2015.
58. Shaw WC. The influence of children's dentofacial appearance on
their social attractiveness as judged by peers and lay adults. Am
J Orthod 1981;79:399-415.
59. Glossary of dental terms. BS 4492:1983. London, United
Kingdom: BSI; 1983.
60. Harrison JE, O'Brien KD, Worthington HV. Orthodontic treatment
for prominent upper front teeth in children. Cochrane Database
Syst Rev 2007;(3):CD003452.
61. Todd JE, Lader D. Adult dental health: 1988 United Kingdom. Lon-
don, United Kingdom: Office for National Statistics; 1988.
62. Kilpenlainen PV, Philips C, Tulloch JF. Anterior tooth position and
motivation for early treatment. Angle Orthod 1993;63:171-4.
63. Johal A, Sharma NR, McLaughlin K, Zou LF. The reliability of ther-
moform retainers: a laboratory-based comparative study. Eur J Or-
thod 2015;37:503-7.
64. Pratt MC, Kluemper GT, Lindstrom AF. Patient compliance with
orthodontic retainers in the postretention phase. Am J Orthod
Dentofacial Orthop 2011;140:196-201.
65. Bartsch A, Witt E, Sahm G, Schneider S. Correlates of objective pa-
tient compliance with removable appliance wear. Am J Orthod
Dentofacial Orthop 1993;104:378-86.
September 2016
Vol 150
Issue 3