European Journal of Orthodontics, 2022, 210–225

doi:10.1093/ejo/cjab047
Advance Access publication 5 October 2021

Systematic Review

Orthodontic space closure in sliding mechanics:

a systematic review and meta-analysis
Biju Sebastian1, , Aarthi Bhuvaraghan2,3 and Badri Thiruvenkatachari3,

Downloaded from https://academic.oup.com/ejo/article/44/2/210/6381541 by guest on 26 July 2023

1
Saveetha Dental College and Hospitals, 162, Poonamallee High Rd, Velappanchavadi, Chennai, Tamil Nadu 600077,
India, 2Leeds Dental Institute, Worsley Building, The Clarendon Way, University of Leeds, Leeds LS2 9LU, UK, 3Sree
Balaji Dental College and Hospital, Bharath Institute of Higher Education and Management, Pallikaranai, Chennai,
Tamil Nadu 600100, India

Correspondence to: Badri Thiruvenkatachari, Sree Balaji Dental College and Hospital, Velachery Main Road, Pallikaranai,
Chennai Tamil Nadu 600100, India. E-mail: [email protected]

Summary
Background/Objectives:  The aim of this systematic review was to compare the different force
delivery systems for orthodontic space closure by sliding mechanics.
Search methods:  Multiple sources including Cochrane Central, Ovid Medline, Embase etc. were
used to identify all relevant studies.
Selection criteria:  Randomized controlled trials (RCT) of parallel-group and split-mouth designs
were included.
Participants:  Orthodontic patients of any age treated with fixed orthodontic appliances and
requiring space closure.
Data collection and analysis:  Search result screening, data extraction and quality assessment were
performed independently and in duplicate by 2 reviewers. The included studies were grouped
into parallel-arm and split-mouth studies and subgroup analysis was then performed for the type
of retraction subsets; en-masse and individual canine retraction. A traditional meta-analysis, and
network meta-analysis (NMA) for direct and indirect comparisons for the rate of space closure
were performed.
Results:  Thirteen studies, six parallel-arm and seven split-mouth were included. The traditional
meta-analysis comparing Nickel-titanium (NiTi) closed coil springs and elastomeric power chain
for the rate of tooth movement showed statistically significant difference favouring NiTi springs
(MD: 0.24; 95% CI, 0.03–0.45; I2 0%, P = 0.02) and the comparison between NiTi springs and active
ligatures also showed statistically significant result favouring NiTi springs (MD: 0.53; 95% CI, 0.44–
0.63; I2 0%, P ˂ 0.00001) for the rate of tooth movement. NMA for the rate of space closure showed
fairly confident evidence for NiTi coil springs when compared with elastomeric chain and active
ligatures. The NiTi coil spring ranked best between all methods of space closure.
Conclusions:  There is moderate quality evidence in favour of NiTi coil springs for the rate of
space closure when compared with active ligature and low quality of evidence favouring NiTi
springs when compared with elastomeric chain. The ranking from NMA showed NiTi coil springs
to be the best method for space closure with 99% chance. An urgent need for standardization of
study designs and the need for development of an agreed core outcome sets and core outcome
instrument measurement sets is evident.
Registration:  PROSPERO CRD42020157811

© The Author(s) 2021. Published by Oxford University Press on behalf of the European Orthodontic Society.
210
All rights reserved. For permissions, please email: [email protected]
B. Sebastian et al. 211

Introduction Protocol and registration

In modern day Orthodontics, there has been an increased interest in The protocol for this systematic review was registered on the inter-
providing our patients with an efficient and comfortable treatment in national prospective register of systematic reviews (PROSPERO)
the shortest time possible. There are several factors that can influence from the National Institute for Health Research database with the
the duration of orthodontic treatment including the treatment plan de- number CRD42020157811 (www.crd.york.ac.uk/PROSPERO).
vised, type of appliance, treatment mechanics, force levels and patient-
related factors (1–3). Reducing the duration of treatment will have Eligibility criteria
benefits for both patients and clinicians (4), in addition to reducing the The selection criteria for this review (PICO) were the following:
iatrogenic risks like root resorption and decalcification (5–8).
The mechanics used for space closure has been shown to sig- Study design
nificantly influence the duration of orthodontic treatment (9). Since Randomized Controlled Trials with parallel-group and split-mouth
the introduction of preadjusted edgewise appliance, sliding mech- designs.
anics has been the predominant method of space closure (10, 11).
However, space closure with sliding mechanics has drawbacks like Participants
friction, binding and notching (12). In an attempt to overcome these Orthodontic patients of any age treated with fixed orthodontic ap-

Downloaded from https://academic.oup.com/ejo/article/44/2/210/6381541 by guest on 26 July 2023

drawbacks, various force delivery options have been attempted. pliances and requiring space closure.
These include Nickel-titanium (NiTi) closed coil springs, elastomeric
power chain, stainless steel coil springs, activate ligatures (module
Intervention
tied with a ligature and activated), elastics and elastic threads.
Orthodontic space closure with elastomeric power chain, NiTi coil
Unfortunately, there have been limited attempts to compare these
springs or any other method.
methods and hence the choice of using these systems depends on the
clinician’s preference and not on sound evidence.
With the use of several different force delivery systems in clin- Comparison
ical orthodontics, a traditional meta-analysis only allows multiple Orthodontic space closure with NiTi coil springs, stainless steel
pair-wise comparisons to test the interventions. This produces sev- coil springs, activate ligatures, elastics, elastic thread or any other
eral parts of evidence for the same type of tooth movement evaluated method or untreated control group.
and therefore, fails to give us a summary of the whole picture. A net-
work meta-analysis (NMA) has the ability to quantitatively compare Primary outcome
interventions that are not directly compared in clinical trials. This Rate of tooth movement (measured in mm/month), measured either
direct comparison on a common comparator (intervention or con- from study models, radiographs or clinically.
trol) not only provides us with more usable information but also
supports adjustment for bias that would otherwise exist if the trials Secondary outcomes
were compared directly. Additionally, analysing direct and indirect (1) Force level measured in grams, (2) Anchorage loss, measured
evidence strengthens the evidence base. More importantly, the NMA in mm, (3) Pain perception, measured using VAS pain scale, (4)
gives us a more concise assessment of the entire evidence base within Periodontal effects, measured by pocket depth in mm and bleeding
a single analysis and allows the interventions to be ranked in order on probing (dichotomous, yes/no), (5) Cost measured in dollars, (6)
of effectiveness. This can make complicated systematic reviews much Tipping of teeth, measured in degrees, (7) Root resorption, measured
easier to interpret for clinicians and academic readers. in mm and percentage, (8) Harms, descriptively reported.
The previous systematic review involving traditional meta-analy-
sis assessed the effectiveness of NiTi closed coil springs versus elasto- Information sources, search strategy, and study
meric power chain. They included only four studies; two split-mouth selection
design and two parallel-group randomized controlled trials (RCT)
Several sources were used to identify all relevant studies from 1946
(13). The overall quality was poor with only one of the four RCTs
to January 28, 2020, independent of language restriction and date
judged as low risk of bias. The review did not assess any other meth-
of publication. The Cochrane Central Register of Controlled Trials
ods of space closure or carry out indirect comparison to strengthen
(CENTRAL), Cochrane Database of Systematic Reviews, LILACS,
evidence. This is the first network meta-analysis to evaluate space
Scopus, EMBASE, MEDLINE (Ovid), Web of Science and EPUB
closure methods in orthodontics. Additionally, there have been many
ahead of publications and non-indexed citations were searched.
relevant clinical trials published in the last couple of years (14–18).
PubMed, Web of Science and Google Scholar were also searched
Hence, in this systematic review we aim to compare the different
for completeness. In addition, Dissertation data (www.theses.com),
force delivery systems for space closure during orthodontic treat-
Grey literature (www.opengrey.eu), ClinicalTrials.gov (http://www.
ment in terms of different outcomes: rate of tooth movement, force
ClinicalTrials.gov), and ISRCTN registry (http://www.controlled-
level, effect on anchorage, periodontal effects, cost efficiency and
trials.com) were searched for ongoing and unpublished studies. The
patient-centred outcomes, using both traditional and network meta-
search strategy for different databases is shown in Supplementary
analysis designs.
Table S1a–c. Hand searching was carried out for selected journals:
American Journal of Orthodontics and Dentofacial Orthopedics,
Angle Orthodontist, European Journal of Orthodontics,
Material and methods Orthodontics and Craniofacial Research, Journal of Orthodontics,
This review was carried out according to the Preferred Reporting and British Journal of Orthodontics. An attempt was made to iden-
Items for Systematic reviews and Meta-analyses (PRISMA) guide- tify unpublished studies or ongoing clinical trials by contacting the
lines (19). first-named author of the trial reports. Reference lists of identified
212 European Journal of Orthodontics, 2022, Vol. 44, No. 2

articles and relevant review articles were checked for further pos- and a value of more than 50% was assessed as having significant
sible studies. heterogeneity.
The title and abstract of potentially relevant articles were As not all studies evaluated all systems, a second set of analyses
screened independently and in duplicate by two reviewers (B.S. and used a network meta-analysis approach to make indirect, as well as
B.T.) who were not blinded to the authors or the results of the re- direct, comparisons between studies. A  frequentist approach using
search. All disagreements were resolved after discussion with the the methods described by White et al. (21) was utilised. Specifically,
third reviewer (A.B.). Full-text articles were examined for eligibility the general approach used a model for treatment contrasts (the “con-
independently by two reviewers (B.S. and B.T.) and any discrepan- trast-based” model of Salanti et al. (22)), which considers treatment
cies were resolved after discussion with the third reviewer (A.B.). effects as fixed effects, and heterogeneity between studies as random
effects.
Data items and collection The quality of the resultant evidence was ranked using the
‘Grading of Recommendations, Assessment, Development and
Data extraction for the included studies was carried out independ-
Evaluation’ (GRADE) approach (23).
ently and in duplicate by the two reviewers (B.S. and B.T.) using a
pre-piloted data extraction form. Discrepancies in the extracted data
were resolved by discussion with the third reviewer (A.B.). The data Publication bias

Downloaded from https://academic.oup.com/ejo/article/44/2/210/6381541 by guest on 26 July 2023

extraction form included (1) general information like extractor de- Standard funnel plots and contoured enhanced funnel plots would
tails, time etc., (2) study information like title, year, authors etc., (3) be drawn to identify publication bias if more than 10 studies were
study eligibility including PICO assessment, (4) risk of bias, (5) char- included in the meta-analysis (24).
acteristics of included studies like details of the trial, participants,
interventions etc., and (6) outcomes like rate of tooth movement, Additional analyses
force level, anchorage loss, periodontal effects, cost, tipping of teeth, Sensitivity analyses were planned to address studies at higher risk of
root resorption, pain and harms. bias, publication bias, and other potential sources of heterogeneity
including overriding effects of at least one large study and differ-
Risk of bias/quality assessment in individual studies ences in outcome related to specific force delivery system to isolate
The risk of bias within parallel-group and split-mouth RCTs was their influence on the overall outcome.
evaluated using Cochrane Collaboration’s tool for assessing the
risk of bias, as described in the Cochrane Handbook for Systematic Results
Reviews of Interventions (20). Two independent reviewers (B.S. and
Study selection and characteristics
B.T.) assessed the risk of bias separately. Any potential disagree-
The search strategy identified 175 studies and after removal of du-
ments between the two reviewers were resolved by discussion with
plicates, screening titles and abstracts for inclusion, a total of 25
the third reviewer (A.B.). The overall risk of bias for each study was
potentially relevant full-text articles (13–18, 25–43) were retrieved
summarized as low, unclear, or high.
for further evaluation (Figure 1). The full text of the retrieved stud-
ies was assessed for inclusion using our inclusion criteria. A further
Summary measures, approach to synthesis and twelve studies were excluded due to specific reasons as explained in
analysis Supplementary Table S2 (13, 25–35). A total of 13 randomised con-
The collected data was quantitatively assessed and analysed using trolled trials, six parallel-group studies (14–16, 36–38) and seven
Review Manager software (RevMan version 5.3; The Nordic split-mouth studies (17, 18, 39–43), were included for the final ana-
Cochrane Centre, The Cochrane Collaboration; Copenhagen, lysis. The corresponding authors were contacted for clarification
Denmark). Mean differences with 95% confidence intervals (CIs) of data in five studies (15, 16, 18, 37, 43) and the response was
were calculated for continuous outcomes like rate of tooth move- obtained from three authors (15, 18, 37)
ment, force level, anchorage loss, pain perception, cost, tipping of Of the six parallel-group studies, four performed en-masse re-
teeth and root resorption. In case of dichotomous outcomes, such as traction (14, 36–38) and two performed individual canine retrac-
adverse periodontal effects and harms, risk ratios with 95% confi- tion (15, 16) as the method of space closure. Correspondingly, there
dence intervals were shown. were three en-masse retractions (39–41) and four individual canine
Data collection was completed from the eligible studies. For all retractions (17, 18, 42, 43) in the split-mouth design group. A sum-
missing data, an attempt was made to contact the corresponding mary version of the characteristics of the included studies is shown
author by email with a request to provide further information. A re- in Tables 1 and 2, and a full version is given in Supplementary Tables
minder email was sent two weeks after in case of no response. Clinical S3 and S4.
heterogeneity of the included studies was gauged by reviewing the
study protocol and methodology, treatment interventions, timing of Risk of bias in studies
data collection, measurement techniques and measured outcomes. The methodological quality assessments of the RCTs are given in
A traditional meta-analysis was carried out for trials using Figures 2 and 3.
similar outcome and similar assessment tools. To facilitate the in-
terpretation of the data synthesis, the included studies were grouped Selection bias
into parallel-arm and split-mouth studies. Subgroup analysis was Of the six parallel-group studies (14–16, 36–38), randomization and
then performed for the type of retraction subsets, en-masse and indi- allocation concealment were considered adequate for four studies
vidual canine retraction. A fixed-effects model for analysis with two (15, 36–38). For Fang et al. (14) randomisation was judged as ad-
studies and random-effects model for analysis with more than two equate and allocation concealment as unclear. Davidović et al. (16)
studies was carried out for the pooled data accounting for possible was judged unclear on randomization and high risk on allocation
heterogeneity. Statistical heterogeneity was measured using the I2 test concealment.
B. Sebastian et al. 213

!
168 records identified through 5 records identified through 2 additional records identified from what was initially planned in Nightingale et al. (41) and hence
electronic database searching cross-referencing! via hand-searching
both were judged as high risk. Gender percentage and baseline char-
acteristics were not reported in two studies and these were judged as
having unclear risk (16, 42). Only three studies were judged to be at
low risk (14, 18, 43).

103 after removal of duplicates

Result summary of individual studies, traditional
meta-analysis, and sensitivity analysis
103 screened
A summary version of the summary of the findings in individual
78 excluded
studies is reported in Table 3 and a full version in Supplementary
Table S5. Three of the thirteen studies did not provide standard de-
25 full-text articles assessed for
eligibility viation, standard error, confidence interval or range (37, 39, 40). For
12 full-text articles excluded these studies, SD was imputed from studies carried out in similar
with reasons
13 included studies: 6 parallel-
conditions, with similar methodology and sample size (44). Bokas
group (4 en-masse + 2 et al. (43) provided a range for space closure and therefore, it was

Downloaded from https://academic.oup.com/ejo/article/44/2/210/6381541 by guest on 26 July 2023

individual canine retraction) and
7 split-mouth trials (3 en-masse possible to calculate SD (45).
+ 4 individual canine retraction)
Analyses for split-mouth studies and parallel-group studies were
separately carried out. In addition, analyses of en-masse and canine
Figure 1.  Preferred reporting items for systematic reviews and meta-analyses
retraction groups were individually assessed. Comparisons regarding
(PRISMA) flowchart.
the rate of space closure were made between the following interven-
tions within the subsets:
In the split-mouth design group (17, 18, 39–43), randomization
was judged to be adequate for three studies (17, 41, 42) and un- Parallel-group studies with en-masse space closure
clear for the rest four (18, 39, 40, 43). All split-mouth studies, except
Elastomeric power chain versus NiTi closed coil springs
Khalid et al. (18) were judged as high risk on allocation concealment.
Two studies (36, 37) looked at comparisons between elastomeric
power chain and NiTi springs. The results were statistically signifi-
Performance and detection bias cant, with a mean difference of 0.24 mm (95% CI, 0.03–0.45; I2 0%,
Blinding of the clinicians and patients to intervention was not per- P = 0.02) (Figure 4a) favouring NiTi closed coil springs.
formed because of the nature of clinical trials. Blinding of the out-
come assessors was clearly carried out in all parallel-group studies, NiTi closed coil springs versus Active ligatures
except Davidović et al. (16) and hence this study was judged as high Two studies (14, 36) were included in the meta-analysis and the re-
risk of bias and the rest were categorised as low risk (14, 15, 36–38). sults showed a statistically significant difference with a mean differ-
However, the outcome assessor blinding was not undertaken ence of 0.53 mm (95% CI, 0.44–0.63; I2 0%, P < 0.00001) (Figure
in three split-mouth trials (39, 40, 42) which were judged as high 4b) in favour of NiTi closed coil springs.
risk of bias. Khanemasjedi et al. (17) lacked clarity on this and was
judged as unclear while the remaining were judged as having a low NiTi closed coil springs versus other space closure methods
risk of bias (18, 41, 43). The effectiveness of the NiTi closed coil springs with other space
closure methods were compared using the pooled data derived from
Attrition bias the four included studies (14, 36–38) involving 129 participants. The
Among the parallel-group studies, four trials were judged as having results showed no statistically significant difference (MD: 0.25; 95%
low risk of bias with two (15, 37) reporting no drop outs while the CI, −0.16 to 0.65; I2 93%, P = 0.23) between NiTi closed coil springs
other two (14, 16) had acceptable percentage of dropouts (<20%). and other methods like elastomeric power chain, active ligatures and
Norman et al. (38) was judged as having high risk of bias due to high stainless steel coil springs (Figure 4c). However, the heterogeneity
withdrawal rates and Dixon et al. (36) did not report the reason for was at an unacceptable level (I2 93%; chi-square 45.16, P ˂ 0.00001)
drop outs and hence was judged as unclear risk. (Figure 4c) and so the results have to be taken with caution.
There were no drop outs in four split-mouth trials and they all
were judged as having low risk of bias (18, 39, 40, 43). The with- Sensitivity analysis
drawal rates were high in two studies and hence judged as having A further meta-analysis was undertaken to gauge the inclusion of
high risk of bias (17, 41). Sonis et al. (42) did not provide any infor- studies on the outcome and statistical heterogeneity. Sensitivity
mation on withdrawals and so was judged as having unclear risk. analysis investigating the influence of Norman et al. (38) resulted
in estimates favouring NiTi closed coil springs (MD: 0.44; 95% CI,
Selective reporting 0.26–0.63; I2 53%, P ˂ 0.00001) against other methods of space
All the important outcomes were reported in the included trials (14– closure (Figure 4d). Similarly, the exclusion of Chaudhari et  al.
18, 36–43). Hence, risk of reporting bias in the studies was judged (37) to assess the impact of age variance of participants between
to be low. the studies reported a comparable finding to the original synthesis
(MD: 0.24; 95% CI, −0.30–0.79; I2 95%, P = 0.38) (Figure 4e).

Other bias
Eight studies (15, 17, 36–41) were judged as having a high risk of Split-mouth studies with en-masse space closure
bias with gender bias being a major problem with six of the studies NiTi closed coil springs versus other methods
(15, 17, 36, 37, 39, 40). The random code was broken before finish Meta-analysis performed on this subset using pooled data from
in Norman et  al. (38) and there was a serious protocol deviation two studies (39, 41) showed no statistically significant difference
Table 1.  Summary of Characteristics of studies (Parallel-group RCTs) (Summary version. The full (long) version is shown in Supplementary Table S3).
214

Study Characteristics

a) En-masse retraction
  Dixon 2002
  Methods Location: University Dental Hospital of Manchester, United Kingdom
Maximum duration of study: 4 months
Trial design: Parallel-group randomized controlled trial
  Participants Age: 15.6 ± 2.5 years (For 33 patients analysed)
Gender: 10 males and 23 females (for 33 patients analysed)
Number randomized: 36 patients (12/group)
Number evaluated: 33 patients (Group I-10; Group II-11; Group III-12)
  Interventions Group I: Power chain
Group II: NiTi coil spring
Group III: Active ligature
  Outcomes Primary: Rate of tooth movement (mm/month), measured from study models
  Chaudhari 2015
  Methods Location: CSMSS Dental College & Hospital, Aurangabad, Maharashtra, India
Maximum duration of study: 4 months
Trial design: Parallel-group randomized controlled trial
  Participants Age: Group 1: 19.8 ± 2.6 years; Group 2: 19.45 ± 3 years; Total: 19.62 ± 2.81
Gender: Group 1: 4 males and 16 females Group 2: 5 males and 15 females
Number randomized: 40 (20/group)
Number analysed: 40 (20/group)*
  Interventions Group 1: NiTi closed coil spring
Group 2: Elastomeric chain
  Outcomes Primary: Rate of tooth movement (mm/month), measured clinically
*Information received via e-mail communication with author (unpublished data)
  Norman 2016
  Methods Location: Orthodontic Department University of Manchester Dental Hospital and Orthodontic Department Countess of Chester Hospital, United Kingdom
Maximum duration of study: 5.2 months
Trial design: Parallel-group randomized clinical trial
  Participants Age: Group I: 16.1 ± 2.8 years and Group II: 16.1 ± 1.8 years
Gender: Group I: 10 males and 9 females and Group II: 9 males and 12 females
Number randomized: 40 patients (Group I-19; Group II-21)
Number evaluated: 30 patients (Group I-15; Group II-15)
  Interventions Group I: NiTi coil spring
Group II: Stainless steel coil spring
  Outcomes Primary: Rate of tooth movement (measured in mm/month), measured from study models
  Fang 2017
  Methods Location: Department of Orthodontics, West China Hospital of Stomatology, Sichuan University, China
Maximum duration of study: Start to end of retraction
Trial design: Parallel-group randomized clinical trial
  Participants Age: Group I: 12.8 ± 0.9 years and Group II: 12.7 ± 0.9 years (for 36 patients analysed)
Gender: 8 males and 10 females for each group (for 36 patients analysed)
Number randomized: 42 patients (21/group)
Number evaluated: 36 patients (18/group)
European Journal of Orthodontics, 2022, Vol. 44, No. 2

Downloaded from https://academic.oup.com/ejo/article/44/2/210/6381541 by guest on 26 July 2023

 B. Sebastian et al.

Table 1.  Continued

Study Characteristics

  Interventions Group I: NiTi coil spring

Group II: Elastomeric module
  Outcomes Primary: Rate of tooth movement (mm/month), measured on study model
b) Individual canine retraction
  Talwar 2018
  Methods Location: Department of Orthodontics and Dentofacial Orthopaedics, Nair Hospital Dental College, Mumbai, Maharashtra, India
Maximum duration of study: 3 months
Trial design: Parallel-group randomized clinical trial
  Participants Age: 20.4 ± 3.52* years (range= 15.2–27.5years)
Gender: 10 10 males and 20 females
Number randomized: 30 patients (15/group)
Number evaluated: 30 patients (15/group) *
  Interventions Group I: Standard NiTi closed coil spring
Group II: Elastomeric chain
  Outcomes Primary: Rate of tooth movement (mm/month), measured from study models
*Information received via e-mail communication with author (unpublished data)
 Davidović, 2018
  Methods Location: University of Banja Luka, Faculty of Medicine, Department of Orthodontics, Banja Luka, Republika Srpska, Bosnia and Herzegovina
Maximum duration of study: 4 months
Trial design: Parallel-group randomized clinical trial
  Participants Age: 12–18 years of age at the beginning of treatment
Gender: Not reported
Number randomized: 25 (Allotment per group not mentioned)
Number evaluated: 23 (Group 1–12; Group 2–11)
  Interventions Group 1: Elastic Chain
Group 2: NiTi closed coil spring
  Outcomes Primary: Rate of tooth movement (mm/month), measured clinically

RCT, randomized controlled trial.

215

Downloaded from https://academic.oup.com/ejo/article/44/2/210/6381541 by guest on 26 July 2023

216 European Journal of Orthodontics, 2022, Vol. 44, No. 2

Table 2.  Summary of Characteristics of studies (Split-mouth RCTs) (Summary version. The full (long) version is shown in Supplementary
Table S4).

Study Characteristics

a) En-masse retraction
  Samuels 1993
  Methods Location: University of Liverpool, Liverpool, United Kingdom
Maximum duration of study: 161 days
Trial design: Split-mouth randomized controlled trial
  Participants Age: Mean, 14 years 7 months (range, 11.1–17.1)
Gender: 5 males and 12 females
Number randomized: 17 patients; 38 quadrants (19 quadrants/group)
Number analysed: 17 patients; 38 quadrants (19 quadrants/group)
  Interventions Group I: 150 g NiTi closed coil spring
Group II: Elastic retraction module
  Outcomes Primary: Rate of tooth movement (mm/week), measured from study models

Downloaded from https://academic.oup.com/ejo/article/44/2/210/6381541 by guest on 26 July 2023

  Samuels 1998
  Methods Location: Department of Orthodontics and Restorative Dentistry, The Glenfield Hospital, Leicester, and the Depart-
ments of Orthodontics and Restorative Dentistry, Liverpool University Dental Hospital, Liverpool, United Kingdom
Maximum duration of study: 187 days
Trial design: Split-mouth randomized controlled trial
  Participants Age: Mean, 15 years 9 months (range, 9.8–24.10)
Gender: 6 males and 12 females
Number randomized: 18 patients; 36 quadrants (18 quadrants/group)
Number analysed: 18 evaluated; 36 quadrants (18 quadrants/group)
  Interventions Group I: 200 g NiTi closed coil spring
Group II: 100 g NiTi closed coil spring
  Outcomes Primary: Rate of tooth movement (mm/week), measured from study models
  Nightingale 2003
  Methods Location: Eastman Dental Hospital, London
Maximum duration of study: 182 days
Trial design: Split-mouth randomized controlled trial
  Participants Age: Patients aged 12–18 years
Gender: Not reported
Number randomized: 15 patients; 40 quadrants (20 quadrants/group)
Number analysed: 8 patients; 22 quadrants (11 quadrants/group)
  Interventions Group I: Elastomeric chain; Group II: NiTi coil springs
  Outcomes Primary: Rate of tooth movement (mm/month), measured on study model
b) Individual canine retraction
  Sonis 1994
  Methods Location: Not reported
Maximum duration of study: Not reported
Trial design: Split-mouth randomized clinical trial
  Participants Age: Not reported
Gender: Not reported
Number randomized: 27 patients; 100 quadrants (50 quadrants/group)
Number analysed: Not reported
  Interventions Group I: NiTi closed coil spring; Group II: Conventional 3/16” latex elastics
  Outcomes Primary: Rate of tooth movement (mm/week), measured clinically
  Bokas 2006
  Methods Location: Orthodontic Unit at the University of Mel-
bourne, Melbourne, Australia
Maximum duration of study: 112 days
Trial design: Split-mouth randomized clinical trial
  Participants Age: Age range: 13–14.5 years
Gender: 6 males and 6 females
Number randomized: 12 patients; 24 quadrants (12
quadrants/group)
Number analysed: 12 patients; 24 quadrants (12
quadrants/group)
  Interventions Group I: Elastomeric chain
Group II: Standard NiTi coil spring
  Outcomes Primary: Rate of tooth movement (mm/month),
measured from study models
B. Sebastian et al. 217

Table 2.  Continued

Study Characteristics

  Khanemasjedi 2017
  Methods Location: Department of Orthodontics, Faculty of Dentistry, Ahvaz Jundishapur University of Medical Sciences,
Ahvaz, Iran
Maximum duration of study: 3 months
Trial design: Split-mouth randomized clinical trial
  Participants Age: Mean, 15 years 2 months (range 12.9–22.4) (For 16 patients analysed)
Sex at baseline: 5 males and 11 females (For 16 patients analysed)
Number randomized: 24 patients; 62 quadrants (31 quadrants/group)
Number analysed: 16 patients; 42 quadrants (21 quadrants/group)
  Interventions Group I: Elastic memory chain
Group II: NiTi coil spring
  Outcomes Primary: Rate of tooth movement (mm/month), measured clinically
  Khalid 2018

Downloaded from https://academic.oup.com/ejo/article/44/2/210/6381541 by guest on 26 July 2023

  Methods Location: Orthodontics Department, Armed Forces Institute of Dentistry, Rawalpindi, Pakistan
Maximum duration of study: 4 weeks
Trial design: Split-mouth randomized clinical trial
  Participants Age: Range in age from 18 to 45 years
Gender: 14 males and 18 females
Number randomized: 32 patients; 64 quadrants (32 quadrants/group)
Number analysed: 32 patients; 64 quadrants* (32 quadrant/group)
  Interventions Group A: Elastomeric module
Group B: NiTi closed coil spring
  Outcomes Primary: Rate of tooth movement (mm/month), measured clinically
*Information received via e-mail communication with author (unpublished data)

RCT, randomized controlled trial.

between NiTi closed coil springs and elastomeric power chain or Anchorage loss: summary of two included studies
active ligatures (MD: 0.06; 95% CI, −0.03–0.16; I2 0%, P = 0.18) Two of the included studies reported on anchorage loss and the inter-
(Figure 5a). ventions compared in both studies were NiTi closed coil springs and
elastomeric power chain (37, 43). The Bokas et al. study reported a
Split-mouth studies with individual canine mean anchorage loss rate per month of 0.46 mm (range: 0.81–0.24)
retraction for NiTi spring and 0.45  mm (range: 0.68–0.19) for elastomeric
chain groups respectively and the mean difference was not statis-
Elastomeric power chain versus NiTi closed coil springs
tically significant (P = 0.7) (43). In the Chaudhari et al. study, the
Two studies (17, 43) were analysed for comparing the two interven-
mean rate of anchorage loss in 4 months for NiTi spring group was
tions and no statistically significant difference for the outcome was
1.1 mm while for elastomeric chain group it was 0.82 mm and the
found in the meta-analysis (MD: −0.17; 95% CI, −0.41 to 0.07; I2
difference was statistically significant (P = 0.006) (37). A meta-anal-
14%, P = 0.17) (Figure 5b).
ysis for this outcome was not attempted due to the clinical hetero-
geneity between the included studies.
NiTi closed coil springs versus other methods
A total of three eligible studies (17, 18, 43) were included with a Network meta-analysis
pooled data involving 120 participants. The meta-analysis failed The data suggested that the values for the rate of movement were
to find any statistically significant difference (MD: 0.11; 95% proportional to the time period over which the measurement was
CI, −0.35–0.57; I2 87%, P = 0.65) between NiTi closed coil springs made, with greater movement for a longer measurement period.
and other methods like elastomeric power chain or active ligatures In order to standardise the period of measurement, studies not re-
(Figure 5c). However, the heterogeneity was high (Figure 5c). porting monthly values were converted to the rate of movement per
month. A visual inspection of the converted data suggested similar
Sensitivity analysis values on the converted scale between studies with different original
In order to assess the impact of the study by Khanemasjedi et  al. time points of measurement.
(17) on the effect estimates and statistical heterogeneity, sensitivity A total of nine studies provided data for analysis (14–17, 36, 37,
analysis was carried out which resulted in estimates favouring 39, 41, 43). Information on the number of studies on which data in
NiTi closed coil springs (MD: 0.36; 95% CI, 0.24–0.49; I2 0%, P each treatment group were collected, along with the pattern of the
˂ 0.00001) (Figure 5d). Additional sensitivity analysis was carried treatment combinations is shown in Figure 6.
out to explore the impact of age variability of participants between Information on the number of direct comparisons between dif-
studies by excluding the study by Khalid et  al. (18). This led to a ferent combinations of studies is also shown in the network map in
reduction in heterogeneity even though the summary results did Figure 7. Both direct comparisons only, and a network meta-analysis
not change significantly (MD: −0.15; 95% CI, −0.44–0.15; I2 14%, were performed. The results of both sets of analyses are shown in
P = 0.33) (Figure 5e). Table 4. A graphical representation of the results of the individual
218 European Journal of Orthodontics, 2022, Vol. 44, No. 2

studies where there was a direct comparison between systems, and

also the pooled differences from the network meta-analysis is shown
in Figure 8.
Overall, the traditional meta-analysis and NMA gave fairly
similar results. Both sets of analyses suggested a significant differ-
ence between the NiTi springs and each of the other two methods,
active ligature and elastomeric chain. The other two methods had
significantly lower outcome values than the NiTi springs, with the
difference largest with the active ligature (Table 4, Figure 7). There
was only borderline evidence of a difference between the active liga-
ture and power chain systems in the direct comparisons, but this
difference was significant in the NMA.
The traditional meta-analyses suggested a high degree of hetero-
geneity between the study results for the comparison of NiTi spring
and power chain, but minimal heterogeneity between the two studies

Downloaded from https://academic.oup.com/ejo/article/44/2/210/6381541 by guest on 26 July 2023

comparing NiTi spring and active ligature.
The NMA was also used to rank studies from the best (highest
rate of tooth movement) to the worst (lowest movement). The per-
centage chance of each treatment taking each rank was calculated.
These results are shown numerically in Table 5. There was fairly
strong evidence that the NiTi spring system had the best result, with
a 99.6% chance that this was the best system. Next came the power
chain method in second place. The active ligature was in third place,
with a 99.6% chance that this was the worst-performing method.

Quality of evidence (GRADE) and Publication bias

The GRADE assessment of the quality of the collected evidence
found low-quality evidence in favour of NiTi coil springs for the rate
of space closure when compared with elastomeric chain and mod-
erate for NiTi coil springs in favour of active ligatures active ligature
(Table 6). This was carried out from NMA results. The results for the
former were downgraded to low due to the high risk of bias in the
included studies. The effect size was also small and so no upgrading
was performed. In the active ligature group, the results for evidence
was downgraded for the risk of bias in the studies whereas, as the
effect size was large the overall quality of evidence was improved to
moderate.
Publication bias assessment tests were not attempted as not
enough studies were included in any individual meta-analysis.

Discussion
Thirteen studies (14–18, 36–43) evaluated the effectiveness of the
NiTi closed coil springs using either split-mouth or parallel-group
design. Seven studies (15–17, 36, 37, 41, 43) evaluated elastomeric
power chain, four studies (14, 18, 36, 39) used active ligatures and
one study each analysed stainless steel coil springs (38) and latex
Figure 2.  Risk of bias summary: Review authors’ judgements – low risk of elastics (42) (Table 3, Supplementary Table S5). The results showed
bias (green); unclear risk of bias (yellow); high risk of bias (red).
that the NiTi closed coil springs produced an increased rate of space
closure when compared with any other method.

Quality of included trials
The overall quality of included studies in this review is low with
none of the studies being judged as low risk of bias. The overall risk
of bias of all included parallel-group studies, except Fang et al., (14)
was judged to be high risk of bias (Figure 2). The study by Fang et al
was judged as unclear risk of bias as the authors did not report on
the allocation concealment method. Interestingly, four studies were
Figure 3. Risk of bias graph: review authors’ judgements for the specific judged high risk in the ‘other bias’ category as three of the studies
domains presented as percentages. (15, 36, 37) had issues relating to gender bias and one study (38) was
Table 3.  Summary of findings for RCTs (Summary version. The full (long) version is shown in Supplementary Table S5).

Type of Sp. Number of patients/quad- Duration of Recall Mean rate of tooth movement/
Study closure Groups rant* randomized (analysed) In weeks month or week in mm (SD) P value
B. Sebastian et al.

Parallel group
  Dixon 2002 En-masse I: Power chain 12 (10) 4 0.58 ± 0.30/month I vs III = 0.45
II: NiTi coil spring 12 (11) 0.81 ± 0.51/month II vs III = 0.017
III: Active ligature 12 (12) 0.35 ± 0.23/month I vs II = 0.50
  Chaudhari 2015 En-masse 1: NiTi closed coil spring 20 (20) 4 0.87/month 0.001
2: Elastomeric chain 20 (20) 0.62/month
  Norman 2016 En-masse I: NiTi coil spring 19 (15) 4 0.58 ± 0.24/month 0.024
II: Stainless steel coil spring 21 (15) 0.85 ± 0.36/month
  Fang 2017 En-masse I: NiTi coil spring 21 (18) 4 1.06 ± 0.16/month 0.000
II: Elastomeric module 21 (18) 0.52 ± 0.14/month
  Talwar 2018 Canine retraction I: NiTi closed coil spring 15 (15) 4 1.62 ± 0.14/month 0.000
II: Elastomeric chain 15 (15) 1.33 ± 0.13/month
 Davidović, 2018 Canine retraction 1: Elastic Chain 25 (23) 4 3.10 ± 1/4 months ˃ 0.05
2: NiTi closed coil spring 3.94 ± 1.06/4 months
Split-mouth
  Samuels 1993 En-masse I: 150 g NiTi closed coil 17/19 (17/19) Mean, 6 weeks (range, 5–8 0.26/week 0.828
spring weeks)
II: Elastic retraction module 17/19 (17/19) 0.19/week
  Samuels 1998 En-masse I: 200 g NiTi closed coil 18/18 (18/18) Mean, 49 days(range, 26–days)0.24/week Not reported**
spring
II: 100 g NiTi closed coil 18/18 (18/18) 0.16/week
spring
  Nightingale 2003 En-masse I: Elastomeric chain 15 /20 (8/11) 4–6 0.84 ± 0.52/month 0.46 based on MD/
0.21± 0.13/week week
II: NiTi coil spring 15 /20 (8/11) 1.04 ± 0.80/month
0.26 ± 0.20/week
  Sonis 1994 Canine retraction I: NiTi closed coil spring 27/50 (Not reported) 3–6 0.51 ± 0.07/week ≤0.0001 based on MD/
II: Conventional 3/16” latex 27/50 (Not reported) 0.27 ± 0.05/week week
elastics
  Bokas 2006 Canine retraction I: Elastomeric chain 12/12 (12/12) 4 1.68 (range: 2.00–1.15)/month 0.011
II: Standard NiTi coil spring 12/12 (12/12) 1.85 (range: 2.33–1.11)/month
 Khanemasjedi Canine retraction I: Elastic memory chain 24/31 (16/21) 4 1.89 ± 0.36/month 0.022
2017 II: NiTi coil spring 24/31 (16/21) 1.67 ± 0.39/month
  Khalid 2018 Canine retraction A: Elastomeric module 32/32 (32/32) 4 0.77 ± 0.27/month ˂0.001
B: NiTi closed coil spring 32/32 (32/32) 1.14 ± 0.26/month

SD, standard deviation; MD, mean difference; RCT, randomized controlled trial.
*In split mouth studies, total number of patients with per group randomization of quadrants is given.
**Analysis was performed for patients included in Samuels 1993 and 1998 studies. Individual study results are not reported.
219

Downloaded from https://academic.oup.com/ejo/article/44/2/210/6381541 by guest on 26 July 2023

220 European Journal of Orthodontics, 2022, Vol. 44, No. 2

Downloaded from https://academic.oup.com/ejo/article/44/2/210/6381541 by guest on 26 July 2023

Figure 4.  (a–e) Meta-analyses of Parallel-group randomized controlled trials
(RCT) comparing interventions (rate of space closure).
Figure 5. (a–e) Meta-analyses of Split-mouth randomized controlled trials
(RCT) comparing interventions (rate of space closure).
terminated early due to resource and time constraints and hence the
random code was broken before finish.
In the split-mouth design, all studies except Khalid et  al. (18)
were judged to be high risk of bias (Figure 2). None of the studies
had a strategy for allocation concealment. In addition, three studies
failed to perform assessor blinding (39, 40, 42). Four studies (17,
39–41) were judged high risk on other bias category with gender
bias being a major issue in three of them (17, 39, 40).
Four of the included trials (36, 37, 41, 43) in this review were
also included in a previous systematic review (13). They had re-
ported Dixon et al. (36) as having an overall low risk of bias across
all examined domains and Bokas et al. (43) as unclear risk of bias
due to lack of descriptions on random sequence generation and al-
location concealment. Contrary to this, we judged both the studies
as high risk of bias. In Dixon et al., (36) out of the 33 patients in-
volved, 23 were girls and 10 were boys indicating a clear gender bias.
Additionally, the authors mentioned a dropout rate of 10% in the
study, but the reason for dropouts was not reported. The other study Figure 6.  Network pattern.
(43) lacked information on randomisation sequence generation and
allocation concealment and so was judged as having a high risk of space closure. Although a split-mouth design helps to eliminate bio-
bias in our assessment. logical variables between patients, it can have carry-across effects
from the intervention to the control side as there is no natural barrier
Split-mouth versus parallel-group and this can introduce bias in the results (46, 47). Hence a split-
This review included RCTs of different study designs, such as split- mouth design is not ideal to compare different space closure force
mouth (17, 18, 39–43) and parallel-arm (14–16, 36–38) to evaluate delivery systems as closing space in one quadrant will have an effect
the effectiveness of different force delivery systems on the rate of on the position of teeth in other quadrants (38). Also, the reporting
B. Sebastian et al. 221

including the studies with missing variance. The results were very
similar to the ones with imputations.

Heterogeneity
Many possible sources of heterogeneity were identified in terms of
their study designs, methodological variations mainly due to type of
retraction, the timing of extractions, data collection, measurement
techniques and measured outcomes. The trials included used different
methodologies to assess the effectiveness of the force delivery systems,
either en-masse retraction (14, 36–41) or individual canine retraction
(15–18, 42, 43). This would affect the rate of tooth movement out-
come assessment significantly as en-masse studies measure the rate of
space closure of six anterior teeth while the other group measures the
rate of space closure of canines. Hence, unlike the previous reviews

Downloaded from https://academic.oup.com/ejo/article/44/2/210/6381541 by guest on 26 July 2023

(13, 27), we subgrouped the individual studies accordingly to elim-
Figure 7.  Network map.
inate clinical heterogeneity in our traditional meta-analysis.
There was variation in the follow-up period between studies in
of the split-mouth trials is frequently poor and the information both split-mouth and parallel-group design. For example, in the two
needed to undertake a meta-analysis is often not directly available individual canine retraction studies from parallel-arm design group,
(48). We also faced a similar problem regarding the overall quality of one study reported rate on canine retraction every four months
the included split-mouth trials. In order not to dilute the quality of (16) and the other study reported rate every month (15). For this
reported evidence, we analyzed split-mouth trials with adequate data reason, a traditional meta-analysis was not possible in the subgroup.
separately from parallel-group trials (46). This approach was dif- Similarly, for split-mouth studies, three studies (39, 40, 42) reported
ferent from the previous systematic review comparing the effective- on space closure per week, three studies (17, 18, 43) reported space
ness of NiTi closed coil springs and elastomeric power chain where closure rate per month and one study (41) reported both weekly and
two of four included studies were of split-mouth design (13). In their monthly space closure rate. For NMA, we standardized the period
review, the meta-analysis of the pooled data from the four studies of measurement, i.e., studies not reporting monthly values were con-
found a statistically significant mean difference favouring the NiTi verted to the rate of movement per month. A  visual inspection of
closed coil springs (MD = 0.20 mm/month, 95% CI: 0.12–0.28; I2 the converted data suggested similar values on the converted scale
0%, P ˂ 0.00001). The present review is in accordance, with a statis- between studies with different original time points of measurement.
tically significant difference (MD: 0.24; 95% CI: 0.03–0.45; I2 0%, There was further variation in the intervention groups with
P = 0.02) (Figure 4a) favouring NiTi closed coil springs when com- studies evaluating different forces on NiTi springs, stainless steel
pared with elastomeric power chain in the parallel-arm en-masse coil springs, elastomeric chain, active ligatures, and latex elastics
space closure group. But a similar comparison of the two interven- (Supplementary Table S5). This, in addition to the above clinical
tions in the split-mouth individual canine retraction subset failed to variations restricted the number of traditional meta-analysis that
show any statistically significant difference for the outcome (Figure could be performed. For NMA, two studies (38, 42) were excluded
5b) (17, 43). as they contained data on delivery systems other than the three most
commonly occurring methods. Thus, after exclusion of these add-
Handling of missing outcome data itional systems, there was no other system to make a within study
Four included studies did not provide information on standard devi- comparison with and were, therefore, unable to be included in the
ation. In order to include these studies in our meta-analysis, SD was analysis. One further study (40) compared two different types of
imputed for three trials (37, 39, 40). One trial (43) had provided NiTi springs. As there were no comparisons between different force
range for space closure and SD was estimated from the observed delivery systems, this study was also excluded from all analyses.
range (45). Studies have shown that trial publications do not con-
sistently report the SDs for change in outcomes from baseline (49, Sensitivity analysis
50) and the omission of such studies from a meta-analysis due to in- Among the various other factors postulated to affect the rate of
complete reporting may introduce bias and reduce statistical power orthodontic tooth movement, age has certainly been shown to alter
(50). When there is not enough information available to calculate the speed of tooth movement by affecting bone density and remod-
the standard deviations for the changes, they can be imputed from elling (54, 55). Sensitivity analysis was carried out to assess the im-
observed SD values of other similar outcome measure studies pro- pact of this by excluding from meta-analysis those individual studies
vided they use the same measurement scale, has the same degree of which reported mean age above 18 years (18, 37). But the analyses
measurement error and has the same follow up period (44, 51, 52). failed to show any significant change in the summary results (Figures
Additionally, sensitivity analysis was performed to assess the robust- 4e and 5e). One study (18) measured outcome in terms of space
ness of the results in relation to the assumptions made, as a verifica- opening mesial to canine rather than space closure distal to canine.
tion for the imputation (53), and the results showed that any source The sensitivity analysis performed by excluding this study showed a
of heterogeneity associated with the imputed missing data had min- significant reduction in I2 value (87%–14%).
imal effect on the overall meta-analysis estimate (Figure 4e). In our Further sensitivity analyses were performed to evaluate the reason
network meta-analysis, we used the similar imputation method to for unacceptable heterogeneity with a high I2 value. This was seen
include the two studies with missing SD. However, a further NMA when NiTi spring was compared with other multiple methods of space
was carried out to evaluate the effectiveness of space closure without closure. Sensitivity analysis investigating the influence of two studies
222 European Journal of Orthodontics, 2022, Vol. 44, No. 2

Table 4.  Summary of network meta-analysis results.

Direct comparisons Network meta-analysis

Group 1 Group 2 N. studies Difference (95% CI) (*) P value I2 Difference (95% CI) (*) P value

NiTi spring Active ligature 4 −0.43 (−0.55, −0.31) <0.001 55% −0.42 (−0.56, −0.28) <0.001
Power chain 7 −0.17 (−0.30, −0.03) 0.02 64% −0.17 (−0.29, −0.05) 0.007
Active ligature Power chain 1 0.23 (0.03, 0.46) 0.05 - 0.25 (0.08, 0.42) 0.005

*Differences reported as Group 2 minus Group 1; Statistically significant P values shown in bold text.

springs and active ligature. This result can be considered clinically sig-
nificant as the space closure phase in orthodontics on an uncrowded
dentition with extraction treatment may involve around six to eight
months. If this is averaged out it could halve the time taken for space
closure. However, the difference between NiTi springs and elastomeric

Downloaded from https://academic.oup.com/ejo/article/44/2/210/6381541 by guest on 26 July 2023

chain may not be clinically significant (MD 0.17 mm, 95% CI 0.03–
0.30, P = 0.02) although the results were statistically significant.

Secondary outcome: anchorage loss assessment

Among the included studies, there were only two trials (37, 43)
which looked into anchorage loss and both compared between
elastomeric power chain and NiTi closed coil springs. But the clinical
heterogeneity between the studies prevented pooling of data for ana-
lysis. The Bokas et al. (43) trial had a split-mouth design individual
canine retraction methodology with anchorage reinforcement using
TPA while the Chaudhari et al. (37) was a parallel-group en-masse
Figure 8. Forest plot of direct comparisons from individual studies and
pooled differences from network meta-analysis.
retraction space closure study with no reinforcement of anchorage.
Another notable difference was that measurements were made on
study casts in Bokas et  al. (43) while the other study used lateral
Table 5.  Ranking of treatments. cephalograms. Additionally, the outcome measurement strategy was
also different with Bokas et al., (43) trial giving mean monthly rate
Position NiTi Spring Active Ligature Power chain
and the other study reported the mean anchor loss after four months.
1st 99.6% 0% 0.4% For this reason, a NMA was not possible for this outcome measure.
2nd 0.4% 0.4% 99.2%
3rd 0% 99.6% 0.4% Other secondary outcomes
The reporting of the secondary outcomes was very poor in the in-
Figures are estimated percentage of each treatment being in each position. cluded studies. Harms was discussed in two RCTs out of which
Norman et al. (38) trial considered the incidence of breakages and
(17, 38) on the overall meta-analysis resulted in estimates favouring distortion of coil springs and reported that none were encountered,
NiTi closed coil springs with a reduction in I2 value (Figures 4d and while the Khanemasjedi et al. (17) study identified no harms except
5d). It is important to point out that the original synthesis had re- for the routine side effects of orthodontic treatment such as pain.
ported no statistically significant difference (Figures 4c and 5c). Only one trial (38) considered the cost-effectiveness of interven-
tions, reporting the cost of NiTi coil springs as £32.29 for a pack of
Primary outcome assessment ten and stainless steel springs as £7.34 for a pack of ten. None of the
The results from the six subgroup analyses from our traditional trials carried out any robust cost evaluation analysis.
meta-analysis and NMA were summed up and interpreted accord- None of the trials discussed about the unwanted consequences
ingly on the basis of relevant evidence. The results showed a statis- of treatment like adverse treatment effects including adverse peri-
tically significant difference favouring NiTi closed coil springs. Our odontal effects, root resorption. In addition, patient-related experi-
results comparing the rate of space closure of NiTi close coil spring ences like pain remained unexplored.
and elastomeric chain concurred with the results of a previous re-
view (13) that found the moderate quality of evidence in this regard. Network meta-analysis
In a similar systematic review by Barlow et al. (27) on the efficiency This is the first NMA in the literature for evaluating the rate of
of different force delivery systems, the authors were unable to con- tooth movement using different space closure techniques. The NMA
duct a meta-analysis because of heterogeneity of the data. was carried out to improve the precision of effect sizes, allowing us
NMA on the primary outcome allowed us to carry out direct and to compare interventions that were not possible with traditional
indirect comparisons with more participants included in the overall meta-analysis. The results clearly showed an increase in the confi-
analysis than the traditional meta-analysis. This strengthens the overall dence in the results on the effectiveness of NiTi coil springs in com-
evidence available on the preferred space closure technique. The results parison with other space closure techniques. The NMA ranking also
from meta-analysis shows a 0.43 mm (95% CI 0.31–0.55, P < 0.001) showed fairly strong evidence for NiTi coil as the best method for
per month mean difference in space closure method between NiTi space closure. Active ligature was shown to be the worst.
 B. Sebastian et al.

Table 6.  Comparison of space closure techniques for en-masse retraction in sliding mechanics using GRADE assessment.

Patient or population: Orthodontic space closure in sliding mechanics.

Setting: Secondary and tertiary care.
Intervention: Orthodontic space closure with elastomeric power chain, NiTi springs or any other method.
Comparison: Orthodontic space closure with NiTi coil springs, stainless steel coil springs, activate ligatures, elastics, elastic thread or any other method or untreated control group.

Illustrated comparative risks* (95% CI)

No of participants Certainty of the evi-
Outcomes (studies) Follow up dence (GRADE) Assumed risk different methods Corresponding risk NiTi springs Comments

Space closure with NiTi 172 (7 RCTs) ⊕⊕〇〇 LOW Mean space closure per month Mean space closure per month Data from network meta-analysis.
springs versus power chain ranged from 0.35 to 1.89 mm was 0.17 mm higher (0.04 higher Downgrading was carried out for risk
to 0.30 higher) of bias. No upgrading on effect size was
made
Space closure with NiTi 76 (3 RCTs) ⊕⊕⊕〇 MODERATE Mean space closure with NiTi Mean space closure was 0.44 mm Data from network meta-analysis.
springs versus Active ligature springs ranged from higher (0.26 higher to 0.61 higher) Downgrading was carried out for risk of
0.19 to 0.52 mm bias. Upgrading on effect size was made

*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; GRADE: Grading of recommendations, assessment, development and evaluation.
GRADE Working Group grades of evidence
High certainty: We are very confident that the true effect lies close to that of the estimate of the effect.
Moderate certainty: 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 certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect.
Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect.
223

Downloaded from https://academic.oup.com/ejo/article/44/2/210/6381541 by guest on 26 July 2023

224 European Journal of Orthodontics, 2022, Vol. 44, No. 2

The NMA only included studies where a measurement of variability 2. Skidmore,  K.J., Brook,  K.J., Thomson,  W.M. and Harding,  W.J. (2006)
in the data (e.g. standard deviation) was reported. Two studies (37, 39) Factors influencing treatment time in orthodontic patients. American
were excluded for this reason. However, further NMA of the data with Journal of Orthodontics and Dentofacial Orthopedics, 129, 230–238.
3. Yassir,  Y.A., El-Angbawi,  A.M., McIntyre,  G.T., Revie,  G.F. and
imputed SD for the two studies did not show any difference in result
Bearn,  D.R. (2019) A randomized clinical trial of the effectiveness
compared to without imputation. Additionally, the NMA excluded two
of 0.018-inch and 0.022-inch slot orthodontic bracket systems: part
studies (38, 42) that evaluated other than the three most common force
1-duration of treatment. European Journal of Orthodontics, 41, 133–
delivery systems, and one study (40) was excluded as they compared 142.
two different types of NiTi coil springs. The latter one was excluded 4. Uribe,  F., Padala,  S., Allareddy,  V. and Nanda,  R. (2014) Patients’, par-
as there were no comparisons between different force delivery systems. ents’, and orthodontists’ perceptions of the need for and costs of add-
itional procedures to reduce treatment time. American Journal of Ortho-
Limitations dontics and Dentofacial Orthopedics, 145, S65–S73.
5. Sameshima, G.T. and Sinclair, P.M. (2001) Predicting and preventing root
The review was limited with the number of studies included in the trad-
resorption: Part II. Treatment factors. American Journal of Orthodontics
itional meta-analysis. This was due to the difference in study designs
and Dentofacial Orthopedics, 119, 511–515.
and the variety of study outcome measures used. In addition, there was 6. Segal, G.R., Schiffman, P.H. and Tuncay, O.C. (2004) Meta analysis of the
lack of sufficient data regarding the secondary outcomes. This shows treatment-related factors of external apical root resorption. Orthodontics

Downloaded from https://academic.oup.com/ejo/article/44/2/210/6381541 by guest on 26 July 2023

the urgent need for standardization of study design and the need for & craniofacial research, 7, 71–78.
development of an agreed, standardized core outcome sets and core 7. Tufekci, E., Dixon, J.S., Gunsolley, J.C. and Lindauer, S.J. (2011) Preva-
outcome measurement instrument sets for orthodontic trials (56). The lence of white spot lesions during orthodontic treatment with fixed appli-
development of this combined core outcome and a core outcome meas- ances. The Angle Orthodontist, 81, 206–210.
urement instrument would avoid frustration for researchers trying to 8. Pinto, A.S., Alves, L.S., Maltz, M., Susin, C. and Zenkner, J.E.A. (2018)
Does the duration of fixed orthodontic treatment affect caries activity
combine evidence on a treatment or an orthodontic intervention and
among adolescents and young adults? Caries Research, 52, 463–467.
ultimately allow patients to make more informed decisions.
9. Mandall, N.A. and Malik, O. (2007) Scientific evidence for pre-adjusted
edgewise mechanics. Dental Update, 34, 45–6, 48, 50.
10. Monini, A.d.a.C., Gandini Júnior, L.G., dos Santos-Pinto, A., Maia, L.G.
Conclusions
and Rodrigues,  W.C. (2013) Procedures adopted by orthodontists for
1. The overall quality of evidence was low with eleven of thirteen space closure and anchorage control. Dental Press Journal of Orthodon-
studies judged as high risk of bias. tics, 18, 86–92.
2. There is moderate-certainty evidence favouring NiTi coil springs 11. Banks, P., Elton, V., Jones, Y., Rice, P., Derwent, S. and Odondi, L. (2010)
for the rate of space closure when compared with active ligatures The use of fixed appliances in the UK: a survey of specialist orthodontists.
Journal of Orthodontics, 37, 43–55.
and low-certainty evidence when compared with power chain.
12. Burrow,  S.J. (2009) Friction and resistance to sliding in orthodontics: a
3. The NMA ranking showed NiTi springs as the best method for
critical review. American Journal of Orthodontics and Dentofacial Ortho-
space closure followed by elastomeric chain. The active ligatures
pedics, 135, 442–447.
were graded as the least effective of the three methods. 13. Mohammed, H., Rizk, M.Z., Wafaie, K. and Almuzian, M. (2018) Effect-
4. An urgent need for standardization of study designs and the need iveness of nickel-titanium springs vs elastomeric chains in orthodontic
for the development of an agreed core outcome sets and core space closure: a systematic review and meta-analysis. Orthodontics &
outcome measurement instrument sets for orthodontic trials is craniofacial research, 21, 12–19.
evident from this review. 14. Fang, S., Zhong, Y., Li, M., Luo, J., Khadka, N., Jiang, C., Wang, J., and
Du,  X. (2017). Comparing two methods of orthodontics space closure:
a randomized clinical trial. International Journal of Clinical and Experi-
Supplementary material mental Medicine, 10, 14667–14672.
15. Talwar,  A., and Bhat,  S.R. (2018). Comparative evaluation of nickel-

Supplementary materials are available at the European Journal of
titanium closed coil spring and Elastomeric chain for canine retraction.
Orthodontics online. A randomized clinical trial. IOSR Journal of Dental and Medical Sciences,
17, 70–75.
16. Davidović, M.U., Savić, M.A., and Arbutina, A. (2018). Examination of
Funding
postextraction space closure speed using elastic chains and NiTi closed
The study was self funded. coil springs. Serbian Dental Journal, 65, 179–186.
17. Khanemasjedi,  M., Moradinejad,  M., Javidi,  P., Niknam,  O., Jah-

romi, N.H., and Rakhshan, V. (2017). Efficacy of elastic memory chains
Conflict of interest versus nickel-titanium coil springs in canine retraction: a two-center split-
None declared. mouth randomized clinical trial. International Orthodontics, 15, 561–
574.
18. Khalid, Z., Bangash, A.A., Anwar, A., Pasha, H. and Amin, E. (2018) Ca-
Data availability nine retraction using a closed nickel titanium coil spring and an elastic
module. Journal of the College of Physicians and Surgeons–Pakistan:
The data relevant to this article will be shared on reasonable request
JCPSP, 28, 695–698.
to the corresponding author. 19. Moher,  D., Liberati,  A., Tetzlaff,  J. and Altman,  D.G.; PRISMA Group.
(2009) Preferred reporting items for systematic reviews and meta-analy-
ses: the PRISMA statement. PLoS Medicine, 6, e1000097.
References 20. Higgins, J.P.T., Thomas, J., Chandler, J., Cumpston, M., Li, T., Page, M.J.
1. Mavreas, D. and Athanasiou, A.E. (2008) Factors affecting the duration of and Welch,  V.A (eds.), Cochrane Handbook for Systematic Reviews of
orthodontic treatment: a systematic review. European Journal of Ortho- Interventions, Version 6.0 (Updated July 2019). Cochrane, 2019. Available
dontics, 30, 386–395. from www.training.cochrane.org/handbook. Accessed March 09, 2020
B. Sebastian et al. 225

21. White,  I.R., Barrett,  J.K., Jackson,  D. and Higgins,  J.P. (2012) Consist- 39. Samuels, R.H., Rudge, S.J. and Mair, L.H. (1993) A comparison of the rate
ency and inconsistency in network meta-analysis: model estimation using of space closure using a nickel-titanium spring and an elastic module: a
multivariate meta-regression. Research Synthesis Methods, 3, 111–125. clinical study. American Journal of Orthodontics and Dentofacial Ortho-
22. Salanti, G., Higgins, J.P., Ades, A.E. and Ioannidis, J.P. (2008) Evaluation pedics, 103, 464–467.
of networks of randomized trials. Statistical Methods in Medical Research, 40. Samuels, R.H., Rudge, S.J. and Mair, L.H. (1998) A clinical study of space
17, 279–301. closure with nickel-titanium closed coil springs and an elastic module.
23. Guyatt,  G.H., Oxman,  A.D., Schünemann,  H.J., Tugwell,  P. and
American Journal of Orthodontics and Dentofacial Orthopedics, 114,
Knottnerus, A. (2011) GRADE guidelines: a new series of articles in the 73–79.
Journal of Clinical Epidemiology. Journal of Clinical Epidemiology, 64, 41. Nightingale, C. and Jones, S.P. (2003) A clinical investigation of force de-
380–382. livery systems for orthodontic space closure. Journal of Orthodontics, 30,
24. Sterne, J.A.C, Egger, M. and Moher, D, Higgins, J.P.T., and Green, S. (eds.). 229–236.
Chapter 10: addressing reporting biases. In Cochrane Handbook for Sys- 42. Sonis, A.L. (1994) Comparison of NiTi coil springs vs. elastics in canine
tematic Reviews of Intervention, Version 5.1.0 (Updated March 2011). retraction. Journal of Clinical Orthodontics: JCO, 28, 293–295.
The Cochrane Collaboration, 2011. Available from www.handbook. 43. Bokas,  J. and Woods,  M. (2006) A clinical comparison between nickel
cochrane.org. Accessed March 16,2020 titanium springs and elastomeric chains. Australian Orthodontic Journal,
25. Sonis,  A.L., Van  der  Plas,  E. and Gianelly,  A. (1986) A comparison of 22, 39–46.
elastomeric auxiliaries versus elastic thread on premolar extraction site 44. Weir, C.J., Butcher, I., Assi, V., Lewis, S.C., Murray, G.D., Langhorne, P.

Downloaded from https://academic.oup.com/ejo/article/44/2/210/6381541 by guest on 26 July 2023

closure: an in vivo study. American Journal of Orthodontics, 89, 73–78. and Brady,  M.C. (2018) Dealing with missing standard deviation and
26. Sueri,  M.Y. and Turk,  T. (2006) Effectiveness of laceback ligatures on mean values in meta-analysis of continuous outcomes: a systematic review.
maxillary canine retraction. The Angle Orthodontist, 76, 1010–1014. BMC Medical Research Methodology, 18, 25.
27. Barlow, M. and Kula, K. (2008) Factors influencing efficiency of sliding 45. Walter, S.D. and Yao, X. (2007) Effect sizes can be calculated for studies
mechanics to close extraction space: a systematic review. Orthodontics & reporting ranges for outcome variables in systematic reviews. Journal of
craniofacial research, 11, 65–73. Clinical Epidemiology, 60, 849–852.
28. al-Sayigh, N.M.S. and Ismail, A.J. (2011). Evaluation of space closure rate 46. Lesaffre, E., Philstrom, B., Needleman, I. and Worthington, H. (2009) The
during canine retraction with nickel titanium closed coil spring and elasto- design and analysis of split-mouth studies: what statisticians and clinicians
meric chain. Al-Rafidain Dental Journal, 11, 146–153. should know. Statistics in Medicine, 28, 3470–3482.
29. Mitra,  R., Londhe,  S.M. and Kumar,  P. (2011) A comparative evalu-
47. Pandis, N., Walsh, T., Polychronopoulou, A., Katsaros, C. and Eliades, T.
ation of rate of space closure after extraction using E-chain and stretched (2013) Split-mouth designs in orthodontics: an overview with applica-
modules in bimaxillary dentoalveolar protrusion cases. Medical Journal, tions to orthodontic clinical trials. European Journal of Orthodontics, 35,
Armed Forces India, 67, 152–156. 783–789.
30. Dholakia,  K.D. and Bhat,  S.R. (2012). Clinical efficiency of nonconven- 48. Lesaffre, E., Garcia Zattera, M.J., Redmond, C., Huber, H. and Needle-
tional elastomeric ligatures in the canine retraction phase of preadjusted man,  I.; ISCB Subcommittee on Dentistry. (2007) Reported methodo-
edgewise appliance therapy: an in-vivo study. American Journal of Ortho- logical quality of split-mouth studies. Journal of Clinical Periodontology,
dontics and Dentofacial Orthopedics, 141, 715–722. 34, 756–761.
31. Al Suleiman, M. and Shehadah, M. (2014). Comparison of two methods 49. Streiner, D.L. and Joffe, R. (1998) The adequacy of reporting randomized,
for canine retraction depending on direct skeletal anchorage system (CR- controlled trials in the evaluation of antidepressants. Canadian Journal of
DSAS). International Journal of Dentistry and Oral Health, 1, 21–32. Psychiatry. Revue Canadienne de Psychiatrie, 43, 1026–1030.
32. Patel,  P., Shanthraj,  R., Bhagyalakshmi,  A., Garg,  N. and Vallakati,  A. 50. Wiebe, N., Vandermeer, B., Platt, R.W., Klassen, T.P., Moher, D. and Bar-
(2016) Comparative evaluation of pentraxin 3 levels in GCF during ca- rowman, N.J. (2006) A systematic review identifies a lack of standardiza-
nine retraction with active tieback and NiTi coil spring: an in vivo study. tion in methods for handling missing variance data. Journal of Clinical
Journal of Orthodontic Science, 5, 52–56. Epidemiology, 59, 342–353.
33. Shankar, S., Ranvijay, Chandra, S., and Shahi, A.K. (2017). A comparison 51. Furukawa, T.A., Barbui, C., Cipriani, A., Brambilla, P. and Watanabe, N.
between space closure by canine retraction with active tiebacks and closed (2006) Imputing missing standard deviations in meta-analyses can provide
coil springs: a clinical study with the MBT system. International Journal accurate results. Journal of Clinical Epidemiology, 59, 7–10.
of Medical Research Professionals, 3, 365–370. 52. Fournier, M., Germe, M., Theobald, K., Scholz, G.H. and Lehmacher, W.
34. Thukral, R., Mangat, S., Ganguly, A., Agarkar, S.S., Bali, H. and Grover, S. (2014) Indirect comparison of lixisenatide versus neutral protamine Hage-
(2017) Pentraxin-3 levels in gingival crevicular fluid during canine retrac- dorn insulin as add-on to metformin and sulphonylurea in patients with
tion with nickel-titanium coil spring and active tieback. The Journal of type 2 diabetes mellitus. German Medical Science: GMS e-journal, 12,
Contemporary Dental Practice, 18, 710–713. Doc14.
35. Sar,  S.K., Shetty,  D., Kumar,  P., Juneja,  S. and Sharma,  P. (2019) Leptin 53. Batson, S. and Burton, H. (2016) A systematic review of methods for han-
levels in gingival crevicular fluid during canine retraction: in vivo com- dling missing variance data in meta-analyses of interventions in type 2
parative study. Journal of Orthodontics, 46, 27–33. diabetes mellitus. PLoS One, 11, e0164827.
36. Dixon,  V., Read,  M.J., O’Brien,  K.D., Worthington,  H.V. and Man-
54. Alikhani,  M., et  al. (2018) Age-dependent biologic response to ortho-
dall, N.A. (2002) A randomized clinical trial to compare three methods of dontic forces. American Journal of Orthodontics and Dentofacial Ortho-
orthodontic space closure. Journal of Orthodontics, 29, 31–36. pedics, 153, 632–644.
37. Chaudhari, C. V., and Tarvade, S. (2015). Comparison of rate of retraction 55. Dudic, A., Giannopoulou, C. and Kiliaridis, S. (2013) Factors related to
and anchorage loss using nickel titanium closed coil springs and elasto- the rate of orthodontically induced tooth movement. American Journal of
meric chain during the en-masse retraction: a clinical study. Journal of Orthodontics and Dentofacial Orthopedics, 143, 616–621.
Orthodontic Research, 3, 129–133. 56. Tsichlaki, A. and O’Brien, K. (2014) Do orthodontic research outcomes
38. Norman,  N.H., Worthington,  H. and Chadwick,  S.M. (2016) Nickel ti- reflect patient values? A systematic review of randomized controlled tri-
tanium springs versus stainless steel springs: a randomized clinical trial of als involving children. American Journal of Orthodontics and Dentofacial
two methods of space closure. Journal of Orthodontics, 43, 176–185. Orthopedics, 146, 279–285.