Self-etch primers and conventional acid-etch technique for ...cdn.elsevier.com/promis_misc/YMOD_Model_Orthodontic_Systematic_Review.pdfLondon, United Kingdom, Corfu, Greece, and Bern,

SYSTEMATIC REVIEW

Self-etch primers and conventional acid-etchtechnique for orthodontic bonding: A systematicreview and meta-analysis

Padhraig S. Fleming,a Ama Johal,b and Nikolaos Pandisc

London, United Kingdom, Corfu, Greece, and Bern, Switzerland

aLocubSeniocPrivaOrthoUniveThe aproduReprinLondoKingdSubm0889-Copyrdoi:10

Introduction: The use of self-etch primers has increased steadily because of their time savings and greatersimplicity; however, overall benefits and potential disadvantages and harms have not been assessedsystematically. In this study, we reviewed randomized controlled trials to assess the risk of attachment failure,bonding time, and demineralization adjacent to attachments between 1-stage (self-etch) and 2-stage (acid etch)bonding in orthodontic patients over aminimum follow-up period of 12months.Methods:Data sourceswere elec-tronic databases including MEDLINE, EMBASE, the Cochrane Oral Health Group's Trials Register, and CEN-TRAL, without language restrictions. Unpublished literature was searched on ClinicalTrials.gov, the NationalResearch Register, and Pro-Quest Dissertation Abstracts and Thesis database. Authors were contacted whennecessary, and reference lists of the included studies were screened. Search terms included randomizedcontrolled trial, controlled clinical trial, random allocation, double-blind method, single-blind method,orthodontics, self-etch, SEP, primer, and bonding agent. Randomized clinical trials directly comparing self-etchand acid-etch primers with respect to the predefined outcomes and including patients with full-arch, fixed, andbonded orthodontic appliances (not banded) with follow-up periods of at least 12 months were included. Usingpredefined forms, 2 authors undertook independent data extraction with conflict resolution by the third author.Randomized clinical trial quality assessment based on the Cochrane Risk of Bias tool was also used. Results:Eleven studies met the inclusion criteria; 6 were excluded because of a high risk of bias. In total, 1721 bracketsbonded with acid-etch and 1723 with self-etch primer techniques were included in the quantitative synthesis.Relatively low statistical and clinical heterogeneity was observed among the 5 randomized clinical trials (n 53444 brackets) comparing acid-etch with self-etch primers. A random effects meta-analysis demonstrateda tendency for a higher risk of failure (odds ratio, 1.35; 95% CI, 0.99-1.83; P 5 0.06) with self-etch primers. Asmall but statistically significant time saving was also associated with the self-etch primer technique (weightedmean difference, 23.2 seconds per bracket; 95% CI, 20.7-25.8; P\0.001). There was insufficient evidence toassess the effect of bondingmodality on demineralization rates.Conclusions: There is weak evidence indicatinghigher oddsof failurewith self-etchprimer thanacid etch over 12months in orthodontic patients, and there is strongevidence that a self-etchprimer is likely to result in amodest timesavings (8minutes for full bonding) comparedwithacid etch. Funding: No funding was received for this review. (Am J Orthod Dentofacial Orthop 2012;-:83-95)

Dental bonding was introduced by Bowen1 afterthe pioneering work on enamel preparation tech-niques of Buonocore et al.2 These principles were

m consultant, Queen Mary University, London, United Kingdom.r lecturer/consultant, Queen Mary University, London, United Kingdom.te practice, Corfu, Greece; visiting assistant professor, Department ofdontics and Dentofacial Orthopedics, Dental School, Medical Faculty,rsity of Bern, Bern, Switzerland.uthors report no commercial, proprietary, or financial interest in thects or companies described in this article.t requests to: Padhraig S. Fleming, Locum Consultant, Barts and Then NHS Trust, Dental Institute, Whitechapel, London E1 1BB, Unitedom; e-mail, [email protected], October 2011; revised and accepted, February 2012.5406/$36.00ight � 2012 by the American Association of Orthodontists..1016/j.ajodo.2012.02.023

subsequently applied to orthodontics, revolutionizingappliances physically and cosmetically, with multi-banded systems becoming obsolete and superseded bybonded appliances.3

Further progress has been made in relation to bond-ing with an emphasis on streamlining the process, en-hancing performance in a moist environment, andimproving resistance to demineralization.4 In recentyears, there has been growing interest in 1-step bondingsystems, which do not rely on separate application ofetchant and bonding material. Self-etch bonding sys-tems or self-etch primers (SEPs) are routinely used by29.5% of practitioners in the United States.5 These sys-tems typically incorporate methacrylated phosphoricacid esters; after application to enamel, the phosphate

83

Delta:1_given name

Delta:1_given name

Delta:1_given name

Delta:1_given name

Delta:1_given name

Delta:1_given name

Delta:1_surname

Delta:1_surname

Delta:1_given name

Delta:1_given name

Delta:1_given name

Delta:1_given name

Delta:1_given name

Delta:1_given name

Delta:1_surname

Delta:1_surname

Delta:1_given name

Delta:1_given name

http://ClinicalTrials.gov

mailto:[email protected]

padhraigfleming

Highlight

Item 1: Title. Identify as meta-analysis or systematic review. Explanation: Search filters have been developed to identify systematic reviews, but inclusion of the words ‘‘systematic review’’ or ‘‘meta-analysis’’ in the title may improve indexing and the yield from electronic searching. We advise authors to use informative titles that make key information easily accessible to readers. Ideally, a title reflecting the PICOS approach (Participants, Interventions, Comparators, Outcomes, and Study design) may help readers as it provides key information about the scope of the review.

padhraigfleming

Highlight

Item 2: Abstract. Include a structured abstract. Specific guidance on Abstract reporting within SRs have been developed: Beller EM et al. PRISMA for Abstracts: Reporting systematic reviews in journal and conference abstracts. PLoS Med. Apr 2013;10(4):e1001419.

84 Fleming, Johal, and Pandis

group dissolves and removes calcium ions from hydroxy-apatite, becoming incorporated in the network beforethe primer polymerizes, neutralizing the acid.

The proposed advantages of SEPs include reducedchair-side time, although this is tempered by the require-ment for judicious pumicing before bonding proceduresto minimize the risk of failure5; reduced sensitivity tomoisture; and reduced inventory requirements. How-ever, although the performance of SEPs has been com-pared with conventional acid-etch (AE) techniques inrandomized controlled trials, a comparison of thesetechniques in the context of a systematic review hasnot been undertaken.

OBJECTIVES

The aims of this study were therefore to compare1-step and 2-step bonding procedures with respect toattachment failure rates and time taken to place attach-ments.

MATERIAL AND METHODS

Protocol and registration

The protocol for a systematic review of SEPs wasregistered on the National Institute of Health ResearchDatabase (www.crd.york.ac.uk/prospero, Protocol:CRD42011001601).

Eligibility criteria

The following selection criteria were applied for thereview.

1. Study design: randomized and controlled clinicaltrials, with split-mouth designs included.

2. Participants: patients with full-arch, fixed, andbonded orthodontic appliances.

3. Interventions: SEPs were used to prepare tooth sur-faces before bonding the orthodontic attachmentsin the intervention sample. The control group's ap-pliances were bonded with the conventional, 2-stepAE technique.

4. Exclusion criteria: studies using banded attach-ments and those involving follow-up periods ofless than 12 months were omitted from the review.

5. Outcome measures: the main outcome measure wasfirst-time bond failure with both bonding systems.Secondary outcome measures included time re-quired to place individual brackets and decalcifica-tion. The attachment failures with each enamelpreparation technique were recorded. When avail-able, the time taken for failures to occur was also re-corded. The time taken to place attachments witheach technique and the presence of demineraliza-

July 2012 � Vol - � Issue - American

tion adjacent to the bonded attachments werenoted, in addition to the severity of each lesion.

Information sources, search strategy, and studyselection

The following electronic databases were searched:MEDLINE (1966 to July 2011; Appendix), EMBASE(1980 to July 2011), Cochrane Oral Health Group's TrialsRegister (March 2011), Cochrane Central Register ofControlled Trials (CENTRAL, The Cochrane LibraryIssue 2, 2011). Language restrictions were not applied.Unpublished literature was searched electronically by us-ing ClinicalTrials.gov (www.clinicaltrials.gov) and theNational Research Register (www.controlled-trials.com)with the term “orthodontic” and “bond.” In addition,the Pro-Quest Dissertation Abstracts and Thesis databasewas searched (www.lib.umi.com./dissertations) by using“orthodontic*” and “bond*.” Conference proceedingsand abstracts were also accessed when possible. Authorswere contacted to identify unpublished or ongoing clin-ical trials and to clarify data as required. Reference lists ofthe included studies were screened for relevant research.

Assessment of research for inclusion in the review,assessment of risk of bias, and extraction of data wereperformed independently and in duplicate by 2 investi-gators (P.S.F. and A.J.) who were not blinded to the au-thors or the results of the research. Disagreements wereresolved by discussion and consultation with the thirdauthor (N.P.).

Data items and collection

A data extraction form was developed to record studydesign, observation period, participants, interventions,outcomes, and outcome data of interest, including riskof failure of attachments, time taken to place attach-ments, and severity of demineralization when applica-ble.

Risk of bias/quality assessment in individualstudies

Seven criteria were analyzed to grade the risk of biasinherent in each study, including random sequence gen-eration, allocation concealment, blinding of participantsand personnel, blinding of assessors, incomplete out-come data, selective reporting of outcomes, and otherpotential sources of bias. An overall assessment of riskof bias (high, unclear, low) was made for each includedtrial by using the Cochrane Collaboration risk of biastool. Studies with at least 1 criterion designated to beat high risk of bias were regarded as having a high riskof bias overall and excluded from the meta-analysis.

Journal of Orthodontics and Dentofacial Orthopedics

http://www.crd.york.ac.uk/prospero

http://ClinicalTrials.gov

http://www.clinicaltrials.gov

http://www.controlled-trials.com

http://www.lib.umi.com./dissertations

padhraigfleming

Highlight

Item 3: Rationale. Give the rationale for the review in the context of what is known and what uncertainty exists. Explanation: Authors should clarify whether their report is new or an update of an existing review. Authors might define the importance of the review question from different perspectives (eg, public health, individual patient, or health policy) and mention the current state of knowledge and its limitations. A possible sequence might be to: -define the importance of the review question from different perspectives (eg, public health, individual patient, or health policy). -briefly mention the current state of knowledge and its limitations. -clearly state what the review aims to add. -discuss the extent to which the limitations of the existing evidence base may be overcome by the review.

padhraigfleming

Highlight

Items 10 and 11: Data collection process and data items. Describe methods of data extraction (including piloting and duplicate extraction). List variables for which data were sought, eg, PICOS and funding sources. Explanation: The authors should describe these methods, including any steps taken to reduce bias and mistakes during data collection and data extraction. Reviewers extract information from each included study so that they can critique, present, and summarize evidence in a systematic review. They might also contact authors of included studies for information that has not been reported. If individual patient data were sought, this must be described. We advise authors to describe any steps they used to avoid double counting and piece together data from multiple reports of the same study (duplicate publications). It is advised that review authors be as explicit as possible in the description of these methods, even if some of the data has to be presented in the Appendix. It is important that the reviewers describe what information was sought from each study, even if this was not available in the end. Variables that were not planned in the protocol can be added but must be described accordingly. Any assumptions made about missing or unclear data should also be reported. Authors are advised to tell readers who extracted what data, whether any extractions were completed in duplicate, and if so, whether duplicate abstraction was done independently and how disagreements were resolved. Published reports of the included studies may not provide all the information required for the review. Reviewers should describe any actions they took to seek additional information from the original researchers. The description might include how they attempted to contact researchers, what they asked for, and their success in obtaining the necessary information. Authors should also tell readers when individual patient data were sought from the original researchers and indicate the studies for which such data were used in the analyses.

padhraigfleming

Highlight

Items 7, 8, and 9: Information sources, search strategy, and study selection. Give information sources outlining the most recent search date, provide full strategy for at least one electronic database. Outline the process for selecting studies. Explanation: As a minimum, for each database searched, the database, platform, or provider (eg, Ovid, Dialog, PubMed) and the start and end dates for the search should be made clear. Authors should also report who developed and conducted the search. In addition, the use of supplementary approaches to identify studies, such as hand searching of journals, checking reference lists, searching trials registries or regulatory agency web sites, contacting manufacturers, or contacting authors, should de outlined. Authors should also report if they attempted to acquire any missing information (eg, on study methods or results) from investigators or sponsors. Authors are encouraged to report the full electronic search strategy for at least one major database. Apart from the keywords used to identify or exclude records, any additional limitations relevant to the search, such as language and date restrictions (see also Item 6, Eligibility criteria), should be reported. Methods of screening the retrieved records (typically a title and abstract) should be made clear, with detail on the number of full texts reviewed, and if any types of record (eg, letters to the editor) were excluded. Authors are encouraged to use the PRISMA flow diagram to summarize study selection processes. Efforts to enhance objectivity and avoid mistakes in study selection are important. Thus, authors should report whether each stage was carried out by one or several people, who these people were, and, when multiple independent investigators performed the selection, what the process for resolving disagreements was. The use of at least two investigators may reduce the possibility of rejecting relevant reports. The benefit may be greatest for topics where selection or rejection of an article requires difficult judgments.

padhraigfleming

Highlight

Item 6: Eligibility criteria. Provide study and report characteristics used as eligibility criteria. These should include PICOS and length of follow-up, years considered, and language and publication status. Explanation: Knowledge of the eligibility criteria is essential in appraising the validity, applicability, and comprehensiveness of a review. Carefully defined eligibility criteria influence the development of the search strategy and ensure that studies are selected in a systematic and unbiased manner.

padhraigfleming

Highlight

Item 5: Protocol and registration. Indicate whether a protocol exists and, if so, where it can be accessed Explanation: A protocol is important because it pre-specifies the objectives/methods of the review and may prevent data-dredging. Registration may possibly reduce the risk of multiple reviews addressing the same question, reduce publication bias, and provide greater transparency when updating systematic reviews. Changes to the protocol after its submission might be warranted, but selective outcome reporting is associated with bias. See the PROSPERO International Prospective Register Of Systematic Reviews (http://www.crd.york.ac.uk/PROSPERO/).

padhraigfleming

Highlight

Item 4: Objectives. Provide a clear statement of questions being addressed, ideally using a PICOS format (Participants, Interventions, Comparisons, Outcomes, Study design). The aims of this study were therefore to compare 1-step (Intervention) and 2-step (Comparison) bonding procedures with respect to attachment failure rates and time taken to place attachments (Outcomes) in patients undergoing full-arch, fixed orthodontics (Participants).

padhraigfleming

Highlight

Item 12: Risk of bias in individual studies. Describe methods used for assessing risk of bias in individual studies (including specification of whether this was done at the study or outcome level, or both), and how this information is to be used in any data synthesis. The likelihood that the treatment effect reported in a systematic review approximates the truth depends on the validity of the included studies, as certain methodological characteristics may be associated with effect sizes. Explanation: There has been a recent transition from the assessment of the “quality” of a study to the assessment of the risk of bias of a study. There is a difference between the terms quality (ie, “did the authors do their best”?) and bias (ie, “should I believe their results”?). Bias can also occur in well-conducted studies, while not all methodological flaws introduce bias. Authors should report how they assessed risk of bias: whether it was in a blind manner, whether assessments were completed by more than one person, and if so, whether they were completed independently. For RCTs, the Cochrane Risk of Bias Tool is considered the gold standard in the assessment of risk of bias. It is important that the threshold level of bias or quality permitting inclusion in quantitative assessment is defined. For nonrandomized studies a variety of assessment tools exists, including checklists, individual components, and scales (though use of the latter should be avoided).

Fleming, Johal, and Pandis 85

Summary measures and approach to synthesis

Clinical heterogeneity of the included studies wasgauged by assessing the treatment protocol—particu-larly, participants and settings, materials used, timingof data collection, and measurement techniques. Statis-tical heterogeneity was assessed by inspecting a graphicdisplay of the estimated treatment effects from the trialsin conjunction with 95% confidence intervals. The chi-square test was used to assess for heterogeneity; a Pvalue below 0.1meant significant heterogeneity.6 I2 testsfor homogeneity were undertaken to quantify the extentof heterogeneity before each meta-analysis. I2 valuesabove 50% would signify moderate to high heterogene-ity and might preclude meta-analysis. A weighted treat-ment effect was calculated, and the results forattachment failure were expressed as odds ratios. Fortime required to place attachments, mean differenceswith 95% confidence intervals were calculated for eachtrial and combined by using a random-effects model,which was considered more appropriate in view of thevariations in populations and settings. For continuousoutcomes, mean differences and standard errors wereentered for parallel and split-mouth designs. When nec-essary, standard errors for the split-mouth designs werecalculated.7

Risk of bias across studies

If more than 10 studies were included in the meta-analysis, standard funnel plots and contoured enhancedfunnel plots would be drawn to identify publicationbias.8

Additional analyses

Sensitivity analyses were prespecified to deal withstudies at higher risk of bias, publication bias, and otherpotential sources of heterogeneity including dominanteffects of at least 1 large study and differences in out-come related to specific SEPs to isolate their influenceon the overall outcome. Meta-analyses and sensitivityanalyses were undertaken using the Stata statistical soft-ware package (version 12.1; StataCorp, College Station,Tex) by using “metan” and “metainf” commands.9

RESULTS

Study selection and characteristics

Forty-eight trials were initially deemed potentiallyrelevant to the review (Fig 1). After we reviewed the ab-stracts, initially 13 satisfied the inclusion criteria.10-22

Two of these were subsequently excluded after retrievalof the full-text article because of duplicate publicationof the data20 and comparison of 2 SEPs without a controlgroup involving conventional etch preparation.21

American Journal of Orthodontics and Dentofacial Orthoped

Of the final 11 articles included in the qualitativeanalysis, all were prospective clinical trials (Tables Iand II). Although all of these were variously describedas randomized controlled trials, the randomizationprocedure was considered inadequate in 5 studies.Consequently, allocation concealment was likely tohave been subverted, thus increasing the risk of bias.These studies were excluded from the quantitativesynthesis (Tables III and IV). Of the remaining studies,4 were split-mouth designs,10-12,14 and 2 wereparallel-group randomized controlled trials.16,22

Risk of bias within studies

Of the 7 criteria used to assess risk of bias, similar re-sults were obtained throughout for 3 criteria: complete-ness of data reporting, absence of selective reporting,and blinding of assessors. In particular, complete out-come data were reported in all studies without selectivereporting of results (Tables III and IV; Fig 2). Addition-ally, blinding of assessors was considered unlikely, sincethe researchers themselves were involved in placing theappliances, precluding blinding. Blinding of assessorswas not mentioned in any reports. Nevertheless, someauthors explicitly mentioned attempts to blind the par-ticipants to the mode of bonding, although this is likelyto pertain equally to all split-mouth studies.12,15,16

Nevertheless, the binary primary outcome (bracketfailure) was not easily open to manipulation, limitingthe potential problems of lack of blinding.

Generation of the random sequencewas considered ad-equate in6 studies10-12,15,16,22; allocation concealmentwasalso thought to be reliable in 5 of these studies.7,11,13,16,17

The randomization procedure was considered inadequateor not sufficiently clear in the remaining studies.However, each of these studies was split-mouth in design,which might have negated the importance of the random-izationprocedure.Nevertheless, itwas agreed that these tri-als should be omitted from the quantitative analysis.

Therefore, overall, 6 studies were deemed to be at lowor unclear risk of bias and were initially considered ap-propriate for quantitative synthesis.10-12,15,15,22 Earlycessation was reported in 1 study because of anunacceptable number of failures with the SEP, causinga threat to validity.12 Therefore, after further appraisaland discussion, it was decided to omit this study becauseof the discordant findings resulting in a premature end tothe trial related to the excessive failure rates of up to 72%.

Results of individual studies, meta-analysis, andadditional analyses

The failure risk of attachments was assessed in all 5included studies. In total, 1721 brackets bonded with

ics July 2012 � Vol - � Issue -

padhraigfleming

Highlight

Items 20, 21, and 23: Results of individual studies, meta-analyses, and additional analyses. Present outcomes from individual studies with 95% CIs and combined estimates with CIs, ideally with a forest plot. Outline results of additional analyses, eg, sensitivity analysis or meta-regression. Explanation: Publication of summary data from individual studies allows the analyses to be reproduced and other analyses and graphical displays to be investigated. Additionally, because data extraction errors in meta-analyses are common and can be large, the presentation of the results from individual studies makes it easier to identify errors. For all included studies (and for all included outcomes, if possible) it is important to present the estimated effect with a confidence interval. This information may be incorporated in a table showing study characteristics or may be shown in a forest plot. If authors have conducted one or more meta-analyses, they should present the results as an estimated effect across studies with a confidence interval (ideally in a forest plot). Authors should also provide, for each meta-analysis, a measure of the consistency of the results from the included studies such as I-squared. If no meta-analysis was performed, the qualitative inferences should be presented as systematically as possible (eg, with a forest plot without a summary estimate) with an explanation of why meta-analysis was not done. Any deviations from the planned analysis should be reported. Information on harms is only rarely reported in systematic reviews, even when it is available in the original studies. Selective omission of harms results biases a systematic review and decreases its ability to contribute to informed decision making. Authors should report any subgroup or sensitivity analyses and whether or not they were pre-specified (Items 5 and 16). For analyses comparing subgroups of studies (eg, fixed or removable functional appliances), the authors should report any tests for interactions, as well as estimates and confidence intervals from meta-analyses within each subgroup. Similarly, meta-regression results (see Item 16) should not be limited to P values, but should include effect sizes and confidence intervals, as the first example reported above does in a table. The amount of data included in each additional analysis should be specified if different from that considered in the main analyses. This information is especially relevant for sensitivity analyses that exclude some studies; for example, those with high risk of bias. Also add GRADE assessment. Present a GRADE Summary of Findings (SoF) table. The quality of evidence may be scored as high, medium, low, or very low. Consider reasons for downgrading the quality of evidence (study limitations, indirectness, inconsistency, imprecision, and publication bias) or for upgrading the quality of evidence (magnitude of effect, dose-response effect, or absence of residual confounding). For further guidance on GRADE: 1. Guyatt GH, Oxman AD, Vist GE, et al. GRADE: an emerging consensus on rating quality of evidence and strength of recommendations. BMJ 2008 Apr 26;336(7650):924–926. 2. Guyatt G, Oxman AD, Akl EA, et al. GRADE guidelines: 1. Introduction-GRADE evidence profiles and summary of findings tables. J Clin Epidemiol 2011 Apr;64(4):383–394.

padhraigfleming

Highlight

Item 19: Risk of bias within studies. Present data on risk of bias from individual studies and outcome level assessments, if available. Explanation: Reviewers should assess the risk of bias in the included studies using a standard approach with defined criteria. The results of any such assessments should be reported in detail (ideally explicitly reporting the methodological features evaluated for each study) and not purely providing summary data (eg, ‘‘two of eight trials adequately concealed allocation’’). The Cochrane Collaboration’s new tool for assessing the risk of bias also suggests that authors substantiate these assessments with any relevant text from the original studies.

padhraigfleming

Highlight

Items 17 and 18: Study selection and characteristics. Give details of number of studies included, ideally using a PRISMA flow diagram. Provide details of characteristics recorded for each study (eg, sample size, PICOS, follow-up). Explanation: Authors should report, ideally with a flow diagram, the total number of records identified from electronic bibliographic sources (including specialized database or registry searches), hand searches of various sources, reference lists, citation indices, and experts. Authors should also note the presence of duplicate or supplementary reports so that readers understand the number of individual studies compared to the number of reports that were included in the review. For readers to gauge the validity and applicability of a systematic review’s results, they need to know details of the included studies including PICOS and specific information relevant to the review question. Authors should present enough detail to allow readers to make their own judgments about the relevance of included studies. Authors should avoid, whenever possible, assuming information when it is missing from a study report (eg, sample size, method of randomization).

padhraigfleming

Highlight

Items 13 and 14: Summary measures and Synthesis of results. State the principal summary measure (eg, mean difference, relative risk). Describe how data will be handled and results combined, eg, outlining methods of dealing with statistical heterogeneity. Explanation: The summary effect measure must be pre-specified for each of the primary/secondary outcomes. The most common summary effect measures include the risk ratio, odds ratio, or risk difference (for binary outcomes) and the mean difference or standardized mean difference (for continuous outcomes). For time-to-event outcomes, the hazard ratios is most widely used. The standardized difference in means is used when the studies do not yield directly comparable data. Usually this occurs when all studies assess the same outcome but measure it in a variety of ways (eg, different scales to measure pain). Caution should be exercised when paired or clustered designs are included in the analysis. Ignoring matching in paired designs decreases the power of the analysis. Correct calculation of the estimates requires additional information such as the correlation coefficient between within-patient measurements. For clustered designs, such as multiple observations within patients (eg, multiple bracket failures), ignoring clustering may increase precision spuriously. An adjustment for the produced standard errors is required according to the correlation within patients (intracluster correlation coefficients) unless outcomes are provided at patient and not at tooth level. Authors should report how they planned to evaluate between-study variability (heterogeneity or inconsistency). The consistency of results across trials may influence the decision to combine trial results in a meta-analysis. When meta-analysis is done, authors should specify the effect measure (eg, relative risk or mean difference), the statistical method (eg, inverse variance), and whether a fixed- or random-effects approach or some other method (eg, Bayesian) was used. Ideally the statistical analysis should be pre-specified (before data extraction) and be based on clinical and statistical reasoning.

padhraigfleming

Highlight

Item 16: Additional analyses. Describe any additional analyses, eg, sensitivity or subgroup analyses. Explanation: Authors may perform additional analyses to help understand whether the results of their review are robust, all of which should be reported. Such analyses include sensitivity analysis, subgroup analysis, and meta-regression. It is important to inform readers whether these analyses were performed, their rationale, and which were pre-specified. Sensitivity analyses are used to explore the degree to which the main findings of a systematic review are affected by changes in its methods or in the data used from individual studies (eg, study inclusion criteria, results of risk of bias assessment). Subgroup analyses address whether the summary effects vary in relation to specific (usually clinical) characteristics of the included studies or their participants. Meta-regression extends the idea of subgroup analysis to the examination of the quantitative influence of study characteristics on the effect size. Meta-regression also allows authors to examine the contribution of different variables to the heterogeneity in study findings. Readers of systematic reviews should be aware that meta-regression has many limitations, including a danger of overinterpretation of findings. Even with limited data, additional analyses are possible; however, the appropriate analysis depends on the aims of the review.

padhraigfleming

Highlight

Item 15: Risk of bias across studies. Outline methods of assessing publication bias or selective reporting. Explanation: Reviewers may examine results from the available studies for clues that suggest that there may be missing studies (publication bias) or missing data from the included studies (selective reporting bias). Authors should report in detail any methods used to investigate possible bias across studies. Also refer to GRADE assessment The GRADE assessment allows assessment of the quality of evidence and subsequent recommendations: http://www.gradeworkinggroup.org/publications/jce_series.htm. Reference: Guyatt GH, Oxman AD, Vist GE, et al. GRADE: an emerging consensus on rating quality of evidence and strength of recommendations. BMJ 2008 Apr 26;336:924–926.

Fig 1. PRISMA diagram of article retrieval.


AE and 1723 bonded with SEP techniques were in-cluded in the quantitative synthesis (Table V). Of these,4.5% (77 brackets) and 6.0% (104 brackets) failed withthe AE and the SEP preparation techniques, respec-tively. The random-effects model assumes that thereare different bond failure risks in different settings;the calculated estimate therefore indicates the averageeffect. Meta-analysis of these studies suggested higherodds of bond failures with the SEP technique, althoughthe difference failed to reach statistical significance(Table VI; Fig 3; odds ratio, 1.35; 95% CI, 0.99-1.83).The pooled odds ratio from the random-effects modelindicated that the failure risk was 35% higher in theSEP group than in the AE group. The 95% confidenceinterval indicates that the mean effect size can rangefrom 1% to 83% in the SEP group compared with theAE group, verging on statistical significance (P 50.06). Based on the heterogeneity of the included stud-ies, the prediction intervals indicate that the true effect


size is likely to range from 0.82 to 2.22. The predictioninterval was wider than the 95% confidence intervaland includes the value 1, indicating that in certain set-tings no difference is expected in bond failures with theprotocols. The test for homogeneity confirmed thatmeta-analysis of this outcome among the 5 studieswas reasonable (I2, 0.0%; chi-square, P 5 0.497;t2 5 0.00).

A further meta-analysis was undertaken to gauge theinclusion of the study by House et al12 on the outcome.The results did not change significantly, with the pro-pensity to higher failure rates with SEP remaining statis-tically insignificant; however, the 95% confidenceinterval increased (0.92-2.9). Statistically, heterogeneityalso increased to an unacceptable level (I2, 78.3%; chi-square, P\0.001; Fig 4).

Little heterogeneity was observed, with confidenceintervals overlapping and the effects of individual stud-ies exclusively favoring the AE technique, with the


padhraigfleming

Sticky Note

Fig 1. PRISMA flow diagram showing the process of identifying and selecting eligible studies for the systematic review. Explanation: The flow diagram is divided in 5 parts: (1) the identification of studies from the literature search and the removal of duplicates, (2) the exclusion of ineligible studies on the basis of the title/abstract, together with reasons, (3) the exclusion of ineligible studies on the basis of the full-text, together with reasons, (4) the inclusion of studies in the qualitative synthesis (systematic review), and (5) the inclusion of studies in the quantitative synthesis (meta-analysis), if appropriate. The flow diagram is an integral part of transparent reporting for systematic reviews, enabling a quick overview of the review’s procedures.

Table I. Design, observation period, interventions, and outcome measures of studies included in the quantitative syn-thesis

Study MethodObservation

period Participants Interventions OutcomeAljubouri et al(2004)10

Split-mouthRCT

6 and 12 months 51 participants: 16 male,35 female32\15 years,19 .15 years

389 brackets bondedwith SEP, 388 bondedwith AE (353 paired pergroup)

Bond failureriskBondingtime

Manning et al(2006)11

Parallel-groupRCT

6 and 12 months;overall treatment

34 participants,17 per group: 11 male,23 female. Ages,11-16 years

299 brackets bondedwith SEP and 298 with AE

Bond failurerisk

House et al(2006)12

Split-mouthRCT

1, 6, and 12 months 30 participants: only 20were analyzed becausetrial stopped prematurely

339 brackets bondedwith Ideal 1 SEP and AE

Bond failurerisk

Murfitt et al(2006)15

Split-mouthRCT

12 months 39 participants: 13 male,26 female. Mean age,14.4 (SD, 2.5) years

661 brackets bonded overallwith SEP (331) and AE (330)

Bond failurerisk

Banks andThiruvenkatachari(2008)16

Parallel-groupRCT

Overall treatment 60 participants, 30 per group:23 male, 37 female. Ages,11-18 years

30 participants (438 brackets)with TransBond Plus SEP;30 participants (433 brackets)with AE

Bond failurerisk

Cal-Neto et al(2009)22

Parallel-groupRCT

12 months 28 participants, 14 per group:Mean age, 14.92 years;11 male, 17 female

276 brackets bonded withSEP and 272 bracketswith AE

Bond failurerisk

RCT, Randomized controlled trial.

Table II. Design, observation period, interventions, and outcome measures of studies excluded from the quantitativesynthesis

Study MethodObservationperiods Participants Interventions Outcome

Pandis et al(2006)13

Split-mouthRCT

12 months 62 participants: 23 male,39 female. Mean age,14 years

610 brackets bonded withTransBond Plus SEP,610 bonded with AEand OrthoSolo primer

Bond failure risk


Split-mouthRCT

15 months 62 participants: 23 male,39 female. Mean age,13.7 years

221 molar tubes bonded withTransBond Plus SEP, 223molar tubes bonded withAE and OrthoSolo primer

Bond failure risk

Reis et al(2008)17

Split-mouthRCT

18 months 30 participants: 15 male,15 female. Ages,12-18 years

283 brackets bonded withSEP and 283 with AE

Bond failure risk

Elekdag-Turk et al(2008)18

Split-mouthRCT

6 and 12months

39 participants: 23 male,39 female. Mean age,15.58 years

344 brackets bonded withSEP and 344 with AE

Bond failure risk

Ghiz et al(2009)19

Split-mouthRCT

18 to 24months

25 participants.No demographics given

236 brackets bonded withSEP and 233 brackets with AE

Demineralization

RCT, Randomized controlled trial.


exception of the study of Aljubouri et al,10 who reporteda lower risk of failure with SEP. Sensitivity analysis inves-tigating the influence of this study on the overall meta-analysis resulted in estimates favoring AE further (Fig 5).Statistical analysis of publication bias was not indicated,


since fewer than 10 studies were included in the quanti-tative synthesis.

Time taken to place individual attachments with ei-ther technique was considered in 2 investigations. Sim-ilar results were obtained in both studies, with Aljubouri


Table III. Risk of bias of studies included in the quantitative synthesis, including the study by House et al12

Trial

Randomsequencegeneration

Allocationconcealment

Blindingparticipantsand personnel

Blindingassessor

Free ofincomplete

outcome data

Free ofselectivereporting

Free ofother threatsto validity

Aljubouri et al(2004)10

Low Low Unclear Unclear Low Low Low


Low Unclear Unclear Unclear Low Low Low

House et al(2006)12

Low Low Low Unclear Low Low High


Low Low Low Unclear Low Low Low

Banks andThirvenkatachari(2008)16

Low Low Low Unclear Low Low Low

Cal-Neto et al (2009)22 Low Low Unclear Unclear Low Low Low

Table IV. Risk of bias of studies excluded from the quantitative synthesis

Trial

Randomsequencegeneration

Allocationconcealment

Blindingparticipantsand personnel

Blindingassessor

Free ofincomplete

outcome data

Free ofselectivereporting

Free ofother threatsto validity


High High Unclear Unclear Low Low Low



Reis et al(2008)17


Elekdag-Turk et al(2008)18


Ghiz et al(2009)19



et al10 highlighting a reduction in bonding time of 24.9seconds (95% CI, 22.1-27.7) per attachment. Banks andThiruvenkatachari16 highlighted a mean reduction of22.2 seconds (95% CI, 21.1-23.3) per tooth with Trans-bond Plus. Quantitative analysis of these studies showeda pooled mean reduction of 23.2 seconds per tooth (95%CI, 20.7-25.8) with the 1-step approach (Fig 6), a statis-tically significant finding (P\0.001). The elevated sta-tistical heterogeneity (I2, 68.2%; chi-square, P 5 0.08;t2 5 2.48) should be interpreted with caution, becauseit is related to the lack of studies and the artificially nar-row confidence intervals, since large numbers of teethartificially inflate the precision of the estimates. Thehigh I2 value indicates that, although the observed var-iance is real, estimates lie in a narrow range.7

Risk of bias across studies

Tests for publication bias were not undertaken as nomore than 6 studies were included in an individual meta-analysis.


DISCUSSION

Summary of evidence

Relative to other systematic reviews in orthodontics,this review identified many studies with a potentiallylow risk of bias, permitting meta-analysis. Five studieswere included in the meta-analyses; however, of these,only 1 study dealt with the duration of treatment in itsentirety.16 Scrutiny of the total duration of treatmenthas been advocated in previous reviews to ascertainthe influence of long-term alterations in bond strengthand to provide a more complete assessment of the per-formance of bonding materials.23 Therefore, additionalstudies encompassing a complete course of treatmentwould be desirable to produce more robust conclusionsfrom future research.

The impact of bias on the outcome of systematicreviews is well documented.24 The preponderance ofsplit-mouth research in our review complicated therisk of bias assessment, since we were unable to identifyspecific guidelines relating to the handling of these


padhraigfleming

Highlight

Item 24: Summary of evidence. Summarize the main findings including the strength of the evidence. Relate the findings to key groups. Explanation: Authors should give a brief and balanced summary of the nature and findings of the review. The applicability of the findings to different patients, settings, or target audiences, for example, should be mentioned. Sometimes, authors formally rate or assess the overall body of evidence addressed in the review and can present the strength of their summary recommendations tied to their assessments of the quality of evidence (eg, the GRADE system). Authors need to keep in mind that statistical significance of the effects does not always suggest clinical or policy relevance. Patients and health care providers may be primarily interested in which intervention is most likely to provide a benefit with acceptable harms, while policy makers and administrators may value data on organizational impact and resource utilization.

padhraigfleming

Highlight

Item 22: Risk of bias across studies. Present results of any assessments for publication bias or selective reporting. Explanation: Authors should present the results of any assessments of risk of bias across studies. If a funnel plot is reported, authors should specify the effect estimate and measure of precision used. Authors should describe if and how they have tested the statistical significance of any possible asymmetry. Authors are advised to tell readers if any pre-specified analyses for assessing risk of bias across studies were not completed and the reasons (eg, lack of included studies).

Fig 2. Risk of bias summary outlining judgment of risk ofbias items for studies included in the quantitative synthe-sis, including the study by House et al.12


reports. Although it might be reasonable to assume thatrobust random allocation can be less important in split-mouth research, inherent bias might prompt differenthandling of the appliances; for example, a decisioncould be made to use 1 technique in a less crowdedquadrant or to partially ligate a tooth to prevent attach-ment failure if there is a preference for a particularbonding technique. Therefore, we thought that it wasreasonable to conclude that the omission of robust ran-dom allocation procedures would lead to an unaccept-able risk of bias.

Split-mouth studies offer the advantages of concur-rent experimental and control assignment, limiting


sample-size requirements and increasing precision,and avoiding period effects obviating the need fora “washout period” that can be necessary with analo-gous crossover designs. For the bonding time outcome,we were able to calculate standard errors from theinformation given to adjust for the matching withinpatients in the split-mouth studies. However, due to in-sufficient information, it was not possible to accountfor the matching effects within patients for bond fail-ure estimation. Reporting the details of the 2 3 2 tablefor matched pairs within patients permits calculation ofthe desired estimates, confidence intervals, andvariance or correlation between pairs, by using eitherthe Mantel-Haenszel or conditional likelihood ap-proaches.25 The calculation of the correlation can beused to appropriately adjust the standard errors toaccount for matching.

A separate and opposite problem in split-mouth re-search stems from the nesting of teeth in patients andquadrants, producing clustering effects. Clustering ofoutcome measurements can be managed by applyingspecific statistical methods accounting for the correlateddata, in which either a summary outcome measurementis calculated for each cluster followed by simple statisti-cal tests or by using complex hierarchical regressionmodels for correlated data such as generalized estimat-ing equations or random effects.26-28 Incorrecttreatment of clustered observations as independentmight result in smaller standard errors andconsequently artificially small P values, increasing thechance of false-positive results. Of the 3 split-mouthstudies included in the quantitative synthesis, 2 incor-porated or discussed statistical adjustments to mitigatethis problem.10,15 Meta-analysis of clustered designswould ideally require knowledge of a measure of thecorrelation of the data, such as the intracluster correla-tion coefficient or the coefficient of variation betweenclusters. This would be used to appropriately adjustthe sample size of the constituent studies. This measurewas not reported in these studies and would have neces-sitated the availability of the entire data sets for separateintracluster correlation coefficient or coefficient of var-iation values to be calculated for each trial. Acquisitionof the complete data sets was not attempted; dependingon the within-cluster correlation, this might have artifi-cially altered the results toward the null. Nevertheless, itis important that further studies account for this prob-lem at the outset to facilitate recruitment of a sufficientsample to confer the desired level of power. Addition-ally, it is important that intracluster correlation coeffi-cient values are reported to allow future investigatorsto perform sample-size calculations and adjustmentsof standard errors for the purposes of meta-analyses.29


padhraigfleming

Sticky Note

Fig 2. Summary of the risk of bias for the studies (in this instance RCTs) included in the systematic review. Explanation: According to the Cochrane risk of bias tool, the risk of bias for each RCT is judged based on seven domains (vertical columns) as “low risk of bias” (green plus sign), as “high risk of bias” (red minus sign), or as “unclear risk of bias” (yellow question mark). This figure enables the reader to quickly identify problematic aspects of the included studies that could introduce systematic errors (bias) in the trials’ results.

Table V. Bond failure risk and time taken to place attachments reported in the included studies

StudyIntervention

(number of attachments)

Bond failures (%) Time (s)

AE SEP AE SEPAljubouri et al(2004)10

SEP (389), AE (388):353 paired pergroup reducingto 312 at 12 months

11 (3.1) 6 (1.6) Mean, 106.6Mean difference,24.9 (95% CI,22.1-27.7)

Mean, 81.7


SEP (299), AE (298) 22 (7.4) 21 (7)

House et al(2006)12

Ideal 1 SEP (339),AE (339)

25 (14.8) 123 (72.4)


SEP (331), AE (330) 13 (3.9) 37 (11.2)

Banks andThiruvenkatachari(2008)16

SEP 30 participants(438), AE 30 participants(433)

15 (3.5) 21 (4.8) 97.7 (SD, 9.1; 95% CI,94.3-101.2)

75.5 (SD, 6.7; 95% CI,72.9-78.5)

Cal-Neto et al (2009)22 SEP (276), AE (272) 13 (4.8) 19 (6.9)

Table VI. Summary of findings (SoF) table according to GRADE. Number of bonded brackets (participants), effectestimates, quality of the evidence, and expected bond failures per 1000 brackets bonded with SEP and AE

SEP compared with AE for orthodontic patientsPatient or population: orthodontic patientsSettings: variousIntervention: SEPComparison: AE

Outcomes

Illustrative comparative risks (95% CI)

Relative effect(95% CI)

Number of participants(studies)

Quality of the evidence(GRADE)

Assumed risk Corresponding risk

AE SEPBond failures 45 per 1000 59 per 1000 (44-79) OR 1.35 (0.99-1.83) 3445 (5 studies) 4444 High

High quality (indicated by4): further research is unlikely to change our confidence in the estimate of effect.Moderate quality: further research islikely to have an important impact on our confidence in the estimate of effect and might change the estimate. Low quality: further research is likelyto have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: we are uncertainabout the estimate. The corresponding risk (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of theintervention (and its 95% CI).GRADE, Working group grades of evidence; OR, odds ratio.


The randomized controlled trial by House et al12

presented a further dilemma. This trial was initiallyadjudged to have low or unclear risk of bias andwas well reported. However, because of the excessivenumber of failures in the SEP arm, the premature endof the trial, and the use of a different bonding mate-rial, it was regarded to be at significant odds with allthe other studies. The closest failure risk to the72.4% reported for Ideal-1 in that study was just11.2%.15 The particular system investigated wasalso in its infancy, having been subject to concurrentin-vitro investigation by the same research group.30

It was, therefore, agreed to omit this trial from the


quantitative analysis to prevent skewing the results.Even with the inclusion of this study, the directionof the results remained the same; however, thedegree of statistical heterogeneity increased signifi-cantly, making amalgamation of the data question-able.

The higher failure rate with SEPs was partially offsetby a reduction in chair-side time with this technique. Themagnitude of the time savings was relatively small (23.2seconds on average). This difference translates to a re-duction of 8 minutes overall during placement ofa dual-arch appliance. The time saving encounteredwith SEPs is counterbalanced by the increased likelihood


Fig 3. Random-effects meta-analysis of bracket failure with SEP and AE.

Fig 4. Random-effects meta-analysis of bracket failure with SEP and AE including the study of Houseet al.12


of failure of attachments with this system, with unsched-uled replacement of each additional attachment on anemergency basis likely to necessitate a 5 to 10 minuteappointment. Practitioners should also consider price


differences between individual agents in conjunctionwith the implications of each agent on chair-side timewhen assessing the economic advantages of 1- or 2-step systems.


padhraigfleming

Sticky Note

Fig 3. Forest plot presenting the meta-analysis results for the main outcome of the systematic review (bracket failure). Explanation: The forest plot can be divided in three regions: (1) the list of included studies, together with their raw data (here on the left side), (2) the graphical representation of each study’s results, together with the overall summary (here in the middle), and (3) the calculated effect size (here the odds ratio [OR ratio] and the 95% confidence Interval [95% CI]) and the assigned weight for each study and for the overall summary. For each of the included studies (gray boxes with dots), if its 95% CI (black horizontal line) crosses the vertical no-effect line, then P value > 0.05 for this study. The overall summary (meta-analysis result) is given as a hollow blue diamond under the included studies, which includes both the pooled effect estimate (pooled OR) and the 95% CI. If the diamond crosses the vertical no-effect line, then P value > 0.05 for the pooled meta-analysis estimate. The blue horizontal lines of the diamond indicate the 95% predictive interval (95% PI) for the pooled estimate, not the 95% CI. The forest plot is integral to the transparent reporting of meta-analyses, as it presents the magnitude of the treatment effect for each study, which treatment group is favored, the homogeneity among the included studies, and the overall pooled estimate.

Fig 5. Sensitivity analysis to investigate the influence of individual studies on the overall meta-analysisestimate. The graph shows the pooled estimate (open circle) and its confidence interval (dotted hori-zontal line with vertical breaks) as trials were sequentially excluded from the meta-analysis. For exam-ple, the first open circle and the dotted range at the top of the graph indicate the pooled estimate afterthe exclusion of the study of Aljubouri et al.10

Fig 6. Random-effects meta-analysis of required time to bond with SEP and AE.


Limitations

It is accepted that bond failure can be influencedby a range of factors including demographics,operator experience, tooth location, tooth surface


preparation, handling of attachments, and tooth sur-face both before and after placement. A confoundingeffect of these variables was not demonstrated in 1study.16 Manning et al,11 however, demonstrated


padhraigfleming

Highlight

Item 25: Limitations. Discuss limitations at study and outcome level (eg, risk of bias, lack of relevant studies) and at review level (eg, incomplete retrieval of studies). Explanation: A discussion of limitations should address the validity (ie, risk of bias) and reporting (informativeness) of the included studies, limitations of the review process, and generalizability (applicability) of the review. Readers may find it helpful if authors discuss whether studies were threatened by serious risks of bias, whether the estimates of the effect of the intervention are too imprecise, or if there were missing data for many participants or important outcomes. Limitations of the review process might include limitations of the search (eg, restricting to English-language publications) and any difficulties in the study selection, appraisal, and meta-analysis processes.

padhraigfleming

Sticky Note

Fig 5. Forest plot presenting the sensitivity analysis for the meta-analysis of the main outcome. Explanation: In this kind of sensitivity analysis a single study is excluded for every line from the overall pooled summary (as indicated to the left), and the results are presented as pooled OR (open circle) and the corresponding 95% CI (dotted horizontal line). This analysis helps assess the relative contribution of each study to the overall pooled estimate. The more a single open circle deviates from the others, the more this study contributes to the meta-analysis.

padhraigfleming

Sticky Note

Fig 6. Forest plot presenting the meta-analysis results for the secondary outcome of the systematic review (time to bond).


a higher failure risk in the maxillary arch, whereasCal-Neto et al22 highlighted greater failures of pre-molar than anterior attachments. Murfitt et al15 re-ported increased risk of failure in male patients.The robust application of selection criteria, random-ization procedures, and allocation concealment willreduce the impact of these potential confounders onthe results. When a significant imbalance is noted, mul-tivariate statistical models can be used to offset thesedifferences. The premium on cleaning teeth beforebondingwith SEPs has previously been demonstrated.31

Enamel preparation was carried out in all included stud-ies with pumice slurry11,12,15,16,22 or prophylaxispaste,10 eliminating tooth surface preparation beforebonding procedures as a significant confounder in thisresearch.

Differences in archwire sequences might also havea bearing on bracket failures. Although use of identicalarchwire sequences throughout treatment is both im-practical and unlikely to be sanctioned by ethical reviewcommittees, standardized archwire sequences wouldideally be used during this type of research to limit con-founding effects. Different initial aligning wires (either0.014-in nickel-titanium or stainless steel) were usedin 1 study based on the degree of initial crowding.15

Use of similar archwire sequences over a 12-month pe-riod was alluded to by Aljubouri et al10; similar sequenceswere also used by Banks and Thiruvenkatachari,16

whereas in the trial by Cal-Neto et al,22 0.014-innickel-titanium wires were used for initial aligning ineach patient.

We had hoped to analyze the effect of bonding mo-dality on the risk of decalcification during treatment.However, only 1 study considering this eventuality wasidentified.19 This study lacked information on randomi-zation and allocation concealment and also failed toaccount for clustering. Although information on con-founders—in particular, plaque accumulation—was ob-tained, the statistical analysis did not account for theinteraction of these variables. Therefore, it was unclearwhether enamel preparation techniques have a demon-strable effect on demineralization of enamel. Further re-search on this aspect of enamel preparation techniqueswould be welcome.

CONCLUSIONS

On the basis of this review, we concluded thefollowing.

1. Weak but statistically insignificant evidence sug-gests that the odds of attachment failures differ be-tween SEP and AE orthodontic bonding techniquesover a minimum period of 12 months.


2. Use of 1-step bonding techniques is likely to resultin a modest time saving compared with 2-stagetechniques.

3. Additional high-quality randomized controlled tri-als investigating the overall course of treatmentare required to analyze the effect of bonding modal-ity on demineralization around fixed appliances.

4. In the absence of clear evidence to favor either sys-tem, the choice of bonding modality remains at thediscretion of each operator.

ACKNOWLEDGMENTS

The authors report no commercial, proprietary, orfinancial interest in the products or companies describedin this article.

REFERENCES

1. Bowen RL. Use of epoxy resins in restorative materials. J Dent Res1956;35:360-9.

2. Buonocore MG, Matsui A, Gwinnett AJ. Penetration of resin dentalmaterials into enamel surfaces with reference to bonding. ArchOral Biol 1968;13:61-70.

3. Newman GV, Snyder WH, Wilson CE Jr. Acrylic adhesives for bond-ing attachments to tooth surfaces. Angle Orthod 1968;38:12-8.

4. Eliades T. Orthodontic materials research and applications: part 1.Current status and projected future developments in bonding andadhesives. Am J Orthod Dentofacial Orthop 2006;130:445-51.

5. Keim RG, Gottlieb EL, Nelson AH, Vogels DS. 2008 JCO study oforthodontic diagnosis and treatment procedures. Part 1. Resultsand trends. J Clin Orthod 2008;32:625-41.

6. Higgins JP, Thompson SG, Deeks JJ, Altman DG.Measuring incon-sistency in meta-analyses. BMJ 2003;327:557-60.

7. Borenstein M, Hedges LV, Higgis JPT, Rothstein HR, editors. Intro-duction to meta-analysis. Chichester, United Kingdom: JohnWiley& Sons; 2009.

8. Sterne JAC, Egger M, Moher D. Addressing reporting biases.Cochrane handbook for systematic reviews of intervention. Version5.1.0. Chichester, UK: Cochrane Collaboration; 2011. Available at:http://www.cochrane-handbook.org/. Accessed September 20,2011.

9. Tobias A. Assessing the influence of a single study in the meta-analysis estimate. Stata Technical Bulletin. College Station, Tex:StataCorp; 1999. p. 47.

10. Aljubouri YD, Millett DT, Gilmour WH. Six and 12 months' evalu-ation of a self-etching primer versus two-stage etch and prime fororthodontic bonding: a randomized clinical trial. Eur J Orthod2004;26:565-71.

11. Manning N, Chadwick SM, Plunkett D, Macfarlane TV. A random-ized clinical trial comparing ‘one-step’ and ‘two-step’ orthodonticbonding systems. J Orthod 2006;33:276-3.

12. House K, Ireland AJ, Sherriff M. An investigation into the use ofa single component self-etching primer adhesive system for ortho-dontic bonding: a randomized controlled clinical trial. J Orthod2006;33:38-44.

13. Pandis N, Polychronopoulou A, Eliades T. Failure rate of self-ligating and edgewise brackets bonded with conventional acidetching and a self-etching primer: a prospective in vivo study. An-gle Orthod 2006;76:119-22.


http://www.cochrane-handbook.org/

padhraigfleming

Highlight

Item 26: Conclusions. Provide a general interpretation of the results in the context of other evidence together with implications for future research. Explanation: Conclusions drawn from systematic reviews should be explicitly formulated (not overly positive) and should consider both harms and benefits. If conclusions cannot be drawn because there are too few reliable studies or too much uncertainty, this should be stated. Authors are also advised to make clear recommendations for future research.

padhraigfleming

Highlight

Item 27: Funding. Describe sources of funding and other support (eg, provision of data).


14. Pandis N, Polychronopoulou A, Eliades T. A comparative assessmentof the failure rate of molar tubes bonded with a self-etching primerand conventional acid-etching. World J Orthod 2006;7:41-4.

15. Murfitt PG, Quick AN, Swain MV, Herbison GP. A randomised clin-ical trial to investigate bond failure rates using a self-etchingprimer. Eur J Orthod 2006;28:444-9.

16. Banks P, Thiruvenkatachari B. Long-term clinical evaluation ofbracket failure with a self-etching primer: a randomized controlledtrial. J Orthod 2007;34:243-51.

17. Reis A, dos Santos JE, Loguercio AD, de Oliveira Bauer JR. Eigh-teen-month bracket survival rate: conventional versus self-etchadhesive. Eur J Orthod 2008;30:94-9.

18. Elekdag-Turk S, Cakmak F, Isci D, Turk T. 12-month self-ligatingbracket failure rate with a self-etching primer. Angle Orthod 2008;78:1095-100.

19. Ghiz MA, Ngan P, Kao E, Martin C, Gunel E. Effects of sealant andself-etching primer on enamel decalcification. Part II: an in-vivostudy. Am J Orthod Dentofacial Orthop 2009;135:206-13.

20. Shah J,ChadwickS.Comparisonof1-stageorthodonticbonding sys-tems and 2-stage bonding systems: a review of the literature and theresults of a randomized clinical trial. Orthod Fr 2009;80:167-78.

21. Paschos E, Kurochkina N, Huth KC, Hansson CS, Rudzki-Janson I.Failure rate of brackets bonded with antimicrobial and fluoride-releasing, self-etching primer and the effect on prevention ofenamel demineralization. Am J Orthod Dentofacial Orthop 2009;135:613-20.


22. Cal-Neto JP, Quint~ao CA, Almeida MA, Miguel JA. Bond failurerates with a self-etching primer: a randomized controlled trial.Am J Orthod Dentofacial Orthop 2009;135:782-6.

23. Millett DT, Glenny AM, Mattick CR, Hickman J, Mandall NA. Adhe-sives for fixed orthodontic bands. Cochrane Database Syst Rev2007;18:CD004485.

24. Millett D. Bias in systematic reviews? J Orthod 2011;38:158-60.25. Elbourne DR, Altman DG. Meta-analyses involving cross-over tri-

als: methodological issues. Int J Epidemiol 2002;31:140-9.26. Donner A, Banting D. The statistical analysis of site-specific data in

dental studies. J Dent Res 1988;67:1392-5.27. Donner A, Eliasziw M. Application of matched pair procedures to

site-specific data in periodontal research. J Clin Periodontol1991;18:755-9.

28. Donner A, Klar N, Zou G. Methods for the statistical analysis of bi-nary data in split-cluster designs. Biometrics 2004;60:919-25.

29. Campbell MK, Elbourne DR, Altman DG, CONSORT group. CON-SORT statement: extension to cluster randomised trials. BMJ2004;20:702-8.

30. House K, Ireland AJ, Sherriff M. An in-vitro investigation into theuse of a single component self-etching primer adhesive systemfor orthodontic bonding: a pilot study. J Orthod 2006;33:116-24.

31. Lill DJ, Lindauer SJ, T€ufekci E, Shroff B. Importance of pumiceprophylaxis for bonding with self-etch primer. Am J Orthod Den-tofacial Orthop 2008;133:423-6.


APPENDIX

MEDLINE SEARCH STRATEGY VIA OVID

1. RANDOMIZED CONTROLLED TRIAL.pt. (311155)2. CONTROLLED CLINICAL TRIAL.pt. (82831)3. RANDOMIZED CONTROLLED TRIALS.sh. (0)4. RANDOM ALLOCATION.sh. (72051)5. DOUBLE BLIND METHOD.sh. (111370)6. SINGLE BLIND METHOD.sh. (15222)7. or/1-6 (459019)8. (ANIMALS not HUMANS).sh. (3533433)9. 7 not 8 (420281)

10. CLINICAL TRIAL.pt. (464598)11. exp Clinical Trial/ (647582)12. (clin$ adj25 trial$).ti,ab. (192365)13. ((singl$ or doubl$ or trebl$ or tripl$) adj25 (blind$

or mask$)).ti,ab. (112153)14. PLACEBOS.sh. (29877)15. placebo$.ti,ab. (130730)16. random$.ti,ab. (524593)17. RESEARCH DESIGN.sh. (63258)18. or/10-17 (1113684)19. 18 not 8 (1027643)20. 19 not 9 (622360)21. 9 or 20 (1042641)22. exp ORTHODONTICS/ (39565)23. orthod$.mp. (43453)24. 22 or 23 (49508)25. (self-etch$or self-etching$or SEP$or self-adhesive$

or single component$).mp. (743767)26. (primer$ or bonding agent$).mp. (146371)27. 25 and 24 and 26 (342)28. 27 and 21 (154)


American Journal of Orthodontics and Dentofacial Orthopedics July 2012 � Vol - � Issue -

Documents

Self-etch primers and conventional acid-etch technique for ...cdn.elsevier.com/promis_misc/YMOD_Model_Orthodontic_Systematic_Review.pdfLondon, United Kingdom, Corfu, Greece, and Bern,