Registration and Protocol

The protocol for this systematic review was registered on the National Institutes of Health Research Database under the registration number CRD42020171507. The review is reported according to the PRISMA statement (Preferred Reporting Items for Systematic Reviews and Meta-analyses).¹³

Eligibility Criteria

Information Sources, Search Strategy, Study Selection

A search of relevant literature was conducted by two reviewers (HW and GF) independently in the Cochrane Library, PubMed, Web of Science, and Embase databases from the earliest available date to November 20, 2021. Related orthodontic and adhesive journals and the reference list in each retrieved study were manually searched for additional relevant studies. In addition, gray literature was searched in the National Research Register, and Open Grey with the term “flash-free.” Specific search strategies for each database are presented in Table 1.

Table 1. Database Search Strategy

View Fullscreen

The eligibility of identified studies was initially checked by screening their titles and abstracts in Endnote. Potentially eligible articles were read in full text and judged against the inclusion/exclusion criteria for a final judgment by the two authors (HW and GF). Any disagreements about study inclusion were resolved through discussion.

Data Collection Process and Data Items

The data extraction process was conducted independently by the two authors (HW and GF) with standardized sheets in Word. Any disagreements were resolved through discussion. When any problems arose, the study investigators were contacted for additional details (every 2 weeks for 2 months).

Information collected from the included studies consisted of publication details, participants, control groups, outcomes measured, study design, enamel pretreatment, and study duration.

Study Risk of Bias Assessment

The Cochrane scale was employed (RoB 2.0 tool).¹⁴ The risk of bias was considered low, some concerns, or high for each assessed criterion. The assessment was conducted independently by the two authors (HW and GF), and disagreements were resolved through discussion. When agreements could not be reached through discussion, a third author's opinion (JS) was considered.

Effect Measures and Synthesis Methods

If it was considered not appropriate to make a quantitative synthesis of studies, a qualitative synthesis was performed. As the bracket bonding in studies was performed by different operators with inevitable methodological differences, a priori choice of a random-effects model was reasonable to account for between-study variance. To conduct the meta-analyses, for continuous data, the means with their corresponding standard deviations (SD) and sample sizes of each outcome were statistically pooled to calculate the mean difference (MD) with corresponding 95% confidence intervals (CI) using the inverse variance method. For dichotomous data, the number of events and the sample sizes were pooled together to calculate the risk ratio (RR) with corresponding 95% CIs using the inverse variance method. Heterogeneity across studies was assessed with the I² statistic.

Sensitivity analysis was conducted to evaluate the robustness of the overall results by the one study removal method. The sensitivity analysis and meta-analysis were implemented by Review Manager 5.3 (Cochrane Collaboration).

Reporting Bias Assessment

The Egger test and Begg test were used to assess publication bias.

Certainty Assessment

The certainty of evidence of each outcome was appraised using GRADEpro Guideline Development Tool¹⁵ according to the study design, risk of bias, inconsistency, indirectness, imprecision, and other considerations. The assessment was conducted independently by the two authors (HW and GF), and disagreements were resolved through discussion.

Study Selection and Characteristics

The database search identified 143 articles, and no articles were identified through additional sources. After the removal of duplicates, 78 articles were reviewed with titles and abstracts, and 55 of them were excluded. The remaining 23 articles were regarded as potentially eligible. After reading the full texts, 17 studies were excluded and six studies were finally included in the review (Figure 1).

View Figure

• Download as Powerpoint
• Download Figure

The main characteristics of the selected studies are presented in Table 2. The studies included were published from 2018 to 2020.^16–21 All of them were randomized controlled trials. Among the studies included, one study made a comparison between flash-free and OPC,²¹ while the others made comparisons between flash-free and APC.^16–20 Two studies compared the bonding time and adhesive remnant index (ARI),^16,18 three reported bracket failure rates^6–18 and three^19–21 reported enamel demineralization and periodontal status. The 17 excluded studies with reasons are shown in Table 3.

Table 2. Main Characteristics of Included Clinical Studies

View Fullscreen

Table 3. 17 Excluded Studies With Reasons at the Full-Text Phase

View Fullscreen

Risk of Bias in Studies

Three studies^16,17,19 were judged to have a low risk of bias, and the other studies^18,20,21 were judged to have some concerns because of lack of information about selection bias, performance bias, and detection bias (Table 4).

Table 4. Risk of Bias of the Included Studies

View Fullscreen

Results of Syntheses

A qualitative analysis was conducted for three studies^19–21 investigating enamel demineralization and periodontal status with methodological variation. Two studies^20,21 reported no significant differences in enamel demineralization between flash-free and conventional bonding systems, while the study by Almosa et al.¹⁹ reported that the mean values of demineralization were significantly decreased in flash-free compared to APC. Tan et al.²¹ reported that the effects of flash-free and OPC on periodontal health did not differ from each other, while Yetkiner et al.²⁰ reported that, though the quantity of plaque around flash-free and APC did not differ, less pathogenic bacteria were detected in the plaque around flash-free.

The meta-analysis results showed that the flash-free significantly reduced bonding time compared to APC (mean difference [MD]: −1.56; 95% CI: −2.56 to −0.56; P = .002, Figure 2), and no significant difference was found in failure rates (RR: 1.54; 95% CI: 0.27 to 8.89; P = .63, Figure 3) and ARI (MD: −0.50; 95% CI: −1.14 to 0.14; P = .13, Figure 4) between them. The results of failure rates in two articles by Grünheid et al.^16,17 were from the same research, so only the article in 2018¹⁶ was included in the meta-analysis of failure rates.

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

Substantial heterogeneity existed among the studies included in the meta-analyses of bonding time and failure rates. However, the studies presented a consistent direction of effect in the meta-analyses. Subgroup analyses and sensitivity analyses could not be performed because only two studies were included in each meta-analysis.

Reporting Biases

Publication bias was not estimated because of the limited number of studies included in the meta-analysis. No selective reporting was found within studies.

Certainty of Evidence

Low-to-moderate certainty of evidence for the outcomes is expected. Evidence was mainly downgraded because of shortcomings in the risk of bias and inconsistency of some studies (Table 5).

Table 5. GRADE Assessment for Certainty of Evidence^a

View Fullscreen

Summary of Evidence

The problem of flash has been a challenge concerning orthodontists for a long time. Careful and complete removal of flash is time-consuming and any flash left may cause the development of enamel demineralization and periodontal inflammation with increased plaque accumulation.⁶ The flash-free bonding system was claimed to solve this problem well. However, its actual bonding performance was questioned because of the modifications in its structure and constituents.²² Thus, this review and meta-analysis was conducted to compare flash-free with conventional bonding systems comprehensively. The review included six studies,^16–21 all of which were RCTs. A critical appraisal of the trials using the Cochrane RoB 2.0 tool found that three studies^18,20,21 were judged to have some concerns. The overall assessment of the evidence was rated as low to moderate according to Grade due to limitations in risk of bias and inconsistency.

According to the results of meta-analysis, flash-free was superior to the APC bonding system in requiring less bonding time. This result was expected since the procedure of excess adhesive removal was eliminated in the flash-free bracket bonding process and, although the elimination of this procedure saves only a few seconds per tooth, it makes sense in clinic because cumulative time savings add up to significant differences when considering a full-mouth application.¹⁷ Bracket bonding is probably the longest appointment during orthodontic treatment and reduced chair time can make work more efficient and improve patient satisfaction.¹⁶

Bond strength performance is one of the most important characteristics of an adhesive, which could be evaluated by clinical bond failure rates. Some considered that the nonwoven mesh at the bracket base, as a new design feature of the flash-free system, had a lower material density and might affect the overall bond strength of the adhesive.¹¹ However, according to the results of meta-analysis, no statistically significant difference in bond failure rates was found between flash-free and APC. Additionally, the bond failure rate values of flash-free reported by the two studies were both clinically acceptable. These results indicated that flash-free probably had comparable bond strength to APC. This was consistent with the results of direct measurements of shear bond strength in vitro, which displayed no significant difference between flash-free and conventional bonding systems.^5,8,9,23

It is noteworthy that substantial heterogeneity existed among the studies included in the meta-analyses of bonding time and failure rates. According to the Cochrane handbook,¹⁴ the included articles will inevitably differ in a meta-analysis, especially in experimental methodology. Any kind of variation between studies in a meta-analysis is called heterogeneity. Considering the methodological differences among the studies in the meta-analyses, such heterogeneity was inevitable. However, despite considerable heterogeneity, all studies included showed a consistent direction of effect in the meta-analysis that, to some extent, supported the conclusion. Existing heterogeneity is a reminder that the results should be treated with caution, and further well-designed research with methodological rigor is deemed necessary in the future.

ARI was evaluated to compare the bonding performance of flash-free with conventional adhesives comprehensively. Though not precise, ARI can give a rough indication of where bond failure happens. According to the results of meta-analysis, no differences in ARI were found between flash-free and APC on brackets that failed clinically. This suggested that bond failures of flash-free did not happen more frequently at the bracket-adhesive interface than APC, which further demonstrated that the nonwoven mesh at the bracket base did not weaken the bond strength of flash-free. In summary, the available evidence suggested that the novel mesh structure in flash-free might not have a negative influence on bond strength and that the clinical bond failure rates of the flash-free and APC were not significantly different.

The study by Almosa et al. found that enamel demineralization around the flash-free brackets was significantly reduced compared with that of the APC.¹⁹ However, the studies by Tan et al.²¹ and Yetkiner et al.²⁰ reported that no significant differences of demineralization were found between flash-free and conventional adhesives; however, they only made an estimation of demineralization degree using relatively rough methods,^20,21 while Almosa et al.¹⁹ sectioned the teeth and measured the demineralization depth accurately under scanning electron microscope (SEM), which may be considered more reliable. Less demineralization of flash-free adhesive could be attributed to the presence of less excess adhesive on smooth surfaces. A previous study reported that flash-free had significantly lower excess adhesive measurements than APC.¹⁰ Additionally, further in vitro examination by SEM described that the excess adhesive of flash-free displayed a smooth, nontextured surface, with the adhesive spreading out and conforming to the enamel surface, while APC and OPC presented a ruffled surface with a more irregular transition from adhesive to enamel.¹¹

Tan et al.²¹ reported that the effects of flash-free and OPC on periodontal health were not different from each other. However, Yetkiner et al.²⁰ reported that, although the quantity of plaque on flash-free and conventional brackets was not different, the constituents of plaque differed, with less pathogenic bacteria detected around flash-free brackets.

Limitations

The main limitation of this review was the small number of studies included. This was probably due to the fact that the flash-free bonding system is a novel adhesive product and lacks a large number of original studies, which is a problem that all meta-analyses concerning new materials face.

Though three included studies^8,20,21 reported the same outcome for enamel demineralization, meta-analysis of them could not be conducted because of their methodological and outcome measurement differences. Tan et al.²¹ and Yetkiner et al.²⁰ detected demineralization with the DIAGNOdent Pen (Kavo, Biberach, Germany) or quantitative light-induced fluorescence imaging,^20,21 while Almosa et al.¹⁹ measured the demineralization depth under SEM.

For risk-of-bias assessment, three studies^18,20,21 were judged to have some concerns because of lack of information in bias arising from randomization, bias due to deviations from intended interventions, and bias in measurement of outcome. Thus, it is critical that future RCTs have more uniform study designs and methodological rigor.