What Is the Most Effective Technique for Bonding Brackets on Ceramic—A Systematic Review and Meta-Analysis

Background: There has been an increase in demand for orthodontic treatment within the adult population, who likely receive restorative treatments using ceramic structures. The current state of the art regarding the most effective method to achieve an appropriate bond strength of brackets on ceramic surfaces isn’t consensual. This systematic review aims to compare the available surface treatments to ceramics and determine the one that allows to obtain the best bond strength. Methods: This systematic review followed the PRISMA guidelines and the PICO methodology was used, with the question “What is the most effective technique for bonding brackets on ceramic crowns or veneers?”. The research was carried out in PubMed, Web of Science, Embase and Cochrane Library databases. In vitro and ex vivo studies were included. The methodological quality was evaluated using the guidelines for reporting of preclinical studies on dental materials by Faggion Jr. Results: A total of 655 articles searched in various databases were initially scrutinized. Sevety one articles were chosen for quality analysis. The risk of bias was considered medium to high in most studies. The use of hydrofluoric acid (HF), silane and laser afforded the overall best results. HF and HF plus laser achieved significantly highest bond strength scores in felsdphatic porcelain, while laser was the best treatment in lithium disilicate ceramics. Conclusions: The most effective technique for bonding brackets on ceramic is dependent on the type of ceramic.


Introduction
In recent years there has been an increase in demand for orthodontic treatment within the adult population. As of 2015, according to the American Association of Orthodontics, the demand within this age group has doubled over a four year period and this number is set to increase further in the future [1]. This can be attributed not only to evergrowing aesthetic concerns [2] but also to the expeditious evolution of orthodontic techniques [1]. In this age group, there is a high likelihood that an orthodontist will encounter complex The results are described in detail in Table 2. The sample size (n) ranged from 8 to 960, obtaining a total sample of n = 7246. The final selection of studies was 64 in vitro, 5 ex vivo e 2 in vitro/ex vivo, from 2011 to 2021.
All the articles evaluated various methods of conditioning the ceramic surface to obtain an adequate bond strength when bonding brackets. The types of adhesion technique mostly present in the included articles are application of orthophosphoric acid or hydrofluoric acid in various concentrations, silane application, sandblasting/air abrasion with aluminum oxide or silicon dioxide, diamond bur roughening, single bond universal adhesive and the application of different types of lasers such as Er:YAG laser, CO2 laser, Er:CrYSGG laser, Nd:YAG laser, Cr:YSGG laser, FS laser.
Regarding the type of brackets, metallic, ceramic, polycarbonate, sapphire, and zirconia brackets were included.
All articles used shear bond test for the application of force, except for one study that used tensile strength test [19] and another one that used the adhesion strength test [20].

Risk of Bias
The results of the quality assessment of the in vitro studies included are reported in Figure 2. The results are described in detail in Table 2. The sample size (n) ranged from 8 to 960, obtaining a total sample of n = 7246. The final selection of studies was 64 in vitro, 5 ex vivo e 2 in vitro/ex vivo, from 2011 to 2021.
All the articles evaluated various methods of conditioning the ceramic surface to obtain an adequate bond strength when bonding brackets. The types of adhesion technique mostly present in the included articles are application of orthophosphoric acid or hydrofluoric acid in various concentrations, silane application, sandblasting/air abrasion with aluminum oxide or silicon dioxide, diamond bur roughening, single bond universal adhesive and the application of different types of lasers such as Er:YAG laser, CO 2 laser, Er:CrYSGG laser, Nd:YAG laser, Cr:YSGG laser, FS laser.
Regarding the type of brackets, metallic, ceramic, polycarbonate, sapphire, and zirconia brackets were included.
All articles used shear bond test for the application of force, except for one study that used tensile strength test [19] and another one that used the adhesion strength test [20].

Risk of Bias
The results of the quality assessment of the in vitro studies included are reported in Figure 2. G4 produced maximum bond strength comparable or even better than the control group followed by G3 and G5. G2 produced least SBST and hence not suitable for bonding Orthodontic brackets in a clinical scenario.
Dilber et al., 2016 [22] In vitro Three surface conditioning methods: G1: fine diamond burr; G2: fine diamond burr + air abrasion with 30 µm SiO 2 + silane G3: fine diamond burr +9.5% HF acid + silane Specimens were mechanically roughened with fine diamond burrs placed with their shafts parallel to the specimen axes. Then, they were washed and rinsed thoroughly to remove the debris, and air-dried
All CAD/CAM materials tested benefitted from additional surface conditioning either with HF acid or silica coating and silanization; Weibull parameters indicated more reliable adhesion of metal brackets to feldspathic ceramic when their graze was removed with fine diamond bur and then conditioned with either hydrofluoric acid or silica coating followed by silanization compared to those of other material conditioning combinations; In group 6 (control), the buccal tube was positioned directly on the untreated ceramic surface only using the luting composite, which was polymerized by light curing (n = 10)

Metallic SBST
The highest mean value of SBST was examined in group 1 (61.56 MPa), followed by group iii (45.53 MPa), group 2 (41.65 MPa), and group 4 (23.14 MPa). The comparison between groups 1-4 (with coupling agent) and group 5 (without coupling agent) revealed statistically significant differences (p ≤ 0.002), with the exception of the comparison between groups 4 and 5. Within groups 1-4, statistically significant results were determined between groups 1 and 4 as well as between groups 3 and 4 (p < 0.001). The SBST of group 6 was not calculated as the buccal tubes debonded after the incubation period.
A suitable coupling agent system produced clinically acceptable shear bond strengths capable of withstanding orthodontic forces.
Kurt et al., 2019 [24] In vitro G1: HF acid 9,6% for 2 min + silane; G2: Sandblasting with Al 2 O 3 applied from a distance of 10 mm for 10 s in circling motions at 2.5 bar pressure + silane; G3: Silica coating with cojet under 2.5 bar pressure, at a 10-mm distance for 10 s + silane; G4: Roughening with diamond burr at 40     Four primer groups (n = 20 per group), and each primer was divided into two subgroups (n = 10 each) to examine by thermocycling protocols.
Surface treatment with a zirconia primer increases the SBST relative to no-primer or silane primer application between orthodontic brackets and zirconia prostheses.
Ihsan et al., 2019 [33] In vitro G1: transbondtm XT primer; G2: single bond universal adhesive for 20 s, and also air dried for 5 s, and then light cured for 10 s; G3: theracem, was done in the same way as described with the previous groups except that no priming or bonding agent to the zirconia surfaces was needed according to manufacturer instructions.

Zirconia 30
Single bond universal adhesive group (n = 10); Theracem group (n = 10).   There were significant effects on SBST of metal bracket to the ceramic veneering materials due to the factor of different types of ceramic materials, surface treatment, resin bonding materials, interaction between types of ceramic materials, and types of adhesive resin cement (p < 0.05). The mean SBST of metal bracket bonded to vitablocs™ mark II was higher than bonded to IPS e.max ® CAD and bonded to IPS d.SIGN ® porcelain (p < 0.05). The mean SBST of metal bracket bonded to IPS d.SIGN ® porcelain for PFM was significant lower than the mean SBST of metal bracket bonded to vitablocs™ mark II ceramic materials (p < 0.05). Also, the mean SBST of metal bracket bonded to IPS e.max ® CAD ceramic reveals significantly lower than the mean SBST of metal bracket bonded to vitablocs™ mark II ceramic materials (p < 0.05).
Etching ceramic surface enhanced ceramic-bracket bond strength. However, bond strengths in nontreated ceramic surface groups were still higher than bond strength required for bonding in orthodontic treatment. Metallic SBST There were statistically significant differences among groups (p = 0.002).
Er:YAG laser application did not allow for elimination of the hydrofluoric acid step. Gonçalves et al., 2011 [38] In vitro G1: 10% hydrofluoric acid for 20 s + silane; G2: 10% hydrofluoric acid for 60 s + silane (after application of silane on the ceramic surface, metallic brackets were bonded to the cylinders using Transbond XT).

Feldspathic ceramic 60
The specimens for each etching time were assigned to four groups (n = 15), according to the light source: xl2500 halogen light, Ultralume 5 LED, Acucure 3000 argon laser, and Apollo 95e plasma arc. Light-activation was carried out with total exposure times of 40, 40, 20 and 12 s, respectively.

Lithium disilicate 40
The specimens were randomly assigned to two experimental groups (n = 20), G1 specimens were treated with two-step surface conditioning system (IPS ceramic etching gel™ and Monobond plus™) and G2 specimens were treated with one-step surface conditioning system (Monobond etch and prime™). Ten randomly selected specimens from each group were subjected to thermo-cycling and the remaining ten served as baseline.

N = 20 NR SBST
The specimens treated with two-step conditioning system had higher bond strength than one-step conditioning system.
Traditional two-step conditioning provides better bond strength. The clinical importance of the study is that, the silane promoted adhesion significantly reduces on exposure to thermo-cycling. Franz

NR Metallic SBST
In all groups, the mean SBST values were statistically significantly lower (p < 0.001) after thermocycling than before.
Furthermore, specimens in groups S@SB2 and S@SBU, both of which had silane pretreatment, showed statistically significantly higher mean SBST values than did the corresponding groups without silane pretreatment (p < 0.05).

Feldspathic porcelain 100
The specimens were divided into four equal groups (n = 25). Porcelain surfaces were conditioned with different protocols. In G1, hydrofluoric acid and Embrace First-Coat primer were used. G2, hydrofluoric acid and silane were utilized. G3 and G4, sandblasting with aluminum oxide powder was done instead of etching.

NR Metallic SBST
Embrace First-Coat and silane exhibited a comparable SBST. The sandblasting process significantly increased SBST.
No significant difference was found in bond SBST utilizing either hydrofluoric acid and Embrace First-Coat or sandblasting and silane. With regard to CSBS, the use of sandblasting and Embrace First-Coat revealed the highest significant CSBS value, followed by sandblasting and silane. Etching with hydrofluoric acid prior to application of either primer exhibited the least CSBS values; however, no significant difference was found between them. The SBST was significantly higher than CSBS.
Embrace First-Coat primer could be used successfully as an alternative to silane. Sandblasting provides higher bond strength than did hydrofluoric acid. Cyclic loading significantly decreased bond strength.    Two groups of 90 specimens, according to the primer used. Each group was further divided into three subgroups according to the surface treatment to be received, thus there were 6 study groups; three with 3acryloxypropyltrimethoxysilane (ACPS) silane primer, namely 1a (pretreatment with hydrofluoric acid, HF), 1b (pretreatment with grit-blasting) and 1c (pretreatment with tribochemical silica-coating) and 3 with a novel silane system (ACPS + bis-1,2(triethoxysilyl) ethane (BTSE)) assigned as 2a (HF), 2b (grit-blast), and 2c (tribochemical silica coating).

NR NR SBST
The highest SBST at baseline (26.8 + 1.7 MPa) and after thermocycling (24.6 + 1.7 MPa) was observed in group 2c, and the lowest (9.6 + 1.5 MPa and 4.5 + 1.1 MPa) was found in G1a.  Deglazing combined with HF etching produced the highest bond strength, but CO 2 laser irradiation provided adequate bond strength and allowed for elimination of the HF step. Deglazing is not recommended as a preliminary step before CO 2 laser conditioning.

NR NR AST
ANOVA showed a significant influence of the grit-blasting distance, silane blend and artificial aging on the shear bond strength values (p < 0.05).
The highest adhesion strengths were obtained for baseline specimens irrespective of the grit-blasting distance or the silane primer blend system used.
Grit-blasting at 10 mm and silane primer blend of 1.0 vol. % 3-MPS and 0.5 vol. % BTSE provided acceptable orthodontic bonding with least surface damage to zirconia surface. Adhesion strength values significantly decreased following thermo-cycling, irrespective of the grit-blasting distance and the silane primer blend system used.    Feldspathic porcelain 56 n = 14-universal adhesive; n = 14-universal adhesive/silane; n = 14-conventional adhesive; n = 14-conventional adhesive/silane.
SBST of bracket to porcelain mainly depends on the use of silane rather than the type of adhesive. Both universal and conventional adhesives yield significantly higher SBST in the presence of silane compared to that in the absence of silane  SBST of the lithium silicate infused with zirconia groups were significantly less than the chemically pre-treated lithium disilicate group, however both materials, when chemical pre-treatment protocol was used, were not statistically different than the enamel control.
Orthodontic bonding to lithium silicate infused with zirconia yielded a weaker shear bond strength than bonding to traditional lithium disilicate, however, when the surface was pretreated with hydrofluoric acid etch it provides a bond strength that is within an acceptable clinical range.
Ceramic         CO-SBU showed the highest shear bond strength. Sandblasting with either AL or CO improved the mechanical bonding by increasing the surface area, and all primer groups showed clinically acceptable increase of SBST for orthodontic treatment.

NR SBST
The heat-treated showed statistically significant higher bond strength in all the sub-groups, and the acid-neutralized samples showed higher bond strength using both types of cement; however, the increase was statistically significant only in resin-based cement-bonded samples. Resin-based cement-bonded samples showed higher bond strength than water-based cement-bonded samples.
Pre-etching heat treatment and post-etching acid neutralization of the cementing surface of lithium disilicate glass-ceramic significantly improve the initial bond strength to orthodontic brackets.   Both dental materials may be recommended for orthodontic bracket bonding to ceramic surfaces, with equally successful results. However, further testing on an increased number of specimens may be considered for more accurate data.

NR SBST
Changed as the most in ceramic in laser irradiation. Bonding strength according to the etching method was the most in laser irradiation and acid etching in ceramic and in zirconia.
Ceramic crown with acid treatment was recommended because of relatively high in bonding strength.
Hellak et al., 2016 [77] In vitro/Ex vivo  Metallic SBST Significant differences in SBST were found between the control group and experimental groups.
Transbond XT showed the highest SBST on human enamel. Scotchbond Universal on average provides the best bonding on all other types of surface (metal, composite, and porcelain), with no need for additional primers. It might therefore be helpful for simplifying bonding in orthodontic procedures on restorative materials in patients. G1: nonglazed zirconia treated with SB + ZP (n = 10); G2: glazed zirconia treated with SB + etching + ZP (n = 10); G3: glazed zirconia treated with SB + etching + PP (n = 10); G4: glazed zirconia treated with SB + etching + ZP + PP (n = 10).

NR Metallic SBST
Group G2 showed significantly lower shear bond strength than did the other groups. No statistically significant differences were found among groups G1, G3, and G4.
Porcelain primer is the more appropriate choice for bonding a metal bracket to the surface of a full-contour glazed zirconia crown with resin cement.  Only two studies not reported a structured abstract, calculation of the sample size [59,75] or scientific background and rationale [38,76]. Regarding the randomization process, only two studies reported these items [4,23,47]. All studies not reported researcher blinding to the interventions. Y-yes; N-no. Only a few studies reported the estimated size of outcomes [5,7,27,30,46]. No studies reported information relative to the protocol domain, except for three [15,43,74].

Meta-Analysis
For the quantitative analysis, only studies that used metallic brackets adhered to felspathic ceramics and lithium disilicate were selected. These studies were pooled regarding the main surface treatment used, although different protocols (concentrations, applications times, energies . . . ) were used. Studies that presented other bracket types presented highly heterogeneous methodologies, making impossible its comparison. Also, regarding the other ceramic types, it was not possible to find studies with similar methodologies to be compared.
The meta-analysis regarding the feldspathic ceramics ( Figure 3) presents the lower adhesion values for the treatments with fine bur (T1) and orthophosphoric acid (T3), without statistically significant differences between them, but significantly lower than all other treatments (p < 0.001). With increased adhesion values the sandblasting technique alone (T2), presents statistically significant differences (p < 0.001) for all groups, including the sandblasting + hydrofluoric acid group (T6), although less significant (p < 0.05). The group that uses LASER (T5) for surface preparation presents the following highest adhesion value with statistically significant differences (p < 0.001) T1, T2, T3, T4 and T7 groups and p < 0.05 to T5 group. The highest adhesion values were found in the LASER with hydrofluoric acid (T7) or hydrofluoric acid alone (T4) groups, without statistically significant differences between them, but being significantly higher than the others (p < 0.001).
Only two studies not reported a structured abstract, calculation of the sample size [59,75] or scientific background and rationale [38,76]. Regarding the randomization process, only two studies reported these items [4,23,47]. All studies not reported researcher blinding to the interventions. Y-yes; N-no. Only a few studies reported the estimated size of outcomes [5,7,27,30,46]. No studies reported information relative to the protocol domain, except for three [15,43,74].

Meta-Analysis
For the quantitative analysis, only studies that used metallic brackets adhered to felspathic ceramics and lithium disilicate were selected. These studies were pooled regarding the main surface treatment used, although different protocols (concentrations, applications times, energies…) were used. Studies that presented other bracket types presented highly heterogeneous methodologies, making impossible its comparison. Also, regarding the other ceramic types, it was not possible to find studies with similar methodologies to be compared.
The meta-analysis regarding the feldspathic ceramics ( Figure 3) presents the lower adhesion values for the treatments with fine bur (T1) and orthophosphoric acid (T3), without statistically significant differences between them, but significantly lower than all other treatments (p < 0.001). With increased adhesion values the sandblasting technique alone (T2), presents statistically significant differences (p < 0.001) for all groups, including the sandblasting + hydrofluoric acid group (T6), although less significant (p < 0.05). The group that uses LASER (T5) for surface preparation presents the following highest adhesion value with statistically significant differences (p < 0.001) T1, T2, T3, T4 and T7 groups and p < 0.05 to T5 group. The highest adhesion values were found in the LASER with hydrofluoric acid (T7) or hydrofluoric acid alone (T4) groups, without statistically significant differences between them, but being significantly higher than the others (p < 0.001). The meta-analysis that evaluates lithium disilicate ceramics (Figure 4) presents the statistically significant lowest adhesion values for the orthophosphoric acid (T3) group (p < 0.001). Still with low adhesion values, but higher than the previous ones, we find the fine bur group (T1), with statistically significant differences regarding all the other groups (p < 0.001). With increased adhesion values, we have the sandblasting technique (T2) and the hydrofluoric acid alone (T4) groups, without statistically significant differences between them, but with statistically significant differences (p < 0.001) with all other groups. The highest adhesion values are found in the LASER alone group (T5), with statistically significant differences from all other groups (p < 0.001). The meta-analysis that evaluates lithium disilicate ceramics (Figure 4) presents the statistically significant lowest adhesion values for the orthophosphoric acid (T3) group (p < 0.001). Still with low adhesion values, but higher than the previous ones, we find the fine bur group (T1), with statistically significant differences regarding all the other groups (p < 0.001). With increased adhesion values, we have the sandblasting technique (T2) and the hydrofluoric acid alone (T4) groups, without statistically significant differences between them, but with statistically significant differences (p < 0.001) with all other groups. The highest adhesion values are found in the LASER alone group (T5), with statistically significant differences from all other groups (p < 0.001). For the two ceramic types evaluated in the meta-analysis, the surface presenting the lowest results is the orthophosphoric acid, with adhesion values close to 0 MPa, such as 3.99 MPa ± 0.48 for felspathic ceramics and 0.7 MPa ± 0.07 for lithium disilicate. These low adhesion results are also observed in surface treatments using only fine drill wear, with 5 MPa ± 0.51 and 6.9 MPa ± 0.91; and sandblasting with 9.13 MPa ± 0.97 and 9.

Discussion
The main purpose of this review was to identify the most efficient and reliable bonding protocol for orthodontic brackets to ceramic surfaces. As this is a complex and sensitive process it is essential to determine the best protocol to achieve the best results [2,4,10,12].
The last systematic review regarding this topic was published in 2014. This previous paper, that solely included in vitro studies, concluded that the best protocol would be etching with 9.6% hydrofluoric acid for 60 s, rinsing for 30 s, air-drying, and finally applying the silane [78]. With new articles emerging in recent years a new systematic review is warranted. Since we included papers published from 2011, all recent literature was scrutinized and included if relevant.
As previously stated, to ensure an acceptable shear bond strength (SBS) capable of resisting not only chewing but also forces induced by orthodontic appliances, optimal ceramic surface conditioning techniques are necessary. The present results revealed that the most studied conditioning methods include 37%/37.5% orthophosphoric acid, 4%/9%/9.5%/9.6%/10% hydrofluoric acid, silane application, sandblasting/air abrasion with aluminum oxide, diamond bur roughening, single bond universal adhesive and the use of different types of LASER, such as Er:YAG laser, CO2 laser, Er:CrYSGG laser, Nd:YAG laser, Cr:YSGG laser, FS laser.

Design and Bracket Material
The included studies present several different combinations of ceramic surface conditioning techniques to understand which one achieves a better SBS value. Some studies For the two ceramic types evaluated in the meta-analysis, the surface presenting the lowest results is the orthophosphoric acid, with adhesion values close to 0 MPa, such as 3.99 MPa ± 0.48 for felspathic ceramics and 0.7 MPa ± 0.07 for lithium disilicate. These low adhesion results are also observed in surface treatments using only fine drill wear, with 5 MPa ± 0.51 and 6.9 MPa ± 0.91; and sandblasting with 9.13 MPa ± 0.97 and 9.7 MPa ± 1.05 for feldspathic ceramics and lithium disilicate respectively.
The treatment with the highest values for lithium disilicate ceramics is the LASER treatment with 19.87 MPa ± 2.01, while for feldspathic ceramics it is the LASER treatment with hydrofluoric acid with 26.79 MPa ± 2.7 and the treatment with hydrofluoric acid alone with 27.32 MPa ± 2.89.
When comparing the same surface treatments on the two types of ceramics, substantially different adhesion values are obtained, as an example of hydrofluoric acid with such different performances as 27.32 MPa ± 2.89 for feldspathic and 9.18 MPa± 1.05 for disilicate. The LASER treatment also presents some differences when we compare feldspathic ceramics with lithium disilicate with 13.56 MPa ± 1.38 and 19.87 MPa ± 2.01, respectively.

Discussion
The main purpose of this review was to identify the most efficient and reliable bonding protocol for orthodontic brackets to ceramic surfaces. As this is a complex and sensitive process it is essential to determine the best protocol to achieve the best results [2,4,10,12].
The last systematic review regarding this topic was published in 2014. This previous paper, that solely included in vitro studies, concluded that the best protocol would be etching with 9.6% hydrofluoric acid for 60 s, rinsing for 30 s, air-drying, and finally applying the silane [78]. With new articles emerging in recent years a new systematic review is warranted. Since we included papers published from 2011, all recent literature was scrutinized and included if relevant.
As previously stated, to ensure an acceptable shear bond strength (SBS) capable of resisting not only chewing but also forces induced by orthodontic appliances, optimal ceramic surface conditioning techniques are necessary. The present results revealed that the most studied conditioning methods include 37%/37.5% orthophosphoric acid, 4%/9%/9.5%/9.6%/10% hydrofluoric acid, silane application, sandblasting/air abrasion with aluminum oxide, diamond bur roughening, single bond universal adhesive and the use of different types of LASER, such as Er:YAG laser, CO 2 laser, Er:CrYSGG laser, Nd:YAG laser, Cr:YSGG laser, FS laser.

Design and Bracket Material
The included studies present several different combinations of ceramic surface conditioning techniques to understand which one achieves a better SBS value. Some studies prove that although the ceramic surface conditioning method is the most important factor in achieving acceptable clinical values for SBS, it is not exclusive. Factors such as the material and design of the bracket, type of ceramic surface, and etch time also affect SBS. Mehmeti et al. states that the bracket type used significantly affects the SBS value and is a valid clinical concern [57]. On the other hand, Guida et al. showed that the failure rate is closely related to the glass-ceramic surface conditioning and that the bracket type is inconsequential [73]. According to Mehmeti et al., metallic brackets seemingly provide stronger adhesion with all-zirconium surfaces when compared to ceramic polycrystalline brackets, which can be attributed to their improved base surface design [59]. However, this is opposed to the findings of Al-Hity et al. which revealed that bonding strength of ceramic brackets on porcelain significantly exceeds that of metal brackets [19]. Different testing protocols and materials used can explain the contradictory results, since these two factors have a profound impact on the obtained results.

Orthophosphoric Acid, Fine Burr and Sandblasting
In our systematic analysis, the lowest adhesion values were verified with orthophosphoric acid, fine burr, and with slightly higher values, sandblasting treatments. Although these treatments created microroughness that could improve adhesion, their use alone presented unsatisfactory results. According to three authors (Mohammed et al., Mehta et al. and Girish et al.), the sandblasting method in association with the application of silane reaches the maximum SBS, while the use of 37% orthophosphoric acid has the lowest SBST and is deemed unsuitable for bonding ceramic brackets [21,27,31]. In this situation, we can attribute the good SBS scores to the use of silane, which alone presents high bond strength forces.
Other studies, regarding surface roughening revealed that the use of sandblasting or diamond burs along with the application of hydrofluoric acid significantly improved bond strength [52]. Sandblasting with SiO 2 was shown to have no advantage when compared to sandblasting with AL 2 O 3 [70].

Hydrofluoric Acid
The etching process partially dissolves the ceramic matrix, increasing the surface area by creating microchannels, this allows for the penetration of resin cement, thus providing finer conditions for increased bond strength.
However, since the available brands of porcelain have dissimilar particle sizes and crystalline structure, different outcomes are to be expected when testing various ceramic surfaces and brands. The heterogeneity of the reviewed studies can be attributed to structural differences in porcelain surfaces (besides the brackets' base designs), which may result in higher or lower bond strength. As example, a paper by Kurt et al. published in 2019, reported that the highest SBS value was found in feldspathic ceramics previously treated with hydrofluoric acid [24]; however, Saraç et al. demonstrated that for any conditioning method, leucite-reinforced ceramic, in general, showed a higher SBS when compared to feldspathic and fluoroapatite ceramics [47].
As stated above, the etching agent HF increases the available surface area for adhesion. Higher HF concentrations promote more ceramic dissolution, which may be linked to higher bond strength values [79]. Such results support the use of HF as surface treatments when bonding ceramic restorations [80]. This can explain the results obtained in the feldspathic ceramics group, where the HF groups (alone or in combination with a laser) presented higher adhesion values. However, the HF promoted significantly lower adhesion values in the disilicate lithium group. Lithium silicate is more susceptible to HF action than feldspathic. HF concentrations above 5% used for more than 20 s significantly influence the characteristics of the material, promoting a decrease in the material strength [81]. Additionally, higher HF concentrations can also result in worse adhesion, as shown in an in vitro study by Pérez et al. [82].
The use of HF also produces insoluble fluorosilicate salts that remain on the material's surface (if not removed by other methods, such as ultrasonic cleaning), which can affect the adhesion [83]. Also, the overall reduced number of studies included for this material and the different experimental methodologies used can affect the observed results. Taken together, such factors and differences in the material composition regarding feldsphatic ceramics can explain the obtained values for the disilicate lithium group.
Also, the acid etching time was inconsistent as different studies used different methodologies. According to Falkensammer et al. this factor is not preponderant for achieving SBS, according to their study an etching time of 30 s was as effective as standard conditioning (60 s) [70]. However, Costa et al. revealed that an etching time of 60 s significantly improved the SBS of brackets to feldspathic ceramic surfaces [34].

Silane
The use of silane improves the bond strength of brackets to ceramic surfaces [23,67]. Silane forms chemical bonds with both organic and inorganic surfaces, resulting in a stronger connection between surfaces. Furthermore, Zhang et al. reported that HF acid etching followed by silane was the best suited method for bonding on silica based ceramics and, according to Tahmasbi et al. SBS of bracket to porcelain mainly relies on the use of silane rather than the type of adhesive chosen [9,25].

Adhesive System
The chosen adhesive protocol will influence the bond strength of brackets to ceramic surfaces. According to the results of the studies reviewed, ceramic surfaces treated with blasting aluminum oxide followed by Single Bond Universal™ application had an improved SBS and caused less cohesive damage to the ceramic [51].

LASER
Recent publications studied alternatives that involve irradiating the ceramic surface with different laser types. The bond strength obtained through the combination of Er:YAG laser and HF acid on the ceramic surface may be sufficient for bonding brackets [28]. Also, according to Cevik et al. hydrofluoric acid and phosphoric acid etching methods were not suitable as surface treatment methods for feldspathic porcelains [17]. Contrarily, other studies revealed that the Er:YAG laser with the recommended settings (intensity and duration) is not a suitable alternative to the application of HF, however the laser Nd: YAG has been shown more promising results [30,65].
The results of this systematic review indicated that laser irradiation and/or HFetching are the two surface treatments that allow greater resin-ceramic bonding. Laser irradiation emits a wavelength which is absorbed by ceramic materials, creating microretentions which improve resin-ceramic bonding [84]. Feitosa et al. compared 5 types of surface treatment and have found that Er:YAG laser promotes higher surface roughness, producing an improvement in the tensile strength. Regarding laser application time, these authors suggested times greater than 5 s, since some regions on the laser-treated surface had a similar morphologic appearance to the control group [85]. An article published in 2013 compared fractional CO 2 laser with different intensities with hydrofluoric acid, showing that 10 and 15 W laser were higher shear bond strength than HF-etching with better results in deglazed specimens [29]. More recently, Mirhashemi et al. suggested that laser combined with HF promotes higher shear bond strength than laser groups only [30].
In lithium disilicate ceramic crowns, the results revealed that irradiation with different types of lasers can be effective in obtaining an adequate SBS. Conditioning with Er,Cr:YSGG and CO 2 laser has the potential to be used in clinical settings alternative to HF+S when bonding to metallic brackets [66]. However, contrary to the previously mentioned statements, the study by Alavi et al. concluded that neither CO 2 nor Nd:YAG lasers resulted in adequate surface changes for bonding ceramic brackets when compared to conditioned samples with HF [16]. This is also confirmed by Mirhashemi et al. who demonstrated that although conditioning with Er:CrYSGG met SBS requirements for orthodontic brackets, the SBS must be improved through refinement of the irradiation details [30]. Regarding zirco-nia crowns, FS laser at 200 mW and 60 µm is ideal treatment for conditioning, producing good SBS while also having a more sustainable energy consumption [53].
Importantly, no studies regarding the combined use of HF with laser (T7) included lithium disilicate ceramics, so we cannot ascertain if high bond values similar to the ones observed in the feldspathic ceramics could be obtained, or if the ceramic type is a decisive factor, like for the HF treatment.
Due to the lack of homogeneity in methodology within the currently available literature investigating the bond strength of orthodontic brackets to ceramic surfaces, the present review results present some limitations. To overcome this, calibrated studies analyzing the same parameters using the same protocols should be performed, hence providing stronger evidence. Further research focusing on surface changes, the architecture of the bracket base and the type of the adhesive resin should be performed.

Conclusions
Surface treatment protocols cannot be universal for all ceramic and/or all bracket types. Based on our results, we can conclude that for felspathic ceramics, the surface treatment which provides the best adhesion values is the use of hydrofluoric acid alone or concomitantly with LASER. For lithium disilicate ceramics, the treatment with the best results is the use of LASER alone, although combination with HF was not evaluated.
Lower bond strengths were observed in the orthophosphoric acid and fine burr groups. Further high-quality studies with similar methodologies regarding the ceramic type, surface protocol, surface changes, the architecture of the bracket base and the type of the adhesive resin are required.