ONLINE ONLY

Loss of surface enamel after bracket debonding:

An in-vivo and ex-vivo evaluation
Huib Berghauser Pont,a Mutlu Özcan,b Bora Bagis,c and Yijin Rend
Groningen, The Netherlands, Z€
urich, Switzerland, and Trabzon, Turkey

Introduction: The objective of this study was to evaluate the surface enamel after bracket debonding and
residual resin removal. Methods: Thirty patients (female, 20; male, 10; mean age, 18.4 years) who completed
orthodontic treatment with fixed appliances (Twin Brackets, 3M Unitek, Monrovia, Calif) (n 5 525) were in-
cluded. The amounts of adhesive left on the tooth surfaces and the bracket bases were evaluated with the
adhesive remnant index (ARI). ARItooth (n 5 498) was assessed on digital photographs by 2 operators. After
resin removal and polishing, epoxy replicas were made from the maxillary anterior teeth (n 5 62), and enamel
surfaces were scored again with the enamel surface index. Elemental analysis was performed on the
debonded bracket bases by using energy dispersive x-ray spectrometry mean area scanning analysis.
The percentages of calcium and silicon were summed up to 100%. Tooth damage was estimated based on
the incidence of calcium from enamel in relation to silicon from adhesive (Ca%) and the correlation between
the ARIbracket and Ca%. Results and Conclusions: While ARItooth results showed score 3 as the most frequent
(41%) (P \0.05), followed by scores 0, 1, and 2 (28.7%, 17.9%, and 12.4%, respectively), ARIbracket results
showed score 0 more often (40.6%) than the other scores (P \0.05). Maxillary anterior teeth had
significantly more scores of 3 (49%) than the other groups of teeth (10%-25%) (chi-square; P \0.001). There
were no enamel surface index scores of 0, 3, or 4. No correlation between the enamel surface index and
ARItooth scores was found (Spearman rho 5 0.014, P 5 0.91). The incidence of Ca% from the scanned
brackets showed significant differences between the maxillary and mandibular teeth (14% 6 8.7% and
11.2% 6 6.5%, respectively; P \0.05), especially for the canines and second premolars (Kruskal-Wallis
test, P \0.01). With more remnants on the bracket base, the Ca% was higher (Jonckheere Terpstra test,
P\0.05). Iatrogenic damage to the enamel surface after bracket debonding was inevitable. Whether elemental
loss from enamel has clinical significance is yet to be determined in a long-term clinical follow-up of the studied
patient population. (Am J Orthod Dentofacial Orthop 2010;138:387.e1-387.e9)

scratches, or loss of enamel.2-7 The dilemma in adhesion

A
fter orthodontic therapy with fixed appliances,
from the clinical standpoint, a major concern of brackets in orthodontics is that it should be strong
is to avoid cohesive failures in the enamel during enough to prevent failure during all treatment but also
debonding brackets and at the same time to obtain tooth low enough so that enamel damage would be none or
surfaces without adhesive.1 Bonding onto and removal minimal during bracket removal after treatment.
of brackets from the enamel surfaces are potential risks Debonding forces can be influenced by many factors:
for topographic changes in the form of cracks, scarring, type of enamel conditioning agents (phosphoric acid,
a
self-etching primers, polyacrylic acid),8 adhesive resin,
Graduate student, Department of Orthodontics, University Medical Center
Groningen, University of Groningen, Groningen, The Netherlands. cement, polymerization methods, bracket type, or
b
Professor, Dental Materials Unit, Center for Dental and Oral Medicine, Clinic bracket base architecture.9,10 Usually, an increase in
for Fixed and Removable Prosthodontics and Dental Materials Science, Univer- debonding force causes an increased risk of enamel
sity of Z€
urich, Z€urich, Switzerland.
c
Assistant professor, Department of Prosthodontics, Faculty of Dentistry, Kara- damage.11 Even though the dental literature contains
deniz Technical University, Trabzon, Turkey. many studies looking at these possible factors on bracket
d
Professor, Department of Orthodontics, University Medical Center Groningen, adhesion, their clinical significance remains scarce,
University of Groningen, Groningen, The Netherlands.
The authors report no commercial, proprietary, or financial interest in the prod- since, in in-vitro studies, data are usually obtained with-
ucts or companies described in the article. out considering major intraoral factors—eg, saliva,
Reprint requests to: Mutlu Özcan, University of Z€urich, Dental Materials Unit, masticatory forces, temperature, and pH changes.12
Center for Dental and Oral Medicine, Clinic for Fixed and Removable Prostho-
dontics and Dental Materials Science, Plattenstrasse 11, CH-8032, Z€urich, Furthermore, clinical debonding with pliers creates
Switzerland; e-mail, mutluozcan@hotmail.com. a combination of shear, tensile, and torque forces. There-
Submitted, October 2009; revised and accepted, January 2010. fore, 1 test method (shear or tensile) in vitro might not
0889-5406/$36.00
Copyright Ó 2010 by the American Association of Orthodontists. represent in-vivo debracketing techniques and, conse-
doi:10.1016/j.ajodo.2010.01.028 quently, in-vivo bond strengths and failure types.12
387.e1
387.e2 Pont et al American Journal of Orthodontics and Dentofacial Orthopedics
October 2010

Several debonding techniques were introduced,13-18 but maxillary lateral incisor, 8.97 mm2; maxillary canine,
there were insufficient data regarding diverse susce- 10.19 mm2; maxillary premolar, 9.61 mm2; maxillary
ptibility to debonding damage in the maxillary or molar, 10.21 mm2; mandibular incisor, 9.81 mm2; man-
mandibular dental arches or in specific tooth groups.19,20 dibular canine, 10.129 mm2; mandibular premolar, 9.61
Bracket debonding strength for incisors in vitro requires mm2; and mandibular molar, 10.21 mm2. During the
significantly higher shear strength compared with orthodontic treatment, the patients received the corre-
molars.20 However, the effect of enamel damage relative sponding stainless steel brackets on all teeth in the man-
to the tooth type is unknown.20 Because of variations dible and the maxilla except on the second molars.
in enamel thickness or enamel prism orientation, Direct bracket bonding was done by 5 operators. A stan-
different failure types could be expected depending dardized protocol was followed for adhering the
on the tooth type.21,22 Since etched anterior teeth brackets to the tooth surfaces. Enamel surfaces were
have greater mean areas than do posterior teeth, it cleaned with water and fluoride-free pumice (Zircate
can be expected that anterior teeth have greater Prophy Paste, Dentsply Caulk, Milford, Del) with
potential for enamel damage than do posterior teeth.22 a prophylaxis brush (Hawe Prophy-Cup Latch-Type,
Possible failure types after bracket debonding are KerrHawe SA, Bioggio, Switzerland), rinsed with
either adhesive between the enamel and the adhesive water, and dried with an air syringe. Each tooth surface
resin, partially adhesive and cohesive in the adhesive was etched with 37% H3PO4 (Total etch, Ivoclar
resin (mixed), or adhesive between the bracket base Vivadent, Schaan, Liechtenstein) for 30 seconds, rinsed
and the adhesive resin, where the latter 2 require re- thoroughly with an oil-free air-water spray for 20
moval of the remnants. Unfortunately, remnant removal seconds, and air dried until they appeared frosty. The
can eventually cause further unwanted damage in the bonding adhesive (Heliobond, Ivoclar Vivadent) was
enamel because it is often done with rotating instru- applied with a microbrush, air thinned, and photo-
ments. Various methods are available to remove adhe- polymerized for 10 seconds (Ortholux, 3M Unitek, St
sive remnants: pliers, scalers, abrasive disks, diamond Paul, Minn). Then the brackets were bonded by using
or tungsten carbide burs, stones, or ultrasonic instru- adhesive cement (Transbond XT, 3M Unitek). After
ments.23 Although consensus is lacking in the dental removal of the excess around the bracket margins with
literature, the most common and efficient way of remov- the tip of a probe, they were photo-polymerized from
ing adhesive remnants in daily practice is with tungsten 5 directions: above the bracket, cervico-incisal, inciso-
carbide burs in a low-speed rotating hand piece.15,23 cervical, mesial, and distal for 20 seconds each (Ortho-
Potential detrimental effects of bracket debonding lux, 3M Unitek) (light output, 430-480 nm).
from enamel or elemental loss from surface enamel At the end of treatment, all brackets were debonded
either during bracket debonding or removal of the by placing the debonding pliers (Dentronix E231, Kon-
remnants is an iatrogenic problem.5,7,24 Calcium loss stanz, Germany) at the outer wings of the bracket.7 One
from the enamel surface particularly can result in clinician carried out all debonding procedures. After de-
dental erosion, which is a localized loss of dental hard bonding, the gross adhesive remnants were removed
tissues.25 with a tungsten carbide bur (Komet H22 AGK 016,
Although enamel surface morphology after bracket Lemgo, Germany) in a slow-speed hand piece under
debonding has been the topic of some in-vitro stud- water cooling. Subsequently, the remaining resin was
ies,13,15,18,23 quantitative information relevant to removed with a fine tungsten carbide bur (Komet H46
enamel damage after bracket debonding in vivo is 204 012) in a slow-speed handpiece. After removing
limited.24 Therefore, the objective of this study was to the resin, the tooth surfaces were polished with a polish
evaluate the surface morphology and the elemental cup (Hawe Prophy-Cup, white 967) and polishing paste
loss from surface enamel by using a semidirect method (Zircate Prophy Paste, Dentsply, Konstanz, Germany).
in an in-vivo and ex-vivo study. Final polishing was achieved with rubber points
(Ceramiste Midipoint, Shofu, Ratingen, Germany).
The amounts of adhesive remnants were scored by
MATERIAL AND METHODS using the adhesive remnant index (ARI) on all debonded
Thirty consecutive patients (20 female, 10 male; bracket bases (n 5 525).26 The ARIbracket scoring
mean age, 18.4 years) who completed orthodontic treat- system consists of a 4-point scale from 0 to 3: 0, no
ment with fixed appliances (Twin Brackets, 3M Unitek, adhesive left on the bracket base; 1, less than half of
Monrovia, Calif) were included in this study. The sur- the adhesive remained on the bracket; 2, more than
face areas of the brackets provided by the manufacturer half of the adhesive remained on the bracket; and 3,
were as follows: maxillary central incisor, 10.52 mm2; all adhesive was left on the bracket.10
American Journal of Orthodontics and Dentofacial Orthopedics Pont et al 387.e3
Volume 138, Number 4

To evaluate the adhesive remnants on the tooth sur- between 25 and 30 times; scan mode, area; type of cor-
faces, after debonding the brackets, before removing the rection, ZAF; type of analysis, standardless) was per-
remaining resin, the enamel surfaces were subjected to formed on the entire surface of the bracket base. The
the disclosing medium (GUM Red-cote, Chicago, Ill). percentages of calcium (from enamel) and silcon
This allowed better contrast between the tooth-colored (from adhesive) were summed up to 100%, and calcium
adhesive remnant and the enamel for the failure-site incidence was calculated in relation to silicon in per-
evaluation. Then standardized digital photographs centages (Ca%). In addition, the correlation between
(DX1 with a micro 60-mm lens, Nikon, Tokyo, Japan) ARIbracket and Ca% was determined. No materials
were taken of the dyed tooth surfaces (n 5 498). These used contained calcium (Table I). Therefore, the cal-
photographs were examined by 2 calibrated observers cium found on the resin was assigned to enamel loss.
who were blinded to the objectives of the study. The
sites of the failure types were recorded by using the Statistical analysis
ARI (ARItooth) scoring system. In cases of conflict, con-
The statistical analysis was performed with the
sensus was reached. ARItooth scores ranged from 0 to 3:
SPSS software package (version 11.5, SPSS, Chicago,
0, no adhesive left on the tooth; 1, less than half of the
Ill). Descriptive results for ARI, ESI, and Ca% were cal-
adhesive remained on the tooth; 2, more than half of
culated and expressed as frequencies, percentages,
the adhesive remained on the tooth; and 3, all adhesive
means, and standard deviations. The frequencies of
was left on the tooth, with a distinct impression of the
ARItooth and ARIbracket scores were compared by using
bracket mesh.
the chi-square test. Mann-Whitney U and Kruskal-
Possible damage to the enamel was evaluated in the
Wallis tests were applied to analyze the differences in
maxillary anterior teeth (n 5 62) only, according to the
Ca% between the maxillary and mandibular teeth, and
enamel surface index (ESI) described by Zachrisson and
the tooth types. Ca% in relation to ARIbracket scores
Årtun.5 The ESI consists of a 5-point scale from 0 to 4:
was analyzed by using the Jonckheere Terpstra test. Fur-
0, perfect surface with no scratches and distinct intact
thermore, the nonparametric Spearman correlation test
perikymata; 1, satisfactory surface with fine scratches
was used to examine the correlation between ESI scores
and some perikymata; 2, acceptable surface with several
and ARItooth scores. Statistical significance was defined
marked and some deeper scratches with no perikymata;
as P \0.05 in all tests.
3, imperfect surface with several distinct deep and
coarse scratches but no perikymata; and 4, unacceptable
surface with coarse scratches and deeply marked ap- RESULTS
pearance. For ESI scoring, immediately after the resid- Overall, while ARItooth results showed that score 3
ual adhesive removal and polishing, impressions (204 of 498) was the most frequent (41%) (P \0.05),
(President, Colténe/Whaledent, Altstätten, Switzerland) followed by 0, 1, and 2 (28.7%, 17.9%, and 12.4%, re-
were made from the entire labial surface and thereafter spectively). ARIbracket results showed score 0 (213 of
epoxy resin (Epofix Resin, Struers, Ballerup, Denmark) 525) most often (40.6%) (P \0.05), followed by 3, 2,
replicas were obtained. Secondary mode images were and 1 (29.1%, 18.1%, and 12.2%, respectively) (Table
made with cold field emission scanning electron micro- II). The highest incidence of ARItooth score 3 (29%)
scope (FE-SEM 6301F, Jeol, Tokyo, Japan) (magnifica- was observed in the central incisors, and the lowest
tion, 10 times) at 25 kV from epoxy replicas that were percentage of score 0 in the first molars (3%).
sputter-coated with 200 A Au (BAL-TEC sputter The maxillary anterior teeth had significantly more
coater; type 07 120B, Balzers, Liechtenstein) and fixed scores of 3 (49%) than other groups of teeth (16%-
to the specimen holder with a photo-polymerized 25%) (chi-square, P \0.001) (Table III). Typical failure
dental composite (Kulzer TransLUX EC, Wehrheim, types after bracket debonding are shown in Figure 1.
Germany). In the 62 teeth of the maxilla (28 central incisors, 28
Enamel damage was also quantified ex vivo from the lateral incisors, 6 canines), there were no ESI scores of
debonded brackets that received ARIbracket scores 1, 2, 0, 3, or 4. Both scores 1 (8%) and 2 (14%) were ob-
or 3 (n 5 306). Energy dispersive x-ray spectrometry served the least in the canines (Table IV). In general,
(EDX) mean area scan analysis (JSM-6400, Jeol) all cleaned teeth had acceptable to satisfactory enamel
(accelerating voltage, 20 kV; beam current, 6.1 nA; surfaces after debonding, with ESI scores of either 1
type of detector, Si(Li)-liquid N2 cooled, secondary or 2 (Fig 2). No correlation between ESI and ARItooth
mode, ThermoNoran System Six, SelectScience Ltd, scores was found (Spearman rho 5 0.014, P 5 0.91).
Bath, UK); spectra acquisition time, 100 seconds; detec- All evaluated brackets with ARIbracket scores of 1, 2,
tor dead time, 25%; resolution, 143 eV; magnification, and 3 had calcium incidence at varying degrees. Ca%
387.e4 Pont et al American Journal of Orthodontics and Dentofacial Orthopedics
October 2010

Table I. Brands, compositions, manufacturers, and batch numbers of the materials used in this study
Brand name Composition Manufacturer Batch number

Twin Brackets 17-4 stainless steel alloy 3M Unitek, Monrovia, Calif 3017-916
3017-917
3017-918
3017-9200
Heliobond Monomer matrix: dimethacrylate Ivoclar Vivadent, H29583
\60% Bis-GMA Schaan, Liechtenstein 154518
\40% triethyleneglycol
Transbond XT Silane-treated quartz (70-80 wt%) 3M Unitek, Monrovia, Calif 147
silane-treated silica \2%
BisDMA 10-20wt%
BisGMA 5-10wt%

Frequencies (%) of ARI scores for tooth/bracket surfaces (ARItooth/bracket) in the maxilla and mandible
Table II.
obtained from the teeth and their corresponding brackets
Score 0 Score 1 Score 2 Score 3
ARItooth/bracket (n 5 143/213) (n 5 89/64) (n 5 62/95) (n 5 204/153)
n 5 498/525 %tooth/%bracket %tooth/%bracket %tooth/%bracket %tooth/%bracket

Central incisor 16/28 12/22 19/11 29/14

Lateral incisor 18/22 16/28 32/17 23/16
Canine 21/22 20/22 26/22 22/18
First premolar 20/14 33/14 13/27 16/18
Second premolar 22/10 18/6 10/17 9/23
First molar 3/4 1/8 0/6 1/11

ARItooth score 0, no adhesive was left on the tooth; score 1, less than half of the adhesive remained on the tooth; score 2, more than half of the adhesive
remained on the tooth; score 3, all adhesive was left on the tooth with a distinct impression of the bracket mesh. ARIbracket score 0: no adhesive left on
the bracket base; score 1, less than half of the adhesive remained on the bracket; score 2, more than half of the adhesive remained on the bracket; score
3, all adhesive was left on the bracket.

from scanned bracket bases showed a total mean inci- Since adhesion has 2 aspects—one to the tooth sur-
dence of 12.6% (6 7.8%), with significant differences face and the other to the bracket base—evaluation of the
between the maxillary and mandibular teeth (14% 6 ARI scores provides information on the site of bond
8.7% and 11.2% 6 6.5%, respectively; P \0.01). failure as either adhesive at the bracket-adhesive resin
Ca% especially for the canines and the second premo- interface or the adhesive at the enamel-adhesive resin
lars showed significant differences between the maxilla interface. Macroscopic evaluation could also show
and the mandible (P \0.05) (Kruskal-Wallis test, cohesive failures in the enamel or in the adhesive resin.
P \0.01) (Fig 3). The mean Ca% differed significantly In this study, no macroscopically cohesive failures in the
in relation to the ARIbracket scores: the more ARI enamel were observed. ARItooth and ARIbracket scores
remnants on the bracket base, the higher the Ca% agreed; ie, the ARItooth scores of 3, 2, 1, and 0 had sim-
(Jonckheere Terpstra test, P \0.05). ilar frequencies to the ARIbracket scores of 0, 1, 2, and 3.
However, ARItooth and ARIbracket percentages did
not indicate that the failure type on the bracket base
DISCUSSION was a mirror image of the adhesive remnants on the tooth
It was previously emphasized that the outermost surface. This means that some failures were experienced
layer of enamel should be left as intact as possible, since solely in the adhesive resin. Although flexural behavior
it has greater microhardness and contains more minerals is different compared with metal brackets, similar obser-
and fluoride than the deeper zones.27 Consequently, the vations were noted in previous studies with ceramic and
loss of surface enamel and associated exposure of the polycarbonate brackets after debonding.10,28
enamel prism endings to the oral environment might Overall, ARItooth results showed that score 3 was the
cause a decrease in the resistance of enamel to the most frequent (41%), indicating adhesive failure be-
organic acids in plaque. This eventually makes enamel tween the bracket base and the adhesive resin with
more prone to demineralization.27 a distinct impression of the bracket mesh on the tooth
American Journal of Orthodontics and Dentofacial Orthopedics Pont et al 387.e5
Volume 138, Number 4

Table III.Frequencies (%) of ARI scores for tooth exposed to fatigue loads from chewing forces. The inci-
surfaces (ARItooth) depending on location (anterior vs dence of other failure types indicates this possibility.
posterior) per jaw (maxilla vs mandible) The high incidences of ARItooth score 0 on the first
Score 0 Score 1 Score 2 Score 3
molars and score 3 on the central incisors could be ex-
(n 5 143) (n 5 89) (n 5 62) (n 5 204) plained based on the etching pattern variations. The dif-
ARItooth %tooth %tooth %tooth %tooth ference in etched enamel morphology of different tooth
Maxillary anterior 20 25 45 49
types could affect composite resin bond strengths.22 In
Mandibular anterior 36 23 33 25 an in-vitro study, prismless enamel with a greater depth
Maxillary posterior 31 26 6 10 on the posterior teeth was found to be more common,21
Mandibular posterior 13 26 16 16 thus producing an inferior etch pattern.32 Other factors
ARItooth score 0, no adhesive was left on the tooth; score 1, less than relevant to low bond strength of brackets to posterior
half of the adhesive remained on the tooth; score 2, more than half teeth were also attributed to the mismatch (poor adapta-
of the adhesive remained on the tooth; score 3, all adhesive was left tion) between the bracket base architecture and the
on the tooth with a distinct impression of the bracket mesh. buccal surface convexity of the posterior teeth.33 This
results in an uneven composite layer with inferior me-
surface. No attempt was made to condition the bracket chanical properties. Moreover, moisture control in the
base surfaces for practical reasons because it prolongs posterior region of the mouth was found to be less favor-
chair-side time. This approach, however, could have able during bonding procedures.34,35 Although it cannot
affected the results and decreased the incidence of be proved that one of these factors is the only cause for
ARIbracket score 0.10,28 The second most common ARItooth score 0 in molars, when the failures were
failure type was adhesive failure between the enamel evaluated based on the quandrant and the ARItooth
and the adhesive resin (score 0). Score 0 implies weak scores of 0 and 3 incidences were summed, the
adhesion between the adhesive and the enamel, and mandibular posterior teeth had the lowest percentages
score 3 means weak adhesion between the bracket and of these scores (29%). Therefore, it cannot be
the adhesive resin. Since no brackets spontaneously generalized that the mandibular teeth or the molars are
debonded during orthodontic treatment, sufficient more prone to bracket debonding or that adhesion is
adhesion, at least for the duration of the treatment, lower to those teeth.35 Interestingly, however, mandibu-
was achieved with the bonding materials and the lar (61%) and maxillary (69%) anterior teeth had more
protocol used. Therefore, although an ARItooth score frequent sums of ARItooth scores of 0 and 3. Most prob-
of 0 or 3 indicates weak bond strength of the brackets ably, the debonding forces varied individually, depend-
in in-vitro orthodontic studies, it should not necessarily ing on crowding and severity of the malocclusion.
lead to early clinical failure.13,15,18,25 Therefore, in-vitro bracket bonding studies and the fail-
Adhesion between the tooth-adhesive-bracket as- ure characteristics should be evaluated with caution.
semblies is primarily influenced by fatigue and the cy- Future studies are warranted to study the bracket de-
clic forces on the bonded interfaces during chewing, bonding phenomena correlated with the clinical severity
resulting in higher debonding failures at the adhesive- before orthodontic treatment. Currently, the methodo-
enamel interface.29,30 Temperature change, humidity, logic quality of clinical trials evaluating debonding
and acidity (pH) might all additionally affect adhesive and bracket failure is generally poor to make compari-
strength, and it is almost impossible to simulate these sons with other clinical trials.36
factors ex vivo.29,30 Considering these factors, it is When failure type findings are coupled with the
conceivable that clinical debonding values could be Ca%, more explanations can be made. Information rel-
lower than those reported in in-vitro studies.31 In this evant to calcium loss was derived from brackets with
context, when failure types per tooth and site were eval- ARIbracket scores of 1, 2, or 3 or from ARItooth scores
uated, maxillary anterior teeth had significantly more of 0, 1, or 2. Ca% was significantly higher in the maxil-
frequent ARItooth scores of 3 (adhesive failure between lary teeth than in the mandibular teeth, indicating better
cement and bracket base) than did the other groups of adhesion on the maxillary teeth. Since moisture control
teeth. This could be attributed to the absence of direct was obtained with only suction and cotton rolls and no
occlusal forces applying shearing forces between the rubber dam was used during bonding brackets, inferior
enamel and the adhesive resin on the maxillary anterior adhesion in the mandible could have been expected.35
teeth compared with the mandibular anterior or poste- However, an insignificant amount of Ca% in the molars
rior teeth. However, depending on the severity of the in the maxilla and the mandible could not cause the
case (eg, crossbite or reverse deepbite in Class III mal- molars to be covered with more salivary proteins34 or
occlusion situations), maxillary teeth might also be the etch pattern to be less favorable on the molars.22
387.e6 Pont et al American Journal of Orthodontics and Dentofacial Orthopedics
October 2010

Fig 1. A-D, Intraoral digital photographs of the stained teeth with the disclosing medium after
bracket debonding. Typical clinical pictures of failure types of ARItooth: A, score 0; B, score 1; C, score
2; D, score 3. See Table I for detailed descriptions of the scoring system.

Table IV. Frequencies (%) of ESI (n 5 62) in the maxilla obtained from the anterior teeth (central incisors, lateral
incisors, and canines)
Score 0 Score 1 Score 2 Score 3 Score 4
(n 5 0) (n 5 40) (n 5 22) (n 5 0) (n 5 0)
ESI %tooth %tooth %tooth %tooth %tooth

Maxillary central incisors 0 40 55 0 0

Maxillary lateral incisors 0 53 32 0 0
Maxillary canines 0 8 14 0 0

Score 0, perfect surface with no scratchesand distinct intact perikymata; score 1, satisfactory surface with fine scratches and some perikymata; score
2, acceptable surface with several marked and some deeper scratches with no perikymata; score 3, imperfect surface with several distinct deep and
coarse scratches and no perikymata; score 4, unacceptable surface with coarse scratches and deeply marked appearance.

The mean Ca% incidence in relation to silicon dif- been due to the variations in their enamel thickness.37
fered significantly in relation to the ARIbracket scores; Enamel thicknesses in these teeth need to be further
the more adhesive remnants on the bracket base, the evaluated. The thickness of notable mineral-like parti-
higher the Ca%. Nevertheless, based on this finding, it cles detached from enamel, present in the adhesive rem-
can be stated that, even though the ARItooth score of nants on the bracket base after debonding, was
0 is often considered to represent a weak bond or a lower estimated to be between 5 and 25 mm after ceramic
hazard to the enamel, calcium loss is still possi- bracket debonding.38 In this study, stainless steel alloy
ble.15,18,26 This further indicates cohesive failures in brackets were used that are free of calcium and silicon.
the enamel prisms that could be detrimental for Since the whole bracket base was scanned, the lack of
possible demineralization or erosion. Therefore, after calcium and silicon in the bracket composition allowed
bracket debonding, with ARItooth scores of 0, 1, or 2, us to relate the incidence of calcium loss from the
these teeth need to be monitored for higher calcium enamel to the presence of silicon in the adhesive resin
loss from their enamel. Calcium loss for the canines debonded from enamel. If ceramic brackets had been
and second premolars showed significant differences used instead of metallic ones, better polymerization of
between the maxilla and the mandible; this could have the adhesive resin could have been expected due to
American Journal of Orthodontics and Dentofacial Orthopedics Pont et al 387.e7
Volume 138, Number 4

Fig 2. Representative photographs of the tooth surfaces immediately after bracket debonding (1A
and 2A) and after adhesive remnant cleaning (1B). ESI score 1, satisfactory surface with fine
scratches and some perikymata; 2B, ESI score 2, acceptable surface with several marked and
some deeper scratches without perikymata) (scanning electron microscope at secondary mode;
magnification, b 5 a 3 4).

the light transmission property of transparent ceramic applied to the outer wings of the bracket transferred
brackets that might also affect the final failure type after the least amount of stress to the enamel, whereas
debonding. However, the presence of silicon in those forces applied to the base of the bracket and to the
ceramics would then interfere with the silicon in the adhesive zone created stress concentration regions in
adhesives. the enamel that would cause separation at the
Although the thickness of enamel loss was not mea- adhesive-enamel interface.7 Therefore, in this study,
sured, the ARI scores provide a rough estimate of the the pliers were applied at the outer wings of the bracket.
amounts of adhesive remnants, since it is only a sur- Calcium scores, however, were not significantly differ-
face-area assessment. Furthermore, scores of 0, 1, and ent when the pliers were applied at either the bracket
2 do not imply adhesive or mixed failures between the base or the tie wings.24 Although the same bracket
enamel and the adhesive, since elemental loss of enamel type and adhesive resin (Transbond XT) was used in
was evident. the study of Brosh et al,24 no bonding agent was applied.
The torque forces created during debonding the In that study, no significant difference was found by
brackets might also affect the results. However, the tooth type, when also first molars were not involved.
opinions on this aspect are controversial.7,24 In The nonsignificant difference might be related to the
a photoeIastic stress analysis, it was found that forces lack of a bonding agent.
387.e8 Pont et al American Journal of Orthodontics and Dentofacial Orthopedics
October 2010

bonding. This was identified macroscopically and could

be considered a limitation of this study. Nevertheless,
the baseline information regarding preexisting defects
such as scratches and grooves was considered the condi-
tion of the enamel immediately after bracket removal.
However, ESI scores might be affected when brackets
are bonded to restorative materials; this is typical espe-
cially for adult patients. Similar to bracket removal,
during finishing and polishing, enamel surfaces next
to the restorations might involve some defects. For
such situations, ESI scores should be adapted in future
investigations.
Although scanning electron microscope images
clearly show the enamel damage, they lack a quantifica-
tion scale but provide only subjective information.
Volumetric quantification of enamel loss using
3-dimensional measurement techniques might provide
more detailed information about the debonded enamel
Fig 3. Calcium loss in weight (%) per tooth type based
on the bracket base measurements for the maxilla and
surface in future studies.39 Hence, our results imply
the mandible (*P \0.05). Note the number of evaluated from the clinical point of view that damage to the
teeth below the x-axis. enamel is inevitable in orthodontic applications, but,
since no correlation between ESI and ARItooth scores
was found, it can be said that, at least macroscopically,
The failure site at the bracket-adhesive interface the damage can be restored to acceptable or satisfactory
macroscopically indicates safe debonding and less levels with the method that is most frequently recom-
chance of enamel loss. However, this then requires me- mended for use in daily clinical practice. These results
ticulous removal of the resin remnants from the enamel from failure sites and calcium loss could be helpful
surface. In this study, there were no ESI scores of 0, 3, or information for future simulation studies.
4. In general, all cleaned teeth had acceptable to satis-
factory enamel surfaces after debonding with ESI scores CONCLUSIONS
of either 1 or 2, with no correlation between the amounts
From this study, the following can be concluded:
of adhesive remnants. Nevertheless, the results demon-
strated only minor iatrogenic damage to the enamel sur- 1. With the adhesive materials and the bonding proto-
face after bracket debonding with the finishing and col used, after debonding metal brackets, mainly
polishing protocol used. In addition to the elemental adhesive failures between the adhesive resin and
loss of surface enamel, we also demonstrated that the bracket base were observed (ARItooth score of 3),
use of rotary instruments for removing remaining resin and they were more frequent in the maxillary ante-
leaves the enamel surface with scratches. Irregularities rior teeth. On the bracket aspect, adhesive failure
caused by these rotary instruments result in greater between adhesive resin and bracket base was also
enamel surface roughness.16 This might cause more pla- more common (ARIbracket score of 0).
que accumulation, which in turn might increase the risk 2. The highest incidence of ARItooth score 3 was
of developing dental caries or gingivitis, depending on observed in the central incisors, with the lowest
the location. However, the iatrogenic damage created percentage of score 0 in the first molars.
by rotary devices mainly depends on the operator’s con- 3. After adhesive remnant removal, all teeth had
trol. Calcium loss was measured only after bracket re- acceptable to satisfactory enamel surfaces with
moval, but, after polishing, more calcium loss could ESI scores of either 1 or 2. No correlation between
be expected that should be studied in the future. Due ESI and ARItooth scores was found.
to easy access, the ESI scores were obtained only 4. Maxillary teeth tend to show more calcium loss
from the anterior teeth. Also, no loops were used while than do mandibular teeth. The mean calcium loss
removing the adhesive remnants. Therefore, the results differed significantly in relation to the ARIbracket
might not represent the posterior teeth. In our patient scores; the more ARI remnants on the bracket
population, enamel surfaces had no restorations, and base, the higher the incidence of Ca% in relation
no distinct enamel defects were noted before bracket to silicon.
American Journal of Orthodontics and Dentofacial Orthopedics Pont et al 387.e9
Volume 138, Number 4

We thank Ietse Stokroos for making the scanning 18. Eminkahyagil N, Arman A, Çetinsxahin A, Karabulut E. Effect of
electron microscope images; Heleen Boom, Manon resin-removal methods on enamel and shear bond strength of
rebonded brackets. Angle Orthod 2006;76:314-21.
Borsje, and Katrina Finnema for their assistance during
19. Arici S, Minors C. The force levels required to mechanically
the clinical procedures; Cengiz Tan at the Middle East debond ceramic brackets: an in vitro comparative study. Eur J
Technical University for conducting the energy disper- Orthod 2000;22:327-34.
sive x-ray spectrometry analysis; and the reviewers 20. Knoll M, Gwinnett AJ, Wollf MS. Shear bond strengths of
for their comments on the energy dispersive x-ray brackets bonded to anterior and posterior teeth. Am J Orthod
1986;89:476-9.
spectrometry analysis.
21. Whittaker DK. Structural variations in the surface zone of human
tooth enamel observed by scanning electron microscopy. Arch
Oral Biol 1982;27:383-92.
REFERENCES
22. Mattick CR, Hobson RS. A comparative micro-topographic study of
1. Campbell PM. Enamel surfaces after orthodontic bracket debond- the buccal enamel of different tooth types. J Orthod 2000;27:143-8.
ing. Angle Orthod 1995;65:103-10. 23. Hong YH, Lew KKK. Quantitative and qualitative assessment of
2. Fitzpatrick DA, Way DC. The effects of wear, acid etching, and enamel surface following five composite removal methods after
bond removal on human enamel. Am J Orthod 1977;72:671-81. bracket debonding. Eur J Orthod 1995;17:121-8.
3. Gwinnett AJ, Gorelick L. Microscopic evaluation of enamel 24. Brosh T, Kaufman A, Balabanovsky A, Vardimon AD. In vivo
after debonding: clinical application. Am J Orthod 1977;71: debonding strength and enamel damage in two orthodontic
651-65. debonding methods. J Biomech 2005;38:1107-13.
4. Brown CR, Way DC. Enamel loss during orthodontic bonding and 25. ten Cate JM, Imfeld T. Dental erosion, summary. Eur J Oral Sci
subsequent loss during removal of filled and unfilled adhesives. 1996;104:241-4.
Am J Orthod 1978;74:663-71. 26. Årtun J, Bergland S. Clinical trials with crystal growth condition-
5. Zachrisson BU, Årtun J. Enamel surface appearance after various ing as an alternative to acid-etch enamel pretreatment. Am J
debonding techniques. Am J Orthod 1979;75:121-37. Orthod 1984;85:333-40.
6. Rouleau BD, Marshall GW, Cooley RO. Enamel surface evalua- 27. Øgaard B. Oral microbiological changes, long-term enamel alter-
tions after clinical treatment and removal of orthodontic brackets. ations due to decalcification and caries prophylactic aspects. In:
Am J Orthod 1982;81:423-6. Brantley WA, Eliades T, editors. Orthodontic materials: scientific
7. Bennett CG, Shen C, Waldron JM. The effects of debonding on and clinical aspects. Stuttgart, Germany: Thieme; 2001. p. 124-39.
the enamel surface. J Clin Orthod 1984;18:330-4. 28. Faltermeier A, Behr M. Effect of bracket base conditioning. Am J
8. Shinya M, Shinya A, Lassila LVJ, Gomi H, Varrela J, Vallittu PK, Orthod Dentofacial Orthop 2009;135:1-5.
et al. Treated enamel surface patterns associated with five ortho- 29. Øilo G. Bond strength testing, what does it mean? Int Dent J 1993;
dontic adhesive systems. Surface morphology and shear bond 43:492-8.
strength. Dent Mater J 2008;27:1-6. 30. Pickett KL. In vivo orthodontic bond strength: comparison with in
9. Ireland AJ, Hosein I, Sherriff M. Enamel loss at bond-up, debond vitro results. Angle Orthod 2001;71:141-8.
and clean-up following the use of a conventional light-cured com- 31. Reynolds IR, von Fraunhofer JA. Direct bonding in orthodontic
posite and a resin-modified glass polyalkenoate cement. Eur J attachments to teeth: the relation of adhesive bond strength to
Orthod 2005;27:413-9. gauze mesh size. Br J Orthod 1975;3:91-5.
10. Özcan M, Finnema K, Ybema A. Evaluation of failure character- 32. Kodaka T, Kuroiwa M, Higashi S. Structural and distribution pat-
istics and bond strength after ceramic and polycarbonate bracket terns of surface ‘prismless’ enamel in human permanent teeth.
debonding. Eur J Orthod 2008;30:176-82. Caries Res 1991;25:7-20.
11. Östman-Andersson E, Marcusson A, Hörstedt P. Comparative 33. Evans ZB, Powers JM. Factors affecting in-vitro bond strength of
SEM studies of the enamel surface appearance following the no mix orthodontic cements. Am J Orthod 1985;87:508-12.
use of glass ionomer cement and a diacrylate resin for bracket 34. Hormati AA, Fuller JL, Denehy GE. Effects of contamination and
bonding. Swed Dent J 1993;17:139-46. mechanical disturbance on the quality of acid-etched enamel. J
12. Eliades T, Bourauel C. Intraoral aging of orthodontic materials: Am Dent Assoc 1980;100:34-8.
the picture we miss and its clinical relevance. Am J Orthod Den- 35. Mardaga WJ, Shannon IL. Decreasing the depth of etch for direct
tofacial Orthop 2005;127:403-12. bonding in orthodontics. J Clin Orthod 1982;16:130-2.
13. Retief DH, Denys FR. Finishing of enamel surfaces after debond- 36. Mandall NA, Millett DT, Mattick CR, Hickman J,
ing of orthodontic attachments. Angle Orthod 1979;49:1-10. Worthington HV, Macfarlane TV. Orthodontic adhesives: a sys-
14. Oliver RG, Griffiths J. Different techniques of residual composite tematic review. J Orthod 2002;29:205-10.
removal following debonding. Time taken and surface enamel ap- 37. Smith TM, Olejniczak AJ, Reid DJ, Ferrell RJ, Hublin JJ. Modern
pearance. Br J Orthod 1992;19:131-7. human molar enamel thickness and enamel-dentine junction
15. Zarrinia K, Eid NM, Kehoe MJ. The effect of different debonding shape. Arch Oral Biol 2006;51:974-95.
techniques on the enamel surface: an in vitro qualitative study. Am 38. Stratmann U, Schaarschmidt K, Wegener H, Ehner U. The extent of
J Orthod Dentofacial Orthop 1995;108:284-93. enamel surface fractures. A quantitative comparison of thermally
16. Eliades T, Gioka C, Eliades G, Makou M. Enamel surface rough- debonded ceramic and mechanically debonded metal brackets by
ness following debonding using two resin grinding methods. Eur J energy dispersive micro- and image-analyses. Eur J Orthod 1996;
Orthod 2004;26:333-8. 18:655-62.
17. Hosein I, Sherriff M, Ireland AJ. Enamel loss during bonding, 39. Lee YK, Lim YK. Three-dimensional quantification of adhesive
debonding, and cleanup with use of a self-etching primer. Am J remnants on teeth after debonding. Am J Orthod Dentofacial
Orthod Dentofacial Orthop 2004;126:717-24. Orthop 2008;134:556-62.