Areas of local decalcification of enamel without

ONLINE ONLY Orthodontic treatment-related white spot lesions: A 14-year prospective quantitative follow-up, including bonding material assessment Dmitry Shungin, a Alexandra Ioannidis Olsson, b and Maurits Persson c Umeå and Ronneby, Sweden Introduction: White spots (WS) related to orthodontic treatment are severe cariologic and cosmetic complications, but they are shown to be partially reduced by remineralization or abrasion in short-term follow-ups. In this prospective study, we quantitatively analyzed changes in WS in general and in treatment-related white spot lesions (WSL) during orthodontic treatment and at a 12-year follow-up after treatment. In addition, we quantitatively compared the effects of an acrylic bonding material vs a glass ionomer cement (GIC) on WSL. Methods: Sum areas of WS and WSL were calculated on scans of standardized photos of the vestibular surfaces of 4 teeth in consecutive orthodontic patients (median treatment time, 1.7 years) bonded with the 2 materials in a split-mouth design. Comparisons were made in 59 patients before treatment (BF), at debonding (T0), at 1 year (T1), and at 2 years (T2), and in 30 patients at a 12-year follow-up (T3) with the Friedman test followed by pairwise comparisons with the Wilcoxon matched-pairs signed rank test. Differences of the effects of acrylic vs GIC on the sum areas of WSL were tested for each observation period with the Mann- Whitney U test. Results: Increases in the sum areas of WS and WSL from BF to T0 (P \0.001) were followed by significant decreases at T1 (P \0.001) and T2 (P \0.01 for WS; P \0.001 for WSL). Significant changes were also found in the sum areas for WS at T3 compared with T2 (P \0.01), but not for WSL (P 5 0.328). The sum areas of WS and WSL at T3 did not return to BF levels (P \0.001). Sum areas of WSL were higher for surfaces bonded with acrylic compared with GIC for each observation period from BF to T2 (P.0.001), and from T2 to T3 (P.0.05). Conclusions: Although significantly reduced during the 12-year follow-up and significantly lower with the GIC than the acrylic material at bonding, WSL are a cariologic and cosmetic problem for many orthodontic patients. (Am J Orthod Dentofacial Orthop 2010;138:136.e1-136.e8) Areas of local decalcification of enamel without cavity formation are well-known complications of orthodontic treatment with a fixed appliance. 1-3 Prevalence of these white spot lesions (WSL) in patients after orthodontic treatment varies from 15% to 85%, 4 with most studies reporting 50% to 70%. 1-3,5-8 This high prevalence is explained by difficulties in performing oral hygiene procedures on bonded dental arches along with long-time accumulation and easier retention of bacterial plaque on tooth a Research fellow, Department of Odontology/Orthodontics, Umeå University, Umeå, Sweden; research assistant, Department of Pediatric Dentistry, Northern State Medical University, Arkhangelsk, Russia. b Orthodontist, Public Dental Health Service, County of Blekinge, Ronneby, Sweden. c Professor emeritus, Department of Odontology/Orthodontics, Umeå University, Umeå, Sweden. The authors report no commercial, proprietary, or financial interest in the products or companies described in this article. Reprint requests to: Dmitry Shungin, Department of Odontology/Orthodontics, Umeå University, SE-90187 Umeå, Sweden; e-mail, dmsh0001@student. umu.se. Submitted, January 2009; revised and accepted, May 2009. 0889-5406/$36.00 Copyright Ó 2010 by the American Association of Orthodontists. doi:10.1016/j.ajodo.2009.05.020 surfaces around fixed orthodontic elements. 6,9 The incidence and severity of WSL are lower if cariespreventive methods are used during orthodontic treatment, such as fluoride mouthwashes, 10-14 but patients compliance in using them is often unsatisfactory. 6 Some studies have shown that fluoride-releasing bonding materials, such as glass ionomer cements (GIC), can decrease the incidence of white spots (WS) after fixed appliance treatment when compared with acrylic-based materials, but results of long-term studies are lacking. 5,15-18 The term white spot lesion was defined by Fejerskov et al 19 as the first sign of a caries lesion on enamel that can be detected with the naked eye. Depending on the oral environment, these spots can either develop into cavities, stay stable for a long time, or heal to a certain extent. Postorthodontic WSL were shown to decrease during the first and second years after debonding. 5,20,21 The observed clinical reduction or healing of WSL after orthodontic treatment can be explained by removal of an etiologic factor (cariogenic plaque adhered to fixed orthodontic elements) combined with abrasion of the surface enamel during tooth brushing 19 and also 136.e1

136.e2 Shungin, Olsson, and Persson American Journal of Orthodontics and Dentofacial Orthopedics August 2010 Table. Number of participants, sex, age at the examinations, and median times to the debonding examination Examination Participants (n) Total Female Male Median age/ minimum-maximum (decimal years) Median time to debonding/ minimum-maximum (decimal years) Before treatment 59 38 21 13.6/10.8-19.1 1.7 (median treatment time)/1.34-2.76 At debonding 59 38 21 15.3/12.5-20.9-1-year follow-up 58 37 21 16.5/13.7-21.9 1.0/0.83-1.31 2-year follow-up 58 37 21 17.4/14.7-23.1 2.1/1.78-2.27 12-year follow-up 30 18 12 27.9/25.0-34.5 12.4/11.29-13.16 by remineralization. 22 Årtun and Thylstrup 23 attributed the decrease in WSL to be primarily a result of surface abrasion in addition to some redisposition of minerals. However, areas of demineralization will still remain on the teeth for up to 5 years after orthodontic treatment and thus be a cosmetic problem, but there is a shortage of conclusive long-term studies on changes in WSL after orthodontic treatment. 3 In addition to being caused by a cariologic process, white spots on the surface of enamel can have other origins eg, a result of developmental disturbances during enamel formation. In 13-year-olds, these enamel defects are reported to have a prevalence of 63% 24 and can cause similar cosmetic problems and thus be confused with WSL induced by orthodontic treatment. 25 As shown earlier, a differentiation between them can and should be made. 26 To assess WSL in orthodontic practice, a method introduced by Gorelick et al 25 has become the standard method of visual assessment. 5,8 The major disadvantage of this qualitative method is that it is insensitive to small and moderate changes in WSL. Quantitative methods of assessment of demineralized spots, such as quantitative light-induced fluorescence, have been described, but they require special equipment and skills. 20,27 Another approach to quantitative investigation of WS is to measure the area of demineralized enamel on photographic images of the teeth. 28,29 The accuracy of such measurements is similar to the more sophisticated methods of enamel demineralization assessment. 30,31 In this study, we quantitatively analyzed changes in WS in general as well as treatment-related WSL during orthodontic treatment and at a 12-year follow-up after treatment: 14 years in all. Thus, we analyzed the risk of prolonged cosmetic and cariologic complications. We also quantitatively compared the effect of an acrylic bonding material vs a GIC on these lesions. The first null hypothesis was that there were no differences in areas of WS and WSL between the examinations during 14 years. The second null hypothesis was that there are no differences in the areas of WSL related to the 2 bonding materials. MATERIAL AND METHODS This was a prospective study of 60 consecutive orthodontic patients from the Department of Orthodontics, Umeå University clinic, in Sweden, examined when their fixed preadjusted appliances were placed, at debonding, and 1, 2, and 12 years after debonding. One patient declined to participate in the follow-up 1 year after debonding and another one at 2 years after debonding. Before the examinations at 12 years after debonding, a patient with incomplete documentation was excluded from the study, resulting in 59 participants, who were then traced in the Swedish Population Register. Forty-one patients, still living within 50 km of the clinic, were invited to participate in the examinaiton. Five of the 41 did not reply, and 6 declined to participate. Thus, 30 patients were examined at 12 years after debonding (Table). Four teeth maxillary left and right lateral incisors, and mandibular left and right canines in each patient were evaluated. Teeth in the maxilla and the mandible were randomly allocated to be bonded with a no-mix acrylate (Unite, 3M Unitek, Monrovia, Calif) or a GIC (AquaCem, De Trey Division, Dentsply, Weybridge, Surrey, United Kingdom) in a split-mouth design. 5 The oral examination 12 years after debonding was supplemented by a questionnaire on dental treatment and caries prevention after orthodontic treatment. The questionnaire showed that no patient had received enamel bleaching, no examined teeth were treated, and no systematic fluoride supplement had been used after debonding. Color photos of the studied teeth were taken actual size (1:1 scale) with a Nikon F4 camera and medical Nikkor lens 120-mm f/4.0 with incorporated ring flash (both, Nikon, Tokyo, Japan). Although it was impossible to use the same film type during the whole study period, positive films with the same characteristics were used.

American Journal of Orthodontics and Dentofacial Orthopedics Shungin, Olsson, and Persson 136.e3 Volume 138, Number 2 Fig 1. Border of a WS, classified as a lesion related to orthodontic treatment (WSL), hand-outlined with the mouse in the FACAD software. To minimize reflection and glare on the enamel surface that could mask its original color, 2 photos of each air-dried tooth surface were taken: parallel to the center of tooth surface, and at an angle of 2 to 3. Thus, a pair of slides of the same vestibular tooth surface of the each studied tooth at every examination was available for assessment, for a total of 1056 pairs of slides. To choose the appropriate scanning resolution, the difference between resolutions of 2400 and 3200 dpi was tested in 25 slides scanned with the Epson Perfection 4990 Photo scanner (Seiko Epson, Nagano, Japan). High agreement was found between measurements of areas of spots on slides scanned at the 2 resolutions (ICC 5 0.998), but scanning with 3200 dpi resolution was more time-consuming. Thus, all slides were scanned at 2400 dpi (24-bit color normal), named, and saved in tagged image file format (TIFF) on an external hard drive. Both images from each pair were enlarged by 600% in the FACAD cephalometric software (version 2.2, Ilexis AB, Linköping, Sweden) and visually compared; the one without reflection and glare or with minimal reflections was chosen for further assessment. To ensure blindness during the image-analysis stage, all selected images were renamed according to a table of random numbers, and the renaming key was sealed in an envelope and stored by an independent person until the end of all measurements. In this way, tooth surfaces from various subjects photographed at different examinations were mixed up, and the researchers were unaware of which image belonged to which patient at which examination. To calculate the area of WS, each image was enlarged by 600% in the FACAD 2.2 cephalometric software. The border of every WS on the tooth surface was hand-outlined by using a mouse (IBM MU29j, International Business Machines, Singapore, China) as Fig 2. Calculation of the AUC for the acrylic bonding material from pretreatment to debonding. In the same way, the AUC was calculated for both materials at each examination for every patient. A, Sum area of WSL at pretreatment; B, sum area of WSL at debonding; C, treatment time (or time between examinations). AUC 5 0.5*(A 1 B)*C is the formula for the area of the trapezoid formed by the curves. shown in Figure 1. The area within the WS s border was automatically calculated in square pixels by the cephalometric software and transferred into the SPSS database (version 15.0, SPSS for Windows, SPSS, Chicago, Ill). To analyze WSL of orthodontic origin, all images of tooth surfaces with hand-outlined WS were assessed a second time. All WS located in the incisal fourth of the crown were considered unlikely to be caused by orthodontic treatment; they were excluded from evaluation. WS reflecting the position of the brace (Fig 1) or the position of the gingival margin (arch, banana, or kidney shaped) were considered treatment-related WSL and were included in the analysis. 32 The data on their areas were again transferred to the SPSS software. The repeatability of the WS and WSL area measurement technique, assessed on 25 slides with a time interval of 3 weeks, was high (ICC 5 0.877). Individuals rather than tooth surfaces were of particular interest in this study, the sum areas of all outlined WS and WSL on the 4 studied tooth surfaces were calculated for each patient at each examination (in square pixels). The sum area was used to compare WS and WSL at the several examinations. To adjust for possible differences in the betweenexamination times in different patients, the area under

136.e4 Shungin, Olsson, and Persson American Journal of Orthodontics and Dentofacial Orthopedics August 2010 Fig 3. Sum area of WS at different examinations (**P \0.01; ***P \0.001, Wilcoxon matched-pairs signed rank test). the curve (AUC) was used to compare the effects of GIC vs acrylic bonding materials on WSL (Fig 2). 33 The AUC was calculated as a trapezoidal area under the WSL curve, and comparisons of the AUC of the sum area of WSL between materials were performed for every studied period. Statistical analysis The Friedman analysis was used to test for differences in the sum areas of WS and WSL at the various examinations. The Wilcoxon matched-pairs signed rank test was used for pairwise comparison between these examinations. Results of the first 4 examinations were analyzed for 59 participants; all comparisons at the examination 12 years after debonding were performed for 30 patients. Differences in the AUC of sum areas of WSL for GIC vs acrylic bonding materials were tested with the Mann-Whitney U test. The level of significance was set at 5%, and the Bonferroni adjustment was applied for multiple comparisons. The ICC was used to assess the agreement of handoutlining of the spots on slides scanned at the 2 resolutions and the repeatability of hand-outlining the spots. 34 RESULTS Statistically significant differences in the sum areas of WS between first 4 examinations in 59 patients were identified with the Friedman test (P\0.001). When compared pairwise with the Wilcoxon matched-pairs signed rank test (Fig 3), a significant increase in the sum area of WS from before treatment to debonding was found (P \0.001). This was followed by a significant decrease in the sum area of WS after the 1-year follow-up from debonding (P \0.001) and from the 1-year follow-up to the 2-year follow-up (P\0.01). In addition, the sum areas of WS before treatment were significantly lower than 2 years after debonding (P \0.01). The Friedman test showed statistically significant differences between pretreatment, 2 years, and 12 years after debonding for the 30 patients (P \0.001). At the 2-year follow-up, the sum area of WS was significantly higher than at the 12-year follow-up (P\0.01, Fig 3). In addition, the sum area of WS after 12 years from debonding did not reach the pretreatment level and was higher (P \0.001). There were significant differences in sum areas of WSL between the first 4 examinations in the 59 patients (P\0.001, Friedman test). Pairwise comparison (Fig 4) showed significant increases in sum areas of WSL from pretreatment to debonding (P \0.001) followed by decreases to the 1-year follow-up after debonding (P \0.001). At the 2-year follow-up after debondig, further significant decreases in sum area of WSL was found (P \0.001), but the sum areas were still significantly higher than at the pretreatment examinations (P \0.001). Differences in the sum areas of WSL attributed to orthodontic treatment between pretreatment, 2 years,

American Journal of Orthodontics and Dentofacial Orthopedics Shungin, Olsson, and Persson 136.e5 Volume 138, Number 2 Fig 4. Sum area of WSL attributed to orthodontic treatment at different examinations (***P \0.001; NS, not significant at 5% level; Wilcoxon matched-pairs signed rank test). and 12 years after debonding examinations were significant in the 30 patients (P \0.001, Friedman test). At the 12-year follow-up after debonding, no significant change was found in sum areas of WSL compared with the 2-year follow-up (P 5 0.328, Fig 4). As with the 2-year follow-up, the sum areas of WSL at the 12-year follow-up were significantly higher than before treatment (P \0.001). The AUC of the sum areas of treatment-related WSL was significantly higher for the acrylic bonding material than for the GIC for each of the 3 periods from before treatment to the 2-year follow-up (P \0.001, Mann-Whitney U test; Fig 5). Similarly, the AUC of the sum area of treatment-related WSL for the last observation period (2-12 years after debonding) was significantly higher for the acrylic material than for the GIC (P \0.05). There were no systematic differences in the sum areas of WS and WSL before treatment and at the 2-year follow-up examinations or in the AUC of the sum areas of WSL for all observation periods for both materials between the 30 patients who participated in the 12-year follow-up examination and the 29 patients who did not (P.0.05 for all, Mann-Whitney U test). DISCUSSION This study was undertaken to quantitatively assess changes of WS and treatment-related WSL during a 12-year follow-up (14 years all in all). Specific changes related to bonding material were also studied. Our major finding was that, even after 12 years from debonding, the sum area of WSL did not return to pretreatment levels. This finding strengthens the results of the study that was previously considered to have the longest postdebonding follow-up period and qualitatively described WSL 5 years after orthodontic treatment. 3 Selection of the teeth for evaluation was based on the available information about tooth-specific prevalence of WSL when this study was planned. At that time, the maxillary lateral incisors and the mandibular canines had been reported to be the most susceptible to WSL formation. 25 In most previous studies, the 4-grade scale introduced by Gorelick et al 25 was used to assess WS. 3,5,13 This approach can be insensible to moderate and small changes in areas of WS and thus could lead to deceiving results in comparative and long-term studies. In our study, a photographic technique was used for WS assessment. This approach is reliable because it is possible to calculate WS areas with good precision on the photographic images even if the slides are taken at slightly different angles. 29 Scanning of both 35-mm slides and digital images can be used to measure enamel demineralization. 28 Furthermore, levels of validity, repeatability, and agreement of photographic measurements have been shown to be equal to the quantitative light-induced fluorescence method of assessment of enamel demineralization. 30,31 In our study, the

136.e6 Shungin, Olsson, and Persson American Journal of Orthodontics and Dentofacial Orthopedics August 2010 Fig 5. Effect of bonding material (GIC vs acrylic) on formation of WSL related to orthodontic treatment. The AUC is used for comparisons between materials (*P \0.05; ***P \0.001, Mann-Whitney U test). repeatability of outlining the WS areas with the mouse on the digital images of the scanned slides was also high (ICC 5 0.877). Because this coefficient includes operator error, we suggest that photographic images of tooth surfaces can be routinely used for WS and WSL evaluation, if the photos are taken with a standardized protocol. For the areas of WS calculated in square pixels, possible quantitative comparisons with other studies can be made only if photography, scanning, and outlining conditions are considered (magnification at photo shooting is 1:1, scanning resolution is 2400 dpi, magnification of the image during hand-outlining of the spots is 600%). For this study, 1 mm 2 of WS or WSL area equals approximately 22677.2 square pixels. WS is a broad and sometimes misused term that describes all spots and opacities of white color on tooth surfaces. In this study, we attempted to distinguish between WS and the more specific term WSL induced by orthodontic treatment. In previous qualitative studies, it was not clearly stated whether all WS on tooth surfaces were considered WSL during assessment, or whether clinical judgments were applied to identify and include only WSL attributed to orthodontic treatment. 5,8 Inclusion of WS of nonorthodontic origin might have influenced the results of previous studies. It was found that WSL can be confused with developmental defects of enamel eg, demarcated and diffused white opacities, opaque hypoplasia, fluorosis, and so on. 35 Differentiation of WSL from spots of developmental or other origin can be done mainly on the basis of location and shape, especially when photographic images of the teeth, but not patients, were examined in random order. 25,26,32 In most tooth surfaces in our study, the origin of the WS was obvious. They were classified as treatment-related WSL because they reflected the location of cariogenic plaque (close to gingival margin or borders of braces; banana, kidney, or arch shaped). 32 WS were classified as not related to the treatment if they were located on the incisal fourth of the crown, in some cases touching the incisal edge. For some WS, it was difficult to decide whether they were caused by the treatment. We decided to include the doubtful WS in our analysis and classify them as treatment-related WSL because it was possible that they were induced by treatment. Interestingly, treatment-related WSL were found in the before-treatment group (Fig 4). This can partly be explained by pretreatment initial caries lesions 32 or developmental defects of enamel 35,36 located in the typical areas for orthodontically induced WSL; thus, they were classified as such during our blind assessment of randomized images. In addition, this group could also include previously mentioned doubtful WS that were classified as WSL but, in reality, could have had another origin.

American Journal of Orthodontics and Dentofacial Orthopedics Shungin, Olsson, and Persson 136.e7 Volume 138, Number 2 The sum area of WS after debonding was almost 6 times higher than before treatment, decreased by 2 times after 1 year, and had further significant decreases after 2 and 12 years from debonding. When only treatmentrelated WSL were analyzed, a sharp increase in the sum area at debonding was also followed by sharp decreases at 1 and 2 years after debonding. The same pattern was shown in quantitative studies 20,21 and qualitative studies, 3,6 supporting the theory that the greatest decrease in WS is during the first year after orthodontic treatment. This dramatic initial reduction of WS and WSL after debonding can probably be explained by remineralization (healing) as well as surface abrasion of the nonmineralized lesion predominantly during tooth brushing. 23 An earlier study, when the number of teeth with WSL was used as an outcome measure, showed that after 2 years from debonding the number of teeth with WSL was higher than before treatment. 5 Even though it is difficult to directly compare results of our study with that one because of differences in outcome measures, our results show the same pattern that, after 2 years from debonding, orthodontic-treatment-related WS were still present on the tooth surfaces. This was also the case at 5 years after debonding. 3 Mattousch et al 21 described no changes in WSL between 6 months and 2 years after debonding using a quantitative technique. In our study, no significant difference was found in the sum areas of treatment-related WSL between 2 and 12 years after debonding. Although there was a difference in time when no further reduction of WSL was found in both studies, this period supports observations that orthodontically induced WSL can be stable in time. This is also further supported since the sum areas of neither WS nor orthodontic WSL did not reach pretreatment levels even after 12 years from debonding in our study, although the appearance of new lesions during 12 years of follow-up because of inadequate oral hygiene or consumption of sugary drinks cannot be ignored. No systematic fluoride supplement was used during or after orthodontic treatment in this study, and no patient reported enamel bleaching during the follow-up period. It was shown that these factors could have reduced the severity of lesions at debonding and follow-ups. 9,10,37,38 Orthodontic treatment time can influence the severity of WSL and thus should be considered when comparing the effect of bonding material on WSL formation in patients with different treatment times. 5,6 The AUC, which was used to compare the effects of GIC vs acrylic bonding materials on WSL, takes into account differences between the examination times in patients. 33 Short-term observations showed that using a GIC for bracket bonding can decrease WSL formation during orthodontic treatment compared with an acrylic bonding material. 5,16,18 Our results showed the same pattern for the long-term period. Acrylic-bonded surfaces had greater sum areas of WSL than did those bonded with the GIC during all separately analyzed periods, including the period from 2 to 12 years after debonding. These can probably be explained because the GIC releases fluoride, preventing lesion formation. 5,19,39 It is also possible that acrylic bonding material is associated with deeper lesions that are less likely to remineralize or be abraded during long-term follow-up. CONCLUSIONS Both null hypotheses were rejected. The area of WS, quantitatively assessed, increased significantly during orthodontic treatment, decreased markedly during the first and second years posttreatment, but did not reach the pretreatment level even 12 years after debonding. WSL classified as related to orthodontic appliance treatment showed similar results and might be a cariologic and cosmetic problem 12 years after debonding. WSL areas were significantly reduced in the short-term and long-term perspectives with the use of a fluoridereleasing GIC compared with an acrylic bonding agent. We thank Hans Stenlund for statistical help and Agneta Marcusson, Lars-Inge Norevall for collecting material for the first follow-up periods, and Ronny Fors for valuable comments. REFERENCES 1. Mizrahi E. Surface distribution of enamel opacities following orthodontic treatment. Am J Orthod 1983;84:323-31. 2. Årtun J, Brobakken BO. Prevalence of carious white spots after orthodontic treatment with multibonded appliances. Eur J Orthod 1986;8:229-34. 3. Øgaard B. Prevalence of white spot lesions in 19-year-olds: a study on untreated and orthodontically treated persons 5 years after treatment. Am J Orthod Dentofacial Orthop 1989;96:423-7. 4. Mitchell L. Decalcification during orthodontic treatment with fixed appliances an overview. Br J Orthod 1992;19:199-205. 5. Marcusson A, Norevall LI, Persson M. White spot reduction when using glass ionomer cement for bonding in orthodontics: a longitudinal and comparative study. Eur J Orthod 1997;19:233-42. 6. Geiger AM, Gorelick L, Gwinnett, Griswold PG. The effect of a fluoride program on white spot formation during orthodontic treatment. Am J Orthod Dentofacial Orthop 1988;93:29-37. 7. Basdra EK, Huber H, Komposch G. Fluoride release from orthodontic bonding agents alters enamel surface and inhibits enamel demineralization in vitro. Am J Orthod Dentofacial Orthop 1996;109:466-72. 8. Banks PA, Chadwick SM, Asher-McDade C, Wright JL. Fluoridereleasing elastomerics a prospective controlled clinical trial. Eur J Orthod 2000;22:401-7.

136.e8 Shungin, Olsson, and Persson American Journal of Orthodontics and Dentofacial Orthopedics August 2010 9. Zachrisson BU. Fluoride application procedures in orthodontic practice, current concepts. Angle Orthod 1975;45:72-81. 10. Øgaard B, Alm AA, Larsson E, Adolfsson U. A prospective, randomized clinical study on the effects of an amine fluoride/stannous fluoride toothpaste/mouthrinse on plaque, gingivitis and initial caries lesion development in orthodontic patients. Eur J Orthod 2006;28:8-12. 11. Øgaard B, Larsson E, Henriksson T, Birkhed D, Bishara SE. Effects of combined application of antimicrobial and fluoride varnishes in orthodontic patients. Am J Orthod Dentofacial Orthop 2001;120:28-35. 12. Benson PE, Parkin N, Millett DT, Dyer FE, Vine S, Shah A. Fluorides for the prevention of white spots on teeth during fixed brace treatment. Cochrane Database Syst Rev 2004:CD003809. 13. Geiger AM, Gorelick L, Gwinnett AJ, Benson BJ. Reducing white spot lesions in orthodontic populations with fluoride rinsing. Am J Orthod Dentofacial Orthop 1992;101:403-7. 14. Demito CF, Vivaldi-Rodrigues G, Ramos AL, Bowman SJ. The efficacy of a fluoride varnish in reducing enamel demineralization adjacent to orthodontic brackets: an in vitro study. Orthod Craniofac Res 2004;7:205-10. 15. Gorton J, Featherstone JD. In vivo inhibition of demineralization around orthodontic brackets. Am J Orthod Dentofacial Orthop 2003;123:10-4. 16. Banks PA, Burn A, O Brien K. A clinical evaluation of the effectiveness of including fluoride into an orthodontic bonding adhesive. Eur J Orthod 1997;19:391-5. 17. Vorhies AB, Donly KJ, Staley RN, Wefel JS. Enamel demineralization adjacent to orthodontic brackets bonded with hybrid glass ionomer cements: an in vitro study. Am J Orthod Dentofacial Orthop 1998;114:668-74. 18. Donly KJ, Istre S, Istre T. In vitro enamel remineralization at orthodontic band margins cemented with glass ionomer cement. Am J Orthod Dentofacial Orthop 1995;107:461-4. 19. Fejerskov O, Nyvad B, Kidd EAM. Clinical and histological manifestations of dental caries. In: Fejerskov O, Kidd EAM, editors. Dental caries: the disease and its clinical management. Copenhagen, Denmark: Blackwell Munksgaard; 2003. p. 71-99. 20. Al-Khateeb S, Forsberg CM, De Josselin De Jong E, Angmar- Mansson B. A longitudinal laser fluorescence study of white spot lesions in orthodontic patients. Am J Orthod Dentofacial Orthop 1998;113:595-602. 21. Mattousch TJ, van der Veen MH, Zentner A. Caries lesions after orthodontic treatment followed by quantitative light-induced fluorescence: a 2-year follow-up. Eur J Orthod 2007;29:294-8. 22. ten Cate JM, Arends J. Remineralization of artificial enamel lesions in vitro. III. A study of deposition mechanism. Caries Res 1980;14:351-8. 23. Årtun J, Thylstrup A. Clinical and scanning electron microscopic study of surface changes of incipient caries lesions after debonding. Scand J Dent Res 1986;94:193-201. 24. Suckling GW, Pearce EI. Developmental defects of enamel in a group of New Zealand children: their prevalence and some associated etiological factors. Community Dent Oral Epidemiol 1984; 12:177-84. 25. Gorelick L, Geiger AM, Gwinnett AJ. Incidence of white spot formation after bonding and banding. Am J Orthod 1982;81: 93-8. 26. Kanthathas K, Willmot DR, Benson PE. Differentiation of developmental and post-orthodontic white lesions using image analysis. Eur J Orthod 2005;27:167-72. 27. Aljehani A, Tranaeus S, Forsberg CM, Angmar-Mansson B, Shi XQ. In vitro quantification of white spot enamel lesions adjacent to fixed orthodontic appliances using quantitative light-induced fluorescence and DIAGNOdent. Acta Odontol Scand 2004;62:313-8. 28. Benson PE, Shah AA, Willmot DR. Measurement of white lesions surrounding orthodontic brackets: captured slides vs digital camera images. Angle Orthod 2005;75:226-30. 29. Benson PE, Pender N, Higham SM. Enamel demineralisation assessed by computerised image analysis of clinical photographs. J Dent 2000;28:319-26. 30. Benson PE, Pender N, Higham SM. Quantifying enamel demineralization from teeth with orthodontic brackets a comparison of two methods. Part 1: repeatability and agreement. Eur J Orthod 2003;25:149-58. 31. Benson PE, Pender N, Higham SM. Quantifying enamel demineralization from teeth with orthodontic brackets a comparison of two methods. Part 2: validity. Eur J Orthod 2003;25:159-65. 32. Nyvad B, Fejerskov O, Baelum V. Visual-tactile caries diagnosis. In: Fejerskov O, Kidd EAM, editors. Dental caries: the disease and its clinical management. 2nd ed. Copenhagen, Denmark: Blackwell Munksgaard; 2008. p. 49-68. 33. Matthews JNS, Altman DG, Campbell MJ, Royston P. Analysis of serial measurements in medical research. Br Med J 1990;300: 230-5. 34. McGraw KO, Wong SP. Forming inferences about some intraclass correlation coefficients. Psychol Methods 1996;1:30-46. 35. Commission on Oral Health, Research & Epidemiology. Report of an FDI working group. A review of the developmental defects of enamel index (DDE index). Int Dent J 1992;42:411-26. 36. Dummer PM, Kingdon A, Kingdon R. Prevalence and distribution by tooth type and surface of developmental defects of dental enamel in a group of 15- to 16-year-old children in South Wales. Community Dent Health 1990;7:369-77. 37. Stecksén-Blicks C, Renfors G, Oscarson ND, Bergstrand F, Twetman S. Caries-preventive effectiveness of a fluoride varnish: a randomized controlled trial in adolescents with fixed orthodontic appliances. Caries Res 2007;41:455-9. 38. Knösel M, Attin R, Becker K, Attin T. External bleaching effect on the color and luminosity of inactive white-spot lesions after fixed orthodontic appliances. Angle Orthod 2007;77:646-52. 39. Chung CK, Millett DT, Creanor SL, Gilmour WH, Foye RH. Fluoride release and cariostatic ability of a compomer and a resin-modified glass ionomer cement used for orthodontic bonding. J Dent 1998;26:533-8.