In vivo inhibition of demineralization around orthodontic brackets☆☆☆★★★

Material and methods 

The study consisted of 2 parallel groups in a double-blind, randomized, controlled, clinical trial, with human subjects' approval UCSF number H9136-16814-01. Patients 11 to 18 years of age who were scheduled to have 2 or more first premolars extracted were invited to participate and sign a consent form. Eight participants (38%) were male, and 13 (62%) were female. There were 9 Hispanics (43%), 5 blacks (24%), 4 whites (19%), and 3 Asians (14%). The age range was 11 to 18 years (mean, 13.19 years; SD 1.91).

Only caries-free first premolars were included in this study, because they accumulate significantly more plaque than any other teeth.7 Participants were randomized into control and intervention treatment groups that balanced age, sex, risk status, and factors not measured.15 To avoid selection bias, randomization of the patients into test and control groups was based on a list compiled by a statistician not directly involved in the study using a computerized statistics program. The group allocation was revealed by an assistant not directly involved in the study only after each patient had been seated for the initial bonding appointment. All patients received a full-mouth cleaning to remove plaque in preparation for bonding.

The patients were instructed to brush twice daily with a dentifrice provided to them, containing 1100 ppm fluoride as sodium fluoride. The patients or their parents were asked to complete a log of their daily tooth-brushing schedules, and preweighed tubes of toothpaste were given to them. They also received detailed verbal and written instructions and check-off sheets for their at-home saliva sampling.

At the start of the experiment (day 0), each patient had 2 caries-free first premolars bonded with either fluoride-releasing resin-modified glass ionomer (Fuji Ortho LC, GC America Inc, Chicago, Ill) or nonfluoride-releasing composite resin (Transbond XT, 3M Unitek, Monrovia, Calif). There were no visible signs of caries, fluorosis, or developmental defects in the teeth used. Both cements were standardized to measure 1 mm of material extruded from the applicator tip. All teeth in this experiment were bonded by 1 investigator (J.G.), and the premeasured automatically mixed capsule form of the Fuji LC was used (Transbond XT did not require mixing). Patients were not told which bonding agent was used on the teeth involved in the study. Also, the extracted teeth were scrambled and renumbered by a technician not directly involved in the study to ensure blinding of the investigator before microhardness testing. Extractions were scheduled to take place 4 weeks after initial bracket placement.

After extraction, the teeth were placed in 0.1% thymol solution in deionized water and sterilized overnight with a gamma irradiation (Cs 137) at a dose above 173 kilorad. Then the collection media was replaced with fresh deionized water and thymol. The roots were removed 2 mm apical to the cementoenamel junction, and the crowns were hemi-sectioned vertically into mesial and distal halves with a 15 HC (large) wafering blade on an Isomet low-speed saw (Buehler, Lake Bluff, Ill). The samples were sectioned directly through the slot of the bracket, leaving a gingival portion and an incisal portion.

The teeth were embedded in epoxy resin (Ladd Research Industries, Rochester, NY), leaving the cut face exposed. After polishing, the exposed flat hemi-sectioned surface was indented to test for microhardness cross-sectionally with a microhardness tester (Leitz, Wetzler, Germany) and microscopic examination.17, 18, 19, 20 The first indention was located 15 μm deep toward the dentin from the enamel surface (lingual) and approximately 100 μm below (gingival) the cement margin. Subsequent indentations continued into the underlying enamel, increasing in depth from the outer surface by 5 μm each time to a depth of 50 μm in a V-shaped pattern. After this, indentions were made at 25 μm deep intervals into underlying sound enamel in a straight line perpendicular to the outer surface up to a total depth of 300 μm.17, 20, 21 The volume percentage of mineral (VPM) for each indent was then normalized based on sound underlying enamel set at 85%. This is an internal calibration of the measurements in the ΔZ formula that allows for normalization of the microhardness data on a per-tooth basis, so that tooth-to-tooth variability is eliminated. Also, all data were verified for reproducibility of the measurement method and for quality assurance before revealing the group assignments by repeating any outlier measurements. Measurement of indentation lengths/demineralization was calculated with Image pro plus 4.0 software (Media Cybernetics, Silver Springs, MD) to capture and measure the image through a microscope (Olympus BX50, Melville, NY) at 500× magnification.

The overall relative mineral loss, ΔZ, for each sample was calculated by creating a hardness profile curve; this was done by plotting normalized VPM against distance from the enamel surface. The area under the curve that represents ΔZ (μm × VPM) was calculated with Simpson's integration rule.20 Also, the ΔZ values for each lesion in each group were combined to give a mean ΔZ and a standard deviation for each group.

Additionally, saliva samples were collected at days 0 (baseline), 1, 2, 3, 7, 14, 21, and 28 (extraction day). Each subject was asked to chew on a 2 × 2-in square of Parafilm (Structure Probe, Inc, West Chester, Pa) and then to spit 2 ml of saliva into a prelabeled sterile 50-mL centrifuge tube. One investigator (J.G.) collected saliva samples at all sampling intervals except for days 1, 2, and 3, when the subjects collected without supervision. All samples were coded and relabeled by a technician not involved in the study. After collection, the saliva was stored at 4°C no longer than 1 week for fluoride analysis. The Taves22 microdiffusion method was used to evaluate the whole saliva samples for fluoride content. Saliva fluoride content (μg) was calculated from a standard curve. These standards were microdiffused at the same time as the samples, and fluoride concentrations were measured by a fluoride ion-specific electrode (Model 960900, Orion Research, Boston, Mass).

Results 

Of the 25 potential subjects, 24 agreed to take part in the study. Data are complete for 21 patients—11 in the test group and 10 in the control group. Three patients did not complete the study. Although the sample size was small, the difference in demineralization directly around the brackets for the test group (fluoride-releasing glass ionomer) and the control group (nonfluoride composite resin) was sizeable.

Most patients (14 of 21, or 67%) remembered to return their saliva logs with the day and time of the at-home saliva collection. There was a lower rate of return for the toothpaste and the tooth-brushing logs because many mistook them as gifts. However, for the 7 tubes of toothpaste returned (7 of 21, 33%), there was a decrease in the volume of toothpaste for the 4-week period that suggested daily usage.

The microhardness testing results (Table) showed significantly more decalcification in the nonfluoride control group, despite an extreme outlier in the test group (not shown in the Table). When the data were analyzed, excluding this outlier, with a 2-sample t test with equal variances, there was a high statistical significance of P = .0002 between the test and the control groups. The ΔZ value for the outlier in the test group (subject #21) was 1853, indicating considerable demineralization. The statistical significance between the groups was evaluated for the data, including the extreme outlier, with the 2-sided Wilcoxon-Mann-Whitney test for nonparametric data. Even with this outlier included in the analysis, the statistical significance for the difference between the test and the control groups was P = .0048, which is still highly significant with P < .005.

Table. Relative mineral loss (ΔZ), volume percentage × μm, as values per subject for control and test groups

	Subject #	Control ΔZ Vol% × μm	Subject #	Test ΔZ Vol% × μm
	2	360	1	267
	3	897	5	−58
	4	274	7	512
	6	1228	9	−135
	8	1194	11	575
	10	712	13	490
	12	766	14	−227
	15	795	18	−292
	17	987	19	219
	16	842	20	244
	Mean (SD)	805 (310)	Mean (SD)	160 (319)

Outlier (subject #21) excluded from test group.

For the whole saliva fluoride analysis, 8 saliva samples were collected from each patient. There was no significant difference between the test and the control groups for intraoral whole saliva fluoride levels (P = .06) when the data at each time point were analyzed with a 2-sample t test assuming equal variances.

Discussion 

The hypothesis of this study was that carious lesions around brackets can be prevented or minimized by using a fluoride-releasing glass ionomer cement in addition to daily fluoride dentifrice use. The high statistical significance of the results indicates that the sample size for the 2 groups was adequate. The use of the fluoride dentifrice alone allowed measurable demineralization (ΔZ > 300) after only a month for 9 of the 10 subjects in the control group. This result was comparable with that previously reported by O'Reilly and Featherstone.6 The striking result is that the test group with the fluoride-releasing glass ionomer cement had significantly less demineralization, with 7 of the 11 subjects demonstrating effectively no demineralization (ΔZ < 300), 3 showing minimal demineralization (ΔZ > 300), and 1 outlier having marked mineral loss (ΔZ=1853). The 3 subjects who had minimal demineralization illustrate that, even with local fluoride therapy, a bacterial challenge can overcome the protection afforded by fluoride and remineralization. However, the demineralization in these subjects was inconsequential for this time period. The microhardness data were analyzed with and without the extreme outlier to test for her effect. That a test group subject had a carious lesion significantly larger than that of any other participant, even those in the control group, most likely indicates that a high caries challenge can overcome even the local protective effects of the fluoride from the glass ionomer. Based on discussions with the parents and the subject, we discovered that, during the test period, this 13-year-old patient was free from parental supervision much of the time, participated in more than the usual physical activity, and had markedly increased her sugar consumption.

A Wilcoxon signed rank test was performed to assess the potential difference between the 2 groups while taking into account and minimizing the distorting effect of this outlier, and the result showed a statistical significance of P < .005. This high statistical significance was even more impressive when this outlier was excluded, and the data were analyzed with a t test for equal variances (P = .0002).

Compliance with returning written logs and toothpaste was lower than expected, but the investigator did not stress the importance of returning them, and the patients were not reminded. The emphasis was placed on compliance with appointment days and times, and only 2 patients missed a single appointment each, so that the failure rate was 2% for 105 total appointments. This suggests that most patients adhered to the research protocol, and any differences in compliance were minimized by the randomization process.

The microhardness results show that teeth bonded with glass ionomer cement have significantly less enamel mineral loss when compared with teeth bonded with composite resin in a group of adolescents. This suggests that, at least in the short term, teeth bonded with glass ionomers are significantly more resistant to demineralization because of dental caries than those bonded with composite, even in patients known for their high caries risk.12, 13 Often patients have braces on all teeth, not just 2 study teeth, with wires and elastics compounding the plaque buildup, so that the difference in the effect of the 2 bonding agents would probably be even more apparent.

The secondary aim of this study was to assess whether the fluoride-releasing effect (if any) of the glass ionomer was restricted to the area around the bracket, or whether there was a whole-mouth increase in fluoride levels. All patients were asked not to use any fluoride supplements during the study with the exception of the fluoride-containing toothpaste they were given for the study.

It has been shown in many in vitro studies that glass ionomers exhibit a high initial release of fluoride that then tapers off. In this study, no such trends were found as measured by fluoride in whole mixed saliva in the mouth, and no statistically significant differences at any time between the test and the control groups. This suggests that the fluoride released from the cement was not enough to affect the whole mouth; thus, the effects seen in the microhardness studies were local. It is likely that intraoral fluoride levels would be higher if there had been brackets on every tooth as is typical during orthodontic treatment.

Despite the absence of a more global, whole-mouth effect on the fluoride levels in patients bonded with glass ionomers, the results from the microhardness testing clearly indicate that the local ability of fluoride released from glass ionomer cement is significant when compared with traditional composite resin cement. The onset of demineralization due to dental caries on the tooth surface around the bracket margin was successfully inhibited in the test group using a fluoride-releasing glass ionomer.

Conclusions 

The use of a glass ionomer cement significantly reduced enamel mineral loss due to dental caries around orthodontic brackets in patients' mouths compared with composite resin during a 4-week period.