Material and methods

Eligible participants were selected from patients of both sexes who were having orthodontic treatment with fixed appliances for less than 16 months at the orthodontics clinic at School of Dentistry of Ribeirão Preto, University of São Paulo, Brazil; had good general health; and had not used antibiotics or antimicrobial mouthwashes within 3 months before the study. Thirty-three patients (ages, 11-33 years) who met these inclusion criteria were enrolled as subjects. The study purposes were fully explained to the patients or their legal representatives, who signed a written informed consent form for participation. The research protocol was reviewed and approved by the institutional research ethics committee (process number 2008.1.163.58.8).

One week before the beginning of the study, each patient’s plaque index was determined by 1 operator (M.C.D.A.) according to the method of Silness and Löe to confirm that all patients had a similar amount of bacterial biofilm. Bacterial biofilm is a more accurate description of bacterial colonization, yet refers to the same clinical phenomenon. Biofilm deposits were eliminated with meticulous rubber cup and pumice prophylaxis. The patients were instructed to brush their teeth 3 times a day after meals using a toothbrush (Professional, Colgate-Palmolive Indústria e Comércio, São Paulo, São Paulo, Brazil) and fluoride-containing dentifrice (Colgate Máxima Proteção Anticáries, Colgate-Palmolive Indústria e Comércio) supplied by the researchers throughout the experimental period.

The 33 patients were randomized into 2 groups by using SAS statistical software (version 9.1.3 for Windows; SAS Institute, Cary, NC). In all patients, 3 new sterile edgewise metallic orthodontic brackets (0.022 × 0.028-in slot) (Generus, GAC International, Bohemia, NY) were bonded with orthodontic light-cured adhesive (Transbond XT, 3M Unitek, Monrovia, Calif) to 3 premolars selected randomly by using the SAS software. Sixteen patients used a 0.12% chlorhexidine gluconate mouthwash (Periogard, Colgate-Palmolive, São Bernardo do Campo, São Paulo, Brazil) as an antimicrobial agent, and 17 patients used a placebo mouthwash (Fármacia de Manipulação Doce Erva, Ribeirão Preto, São Paulo, Brazil) with a similar color, taste, and composition to Periogard, without chlorhexidine, as the control. In both groups of patients, oral rinses were done with 10 mL of the test solution for 30 seconds twice a week (Tuesdays and Fridays) for 30 days. On Tuesdays, mouth rinsing was performed at the dental school under the researcher’s (M.C.D.A.) supervision; on Fridays, mouth rinsing was performed at the patient’s home (under parental supervision, for children) at night 1 hour after toothbrushing. Both solutions were stored in 120-mL plastic bottles and were given to the subjects the week after the prophylaxis. The patients were blinded to which mouthwash was being used.

After 30 days, the brackets were removed by 1 trained operator (M.C.D.A.) in a blinded fashion and sent for analysis. Two brackets from each patient of both groups were subjected to analysis by the checkerboard DNA-DNA hybridization technique. For that purpose, they were placed into individually labeled plastic tubes containing 150 μL of Tris-EDTA (TE) buffer solution (pH 7.6) and 100 μL of 0.5 mol/L of sodium hydroxide (NaOH) to promote lysis of the bacterial cells. After this procedure, the plastic tubes were agitated in a shaker (Mixtron; Toptronix, São Paulo, São Paulo, Brazil) for 30 seconds for microbial desorption. Next, the brackets were removed with sterile clinical pliers, and the plastic tubes containing the bacterial suspension were stored frozen at −20°C for further analysis. The other bracket was subjected to quantification of bacterial endotoxin. It was immersed in labeled nonpyrogenic microcentrifugal plastic tubes containing 100 μL of pyrogen-free water (Milli-Q; Millipore, Billerica, Mass) and vigorously agitated in a shaker for 30 seconds for material desorption. Next, the bracket was removed, and the plastic tubes containing the bacterial suspension were stored frozen at −20°C for further analysis by the limulus amebocyte lysate assay (QCL-1000; Cambrex Bio Science Walkersville, Walkersville, Md).

The presence of 16 gram-negative periodontal pathogenic microorganisms of the orange and red complexes and T socranskii were determined in each sample by using the checkerboard DNA-DNA hybridization method described by Socransky et al.

Briefly, the samples were boiled for 10 minutes and neutralized with 0.8 mL of 5 mol/L of ammonium acetate. The released DNA was then fixed in individual lanes of a positively charged nylon membrane (Boehringer Mannheim, Indianapolis, Ind) by using a checkerboard slot blot device (Minislot 30, Immunetics, Cambridge, Mass). Sixteen digoxigenin-labeled whole genomic DNA probes (Roche Applied Science, Indianapolis, Ind) were constructed and hybridized perpendicularly to the lanes of the clinical samples by using a Miniblotter 45 apparatus (Immunetics). Bound probes were detected with phosphatase-conjugated antibody to digoxigenin (Roche Applied Science). After incubation in a solution containing the CDP-Star substratum (Amersham Pharmacia Biotech, Piscataway, NJ), the membranes were placed in an autoradiography cassette under a radiographic film (X-Omat, Kodak, Rochester, NY) developed for chemiluminescence signal detection. Signals were evaluated visually by comparing with the standards of 10 ⁵ and 10 ⁶ bacterial cells of the test species on the last 2 lanes of the same membrane. This provided the approximate number of cells per sample for each bacterial strain evaluated; this was equal to the sum of the values obtained in the 2 brackets removed from each patient. The data were read twice by a blinded examiner (M.F.) (kappa, >0.8). The sensitivity of this assay was settled to allow detection of 10 ⁴ cells of a bacterial species by adjusting the concentration of each DNA probe. This procedure was carried out to provide the same sensitivity of detection for all species (ie, the concentrations were adjusted so that all probes had a similar signal intensity). To facilitate the semi-quantitative examination of chemiluminescence signals for each microorganism in each sample, the intensity of the contamination of the brackets by the different bacterial species was evaluated at the following levels: 0 (not detected), 1 × 10 ⁴ , 1 × 10 ⁵ , 5 × 10 ⁵ , 1 × 10 ⁶ , and 1 × 10 ⁷ .

The amount of bacterial endotoxin present on the metallic brackets was quantified with the limulus amebocyte lysate assay, according to the manufacturer’s instructions. A standard curve was used to determine the concentration of endotoxin in the samples collected from the brackets. Briefly, 50 μL of solutions prepared at known bacterial endotoxin concentrations in endotoxin unit per milliliter (EU/mL) (1.0 EU/mL, 0.5 EU/mL, 0.25 EU/mL, 0.1 EU/mL), 50 μL of negative control (nonpyrogenic water), and 50 μL of each sample obtained from the brackets (defrosted and diluted in nonpyrogenic water at a ratio of 1:50,000, as determined in a pilot experiment) were pipetted in duplicate in the wells of a nonpyrogenic 96-well plate (Corning, Corning, NY). Next, 50 μL of limulus amebocyte lysate were added to each well, and the plates were agitated for 15 seconds in the shaker (Mixtron) and incubated at 37°C for 10 minutes, and 100 μL of chromogenic substrate preheated to 37°C were added to each well followed by agitation and incubation at 37°C for an additional 6 minutes in a light-proof environment. Finally, 100 μL of block reagent (25% v/v glacial acetic acid in water) were added to each well to interrupt the reaction. The absorbance of each sample was measured in an ELISA plate reader (Bio-Rad 450; Ultramark, Tokyo, Japan) at a wavelength of 410 nm, and the concentration of bacterial endotoxin was expressed as EU per milliliter.