Clinical management and prevention of dental caries in athletes: A four-year randomized controlled clinical trial

The aims of this four-year randomized controlled clinical trial were to gain insights into management and prevention of dental caries and the effect of stannous fluoride products in athletes. Fifty-four participants were randomized into test and control groups. The test group used special stannous fluoride products. The primary endpoint dental caries was assessed by the ICDAS-II-System and analyzed both by a linear mixed model for repeated measures and a generalized linear mixed model. During the observation period an increase in caries-free surfaces from 64.91 ± 6.42 at baseline to 73.22 ± 4.43 was observed. In surfaces with caries superficialis and caries media, a decrease from 13.94 ± 5.70 and 2.96 ± 2.55 surfaces at baseline to 7.89 ± 3.18 and 0.46 ± 0.78 after 2.5 years was noted, respectively. The analysis showed no effect of stannous fluoride products, but a significant difference for the time of examination (p < 0.0001). In addition, it could be shown that at any time of examination, the odds of developing caries media on a new surface was significantly lower than at baseline (up to 25-times). Due to biannual dental examinations, professional tooth cleaning and restorative treatment the number of caries-free surfaces increased and the odds of a new surface to be afflicted with caries media decreased 25-fold.

In recent years there has been an increasing interest in the oral health of elite and amateur athletes. Although the research in this field is limited, there is a general consensus that oral health of athletes is poor and can be compared to non-athletes with low socio-economic status 1,2 . The main dental problems of the athletes are caries (15-75%), dental erosion (36-85%), trauma (14-47%), periodontal problems (15%) and pericoronitis/impacted third molars. Furthermore, in some studies, a negative impact of poor oral health or trauma on sports performance was reported 3 . This impact on performance might arise from pain, systemic inflammation due to active periodontitis 4 , and/or reduced confidence because of an impaired oral status 5 . The crucial factors that influence the oral health of athletes are nutritional factors, host regulation and behavioural characteristics 1,2 . Training and exercise are often related to frequent consumption of high carbohydrate-containing sports nutrition or sports drinks 6,7 . High carbohydrate intake promotes the development of carious lesions, whereas acidic sports drinks with low pH contribute to the development of dental erosion [8][9][10] . A fluid and electrolyte deficit during exercise, due to water and sweat loss, may affect exercise performance, reduces salivary flow and induces dehydration and a dry mouth 11,12 . In this case, the negative effect of high carbohydrate-containing sports nutrition or sports drinks on oral health is assumed to increase manifold.
In contrast to complex and time-consuming therapeutic interventions, prevention of dental caries can be achieved easily and cost-effectively. In caries prevention, topical application of fluoride compounds is considered to be the most important preventive agent. On the one hand, fluoride ions can be partly adsorbed and incorporated into the crystal surface of the enamel and offer direct protection against demineralization of the tooth hard tissue 13 . On the other hand, in contact with calcium-containing saliva and the tooth surface, fluorides can form a calcium-fluoride precipitate on the tooth surface. These calcium-fluoride deposits are capable of reducing the SCIEnTIFIC REPORTS | (2018) 8:16991 | DOI: 10.1038/s41598-018-34777-x surface micro-hardness of enamel, and thus counteracting the carious process 13,14 . With regard to the prevention of erosive wear due to acidic food or beverages, it is known that the combination of tin and fluoride shows an even higher efficacy in comparison to tin or fluoride used alone 15,16 . Stannous fluoride reacts with hydroxyapatite to form CaF 2 , Sn 2 OHPO 4 , Sn 3 F 3 PO 4 , and Ca(SnF 3 ) 2 17 . The tin ions precipitate at the tooth surface and within the acquired enamel pellicle to form a protective layer, which is more acid-resistant than pure CaF 2 18,19 . Thus, according to the Patient, Intervention, Comparison, Outcome format (PICO) the aim of this randomized controlled clinical trial (RCT) was to first gain insights into the clinical management and prevention of dental caries in a population of athletes. The intervention of the RCT includes: i) A randomized daily usage of test products: special stannous fluoride toothpaste/mouth rinse and ii) half-year follow-up appointments for all participants; and professional tooth cleaning and oral hygiene instructions for all participants. The control group used their conventional fluoride-containing oral hygiene products. Related to the outcomes it was hypothesized that structured biannual dental visits in combination with professional tooth cleaning and oral hygiene instructions might affect caries experience. Furthermore, it was hypothesized that the use of special stannous fluoride toothpaste/mouth rinse might have an impact on caries development. Based on the data, the aim was to give recommendations for caries disease management in sports medicine and dentistry. Of the 54 athletes, 41 were male and 13 were female. The mean age of the athletes was 36.53 ± 9.49 years (range 20-60 years). To compare the similarities between the test and control group according to the inclusion criteria, general data and nutritional habits are depicted in Table 1. Table 2 shows the number of teeth, the number of missing teeth, the surfaces with direct and indirect restorations (ICDAS F 3+4+6 ) as well as the caries prevalence at baseline. Of the altogether 54 test subjects at baseline, 29 (53.70%) participated in all follow-up examinations. With 19 patients withdrawing from the study, the overall dropout rate was 35.19% with 7 subjects resigning from the control, and 12 subjects from the test group. Figure 1 depicts the CONSORT flow diagram, the adverse events as well as the reasons for withdrawal (16). Six control subjects missed one follow-up appointment each, but continued to participate in the remaining follow-up examinations afterwards.

Development of caries-free surfaces (ICDAS D0).
At baseline the mean of caries-free surfaces of the participants was 64.91 +/− 6.42. During the observation period an increase in caries-free surfaces was observed up to the 2.5 year recall (t = 2.5) showing 73.22 +/− 4.43 caries-free surfaces. After two and a half years, the number of caries-free surfaces remained constant, and the mean of caries-free surfaces recorded at the four-year recall (t = 4.0) was 72.66 +/− 6.03.
The analysis by the linear mixed models for repeated measures (MMRM) showed a significant difference for the time of examination (p < 0.0001, Table 4). Group-specific differences were not observed, as the intervention with special stannous fluoride-containing oral hygiene products did not show a statistical significant effect on the development of caries-free surfaces (p = 0.8687, Table 4). The analysis by the linear mixed model for repeated measures (MMRM) showed a significant difference for the time of examination (p < 0.0001, Table 3). Group-specific differences were not observed, as the intervention with special stannous fluoride-containing oral hygiene products did not show a statistically significant effect on the development of surfaces with caries superficialis (p = 0.5669, Table 4).

Development of carious surface: Caries media (ICDAS D 3+4)
. With regard to caries media (ICDAS D 3+4 ) lower prevalence was recorded in comparison to caries superficialis ( Fig. 3). At baseline the surfaces with caries media of the participants showed a mean of 2.96 +/− 2.55. Due to subsequent restorative treatment of surfaces with caries media, their number decreased over the four-year observation period. After two years the surfaces with caries media showed a mean of 0.73 +/− 1.13, and after the four-year observation period a mean of 0.46 +/− 0.78 (Table 3).
Analysis by the generalized linear mixed model (GLMM) showed a significant difference for the time of examination (p < 0.0001). In addition, the odds ratios for the time of examination were calculated and it could be shown that at any time of examination, the odds of developing a new surface with caries media was significantly lower than at baseline (Table 5). Associated with periodic dental examinations, the odds of developing a new surface with caries media was four times lower after six months, 17x lower after two years, and 25x lower after four years ( Table 5).

Development of carious surfaces: Caries profunda (ICDAS D 5+6 ). Deep and cavitated carious
lesions, caries profunda (D 5+6 ), were only recorded in individual participants (Fig. 3) and restorative treatment was subsequently carried out. The mean number of surfaces was low throughout the four-year observation period and variations could not be observed. Therefore, only a descriptive analysis can be provided, since the GLMM did not converge. After three (t = 3.0) and three and a half years (t = 3.5) no caries profunda was recorded. The maximum value of surfaces with caries profunda recorded in this study was n = 4 after two years of examination (t = 2.0). At any other time points, the number of surfaces with caries profunda was n = 1.

Discussion
This four-year randomized controlled clinical trial is the first investigation that gives insight into management strategies for the prevention of caries in athletes over a four-year period. To the best of our knowledge, no comparable data on longitudinal observations in caries management exist in the literature so far 2 .
With special regard to caries prevalence, we used the ICDAS-II score which allows for detection of different stages of carious decay: no decay, superficial decay, medium decay and deep lesions. Therefore, where preventive care is indicated, a distinction can be made between lesions that require restorative treatment and lesions that are located in the enamel without cavitation 20,21 . Also the reversal of initial carious lesions due to remineralization of the enamel is detectable as we could see a decrease in surfaces with caries superficialis from 13 baseline to 8.23 +/− 4.24 after four years. The remarkable increase of caries-free surfaces has to be interpreted in the context of restorative treatments as the scoring of the tooth surface switches from an unrestored (F 0 ) and decayed surface (D 3,4,5 or 6 ) to a restored (F 3,4, or 6 ) and caries-free surface (D 0 ). Nevertheless, the ICDAS II-Score is a valid and reliable tool to record and monitor carious lesions at different stages 22,23 , and therefore is the right tool for data assessment in longitudinal clinical studies that focus on caries development 21 .
In this study, it could be shown that the time of examination had significant influence on the increase of sound enamel surfaces and on the decrease of surfaces with carious lesions (p < 0.0001, Tables 3 and 4). Up to 8.9 +/− 1.05 additional caries-free surfaces were recorded after three and a half years of observation (t = 3.5). Regular dental visits in combination with professional tooth cleaning and oral hygiene instruction seem to have a beneficial effect on caries management in the population investigated here. However, in this population and under the conditions of this trial the special stannous-fluoride test products did not show an additional effect on caries development ( Table 4). The athletes in this investigation belong to a group with a high awareness of (oral) health and represent a group with high socio-economic status. The additional effect of stannous fluoride might have been visible if the participants of this study would have been chosen from a social deprived population with a high caries risk. Compared to the age group of 35-44 years of the German Study on Oral Health conducted in 2016 it was obvious that a low socio-economic status is associated with a 2.2 points higher DMFT-index.
The significant increase in caries-free surfaces and reduction in surfaces with caries media and caries profunda was due to restorative treatment of carious surfaces. All participants with carious lesions requiring treatment (surfaces with ICDAS scores D 3, D 4 , D 5 and D 6 ) were scheduled for treatment appointments in our department, or they consulted their family dentist. The results show that surfaces with caries media could be reduced significantly    (Fig. 3). Additionally, it could be shown that at any time of examination, the odds of developing a new surface with caries media was significantly lower that at baseline (Table 5). Associated with periodic dental examinations, the risk of developing a new surface with caries media was four times lower after six months, 17x lower after two years and 25x lower after four years (Table 5). This leads to the conclusion that regular dental care for athletes reduces the caries risk of the individual. When comparing the results of caries prevalence in the athletes from this longitudinal study with the German Study on Oral Health conducted in 2016, the caries prevalence of athletes after the four-year period were distinctly lower than in the German population (DMFT-values) (Jordan and Micheelis, 2016). Another known risk factor for caries is the frequent consumption of carbohydrate-containing sports nutrition to maintain athletic performance 11,24 . At baseline, the nutritional habits during training were not significant different between the two groups (Table 1). Over the observation period, a significant reduction of carious surfaces was observed in all participants, as explained above. The participants, who strictly adhered to the repetitive dental visits in this RCT, acquired an increased awareness of oral health related to therapeutic interventions, effective oral hygiene and/or dietary interventions 25 . As a further strategy to keep the number of surfaces with carious Figure 1. Depicts the CONSORT flow diagram, the adverse events as well as the reasons for withdrawal. Data were analysed according to the "Intention-to-treat analysis". With 19 patients withdrawing from the study, the overall dropout rate was 35.19% with 7 subjects resigning from the control, and 12 subjects from the test group. Six control subjects missed one follow-up appointment each, but continued to participate in the remaining follow-up examinations afterwards. Drawing conclusions from these findings, we became aware of the following weaknesses and limitations of the study: i) The number of test subjects and controls that were included does not allow to draw confirmatory conclusions from the data; ii) in RCTs evaluating oral hygiene formulations over years, the choice of controls is problematic, because a placebo or negative control (here fluoride-free toothpaste) is unethical, and therefore not possible and iii) between the two groups the loss to follow-up is unequal over the four years of investigation. With 19 patients withdrawing from the study, the overall dropout rate was 35.19% with 7 subjects resigning from the control, and 12 subjects from the test group. Six control subjects missed one follow-up appointment each, but continued to participate in the remaining follow-up examinations afterwards (Fig. 1).
Despite all limitations, we tried to provide a structured prevention regimen and regular follow-up appointments for all participants, in order to raise awareness of oral health. However, there were many reasons to drop-out. Some of the participants, being university students, moved away or were unwilling to participate any longer, two women got pregnant (both test group) and one participant suffered from a chronic disease and was not able to participate any longer (test group). Furthermore, limitations of clinical monitoring using the ICDAS score were the proximal surfaces of the teeth. First, in contact with neighboring teeth a sufficient and valid monitoring was not possible. Second high difficulties in obtaining ethical approval for taking systematically radiographs were existent and therefore a longitudinal monitoring of approximal surfaces was not included in the study. To address these shortcomings, further clinical research and, if possible, greater cohorts of participants would be preferable to support the data obtained in this study. It can be concluded that caries-free surfaces increased, due to regular biannual dental examinations with restorative treatment, professional tooth cleaning and oral hygiene instructions, and the odds of developing a new surface with caries media decreased to the factor 25 after four years. Therefore, management and prevention of caries in athletes might be realized and can be summarized as follows: i) periodic scheduling of athletes: 6-months intervals; ii) professional tooth cleaning and oral hygiene instructions; iii) general recommendation of fluoride-containing oral hygiene products. The increase of caries-free surfaces has to be interpreted in the context of restorative treatments as the scoring of the tooth surface switches from an unrestored (F 0 ) and decayed surface (D 3,4,5 or 6 ) to a restored (F 3,4, or 6 ) and caries-free surface (D 0 ).  Table 4. Results of the linear mixed models for repeated measures (MMRM) for caries-free surfaces and surfaces with caries superficialis. Fifty-four athletes from sports clubs or the university, who perform five or more hours of endurance training per week, participated in this investigation and gave written informed consent. The study was officially announced in the sports clubs and at the university, and athletes that met the inclusion criteria and were willing to participate were invited for study appointments. The participants were randomized by block randomization (sequentially numbered envelopes) into either test or control group, as is depicted in the CONSORT flow diagram (Fig. 1) 27 . The random allocation sequence was generated by the statistician; enrolment and assignment of participants to interventions was done by the principal investigator. The participants in the test group were instructed to use a special stannous fluoride-containing [(AmF)/NaF/SnCl 2 ] mouth rinse (500 ppm F -, 800 ppm Sn 2+ ), 1 × 30 s per day and a special toothpaste containing NaF/Sn(2 + ) and the biopolymer chitosan (elmex EROSIONSSCHUTZ, CPGABA GmbH, Hamburg, Germany) for daily oral hygiene at home 28,29 .

Caries-free surfaces Standard
The participants in the control group did not get any products with the exception of the instruction to use fluoridated toothpaste (1500 ppm) with their conventional oral hygiene products at home. They were explicitly told not to use stannous fluoride containing oral hygiene products.
Inclusion and exclusion criteria. The inclusion criteria were: participants who i) were older than 18 years, ii) gave written informed consent, iii) declared that they performed endurance sports (running, cycling, swimming, track and field, triathlon) with a cumulative weekly training time of five or more hours, and iv) were in good general health and not restricted in practicing oral hygiene.
The exclusion criteria were: participants who (i) were under the age of 18, ii) gave no written informed consent, iii) performed a cumulative weekly training of less than five hours, iv) were restricted in practicing oral hygiene, v) were pregnant or nursing, vi) had been or were still taking part in another clinical study within the last 30 days, vii) took antibiotics within the last 30 days, viii) were dental students or dental staff members.
Clinical oral investigation. One blinded calibrated examiner (TW) performed the standardized clinical examination every six months. The study protocol was comprised of anamnesis in written form, intraoral inspection, assessment of caries index (ICDAS II), standardized photographs and a professional tooth cleaning session including oral hygiene instructions. The aim of this preventive setting was to generate an increased awareness of oral health related to therapeutic interventions, effective oral hygiene and/or dietary interventions. Intraoral examination was done after professional tooth cleaning using a dental operating light, binocular loupes (magnification 2.5x), plain mirrors, and diagnostic probes. Calibration was done using the ICDAS e-learning and online calibration tool.
Assessment of caries prevalence. For assessment of caries prevalence, the International Caries Detection and Assessment System II (ICDAS II) was applied on three sites per tooth (buccal, oral, occlusal/incisal) 30 . In individual cases when valid medical indications were seen, bitewing radiographs were taken. However due to the inhomogeneity of this data, it could not be considered in the statistical analysis. ICDAS II values were obtained at every recall appointment, and depending on the number of teeth, a maximum of 84 surfaces could be recorded per participant. Using the ICDAS criteria, the sites were recorded by a 0 to 6 scoring system: 0 = sound, no caries; 1 = first visible sign of non-cavitated caries after 5 sec air-drying, caries superficialis; 2 = first visible sign of non-cavitated caries when tooth is wet and dry, caries superficialis; 3 = microcavitation in enamel, caries media; 4 = caries lesion extending into dentine (underlying shadow), caries media; 5 = small cavitated lesion, and dentine is visible in <50% of the surface, caries profunda; 6 = large cavitated lesions, and dentine is visible >50% of the surface, caries profunda. If carious lesions requiring treatment were observed (ICDAS codes D 3-6 ), subsequent caries removal and restorative treatment was carried out. Statistical analysis. This study, being a RCT with a four-year observation period, had to be prepared for exclusion and Nstudy withdrawal, and the subsequent loss of test persons' data. The number of subjects resigning from the study, as well as their reasons for doing so, were documented in the CONSORT flow diagram (Fig. 1). At the end of the trial all 54 patients were analyzed according to the intention-to-treat principle (ITT), i.e. all randomized patients were included regardless of any protocol violations and analyzed according to the group they were allocated to originally. Hence, also those patients who dropped out during the trial or were lost to follow-up were included in the analyses. The data were analyzed using descriptive statistics, evaluating mean, standard deviation, minimum, median and maximum for all baseline and follow-up parameters.
The statistical analysis of the ICDAS II system with regard to the carious decay of athletes was assessed by calculating cumulative D-S 1-6 indices combining the cut-off points: D 0 -S (ICDAS code 0), D 1+2 -S (ICDAS codes 1 and 2 representing caries superficialis), D 3+4 -S (ICDAS codes 3 and 4, representing caries media) and D 5+6 -S (ICDAS codes 5 and 6, representing caries profunda). The primary endpoint dental caries for the ICDAS codes D 0 and D 1+2 was analyzed by a linear mixed model for repeated measures (MMRM), which allowed us to include patients with an incomplete follow-up into the analysis. Due to the non-continuous character of the ICDAS codes D 3+4 and D 5+6 a generalized linear mixed model (GLMM) with the dependent variable "number of surfaces with caries media/profunda >0" (yes/no) was used. Effect estimates for comparing each time point between groups were calculated by Least Squares MEANS (LSMEANS) statements together with the 95% confidence intervals and descriptive p-values.
To compare the time points of investigation, odds ratios were calculated by the use of LSMEANS. Further pairwise comparisons between the test and control group were made using U-test for continuous and ordinal data, while categorical data were compared using the χ 2 -test. All analyses were done using the software package SAS ® System 9.4 (SAS Inc., Cary/NC, USA), or higher. RStudio Desktop 1.1.383 was used to create the graphics.