Effect of Erythritol and Xylitol on Dental Caries Prevention in Children

Transcription

1 Original Paper Received: August 1, 2013 Accepted after revision: January 7, 2014 Published online: May 21, 2014 Effect of Erythritol and Xylitol on Dental Caries Prevention in Children Sisko Honkala a Riina Runnel b Mare Saag b Jana Olak b Rita Nõmmela b Silvia Russak b Pirkko-Liisa Mäkinen c Tero Vahlberg d Gwen Falony e, f Kauko Mäkinen c Eino Honkala a, c a Faculty of Dentistry, Kuwait University, Safat, Kuwait; b Department of Stomatology, Faculty of Medicine, University of Tartu, Tartu, Estonia; c Institute of Dentistry and d Department of Biostatistics, Faculty of Medicine, University of Turku, Turku, Finland; e Department of Structural Biology and f Department of Bioscience Engineering, Vrije Universiteit Brussel, Brussels, Belgium Key Words Clinical trial Dental decay Double-blind study Estonia Intervention study Sugar alcohol Abstract Objective: The aim of this study was to test the efficacy of long-term, daily intake of erythritol and xylitol candy, compared with sorbitol candy, on the development of enamel and dentin caries lesions. Methods: The study was a doubleblind randomized controlled prospective clinical trial. Altogether 485 primary school children, first- and second-graders at baseline, from southeastern Estonia participated in this 3-year intervention. Each child consumed four erythritol, xylitol or sorbitol (control) candies three times per school day. The daily intake of polyol was about 7.5 g. The International Caries Detection and Assessment System (ICDAS) was used in the clinical examinations by four calibrated examiners at baseline and at 12, 24 and 36 months. Results: The annual examination analyses and the follow-up analyses confirmed that the number of dentin caries teeth and surfaces at 24 months follow-up and surfaces at 36 months follow-up karger@karger.com S. Karger AG, Basel /14/ $39.50/0 was significantly lower in the mixed dentition in the erythritol group than in the xylitol or control group. Time of enamel/dentin caries lesions to develop and of dentin caries lesions to progress was significantly longer in the erythritol group compared to the sorbitol and xylitol groups. Also the increase in caries score was lower in the erythritol group than in the other groups. Conclusions: In the follow-up examinations, a lower number of dentin caries teeth and surfaces was found in the erythritol group than in the xylitol or control groups. Time to the development of caries lesions was longest in the erythritol group. Trial registration: ClinicalTrials.gov Identifier NCT S. Karger AG, Basel Several clinical trials have demonstrated the effectiveness of xylitol, a sugar alcohol of the pentitol type, in preventing caries [Scheinin and Mäkinen, 1975; Scheinin et al., 1985; Kandelman and Gagnon, 1990; Mäkinen et al., 1995]. The benefits of sugar-free chewing gum may evolve from decreased lactic acid production and increased salivary flow, which might increase buffering of acids in Sisko Honkala Faculty of Dentistry, Kuwait University PO Box 24923, Safat (Kuwait) hsc.edu.kw or gmail.com

2 plaque, and increased super-saturation of saliva with the mineral ions and enhanced clearance of sugars from the mouth [Mäkinen, 2010, 2011]. Increased polyol levels of dental plaque may be associated with elevated amounts of soluble calcium in plaque [Steinberg et al., 1992; Mäkinen et al., 1996]. This calcium has been suggested as contributing to the natural repair (remineralization) of incipient enamel caries lesions. In addition, xylitol has a suppressing effect on Streptococcus mutans bacteria in plaque [Trahan et al., 1992; Lingström et al., 1997]. Two field studies revealed that also the use of xylitol candy can have caries-preventive and plaque reduction effects [Alanen et al., 2000; Honkala et al., 2006]. Some studies have shown that erythritol, a tertitol type of polyol, has a similar effect on the risk factors of caries as xylitol [Kawanabe et al., 1992; Mäkinen et al., 2005]. Xylitol and erythritol are considered to be non-acidogenic substances [Mäkinen, 2010]. Some trials have suggested that sorbitol, a hexitol type of polyol, has also a cariespreventive effect [Glass, 1983; Machiulskiene et al., 2001], although sorbitol can be fermented (at a slow rate) by all streptococci [Mäkinen, 2010]. Most of the clinical trials comparing the effect of sorbitol and xylitol have shown the superiority of xylitol in caries prevention [Söderling et al., 1989; Steinberg et al., 1992; Mäkinen et al., 1996]. This study for the first time aimed at testing the efficacy of long-term, daily intake of erythritol and xylitol candy, compared with sorbitol candy, on the development of incipient caries lesions and new dentin caries lesions among primary school children. Subjects and Methods The study was set up as a double-blind randomized controlled prospective intervention trial. Sample size calculation was based on the caries data of the Kuwait xylitol study [Honkala et al., 2006] as the caries level (DMFT) among 12-year-olds in Kuwait was the same (2.7) as in Estonia [World Health Organization, 2007]. Based on the estimated difference (1.5 carious surfaces) in the caries increment between the study (2 carious surfaces) and the control (3.5 carious surfaces) groups (common standard deviation = 4) after the 2 years of intervention using α error 0.05 and β error 0.20, the sample size should have been 113 per group. When the estimated drop-out during the study was estimated as 25%, the sample size required was 151 [(1/0.75) 113] in each of the three intervention groups. The initial study population consisted of 485 first- and secondgrade primary school children enrolled from 10 schools (10% of all schools) in the region around Tartu, southeastern Estonia [Honkala et al., 2011; Runnel et al., 2013a], where the fluoride content in drinking water is low [Karro et al., 2006]. The mean age of children was 9.1 years in the erythritol group and 8.7 years in the xylitol and Table 1. Number of children (n) examined according to groups and age at baseline (2008) and at 12 months (2009), 24 months (2010) and 36 months (2011) Erythritol SorbitolXylitol Total Age: sorbitol groups. Table 1 indicates the number of children examined during the study. A CONSORT flow diagram is presented in figure 1. Only pupils whose parents/caretakers returned the signed informed consent form were included in the study. Children who did not return the signed informed consent form, did not want to continue the intervention or were away from school on the day of examination were excluded from the study. At enrollment (January to February 2008), school classes were randomly divided into three groups: erythritol (n = 165), xylitol (n = 156) and sorbitol (n = 164; control). The list of all classes from all participating schools (n = 10) was used as a sample frame. The statistician (T.V.) allocated the classrooms according to computer-generated random numbers, but inside the schools the first-grade pupils were allocated into a different group than the second-graders to reduce school bias. There were 2 5 classrooms per school in the sample. Children were assigned from 6 schools/9 classrooms in the erythritol group, from 9 schools/11 classrooms in the sorbitol group and from 9 schools/10 classrooms in the xylitol group. None of the children changed from their group to another intervention group during the 3-year intervention, although a few children changed school ( fig. 1 ). Concealment of allocation was maintained by the Cargill company up to the main statistical analyses and fixing of the data. At baseline, subjects were blindly assigned to examiners. The subject-examiner assignment was fixed for the duration of the study. Double-blind clinical examinations of the children in all groups were completed four times (at baseline and at 12, 24 and 36 months, in 2008, 2009, 2010 and 2011, respectively) by four trained and calibrated investigators (E.H., S.H., J.O., R.R.). The clinical examinations were conducted at four standard dental units of the Department of Stomatology, University of Tartu. The International Caries Detection and Assessment System (ICDAS II) was used in the clinical examinations [Ismail et al., 2007]. Enamel (Dd 1, Dd 2, Dd 3 ) and dentin caries lesions (Dd 4, Dd5, Dd6 ), filled teeth and surfaces (FfTt and FfSs) and missing teeth and surfaces (MmTt and MmSs) were determined in all four annual examinations on all surfaces of the teeth (max. 28 teeth, 128 surfaces). If a tooth was exfoliated (or extracted), all the visible surfaces from earlier assessment were recorded as missing when calculating caries increments. Calibrations were arranged during 2 days before every annual examination. Consistency of the ICDAS codes by each of the examiners and between the examiners was very high (κ > 0.9) [Runnel et al., 2013b]. During the first, second and third year, pupils consumed xylitol-, erythritol- or sorbitol-containing candies during school days ( 200 days per year). Each child consumed four candies three Effect of Erythritol and Xylitol in Caries Prevention 483

3 Enrollment Assessed for eligibility (n = 522) Randomized (n = 485) Excluded (n = 37) Reason: were not at school at the day of the examination Color version available online Allocation Erythritol group (n = 165) Received allocated intervention Sorbitol group (n = 164) Received allocated intervention Xylitol group (n = 156) Received allocated intervention Follow-up Lost to follow-up (n = 43) Discontinued intervention: changed school or were absent on the day of the examination Lost to follow-up (n = 38) Discontinued intervention: changed school or were absent on the day of the examination Lost to follow-up (n = 30) Discontinued intervention: changed school or were absent on the day of the examination Analysis Fig. 1. CONSORT flow diagram of the Tartu caries study (166/T-7), ClinicalTrials. gov Identifier NCT Analyzed (n = 122) Analyzed (n = 126) Analyzed (n = 126) times per school day, not knowing which group they belonged to. One piece of candy weighed around 0.7 g and contained about 90% of erythritol, xylitol or sorbitol. Otherwise, the contents of the candies were similar. Candies were manufactured and provided by Cargill R&D Center Europe. The daily intake of polyol was about 7.5 g; no side effects were expected or experienced with that amount. The teachers distributed and supervised the use of the products before the first lesson in the morning (8 a.m.), immediately after the school lunch (10.30 a.m.) and at the end of the school day (1.30 or 2.15 p.m.). They were trained by R.N. before the start of the intervention. Teachers received a honorarium each year for this extra work. All pupils received dental health education on oral hygiene and diet in connection with the annual examinations. Each half year, every child was also given a toothbrush and fluoride toothpaste (Colgate Total with 0.24% sodium fluoride and sorbitol as a sweetener). At each examination, children were recommended to brush their teeth more than once a day. In Estonia, among 11-yearolds about 60% report to brush more than once a day [Honkala et al., 2012]. However, toothbrushing frequency was not confirmed in this study population. A group representing the investigating team made three annual site visits to the schools during the intervention to enhance compliance of subjects to the study. These visits checked the awareness of the subjects and the teachers of the details of the program. All the classes were visited and information about the study was explained to the children. Discussion was also arranged with all the teachers involved. The study was conducted according to the ethical principles of the Declaration of Helsinki. The Research Ethics Committee of the University of Tartu approved the study (166/T-7). Approval of the School Management Authority and school principals were also received. The study was listed to the register of clinical trials (www. clinicaltrials.gov) at initiation as ClinicalTrials.gov Identifier NCT The data were analyzed by SPSS (version 19.0) and by SAS (9.2 or higher) programs. Statistical analyses were done by intention to treat, with pupils analyzed in the group to which they had been randomized. One-way analysis of variance was used to test the difference in mean ages between groups. The difference in gender distribution between groups was tested with the χ 2 test. In the analyses, mixed dentition was studied. The ICDAS codes 1, 2 and 3 were combined to enamel caries teeth (Dd 1 3 Tt) and surfaces (Dd 1 3 Ss). The ICDAS codes 4, 5 and 6 were combined to dentin caries teeth (Dd 4 6 Tt) and surfaces (Dd 4 6 Ss). Caries indices (D 4 6 MFT + d 4 6 mft and D 4 6 MFS + d 4 6 mfs) were calculated as the sum of decayed, missing and filled teeth and surfaces. The number of enamel and dentin caries teeth and surfaces, the number of teeth and surfaces with fillings and the number of caries experience teeth and surfaces at baseline and at 12, 24 and 36 months were compared between the groups using negative binomial regression. Negative binomial regression models were adjusted for gender, age (categorized as 7 years old or younger, 8 years old, 9 years or older) and school. Data were also analyzed including the classroom as a random effect (clustering effect). However, classroom effect was not statistically significant, and it did not affect the results and thus was excluded from the statistical models. The natural log of the number of teeth or surfaces present was included as an offset when analyzing the number of enamel/dentin caries and filled teeth or surfaces. Pearson χ Honkala et al.

4 goodness-of-fit statistics were used to assess the fit of the models. The changes in caries indices were analyzed with negative binomial regression using generalized estimating equations with exchangeable correlation structure to account for dependency between repeated measures. The Bonferroni method was used to adjust p values in pairwise comparisons between years. The significance level was set at p < For each surface within each subject, a survival time was performed. Analysis of the following events was conducted: (1) Time to enamel/dentin caries development: transition of caries score 0 to 1, 2, 3, 4, 5 or 6 (time = time from the start of the trial to the time of enamel/dentin caries development). (2) Time to dentin caries development: transition of caries score 0, 1, 2 or 3 to 4, 5 or 6. (3) Time to increase in caries score: transition from any caries score to increase in score of 1 or more. (4) Time to decrease in caries score: transition from any caries score to decrease in score of 1 or more. Since the exact values of the events were not available, the time was defined by lower and upper bounds. The examinations for each subject took place exactly 12 months apart. Therefore, for time to caries development, the lower bound was calculated as twelve times the number of examinations where the surface was sound. The upper bound was the lower bound plus twelve. The number of surfaces included in the analysis was presented for each intervention group. Differences in the survival times between three intervention groups, between the erythritol and the sorbitol group, and between the xylitol and the sorbitol group were tested with χ 2 test. The analysis included an accelerated failure time model of the interval-censored data using SAS Proc LIFEREG. The distribution of the data was specified as log-logistic as this allowed the rate of decay to increase or decrease over time [Hannigan et al., 2001]. The model was fitted using the maximum likelihood method and included terms for the intervention group, age of the subject, time to eruption and school. For each statistically significant term included in the model, the parameter estimate, the standard error, the p value and the acceleration factor were presented. Survival curves were generated for each intervention group for comparing time of caries development between the groups during the 3-year follow-up. R e s u l t s The mean ± SD age of the children in the erythritol group (8.6 ± 0.5 years) was slightly higher than in the xylitol group (8.2 ± 0.5 years) or in the control group (8.1 ± 0.6 years) (p < ). The percentage of girls was 43.0% in the erythritol group, 45.5% in the xylitol group and 48.2% in the control group, respectively (p = 0.645). There were no statistically significant differences between the intervention groups neither at baseline nor at 12 months in the caries indicators of the mixed dentition ( table 2 ). At 24 months, the xylitol group had a statistically significantly higher number (and percentage) of dentin caries teeth (Dd 4 6 Tt) (relative risk = 1.96, 95% CI ) and surfaces (Dd 4 6 Ss) (relative risk = 2.33, 95% CI ) than the erythritol group, and at Table 2. Total number of teeth and surfaces, number (%) of decayed and filled surfaces and mean [SEM] of DMFT/DMFS indices in the mixed dentition in 2008, 2009, 2010 and 2011 between the intervention groups Erythritol SorbitolXylitol 2011 (n = 126) 2010 (n = 132) 2009 (n = 145) 2008 (n = 156) 2011 (n = 126) 2010 (n = 137) 2009 (n = 149) 2008 (n = 164) 2011 (n = 122) 2010 (n = 132) 2009 (n = 142) 2008 (n = 165) Teeth 3,828 3,447 3,213 3,109 3,715 3,588 3,258 3,103 3,597 3,499 3,160 3,151 Surfaces 17,220 15,582 14,518 14,095 16,716 16,202 14,681 14,029 16,183 15,803 14,246 14,268 Dd1 3Tt 634 (16.6) 460 (13.3) 428 (13.3) 404 (13.0) 611 (16.4) 511 (14.2) 452 (13.9) 371 (12.0) 609 (16.9) 514 (14.7) 487 (15.4) 430 (13.6) Dd1 3Ss 835 (4.8) 610 (3.9) 548 (3.8) 508 (3.6) 790 (4.7) 671 (4.1) 590 (4.0) 464 (3.3) 806 (5.0) 674 (4.3) 623 (4.4) 547 (3.8) Dd4 6Tt 267 (7.0) 156 (4.5) 77 (2.4) a 58 (1.9) 345 (9.3) 201 (5.6) 141 (4.3) 84 (2.7) 311 (8.6) 216 (6.2) 148 (4.7) a 92 (2.9) Dd4 6Ss 468 (2.7) 232 (1.5) 110 (0.8) b 64 (0.5) c 621 (3.7) 308 (1.9) 197 (1.3) 115 (0.8) 549 (3.4) 354 (2.2) 238 (1.7) b 130 (0.9) c FfTt 774 (20.2) 598 (17.3) 467 (14.5) 380 (12.2) 601 (16.2) 630 (17.6) 545 (16.7) 408 (13.1) 614 (17.1) 608 (17.4) 503 (15.9) 389 (12.3) FfSs 1,234 (6.5) 964 (6.2) 730 (5.0) 573 (4.1) 986 (5.9) 1,015 (6.3) 842 (5.7) 638 (4.5) 1,009 (6.2) 1,003 (6.3) 795 (5.6) 559 (3.9) Caries indices d4 6mft + D4 6MFT 6.10 [0.27] 4.94 [0.25] 3.93 [0.24] 3.33 [0.25] 6.02 [0.26] 5.20 [0.25] 4.64 [0.27] 3.66 [0.26] 5.65 [0.28] 5.23 [0.27] 4.58 [0.27] 3.66 [0.22] d4 6mfs + D4 6MFS [0.65] 8.35 [0.52] 6.26 [0.47] 5.16 [0.44] [0.80] 8.75 [0.48] 7.53 [0.50] 6.23 [0.54] [0.74] 9.22 [0.57] 7.76 [0.55] 5.55 [0.40] Dd1 3Tt = Number of teeth with enamel caries; Dd1 3Ss = number of tooth surfaces with enamel caries; Dd4 6Tt = number of teeth with dentin caries; Dd4 6Ss = number of tooth surfaces with dentin caries; FfTt = number of teeth with fillings; FfSs = number of tooth surfaces with fillings; d4 6mft + D4 6MFT = sum of decayed, missing and filled teeth; d4 6mfs + D4 6MFS = sum of decayed, missing and filled tooth surfaces. For difference between groups, negative binomial regression adjusted for gender, age and school: a xylitol vs. erythritol p = 0.004; b xylitol vs. erythritol p = 0.002; c xylitol vs. erythritol p = Effect of Erythritol and Xylitol in Caries Prevention 485

5 Table 3. Pairwise comparisons (p values a ) of the differences in caries indices in mixed dentition within groups between the years 2009 and 2008, 2010 and 2008, and 2011 and 2008 Erythritol SorbitolXylitol Dd 1 3 Tt <0.001*** <0.001*** <0.001*** 0.042* 0.008** <0.001*** 0.003** ** Dd 1 3 Ss <0.001*** <0.001*** <0.001*** * <0.001*** 0.014* 0.040* 0.001** Dd 4 6 Tt <0.001*** <0.001*** <0.001*** <0.001*** <0.001*** <0.001*** <0.001*** <0.001*** <0.001*** Dd 4 6 Ss <0.001*** <0.001*** <0.001*** <0.001*** <0.001*** <0.001*** <0.001*** <0.001*** <0.001*** FfTt b <0.001*** <0.001*** <0.001*** <0.001*** FfSs c <0.001*** <0.001*** <0.001*** ** <0.001*** d 4 6 mft + D 4 6 MFT <0.001*** <0.001*** <0.001*** 0.002** <0.001*** <0.001*** 0.003** <0.001*** <0.001*** d 4 6 mfs + D 4 6 MFS <0.001*** <0.001*** <0.001*** <0.001*** <0.001*** <0.001*** 0.001** <0.001*** <0.001*** For significant difference within different intervention groups, negative binomial regression adjusted for gender, age and school using GEE estimation: * p < 0.05, ** p < 0.01, *** p < a The Bonferroni method was used to adjust p values in pairwise comparisons. b Group by time effect, p < c Group by time effect, p = Table 4. Number and percentage of surfaces according to the analysis of caries development Analysis Erythritol Sorbitol Xylitol p value a Erythritol vs. sorbitol, p value b Xylitol vs. sorbitol, p value b Enamel/dentin caries development 860/18,763 (4.6%) 1,022/18,596 (5.5%) 948/16,414 (5.8%) <0.001 < Dentin caries development 248/19,513 (1.3%) 333/19,406 (1.7%) 342/17,178 (2.0%) <0.001 < Increase in caries score 1,046/19,645 (5.3%) 1,202/19,577 (6.1%) 1,163/17,366 (6.7%) <0.001 < Decrease in caries score 401/1,313 (30.5%) 456/1,531 (29.8%) 449/1,584 (28.3%) a χ 2 test. b Fisher s exact test (two-tailed). 36 months a higher number of dentin caries surfaces (Dd 4 6 Ss) (relative risk = 1.93, 95% CI ). The changes within groups were all significant except for the number of enamel caries surfaces (Dd 1 3 Ss) in in the sorbitol group, for the number of enamel caries teeth (Dd 1 3 Tt) in in the xylitol group, for the number of restored teeth (FfTt) and surfaces (FfSs) in and in the sorbitol group and in and in the xylitol group, and for the number of restored teeth (FfTt) in in the xylitol group ( table 3 ). The differences in the changes in the number of restored teeth (FfTt, group by time effect; p = ) and surfaces (FfSs, group by time effect; p = 0.003) between groups in mixed dentition were significant. In the erythritol group, 4.6% of tooth surfaces developed enamel or dentin caries over 3-year follow-up ( table 4 ). The respective figures were 5.5% for the sorbitol and 5.8% for the xylitol group (p < ). The time of 486 enamel/dentin caries lesion to develop was significantly longer in the erythritol group compared to sorbitol and xylitol groups ( fig. 2 ). There were no statistically significant differences between the intervention groups regarding school, gender, age or time of tooth eruption. In the erythritol group, statistically significantly less enamel caries tooth surfaces (1.3%) developed to dentin caries over 3-year follow-up when compared to sorbitol (1.7%) and xylitol (2.0%) (p = ; table 4 ). The time of dentin caries lesion to progress was statistically significantly longer in the erythritol group when compared to the sorbitol and xylitol groups (p = ; fig. 3 ). The risk of caries development was higher among younger school children and among boys. The percentage of surfaces with an increase in any caries code was statistically significantly lower in the erythritol group than in the other groups (p = ; table 4 ). Decreases in caries codes were similar in all intervention groups ( %). Honkala et al.

6 Fig. 2. Time to enamel/dentin caries development between intervention groups during the 3-year follow-up. Survival distribution function estimate Months Erythritol Sorbitol Xylitol Color version available online Fig. 3. Time to dentin caries development between intervention groups during the 3-year follow-up. Survival distribution function estimate Months Erythritol Sorbitol Xylitol Color version available online D i s c u s s i o n The main finding of this 3-year clinical intervention trial was that there was a statistically significantly lower number of dentin caries surfaces and teeth in the erythritol group compared with the xylitol group in the mixed dentitions in the 24-month follow-up examinations in 2010, and a lower number of dentin caries surfaces at the end of the trial in The time to development of enamel/dentin caries lesions (surfaces) and dentin caries lesions was statistically significantly longer in the erythritol group than in the sorbitol or xylitol group. Also the increase in caries score was lower in the erythritol group than in the other groups. A previous study among 10-year-old Finns showed no additional caries-preventive effect of a school-based 2-year intervention of xylitol/maltitol (4.7/4.6 g) or erythritol/maltitol (4.5/4.2 g) lozenges in the permanent teeth of children with low caries level when compared with comprehensive prevention [Lenkkeri et al., 2012]. In most xylitol studies [Mäkinen, 2010], sugar substitutes were distributed in the form of chewing gum when chewing itself might have made a significant contribution to the observed caries reduction. Clear differences in dentin caries existed, but there were no statistically significant differences in the number of restorations or caries experience indices at the end of the intervention. The children who were detected to have any treatment need in the annual examinations were referred for restorative treatment to their own dentists. Effect of Erythritol and Xylitol in Caries Prevention 487

7 There were no indications of a difference in the cariespreventive effects between the intervention groups in the number of enamel caries lesions. This absence of effect could reflect the unstable stage of the enamel caries lesions, which can remineralize, stay the same or proceed into dentin caries. It has been reported that about 30% of enamel caries lesions on the occlusal surfaces, but only a few smooth surface lesions, progress to dentin caries over 4 years [Warren et al., 2006]. Age, gender and school were controlled in the analyses because differences existed at baseline regarding these variables. The numbers of surfaces and teeth with enamel and dentin caries, restorations and caries experience were lower at the end of the trial than at the beginning because of the natural exfoliation of the primary teeth in the mixed dentition stage. The confounding effect of exfoliation was controlled in this study by comparing the caries indices only on the tooth surfaces which were present on the two consecutive annual examinations. This way the exposure time of all teeth for the effect of the intervention was the same, namely one year. Every child was given a toothbrush and fluoride toothpaste (not containing any xylitol or erythritol) each half year. In addition to having a preventive effect on caries [Bratthall et al., 1996], fluoride in toothpaste has been shown to possess a synergistic inhibition effect with xylitol on glycolysis by mutans streptococci [Maehara et al., 2005], which might also exist with other polyols. This could partly explain the caries-preventive effect in all the intervention groups, especially since water fluoride concentration was rather low in the area where test subjects lived. The strongest caries-preventive effect on dentin caries in the erythritol group became significant only after 24 months of intervention. This might have happened because the development of dentin caries often takes 2 3 years. Erythritol has been reported to be a totally safe and promising sweetener and to have several advantages, being non-caloric, less laxative than any other polyol, including sorbitol and xylitol, non-glycemic, and having anti-oxidant properties [Munro et al., 1998; de Cock and Bechert, 2002]. The lowest solubility (longer retention time) and lowest molecular weight (higher diffusion rate) of erythritol [Munro et al., 1998] might play an important role. It has been shown in an animal study that erythritol cannot be used by S. mutans or other oral bacteria for plaque formation and lactic acid production [Kawanabe et al., 1992; Söderling and Hietala-Lenkkeri, 2010]. Erythritol has also reduced dental plaque and the number of 488 mutans streptococci in plaque and in saliva in earlier short-term trials [Mäkinen et al., 2001, 2005]. This longterm study also examined plaque weight and microbiology with results similar to those reported elsewhere [Runnel et al., 2013b]. Most of the previous clinical trials have demonstrated the caries-preventive effect of xylitol [Mäkinen, 2010]. However, in this study there was no difference in the occurrence of dentin caries between the xylitol and sorbitol groups. It was unexpected that the caries-preventive effect of xylitol compared to the control group was not significant. This might be because of the relatively mild treatment (candies given only three times per day and latest around 2 p.m.) for only 200 days per year and because good preventive and restorative care (referral) was provided to these children. In addition, only a few new dentin caries lesions developed in all the intervention groups. All intervention groups benefitted from the health education at schools and in connection of the annual examinations, and consequently oral health behavior might have improved in all groups. The teachers were very motivated in distributing candy and might also have paid much more attention to the prevention of dental caries than before. The compliance in the study was very good because the teachers distributed and supervised the distribution of candy. There was no evidence of differences in compliance between the groups. The annual site visits to the participating schools by the investigators could also have improved the motivation of the children and teachers to prevent dental caries. Caries was diagnosed in this study with the ICDAS method [Ismail et al., 2007], which has recently been accepted as a global caries detection system [EGOHID II, 2008]. Its high validity has been shown with in vitro [Jablonski-Momeni et al., 2008] and in vivo [Ekstrand et al., 2007] methodological studies. The results of the baseline examinations of our study in 2008 have been reported separately [Honkala et al., 2011; Runnel et al., 2013a]. The high consistency of the caries diagnoses was confirmed when intra- and inter-examiner repeatability levels were high. In conclusion, the present study showed a lower number of dentin caries teeth and surfaces in the mixed dentition in the erythritol group than in the xylitol and sorbitol groups after 24 and 36 months. Time to the development of caries lesions was longer and increase in caries score lower in the erythritol group than in the other groups. Sorbitol seemed to have a better caries-preventive effect than xylitol in this study population. Honkala et al.

8 Acknowledgements The funding provided by Cargill R&D Center Europe (Vilvoorde, Belgium) to this study (ClinicalTrials.gov Identifier: NCT ) is gratefully acknowledged. Special thanks go to the children who participated in the study and the teachers who helped in implementing it. The funders had no role in the study design, data collection and analysis, decision to publish or preparation of the manuscript. Author Contributions E.H., K.M. and M.S. designed the study; R.N. and M.S. organized the practicalities of the study; S.H., R.R., J.O. and E.H. performed the clinical examinations; P.L.M. and K.M. performed the biometric measurements; S.R. implemented the oral health education sessions; E.H., K.M., M.S., R.R., S.H. and P.L.M. conducted the school visits; T.V., G.F. and E.H. analyzed the data; S.H., E.H., K.M., R.R. and T.V. wrote the paper. Disclosure Statement The travel costs from Finland and the accommodation in Tartu for S.H., E.H., K.M. and P.L.M. were paid by Cargill R&D Center Europe (Vilvoorde, Belgium), which also provided a consultation payment to E.H., K.M. and M.S. The authors declare no conflict of interest with respect to the authorship and/or publication of this article. References Alanen P, Isokangas P, Gutman K: Xylitol candies in caries prevention: results of a field study in Estonian children. Community Dent Oral Epidemiol 2000; 28: Bratthall D, Hänsel-Petersson G, Sundberg H: Reasons for the caries decline: what do the experts believe? Eur J Oral Sci 1996; 104: de Cock P, Bechert CL: Erythritol. Functionality in noncaloric functional beverages. Pure Appl Chem 2002; 74: EGOHID II: Health Surveillance in Europe. Oral Health Interviews and Clinical Surveys: Guidelines. European Commission, Lyon I University Press, Available at: (accessed 2013). Ekstrand KR, Martignon S, Ricketts DJ, Qvist V: Detection and activity assessment of primary coronal caries lesions: a methodologic study. Oper Dent 2007; 32: Glass RL: A two-year clinical trial of sorbitol chewing gum. Caries Res 1983; 17: Hannigan A, O Mullane DM, Barry D, Schafer F, Roberts AJ: A re-analysis of a caries clinical trial by survival analysis. J Dent Res 2001; 80: Honkala E, Honkala S, Shyama M, Al-Mutawa SA: Field trial on caries prevention with xylitol candies among disabled school students. Caries Res 2006; 40: Honkala E, Runnel R, Honkala S, Olak J, Vahlberg T, Saag M, Mäkinen KK: Measuring dental caries in the mixed dentition by ICDAS. Int J Dent 2011; 2011: Honkala S, Honkala E, Levin KA: Part 2. Key data/ chapter 4: Oral health; in Currie C, Zanotti C, Morgan A, Currie D, de Looze M, Roberts C, Samdal O, Smith ORF, Barnekow V (eds): Social Determinants of Health and Well-Being among Young People. Health Behaviour in School-Aged Children (HBSC) Study: International Report from the 2009/2010 Survey. Copenhagen, WHO Regional Office for Europe, 2012, pp Available at: data/assets/pdf_ file/0003/163857/social-determinants-ofhealth-and-well-being-among-youngpeople.pdf (accessed 2013). Ismail AI, Sohn W, Tellez M, Amaya A, Sen A, Hasson H, Pitts NB: The International Caries Detection and Assessment System (ICDAS): an integrated system for measuring dental caries. Community Dent Oral Epidemiol 2007; 35: Jablonski-Momeni A, Stachniss V, Ricketts DN, Heinzel-Gutenbrunner M, Pieper K: Reproducibility and accuracy of the ICDAS-II for detection of occlusal caries in vitro. Caries Res 2008; 42: Kandelman D, Gagnon G: A 24-month clinical study of the incidence and progression of dental caries in relation to consumption of chewing gum containing xylitol in school preventive programs. J Dent Res 1990; 69: Karro E, Indermitte E, Saava A, Haamer K, Marandi A: Fluoride occurrence in publicly supplied drinking water in Estonia. Environ Geol 2006; 50: Kawanabe J, Hirasawa M, Takeuchi T, Oda T, Ikeda T: Noncariogenicity of erythritol as a substrate. Caries Res 1992; 26: Lenkkeri AM, Pienihäkkinen K, Hurme S, Alanen P: The caries-preventive effect of xylitol/maltitol and erythritol/maltitol lozenges: results of a double-blinded, cluster-randomized clinical trial in an area of natural fluoridation. Int J Paediatr Dent 2012; 22: Lingström P, Lundgren F, Birkhed D, Takazoe I, Frostell G: Effects of frequent mouthrinses with palatinose and xylitol on dental plaque. Eur J Oral Sci 1997; 105: Machiulskiene V, Nyvad B, Baelum V: Caries preventive effect of sugar-substituted chewing gum. Community Dent Oral Epidemiol 2001; 29: Maehara H, Iwami Y, Mayanagi H, Takahashi N: Synergistic inhibition by combination of fluoride and xylitol on glycolysis by mutans streptococci and its biochemical mechanism. Caries Res 2005; 39: Mäkinen KK: Sugar alcohols, caries incidence, and remineralization of caries lesions: a literature review. Int J Dent 2010; 2010: Mäkinen KK: Sugar alcohol sweeteners as alternatives to sugar with special consideration of xylitol. Med Princ Pract 2011; 20: Mäkinen KK, Allen P, Bennett CA, Hujoel PP, Isokangas PJ, Isotupa KP, Allen P, Mäkinen PL: Xylitol chewing gums and dental caries: a 40-month cohort study. J Dent Res 1995; 74: Mäkinen KK, Chen CY, Mäkinen PL, Bennett CA, Isokangas PJ, Isotupa KP, Pape HR Jr: Properties of whole saliva and dental plaque in relation to 40-month consumption of chewing gums containing xylitol, sorbitol and sucrose. Caries Res 1996; 30: Mäkinen KK, Isotupa KP, Kivilompolo T, Mäkinen PL, Toivanen J, Söderling E: Comparison of erythritol and xylitol saliva stimulants in the control of dental plaque and mutans streptococci. Caries Res 2001; 35: Mäkinen KK, Saag M, Isotupa KP, Olak J, Nõmmela R, Söderling E, Mäkinen PL: Similarity of the effects of erythritol and xylitol on some risk factors of dental caries. Caries Res 2005; 39: Munro IC, Berndt WO, Borzelleca JF, Flamm G, Lynch BS, Kennepohl E, Bär EA, Modderman J: Erythritol: an interpretive summary of biochemical, metabolic, toxicological and clinical data. Food Chem Toxicol 1998; 36: Runnel R, Honkala S, Honkala E, Olak J, Vahlberg T, Mäkinen KK, Saag M: Caries experience in permanent dentition among the first and second grade schoolchildren in southeastern Estonia. Acta Odontol Scand 2013a;71: Effect of Erythritol and Xylitol in Caries Prevention 489

9 Runnel R, Mäkinen KK, Honkala S, Olak J, Mäkinen PL, Nõmmela R, Vahlberg T, Honkala E, Saag M: Effect of three-year consumption of erythritol, xylitol and sorbitol candies on various plaque and salivary caries-related variables. J Dent 2013b;41: Scheinin A, Banoczy J, Szöke J, Estari I, Pienihäkkinen K, Scheinin U, Tiekso J, Zimmermann P, Hadas E: Collaborative WHO xylitol field studies in Hungary. I. Three-year caries activity in institutionalized children. Acta Odontol Scand 1985; 43: Scheinin A, Mäkinen KK (eds): Turku Sugar Studies I XXI. Acta Odontol Scand 1975; 33(suppl 70): Söderling E, Hietala-Lenkkeri AM: Xylitol and erythritol decrease adherence of polysaccharide-producing oral streptococci. Curr Microbiol 2010; 60: Söderling E, Mäkinen KK, Chen CY, Pape HR Jr, Loesche W, Mäkinen PL: Effect of sorbitol, xylitol, and xylitol/sorbitol chewing gums on dental plaque. Caries Res 1989; 23: Steinberg LM, Odusola F, Mandel ID: Remineralizing potential, antiplaque and antigingivitis effects of xylitol and sorbitol sweetened chewing gum. Clin Prev Dent 1992; 14: Trahan L, Söderling E, Drean MF, Chevrier MC, Isokangas P: Effect of xylitol consumption on the plaque-saliva distribution of mutans streptococci and the occurrence and longterm survival of xylitol-resistant strains. J Dent Res 1992; 71: Warren JJ, Levy SM, Broffitt B, Kanellis MJ: Longitudinal study of non-cavitated carious lesion progression in the primary dentition. J Public Health Dent 2006; 66: World Health Organization: Oral health country/ area profile programme, Available at: Health-Profiles/According-to-Alphabetical/ (accessed 2013). 490 Honkala et al.