Caries-preventing effect of a hydroxyapatite-toothpaste in adults: a 18-month double-blinded randomized clinical trial

Background Dental caries is a worldwide challenge for public health. The aim of this 18-month double-blinded, randomized, clinical trial was to compare the caries-preventing effect of a fluoride-free, hydroxyapatite toothpaste (test) and a toothpaste with sodium fluoride (1450 ppm fluoride; positive control) in adults. Methods The primary endpoint was the percentage of subjects showing no increase in overall Decayed Missing Filled Surfaces (DMFS) index. The study was designed as non-inferiority trial. Non-inferiority was claimed if the upper limit of the exact one-sided 95% confidence interval for the difference of the primary endpoint DMFS between test and control toothpaste was less than the predefined margin of non-inferiority (Δ ≤ 20%). Results In total, 189 adults were included in the intention-to-treat (ITT) analysis; 171 subjects finished the study per protocol (PP). According to the PP analysis, no increase in DMFS index was observed in 89.3% of subjects of the hydroxyapatite group and 87.4% of the subjects of the fluoride group. The hydroxyapatite toothpaste was not statistically inferior to a fluoride toothpaste with regard to the primary endpoint. Conclusion Hydroxyapatite was proven to be a safe and efficient anticaries agent in oral care. Clinical trial registration NCT04756557.

The modes of action of hydroxyapatite in the oral cavity are based on physical, biochemical, and biological principles (23). Two key mechanisms involved in how toothpastes prevent dental caries are inhibiting demineralization and remineralization of initially demineralized tooth surfaces (25). In this respect, several in vitro and in situ studies that investigated hydroxyapatite toothpastes have been published (26)(27)(28). It has been shown that hydroxyapatite toothpastes remineralize enamel and dentin (8-10, 29) and inhibit demineralization (9,19). Additionally, clinical studies have shown caries-preventing effects (6,7,26). Caries studies comparing hydroxyapatite and fluoride are important as fluoride has been used in oral care as an anticaries agent for decades (25,30). However, it is known that chronic exposure of fluoride from various sources (drinking water, oral care products, food, etc.) has certain drawbacks. Fluoride can cause dental fluorosis and other side effects (31- 35). As a consequence, the concentration of fluoride in toothpastes is limited and children up to 6 years are only allowed to use small amounts of toothpaste (pea size, grain of rice size) (36, 37). Reduced toothpaste amounts can lead, however, to a reduced cleaning efficacy compared with larger toothpaste amounts (i.e., full length of brush) (38).
The use of fluoride-free hydroxyapatite toothpastes has in recent years been shown to be a clinically proven approach to caries prevention (26). Paszynska et al. (6) showed the noninferiority of a hydroxyapatite toothpaste to a toothpaste with 500 ppm fluoride (as amine fluoride) in children (6). Schlagenhauf et al. (7) have proven the non-inferiority of a toothpaste with hydroxyapatite to a toothpaste with 1400 ppm fluoride [as amine fluoride and stannous(-II-)fluoride] in adolescents undergoing orthodontic therapy (7). An important advantage of using hydroxyapatite in caries prevention is its pH buffering effect and its protective calcium-and phosphate-releasing properties in cariogenic biofilms (39).
The efficacy of hydroxyapatite-based toothpastes in preventing caries has been demonstrated in adolescents (7) and children (6) but not in adult populations. To fill this gap, the aim of this randomized clinical trial was to test the caries-preventing effect of a fluoride-free, hydroxyapatite toothpaste compared with a toothpaste with 1,450 ppm fluoride in adults. While the clinical studies from Paszynska et al. (6) and Schlagenhauf et al. (7) used commercial control toothpastes, the toothpastes formulations used in this study by the two study groups (test and control) just differ in the active ingredients, i.e., hydroxyapatite and fluoride. Thus, this caries study was aimed to directly compare hydroxyapatite and fluoride in toothpastes for daily oral care.
. Materials and methods

. . Objectives
The study objective was the investigation of the cariespreventing effect in adults aged 18-45 regularly using a fluoridefree, hydroxyapatite toothpaste (test toothpaste) or the fluoridated control toothpaste. The aim of the study was to compare the caries preventive effect of the test and control toothpaste to prove the non-inferiority of the test toothpaste compared with the control toothpaste.

. . Study design
The study was performed as a two-centered, double-blinded, randomized, and active-controlled parallel-group study with two arms (control and test). The study was performed at two study centers, i.e., Universities of Medical Sciences, Poznan and Bialystok, Poland.

. . Inclusion and exclusion criteria
The following inclusion and exclusion criteria were applied:

A) Inclusion criteria
• Provision of written informed consent.

B) Exclusion criteria
Medical reasons: • Untreated caries (subjects with untreated caries in need of restoration can become eligible after restorative therapy). • Severe periodontitis at the baseline visit (pocket depth on at least one tooth ≥ 5.5 mm). • Undergoing orthodontic treatment.
• Known allergy to one of the ingredients of the study toothpastes. • Systemic disorders interfering with salivary function or flow rate. • Regular medication intake interfering with salivary function or flow rate.
Other reasons: • Participation in any other clinical study within the past 3 months or ongoing. • Lack of intellectual or physical ability to conduct the study per protocol.

. . Toothpastes (blinded)
The test toothpaste and the active control toothpaste were provided in neutral plastic tubes all having the same shape; thus, both the subjects and the investigators were blinded. It is pertinent to mention that the subjects were requested not to discuss anything related to the toothpastes with any of the study team member who was involved in the clinical examinations. The neutral plastic tubes were labeled with a random subject number for each subject. The toothpastes were produced by an external certified laboratory. The composition of the toothpastes for both groups was identical (except for the main active ingredients, i.e., hydroxyapatite and fluoride). The compositions of the study toothpastes were as follows.

. . Application
Tooth brushing was performed twice a day (morning and evening; after the meals) for 3 min per episode. No other fluorideand/or hydroxyapatite-containing oral care products, professional or home use, such as mouthwashes, gels, or fluoridate supplements, were allowed during the trial.

. . Toothbrushes
Electric (powered) toothbrushes (Oral-B; P&G, Schwalbach, Germany) were used by the subjects, and the brushing heads were changed every 2 months.

A) Decayed Missing Filled Surfaces (DMFS) index
DMFS Index calculation (41): There are five surfaces on the posterior teeth: facial, lingual, mesial, distal, and occlusal. There are four surfaces on anterior teeth: facial, lingual, mesial, and distal. The third molars were not counted.
• When a carious lesion or both a carious lesion and a restoration were present, the surface was listed as a D.
• When a tooth was extracted due to caries, it was listed as an M.
• When a permanent filling was present, or when a filling was defective but not decayed, this surface was counted as an F. Surfaces restored for reasons other than caries were not counted as an F. The total count was 128 surfaces for a maximum of 28 teeth.

B) Examinations with DIAGNOcam
DIAGNOcam was used according to the instructions of the manufacturer (KaVo Dental, Biberach, Germany). The following classification was used (42): 0 = Light transmission unchanged 1 = Shadow visible in enamel 2 = Shadow visible in dentin C) Plaque control record The Plaque Control Record (PCR) is an established method of recording the presence of the plaque on individual tooth surfaces (40). At the study visits, a plaque-disclosing solution (Gum Red-Cote, Sunstar Europe, Switzerland) was painted on tooth surfaces. After the subject had rinsed, the investigator (using an explorer or a tip of a probe) examined each stained surface for plaque accumulations at the dentogingival junction. After all teeth were examined and scored, the index was calculated by dividing the number of plaque-containing surfaces by the total number of available surfaces (Plaque Index Calculation = The number of plaque-containing surfaces/The total number of available surfaces).
. . Course of the study Visit 1: Study Day 0 Screening and baseline examination Potential subjects were informed by the investigator about the nature, significance, and scope of the clinical trial according to the . /fpubh. . requirements described in the written subject information. Before study enrollment, the willingness of the subjects to properly follow the study protocol throughout the 18 months of the study was assessed. Subjects were enrolled as study participants only when they had given their written informed consent. Subjects had to meet all inclusion criteria and no exclusion criteria. Subjects disqualified due to untreated caries became eligible after restorative therapy. Once informed consent had been given, an initial examination took place that covered the following aspects: • Screening subjects for study eligibility (inclusion and exclusion criteria). • Collection of demographic data.
• Baseline data collection performed in the following sequence: PCR, DMFS, DIAGNOcam.
After the analysis of the PCR record using a plaque-staining solution, a professional tooth cleaning was performed. The professional tooth cleaning was performed at visits 1 and 4 only, and there was no fluoride application after the cleaning.
Finally, the study subjects received an electric toothbrush with three brushing heads (replaced every 2 months) and the allocated toothpaste (test or control) by a trained study nurse or dentist not involved in clinical study examinations.
The proper use of the assigned electric toothbrush and the issued toothpaste were also instructed by this study nurse or a dentist not involved in the clinical study examinations.
Visit 2: Study day 182 1st follow-up examination One-hundred eighty-two days after the baseline visit, the PCR and DMFS were reassessed as described for the baseline visit. Subsequently, a study nurse or a dentist who has not involved in clinical study examinations handed out three new brushing heads for the electric toothbrush and a new supply of the assigned experimental toothpaste (test or control) for the next 182 days. Finally, subjects received a new appointment for visit 3.
Visit 3: Study day 364 2nd follow-up examination Three hundred sixty-four days after the baseline visit, the procedure performed on Visit 2 was repeated, after which the subjects received a new appointment for Visit 4.

Visit 4: Study day 546 Final visit
Five hundred forty-six days after the baseline visit, the following assessments were repeated: PCR, DMFS after professional cleaning, and DIAGNOcam as described from the baseline visit.
An overview of the study course in detail is presented in Table 1.

. . Methods of determining safety
Safety assessments were performed in all subjects at every visit. The clinical examiner visually examined each subject's oral cavity and perioral area as well as questioned the subjects on any adverse events. All observed or voluntarily reported adverse events (AEs) and serious adverse events (SAEs) regardless of the experimental group were recorded.

. . Monitoring
The monitor (Dr. Egmont Zieseniß, Inpharm Consulting, Germany) reviewed the case report forms with the investigator and his staff at regular intervals: 1. First monitoring visit after 30% recruitment. 2. Second monitoring visit after 12 months. 3. Third monitoring visit at study close-out.
During the study, the monitor reviewed the case report forms at the above intervals to verify completeness, plausibility and consistency of the data, protocol adherence, and the progress of enrolment, and to ensure that study supplies are being stored, dispensed, and accounted for according to specifications.

. . Determination of sample size and statistical analysis
It was assumed that the primary endpoint (i.e., no increase in DMFS index, i.e., DMFS = Visit 4 -Visit 1 ≤ 0) will occur in about 60% of study subjects. A sample size of 77 study subjects per arm (two arms) was calculated to be sufficient to reject the null hypothesis that the test toothpaste is inferior to the control toothpaste, using a non-inferiority margin of = 20% for the primary endpoint one-sided chi-square test (α = 5%, power = 80%) as described in a previous study (7). SAS 9.4 (proc power two sample freq test = PCHI) was used for the sample size calculation.
A non-inferiority testing was performed. The non-inferiority margin was set to 20% between the groups. The primary endpoint was analyzed for the per-protocol (PP) population and, in addition, repeated for sensitivity reasons for the ITT population. The confidence interval approach was used to test non-inferiority (43). Non-inferiority was claimed if the lower limit of the one-sided 95% confidence for the difference between test and control toothpaste was less than ≤ 20%. The exact one-sided and two-sided 95% confidence intervals were calculated according to the method of Chan and Zhang (44) using SAS 9.4. In addition, a logistic regression analysis was performed with the primary endpoint as a dependent variable and toothpaste, center, gender, and age as independent variables (covariates).
The confidence interval approach was also performed to test non-inferiority of the secondary endpoint's percentage of subjects experiencing no increase in the overall number of caries lesions (as examined using DIAGNOcam) as explorative statistical evaluation. Moreover, a logistic regression analysis was performed for this secondary endpoint as a dependent variable and toothpaste, center, gender, and age as independent variables (covariates).
. /fpubh. . A two-sided independent t-test was applied for the secondary endpoint PCR (=PCR Visit 4-PCR Visit 1) to compare test vs. control toothpaste. In addition, an analysis of covariance (ANCOVA) was performed for the secondary endpoint PCR using toothpaste, center, gender, and age as independent variables (covariates).

. . Ethical committee and study registration
The study for both centers was approved by the ethical commission of the Poznan University of Medical Sciences, Poznan, Poland (No. 691/20; November 04, 2020). The study was registered at ClinicalTrials.gov: NCT04756557.
The primary endpoint was analyzed for the PP population and, in addition, repeated for sensitivity reasons for the ITT population.

. . Patient flowchart and analysis sets
The patient flow chart according to the CONSORT statement is shown in Figure 1.   In the following, the results are reported for the PP population ("analyzed subjects") if not mentioned otherwise.

. . Demographic data
The demographic data of the PP population (n = 171) are summarized in Table 2. . . Primary e cacy parameter-DMFS index . .

. Descriptive statistics and figures
The descriptive statistics of the DMFS scores at visit V1 (baseline), visit V2 (6 months), visit V3 (12 months), and visit V4 (end of study, after 18 months) differentiated by toothpaste is shown in Table 3 for the PP population.
The descriptive statistics (Table 3) indicate that the mean DMFS index differed only slightly between groups (test toothpaste vs. control toothpaste) and changed only slightly in both groups from baseline (V1) to the end of study (V4).
The descriptive statistics of the differences in the DMFS Index (Delta V2 -V1, Delta V3 -V1, and Delta V4 -V1) are shown in Table 4 for the PP population. Figure 2 illustrates the mean values and corresponding 95% confidence intervals (95% CI) of the differences between Delta DMFS V2 -V1, Delta DMFS V3 -V1, and Delta DMFS V4 -V1 differentiated by group (control vs. test toothpaste).
For the test toothpaste, the 95% confidence intervals of the differences include the value 0. This indicates that in this group .
/fpubh. .   the DMFS index did not change (statistically significant) during the application period compared with baseline. In contrast, in the group "Control Toothpaste" the 95% confidence interval for the difference DMFS V4 -V1 did not include the value 0 (DMFS V4 -V1 > 0). This indicates a (statistically significant) increase in DMFS in this group.
In addition, Figure 3 shows the percentage of subjects with an increase in DMFS index at visits V2, V3, and V4 compared with the baseline (V1) differentiated by group (control vs. test toothpaste).

. . . Confirmative statistical test (primary analysis) and additional statistical tests
The primary efficacy endpoint was the percentage of subjects showing no increase in DMFS Index ( DMFS = DMFS Visit4 -DMFS Visit1 ≤ 0) during the application period of 18 months (Visit 4 -Visit 1) (Table 5). For comparison of the efficacy of the test toothpaste and control toothpaste, the difference in the percentages of subjects with DMFS ≤ 0 was calculated: DMFS% CT = DMFS% C -DMFS% T With DMFS% C = percentage of subjects in the control group with DMFS ≤ 0 DMFS% T = percentage of subjects in the test group with DMFS ≤ 0.
The upper limit of the one-sided 95% confidence interval for the difference in percentage of subjects without an increase of DMFS ( DMFS% CT = −1.93%) was 6.84% and clearly below the non-inferiority margin of ≤ 20% in the PP analysis. Thus, the test toothpaste (hydroxyapatite) can be considered non-inferior to the control toothpaste (1,450 ppm fluoride). This was also true for the upper limit of the two-sided 95% confidence interval (8.24%). In the ITT population, the upper limit of the onesided 95% confidence interval for the difference in percentage of subjects without an increase of DMFS ( DMFS% CT = −2.00%) was 5.68% and also clearly below the non-inferiority margin of ≤ 20%. This was also true for the upper limit of the twosided 95% confidence interval (7.27%), i.e., the "sensitivity analysis" is in accordance with the PP analysis. Superiority cannot be assumed as the confidence intervals include zero. The two-sided 95% confidence intervals for DMFS% CT in the PP population  (−12.2%, 8.24%) and ITT population (−11.4%, 7.27%) lay within the range of (−20%, +20%). Assuming a margin of equivalence for DMFS% CT of (−20% to +20%), this indicates equivalence of the test and control toothpaste.
In addition, a logistic regression analysis was performed with the primary endpoint (increase in DMFS vs. no increase in DMFS) as dependent variable and toothpaste, center, sex, and age as independent variables (covariates).
The results for the PP population confirmed that the "risk" of an increase in DMFS did not differ significantly between control and test toothpastes.
Furthermore, the risk of an increase in DMFS did not differ significantly between centers, was not dependent on age, but was significantly higher in men compared to women. The corresponding odds ratio revealed that the "risk" of an increase in DMFS was 4.7-fold higher in men: An increase in DMFS was observed in 13 of 57 (22.8%) men and only in 7 of 114 (6.1%) women of the PP population.

. . . NCL-Descriptive statistics and figures
The overall number of caries lesions was assessed by DIAGNOcam (Table 6).
Apparently, the percentage of surfaces classified as "score 2" (shadow visible in dentin) was very low at baseline (V1) in the group "control toothpaste" (0.9%) and in the group "test toothpaste" (0.6%) and did not change significantly during the application period in either groups at V4 (0.9%, 0.5%, respectively), whereas the percentage of surfaces classified as "score 1" (shadow visible in enamel) increased slightly in the "control toothpaste" (5.9% vs. 6.5%) and "test toothpaste" (6.4% vs. 6.9%) groups. Please note that NCL was not determined at V2 and V3. Figure 4 indicates that the overall percentage of caries lesions NCL% differed only slightly between groups at V1 and V4 and increased only slightly between visit V1 (baseline) and V4 (after 18 months of applications of the toothpastes).

. . . NCL-explorative statistical tests
The secondary efficacy endpoint for NCL is the proportion of subjects showing no increase in NCL% ( NCL = NCL% Visit4 -NCL% Visit1 ≤ 0) during the application period of 18 months (Visit 4 -Visit 1) ( Table 7). For comparison of efficacy of test toothpaste and the control paste, the difference of the percentages of subjects with NCL ≤ 0 was calculated: NCL% CT = NCL% C -NCL% T With NCL% C = percentage of subjects in the control group with NCL ≤ 0 NCL% T = percentage of subjects in the test group with NCL ≤ 0.
The upper limit of the one-sided 95% confidence interval for the difference in proportion of subjects without an increase of NCL ( NCL% CT = −3.24%) of 9.27% was clearly below the noninferiority margin of 20% in the PP analysis. This is also true for the upper limit of the two-sided 95% confidence interval (11.60%). This indicates also for the secondary endpoint NCL% that the test toothpaste is non-inferior to the control toothpaste.
In addition, a logistic regression analysis was performed for the secondary endpoint NCL (increase in NCL vs. no increase in NCL) as dependent variable and toothpaste, center, sex, and .
/fpubh. .  age as independent variables (covariates). The results for the PP population confirmed that the "risk" of increase in NCL did not differ significantly between control and test toothpaste. Furthermore, the risk did not differ significantly between centers and was not dependent on age and sex.

. . . PCR-descriptive statistics and figures
The descriptive statistics (Tables 8-10) and Figure 5 indicate that the PCR scores differed only slightly between groups (test toothpaste vs. control toothpaste) and decreased slightly in both groups from baseline (V1) to the end of study (V4).
In addition, descriptive statistics for percentage changes in PCR from baseline to V2, V3, and V4 differentiated by group (control vs. test toothpaste) were calculated. Due to skewed distributions and extreme values, respectively, mean and standard deviation may be misleading, median, 25 and 75% percentiles should be interpreted instead. The medians indicate that percentage changes in PCR scores differed only slightly between groups (test toothpaste vs. control toothpaste), but decreased in both groups between V1 and V4, especially between V1 (baseline) and V2 (after 6 months) (Table 10).

. . . PCR-explorative statistical tests
A two-sided Mann-Whitney-Wilcoxon test was applied (instead of t-test) for the secondary endpoint PCR to analyze the difference between test vs. control toothpaste, because Shapiro-Wilk test rejected normality (p = 0.003). Mann-Whitney test confirmed that PCR (V4-V1) did not differ significantly .

FIGURE
Boxplots of NCL% at V and V di erentiated by group (control vs. test toothpaste).   a does not sum up to 171 because in single subjects (n = 2 / n = 3) PCR was 0 at V1 and > 0 at V2, V3, or V4, respectively. If PCR was 0 at V1 and 0 at V2, V3, or V4, respectively, the corresponding PCR% was set to 0.

. . Safety
All randomized subjects (n = 194), who received the control or test toothpaste at baseline (visit V1), were included in the Safety Analysis Set. No serious and no severe adverse events were reported. There were no statistically significant differences in causality and outcome of adverse events between the two toothpastes groups (p > 0.05, two-sided Fisher's exact test). In . /fpubh. . almost all AEs, the causality of AEs with the application of the toothpaste was rated as "unlikely" (control toothpaste: 99.0%, test toothpaste: 99.4%).

. Discussion
In the present clinical trial, the effectiveness of hydroxyapatite toothpaste in preventing caries development was compared with that of fluoride (1,450 ppm) toothpaste, to determine the non-inferiority of the hydroxyapatite toothpaste to the fluoride toothpaste. Analysis of both the primary and the secondary outcome measures demonstrated non-inferiority of the two products to each other. This result is in agreement with two previously published clinical trials (6, 7); thus, this is the third clinical trial showing that fluoride-free hydroxyapatite toothpastes are non-inferior to fluoride toothpastes in caries prevention. Although not statistically significant in all three clinical trials, there was a tendency that the hydroxyapatite toothpastes were slightly more efficient than the corresponding fluoride control toothpastes.
The present study used a toothpaste with sodium fluoride (1,450 ppm fluoride) as control toothpaste. Sodium fluoride is one of the most frequently used fluoride compounds in fluoridated toothpastes worldwide (25). Walsh et al. (30) published a comprehensive systematic review on the use of fluoride toothpastes for cavity prevention (30). Even though most studies have been conducted and published more than 30 years ago (30,45), for a comparability of those studies with recently published data, the methodology should at least be comparable. The authors defined, apart from fluoride toothpaste or concentration, the following inclusion criteria: The studies had to be randomized clinical trials with a study duration of at least 1 year. For caries detection, DMFS/T (Decayed Missing Filled Surfaces/Teeth) was used. These inclusion criteria are in line with the present study, in which the study duration was 18 months, and primary outcome was measured using DMFS, supplemented with a laser nearinfrared transillumination by means of DIAGNOcam. Although a transillumination is considered a qualitative method of caries diagnostics, in the present study a classification was applied as published before (42), which enabled the quantitative assessment of caries lesions. The sample size of the present study was also comparable with those from the studies included in the review of Walsh et al. (30). Cardoso et al. (46) which is the most recent study included in the Walsh et al. (30) review, included approximately 50 subjects per group (46). The same sample size per group (n = 50) was included in Rao et al. (47). In general, sample size calculations were seldom provided by the other papers. Apart from this, no study included in Walsh et al. (30) compared a placebo toothpaste or any other fluoride toothpaste with a 1,450-1,500 ppm fluoride toothpaste in adults (30). Three studies compared toothpastes with 1,000 ppm fluoride to placebo toothpastes (48)(49)(50). All of those studies showed a significant caries prevention after 12 months of fluoride toothpaste use.
Approximately almost 90% of the subjects remained caries-free (no increase in DMFS) in the present study, representing the cariespreventing effect of both toothpastes within the study duration. This is higher than in data from Rao et al. (47): 31% in the placebo group, 53% in the fluoride group (1,190 ppm fluoride), and 72% in the casein phosphopeptide group remained caries-free after 2 years of using their respective toothpaste (47). However, this study was on children, which might be one reason for higher caries levels compared with the present study. For this reason, a direct comparison is not possible, e.g., due to different populations and time points of DMFS examination.
The primary parameter in the present study was the DMFS index. This index was established in clinical studies that investigated the caries-preventing effect of toothpastes (30). Based on the statistical analysis, a remineralizing effect was not excluded in the present study (DMFS Visit4 ≤ DMFS Visit1 ). According to the per-protocol analysis, no increase in DMFS index was observed in 89.29% of subjects of the hydroxyapatite group and 87.36% of subjects of the fluoride group. According to the intention-to-treat analysis, no increase in DMFS index was observed in 90.43% of the subjects of the hydroxyapatite group and 88.42% of the subjects of the fluoride group (Table 5). Assuming a margin of equivalence for the primary endpoint of (−20 to +20%), the results would indicate not only non-inferiority but equivalence of hydroxyapatite toothpaste and fluoride toothpaste.
When comparing the mean DMFS increase on the tooth surface level during the 18-month study duration with the previously published studies (48)(49)(50) (Table 11).
In addition to DMFS, the overall number of caries lesions was determined with DIAGNOcam. DIAGNOcam is a modern device for laser diode near-infrared light transillumination of dental tissues. The appliance combines fiber optic transillumination with near-infrared (NIR) light and a digital camera to non-ionizing imaging of dental tissues. DIAGNOcam enables the visualization of caries lesions on occlusal and approximal surfaces and measures their severity in real time. It may be used without any limitation to monitor the caries process. It is a non-invasive and non-destructive tool (51). This novel infrared digital imaging transillumination technology shows a good accuracy and even a higher sensitivity and accuracy compared with bitewing radiographs in both in vitro (52-54) and in vivo studies (55,56). The trend showing that the hydroxyapatite toothpaste is non-inferior to the fluoride toothpaste observed in the DMFS indices was clearly confirmed with DIAGNOcam (Tables 6, 7). According to the per-protocol analysis, no increase in percentage of caries lesions (DIAGNOcam) was observed in 60.71% of the subjects of the hydroxyapatite group and 57.47% of the subjects of the fluoride group (Table 7). It is pertinent to mention that in both study groups no subject was withdrawn for developing new caries.
The difference in DMFS (visual) and NCL based on DIAGNOcam (near-infrared light transillumination) can be explained by the sensitivity of this method. While decayed surfaces can only be detected on the outer area of the tooth visually by an experienced clinician, transillumination methods can also detect lesions underneath the surface (57,58). A notable trend was also observed in the PCR index of the hydroxyapatite group: Unlike the fluoride group, the PCR scores in the hydroxyapatite group showed a tendency to a lowering of scores during the course of the study ( Figure 5). This is in line with previously published in situ studies on hydroxyapatite-based mouthwashes that showed that hydroxyapatite particles reduce the initial bacterial colonization on enamel surfaces (12, 13).
A strength of the present study is that the same toothpaste composition was used for the hydroxyapatite toothpaste and the fluoride toothpaste except that they differed in their main active ingredients (i.e., hydroxyapatite vs. fluoride). All other ingredients remained the same. While the non-inferiority of hydroxyapatite to fluoride toothpastes was shown in caries risk groups [children (6) and orthodontic patients (7)], this is the first study of this nature conducted in adults. Most studies on the caries-preventing effect of fluoride toothpastes were limited to children and adolescents (30). Additionally, the toothpastes were analyzed over 18 months in the present study which is a longer period than in previous anticaries studies on hydroxyapatite toothpastes (6,7).
Secondly, as we tested the influence of both toothpastes on caries progression, the patients were not monitored according to dietary habits, and they kept their sugar intake under real-life conditions. There were no specific inclusion or exclusion criteria concerning the diet. However, the subjects had comparable diet behaviors due to living in urban areas in both study centers (59, 60). Recently published meta-analysis on sugar intake (including all mono-and disaccharides) in Eastern European countries showed high consumption of sweets, confectioneries, and drinks with high added sugar among the Polish adult population (59, 60). In addition, the amount of carbohydrates taken into the body through oral cavity has also deteriorated during the pandemic Covid-19 quarantine (61).
A limitation of this study is that the study population was relatively homogeneous, consisting of healthy, predominantly young adults without significant oral health problems. However, previously published RCTs on hydroxyapatite toothpastes were performed in caries risk groups, i.e., orthodontic patients (7) and children (6). A further limitation of the present study is that it was performed in a clinical setting with dental visits every 6 months; however, this is not the case for the whole population. Furthermore, no placebo toothpaste was used, but this is because it would be unethical to expose subjects to a risk of developing caries using a toothpaste without an anticaries agent. Additionally, based on evidence-based medicine, a placebo is only ethically reasonable, when there is no other known active treatment for the investigated disease. Another important note is that no bitewing radiographs were used to monitor caries progression. As studies have shown a better accuracy and sensitivity of DIAGNOcam compared to bitewing radiographs (57), there is no need to include diagnostic measurements based on X-rays and expose the subjects to a possible risk of radiation. In addition, it is recommended that adults designated as low caries risk (as in the present study) should have bitewing radiographs made at approximately 24-month intervals (62, 63), and the duration of the present study was 18 months.
There are several reasons why hydroxyapatite toothpastes represent an alternative to fluoride toothpastes. Hydroxyapatite is a safe active ingredient (3). Fluoride, on the other hand, can cause dental fluorosis documented also in areas of low fluoride content in the drinking water (64). It is important to note that both children and adults may ingest excessive amounts of fluoride. Children often use unappropriated toothpaste amount than recommended (65). Moreover, a recently published study indicates the negative impact of fluoride on thyroid functions in pregnant women and neuronal development in unborn (66). Therefore, reducing the overall fluoride intake is beneficial. Although those studies were often performed with populations having either water fluoridation or high fluoride concentrations in the groundwater, a risk of the chronic fluoride exposure through fluoride from fluoridated toothpastes cannot be fully excluded (i.e., there are no studies showing the safety of fluoride toothpastes in this field).
Hydroxyapatite has been shown to be not only efficient in caries prevention, but also in other fields of oral prevention, such as reduction of dentin hypersensitivity (21, 22) and improvement of periodontal health (20), biofilm control (12, 13), and whitening (24). In contrast to other active ingredients, hydroxyapatite is safe if accidentally swallowed and does not interfere with the oral microbiome and does not stain the tooth surface.
In summary, the three clinical caries studies [Paszynska et al. (6), Schlagenhauf et al. (7), and the present study] as well as a number of in situ (8, 12, 13, 67) studies and in vitro studies (9,17,29,39) in the field of remineralization and biofilm control, have demonstrated that hydroxyapatite serves as an efficient and safe alternative to fluoride in anticaries toothpastes for all age groups.

. Conclusion
The results of this long-term double-blinded, randomized clinical trial in adults clearly show the non-inferiority of the fluoride-free hydroxyapatite toothpaste to the toothpaste with 1,450 .
/fpubh. . ppm fluoride with regard to the primary endpoint DMFS index. According to the per-protocol analysis, no increase in DMFS index was observed in 89.3% of subjects of the hydroxyapatite group and 87.4% of subjects of the fluoride group. In conclusion, hydroxyapatite was proven to be a safe and efficient anticaries agent in oral care.

Data availability statement
The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author/s.

Ethics statement
The studies involving human participants were reviewed and approved by the Poznan University of Medical Sciences, Poznan, Poland (No. 691/20; November 4, 2020). The patients/participants provided their written informed consent to participate in this study.