ABSTRACT
Introduction
Children with more caries in the primary dentition have been observed to have greater caries risk in the permanent dentition. In a previous trial, we determined lesion arrest in primary teeth, where 38% SDF gel and 2.5% NaF varnish were applied sequentially. This follow‐up study determined the prevalence of children with any caries in untreated first permanent molars.
Materials and Methods
The design was a retrospective cohort. Participants were drawn from five Centros Educativos Iniciales, Puno, Peru. Examinations were conducted 19 months after the second treatment. Any caries prevalence (D2–4MFS > 0) and D2–4MFS were compared between groups.
Results
The exposed group was 219 children in the initial trial; the unexposed group was in the same classes but did not participate. 181/237 (76.4%) children treated at baseline were followed up; 104 unexposed children were examined. Prevalence of any caries in first permanent molars was lower in exposed children (4.4%, 95% CI 1.4, 10) compared to unexposed (66.6%, 95% CI 53.6, 77.2; p < 0.001). D2–4MFS in first permanent molars was lower in the exposed (median [IQR] = 0 [0, 0]) vs. unexposed (median [IQR] = 1 [0, 3]; p < 0.001).
Discussion
Children who received two sequential applications of SDF gel/NaF varnish at 3–4y were observed to have fewer caries in first permanent molars at 6–7 y than unexposed children.
Keywords: children, dental caries, dental caries arrest, silver diamine fluoride, sodium fluoride varnish
Why This Paper Is Important to Pediatric Dentists
This paper is important to pediatric dentists because combined sequential treatment with 38% SDF gel and 2.5% NaF varnish is highly efficacious in the primary dentition.
Within the limits of an observational design, the combined sequential treatment with 38% SDF gel and 2.5% NaF varnish appears to have provided continued benefit in the untreated permanent dentition.
1. Introduction
Severe early childhood caries can be difficult to manage in young pre‐cooperative children, especially when there are many caries lesions to be treated. 38% silver diamine fluoride (SDF) has been demonstrated to arrest caries lesions, but multiple applications are more effective than a single treatment [1]. Most studies have repeated treatment at about 6‐months. Similarly, SDF has been shown to be effective in preventing caries in permanent teeth [2, 3, 4, 5].
An observational study suggested that caries experience in primary dentition, particularly decay in the primary second molars, predicts greater risk for caries in the permanent dentition [6]. Similarly, previous investigations have suggested that treating active caries lesions in the primary dentition results in protection of intact surfaces at risk [3], at least in the primary dentition, perhaps because of the so‐called “zombie effect” where dead bacteria serve as a reservoir for silver ions [7]. One observational study, using insurance claims, found only a weak connection between treatment of the primary dentition and caries risk in the permanent dentition [8].
Although quite safe, the liquid form of SDF is difficult to control in young children, increasing the possibility of inadvertent exposure of the soft tissues. Moreover, SDF alone is bitter, and children with many teeth to be treated may respond negatively to the taste. A gel version of 38% SDF became available in the United States in early 2023. The gel is applied similarly to traditional SDF liquid and can be covered with sodium fluoride varnish (NaF varnish) to mask the taste of the gel alone and enhance cooperation. In contact with the surface of active lesions, the gel behaves similarly to the traditional liquid [9].
The objective of this study was to determine the prevalence of dental caries and the number of carious surfaces in first permanent molars at 19 months of follow‐up without further treatment. We hypothesized that the large number of treated teeth that remain in the mouth serves as a reservoir of continuing anti‐caries activity.
2. Materials and Methods
2.1. Study Design
The design was a retrospective cohort study. We followed the STOBE Checklist for Cohort Studies in reporting the findings.
2.2. Study Setting and Participants
The study was conducted at five preschools (Centros Educativos Iniciales), Puno, Peru. Exposed were 219 healthy children in the earlier trial who were 3–4 years old when initially treated; unexposed were 104 children in the same classes who did not participate, identified from study records and the schools' database. In the previous trial, we determined lesion arrest in primary teeth, where 38% SDF gel (Advantage Arrest Silver Diamine Fluoride 38% GEL, Elevate Oral Care, West Palm Beach, FL) and 2.5% NaF varnish (FluoriMax Varnish, Elevate Oral Care, West Palm Beach, FL) were applied sequentially. The exposed children in that trial received their second SDF/FV treatment in December 2022, with outcomes data on caries lesion arrest rate in primary teeth in July 2023. For the current study, follow‐up examinations were conducted from June 17, 2024 through July 20, 2024. The follow‐up times since the initial treatment and since the 5 month second treatment are given in Table 1. Parents consented to the examinations in Spanish. The study population was urban Hispanic American/South American Native. Peru is an upper‐middle‐income country undergoing a nutrition transition towards a Westernized dietary pattern.
TABLE 1.
Characteristics of the children in the Peru cohort study.
| Characteristic | Overall N = 285 | Unexposed N = 104 | Exposed treated N = 181 | p‐value |
|---|---|---|---|---|
| Age, n (%) | 0.554 a | |||
| 5 | 58 (20.4) | 18 (17.3) | 40 (22.1) | |
| 6 | 134 (47.0) | 49 (47.1) | 85 (47.0) | |
| 7 | 93 (32.6) | 37 (35.6) | 56 (30.9) | |
| Median (IQR c ) | 6.0 (6.0, 7.0) | 6.0 (6.0, 7.0) | 6.0 (6.0, 7.0) | 0.293 c |
| Gender, n (%) | 0.098 a | |||
| Female | 139 (48.8) | 44 (42.3) | 95 (52.5) | |
| Male | 146 (51.2) | 60 (57.7) | 86 (47.5) | |
| Permanent first molars, n (%) | 0.626 a | |||
| 0 | 104 (36.5) | 36 (34.6) | 68 (37.6) | |
| 1 | 33 (11.6) | 11 (10.6) | 22 (12.2) | |
| 2 | 32 (11.2) | 11 (10.6) | 21 (11.6) | |
| 3 | 19 (6.7) | 5 (4.8) | 14 (7.7) | |
| 4 | 97 (34.0) | 41 (39.4) | 56 (30.9) | |
| Median (IQR c ) | 2.0 (0.0, 4.0) | 2.0 (0.0, 4.0) | 2.0 (0.0, 4.0) | 0.309 b |
| No. of permanent teeth | ||||
| Median (IQR c ) | 3.0 (0.0, 6.0) | 3.0 (0.0, 7.0) | 3.0 (0.0, 6.0) | 0.345 b |
p‐value, Chi‐square test.
p‐value, Wilcoxon rank sum test.
IQR, 25th and 75th percentiles.
The Universidad Nacionaldel Altiplano‐Puno Institutional Review Board approved this study (051‐CIEI‐UNAPuno). The local study name was Kiru Poderoso, a Quechua language expression that means “mighty teeth.” The original clinical trials registration number is NCT05395065.
2.3. Examinations
As in the initial trial [1], dental caries lesions were assessed using the modified World Health Organization (WHO) criteria [10] for permanent teeth and the National Institute of Dental and Craniofacial Research (NIDCR) EC4 guidelines for primary teeth [11]. As described in the previously published trial report, the original WHO criteria [10] were based on classifying each tooth as decayed, missing, or filled and were modified in 3 ways: (a) surface‐level assessments made (rather than tooth level assessments); (b) pre‐cavitated and cavitated lesions confined to enamel were identified (rather than classifying these lesions as sound); and (c) cavitated lesions that extended directly into the pulp were identified (rather than classifying these lesions as decayed into dentin). Decayed, missing, and filled teeth and surfaces were recorded for permanent (DMFT/DMFS) and primary teeth (dmft/dmfs) using the following codes for decayed and filled surfaces: sound (o); uncavitated enamel lesion (d1/D1); cavitation confined to enamel (d2/D2); cavitation into dentin, lesion with no pulp or pulp chamber exposure visible (d3/D3); cavitation into pulp (d4/D4); sealed (s), or filled (f). Teeth that could not be scored at the surface level were scored at the tooth level as unerupted (u), missing due to caries (m), missing due to trauma (t), missing due to exfoliation (x), missing due to other or unknown reasons (o), or crowned (c). Teeth were dry brushed before the examination. Teeth were dried with gauze, and the exams were performed using a mouth mirror with artificial light. No radiographs were used.
2.4. Examiner Training and Calibration
The children were examined clinically by trained and calibrated examiners. Classroom training was followed by clinical training, then calibration with the two local examiners doing exams along with the gold standard examiner (MR). Two examiners (TP and CV) and the gold standard examiner (MR) examined 8 children. The kappa statistic was used to summarize the reliability for the prevalence of any decayed, missing due to caries, or filled surface for first permanent molars and permanent teeth, and the intraclass correlation coefficient was used to summarize the reliability for the number of decayed, missing due to caries, or filled surfaces for first permanent molars, permanent teeth, and primary teeth. The kappa statistic ranged from 0.96 to 0.99, and the intraclass correlation (ICC) was 0.93 to 0.99, indicating excellent reliability. Repeated examinations were not performed to assess intra‐examiner reliability to minimize participant burden.
2.5. Bias
It is not possible to blind examiners to the presence of teeth previously treated with 38% SDF, potentially introducing examiner bias, although some untreated children may also have darkened cavities. We attempted to overcome this limitation with rigorous training and calibration.
2.6. Statistical Methods
Based on an a priori calculation, expected sample sizes of 200 exposed (treated) children (70% of children treated at baseline in our prior study) and 68 unexposed children, the power was 80% to demonstrate a reduction in the exposed of 22% to 45% in the prevalence of any caries in the first permanent molars assuming a prevalence of 40% to 80% in the unexposed based on a two‐sided chi‐square test at a 0.05 significance level. To demonstrate a difference in the number of decayed, missing due to caries, or filled surfaces, the power was 80% for an effect size of 0.4 based on a two‐sided two‐sample t‐test at a 0.05 significance level.
2.7. Data Analysis
Descriptive statistics were computed by group (exposed or unexposed) using the median and interquartile range (25th and 75th percentiles) for quantitative variables and frequency and percent for categorical variables. Chi‐square test, Fisher's exact test, and Wilcoxon rank sum were used to evaluate the comparability of the groups by age, gender, number of first permanent molars, and number of permanent teeth.
Primary outcomes were prevalence of any decayed surfaces (D2, D3, or D4 lesions), missing surfaces due to caries or filled surfaces in first permanent molars (any D2–4MFS > 0) and the number of decayed surfaces, missing surfaces due to caries or filled surfaces in first permanent molars (D2–4MFS). Secondary outcomes were any D2–4MFS > 0 and D2–4MFS in permanent teeth, and any d2–4mfs > 0 and d2–4mfs in primary teeth. The prevalence was compared between groups using Fisher's exact test and exact 95% confidence intervals, and D2–4MFS and d2–4mfs were compared between groups using a Wilcoxon rank sum test. In addition, log‐binomial regression using generalized estimating methods and robust SEs was used to compare the prevalence, adjusting for age, gender and the number of surfaces present or missing due to caries. Log‐linear regression using generalized estimating methods and robust SEs was used to compare the amount of decay adjusting for age, gender and number of surfaces present or missing due caries, where the natural logarithm of the number of surfaces was entered as an offset in the regression model [12, 13]. All analyses were performed with R software, Version 4.4.3 (The R Foundation for Statistical Computing) and a significance level 0.05 was used for statistical significance.
3. Results
Of 237 exposed children treated at baseline, 181 (76.4%) were followed up; 104 unexposed (untreated) children were also examined. The median (range) follow‐up time from initial and second treatments for exposed children was 23.6 (23–24) months and 19 (18–20) months, respectively. The main reasons for loss to follow‐up were that the child moved, the school was on strike, or the child did not appear for the examination (Table 1).
The caries prevalence in the permanent first molars (any D2–4MFS > 0) was lower in exposed children (4.4%, 95% CI 1.4, 10) vs. unexposed (66.6%, 95% CI 53.6, 77.2; p < 0.001) (Table 2). D2–4MFS in first permanent molars was lower in the exposed children (median [IQR] = 0 [0, 0]) vs. unexposed (median [IQR] = 1 [0, 3]; p < 0.001). The amount of decay in primary teeth (d2–4mfs) was significantly lower in the exposed children (median [IQR] = 23 [13, 36]) as compared to unexposed children (median [IQR] = 29 [16, 40]; p = 0.012).
TABLE 2.
Dental caries in permanent first molars, permanent teeth, and primary teeth.
| Outcome | Exposed (Treated) | Unexposed | p‐value b | ||||
|---|---|---|---|---|---|---|---|
| N | % (n) | 95% CI a | N | % (n) | 95% CI a | ||
| Any decay (D2–4MFS > 0 or d2–4mfs > 0) | |||||||
| Permanent first molars | 113 | 4.4 (5) | 1.4, 10.0 | 68 | 66.6 (45) | 53.6, 77.2 | < 0.001 |
| Permanent teeth | 125 | 4.0 (5) | 1.3, 9.1 | 73 | 61.6 (45) | 49.5, 72.8 | < 0.001 |
| Primary teeth | 181 | 100 (181) | 98, 100 | 104 | 100 (104) | 96, 100 | > 0.99 |
| N | Median | IQR c | N | Median | IQR c | p‐value d | |
| Amount of decay (D2–4MFS or d2–4mfs) | |||||||
| Permanent first molars | 113 | 0 | 0, 0 | 68 | 1 | 0, 3 | < 0.001 |
| Permanent teeth | 125 | 0 | 0, 0 | 73 | 1 | 0, 3 | < 0.001 |
| Primary teeth | 181 | 23 | 13, 36 | 104 | 29 | 16, 40 | 0.012 |
95% CI, exact (binomial) 95% confidence interval.
p‐value, Fisher's exact test.
IQR, 25th and 75th percentiles.
p‐value, Wilcoxon rank sum test.
The differences between exposed and unexposed were still highly significant after adjusting for age, gender and number of surfaces present or missing due to caries (Table 3). The prevalence ratio and rate ratios are close to zero for the permanent first molars and permanent teeth because the difference between exposed and unexposed is very large.
TABLE 3.
Log‐binomial regression results for dental caries in first permanent molars, permanent teeth, and primary teeth at follow‐up.
| Outcome | Model 1 a | Model 2 b | ||
|---|---|---|---|---|
| PR (95% CI) d | p‐value | PR (95% CI) d | p‐value | |
| Any decay (DMFs > 0 or dmfs > 0) c | ||||
| Permanent first molars | 0.075 (0.031, 0.18) | < 0.001 | 0.077 (0.032, 0.18) | < 0.001 |
| Permanent teeth | 0.075 (0.030, 0.18) | < 0.001 | 0.077 (0.031, 0.18) | < 0.001 |
| RR (95% CI) e | p‐value | RR (95% CI) e | p‐value | |
| Amount of decay (DFMs or dmfs) | ||||
| Permanent first molars | 0.083 (0.026, 0.26) | < 0.001 | 0.086 (0.02, 0.26) | < 0.001 |
| Permanent teeth | 0.082 (0.026, 0.26) | < 0.001 | 0.084 (0.02, 0.25) | < 0.001 |
| Primary teeth | 0.84 (0.73, 0.96) | 0.010 | 0.86 (0.76, 0.98) | 0.001 |
For any decay, model 1 adjusts for the number of surfaces present for permanent first molar and permanent teeth, and the number of surfaces present or missing due to caries for primary teeth. There were no permanent teeth missing due to caries. For the amount of decay, the natural logarithm of the number of surfaces present or missing due to caries was entered as an offset in the log‐regression model.
Model 2 additionally adjusts for child age and gender (female or male).
All children had decay (dmfs > 0) in their primary teeth, and hence regression results are not reported for any decay (dmfs > 0).
PR (95% CI), prevalence ratio and 95% confidence interval for any decay.
RR (95% CI), rate ratio and 95% confidence interval.
4. Discussion
In this observational study, we determined the prevalence of dental caries and the number of carious surfaces in first permanent molars at 19 months of follow‐up after treatment of extensive carious lesions in the primary teeth, without further treatment. The prevalence of caries in the first permanent molars was more than 15‐fold greater in the unexposed (untreated) group than in the exposed group, and this difference persisted after adjustment for age, gender and number of surfaces present or missing due to caries.
The children in this study had extraordinarily high levels of caries lesions in the primary dentition. Previous work suggested a correlation between caries in the primary dentition and the permanent dentition because of the shared microbial burden [6], while one study suggested a weak link between treatment of the primary dentition and lower risk to the permanent dentition [8]. However, these connections are mediated by behavioral and environmental factors, such as diet and oral hygiene. Nevertheless, this inexpensive approach, if confirmed in studies with more rigorous designs, combined with use of SDF and/or glass ionomer sealants in the permanent teeth, may result in many fewer children with active disease. The public health benefit from the school‐based treatment cannot be understated.
4.1. Limitations
We initially carried out a rigorous clinical trial to ascertain the rate of caries lesion arrest with sequential treatment using an 38% SDF gel and 2.5% NaF varnish. In the trial report previously published [1], we acknowledged that the single‐arm design limits any formal comparison with the gel alone or traditional SDF liquid and that uncontrolled trials can produce greater mean effect estimates than a controlled trial. A factorial design comparing the gel and traditional SDF liquid, with or without NaF varnish, was not possible in this setting because of practical and regulatory factors. Moreover, such a study design would have allowed for a priori testing of the hypothesis in this retrospective observational study. Although the results of this follow‐up cohort study at about 19 months post‐treatment are impressive, the lack of a randomized design limits generalizability. Nonetheless, the study cohorts are quite similar, lending credence to the findings. The study population has caries prevalence higher than most studies in the literature. A randomized controlled trial comparing this SDF gel and the traditional liquid is ongoing elsewhere [14].
Author Contributions
T.C.P.C. conceived the study and served as an examiner. J.L.C. contributed to the design of the study. C.M.V.R. served as an examiner and assembled and managed the data files. M.R. trained the examiners and served as the gold standard examiner. L.L.M. was the study biostatistician. P.M. helped conceive the study, contributed to the design, and wrote the first draft of the paper. All of the authors contributed to the final preparation of the paper and approved of its content.
Funding
This research was supported, in part, by Advantage Silver Dental Arrest LLC and Elevate Oral Care LLC. and an unrestricted donation to the University of Washington. The sponsors provided the products used in this study but had no role in carrying out the research or in the analysis of the data. Colgate‐Palmolive‐Perú donated toothbrushes and toothpaste for the children enrolled in this study. This project was also supported by the University of Washington Institute of Translational Health Sciences; NIH awards to the institute did not provide direct funding for this project.
Disclosure
Institutional Review: This research was approved by the Institutional Review Board of the Universidad Nacional del Altiplano‐Puno (051‐2022‐CIEI‐UNAPuno).
Conflicts of Interest
Dr. Milgrom is a member of Advantage Silver Dental Arrest LLC. The remaining authors declare no conflict of interest.
Acknowledgments
The authors acknowledge the contributions of Paola Balabarca. Maria Mamani and Pedro Ancco, who assisted in carrying out this research.
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.
References
- 1. Padilla Cáceres T. C., Cervantes‐Alagón S., Castillo J. L., et al., “Using Sequential Applications of a Novel Silver Diamine Fluoride Gel and Sodium Fluoride Varnish to Arrest Severe Early Childhood Caries Lesions: A Clinical Trial With Single Group Assignment,” Journal of the American Dental Association 155, no. 6 (2024): 526–535. [DOI] [PubMed] [Google Scholar]
- 2. Alqalaleef S. S., Alnakhli R. A., Ezzat Y., AlQadi H. I., Aljilani A. D., and Natto Z. S., “The Role of Silver Diamine Fluoride as a Dental Caries Preventive and Arresting Agent: A Systematic Review and Meta‐Analysis,” Frontiers in Oral Health 5 (2024): 1492762. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Liu B. Y., Lo E. C., Chu C. H., and Lin H. C., “Randomized Trial on Fluorides and Sealants for Fissure Caries Prevention,” Journal of Dental Research 91, no. 8 (2012): 753–758. [DOI] [PubMed] [Google Scholar]
- 4. Chu C. H., Lo E. C., and Lin H. C., “Effectiveness of Silver Diamine Fluoride and Sodium Fluoride Varnish in Arresting Dentin Caries in Chinese Pre‐School Children,” Journal of Dental Research 81, no. 11 (2002): 767–770. [DOI] [PubMed] [Google Scholar]
- 5. Fung M. H. T., Duangthip D., Wong M. C. M., Lo E. C. M., and Chu C. H., “Randomized Clinical Trial of 12% and 38% Silver Diamine Fluoride Treatment,” Journal of Dental Research 97, no. 2 (2018): 171–178. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Skeie M. S., Raadal M., Strand G. V., and Espelid I., “The Relationship Between Caries in the Primary Dentition at 5 Years of Age and Permanent Dentition at 10 Years of Age—A Longitudinal Study,” International Journal of Paediatric Dentistry 16, no. 3 (2006): 152–160. [DOI] [PubMed] [Google Scholar]
- 7. Wakshlak R. B., Pedahzur R., and Avnir D., “Antibacterial Activity of Silver‐Killed Bacteria: The “Zombies” Effect,” Science Reports 5 (2015): 9555. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Heller K. E., Eklund S. A., Pittman J., and Ismail A. A., “Associations Between Dental Treatment in the Primary and Permanent Dentitions Using Insurance Claims Data,” Pediatric Dentistry 22, no. 6 (2000): 469–474. [PubMed] [Google Scholar]
- 9. Kiesow A., Menzel M., Lippert F., Tanzer J. M., and Milgrom P., “Dentin Tubule Occlusion by a 38% Silver Diamine Fluoride Gel: An Invitro Investigation,” British Dental Journal Open 8, no. 1 (2022): 1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. World Health Organization , ed., Oral Health Surveys: Basic Methods, 5th ed. (World Health Organization, 2013). [Google Scholar]
- 11. Warren J. J., Weber‐Gasparoni K., Tinanoff N., et al., “Examination Criteria and Calibration Procedures for Prevention Trials of the Early Childhood Caries Collaborating Centers,” Journal of Public Health Dentistry 75, no. 4 (2015): 317–326. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Hardin J. W. and Hilbe J. M., Generalized Estimating Equations (Chapman and Hall/CRC, 2003). [Google Scholar]
- 13. Nahhas R. W., Introduction to Regression Methods for Public Health Using R (Chapman and Hall/CRC, 2025). [Google Scholar]
- 14. Chu C. H., Arresting Early Childhood Caries With a Silver Diamine Fluoride Gel , ClinicalTrials.gov ID NCT06241261, Updated 2026‐03‐05, accessed August 9, 2026, https://clinicaltrials.gov/study/NCT06241261.
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.
