Design and ethical consideration
Ethical approval for this randomized, controlled, parallel-groups field trial was granted by the Research Ethics Committee of the Faculty of Dentistry, Alexandria University (#0274-07/2021). This trial was registered in the clinicaltrials.gov registry (#NCT05642494) on 8/12/2022. The research procedures, risks and benefits were explained to the parents, and they were asked to sign an informed consent form. The study was conducted in accordance with the Helsinki declaration [21] and reported following the CONSORT guidelines [22]. Participants were referred for treatment for teeth showing ECC progress after the intervention.
Participants, settings, and location
Children were invited to join the study if they were healthy, ≤ 4 years of age, with at least one active carious lesion on a primary tooth, with ICDAS codes 3 and higher as reported by Duangthip et al [12]. Teeth with spontaneous or elicited pain due to caries, teeth showing any sign of pulpal infection, swelling and/or abscess were excluded. Also, children were excluded if they had allergies to silver or any material used in the study. All eligible teeth per child were included. The study was conducted in a rural area in Alexandria, Egypt. Children were recruited from 4 nurseries in 3 villages.
Interventions
The children received one of two interventions with an allocation ratio = 1:1. The control group received 38% SDF solution (Advantage Arrest, Elevate Oral Care, FL, USA). Petroleum jelly was applied on the lips and perioral skin as a protective barrier to avoid staining. Gross debris was removed from the carious cavities to allow better SDF contact with denatured dentin. No attempt was made to remove carious tissues before SDF application. The areas to be treated were isolated with cotton rolls and dried with dry cotton pellets. One drop of SDF was carefully applied with a microbrush. Carious lesions were painted for 10 seconds, and the excess was removed using cotton pellets. The solution was left in contact with the tooth surface for one minute before the child was allowed to close their mouth [23]. The application was done at baseline and after 6 months[15].
The test group received the same 38% SDF solution as the control group and 5% NaF varnish (ALPHA-PRO ® WHITE VARNISH, USA). After the application of SDF using the method described in the control group, it was left to dry for 1 minute.
Partial isolation of the whole oral cavity was done using cotton rolls. 0.5 ml of 5% NaF varnish was painted over the carious lesion for one minute and on all remaining teeth in the oral cavity for 2–4 minutes [24]. SDF + NaF were applied at baseline and after 6 months. The 5% NaF varnish alone was re-applied after 3 months from baseline [15]. After the application of the fluoride agents in both groups, the child was instructed not to drink or eat for at least 30 min.
Outcome assessment
Primary outcome
The primary outcome was the arrest status of carious lesions after six months. Three examiners (EBA, MQ, and RY) were trained on assessing caries using the ICDAS criteria [20] and lesion activity using the ICDAS- lesion activity assessment (LAA) criteria [25]. The training was based on the online resources for ICDAS method of caries examination [26]. This was followed by assessing caries in 10 children for training and consensus building. The inter and intra examiner agreement levels of the three examiners were then checked by examining 15 other children not included in the study with a repeated examination after a week. The Kappa statistic for the inter and intra examiner agreement ranged from 0.89 to 0.96 which denotes excellent agreement [27].
The oral examinations were conducted in the nurseries under natural daylight where children were seated in front of a window. No magnification or radiographs were used, and potential lesions were cleaned and dried by a piece of cotton before examination. Assessing lesion activity was done by visual inspection and tactile detection using a 0.5mm ball-ended Community Periodontal Index (CPI) periodontal probe and a disposable dental mirror. Lesions were assessed at baseline and after 6 months. At baseline, lesions with ICDAS code 3 were recorded as active if the surface was whitish/ yellowish opaque with loss of luster; or felt rough when the tip of the probe was moved. All lesions with ICDAS code 4 were recorded as active. After 6 months, lesions coded as ICDAS 3 or 4 were classified as arrested if there was no progress to cavitation and the lesions did not become ICDAS 5 or 6 [12].
Lesions that were cavitated at baseline with ICDAS codes 5 or 6 were recorded as active if softness was detected upon gentle probing [28]. After 6 months, if the wall or floor of the lesion was soft and easily penetrated by the probe using light force, it was diagnosed as active. A lesion where all surfaces were hard and smooth was diagnosed as arrested [12, 28].
Secondary outcomes
After the application of the fluoride agents, parents were asked if they were satisfied with the appearance of their children’s teeth and the responses were recorded as satisfied or unsatisfied [29]. A week after applying the interventions, parents were asked about adverse effects of SDF including tooth or gum pain, gum swelling, gum bleaching, and systemic toxicity such as nausea, vomiting, or generalized discomfort [30]. Staining of the lesions was recorded as present after applying the interventions if any black stain appeared where the agents were applied or absent if no change was observed in lesion color [30].
Confounders
In addition to assessing the ICDAS code and lesion activity, the clinical examination assessed oral hygiene using the plaque index (PI) of Silness and Loe on 6 index teeth (52, 55, 64, 72, 75, and 84) [31] and caries experience by recording the number of decayed, missing and filled surfaces (dmfs) using the World Health Organization (WHO) criteria [32].
The Arabic version of the WHO questionnaire for oral health assessment of children [33] was used to collect information by interviewing the mothers. The questionnaire assessed the child’s demographic characteristics including age in months, sex, and mother’s education (No formal schooling, Primary school completed, Middle school completed, High school, College/university completed). The questionnaire also assessed oral health behaviors such as toothbrushing (Never, Several times a month (2–3 times), Once a week, Several times a week (2–6 times), Once a day, 2 or more times a day), dental visits during the previous year (Once, Twice, Three times, Four times, More than four times, no visit during the past 12 months, never received dental care/visited a dentist, I don’t know/don’t remember) and sugar consumption (at least once daily versus less frequent consumption) of eight types of sugary products: fruit, biscuits and cakes, carbonated beverages, jam and honey, sugar-added chewing gums, candies, sugar-sweetened milk and sugar-sweetened hot drinks.
Sample size determination
Sample size was estimated assuming 5% alpha error and 80% study power. Caries arrest rate after 6 months of a single application of 38% SDF was reported to be 43.9% [23]. After two applications of 5% NaF varnish, caries arrest rate after 12 months was calculated in a previous study to be 20.9% [28]. Assuming an additive effect of 2 applications of 5% NaF and 38% SDF, the caries arrest rate at 6 months in the SDF + NaF group was calculated to be 64.8%. The required number of children per group was 88 as calculated by G*power 3.0.10 [34]. Expecting a drop-out of 20%, the number of children to be recruited per group was estimated to be 106 ≈ 110. Total sample size = number of groups X number per group = 2 X 110 = 220 children.
Randomization, allocation, and blinding
The children were equally allocated to the two groups via block randomization, using a computer-generated random sequence in blocks of 4 [35] .The allocation sequence was kept in opaque sealed envelopes by a trial independent individual (NN). The child was examined by an investigator (EBA) to confirm eligibility and conduct the clinical examination. The trial independent person (NN) opened the envelope and informed EBA of the allocated group to apply the intervention. Blinding the operator was not possible as one group received one agent and the other group received two agents. However, the children/ families, the outcome assessors and the biostatistician were blinded to the intervention type.
Statistical Analysis
Data were analyzed using IBM SPSS Statistics for Windows, Version 23.0. Armonk, NY: IBM Corp and significance was set at p < 0.05. Mother’s education was categorized into “high school and more” if she completed high school or college/university education and “less than high school” including all other categories. Daily tooth brushing was recoded into “Yes” if the child brushed once or more per day while all other categories were recoded into “No”. Dental visits last year were recoded into “Yes” if the child visited the dentist once or more during the last year and “No” if the child made no visit during the past 12 months or ever or if the parent did not know/ remember. A sugar consumption score was created by adding the points of each of the eight sugary products consumed at least once daily. The score ranged from 0 to 8 with higher scores indicating greater daily sugar consumption [36]. The plaque index score was calculated by averaging the scores of index teeth and dmfs was calculated as the sum of surfaces that were decayed, missing due to caries, and filled. Data were assessed for normality. The Mann–Whitney U test was used to compare non-normally distributed variables and t test was used for normally distributed variables. Categorical variables were compared using X2 test. Analysis was conducted at lesion level using the intention to treat approach. Lesions in children lost to follow up and surfaces that were restored or extracted by 6 months were recorded as active.
The effect of the interventions on caries arrest (yes/ no) was assessed in bivariate analysis followed by multi-level binary logistic multivariable regression analysis to control for confounders. Multi-level analysis accommodates the clustering of teeth within children since all eligible teeth per child were included. The explanatory variable (intervention) and confounders were introduced as fixed effect variables at two levels. Level 1 factors included lesion-related variables: type of intervention and ICDAS code at baseline. Level 2 factors were child related confounders (defined above). The dmfs score was not included due to collinearity with ICDAS codes. Children were included as random effect variables. The interaction between the intervention and baseline lesion severity was assessed by calculating the p value for interaction. This was followed by splitting the sample into a subgroup of moderate lesions (ICDAS codes 3 or 4 at baseline) and subgroup of advanced lesions (ICDAS codes 5 or 6). A multilevel, binary logistic, multivariable model was developed for each subgroup. Adjusted odds ratios (AORs), p value and 95% confidence intervals (CIs) overall and per subgroup were calculated.