The prevalence of dental fluorosis in children living in fluoride endemic locations has been shown to vary substantially across studies conducted with similar rural study settings in Asia. In epidemiological surveys from China, India, and Indonesia; the overall prevalence of dental fluorosis regardless of severity levels were 38.2%, 69.4%, and 96.0%, respectively [27–29]. In Thailand, the overall prevalence ranged from 5% (very mild to severe dental fluorosis) in Panomsarakham District, Chachoengsao Province [12], to 70.9% (Thylstrup & Fejerskov Index, level 1–4) in Chiang Mai Province [15]. The finding of 54.3% prevalence in this study was considered relatively high given the context of Thailand and especially when compared to the 5% prevalence in Chachoengsao Province which was only 150 kilometers east of the current study site. The substantially lower prevalence in Chachoengsao Province could be related to the high proportion of 87.5% of children living in areas with water supplies containing fluoride less than 0.7 ppm. In this study, however, 89.6% of the children had household water sources with fluoride contents of ≥ 7 ppm. This large disparity of the prevalence in these two comparable settings in Thailand implied the crucial effect of endemic fluoride on dental fluorosis occurrence at the population level.
In addition, this plausible effect of natural fluoride in groundwater use for household consumption on the overall and severity-specific prevalence in this study was also comparable with the ones observed in Birigui, SP, Brazil. In that study with a socio-environmental setting and methods of exposure and outcome measurement resembled the current study; the overall prevalence was 58.9% and severity-specific prevalence values were 44.4% for very mild, 11.9% for mild, 2.4% for moderate, and 0.2% for severe dental fluorosis [19]. This evidence further demonstrated the quality of consistency in the effect of endemic fluoride on the prevalence of dental fluorosis which was observed in different groups of children, at different times, and in different places [30, 31].
The biological gradient between fluoride concentrations in groundwater used for household water supply and dental fluorosis occurrence in children was suggested by the unidirectional positive relationship of these attributes in the current study. The finding of 23.3% prevalence with only the very mild dental fluorosis among children with time-averaged fluoride concentrations of < 0.7 ppm (the referent category) was evidence that somewhat reassured the safety of this recommended optimal fluoride level in this setting and the others [12, 19]. When the time-averaged fluoride concentrations increased to the range of 0.7–1.49 ppm (index category 1), the prevalence among children in this group also increased to 37.7%, with the additional higher levels of mild and moderate severity. Although the fluoride concentrations in this range did not surpass the WHO's recommended limit of 1.5 ppm [4], the results of this study were concerning as the prevalence exceeded one-third of the children and 14.7% of the severity was beyond the very mild level. In the extreme group with the fluoride ≥ 1.5 ppm (index category 2), the prevalence further rose to 64.1% or approximately 2.8 times the prevalence of those with fluoride concentrations of < 0.7 ppm. The severity beyond the very mild level also grew to 15.7%. This finding of the biological gradient suggested the rational use of fluoride concentrations in household water sources as an indicator for the possible occurrence of dental fluorosis and related severity.
Multivariable regression models constructed according to the DAGs, which displayed various assumptions regarding the main association of interest between the time-average fluoride concentrations and dental fluorosis occurrence given a different set of socio-behavioral determinants being simultaneously considered in each model, provided insight into the etiologic pattern of dental fluorosis in this fluoride endemic setting (Table 3). When the measure of association between fluoride concentrations and dental fluorosis was compared before and after adjusting for the child’s demographic factors (sex and age) and caregiver’s factors (education and sufficiency of family income), the difference was negligible, indicating that confounding caused by these factors was extremely minimal. Lack of association between child’s sex and dental fluorosis; indicated by the findings of crude prevalence ratio of 1.0 in Table 2 and adjusted prevalence ratios of 1.06 (Model 2) and 1.07 (Model 6) in Table 3), would explain its trifling influence on the main association between fluoride concentrations and dental fluorosis. The insignificant difference in dental fluorosis occurrence between Thai male and female children, aged 8 to 10 years old, and the lack of association between child’s sex and dental fluorosis–indicating by statistically non-significant crude and adjusted odds ratios of 1.2 and 0.9–were consistently observed in a previous study [32]. Regarding the child’s age, several studies conducted in fluoride endemic areas have contrastively demonstrated a positive linear association between age and dental fluorosis prevalence [33]. The association in those studies might be attributable to the more diverse age categories ranging from 3 to 18 years of age among the study participants and the more visible dental fluorosis in children aged over 10 years old compared to the ones below the age of 8 [33]. Although children cared for by the caregivers with less than a secondary school education had a 5.1% higher prevalence of dental fluorosis, the negative association between lower caregiver’s education and higher dental fluorosis was not supported by the findings of crude and adjusted prevalence ratios close to 1 and statistical non-significance. The prior investigation in Thai children living in Bangkok consistently revealed that dental fluorosis prevalence among children having caregivers with education higher than bachelor’s degree was not significantly lower than the prevalence among those having caregivers with lower levels of education (odds ratio, 0.85; 95%CI, 0.54–1.33) [32]. The lack of association between caregiver’s education and child dental fluorosis might be attributable to the fact that knowledge regarding dental fluorosis and its prevention has never been included in general education in Thailand. This suggested the need for a community-based educational effort to improve literacy regarding dental fluorosis prevention among caregivers of children with developing dentition. In this study, insufficient family income was hypothesized to be associated with higher dental fluorosis prevalence due to an assumption that children in poorer families might be more likely to expose to fluoride in groundwater than their counterparts whose families might be able to afford alternative water sources for drinking such as bottled water. The results, however, did not support this hypothesis. This might be because all the 8 children from families with sufficient income had a household water supply with time-averaged fluoride concentrations of ≥ 1.5 ppm. Another study in Bangkok unveiled a significant relationship between the higher class of family income and greater dental fluorosis prevalence (odds ratio, 1.77; 95%CI, 1.10–2.86) [32]. It was suggested that families with greater income might be more capable of purchasing fluoride products and the children in these families would be more likely to have increased exposure to fluoride [32]. Therefore, the effect of family income on dental fluorosis may vary by different contexts of studies.
The association between fluoride concentrations and dental fluorosis was heightened after adjusting for the effect of breastfeeding and children’s oral health behaviors (Table 3). The dental fluorosis prevalence ratio comparing the fluoride concentrations of 0.7–149 ppm (index category 1) to < 0.7 ppm (the referent category) considerably increased from 1.62 (Model 1) to 3.08 (Model 4) and 3.44 (Model 5), indicating confounding magnitudes of -47.4% by breastfeeding and − 52.9% by children’s oral health behaviors. The fluorosis prevalence ratio comparing the concentrations of ≥ 1.5 ppm (index category 2) to the referent category also greatly increased from 2.75 (Model 1) to 5.30 (Model 4) and 6.46 (Model 5), showing confounding magnitudes of -48.1% by breastfeeding and − 57.4% by children’s oral health behaviors. The large confounding magnitudes created by these factors highlighted their influential role in the etiologic mechanism of dental fluorosis in the fluoride-endemic environment, accentuating the need to consider these factors when measuring the exposure-outcome association between natural fluoride in water used for consumption and dental fluorosis in settings similar to the current ones. Regarding the role of breastfeeding, a prolonged period of breastfeeding has been demonstrated to protect against dental fluorosis [35–38]. Breast milk was shown to have only a trace amount of fluoride regardless of the quantity of fluoride consumed by mothers [39]. Breastfeeding during the first two years of life, which aligns with the active period of enamel formation, thus appears to prevent dental fluorosis [37]. In the current study, breastfeeding might reduce the exposure to fluoride through the use of water containing natural fluoride to prepare powdered formula. Concerning the role of children’s oral health behaviors, early toothbrushing before 2 years old was shown to be strongly associated with mild-to-moderate fluorosis [3]. This habit was also common in this study, having 89.8% of children brushing once a day or more. Fluoride exposure could be heightened in this context as this habit coincided with 94.3% of children using fluoride toothpaste and 51.4% of children using more than pea-sized toothpaste amounts. Ingesting toothpaste during the period of developing enamel could lead to dental fluorosis [40] and this would explain the collective influence of children’s oral health behaviors on dental fluorosis occurrence in this setting. Ultimately, the association between the time-averaged fluoride concentrations and dental fluorosis attenuated after taking full control of all covariates (Model 6), ruling out alternative explanations given information of all other covariates for the observed effect of higher classes of fluoride concentrations (index category 1 and 2) on increased dental fluorosis prevalence beyond the ones of the referent category. By comparing the prevalence ratios for both index categories of fluoride concentrations in Model 6 to Model 1, the adjusted estimates differed only slightly from their corresponding crude estimates and the confounding magnitudes of -1.2% and − 3.5% created by all covariates simultaneously considered in Model 6 were trivial. The limited confounding effect of all covariates allowed a conclusion to be drawn based on the crude estimates [41] that consumption of groundwater containing natural fluoride concentrations beyond 0.7 ppm increased the prevalence of dental fluorosis, particularly fluoride levels of 1.5 ppm or higher significantly increased the prevalence by 2.75 times compared to the ones of fluoride levels below 0.7 ppm.
The strength of this epidemiological study primarily lied in its study participants who could reflect the population-level dental fluorosis problem and pave the path for public health initiatives needed to address the condition in future generations of children in these fluoride endemic areas. The case-control study design along with the application of causal directed acyclic graphs not only allowed evaluation of the temporal association between natural fluoride in groundwater and dental fluorosis but also provided a plausible elucidation of the dental fluorosis’s etiologic pattern taken various socio-behavioral determinants into account. The evidence of high dental fluorosis prevalence necessitated multi-level public health initiatives to manage the problem. Children who were affected by dental fluorosis in the forms of physical damage or apparent stain on dental enamel should be provided with access to public dental care. Public communication to inform and educate the residents, especially the families of children at the ages of developing dentition, about dental fluorosis and its prevention should be carried out by local public health officers and village health volunteers. Preventive measures including prolonged breastfeeding, toothbrushing with water for children younger than 2 years, and the use of pea-sized fluoride toothpaste for 2–6 years old children should be recommended. Fluoride mapping based on the annual records of fluoride concentrations in the groundwater used for household water supply should be utilized to identify water sources with ≥ 0.7 ppm fluoride which must be avoided for drinking and cooking. Engaging all community stakeholders to have shared accountability in developing a solution for safe water allocation to all residents should be implemented. School-based dental fluorosis surveillance should continue to monitor the situation of dental fluorosis and further provide information for pertinent public health actions.