Early childhood caries (ECC) is a disease that involves development of one or more caries lesions, with or without cavitation, by the age of 71 months [1]. Epidemiological studies have indicated that ECC has a strong social element [2–4]. Indeed, higher prevalence in non-developed countries [5, 6] and in socioeconomically disadvantaged groups living in low, middle [2–4] and high-income countries [7, 8] have been documented.

Increased concentration of ECC within a specific group signifies a polarizing phenomenon, defined as a greater number of cases within a small segment of the population [9–11]. Polarization of caries has often been associated with socioeconomic deprivation in populations at social risk [9, 11–13]. This phenomenon may result from the non-homogenous distribution of strategies to control caries in populations, such as improvement in socioeconomic status and access to preventive dental care, fluoride, and dental education programs [9].

Polarization of ECC is well documented; the majority of young children are caries-free, and most carious lesions are concentrated in a small number of children [14–16]. The concentration of caries disease in a segment of the population has also been demonstrated in Brazil, in preschool [11] and school age children [9, 17]. The phenomenon is also associated with socioeconomic determinants in this population. In addition, polarization of caries disease has also been documented among Brazilian adolescents [10, 18] associated with skin color and the presence of tooth pain in the previous half year [10].

In a previous study involving children 18 to 42 months of age [19], we observed that caries disease was polarized in a population characterized by low socioeconomic status in the peripheral neighborhoods of São Luís, Maranhão Brazil. Given that socioeconomic factors have been associated with ECC risk and its polarization among children, it is important to understand what factors mediate susceptibility to the disease. Elucidation of these factors may reveal why, within a group of children who are all vulnerable, some have more caries than others.

Identifying modifiable proximal factors that mediate the association between socioeconomic status (SES) and ECC polarization may help design interventions to ameliorate the burden of this disease, especially among low-income groups. To the best of our knowledge, no previous study has explored proximal factors that may mediate the association between SES and caries polarization. The objective of this study was to use a hierarchical approach to identify factors associated with ECC polarization within a vulnerable low-income population.

Among the 260 children whose parents signed informed consent forms, 16 (6.2%) were excluded from the study because they did not complete all steps of the investigation. In our final analyses, we used data from 244 children. In this study, ECC prevalence was 32%. Mean dmfs index was 0.42 and the standard deviation (SD) was 0.69 in the whole sample. Separating the dmfs index into their components, only the component cavity was observed in our sample.

In the sixth block, the variable of the mother or caregiver brushing the child’s teeth for them was a significant protective factor in the few caries group (p = 0.004; PR = 0.27; 95% CI 0.11–0.66) but was not significant in the high caries group. The biological variables in the seventh (VPI) and eighth blocks (S. mutans contaminated saliva) were not associated with childhood caries.

This study revealed that high sucrose consumption was associated with caries polarization among children living in vulnerable social condition after controlling for confounding variables through a hierarchical approach.

The phenomenon of caries polarization among children has been associated with social deprivation [24]. However it has been shown that even in populations with homogenous social background caries tend to concentrate in certain population subgroups [9, 24]. Reasons for this are poorly understood. In the present study we showed that higher sugar consumption is possibly the factor that mediates the association between low SES and caries polarization.

Indeed, studies have indicated that low socioeconomic level can alter dietary patterns [25] so sugar consumption may be a pathway to the association between low family income and caries polarization, since both socioeconomic factors [3, 4] and sucrose consumption [4, 15, 18, 19, 25, 26] also present robust associations with ECC. It is possible that synergy between deprived social conditions and high-sugar dietary patterns may explain these findings.

In the present study, we used the SiC index to identify the high caries group as a proxy for disease polarization, showing a mean of 3.82 lesions. Other studies involving children up to five years of age showed a SiC value similar to ours, such as a SiC value of 4.14 in Mexican children [27] and of 4.09 in Sri Lankan children [12].

Ours results showed that age was associated with ECC, in the high caries groups, confirming results of previous studies indicating that caries severity increases with age [19, 28–32].

The demographic, economic, and social variables (income, schooling, health information) in the second level were not associated with ECC. These data differ from results of two prior studies [2, 3], which found that social variables were associated with ECC. This is likely explained by the homogeneity of social variables in the cohort included in this study. Although the homogeneity lessened the power of social variables, homogeneous social variables are desirable in a study investigating variables that explain polarization of ECC incidence to reduce confounding by SES.

In the few caries group, the non-polarization phenomenon may be attributable to several protective factors, including access to a dentist, being breastfed after the age of 1 year and teeth brushing by the mother. Thus, it seems that even in deprived low-income groups, access to health care, protective feeding practices and oral hygiene might buffer the deleterious effect of high sugar consumption on ECC.

Access to the dentist was protective factor only in the few caries group (PR = 0.20). It’s possible that children of high caries group may have gone to the dentist for restorative care or even due to toothache. In this sense, it has been shown that access to dental treatment among preschoolers occurs especially for emergency care [33].

Similarly, teeth brushing performed by the mother were also protective only for the few caries group (PR = 0.27). The high caries group may not have had adequate hygiene practices, resulting in higher DMFS scores.

Nocturnal breastfeeding after age 1 (PR = 0.33) was a protective factor of borderline significance in the group with few caries (p = 0.051). It is possible that the small sample size could have contributed to the borderline association between nocturnal breastfeeding and ECC observed in the group with few caries. In the high caries group, however, this practice was not found to protect from the disease, possibly because children with greater number of lesions had other deleterious dietary practices superimposed on breastfeeding, such as consumption of more sucrose. Others epidemiological studies failed to show that breastfeeding after age 1 were associated with ECC [34–36], including previous analysis of data involving the same population of children [19].

The fact that we used a convenience sample is a limitation to the external validity of this study’s findings. However, it is possible that our results could be extrapolated to other low SES populations living in similar conditions.

One of the strengths of our study is that since ECC is considered a social disease, studying a population with homogenous social and economic conditions favors the identification of factors that stratify risk within vulnerable groups and allows for better control of confounding factors. Furthermore, the use of a hierarchical model permitted us to evaluate mediation of more proximal factors on the association between distal variables such as SES and ECC.

In conclusion, higher sucrose consumption increases the likelihood of ECC in socially vulnerable children. Caries population prevention strategies based on teeth brushing and visits to a dentist seems not to be enough to reduce inequalities in oral health because they do not seem to be reaching the most vulnerable groups among the vulnerable. Another possible explanation is that these preventive practices are not sufficient to counterbalance the deleterious effect of high sugar consumption in the high-caries group of children. Public policies focused on warning about the risk of added sugars consumption might be an additional strategy to control ECC in these deprived low-income populations.