Caries experience in children under 5 years old in the Xingu indigenous park in Brazil

Aim: The present study sought to investigate dental caries experience and its association with sociodemographic, postnatal and breastfeeding variables in children in the agerange from 6 to 71 months of age, in the Xingu Indigenous Park, Mato Grosso, Brazil. Methods: This was an analytical cross-sectional study that used secondary data pertaining to 402 indigenous children of the Low, Middle and Eastern Xingu regions, who participated in the Oral Health Epidemiological Survey in 2013. The dependent variable was dental caries, dichotomized by the median (dmf-t≤1 and dmf-t>1). The data of independent variables were obtained by means of instruments of the Local Health Information System of the Xingu Indigenous Special Sanitary District (DSEI). Raw analyses were performed to test the association of the independent variables with the dependent variable. The variables were tested in the multiple logistic regression model. Results: The mean value of the dmf-t index was 2.60 and the prevalence of affected children was 51%. In the multiple analysis, only children older than 36 months (OR: 6.64; CI95%: 4.11 to 10.73) and those that were breastfed for a longer period of time (OR: 1.88; CI95%: 1.16 to 3.02) showed significant association with the dmf-t>1 index. Conclusion: Childhood dental caries among indigenous children was associated with age and breastfeeding prolonged for over 26 months, therefore, pointing out the need to offer dental follow-up care at earlier ages.


Introduction
In Brazil, there are approximately 820 thousand indigenous children, who represent almost a third of the total number of Brazilian children 1 , and they are the ones who suffer most with health problems, when compared with non-indigenous children 2 .
As far as oral health is concerned, an unequal distribution of caries disease among children is suspected. The data of the last national oral health survey, known as "SB Brasil 2010," revealed that 5-year-old indigenous children have double the number of cases of dental caries when compared with non-indigenous children in the same age group 3 . This warning has instigated the need for investigations into the area, and it would be interesting to find out about the trajectory of the disease and its early onset, which occurs before children reach the age of 71 months of life 4 .
Early childhood caries causes pain, difficulties with eating, speaking, sleeping, and socializing. It may compromise children's development and growth and increase the probability of developing disease in their permanent dentition 5 .
This disease develops as a result of the interaction of multiple individual, social and environmental components 6 , such as sex, age, dietary habits, social conditions, and access to oral hygiene instructions [7][8][9] . Moreover, the causal relations between the time of breastfeeding and caries disease, especially among younger children, must be emphasized 10-12 . Reports in the literature have indicated that the concentration of lactose in human milk has the capacity to reduce the pH of dental plaque, thereby contributing to the establishment of caries, particularly when plaque is not removed from the tooth surface by means of correct brushing 13 . Indigenous women of some Brazilian ethnicities are known to have the habit of offering breastfeeding on free demand and for an indefinite period of time 14 , and this is a behavior that must be checked in the indigenous communities, in view of the occurrence of oral health problems, such as dental caries 15 .
The prevalence of dental caries has remained high among the people of the Xingu Indigenous Park (PIX) over the last few years 16 , and the oral health conditions of indigenous children under the age of 5 years, together with the factors associated with them, have hardly been investigated.
Therefore, the aim of the present study was to find out about dental caries experience among children of PIX, under the age of 5 years, and test the hypothesis that caries would be associated with the sociodemographic conditions, post natal and breastfeeding variables. This information could contribute to subsidize strategies for the prevention and control of caries disease among the indigenous children.

Ethical aspects
The project was summited to and approved by the  17 .
The major portion of the PIX population consists of youngsters, of whom 38.2% under the age of 15 years; and approximately 5% are persons over the age of 60 years. As regards proportion between sexes/ethnicity in the territory, a higher number of women (51%) is observed, and predominance of the Kaiabi ethnicity, of which 43% of the local population is composed 18 .

Eligibility Criteria
The survey involved 44 villages in the Low, Middle and East Xingu regions, between the months of April 2012 and April 2013. In the present study, the dmf-t data considered were those of all the children between the ages of 6 and 72 months, who participated in the epidemiological survey. Of the total of 431 existent children under five years of age, 29 (7.7%) did not participate in the inquiry because they refused or were absent during the period when the exams were performed.

Calibration of examiners
The examiners were calibrated in accordance with the standards adopted in epidemiological surveys, which consisted of the following steps: a) training in the theoretical framework of the variables used; b)acceptance of the criteria adopted for defining each observation of the exam and its respective codes; c) application of the criteria in real situations; that is, the calibration itself; and e) calculation of intra-and inter-examiner errors, based on the coefficient. The pondered Kappa coefficient for each examiner considered the minimum acceptable limit to be the value 0.65. The Kappa inter-examiner value obtained was 0.90. The data were noted on paper forms by 3 Indigenous Oral Health Agents, who acted as note-takers.

Study variables and instruments
In the survey, clinical oral exams were performed to identify decayed, missing and/or filled primary teeth, in order to calculate the dmf-t index. The sociodemographic, post-natal and breastfeeding data were obtained from the Local Health Information System (SLIS) and DSEI Xingu, instruments of records dated 2013. The sociodemographic (basic pole, sex, age, Village population size and presence or absence of Indigenous Health Agent or Indigenous Sanitation Agent) were obtained from the demographic census of the villages of DSEI Xingu 17 . The post-natal data (type of delivery and birth weight) were obtain from the birth spreadsheet records, and data about time of breastfeeding were localized on the daily follow-up map of children under the age of five years. This instrument is used by the Indigenous Health Agents and nurses for filling out the breastfeeding data that are inserted into the Local Health Information System of DSEI Xingu.

Statistical Analysis
The sample size of 402 participants provided a test power of 90% (β = 0.10) with a significance level of 5% (α = 0.05) for the effect size found in the study (Odds Ratio of 2.0 and 30% of response in the unexposed group).
The time of breastfeeding was calculated by measuring the time interval from the child's birth up to the time when the inquiry was conducted. Therefore, if the child were 24 months old on the date of the exam and was still breastfeeding, this was the total period of breastfeeding recorded. Therefore, the time of exclusive breastfeeding was not considered, instead, the time of breastfeeding either associated with or without complementary feeding was considered.
The data collected were tabulated in a Microsoft Excel® 2010. The variable dental caries was described by stratifying the data into age groups and presented as absolute and relative frequency distributions, mean dmf-t values, standard deviations, and dental caries prevalence. The dental caries index (dmf-t) was considered the dependent variable, dichotomized by the median of dmft (value=1) into: dmf-t≤1 and dmf-t>1).The independent variables, child's age and time of breastfeeding were also dichotomized by the median into ≤36 months and >36 months and ≤26 months and >26 months, respectively. The other variables were grouped in the following manner: presence of Indigenous (AIS) or Sanitary (AISAN) Health Agent in the village: yes or no; basic pole: Diauarum, Pavuru or Wawi; child's sex: female or male; type of delivery: vaginal or cesarean; birth weight: <2500g or ≥2500g; village population size: small, for villages with population smaller than or equal to 50 inhabitants, medium for villages with population between 50 and 250 inhabitants or large, for villages with population larger than 250 inhabitants.
Association of the independent variables with the dependent variable was tested by raw analysis. The variables with p<0.20 in the simple analysis were tested in the multiple logistic regression model by using the stepwise procedure. The Odds Ratio (OR) and respective confidence intervals of 95% (IC) were estimated for the variables that remained in the final model, at the level of significance of 5%. All statistical tests were performed with SAS program 9.3 (Institute Inc., Cary, NC, USA). A residual analysis was performed and applied the Hosmer and Lemeshow Goodness-of-Fit test that showed a good fit of the model (p=0.6880).

Results
Of the total number of 402 indigenous children studied, aged between 6 and 71 months, 208 (51.70%) were of the female sex. Over half of the children had had previous caries experience, with a mean dmf-t index of 2.60; with children in the group between 60 and 71 months old showing the highest dmf-t index (5.81). The highest caries prevalence was found in children from 48 to 59 months old, with 87.87% of these children being affected by the disease (Table 1).  Table 2 shows that there were a higher number of participants in the low region of Xingu Indigenous Park (54.98%), with predominance of the female sex (51.74%), and children delivered by vaginal birth (92.54%). The majority of the children had a birth weight equal to or higher than 2500 g (88.81%), and a time of breast feeding shorter than or equal to 26 months (51.00%). The majorities of the villages were of medium size (53.48%), and had the presence of an AIS/AISAN (97.51%).
In the multiple logistic regression, the children over the age of 36 months had 6.64 times more chance of having a caries index higher than 1 than those of an age lower than or equal to 36 months. Furthermore, children who were breastfeed for longer than 26 months (OR=1.88) had more chance of having a dmf-t index higher than 1 (Table 2).

Discussion
The growing interest in childhood dental caries has been emphasized in the latest national researches that have investigated the indigenous populations [19][20][21] .
Among the children under 71 months old, in the present research over half of them were found to have experienced caries, which showed association with age and time of breastfeeding.
In indigenous children, the association of dental caries experience with the increase in their age is plausible since a previous study found that among the indigenous women of Kaiowa ethnicity in Mato Grosso do Sul, the mean dmf-t tripled with the increase in children's age, especially between 12 and 24 months of life 15 .
While among the indigenous Baniwa of Alto Rio Negro, the mean dmf-t index more than doubled between the ages of 36 and 48 months 19 ; and among the indigenous Kotiria in Alto Rio Negro, the mean dmf-t index was 4.72 among children from 1 to 5 years of age 20 . When all the ethnicities studied in the Xingu Indigenous Park were considered, an increase was also observed in the mean dmf-t index in the age group from 0 to 36 months and from 48 to 60 months old 21 . The justification for the association between these two variables is based on the cumulative nature of the caries index over the course of years 22 and on the relations between the development of caries lesions and chronology of tooth eruption. This is because the tooth surfaces are affected as a result of the time during which they remain exposed to the risk factors present in the oral cavity 23 . This is a natural fact that has been observed, however, the factors that cause concern are the magnitude of differences in the prevalence rates and levels of severity of the disease among the age groups studied.
When the epidemiological findings were compared with the results of the last national oral health survey in Brazil, the mean dmf-t of children from 60 to 71 months of age in this study was observed to be higher than the national mean value of non-indigenous children (2.43 teeth with caries experience) 24 . Similarly, the prevalence of children with caries was also higher than the Brazilian mean (53.4%) for the age group from 60 to 71 months, demonstrating the possibility of inequalities existent between indigenous and non-indigenous populations, and revealing the need to implement strategies of preventive actions directed towards this segment of the population.
Indeed, progressive reduction in dental caries had been expected in the Low, Middle and East Xingu regions, as a result of a series of public and social actions in the region, especially the increase in financial investments, from the time the Xingu Indigenous Special Sanitary District (DSEI) was created in 1999. This strategy, together with the training and presence of the Indigenous Health Agents (AIS), and Indigenous Sanitation Agents (AISAN) has improved and extended the structure of the health services 25 . In spite of the importance of these professionals in the villages, no association of this variable with the reduction in prevalence of childhood caries experience was found in the present study. The competence of these professionals is not confined exclusively to oral health actions, as they need to provide general guidance about health care during breastfeeding and healthy eating habits of small children. Moreover, public health surveillance forms part of public health practice and involves actions of health promotion developed by these professionals 25 . Therefore, we suggest that reflections about educational methods in oral health should be shared with the technical professionals that work in the villages.
The high dental caries indexes in both PIX and other indigenous populations are influenced by various determinants such as living conditions [7][8][9] . The sociodemographic conditions were investigated from this aspect relative to the outcome. The variable village population size was found to be associated with the dmf-t index exceeding 1, however, it did not remain in the adjusted multiple logistic analysis model. Village population size is a factor that depends on the sociocultural characteristics of each ethnicity, since some ethnic groups form large communities, and others live in small family nuclei, exhibiting extensive ethnic diversity (16 ethnicities). Therefore, association between village population size and dental caries could be related to the different factors, e.g., economic, sociocultural, localization, infrastructure, type of access and use of health services, and the different ethnicities of residents in the park. Alves Filho et al. 24 (2013) and Arantes and Frazão (2016) 26 have previously demonstrated the association of these variables with the occurrence of dental caries in other indigenous populations.
The authors point out that that the World Health Organization recommends the practice of breastfeeding up to the age of two years; moreover, it should be complemented with other foods as from six month of age, as this is fundamental for healthy development 18 . In the present study, almost half of the children (46.7%) of the villages in the Low, Middle and East Xingu regions were breastfed for extended periods of over 26 months. This clearly benefitted both mothers and children, as has been described in the literature 12 . Nevertheless, although this practice apparently was advantageous to indigenous community, the children breastfed for a longer period of time, and older children were observed to be those most affected by caries experience.
Each ethnic group has its own peculiarities that influence the practice of breastfeeding 14,19. Among the indigenous of Xakriabá ethnicity, interior of the state of Minas Gerais, the median time of breastfeeding was from six to twelve months, with high prevalence of early weaning, in which 49% of the children no longer received mother's milk before they completed one year of life. The short duration of breastfeeding was explained by the introduction of water and teas, according to guidance provided by the village elders and contact with non-indigenous people 14. Whereas in the Xingu Indigenous Park, indigenous children culturally tended to be breastfed from the first year through to the third year of life, irrespective of the introduction of solid foods 27,28 .
The relationship of association between breastfeeding for longer than 24 months and being affected by dental caries has previously been shown in the literature. Chaffee et al. 10 (2014) and Tham et al. 11 (2015) affirmed that dental caries was associated with the time of breastfeeding when the child was offered the breast frequently during the day and night. This led to reduced saliva production and facilitated milk stagnating in the mouth, so that occurrence of the two factors with absence of oral hygiene led to caries.
There is a hypothesis that mother's milk has factors that provide protection against dental caries, such as the presence of casein protein and imunoglobulin A (IgA), which inhibit the growth and adhesion of cariogenic bacteria on tooth surfaces, particularly Streptococcus mutans. In breastfed children, another complementary mechanism of initial protection against caries disease is the healthy oral microbiome established by exposure to mother's milk and her skin 29  Therefore, breastfeeding in early childhood may well protect against dental caries, however, the time of breastfeeding and behavioral components (such as cleaning the teeth and supporting structures) may significantly contribute to the increase in dental caries in children breastfed after 12 months of age 11 . Furthermore, the risk of dental caries increases, depending on the carbohydrate content, acidity and frequency of consumption of these types of foods. When associated with poor oral hygiene, these foods become essential factors for development of the disease, making them factors that could be confounded with the time of breastfeeding 11 . Therefore, it was perceived that breastfeeding of itself was not the sole factor contributing to caries disease progression, but that other factors, including its practice in conjunction with a cariogenic diet, were involved.
The above-mentioned conditions are exacerbated by the absence of behaviors relative to oral health care, especially as the children get older, new primary teeth erupt, and they are exposed to certain groups of foods, possibly leading to changes in the microbiome and increasing the risk of developing dental caries.
The present study had some important limitations related to the methodological approach, with respect to controlling potential confounding factors, such as cariogenic diet and oral hygiene that were not studied, and could have had a negative impact on the children's oral health. Another limitation was the cross-sectional study design that did not allow causal inferences to be made, therefore, only associations between the independent variables and the dependent variable -dental caries (dmf-t index) could be demonstrated.
However, the study was conducted with the intention of identifying caries experience in indigenous children under the age of five years, and possible factors related to this outcome, in order to produce new knowledge, based on strategies with actions to improve the health of this segment of the population The findings on dental caries in indigenous children of PIX demonstrated the need for actions promoting oral health, preventing oral diseases and providing dental care for children in the age group studied, within the Oral Health program of DSEI Xingu. As a coping strategy, we suggest that the oral health teams become more involved in periodic actions of prenatal care and puericulture. Continual guidance and encouragement of exclusive breastfeeding up to 6 months, and complementary feeding without restrictions of time or frequency, up to at least 2 years of age, as recommended by the World Health Organization, must be provided, considering the benefits of breastfeeding to both mother and child 11,12 . Moreover, oral hygiene instructions and healthy dietary practices must be reinforced, considering the sociocultural and local characteristics of the community.
Lastly, we suggest that educational actions implemented should involve members of the family, such as caregivers (sibling, uncles and aunts, grandparents) and that members of the community (midwives, shamans, teachers) should participate in demonstrating the traditional care practices. Since the villages have no fluoridated water, we suggest periodical distribution of oral hygiene products, extension of preventive actions of topical fluoride application, including children under 5 years of age. The IAS/AISAN in the villages could also develop periodical actions of guidance on mother and child self care, which must be guaranteed by DSEI Xingu, with a view to improving the oral health conditions of this group.
This study revealed that dental caries experience in indigenous children in the PIX, between the ages of 48 and 71 months, and of those older than 36 months was associated with breastfeeding for a period of longer than one year. Therefore, implementing preventive measures and oral health care for indigenous children under five years of age, and providing the mothers with guidance about oral care to be performed during the breastfeeding period are indispensable actions for diminishing the incidence of caries in this segment of the population.