Effect of body weight and behavioural factors on caries severity in Mexican rural and urban adolescents

Introduction

Great improvements have been realised in global oral health, even though in the under-privileged populations of both developed and developing countries such success has not yet been achieved¹. Caries aetiology is a complex process². Diet has been reported as one of its fundamental aetiological factors because the intake of fermentable carbohydrates results in a decrease of dental-plaque pH, leading to tooth demineralisation³. The intake of sugared foods and drinks is a common behavioural risk factor that caries shares with other diseases, such as obesity4., 5.. As dental caries and body weight are considered to be diet-related health outcomes, an association between the two might be found. This supposed association has been mainly studied in adolescents from industrialised countries through an index that classifies and associates weight and height, namely the body mass index (BMI)6., 7.. Even though in developing countries malnutrition is traditionally a fatal risk factor for children, obesity has now also become a major health issue. Mexico is the country with the highest prevalence of childhood obesity in the world, recently surpassing the USA in this respect⁸. The relationship between nutritional status and caries is confounded by many factors; the presence of severe carious lesions may affect growth, influencing body weight and height⁹.

In the literature, conflicting findings on the relationship between dental caries and bodyweight are reported. Four main patterns of relationships have been found: no association between BMI and dental caries; a positive relationship; an inverse relationship; and an association with the two-tailed BMI distribution curve¹⁰.

Nowadays, although the prevalence of caries has generally declined, in Latin American countries such as Mexico, the disease is still a health policy problem, mostly affecting the young population with economic (unable to afford dental services) and educational (poor dietary and oral-hygiene habits) deficiencies, as well as social inequalities (limited access to health services)11., 12.. The caries prevalence in this age group oscillates between 43.9% and 69.5%, being influenced by multiple variables, including the area of residence13., 14.. In Mexico, the prevalence of caries has been reported to be higher in rural areas than in urban areas¹⁵. Rural populations in Mexico, comprising 15 million people, present worse oral-health conditions than urban populations because of the higher percentage of poverty, an overall poor health condition and limited access to dental care16., 17.. An environmental and socio-economic homogeneity in rural areas, as a result of similar characteristics and behaviours, was observed and should be considered when preventive strategies are planned17., 18..

Starting from these premises, some questions arise. Is caries prevalence in Mexican adolescents related to body weight? And, if so, how may this affect urban and rural populations? To answer these questions a cross-sectional study was designed with the aim of investigating if Mexican adolescents’ body weight and caries severity are associated, and if rural and urban environments play a role in this supposed association.

Methods

Study population

This survey was a part of a larger evaluation aiming to assess the association between weight and caries in adolescents living in different European and non-European countries, including Mexico, promoted by the Collaboration Centre of the World Health Organization (WHO) for Epidemiology and Community Dentistry of Milan, Italy. Ethical approval was obtained from the University of Sassari, Italy (no. 1073/L 23/07/2012).

The state of Veracruz is the third most populated state in Mexico, and Veracruz City is the major industrial port; 39% of the population lives in rural areas, the highest percentage in the country (mean value 22%). Tepancan is a rural area located in the south of the state of Veracruz. Only 66% of the national population has access to national health services. Over one-third (36%) of the residents in Veracruz are under 18 years of age¹⁹.

General criteria

Adolescents (12–15 years old) from two different areas of the state of Veracruz – one rural (Tepancan) and one urban (Veracruz City) – were examined. The study was approved by the Secretary of Education of Veracruz (Rural S.E.V. 30FTV5610X; Urban S.E.V. 30FIS0030Z) and parents signed an informed consent form. The survey was performed in accordance with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards.

Rural area

The township of Tepancan has only one school and all adolescents attending it share a common background and a similar socio-economic status. All adolescents 12–15 years of age were enrolled. The population consisted of 240 subjects (98 female subjects and 142 male subjects).

Urban area

One per cent of the subjects from Veracruz City (n = 22,086) within the age range 12–15 years was randomly selected¹⁹. A post-hoc power analysis was calculated after the survey, with a λ noncentrality parameter of 18, a critical χ² = 7.81 and a power (1-α error probability) of 0.96. The final sample was 224 adolescents (105 female subjects and 119 male subjects).

The reported fluoride concentration in drinking water in the two areas was well under the recommended level of <0.7 mg/L (0.24 mg/L in the rural area and 0.15 mg/L in the urban area)²⁰.

Clinical examination

A dental operator (C.L.C.) was calibrated to perform all the examinations. Thirty subjects were examined and re-examined after 72 hours, to assess intra-examiner reproducibility calculated as kappa values.

Height (cm) and weight (kg) were measured using a mechanical physician scale with an integrated measuring rod (Rice Lake RL-MPS; Rice Lake Weighing Systems, Rice Lake, WI, USA) and scored in meters and kilograms, respectively. The BMI of each subject was calculated by dividing their weight by the square of their height (kg/m²), according to WHO criteria⁶. Caries was recorded according to the International Caries Detection and Assessment System (ICDAS) criteria, considering the depth of the lesion (caries severity) and the past and current tooth status (caries experience)²¹. Each subject was examined inside the school under natural lighting using a mouth mirror No. 5 and a WHO explorer (LM-Dental; LM-Instruments Oy, Parainen, Finland).

Questionnaire

A pretested questionnaire was administered to each subject by direct interview. The questionnaire was divided into three areas: (i) vital statistics (i.e. gender, age, presence of systemic disease); (ii) dietary patterns (i.e. description of meals, consumption, frequency and amount of sugared drinks and snacks); and (iii) oral-hygiene habits (i.e. toothbrushing, frequency of dental floss and reason for dental visit).

Statistical methods

Data from the clinical examination and questionnaire were entered in the FileMaker Pro 9.0 Runtime database and then exported to a Microsoft Excel spreadsheet (Microsoft Office 2016; Microsoft Italia, Milan, Italy). Subjects were classified, for the analysis, into two age groups: 12–13 years; and 14–15 years. Body mass index was classified for the respective age groups as normal weight (BMI = 15–21 for the 12–13 years age group; and BMI = 16–23 for the 14–15 years age group) and overweight, including obese (BMI >21 for the 12–13 years age group; and BMI >23 for the 14–15 years age group), according to the WHO guidelines⁶. No underweight adolescents were found. Dental examination data were treated as numerical ordinal data for ICDAS (ICDAS 0 = healthy; ICDAS 1–3 = caries in enamel; ICDAS 4–6 = caries in dentine): subjects were classified as healthy if all teeth were coded ICDAS = 0; subjects were classified as having caries in enamel if at least one tooth was coded ICADS = 1–3; and subjects were classified as having caries in dentine if at least one tooth was coded ICADS = 4–6. Subjects affected by systemic diseases, such as diabetes, asthma, cardiovascular diseases, etc., were classified as not in good general health.

Questionnaire data were categorised as follows: the toothbrushing frequency was divided into ‘once a day’, ‘twice a day’ and ‘more than twice a day’; the use of dental floss was split into ‘irregular’ if less than once a day and ‘regular’ if at least once a day; and frequency of dental check-ups was catergorised as ‘irregular’ if subjects attended only when in pain and as ‘regular’ if subjects attended scheduled dental check-ups. The sugared drink intake was divided into ‘never’, ‘less than twice a day’ and ‘more than twice a day’ and sweet snack intake was divided into ‘never’, ‘less than three times a day’ and ‘over three times a day’.

All data were analysed using the software STATA (Mac version 10.1, Stata for Mac 10.1; StataCorp LLC, College Station, TX, USA). Descriptive statistics as cross-tabulations and linear trends were calculated for normal weight and overweight, caries severity (ICDAS) and area of residence, gender, age, toothbrushing frequency, use of dental floss, frequency of dental check-ups and intake of sugared drinks and snacks. The data showed that overweight and area of residence were statistically associated; therefore, a dummy variable as the sum of overweight data and area of residence was generated (‘BMI/Area’).

Multinomial logistic regression models were performed using ICDAS scores as the dependent variable. The Akaike information criterion (AIC) was used to measure the goodness of fit of the statistical model²². The possible modifying effects of covariates on the outcomes were tested by an interaction model (likelihood ratio test statistic). Multicollinearity might sometimes cause problems with regression results. This problem was solved using the DFBETA command in STATA, dropping the information that has too much influence on the regression line. However, after data elaboration, no statistically significant multicollinearity was observed and therefore it was decided to report findings without outliers. For all analyses, the statistical significance level was set at α=0.05.

Results

A total of 464 adolescents (261 male subjects and 203 female subjects, mean age 13.5 ± 0.9 years) participated in the study (240 from the rural area and 224 from the urban area). Caries prevalence was 33.62% in participants from the rural area and 42.41% in participants from the urban area, and caries experience (past and current caries status) was higher in subjects from the rural area (1.38 ± 2.63; range 0–20) than in subjects from the urban area (1.16 ± 1.48; range 0–7), data not shown. Urban subjects had a higher prevalence of caries in enamel compared with the rural subjects (38.39% vs. 20.76%), but a lower prevalence of caries in dentine (4.02% vs. 12.86%) (P < 0.01). Sample distribution across BMI (normal weight and overweight) and caries severity (using the ICDAS scores according to area of residence) is shown in Figure 1. The sample distribution across BMI and caries according to area of residence is displayed in Table 1 as number of subjects and percentage. In the urban area the sample was perfectly split in half (normal weight and overweight), while in the rural area the number of overweight adolescents was nearly double (n = 159) that of normal-weight adolescents (n = 81). No underweight subjects were present in either area. Body mass index and area of residence were statistically significantly associated (χ² = 12.59, P < 0.01). The area of residence was statistically significantly associated (χ² = 24.23, P < 0.01) with caries severity, with the highest number of subjects with caries in dentine in the rural area group; in the overall population, overweight was not statistically significantly associated with caries severity.

Figure 1.

Sample distribution across body mass index (BMI) and caries severity, using the International Caries Detection and Assessment System (ICDAS) scores stratified according to area of residence. NW, normal weight; OW, overweight.

Table 1.

Sample distribution across caries severity, using the International Caries Detection and Assessment System (ICDAS) scores

Variable	Healthy (ICDAS 0) n (%)	Caries in enamel (ICDAS 1–3) n (%)	Caries in dentine (ICDAS 4–6) n (%)	P > |z|
Area
Urban	129 (57.59)	86 (38.39)	9 (4.02)	<0.01
Rural	159 (65.98)	50 (20.76)	31 (12.86)
BMI
Normal weight	121 (62.70)	57 (29.53)	15 (7.78)	0.86
Overweight	167 (61.62)	79 (29.15)	25 (9.22)
BMI/Area
Normal weight urban area	64 (57.14)	41 (36.61)	7 (6.25)	<0.01
Overweight urban area	65 (58.03)	45 (13.15)	2 (2.15)
Normal weight rural area	57 (62.64)	16 (17.58)	8 (8.79)
Overweight rural area	102 (64.15)	34 (21.38)	23 (14.46)

BMI	Rural	Urban
Normal weight	81 (42.97)	112 (58.03)	<0.01
Overweight	159 (58.67)	112 (41.37)

Distribution according to area of residence, body mass index (BMI) and dummy variable BMI/Area. BMI in according to area of residence is also reported Statistically significant values are reported in bold style.

The dummy variable BMI/Area was associated with caries severity (χ² = 27.47, P < 0.01) [i.e. regarding caries in dentine, the percentage of overweight subjects (14.46%) living in the rural area was higher than the percentage of normal-weight subjects (6.25%) living in the urban area, data not shown].

The relationship between dental check-ups and caries status was relevant only in the urban population (P < 0.01). The distribution of behavioural habits such as toothbrushing frequency, dental check-ups and dietary patterns were more homogeneous in rural areas than in urban areas (data not shown). The association between both irregular flossing/higher number of caries in enamel and poor general health and caries in dentine were statistically significant (Table 2).

Table 2.

Sample distribution across caries severity, using the International Caries Detection and Assessment System (ICDAS) scores)

Variable	Healthy (ICDAS 0) n (%)	Caries in enamel (ICDAS 1–3) n (%)	Caries in dentine (ICDAS 4–6) n (%)	P > |z|
Gender
Male	163 (35.13)	74 (15.95)	24 (5.17)	0.81
Female	125 (26.94)	62 (13.36)	16 (3.45)
Age
12–13 years	158 (34.05)	70 (15.09)	24 (5.17)	0.61
14–15 years	130 (28.02)	66 (14.22)	16 (3.45)
Good general health
Yes	278 (59.91)	131 (28.23)	35 (7.54)	0.03
No	10 (2.16)	5 (1.08)	5 (1.08)
Toothbrushing frequency
1/day	66 (14.22)	38 (8.19)	11 (2.37)	0.79
2/day	89 (19.18)	41 (8.84)	13 (2.80)
>2/day	133 (28.66)	57 (12.29)	16 (3.45)
Flossing
Irregular	147 (31.68)	89 (19.18)	21 (4.53)	0.02
Regular	141 (30.39)	47 (10.13)	19 (4.09)
Dental check-ups
Irregular	164 (35.35)	89 (19.18)	27 (5.82)	0.15
Regular	124 (26.72)	47 (10.13)	13 (2.80)
Sugared drinks
Never	25 (5.39)	12 (2.59)	6 (1.29)	0.50
<2/day	132 (28.45)	58 (12.50)	17 (3.66)
>2/day	131 (28.23)	66 (14.23)	17 (3.66)
Sweet snacks
Never	60 (12.93)	22 (4.74)	9 (1.94)	0.78
≤3/day	186 (40.09)	95 (20.48)	25 (5.39)
>3/day	42 (9.05)	19 (4.09)	6 (1.29)

Distribution according to gender, age, general health, oral-hygiene habits (toothbrushing and flossing), dental check-ups and frequency of the consumption of sweets (drinks and snacks). Statistically significant values are reported in bold style.

Intake of sweet snacks was higher in the urban population (P = 0.04): their consumption was nil for 19.61% of this population, moderate for 65.96% (one to three times a day) and high for 14.43% (>3 times a day). In addition, even if the 9.27% of the adolescents did not consume any sugared drinks, the 46.12% had a high intake (Table 1).

Multinomial modelling using ICDAS as the dependent variable and including BMI/Area, good general health and flossing, demonstrated that the dummy BMI/Area variable was always statistically significantly associated with caries both in enamel and in dentine (P = 0.03 and 0.02, respectively). Flossing was statistically significantly associated with the presence of caries in enamel (P < 0.01), while general health status was related to caries in dentine (P < 0.01) (Table 3).

Table 3.

Relative risk ratios (RRR) after multinomial logistic regression, using the International Caries Detection and Assessment System (ICDAS) scores as dependent variable (ICDAS score of 0 = healthy; ICDAS score of 1–3 = caries in enamel; ICDAS score of 4–6 = caries in dentine)

ICDAS	Covariates	RRR	SE	P > |z|	(95% CI)
0	(base outcome)
1–3	BMI/Area	0.73	0.10	0.03	0.56–0.96
	Flossing	0.56	0.12	< 0.01	0.36–0.85
	Good general health	1.04	0.59	0.95	0.34–3.13
4–6	BMI/Area	1.78	0.45	0.02	1.08–2.91
	Flossing	0.90	0.31	0.77	0.46–1.77
	Good general health	4.65	2.74	< 0.01	1.47–14.75
Number of obs. = 464		P < 0.01		Log likelihood = − 389.27

95% CI, 95% confidence interval; BMI, body mass index; obs., observations; SE, standard error. Statistically significant values are reported in bold style.

Discussion

This study aimed to evaluate the possible association between body weight and caries severity in adolescents living in different areas of Mexico, the country where the highest prevalence of childhood obesity is reported. The study failed to prove an association between body weight and caries severity in the overall population; however, when overweight and area of residence were combined, the new variable was proven to be statistically significantly associated with caries severity.

The relationship between childhood obesity and dental caries has been extensively debated in the literature. Even though an overall significant association was found in a systematic review with meta-analysis⁷, a lack of association was mainly described in Latin American countries5., 23.. Nevertheless, obesity appears to be associated with caries in the primary teeth of preschool Mexican children from urban areas²⁴. Different confounding, effect-modifier and frank risk predictors might explain a negative or positive association between overweight and caries presence, such as the age of the study sample²⁵, the intake frequency of sugared products²⁴ and the socio-economic status of the subjects²⁶.

In Mexico, one in three adolescents 12–19 years of age is overweight or obese and the prevalence of overweight has increased by 26% for both genders, to more than 4.1 million schoolchildren²⁷. In this survey, overweight was recorded in two-thirds of the sample living in the rural area, and in half of the sample in the urban area; underweight was not present in either area.

Diet plays a key role in overweight and obesity, as well as in caries development, as a result of the high consumption of foods rich in fats and carbohydrates²⁸. The negative effect of obesity on children’s oral health has been studied²⁹. Dietary habits of children and adolescents are strongly related to the socio-economic status of their family, to the geographical location and to the area of residence (rural/urban)³⁰. In this survey, overweight and area of residence were statistically significantly associated with each other and with caries severity when considered as a single variable; overweight subjects living in the rural area had more severe caries lesions compared with the rest of the sample. This finding differs from those reported in the literature, in which caries severity was higher in studies demonstrating an inverse association between body weight and dental caries than in studies finding a positive or no association¹⁰.

Although the Mexican agricultural society has maintained its traditional dietary habits, based mainly on corn, rice and beans, there has been an increase in the intake of commercial products and snacks; moreover, rural populations have adopted a more sedentary lifestyle³¹. Children from rural areas, socioeconomically disadvantaged, are more likely to be obese than children living in cities, suggesting that the rural environment may favour this condition32., 33..

Behavioural factors differed in rural and urban areas. Urban subjects reported better oral-hygiene habits and more frequent dental check-ups, but a higher consumption of sugared snacks and drinks; in this group a larger number of subjects had caries lesions with low severity compared with adolescents living in rural area. Although access to health services is considered to be easier in urban areas than in rural areas, only 69% of the children and adolescents living in Veracruz City had access to health-care facilities¹⁷. In the rural area, a lack of health service facilities and poor oral-hygiene habits were reported. These findings might explain why, even if the number of children with caries lesions was lower, the severity was higher. In Mexico, public dental health services offer only fillings, extractions and prevention. For other treatments, patients need private practitioners. Access to health care depends on the characteristics of individuals and the community, according to geographical, social and cultural factors³⁴. In addition, it may be considered that dental-care need is perceived differently between rural and urban children.

The strengths and weaknesses of the present study need to be addressed. A limitation is the lack of information regarding the actual health-service facilities in the rural area. A second limitation is the sampling technique in the rural area because a convenience sample of schoolchildren was enrolled, therefore excluding children not attending school. In contrast, in the urban area, a random procedure with the purpose of selecting an equivalent sample was performed. A further limitation might be attributed to the questionnaire: it was administered as a direct interview and it cannot be totally excluded that some adolescents, fearing the judgement of the interviewer, modified their responses. This is the first study that has verified the possible association between weight and caries severity in a sample of Mexican adolescents in relation to the different urban and rural living environments.

It is possible to speculate that, unlike in the urban area, environmental homogeneity is present in the rural area; this has been described in previous studies, in which a reduction in the impact of confounding factors on data analysis and interpretation was detected¹⁸.

The caries value reported in this survey suggests that different preventive strategies need to be planned in order to address the presence of oral disease, not only globally, but also within the same country and region35., 36..

Conclusions

In the present survey, weight was not associated with caries severity for the overall adolescent population. Nevertheless, when overweight was associated with the area of residence, it became a statistically significant risk indicator for caries severity in the adolescents living in the rural area, in which the highest prevalence of overweight subjects was recorded. Moreover, in this population a lower number of caries lesions with a higher severity emerged, probably related to the limited access to oral-health services. In contrast, higher caries prevalence with lower severity was observed in adolescents living in the city, as well as a higher consumption of sugared snacks and drinks. Further studies, aiming to investigate the trend of overweight in rural areas of Mexico, are needed in order to address both general and oral health-preventive strategies.

Acknowledgement

This study did not receive financial support.

Conflict of interest

The authors declare they have no conflict of interest.