Abstract
Objectives
This study aimed to explore the association between early life factors and dental caries among 5-year-old Chinese children.
Methods
Data from 9722 preschool children who participated in the third National Oral Health Survey of China were analysed. Information on early life (birth weight, breastfeeding and age when toothbrushing started), child (sex, ethnicity, birth order and dental behaviours) and family factors (parental education, household income, place of residence, number of children in the family, respondent’s age and relation to the child) were obtained from parental questionnaires. Children were also clinically examined to assess dental caries experience using the decayed, missing and filled teeth (dmft) index. The association of early life factors with dmft was evaluated in negative binomial regression models.
Results
We found that birth weight was not associated with dental caries experience; children who were exclusively and predominantly formula-fed had lower dmft values than those exclusively breastfed; and children who started brushing later in life had higher dmft values than those who were brushing within the first year. Only one in seven of all children received regular toothbrushing twice per day, and only 34.7% had commenced toothbrushing by the age of 3 years.
Conclusions
This study shows certain early life factors play a role in dental caries among Chinese preschool children and provides important insights to shape public health initiatives on the importance of introducing early toothbrushing.
Clinical significance
The early environment, especially the age when parents introduce toothbrushing to their children, can be an important factor to prevent childhood dental caries.
1
Introduction
Untreated tooth decay is the most common health problem worldwide . The Global Burden of Disease (GBD) Study showed 9% of children had untreated decay into dentine in primary teeth − sufficiently severe to require a filling or extraction − with a peak at age 5–6 years . Dental caries can negatively affect children’s well-being, including growth and school performance , and is very costly to treat, thereby placing a major financial burden on families and the healthcare system . Researchers have recently started to examine distal determinants of dental caries, i.e. those beyond child behaviours and oral factors, to identify high-caries risk children early in life .
A life course approach to chronic disease epidemiology recognises how exposures early in the lifespan influence disease risk later in life, and hence, may help explain inequalities in child and adult morbidity and mortality . Early life factors refer to socially-patterned experiences during prenatal and infant life which may lead to changes in the structure or function of organs and systems, thus affecting health in childhood and later life . They can impact on children’s growth and development , and are associated with a number of chronic conditions . Two notable factors in early life are low birth weight and breastfeeding. Low birth weight, a marker of fetal growth, has been linked to a variety of chronic diseases in later life . Impaired fetal nutrition and growth may affect the development of primary teeth, which commences in utero . A systematic review of the relationship between low birth weight and subsequent development of caries identified 4 papers, one reporting a longitudinal study, and none of which supported an association . Three subsequent longitudinal studies show contradicting results . Whilst the benefits of exclusive breastfeeding on child and maternal health are well known , the picture remains unclear with regard to childhood dental caries. Two recent reviews, based mainly on cross-sectional studies, are equivocal . There is growing evidence that breastfeeding beyond 12 months , and nocturnal feeding , may be associated with greater caries experience in primary teeth.
Toothbrushing, a positive health behaviour, is the primary method of mechanical plaque removal , and delivery of fluoride in toothpaste to protect against caries , with professional guidance recommending brushing with a fluoride toothpaste once the first tooth appears and doing so at least twice per day .
Based on the above background, further research is needed to explore possible pathways between the multiple layers of influences potentially accounting for how determinants of early childhood caries operate and traverse individual, familial, community, and socio-cultural contexts , at national, rather than merely at local level.
The aim of this research was to explore the association between early life factors and dental caries in 5-year-old children in China. We hypothesised that childhood dental caries was positively associated with breastfeeding and negatively associated with low birth weight and late commencement of toothbrushing.
2
Materials and methods
2.1
Data source
This cross-sectional study used data from the 3rd National Oral Health Survey of China (2005), which covered the four World Health Organization (WHO) index ages . All 31 provinces of Mainland China participated in the survey, except for Tibet. Participants were selected using multistage stratified cluster sampling. First, each province was divided into urban and rural areas; urban areas were classified into 3 strata by population size, whereas rural areas were classified into 3 strata by Gross Domestic Product (GDP). Second, one city or county was randomly selected from each stratum. Hence, 3 cities from urban areas and three counties from rural areas were selected from each province. Third, 3 streets or townships were randomly chosen from every city or county, respectively. Fourth, two kindergartens in these streets or townships were chosen randomly from the list of kindergartens provided by the council. Fifth and finally, at each kindergarten, a random sample of 20 children aged 5 years was recruited using the list of enrolees. A target sample of 720 participants was initially set per province, for a total of 21,600 children nationally. A total of 23,365 5-year-olds were clinically examined and questionnaire data were available for 12,692 children as only parents or guardians of half the survey participants were invited to complete questionnaires. Ethical clearance was not required as the present study was based on secondary analysis of anonymised data from an existing national health survey.
2.2
Variables
The number of decayed, missing and filled primary teeth (dmft) was the outcome measure. Dental clinical examinations were carried out with participants seated on a chair, using artificial light, plane mouth mirrors and standard WHO CPI probes. All deciduous teeth were examined and dental caries was diagnosed according to the WHO criteria . Unified training sessions were provided to over 200 survey examiners in Kunming city, Yunnan, before the national survey began. For reliability assessment, duplicate examinations were conducted during the main survey. Five percent of participants were re-examined to calculate inter-examiner reliability, and the Kappa score was 0.94.
Three early life factors, namely birth weight, breastfeeding and age of commencement of toothbrushing, were collected through parental questionnaires. Birth weight, a continuous variable (grams), was recoded into two groups: normal (≥2500 g) and low (<2500 g) weight at birth. Child feeding approaches during the first four months of life were reported across five categories: exclusively breastfed, predominantly breastfed, mixed-fed (50/50), predominantly formula-fed, and exclusively formula-fed. The age of commencement of toothbrushing was reported as being within first 6 months, 1st, 2nd, 3rd, 4th, or 5th year of life and never/seldom. The first two categories were combined into one variable (within first year), thus providing six categories for analysis.
A number of child and family factors were included in the analysis as potential confounders based on previous literature . Child factors were sex, ethnicity, birth order and dental behaviours. Children’s ethnicity was assigned by their parents using a list of officially recognised ethnic groups in China, and responses later grouped as Han or minority ethnic group. Dental behaviours included toothbrushing frequency (less often than daily, once a day, and twice a day or more often), last dental visit for any reason (never, within last year, and more than 1 year ago) and frequency of sugar intake. Parents reported their child’s intake frequency of five common sugary items (biscuits, cake or sweet bread, candy or chocolate, sugared water, soft drinks and fruit juice) on 6-point ordinal scales. Each sugary item was scored as follows: twice or more a day , once a day , 2–6 times a week (2/7 = 0.286), once a week (1/7 = 0.143), 1–3 times per month (1/30 = 0.033), seldom/never (0). The weighted scores were chosen to match the lower frequency of consumption in each response category . A total score, ranging from 0 to 10, was generated by aggregating scores for the 5 sugary items. Based on this, participants were grouped into 3 categories: less often than daily, once or twice a day, and more than twice a day. Family factors included parental education, household income, place of residence (urban or rural), number of children in the family, respondent’s age and their relationship to the child (father, mother and other). Parents were asked to provide an estimate of the annual household income with no pre-set categories. Income data were equivalised using the Luxembourg Income Study scale to account for family size . This involved dividing the total household income by the square root of the number of individuals in the family . After equivalisation, household income in Chinese Yuan was categorised into five groups: very low (0–4,999), low (5000–9999), medium (10,000–14,999), high (14,999–19,999) and very high (20,000+).
2.3
Statistical analysis
Post-stratification weights were used to adjust for differences in the age-by-sex-by-ethnicity-by-province distribution between the sample and the general population in the 30 provinces involved in the study, according to the 5th National Demographic Census in 2000. Analyses also took into account the complex survey design (stratification and clustering) to produce corrected standard errors. All analyses were performed in STATA 14 (Stata Corp LP, College Station, TX).
We first present the composition of the sample according to child and family factors. The impact of missing data was evaluated comparing the characteristics of participants with complete data (study sample) and those excluded due to missing values in relevant variables, using the Chi-square test. The proportion of children that were lighter at birth, exclusively breastfed and had started toothbrushing at one year of age or beforehand, was then examined by child and family factors using the Chi-square test.
The association between each early life factor and dmft index was explored in crude and adjusted models using negative binomial regression, as the outcome measure was a count variable with over-dispersion. Rate ratios (RRs) and 95% confidence intervals (95% CI) were therefore reported as the measure of association. The modelling strategy was first to estimate the crude association of each early life factor with dmft, and then, to adjust these estimates for potential confounders at child (sex, ethnicity, birth order, toothbrushing frequency, last dental visit and sugar intake frequency) and family level (parental education, household income, place of residence, number of children in the family, respondent’s age and relation to the child). The adjusted model also included the other early life factors in order to explore their relative contribution to explain variations in dental caries experience.
Undertaking sensitivity analysis, we evaluated the impact of the respondent’s relationship to the child as this might have an influence on the findings . To that end, we repeated our set of regression models with the subgroup of children for whom the mother completed the parental questionnaire.
2
Materials and methods
2.1
Data source
This cross-sectional study used data from the 3rd National Oral Health Survey of China (2005), which covered the four World Health Organization (WHO) index ages . All 31 provinces of Mainland China participated in the survey, except for Tibet. Participants were selected using multistage stratified cluster sampling. First, each province was divided into urban and rural areas; urban areas were classified into 3 strata by population size, whereas rural areas were classified into 3 strata by Gross Domestic Product (GDP). Second, one city or county was randomly selected from each stratum. Hence, 3 cities from urban areas and three counties from rural areas were selected from each province. Third, 3 streets or townships were randomly chosen from every city or county, respectively. Fourth, two kindergartens in these streets or townships were chosen randomly from the list of kindergartens provided by the council. Fifth and finally, at each kindergarten, a random sample of 20 children aged 5 years was recruited using the list of enrolees. A target sample of 720 participants was initially set per province, for a total of 21,600 children nationally. A total of 23,365 5-year-olds were clinically examined and questionnaire data were available for 12,692 children as only parents or guardians of half the survey participants were invited to complete questionnaires. Ethical clearance was not required as the present study was based on secondary analysis of anonymised data from an existing national health survey.
2.2
Variables
The number of decayed, missing and filled primary teeth (dmft) was the outcome measure. Dental clinical examinations were carried out with participants seated on a chair, using artificial light, plane mouth mirrors and standard WHO CPI probes. All deciduous teeth were examined and dental caries was diagnosed according to the WHO criteria . Unified training sessions were provided to over 200 survey examiners in Kunming city, Yunnan, before the national survey began. For reliability assessment, duplicate examinations were conducted during the main survey. Five percent of participants were re-examined to calculate inter-examiner reliability, and the Kappa score was 0.94.
Three early life factors, namely birth weight, breastfeeding and age of commencement of toothbrushing, were collected through parental questionnaires. Birth weight, a continuous variable (grams), was recoded into two groups: normal (≥2500 g) and low (<2500 g) weight at birth. Child feeding approaches during the first four months of life were reported across five categories: exclusively breastfed, predominantly breastfed, mixed-fed (50/50), predominantly formula-fed, and exclusively formula-fed. The age of commencement of toothbrushing was reported as being within first 6 months, 1st, 2nd, 3rd, 4th, or 5th year of life and never/seldom. The first two categories were combined into one variable (within first year), thus providing six categories for analysis.
A number of child and family factors were included in the analysis as potential confounders based on previous literature . Child factors were sex, ethnicity, birth order and dental behaviours. Children’s ethnicity was assigned by their parents using a list of officially recognised ethnic groups in China, and responses later grouped as Han or minority ethnic group. Dental behaviours included toothbrushing frequency (less often than daily, once a day, and twice a day or more often), last dental visit for any reason (never, within last year, and more than 1 year ago) and frequency of sugar intake. Parents reported their child’s intake frequency of five common sugary items (biscuits, cake or sweet bread, candy or chocolate, sugared water, soft drinks and fruit juice) on 6-point ordinal scales. Each sugary item was scored as follows: twice or more a day , once a day , 2–6 times a week (2/7 = 0.286), once a week (1/7 = 0.143), 1–3 times per month (1/30 = 0.033), seldom/never (0). The weighted scores were chosen to match the lower frequency of consumption in each response category . A total score, ranging from 0 to 10, was generated by aggregating scores for the 5 sugary items. Based on this, participants were grouped into 3 categories: less often than daily, once or twice a day, and more than twice a day. Family factors included parental education, household income, place of residence (urban or rural), number of children in the family, respondent’s age and their relationship to the child (father, mother and other). Parents were asked to provide an estimate of the annual household income with no pre-set categories. Income data were equivalised using the Luxembourg Income Study scale to account for family size . This involved dividing the total household income by the square root of the number of individuals in the family . After equivalisation, household income in Chinese Yuan was categorised into five groups: very low (0–4,999), low (5000–9999), medium (10,000–14,999), high (14,999–19,999) and very high (20,000+).
2.3
Statistical analysis
Post-stratification weights were used to adjust for differences in the age-by-sex-by-ethnicity-by-province distribution between the sample and the general population in the 30 provinces involved in the study, according to the 5th National Demographic Census in 2000. Analyses also took into account the complex survey design (stratification and clustering) to produce corrected standard errors. All analyses were performed in STATA 14 (Stata Corp LP, College Station, TX).
We first present the composition of the sample according to child and family factors. The impact of missing data was evaluated comparing the characteristics of participants with complete data (study sample) and those excluded due to missing values in relevant variables, using the Chi-square test. The proportion of children that were lighter at birth, exclusively breastfed and had started toothbrushing at one year of age or beforehand, was then examined by child and family factors using the Chi-square test.
The association between each early life factor and dmft index was explored in crude and adjusted models using negative binomial regression, as the outcome measure was a count variable with over-dispersion. Rate ratios (RRs) and 95% confidence intervals (95% CI) were therefore reported as the measure of association. The modelling strategy was first to estimate the crude association of each early life factor with dmft, and then, to adjust these estimates for potential confounders at child (sex, ethnicity, birth order, toothbrushing frequency, last dental visit and sugar intake frequency) and family level (parental education, household income, place of residence, number of children in the family, respondent’s age and relation to the child). The adjusted model also included the other early life factors in order to explore their relative contribution to explain variations in dental caries experience.
Undertaking sensitivity analysis, we evaluated the impact of the respondent’s relationship to the child as this might have an influence on the findings . To that end, we repeated our set of regression models with the subgroup of children for whom the mother completed the parental questionnaire.
3
Results
A total of 9722 children with complete data in all relevant variables were included in this analysis (representing 77% of those whose parents were invited to complete the questionnaire). The characteristics of the study sample are presented in Table 1 . Children in urban, wealthier and single-child families were more likely to be included in the study sample. No other differences, even for the three early life factors and dental caries experience, were found between the study sample and those excluded because of missing data. The mean dmft was 3.59 (95% CI: 3.37–3.81) and 66.9% (64.5–69.2%) of children had experienced dental caries. The average number of decayed teeth was 3.49 (3.26–3.71), representing 97.2% of the dmft index.
| Factors | Categories | n a | % |
|---|---|---|---|
| Sex | Boy | 4961 | 54.3% |
| Girl | 4761 | 45.7% | |
| Ethnicity | Han | 8823 | 90.4% |
| Minority ethnic groups | 899 | 9.6% | |
| Place of residence | Rural | 4546 | 67.2% |
| Urban | 5176 | 32.8% | |
| Birth order | First child | 7958 | 77.9% |
| Second child | 1585 | 19.4% | |
| Third or later | 179 | 2.7% | |
| Number of children in the family | 1 child | 7189 | 67.7% |
| 2 children | 2232 | 27.5% | |
| 3 or more | 301 | 4.8% | |
| Parental education | Up to primary school | 1437 | 21.2% |
| Junior middle school | 3485 | 41.4% | |
| Senior middle school | 2479 | 21.4% | |
| Higher education | 2321 | 16.0% | |
| Household income | Very low | 2652 | 38.3% |
| Low | 2482 | 25.7% | |
| Medium | 1918 | 15.4% | |
| High | 979 | 7.4% | |
| Very high | 1691 | 13.1% | |
| Respondent’s relationship to child | Father | 2795 | 28.8% |
| Mother | 6734 | 68.8% | |
| Other relatives | 193 | 2.4% | |
| Respondent’s age | <30 years | 2195 | 25.2% |
| 30–39 years | 6937 | 68.6% | |
| 40+ years | 590 | 6.2% | |
| Toothbrushing frequency | Less often than daily | 3450 | 41.1% |
| Once a day | 4530 | 44.7% | |
| Twice a day or more | 1742 | 14.2% | |
| Last dental visit | Never | 7451 | 79.3% |
| Within last year | 1489 | 13.4% | |
| Over 1 year ago | 782 | 7.3% | |
| Sugar intake frequency | Less often than daily | 5805 | 59.3% |
| Once or twice a day | 2467 | 25.0% | |
| More than twice a day | 1450 | 15.7% |
Overall, 3.7% (CI: 3.0–4.5%) of children were lighter at birth, 62.2% (CI: 61.0–63.3%) were exclusively breastfed and only 2% (CI: 1.5–2.6%) were toothbrushing at the age of one year. Low birth weight was significantly more common in girls, later-born children within the family, larger families and less educated parents ( Table 2 ). Exclusively breastfeeding was more common in later-born children, those living in rural areas, and amongst less educated, poorer and younger parents. Early toothbrushing (aged 1 year or below) was more common in first-born children, smaller and urban families, and those children with older, more educated and wealthier parents Toothbrushing was not a social norm with only 34.7% of children reported as having started toothbrushing by 3 years of age.
| Factors | Categories | Low birth weight | Exclusively breastfed | Early toothbrushing |
|---|---|---|---|---|
| % (95% CI) | % (95% CI) | % (95% CI) | ||
| Sex | Boy | 3.2% (2.6–3.9%) | 61.2% (59.4–63.0%) | 1.8% (1.2–2.4%) |
| Girl | 4.2% (3.2–5.2%) | 63.3% (61.9–64.7%) | 2.2% (1.5–3.0%) | |
| P value a | 0.012 | 0.072 | 0.229 | |
| Ethnicity | Han | 3.6% (2.9–4.3%) | 62.7% (61.3–64.1%) | 2.1% (1.5–2.7%) |
| Minority ethnic groups | 4.3% (3.0–5.6%) | 56.9% (50.2–63.7%) | 1.0% (0.0–2.1%) | |
| P valuea | 0.148 | 0.107 | 0.186 | |
| Place of residence | Rural | 4.0% (3.0–5.0%) | 66.4% (64.9–67.8%) | 0.7% (0.1–1.2%) |
| Urban | 3.0% (2.1–3.9%) | 53.6% (52.0–55.2%) | 4.7% (3.4–6.0%) | |
| P value a | 0.128 | <0.001 | <0.001 | |
| Birth order | First child | 3.3% (2.8–3.9%) | 61.2% (59.6–62.9%) | 2.4% (1.8–3.0%) |
| Second child | 4.9% (2.7–7.2%) | 64.8% (62.4–67.2%) | 0.5% (0.1–0.9%) | |
| Third or latter | 4.7% (2.4–7.0%) | 71.1% (59.8–82.4%) | 1.1% (0.0–3.3%) | |
| P value for trend a | 0.008 | 0.022 | 0.009 | |
| Number of children in the family | 1 child | 3.1% (2.6–3.6%) | 61.4% (59.5–63.2%) | 2.6% (1.9–3.3%) |
| 2 children | 4.8% (2.7–6.9%) | 63.4% (61.6–65.3%) | 0.6% (0.2–1.0%) | |
| 3 or more | 5.2% (2.8–7.6%) | 66.7% (58.6–74.7%) | 0.8% (0.0–2.1%) | |
| P value for trend a | 0.006 | 0.154 | 0.001 | |
| Parental education | Up to primary school | 6.1% (3.8–8.4%) | 70.1% (67.4–72.8%) | 0.2% (0.0–0.6%) |
| Junior middle school | 3.1% (2.4–3.9%) | 65.7% (63.4–68.1%) | 0.9% (0.4–1.4%) | |
| Senior middle school | 2.6% (1.5–3.8%) | 55.8% (52.4–59.3%) | 2.6% (1.8–3.3%) | |
| Higher education | 3.3% (2.1–4.6%) | 51.0% (48.2–53.8%) | 6.4% (4.5–8.3%) | |
| P value for trend a | 0.011 | <0.001 | <0.001 | |
| Household income | Very low | 3.9% (3.0–4.8%) | 67.6% (64.8–70.4%) | 0.4% (0.2–0.5%) |
| Low | 4.3% (2.9–5.7%) | 62.7% (59.9–65.5%) | 1.2% (0.5–1.9%) | |
| Medium | 3.0% (1.9–4.2%) | 59.4% (57.4–61.3%) | 2.4% (1.6–3.3%) | |
| High | 3.2% (1.0–5.4%) | 59.8% (53.0–66.6%) | 3.7% (2.1–5.2%) | |
| Very high | 2.9% (1.4–4.5%) | 49.9% (47.2–52.7%) | 6.9% (4.6–9.2%) | |
| P value for trend a | 0.208 | <0.001 | <0.001 | |
| Respondent’s relationship to child | Father | 4.0% (2.9–5.1%) | 60.9% (56.9–64.8%) | 1.7% (1.3–2.1%) |
| Mother | 3.5% (2.8–4.3%) | 62.3% (60.7–64.0%) | 2.2% (1.4–2.9%) | |
| Other relatives | 3.8% (1.1–6.5%) | 73.6% (67.1–80.1%) | 0.3% (0.0–0.9%) | |
| P value a | 0.326 | 0.206 | 0.604 | |
| Respondent’s age | <30 years | 3.3% (2.5–4.1%) | 66.1% (63.8–68.4%) | 1.3% (0.5–2.1%) |
| 30–39 years | 3.5% (2.5–4.5%) | 60.9% (59.4–62.3%) | 2.3% (1.6–2.9%) | |
| 40+ years | 6.5% (3.9–9.1%) | 60.8% (56.4–65.3%) | 1.9% (0.5–3.2%) | |
| P value for trend a | 0.063 | 0.001 | 0.013 |
Stay updated, free dental videos. Join our Telegram channel
VIDEdental - Online dental courses
