Although the American Academy of Pediatrics recommends that fruit juice intake among 1- to 6-year-olds be limited to 4 to 6 ounces per day,³ more than 10 percent of preschoolers in the United States consume at least 12 fluid oz of fruit juice per day.⁴ Overall carbohydrate intake has increased from 46.3 percent of total energy intake in 1965 to 54.2 percent of total energy intake in 1996.² On a positive note, breast-feeding rates have risen since the middle 1970s.⁵ From the 1970s through the 1990s, dental decay among U.S. children and adolescents has seen a marked decrease in prevalence, due largely to the effectiveness of fluoridating drinking water and dentifrices.⁶

The role of refined carbohydrates in the caries process is well-accepted in the dental community.^1,7–9 The most important dietary etiological factor in the promotion of dental caries is sugar consumption, particularly sucrose. Aggregate observations from a review of studies indicate that when sugar intake exceeds 15 to 20 kilograms per person per year, such intake is directly associated with increasing caries prevalence.¹⁰ Earlier research, such as the landmark Vipeholm study of the 1950s,⁷ was conducted when caries prevalence was much higher. Conclusions generally have supported the idea that sugar consumption increases caries prevalence, and the risk of caries increases when sugar is consumed between meals. Indeed, current dental health education for the control of dental caries is based on this earlier research and supports the promotion of dietary restriction of sugars to prevent caries.¹¹

Sixty percent of children have one or more decayed or restored teeth by age 5 years.

Although the United States is in an era of relatively low caries prevalence, 60 percent of children still have one or more decayed or restored teeth by age 5 years, with 20 percent of children experiencing approximately 80 percent of the total dental caries burden.¹² This raises the following question: what are the effects of our changing eating habits on caries in primary teeth in an environment of reduced caries prevalence in the United States?

Some meals, such as breakfast, often are skipped altogether. Teenagers who miss breakfast are more likely to snack during the day¹³ and snacks have the highest sugar content of any type of meal (that is, breakfast, lunch, dinner or snacks).¹⁴ Consumption of whole grains and dairy products has been shown to decrease an individual’s appetite,^15–18 while diets high in sugar cause people to feel hungry and seek more calories.^19,20 Missing meals could have a direct influence on consumption of refined carbohydrates, and skipping meals such as breakfast could lead to increased sugar consumption.

The last national report on diet and dental health in the United States concluded, "There was a direct, strong, and statistically significant relationship between DMF [decayed-missing-filled] experience and the frequency of intake of sugary snacks between meals."²¹ However, this report, which was released two decades ago, was limited to school-aged children and permanent teeth.

Public health advocates widely believe that poor infant feeding practices, particularly feeding with juice in a bottle at bedtime, are associated with the development of caries in primary teeth. Douglass and colleagues²² have suggested that dietary factors other than bottle feeding may promote caries in primary teeth among young children. Al-Dashti and colleagues²³ reported that breast-fed children are more likely to be caries-free compared with children who are bottle-fed. However, researchers generally believe that breast-feeding is associated with dental caries only if it is prolonged.²³

Although human milk is potentially more cariogenic than cow’s milk because of its higher lactose and lower calcium and phosphate levels,²⁴ breast-feeding might mitigate consumption of refined carbohydrates by promoting reduced consumption of juices and drinks high in sugar. Consequently, if breast-feeding prevalence increases, pediatric caries experience should decline.

Our study poses two main questions. First, it attempts to determine if the regular consumption of breakfast or five servings per day of fruits or vegetables is associated with less caries in young children, based on the hypothesis that children who eat five fruits or vegetables, breakfast or both on a regular basis consume fewer snacks and, therefore, are exposed to fewer carbohydrates. Second, this study investigates the relationship between a history of breast-feeding and caries in primary teeth. We hypothesize that breast-feeding could delay an infant’s exposure to high-sugar-content fruit juices, thus promoting a lower prevalence of pediatric caries.

	METHODS

TOP ABSTRACT METHODS RESULTS DISCUSSION CONCLUSION REFERENCES

 
Study population. We used data from 4,236 children who participated in the third National Health and Nutrition Examination Survey, or NHANES III, from 1988 to 1994.²⁵ The NHANES III was conducted by the National Center for Health Statistics, or NCHS, of the Centers for Disease Control and Prevention, or CDC. It used a complex, stratified, multistage probability design capable of producing a nationally representative sample for the noninstitutionalized civilian population of the United States. Details of the sample design and methods used to obtain informed consent from study participants have been described elsewhere.²⁶

The NHANES III oversampled people who were either younger than 6 years or older than 60 years, Mexican-Americans and non-Hispanic blacks. Information regarding sociodemographic characteristics, growth and social development, and oral health was collected through the household interview questionnaire and physical examination. Dietary information was obtained via adult proxy from one 24-hour dietary recall administered in the mobile examination center, or MEC, on the day of the examination, and has been described in detail elsewhere.²⁶ We produced additional information about healthful eating using the Healthy Eating Index, or HEI, Data File, which was created by CDC and NCHS using NHANES III dietary intake data and the U.S. Department of Agriculture guidelines.²⁷

For our study, we obtained data from questionnaire information (the Youth File), dietary information (the HEI File) and a standardized oral health examination (Exam File). Three trained dentists in NHANES III conducted the majority of the dental examinations, all of which were performed in MECs. Dental examiners were calibrated periodically by the survey’s expert dental examiner, and inter-rater reliability between the survey examiners and the reference examiner was considered to be very good ( statistics ranged from 0.96 to 1.00).²⁸ Detailed descriptions about the NHANES III oral health component protocol, quality control and measurement issues have been described elsewhere.^26,29

Caries experience was based on the number of decayed or filled primary dental surfaces as derived from the coronal caries examination.

We identified a total of 13,944 participants who had a record in the NHANES III Youth File as potentially eligible for this analysis. We excluded 824 people who had no recorded examination value for age. We then excluded 8,567 people who were not within the targeted age range (2 through 5 years) for this analysis. From the remaining group, we excluded 317 people who did not receive a completed dental examination. This yielded an analytical sample of 4,236 participants for our study.

Outcome variables. Caries experience and existing untreated decay in primary teeth were the two dependent variables used in our analysis. Coronal dental assessments in NHANES III were made according to visual/tactile examinations, without radiographs at the surface level for both primary and permanent teeth (excluding third molars). Caries experience was based on the number of decayed or filled primary dental surfaces, or dfs, as derived from the coronal caries examination. For our analysis, we defined caries experience as at least one primary tooth surface with a dental filling or untreated decay, which was dichotomized as "yes" or "no." We derived the untreated decay status variable from the number of primary surfaces with caries, or ds, and categorized it into four levels: no untreated decay, one or two surfaces with untreated decay, three through five surfaces with untreated decay and six or more surfaces with untreated decay.

Covariate selection. For these analyses, we included sociodemographic indicators that have been reported to be associated with oral health. Race/ethnicity was categorized as Mexican-American, non-Hispanic black, non-Hispanic white and other. Children who were identified as "other" were included in the total population estimates, but not in the regression analyses. We categorized educational attainment for the identified adult responsible for the child as "some or no high school experience," "completed high school" and "at least some college experience." Poverty status was dichotomized as either equal to or below 200 percent of the federal poverty level or greater than 200 percent, and was calculated by dividing total family income by the adjusted federal poverty income threshold. We categorized dental history as either having had a dental visit within the previous 12 months or not. Sex also was included in the analyses.

We assessed past and current healthful eating practices using the history of breast-feeding, breakfast eating frequency, and eating five servings of fruits and vegetables a day. A child’s history of breast-feeding or receiving breast milk was categorized as "positive" or "negative." We assigned a "positive" score if the respondent reported that the child had ever been breast-fed or received breast milk, regardless of periodicity. A child’s frequency of eating breakfast was dichotomized as either "every day" or "not every day." The "not every day" value was obtained by aggregating the following possible responses from the reporting adult: never, some days, rarely or weekends only. Using fruit and vegetable serving information from the HEI file, we categorized a "5-a-Day" variable as either currently eating fewer than five servings of fruits and vegetables a day or eating five or more servings a day.

Overweight and short stature among children were defined on the basis of the sex-specific 2000 CDC growth charts for the United States.³⁰ Overweight was defined as the 95th percentile or higher of body mass index, or BMI, in kilograms per square meter for age, and at risk of being overweight was defined as the 85th percentile or higher but less than the 95th percentile of BMI for age. The Committee on Clinical Guidelines for Overweight in Adolescent Preventive Services³¹ recommended use of the 95th percentile of BMI to classify adolescents as being overweight. Short stature was defined as less than the fifth percentile of stature for age. We used a database and statistics program (Epi Info 2000) to make the calculations using the 2000 CDC growth charts.³²

Data analysis. We performed all statistical analyses using SUDAAN, a software package that is specifically designed to accommodate complex sample surveys.³³ Sample weights were used to account for the unequal probability of selection and nonresponse of the study participants to produce prevalence estimates, relative odds and related standard errors. We conducted bivariate analyses of the sociodemographic, healthful eating and anthropometric characteristics with dental caries. We used pairwise t-tests to assess for significant differences between groups.

Caries experience increased with age and was greater for Mexican-American children than it was for non-Hispanic black and non-Hispanic white children.

Using the LOGISTIC and MULTILOG functions within SUDAAN, respectively, we used standard and cumulative logistic regression models to compute adjusted odds ratios, or OR, and 95 percent confidence intervals, or CI. The cumulative logistic function uses an ordinal categorical outcome variable of two or more categories, compared with the standard logistic function, which uses a dichotomous outcome variable. The MULTILOG function produces estimates using a proportional odds model with a cumulative link. This statistical methodology has been reported to be the most frequently used model for ordinal logistic regression applications.³⁴ These estimates indicate the odds of a categorical increase in the number of surfaces affected by untreated decay for the indicators included in the model.

We used nonautomated stepwise regression modeling to assess the relationships between the covariates. Parsimonious models were determined by covariate exclusion, with criteria for inclusion set for a Satterthwaite-adjusted F statistic of P < .05. Potential two-way interactions were explored throughout the modeling process. We considered P < .05 to be statistically significant.

	RESULTS

TOP ABSTRACT METHODS RESULTS DISCUSSION CONCLUSION REFERENCES

 
Table 1 shows the sociodemographic, nutrition, body measure and oral health characteristics.²⁵ Table 2 shows the prevalence of dental caries experience (dfs) (a measure of the history of carious activity) and untreated decay (ds) according to characteristics of interest.²⁵ Caries experience increased with age (to 40.2 percent for 5-year-olds) and was greater for Mexican-American children (39.7 percent) than it was for non-Hispanic black and non-Hispanic white children (28.6 percent and 18.1 percent, respectively). Children with the greatest number of tooth surfaces affected by untreated decay ( six surfaces) were more likely to be Mexican-American (15.7 percent), to live in households with an adult who had not completed high school (12.6 percent) and to be at or below 200 percent of the federal poverty line (11.0 percent).

The unadjusted regression models also showed that more untreated decayed surfaces (that is, ds) in children aged 2 through 5 years were significantly associated (P < .05) with lower educational achievement, poverty, being Mexican-American or non-Hispanic black, not receiving breast milk, eating fewer than five servings of fruits and vegetables a day, and not eating breakfast every day (data not shown). During the modeling process, we also found an interaction between poverty and educational attainment.

Because of the interactions between poverty and education and race/ethnicity, we performed additional regression analyses that were stratified by poverty status (Table 3²⁵). For children living at or below 200 percent of the federal poverty level, caries experience was only significantly associated (P < .05) with being Mexican-American or not having had a dental visit within the past 12 months. For children not living in poverty (> 200 percent of the federal poverty line), race/ethnicity, not eating breakfast daily, not eating five fruits and vegetables a day, and not having had a dental visit within the past 12 months were significantly associated with caries experience (P < .05). Among this group of children, the likelihood of experiencing caries was greater for those who did not eat breakfast daily (OR = 3.77; 95 percent CI, 1.80 to 7.89) and who did not eat five servings of fruits and vegetables a day (OR = 3.21; 95 percent CI, 1.74 to 5.95).

Not eating breakfast every day was found to be associated with overall caries experience and untreated decay in the primary dentition in children aged 2 through 5 years.

For children living at or below 200 percent of the federal poverty line, being Mexican-American or not having had a dental visit within the past 12 months continued to be strong indicators of having an increasing amount of untreated decay (OR = 2.12; 95 percent CI, 1.40 to 3.22 and OR = 1.46; 95 percent CI, 1.06 to 2.02, respectively) (Table 3).

Among children aged 2 through 5 years not living in poverty, the odds of having increasing levels of untreated decay associated with not eating breakfast every day (OR = 2.76; 95 percent CI, 1.21 to 6.26) were similar to the independent effects of race/ethnicity (OR = 2.63 and 2.64 for being Mexican-American and non-Hispanic black, respectively). Furthermore, not eating five fruits and vegetables a day was significantly associated with having more untreated decay, but not having had a dental visit within the past 12 months was not significantly associated with the amount of untreated decay (P > .05).

	DISCUSSION

TOP ABSTRACT METHODS RESULTS DISCUSSION CONCLUSION REFERENCES

 
Poor eating practices. In this analysis of more than 4,000 preschoolers, we found that poor eating practices were associated with caries in primary teeth among those not living in poverty. Specifically, not eating breakfast every day was found to be associated with overall caries experience and untreated decay in the primary dentition in children aged 2 through 5 years. The finding that caries was associated with unhealthful eating practices (that is, not eating breakfast or five fruits and vegetables per day) and that these relationships were not homogenous across the poverty strata (that is, the effect is not the same at different levels of household income) are important to consider in health promotion and policy deliberations. Moreover, these findings reinforce the notion that good dietary habits promote oral health; thus, reinforcing good eating practices such as daily breakfast consumption could be an important adjunct to fluoridation to further reduce caries experience in children.

Daily breakfast consumption. We used daily breakfast consumption as an indicator of consistent good nutrition. This is based on several studies that have associated a lack of breakfast consumption with higher rates of caloric intake, poor nutrient intake and obesity.^35,36 Although breakfast consumption among many young children is limited to presweetened cereals, many studies have indicated that these cereals are nutritionally beneficial.³⁷ However, many people view presweetened packaged cereals consumed by children as being potentially cariogenic.

Gibson³⁸ examined this perception using data from the 1995 United Kingdom National Diet and Nutrition Survey of preschool-aged children, and found caries experience to be unrelated to eating presweetened cereals. Furthermore, Gibson reported that children who consumed breakfast cereals were more likely to reduce consumption of sugary snacks or soda compared with children who skipped breakfast. Dairy products also have been shown to have some anticariogenic effects.^39,40 Thus, it is possible that the relationship between sugars in breakfast cereals and caries could be mitigated when children consume presweetened cereals with milk.

Breast-feeding. We used any history of breast-feeding, regardless of frequency, as another indicator of good nutritional practices. We proposed that breast-feeding would likely delay or reduce the consumption of juices and drinks that contain high levels of sugar. Dini and colleagues⁴¹ reported that children aged 3 through 4 years who were never breast-fed were likely to have more caries compared with children who were breast-fed, regardless of duration, within the first 24 months of life.

Currently, differing recommendations exist for promoting prolonged breast-feeding. The American Academy of Pediatric Dentistry has endorsed the recommendation by the American Academy of Pediatrics to promote breast-feeding during the first 12 months of childhood; however, the World Health Organization advocates unrestricted breast-feeding beyond age 2 years.^42–44

Although Dini and colleagues⁴¹ reported that children who were breast-fed beyond 24 months of age were more likely to have caries compared with children who were breast-fed only up to 24 months of age, Weerheijm and colleagues⁴⁵ reported that prolonged breast-feeding "does not lead to a higher caries prevalence." Roberts and colleagues⁴⁶ reported that infant feeding methods are not related to the prevalence of caries in primary teeth, but when caries is present, the magnitude of caries is related to feeding methods. Furthermore, a systematic review by Valaitis and colleagues,⁴⁷ which examined the relationship between early childhood caries and breast-feeding, yielded inconclusive findings. Our findings indicate that there is no relationship between caries and a history of ever breast-feeding.

BMI. For our study, we used anthropometrics measures, such as BMI, to control for potential confounding effects. Published research examining the relationships between body measures and tooth decay has been sparse and inconclusive, with most studies failing to control for the potential confounding effects of diet on tooth decay. Studies have reported an association between BMI and decay in the permanent dentition,^48,49 but no findings of association between BMI and decay in the primary dentition have been reported.⁵⁰ Our findings show that after adjusting for the more important known indicators of caries in primary teeth (such as race/ethnicity and dental visit frequency), healthful eating practices are more significant in the prevalence of caries than are BMI measures.

The data from this study produced mixed findings with regard to previous analyses of the impact of socioeconomic status on oral health status.

Changes in dietary habits also have been linked to increased levels of obesity in the United States and other nations. Fragmented meals and meals away from home both are contributors to overconsumption of foods.⁵¹ Decreased stature and failure to thrive have been reported with excessive fruit juice intake in some children, while in others, excessive comsumption of juice drinks has been found to be related to excessive caloric intake and obesity.⁴ Skinner and Carruth⁵² reported that consumption of 100 percent juice was not associated with either obesity or short stature, but as juice consumption decreases, soda and juice drink intake increases. Consumption of high-carbohydrate liquids may be a risk factor for excessive caloric intake,⁵³ and may have harmful added effects for people with poor oral hygiene.⁵⁴

Scandinavian studies have reported that preschool-aged children were more likely to experience caries in primary teeth when they had poor oral hygiene independent of poor dietary habits,^55,56 and that poor dietary practices at 12 months of age are related to caries incidence at 3 years of age.⁵⁷ Other studies have suggested that there is no relationship between nutrient intake and caries in the primary dentition.

Marques and Messer⁵⁸ reported that sugar consumption was not associated with primary caries and that increasing age and history of fluoride use and dental care visits were better indicators of primary caries experience. In addition, McMahon and colleagues⁵⁹ reported that there was no relationship between carbohydrate intake and caries in the primary dentition among young children who lived in areas with optimal community water fluoridation. However, their study did report that young children being treated with medication mixed with sweetened syrups were more likely to develop caries, and that socioeconomic status was an independent indicator of primary caries prevalence.

The data from our study produced mixed findings with regard to previous analyses of the impact of socioeconomic status on oral health status. It is well-known that minority children or children identified within lower socioeconomic groups, assessed by either the level of the primary caretaker’s education or household income, are more likely to experience caries compared with nonminority children or children in higher socioeconomic groups.^57,60–63 However, our findings suggest that the effect of the primary caretaker’s education on a preschooler’s caries experience is not significant when the effects of healthful eating practices, such as eating breakfast or five fruits and vegetables every day, are accounted for.

Our findings also suggest that poverty status is a significant effect modifier in the relationship between healthful eating practices and primary tooth decay. For children who are at or below the 200 percent federal poverty level, being Mexican-American or not having had a dental visit within the past 12 months were the only independent factors that were significantly related to caries experience or more untreated tooth decay in the primary dentition.

Conversely, for children above the 200 percent federal poverty line, being non-Hispanic black or Mexican-American, not having had a dental visit within the past 12 months, and not eating breakfast or five fruits and vegetables per day all were significantly related to caries experience. The results were similar for having more untreated decay, except that not having had a dental visit within the past 12 months was not a significant indicator.

Caries prevalence. In the United States, caries prevalence in primary teeth among 5- to 9-year-olds has gradually declined from a mean dfs of 5.3 in the early 1980s⁶⁴ to 3.9 in the mid-1990s (B. Dye, unpublished data, 2003). Although the prevalence of caries in primary teeth has been decreasing during the past two decades in older children, Brown and colleagues⁶ reported that the prevalence has remained unchanged among 2- to 5-year-olds since the 1970s. Recent studies have reported that caries prevalence and unmet dental needs continue to be an important public health issue, especially among high-caries-risk preschool-aged children.^65,66

At first glance, our findings that unhealthful eating practices, such as not having breakfast or five fruits and vegetables every day, are significant risk factors for caries in primary teeth only among nonpoor preschool-aged children may seem to be counterintuitive. However, many high-caries-risk children attend Head Start and related programs, where they receive free or subsidized meals, particularly breakfast. Poverty may be the more important cofactor in indicating caries risk, but healthful eating practices are an important factor in the overall, complex process that leads to caries experience in young children. Our findings support the notion that even if the effects of poverty could be mitigated, healthful eating practices among preschoolers would contribute to a further reduction in caries.

The results of our study suggest that dietary habits of children are still a component of the caries process in the primary dentition, and there appears to be a "protective effect" from engaging in healthful eating practices in early childhood, especially among the nonpoor. Unfortunately, the frequency of consumption of breakfast among children has decreased by more than 10 percent during the last 30 years³⁶ and the consumption of dairy products has decreased by 30 percent, while consumption of sugar-containing soft drinks has doubled during this period.² As dietary habits continue to progress along a similar trend within the United States, unhealthful eating practices may become a significant barrier to efforts to help children reach a caries-free status in adulthood.

Study limitations. Limitations of our study are related to the use of a cross-sectional design to examine indicators of untreated tooth decay and caries experience. We were unable to accurately determine weekly caloric intake or dietary composition based on self-reported frequency of breakfast consumption. In addition, we were unable to ascertain the level of parental bias in responding to self-reported nutritional questions. Although survey protocols for caries assessment in NHANES III were more conservative compared with clinical standards, and radiographs were not obtained, the direction of the bias produced from underreporting disease during the dental examination may have underestimated the magnitude of the association between caries and healthful eating practices. The strength of our study is its use of a large, nationally representative sample of preschool-aged children to explore and control for multiple risk factors.

	CONCLUSION

TOP ABSTRACT METHODS RESULTS DISCUSSION CONCLUSION REFERENCES

 
To our knowledge, this is the first epidemiologic study to examine the relationship between dietary habits and tooth decay among preschool-aged children through the use of a nationally representative sample. Overall, our findings suggest that young, nonpoor children with poor dietary habits are more likely to experience caries in primary teeth compared with poor children. Although we did not find an association between caries and breast-feeding, relationships with other surrogates for healthful eating practices during early childhood—such as eating five fruits and vegetables per day and, in particular, eating breakfast every day—are important findings.

The data suggest that improved access to well-balanced meals and to breakfast programs for children may be an important oral health promotion activity. Given the results of our study, future research opportunities include more detailed cross-sectional and longitudinal studies. More in-depth dietary information with longer periods of data collection could assist in defining nutritional risk factors for caries development. The impact of increasing carbohydrate exposure, including consumption of sodas and juice drinks, on the rate of dental caries also should be further investigated.

View larger version (120K): [in this window] [in a new window]   Dr. Dye is the dental epidemiology officer, Analysis Branch, Division of Health and Nutrition Examination Survey, National Center for Health Statistics, Centers for Disease Control and Prevention, 3311 Toledo Road, Room 4416, Hyattsville, Md. 20782, e-mail "bfd1{at}cdc.gov". Address reprint requests to Dr. Dye.

View larger version (110K): [in this window] [in a new window]   At the time this study was conducted, Dr. Shenkin was a resident in pediatric dentistry, University of Iowa, Iowa City. He currently is an assistant clinical professor of Health Policy, Health Services Research and Pediatric Dentistry at the Goldman School of Dental Medicine, Boston University. He also is in full-time private practice in Bangor, Maine.

View larger version (142K): [in this window] [in a new window]   Dr. Ogden is a nutritional epidemiologist, Analysis Branch, Division of Health and Nutrition Examination Survey, National Center for Health Statistics, Centers for Disease Control and Prevention, Hyattsville, Md.

View larger version (111K): [in this window] [in a new window]   Dr. Levy is a professor, Department of Preventive and Community Dentistry, College of Dentistry, and Department of Epidemiology, School of Public Health, University of Iowa, Iowa City.

View larger version (121K): [in this window] [in a new window]   Dr. Kanellis is an associate professor and head, Department of Pediatric Dentistry, College of Dentistry, University of Iowa, Iowa City.