| Sign In to gain access to subscriptions and/or personal tools. |
Carbonated Soft Drinks and Dental Caries in the Primary Dentition
1 Department of Cariology, Restorative Sciences, and Endodontics, School of Dentistry, University of Michigan, 1101 N. University, Ann Arbor, MI 48109-1078, USA; and Correspondence: * corresponding author, woosung{at}umich.edu
We analyzed fluid intake data among children aged 2–10 years from a 24-hour dietary recall interview in the NHANES III (1988–94) to investigate the effect of high consumption of carbonated soft drinks on caries in the primary dentition. We used cluster analysis to determine fluid consumption patterns. Four distinct fluid consumption patterns were identified: high carbonated soft drinks, high juice, high milk, and high water. About 13% of children had a high carbonated soft drink consumption pattern; they also had a significantly higher dental caries experience in the primary dentition than did children with other fluid consumption patterns. A fluid intake pattern comprised mainly of milk, water, or juice was less likely to be associated with dental caries. Findings of this study suggest that high consumption of carbonated soft drinks by young children is a risk indicator for dental caries in the primary dentition and should be discouraged.
Key Words: fluid consumption pattern • carbonated • soft drinks • NHANES III • cluster analysis • dental caries • primary dentition
Patterns of fluid consumption by children are of public health interest, since they may be related to both oral health and general health conditions such as obesity and diabetes (Marshall, 2003). It has been reported that contemporary fluid consumption patterns of children are now more diverse than in past years, since carbonated soft drinks and fruit juices have replaced much of the previous consumption of water and milk among children (Harnack et al., 1999; Heller et al., 1999). The implications of these changes in fluid consumption have not been well-studied, especially the association between carbonated soft drink consumption and dental caries. While the association between the consumption of sugars (all mono-and disaccharides) and dental caries experience in permanent teeth has been well-documented (Ismail et al., 1984; Rugg-Gunn, 1996; Jones et al., 1999), the association between carbonated soft drink consumption and dental caries in the primary dentition is less clear. Some studies have reported significant associations (Grindefjord et al., 1995; Moynihan and Holt, 1996; Levy et al., 2003), while others did not (Heller et al., 2001; Sayegh et al., 2002). These inconsistencies may arise because fluid consumption patterns are complex and methodologically challenging to ascertain. For example, like other diet and nutrition data, fluid consumption sources are highly correlated, which precludes testing the association between fluid intake from a single source and a disease outcome without considering the effect of fluid intake from other sources (Wirfalt and Jeffery, 1997). To develop a more simple measure of complex patterns, some investigators have used cluster analysis (Aldenderfer and Blashfield, 1984) to aggregate dietary patterns and then associate the patterns with disease outcomes (Akin et al., 1986; Tucker et al., 1992; Wirfalt and Jeffery, 1997). The aims of this analysis were: (1) to identify, by cluster analysis, distinct fluid consumption patterns, with a focus on high consumption of carbonated soft drinks, among children aged 2 to 10 yrs; and (2) to assess the association between high consumption of carbonated soft drinks and dental caries in the primary dentition among children in the United States.
The NHANES III Survey Data for Fluid Consumption This analysis used fluid intake data from a 24-hour dietary recall interview in the Third National Health and Nutrition Examination Survey (NHANES III, 1988–94) in the United States (National Center for Health Statistics, 1994). Detailed descriptions of the data preparation and variables used in this analysis can be found elsewhere (Sohn et al., 2001). We defined major fluid sources as milk (and milk drinks), juice (fruit and vegetable juices and other non-carbonated drinks), carbonated soft drinks (sugared and non-sugared), plain water (tap and spring water), and coffee and tea. Fluid intake from sources other than these major sources—such as soup, homemade beverages, and water used for cooking—was all grouped into ‘other foods and beverages’. This definition is consistent with previous studies from other researchers (Ershow and Cantor, 1989; Heller et al., 1999). In this analysis, sugared and non-sugared carbonated soft drinks were not separated; natural juice and fruit-based juice drinks were also not separated.
Dental Caries Data
Sociodemographic Information
Statistical Analysis We conducted the cluster analysis based on the proportions of total fluid intake represented by each of the four primary sources defined above, rather than the absolute amount of fluid intake. With the FASTCLUS procedure, the number of clusters as an outcome must be pre-determined. We did this by evaluating various cluster numbers (2 to 10) by comparing the approximate R-squared values and the within-cluster standard deviations. The R-squared value increased with more clusters, but the within-cluster standard deviations no longer decreased after four clusters, which was the criterion for determining the number of clusters to be used in our analyses. Drinking habits and amounts varied substantially by children’s age in our preliminary analysis. Hence, the cluster analysis procedures were applied to data that were stratified by children’s age groups: 2-year-olds (toddlers), 3- to 5-year-olds (pre-school children), and 6- to 10-year-olds (school-aged children). The crude associations between dental caries and fluid consumption groups, as well as with other sociodemographic characteristics, were analyzed with the chi-square test. Subsequently, a multivariate logistic regression model was constructed to predict dental caries in the primary dentition with fluid consumption patterns and sociodemographic factors. All analyses incorporated sampling weights to adjust for unequal sampling probabilities and non-response bias (National Center for Health Statistics, 1994). We used SUDAAN (Release 9.0.0 SAS Callable, Research Triangle Institute, Research Triangle Park, NC, USA) to estimate variances adjusted for the design effect from the complex, multi-stage cluster sample design of NHANES III (Shah et al., 1997).
Fluid Consumption Patterns Carbonated soft drinks comprised 8.5% of total fluid consumption among children aged 2–10 yrs. Milk and juice each comprised less than 20%, whereas plain water constituted about 32% of total fluid consumption (Table 1  ).
Cluster analysis identified the four groupings with distinct fluid consumption profiles (Table 1  ). There was a clear pattern in each fluid consumption grouping, based on the proportion of major fluid sources. For example, the first (left-most column) grouping constituted, on average, 31.8% of the total fluid intake from carbonated soft drinks, and was labeled as the high carbonated soft drink consumption cluster. The other groupings were similarly named, according to their largest contributing fluid sources: high juice consumption cluster, high milk consumption cluster, and high plain water consumption cluster, respectively. Overall, 619 children (12.7%) were grouped in the ‘high carbonated soft drink consumption’ cluster. Similarly, 21.1% were grouped in the ‘high juice’ cluster, 24.5% in the ‘high milk’ cluster, and 41.7% in the ‘high plain water’ cluster (Table 1).
There were significant associations between the fluid consumption pattern and sociodemographic factors (Table 2
Carbonated Soft Drink Consumption and Dental Caries in the Primary Dentition About 38% of children had at least one decayed or filled tooth surface in the primary dentition; the mean number of decayed and filled primary tooth surfaces (dfs) was 2.94. Only 52% of children with high carbonated soft drink consumption were caries-free, which was from 10% to 15% lower than children in the other groups (Table 2 ). In a logistic regression model, these children were about 1.8 times more likely to experience dental caries in the primary dentition than those in the ‘high milk’ and ‘high water’ clusters, after adjustment for age, race/ethnicity, sex, and SES (Table 3). Children in the ‘high carbonated soft drink consumption’ cluster also showed a significantly higher tendency toward caries experience than those in the ‘high juice’ clusters. While children of the ‘high juice’ cluster showed a higher tendency toward caries experience than those children in the ‘high milk’ and ‘high water’ clusters, the differences were not statistically significant. Children with a high milk consumption pattern had a tendency toward lowest caries experience. The model explained about 15% of variability of dental caries prevalence among children.
Using the cluster analysis method, we analyzed children’s fluid consumption patterns. To our knowledge, this is the first cluster analysis of fluid consumption patterns among children in the United States. We identified four distinct fluid consumption patterns. Because different clustering methods can generate different solutions in the same dataset, several trials with higher numbers of groupings (5 to 10 clusters) were evaluated to provide more detailed fluid consumption patterns (i.e., high-milk-high-juice; high-juice-high-water; and so on). However, more complicated patterns were not necessarily always clearly discernible, or useful in subsequent analyses. Children with high carbonated soft drink consumption had a significantly higher prevalence of dental caries in the primary dentition than did children with any other fluid consumption pattern. We did not separately analyze sugared and non-sugared carbonated soft drinks, due to a small number of children (only 6.7%) who reported consuming non-sugared carbonated soft drinks. A preliminary analysis showed that cluster analysis with non-sugared carbonated soft drinks as a separate entity yielded results that did not differ from those in the combined category. Children with a high carbonated soft drink consumption pattern showed significantly higher caries experience, even compared with those children with a high juice consumption pattern. The findings of this analysis agree with the results from some previous studies (Grindefjord et al., 1995; Moynihan and Holt, 1996; Marshall et al., 2003). In contrast, a previous analysis of the NHANES III data reported no significant association between sugared soda consumption and dental caries in the primary dentition (Heller et al., 2001). The disparity between these results could be explained, in part, by differences in data analytic methods, for the previous analysis (Heller et al., 2001) used individual fluids rather than clusters. Cluster analysis has the advantage of dealing with a greater complexity of fluid consumption patterns when compared with single-fluid-source approaches; it examines all fluid sources, especially those that are consumed in high quantities (Akin et al., 1986). Children who consumed high amounts of carbonated soft drinks may also have undesirable eating patterns and eat high amounts of sugars from other dietary sources. If so, the high carbonated soft drink consumption pattern might serve more as a marker for unhealthy diet patterns, such as high consumption of sugars. The complexity of the modern diet makes the analysis of diet patterns and fluid consumption patterns challenging. Nevertheless, further analysis, such as cluster analysis including both solid food and fluid sources, should investigate more detailed relationships between diet and fluid intake patterns, and their implications for health and disease outcomes such as dental caries and obesity. Fluid consumption patterns of children varied significantly by sociodemographic factors. This finding showed some consistency with results from previous reports (Ershow and Cantor, 1989; Harnack et al., 1999; Sohn et al., 2001). In our results, white children showed a higher tendency toward a high carbonated soft drink consumption pattern, and African-American children showed a tendency toward higher plain water and lower milk consumption patterns than did children from other race/ethnic groups. Unfortunately, the implication of various fluid consumption patterns among different population groups in relation to their health outcomes has received only limited attention. Conducting a large-scale study on fluid consumption and dental caries may not be practical; however, small-scale studies of beverage and diet intake, focusing on specific population groups, could widen our understanding of the mechanism behind these interactions. A few limitations of this analysis should be noted. The NHANES III used a cross-sectional design. Therefore, the results from this analysis should not be interpreted as a cause-effect relationship between and among variables. A fluid consumption pattern from a 24-hour recall interview may not be representative of an individual’s general fluid consumption pattern over time. Although the 24-hour recall interview has been reported to provide reliable group mean estimations (Persson and Carlgren, 1984), it was also reported to be prone to reporting of extreme values, non-reporting, and under-reporting (Yetley et al., 1992). In the NHANES III dietary recall, proxy respondents (caregivers) were allowed for young children (all children until 6 yrs old, and some up to 11 yrs old). For children 1 to 2 yrs old, the caregivers might be well aware of their children’s fluid consumption, but less likely so when the children are older and can obtain drinks for themselves. In summary, we found that a high consumption of carbonated soft drinks in early childhood was significantly associated with an increased risk of dental caries in the primary dentition after adjustment for age, sex, race/ethnicity, and SES, as well as consumption of fluid from other sources. Findings from this study suggest that the high consumption of carbonated soft drinks by young children is a risk indicator for dental caries in the primary dentition, and should be discouraged.
This research was funded independently by the lead author while a student at the University of Michigan School of Public Health. Received for publication August 12, 2004. Revision received August 26, 2005. Accepted for publication October 27, 2005.
Journal of Dental Research, Vol. 85, No. 3,
262-266 (2006) This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
|
|
 |
|
Sign In to gain access to subscriptions and/or personal tools.
TOP
ABSTRACT
INTRODUCTION
