| Sign In to gain access to subscriptions and/or personal tools. |
Predicting Caries in Permanent Teeth from Caries in Primary Teeth: An Eight-year Cohort Study
1 Department of Basic Science and Craniofacial Biology, New York University College of Dentistry, 345 E. 24th Street, New York, NY 10010-4086, USA; and Correspondence: * corresponding author, yihong.li{at}nyu.edu
Several cross-sectional studies report that caries in primary teeth is correlated with caries in permanent teeth. This eight-year cohort study sought to determine if caries in the primary dentition can predict caries in the permanent dentition of the same individuals and, if so, with what degree of prediction accuracy. A total of 362 Chinese children, from 3 to 5 years old at the time of the 1992 baseline study, were re-examined in 2000. The study found statistically significant associations between caries prevalence in primary and permanent dentitions (p < 0.01). Children having caries in their primary teeth were three times more likely to develop caries in their permanent teeth (relative ratio = 2.6, 95% CI = 1.4-4.7; p < 0.001). Caries on primary molars had the highest predictive value (85.4%). This study demonstrates that caries status in the primary teeth can be used as a risk indicator for predicting caries in the permanent teeth.
Key Words: dental caries • primary and permanent teeth • Chinese children
Dental caries is the most prevalent of all chronic diseases among US children (USDHHS, 2000). It affects 18% of all children ages 2-4, 52% of children ages 6-8, and 80% of adolescents age 17 (Kaste et al., 1996). In China, caries prevalence ranges from 67% to 86% among pre-school children (3 to 6 yrs old) and 32% to 48% among adolescents (12 yrs old) (PRCMPH, 1987,1999; Petersen and Guang, 1994; Wang et al., 1994; Douglass et al., 1995; Peng et al., 1997; Wong et al., 1997,2001; Petersen and Esheng, 1998). High caries prevalence still endures as one of the major oral health issues in children. Because dental caries is a disease that is both preventable and costly to treat, previous studies have focused on identifying caries risk predictors, including developmental tooth defects, mutans streptococci infection (time, source, and level), lactobacilli counts, salivary buffer capacity and flow rate, sucrose intake frequency, and past caries experience (Seppä et al., 1989; Alaluusua et al., 1990; Helfenstein et al., 1991; Disney et al., 1992; Steiner et al., 1992; Vehkalahti et al., 1996; van Palenstein Helderman et al., 2001). Most of these studies, however, consisted of a single or several cross-sectional surveys. Information validating the correlation between caries in primary and permanent dentitions of the same individuals is needed. The present eight-year cohort study addresses the following questions: Does the caries status of the primary teeth correlate with that of the permanent teeth in the same individual? If so, which teeth contribute most strength to the correlation? Finally, can caries in the permanent teeth be predicted from caries status in the primary teeth of the same individual? Here, we examined the specificity, sensitivity, predictive value, and efficiency of various risk predictors that might be used to predict future caries in this prospective study.
Study Cohort The study was approved by the University of Alabama at Birmingham Institutional Review Board for Human Use with English Informed Consent and Chinese translation. Parents’ consent was obtained prior to the study. The baseline dental examination for caries status in the primary dentition was conducted from July to October, 1992 (Li et al., 1996). A total of 504 children, 256 boys and 248 girls, participated in this eight-year cohort study. They were randomly selected from 11 villages and 4 kindergartens of two representative communities, outside the Beijing metropolitan area. The children were then 3 to 4 yrs old (mean = 3.5). Gender, age, and socio-economic status (SES) were equally distributed. Forty-nine percent of the families had annual incomes less than ¥3000 RMB ($370 US dollars) and were categorized as the low-SES group. The average fluoride concentration was below 0.26 ppm in the drinking water. Eight years later, in December of 2000, the same group of children (N = 362, 46% boys and 54% girls) was re-examined for caries status in their permanent dentition. Fifty-seven of the children had passed their 13th birthday when the oral examination was conducted. Therefore, the average age of the children examined was 11.7 yrs, ranging from 11 to 13 yrs. The rate of loss to follow-up was 28%, in part the result of two entire villages relocating for regional irrigation projects. More than half (56%) of the families remained at annual income levels below the regional average.
Dental Examination
Statistical Analysis
Caries Status Caries prevalence and the mean values are summarized in Table 1  . Caries prevalence in the primary dentition was 83%; the mean dmft (dmfs) was 6.1 ± 4.7 (12.5 ± 12.4). The caries prevalence associated with increase in age was significant (p < 0.01). No significant difference was observed between genders. Caries prevalence in the permanent dentition among Chinese adolescents was 41%, and the means of both DMFT and DMFS were low. When caries status was compared with SES, higher caries prevalence was found among children with lower SES (86% vs. 81%, p < 0.05; dmft 6.7 vs. 5.7, p < 0.01) in primary dentitions. In the permanent dentition, however, a reverse correlation was observed (35% vs. 47%, p < 0.05). Adolescents in the high-SES group also experienced more caries than adolescents in the low-SES group (Independent t test, p < 0.05).
Caries Correlation and Relative Risk Results from this study illustrated a significant correlation between caries experience in the primary and that in the permanent teeth (r = 0.38, p < 0.001). A greater correlation coefficient was found between caries in primary maxillary anterior teeth and caries in permanent molars (r = 0.49, p < 0.001). Of the children who developed dental caries in permanent teeth, 94% of them experienced caries in their primary teeth (Pearson  2 = 13.7; p < 0.001). Of the children who did not have any caries in their primary teeth, 83% remained caries-free by the age of 12. A relative risk value of 2.6 (95% CI of 1.4 to 4.7, p < 0.001) indicated that children who manifested caries in their primary teeth were three times more likely to develop caries in their permanent teeth than were those children previously free of caries. The study also revealed that the mean decay in the permanent teeth significantly correlated with the mean decay in the primary teeth (p < 0.001). The linear regression analysis demonstrated a trend in which the number of permanent teeth with caries rose during the study period in response to an increase of the mean caries score of the primary teeth (FDMFT = 0.39 + 0.08*dmft; r = 0.27; p < 0.001). In addition, a steady increase in the relative risk (RR) and predictive value for caries was observed as the number of decayed tooth surfaces increased (Table 2).
Caries Prediction When the caries experience in the primary teeth was used to predict future caries of the same individual, the study found an overall sensitivity of 93.9%, an overall specificity of 20.0%, and an overall positive predictive value of 85.4%. To answer the question as to which teeth contribute the most to the positive predictive value, we repeated the statistical analyses on subsets of teeth, by dividing the primary teeth with caries into different groups: maxillary incisors (4 teeth), maxillary anterior teeth (6 teeth), maxillary first and second molars (4 teeth), mandibular first and second molars (4 teeth), and all primary molars (8 teeth). The sensitivity, specificity, predictive value, and efficiency were determined for different tooth combinations according to two categories: caries present in any one of these teeth, or caries present in all of the teeth. Table 3 shows that the highest sensitivity (93.9%) was observed in caries on any of the primary molars with the highest predictive value (85.4%), and the highest specificity (91.6%) was observed in caries on all of the primary molars. An almost-perfect specificity, 97.7% for caries, on all maxillary anterior teeth was paired with a low sensitivity. The highest efficiency (65.8%) was for caries on all mandibular primary molars. The positive predictive value decreased when the prevalence was low, even for high values of sensitivity or specificity.
Another way to determine the predictive value was to compare relative risk values according to caries status of different tooth segments in the primary dentition. As shown in the Fig. , caries presence on all upper incisors (4 teeth) did not have significant predictive value for caries development in the permanent dentition (RR = 1.1; 95% CI = 0.85-1.47; p = 0.43). Caries presence on all upper anterior teeth (6 teeth) had marginally significant predictive value for caries development in the permanent dentition (RR = 1.6; 95% CI = 1.07-2.45; p = 0.07). Caries presence either on all upper primary molars or on lower primary molars had similar predictive values (p < 0.001). A relative risk value for caries presence on all first and second primary molars was 1.8 (95% CI = 1.40-2.30, p < 0.001). The relative risk value was even greater for caries presence on any of the first and second primary molars (RR = 3.4; 95% CI = 1.8-6.1, p < 0.001), with the positive prediction value 94%.
When caries experience in primary teeth is used as a risk indicator to predict caries in permanent teeth, the commonly used and most pragmatic characteristics of tests are sensitivity, specificity, and predictive value. Previous studies reported ranges of sensitivities from 28% to 76%, specificities from 63% to 93%, and predictive values from 61% to 77% for overall positive prediction (Abernathy et al., 1987; Seppä et al., 1989; Wilson and Ashley, 1989; Helfenstein et al., 1991; Disney et al., 1992). Results of this longitudinal study demonstrated that the best indicator of sensitivity (94%) was caries on any of the primary molars. The best indicator of specificity (98%) was caries on all maxillary anterior teeth. There was no single indicator with both high sensitivity and specificity for the identification of individuals at high risk before they develop caries. In addition to the sensitivity, specificity, and positive predictive values, the study also systematically examined the prediction efficiency, calculated from true caries-active and caries-free status, for each of the tooth groups over the grand total. The predictive value of 85.4% suggests that nine out of ten children who had caries in their primary molars will develop caries in their permanent teeth. Accordingly, a combination of caries present on primary molars (the highest sensitivity) and caries-free primary maxillary anterior teeth (the highest specificity) would be the best predictor for distinguishing children as high- or low-risk for caries.
Analysis of the baseline data showed that 78% of children at age 3 and 88.5% at age 4 had at least one decayed tooth. According to the new definition of early childhood caries (ECC) (Drury et al., 1999), 90% of this cohort manifested ECC. But not all of them developed caries in his/her permanent teeth. Since the caries scores ranged from 1 to 20 for dmft and from 1 to 60 for dmfs, a wide spread in positive predictive values was also observed. Based on caries prediction analyses as a function of different caries scores, the study discovered that caries on primary maxillary incisors alone or fewer decayed teeth (dmft Since 1990, several cross-sectional epidemiological studies on caries in Chinese children have been reported (PRCMPH, 1987,1999; Petersen and Guang, 1994; Wang et al., 1994; Douglass et al., 1995; Peng et al., 1997; Wong et al., 1997,2001; Petersen and Esheng, 1998). The caries prevalence in primary dentitions was 76% to 83%, with an average of 4.3 to 4.8 decayed teeth. Caries prevalence in permanent dentitions was 32% to 45%; the average ranged from 0.7 to 1.0. This study found similar results. Interestingly, high caries prevalence and more decayed teeth were observed in the primary dentition of children with lower SES. Eight years later, the caries prevalence in the same cohort was higher among children with higher SES. This discordance of caries status between the primary and the permanent dentitions in Chinese children is remarkable compared with that in populations in developed countries (Kaste et al., 1996). One hypothesis was that the primary teeth, compared with the permanent teeth, are more susceptible to developmental defects (Li et al., 1995), such as enamel hypoplasia, as the result of maternal chronic nutritional deficiency, low birth weight, and prenatal infectious diseases. The second hypothesis was that enamel hypoplastic lesions could foster earlier colonization of cariogenic bacteria, mainly mutans streptococci, in the oral cavity—colonization that could accelerate caries development and progression (Li et al., 1994,1996). As expected, the socio-economic status may not indicate a direct etiologic risk factor for dental caries, but it can significantly undermine tooth susceptibility, especially in the primary teeth, for caries. Since dental caries is a dietary carbohydrate-modified bacterial infectious disease (van Houte, 1994), one would expect a similar caries development pattern in both the primary and the permanent dentitions in the same individual. However, the study observed reversed caries prevalence in the permanent dentition for high-SES children. One explanation could be that the permanent teeth have a longer developmental and maturational period than the primary teeth. They are less affected by prenatal disturbances. The impact of developmental defects on tooth susceptibility to caries might be diminished in children with low SES. Second, there was a substantial difference in the levels of consumption of sugar-containing beverages and sweetness between children of low SES (rural area) and those of high SES (urban area), as a result of overall economic improvement in China (ISO, 2001). An increase in caries prevalence has been reported among high-SES children (Lo et al., 1999; PRCMPH, 1999). In comparison, in children of low SES with well-developed permanent teeth and a more traditional diet, the caries prevalence remained relatively low. Although the study presents several strengths–such as a very homogeneous study cohort with known traditional dietary habits, limited exposure to dental restorative treatments, fluoride, and antibiotics–the main drawback was that caries diagnostic criteria in the study were based on exclusion of non-cavitation (enamel caries lesion) for caries risk prediction. The caries status in primary teeth, the predisposing variable, could have been underestimated. It might have affected the measurement of association between the predisposing variable and the outcome variable, the caries status in permanent teeth, and contribute to the low prediction efficiency observed in the study. Clearly, more sensitive caries diagnostic and detecting techniques are needed for the future study of caries prediction. Overall, this eight-year cohort study demonstrated a significant positive association between caries in primary and caries in permanent dentitions. Future caries development could be predicted based on the overall risk estimation and caries status of specific groups of primary teeth. As evident in this study, assessment of caries status in primary teeth constitutes a valuable prognostic tool with a high level of confidence in predicting future caries, making it possible for preventive regimens to be initiated in anticipation of future caries.
We express our grateful appreciation to Dr. Page W. Caufield for his critical comments on this manuscript. We also thank Drs. Xiang-Yong Pan and Hua Wu and Ms. Shu-Yuan Wang for their technical assistance, and Dr. Howard Sage for his English editing. This study was supported, in part, by the John J. Sparkman Center for International Public Health Education (SCIPHE) of the University of Alabama at Birmingham and by NIH/NIDCR Grant DERR10595. Received for publication October 31, 2001. Revision received June 7, 2002. Accepted for publication June 13, 2002.
Journal of Dental Research, Vol. 81, No. 8,
561-566 (2002) This article has been cited by other articles:
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
|
|
 |
|
Sign In to gain access to subscriptions and/or personal tools.
TOP
ABSTRACT
INTRODUCTION
6) might not necessarily be a sufficient predictive risk indicator for future caries. Children with severe ECC (24%) who had dmft scores equal to or greater than 10 were 3.5 times at greater risk for developing caries in their permanent dentition. Therefore, the current definition for ECC or severe ECC may lack specific and sufficient power to distinguish between those individuals at high risk and those at low risk for caries. 
