Dental development in a South African sub adult population: determination of reference values for permanent tooth formation and emergence

A thesis submitted to the Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, in fulfilment of the requirements for the Degree of Doctor of Philosophy in Medicine. October 2017.Background: Population-based knowledge of human biological growth and development processes is fundamental for assessing the health status of a community. This requires an understanding of the growth pattern for the children as well as the environmental stresses that disrupt or impede their growth. These stresses are usually easy to identify, but data on normal development and growth variation in most populations is surprisingly lacking. Instead, researchers typically compare growth in the population of interest to references formulated for European or US children. The problems associated with using non-population specific references are complex, and their application can lead to misrepresentations of the health status. In addition, the influence of environmental factors on dental development is still debated and the relationship of dental development with life history events, such as sexual maturity, is unclear. Aim: The aim of this study is to develop population-specific reference for permanent tooth formation and emergence among Black Southern Africans, to compare this reference with other population references, and to investigate the influence of sex and nutritional status on dental development. Method: Study design and population This is a cross sectional study. A total of 642 children comprising of 270 males and 372 females from primary and secondary schools were recruited over one and half years. Only participants whose parents and grandparents are indigenous Southern Africans were included. Participants were screened in a mobile dental truck fitted with digital panoramic x ray. Systematic Review A literature search of PubMed, Scopus, Ovid, Database of Open Access Journals and Google Scholar was undertaken. All eligible studies published before December 28, 2016 were reviewed and analyzed. Meta-analysis was performed on 28 published articles using the Demirjian and/or Willems methods to estimate chronological age. The weighted mean difference at 95% confidence interval was used to assess accuracies of the two methods in predicting the chronological age of children. Tooth formation in Southern Africa To investigate tooth formation, all the 642 Black Southern African children comprising of 270 males and 372 females were recruited. The panoramic radiograph of each child was analysed and the dental maturity score of the left mandibular permanent teeth was obtained according to the Demirjian et al. (1973) method. The dental maturity score of each child was converted to dental age using standard tables and percentiles curves for both sexes by Demirjian et al. (1973). The ages of attainment of specific maturity stages were calculated with pr obit analysis and compared by sex and population. Comparisons of age estimation methods For comparison of the common methods used in estimation of age, 540 children (233 males and 307 females out of the 642 children were recruited. This is because all the children aged 16 years and above have reached 100% maturity and hence excluded from the study. Panoramic radiographs of the children were analyzed and the dental maturity scores of the left mandibular permanent teeth were calculated according to the Demirjian et al. (1973), Demirjian and Goldstein (1976) and Willems et al (2001) methods. The dental maturity scores were converted to dental ages using standard tables and percentiles curves for males and females (Demirjian et al. 1973; Demirjian and Goldstein 1976; Willems et al. 2001). The dental ages obtained were compared to the chronological ages of the children and the mean differences obtained by the three methods compared. Nutrition and tooth formation Effect of nutrition on tooth formation was investigated on all the 642 Black Southern African children comprising of 270 males and 372 females were recruited. The Panoramic radiograph of each child was analysed using the Demirjian et al. (1973) method. The dental maturity score of each child was converted to dental age using standard tables and percentiles curves for both males and females by Demirjian et al. (1973). In addition, measures of nutritional status such as, height, weight, mid upper arm circumference and head circumference were obtained from the children. The timing, sequence of emergence and the effect of nutrition on tooth emergence To investigate tooth emergence and the influence of nutritional status on emergence, information on type of teeth and number of teeth emerged were collected from 639 (266 males and 373 females) Black Southern African children aged 5-20 years out of the total 642 children because the emergence data for 3 children were found to be incomplete. An emerged tooth was defined as a tooth with any part of its crown penetrating the gingiva and visible in the oral cavity. Height, weight, mid upper arm circumference and head circumference of the participants were measured. Children with any form of tooth impaction and agenesis were excluded from the study. Life history events and dental development To determine the association between tooth development and life history variables, mean ages of attainment of sexual maturity stages were adapted from Lundeen et al. (2015) and Norris and Richter (2005) to identify if any stage of dental development co-occured with life history events. Southern Africa specific reference values The WITS Atlas was developed using the tooth formation stage with the highest frequency for each tooth. This stage was considered the developmental reference for an age cohort. Southern African tables of conversion of maturity scores were generated separately for males and females using polynomial regression functions (3rd degree). Maturity curves were plotted to determine the dental maturity curves for each sex. The Southern African specific tables of conversion of maturity score were tested on 540 participants aged 5 to 15.99 years and the results compared to the Willems and Demirjian methods of age estimation. Data were analysed with Stata 12 for Windows. The analysis included frequencies and cross-tabulations. Associations between categorical variables were tested with chi square while those between continuous variables were tested with Student’s t-tests. The mean ages of emergence and standard deviation were computed using probit analysis. Sex and population comparisons were done using Student’s t-tests. The height and BMI were converted to z-scores using WHO z-scores for age tables (WHO 1995). A cut-off z-score of ≤−2 for BMI/height was used to place children into underweight/short for age, ≥-2 to 2.0 for normal, and ≥2 for overweight/obese/tall for age categories. Mean age of emergence and mean age of attainment of maturity stages were calculated for each tooth using these BMI subdivisions. Analysis of variance (ANOVA) and Games-Howell were used to determine the differences between the BMI/height subdivisions. A Student’s t-test was used to compare any two means whenever one of the three subdivisions of BMI did not yield a mean age of emergence. Spearman’s rho correlations between total number of teeth, dental maturity scores and anthropometric variables were done. A Shapiro-Wilk W test showed that the dependent variables (total number of teeth emerged and dental maturity) and the predictor variables were not normally distributed. Therefore, a generalized linear model (negative binomial) was used with the number of emerged teeth/dental maturity modelled as the dependent variable and anthropometric variables and age as predictors. Adequacy of fit was checked using the deviance residuals as recommended by McCullagh and Nelder (1989). The deviance residuals showed that it was normally distributed and the plot of the residuals against each of the covariates also showed model fit. As expected, the collinearity test showed that BMI, height and weight were significantly collinear. When these variables were excluded from the model, there was no difference in the values of the output. Hence, the variables were included in the final model for generalized linear regression analysis. The model was built using forward selection. Statistical significance was inferred at p<0.05. Results: Systematic review Meta-analysis revealed that the Willems method has better accuracy globally compared with the Demirjian method. Dental maturity in Black Southern Africans The females show advanced dental maturity and dental ages compared to males (p<0.05). Cross-population comparison shows the Southern African females are advanced in dental maturity compared to European and Asian children. Comparison of methods for estimating dental age The Original Demirjian method significantly overestimated the age of the males by 0.85 years and the females by 1.0 years (p<0.05) with the same mean absolute error of 1.1 years for both sexes. Similarly, the Modified Demirjian method significantly overestimated chronological ages of males (0.90 years) and females (1.21 years) with the highest mean absolute error of 1.1 years and 1.4 years for males and females respectively. The Willems method had the lowest, but still significant mean differences (0.2 years for males and 0.3 years for females) between the dental age and chronological age. It also demonstrated the least mean absolute errors for males (0.70 years) and females (0.68 years). Nutrition and tooth formation Significant advancements were found in the age of attainment of H stage for all the permanent teeth in the overweight group compared to the underweight group (p<0.05). Negative binomial regression analysis indicates that age, height, and BMI are significant predictors of the dental maturity score for males (p<0.05), while age, height, weight, BMI and head circumference are significant predictors of the dental maturity score for females. Tooth emergence Females have all the permanent teeth emerged earlier than males except for the third molars (p<0.05). Generally, Black Southern African children have similar ages and sequence of emergence as children from other sub-Saharan Africa countries. Black Southern African children have earlier mean ages of emergence of permanent teeth compared to children from the USA, Europe, Australia and Asia. Sexual dimorphism was noted in the sequence of emergence of I1/M1 in the mandible with the females having the M1I1 sequence as opposed to I1MI in males. The sequence of emergence of Southern African males is similar in both jaws to males from the USA and Europe but differs from Iranians and Pakistanis. Females show similar patterns of sequence with sub-Saharan African, Australian and US females in the maxilla. They display MI/I1 variation in the mandible. Nutrition and tooth emergence Overweight/obese children generally show significantly earlier emergence times compared to normal weight/severely underweight children (p<0.05). Females and tall children have more emerged teeth than shorter children when corrected for age and sex (p<0.05). The generalized linear regression model (negative binomial) shows that height, weight and BMI have significant associations with the number of emerged teeth (p<0.05). Dental development and life history variables The number of teeth emerged in males correlate strongly with chronological age (r=0.91, p=0.00) and height (r=0.89, p=0.00), moderately with mid-upper arm circumference (r=0.61, p=0.00) and weakly with head circumference (r=0.16, p=0.00). In females, the number of teeth emerged correlates strongly with chronological age (r=0.88, p=0.00) and height (r=0.83, p=0.00), moderately with mid-upper arm circumference (r=0.59, p=0.00), and weakly with head circumference (r=0.38, p=0.00). Similar patterns of correlation are found for dental maturity. The emergence of the maxillary and mandibular M2s co-occurs with the G2 stage of gonad development and the PH2 stage of pubic hair development in males. The M2s emerge coincident with the attainment of Tanner’s B2 breast stage and the PH2 pubic hair stage in females. The age of menarche does not coincide with any of the determined ages for emergence of teeth. Attainment of the H stage of development in the C1 co-occurs with the G2 stage of gonad development and shortly after the pubic hair stage PH2 in the males. In females, the attainment of the H stage of C1 formation occurs shortly before the attainment of the B2 stage of breast development. Furthermore, the H stage of P1 formation coincides with the PH2 stage of pubic hair development, shortly after the attainment of the stage B2 of breast development. The attainment of the H stage in P2 formation coincides with the age of menarche at approximately 13 years. Southern African specific reference A new dental atlas (WITS Atlas) was developed due to the significantly earlier ages of emergence and formation among Black Southern Africans. When compared to the London atlas, the canines, premolars and second molars are at least a year ahead in the WITS Atlas. Third molar formation and emergence occurs three years earlier in the WITS Atlas. Polynomial regression formulae were generated and Southern African specific conversion tables were generated for the males and females. The new tables of maturity scores show no overestimation of the chronological ages of males (0.045, p>0.05) and females (0.08, p>0.05). Compared to the Willems and Demirjian methods, the Southern African specific maturity tables showed the least mean absolute error for both sexes. Conclusion: There is sexual dimorphism in the timing of tooth emergence with females having earlier emergence times. Black Southern Africans show similarities in the ages and sequence of emergence of the permanent teeth with children from other sub-Saharan African countries but, they are advanced relative to children from the USA, Europe, Australia and Asia. Similarly, the Black Southern African children show advanced tooth formation compared to children from Europe, Asia and Australia. The Willems method is more accurate at estimating chronological age for forensic and anthropological purposes compared to the Demirjian methods that significantly overestimate the chronological age of children. Of the three methods tested on Black Southern African children, the Willems method is the most accurate in estimating chronological age. However; it significantly overestimated the chronological age of Black Southern African children. Hence, there is a need for population-specific reference values for use in the age estimation of Black Southern African children The WITS Atlas and new population-specific maturity tables for Black Southern African males and females were developed. The WITS Atlas differs significantly from the London atlas with earlier ages of tooth formation and emergence. The Southern African specific age estimation method shows good accuracy in the estimation of dental ages. By inference, this method could be used in other sub-Saharan African countries because of similarities in tooth formation and emergence times. Contrary to some studies, nutrition was found to have a significant influence on the number of teeth emerged and the timing of emergence. Obese/overweight/tall children tend to have earlier timing of emergence and more emerged teeth compared to their underweight peers. Similarly, obese/overweight/tall individuals attained the H stage of tooth formation of most teeth earlier than their underweight and normal weight age-mates. Emergence of second molars and the H stage of canine and first premolar formation co-occur with the onset of puberty in males and females. Menarche appears to coincide with the attainment of H stage of the mandibular second premolar.LG201