Gingival bleeding and calculus among 12-year-old Chinese adolescents: a multilevel analysis

Oral diseases are the most common noncommunicable diseases, and have become alarming public health problems worldwide [1]. Gingivitis, one of the most common oral health problems, is a mild form of periodontal disease [1, 2]. Plaque-induced gingivitis is considered to be the most prevalent type of gingivitis [3, 4]. The clinical symptoms of gingivitis include redness, oedema and bleeding at the gingival margin [5]. During puberty (girls, 11–13 years; boys, 13–14 years), periodontal tissues may have an amplified response to local factors: dental plaque, calculus, food debris, and materia alba [6]. And regular removal of these local factors can prevent the occurrence and progression of early gingival disease [7, 8].

Dental calculus (mineralized biofilms), located along the gingival margin, is an important gingivitis-contributing factor [9, 10]. Calculus can’t be removed easily from the surface of teeth and provides a substratum for plaque retention in the vicinity to the gingiva [9, 10]. Calculus can also be a surrogate indicator of long-term exposure to biofilm and poor oral hygiene practices [11]. Its presence, along with gingival inflammation, is associated with the initiation and progression of early-onset periodontitis [10].

Gingivitis and periodontitis can be considered as a continuum of the same inflammatory process, and untreated gingivitis can progress to periodontitis complicated with further tissue destruction and bone reabsorption [2, 3, 12]. Therefore, gingivitis management of adolescents is a prevention strategy for advanced periodontal diseases that might persistently play a crucial role in reducing the periodontal disease burden [2, 4]. The greatest share of oral health problems is attributable to the social conditions in which people live and work, referred to as the social determinants of health (SDH) [13‐15]. There are huge disparities in the level of economic development and the availability of healthcare resources in China [16]. The contextual differences are associated with several clinical outcomes of various oral diseases and could thus be related to gingivitis. Other individual SDH factors associated with gingivitis include sex and ethnic inequalities [13, 14], family factors [14, 17], and health behaviors [14, 17, 18]. Therefore, it is important to identify associated individual and contextual factors whose modification can reduce the prevalence of gingival bleeding and calculus in the future, and further, to provide advises on controlling such factors for maintaining a good quality of life [19].

Traditional periodontal examinations are resource-intensive, and a large-scale screening for adolescents is infeasible in China [20]. Recently, self-reported questionnaires have been proposed to predict the probability in periodontal surveys. And the validation of prediction models based on self-reported questionnaires were demonstrated in Taiwan [20], Australia [21], Germany [22], Japan [23], and the United States [24]. Twelve years has been recommended by World Health Organization (WHO) as the indicator age group for international benchmarking of adolescents’ oral health [25, 26]. Little is known about the prevalence of gingival bleeding and calculus in adolescents with large sample size in western China. Therefore, the present study aimed to investigate the associated factors for gingival bleeding and dental calculus in 12-year-old adolescents and to build predictive models for gingivitis management and treatment planning.

A total of 4525 subjects aged 12 years were selected for eligibility in this study with an effective response rate of 94.27%. Among the participants, 48.60% (n = 2199) of the participants were boys, and 51.40% (n = 2326) were girls. The prevalence of gingival bleeding was 46.63% (n = 2110; 95%CI: 40.71, 51.60), that of calculus was 66.94% (n = 3029; 95%CI: 56.85, 67.45). The mean scores of the individual scores of gingival bleeding and calculus were 4.00 and 4.57, respectively.

The details of the total data were showed in Table 2. Among, the training dataset characteristics and the results of univariate analyses were presented in Table 3. As Fig. 2 showed, gingival bleeding in three regions (Guang’an District, Jinniu District, and Pi County) and calculus in two (Yibin County, and Pi County) indicated significant variations (ie, they did not overlap the horizontal line at zero). Therefore, two-level logistic regression models were built with the selected variables by univariate analyses.

Table 4 presented the eight individual variables significantly associated with gingival bleeding (p < 0.05) in two-level logistic regression analysis. The model for gingival bleeding was established including hukou (OR = 0.61, 95% CI: 0.52–0.72), family size (OR = 1.41, 95% CI: 1.19–1.68), parental educational level (Father: OR = 0.53, 95% CI: 0.45–0.63; Mother: OR = 0.71, 95% CI: 0.59–0.86), tooth-brushing frequency (OR = 0.35, 95% CI: 0.26–0.48), dental floss use (OR = 0.58, 95% CI: 0.41–0.83), sugar-containing drink consumption (OR = 2.11, 95% CI: 1.80–2.49), and dental visit (OR = 1.44, 95% CI: 1.19–1.74). Regarding the associated factors of calculus, the same model was used and significant variables including gender (OR = 1.32, 95% CI: 1.13–1.54), hukou (OR = 0.69, 95% CI: 0.59–0.82), family size (OR = 1.34, 95% CI: 1.12–1.59), parental educational level (Father: OR = 0.46, 95% CI: 0.39–0.54; Mother: OR = 0.65, 95% CI: 0.59–0.82), tooth-brushing frequency (OR = 0.57, 95% CI: 0.42–0.78), dental floss use (OR = 0.66, 95% CI: 0.48–0.90) and sugar-containing drink consumption (OR = 1.30, 95% CI: 1.11–1.53) are presented in Table 5. Tables 3 and 4 also displayed the parameters of the fixed and random effects in the two-level logistic regression models and their average intercepts. Finally, the percentage variance explained by the contextual level was 3.13% for gingival bleeding model and 2.27% for calculus model.

Based on these models with the average intercepts and coefficients of the variables before, area under the curve (AUC) values were calculated from the receiver operating characteristic (ROC) curves for gingival bleeding and calculus deposits (Table 6). As to the model of gingival bleeding, the estimated AUC values were 0.73 (95% CI: 0.72, 0.75) for the training data and 0.75 (95% CI: 0.73, 0.78) after applying the model by using the validation data. For the model of calculus deposits, the estimated AUC values were 0.68 (95% CI: 0.66, 0.70) for the training data and 0.67 (95% CI: 0.63, 0.69) for the validation data. Table 6 showed the performance values for all models, and the optimal cut-off values for the probabilities were calculated using maximized Youden indices. Based on these optimal cut-off values for all samples in the gingival bleeding models, the sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), Youden index and predicted prevalence were determined.

This was a cross-sectional study included a relatively large sample size (N = 4525) to assess the prevalence of gingival bleeding and dental calculus in western Chinese adolescents. The study found high prevalence of both gingival bleeding and calculus, 46.63% of participants showed gingival bleeding, and 66.94% had calculus deposits. Here multistage sampling was used in this study, which was considered as a more accurate sampling [35, 36]. It is also a convenient and effective way for finding the survey sample and is particularly suitable for our study, which needed a large sample size [35, 36]. A higher prevalence of calculus was calculated than that of gingival bleeding. These results are contrast with the findings of previous studies [4, 37]. The reason may be the different qualities of water in western China, with high hardness and much salinity [38, 39]. Therefore, the effect of water supply on the oral health is needed to explore in the further study.

In the present study, multi-level models were used to correctly specify the data structure and to demonstrate the significant variables of gingival bleeding and calculus. As screening approach for assessing the associated factors of gingival bleeding and calculus, adolescents could be encouraged to modify bad behaviors to decrease the prevalence. For simple validation method, we randomly divided our study population into three-quarters and one-quarter samples used in the previous studies [20, 40]. Simple random sample was used to select the training data and validation data. Actually, the dataset selection should be considered its multilevel structure [41]. If this is not done, the analysis may give incorrect or potentially misleading results. Because the hierarchical datasets may be close to flat and candidate models can be swamped by random fluctuations [41]. Therefore, we confirmed the distribution of the samples in 6 regions were similar, and no significant bias existed. Table 6 showed the summary findings of our models [20]. Among, Youden’s index provides a measure of validity by considering both sensitivity and specificity [42]. Here the low values of Youden indices suggested the low predictive capacity of the models. The results also implied the models could only be used as a first-line screening tool. Regular dental visit is still needed. Previous study said ethnical and methodological differences would influence the predictive capacity [42]. In addition, this epidemiological survey included only 6 regions. A small region size cannot produce valid and accurate estimates for multi-level logistic regression models [43]. Further we will increase the regions and adolescents included. And factors in questionnaire, especially contextual variables, should be added in the survey as well. Then the prediction models would obtain more accuracy.

From a conceptual perspective, the results contribute to our understanding of the determinants of gingival bleeding and calculus. The popularity of calculus was higher in male gender, however, there was little sex inequalities presented in the outcome of gingival bleeding. Previous studies confirmed that prevalence of gingivitis in puberty was correlated with elevation in systemic levels of the sex hormones (testosterone in boys and estradiol and progesterone in girls) [4, 6, 7]. Increased serum levels (testosterone in boys, and estradiol, progesterone in girls) are positively associated with high levels of Prevotella (P.) intermedia and P. nigrescens [6]. The age peak of gingivitis in girls (11–13 years) was earlier than that in boys (13–14 years) [4, 6, 7]. Therefore, the increased sex hormones in girls may interfere the results. Along with gender, age and income, hukou is quite a common variable on individual behavior in social and economic studies in China [44, 45]. It was proved to be associated with dental health. But previous studies were focused on its impact on dental caries [46, 47]. In this study, the periodontal status was better in urban hukou than that in rural hukou. The reason might be people with rural hukou didn’t have full access to oral-related public health services funded by the tax revenue [46, 47].

Several characteristics, broadly categorized as child-level, family-level and community-level, could influence the periodontal diseases in adolescents [48, 49]. Family-level characteristics affect both the child-level factors, such as self-care practices and service utilization, and periodontal disease status [50, 51]. This is because parents and adolescents share similar cariogenic and periodontal pathogenic microbiota [52, 53]. The results of our analyses suggested that sociodemographic characteristics of adolescents’ family at 12 years of age, as well as family environment, were important determinants of the occurrence of gingival bleeding and calculus in adolescents. In this study, the only child had lower prevalence of gingival bleeding and calculus than child with siblings. The results were similar with many previous studies that proved less adolescent number in one family was easier to have better periodontal condition [54, 55]. In addition, our study observed a positive impact of higher level of parental education on the gingival status of adolescents, and mothers’ and fathers’ education levels, respectively, could be used as a predictor of gingival bleeding and calculus.

As to the self-care practices and habits, unhealthy diet behaviors (high carbohydrate diet) and bad hygiene care (less brushing frequency per day) increased the prevalence of gingival bleeding and calculus. However, no significant difference was found between adolescents brushing once/day and those did not brush at all. This thought-provoking point surprised us. The potential reasons to interpret this phenomenon are as following: first, gingivitis is caused by multiple factors [54], including not only brushing frequency, but also duration of brushing, brushing method, and the type of dentifrice [55]. Second, people with irregular health habits, such as meals, tooth-brushing and bedtime, are more prone to disease [56, 57]. Third, reporting bias and undercounting of errors were unavoidable in this survey.

The periodontal status of the included adolescents in this survey was unsatisfactory, and most of them had calculus. However, a somewhat high number of adolescents claimed to brush their teeth at least once a day. This inconsistency between periodontal health and reported dental hygiene practice may be a result of the adolescents having not acquired the proper brushing techniques. Motivation to implement the instructions given for oral health care and regular enforcement are essential in promoting adolescents’ oral health [58]. In China, dental curricula are not enforced and only provided in certain schools in the health education textbook [59]. Here, the popularity of dental curricula and the relationship between this intervention and oral health were explored. Dental curricula showed limited impact on the oral health improvement, which suggested that better and more effective oral health education was essential.

Only 20% of the students had visited a dentist during the last 12 months, which was a lower proportion than that found in the other surveys [60]. However, a higher gingivitis rate was found in the adolescents who had visited dentists. The trend was consisted with previous studies [35, 36]. This might be due to the common problem-oriented dental-care-seeking behavior in China. It was plausible that adolescents were brought to visit dentists for pain and infection [35, 36]. The limited availability of dental-care resources might be another reason. The dentist to population ratio is around 1:30,000 in China and the number of dentists is even fewer in rural regions [61].

Some limitations need to be analyzed in this investigation. This was a cross-validation analysis, an external validation was not possible because of the limited numbers of variables obtained from other screening database. In addition, the questionnaires were completed by the participants; thus, certain limitations were unavoidable, such as response bias, reporting bias and undercounting of errors. Moreover, more contextual variables that reflect the inequitable distribution of dental-care resources and oral health coverage, water supply and utilization of dental services need to be assessed. At last, this epidemiological survey included only 6 regions in Sichuan Province based on a complex sample design. A small region size may hardly represent the whole province’s population, and generate biased estimates. Because of these limitations, our results should be interpreted with caution.

The findings from this study can serve as a warning sign for adolescents and policy makers in western China. Oral health education and practice should be strengthened, especially for adolescents with high risk. Individual, parents and educators should be instructed about the prevention and management of gingival disease. And oral health educational programs for parents are also imperative. Tackling the inequitable distribution of dental care resources is very important. Due to lack of dentists in some regions, the cultivation of a community-based dentist may be a priority, which can improve the adolescents’ access to dental care.

Gingival bleeding and calculus were prevalent among 12-year-old western Chinese adolescents. Unhealthy diet behaviors and insufficient hygiene care were related to the high prevalence. Individual, family, and school should be instructed about the prevention and management for improving the gingival status of adolescents. Moreover, regular dental visit is also recommended.