Epidemic Trend of Dental Fluorosis in Mainland China, 1995-2020: A Systematic Review and Meta- Analysis

Yuzhen Zhan Chongqing Medical University Stomatology College: Stomatological Hospital of Chongqing Medical University Lianjie Xiong Chongqing Medical University Stomatology College: Stomatological Hospital of Chongqing Medical University Zhoujie Gong Chongqing Medical University Stomatology College: Stomatological Hospital of Chongqing Medical University Ting Xu Chongqing Medical University Stomatology College: Stomatological Hospital of Chongqing Medical University Xiaonan Zhang (  zhangxiaonan@hospital.cqmu.edu.cn ) Chongqing Medical University Stomatology College: Stomatological Hospital of Chongqing Medical University https://orcid.org/0000-0003-0155-776X


Introduction
Dental uorosis (DF) is a speci c esthetic disturbance, which is caused by excessive intake of uoride during the development of teeth (Zhang et al. 2013). Clinically, the manifestations of DF could be divided In China, the second, third and fourth oral epidemiological surveys showed that the national prevalence of dental uorosis was 6.89% in 1995, 11.7% in 2005 and 14.4% in 2015 respectively, showing an increasing trend. Besides, the prevalence of dental uorosis in China is variable across the country. The prevalence of dental uorosis is relatively low in some places, while the prevalence of dental uorosis is particularly high in some areas where the uoride content in drinking water is too high. For instance, the prevalence of DF was 9.58% (Fan et al. 2008) in Yunnan, while 35.17% (Zhu et al. 2006) in Qingdao.
It has been reported that the content of uorine resources in China accounts for 60% of the total reserves in the world (Luo 2011), which probably affects the prevalence of DF. Moreover, China is a vast and multiethnic country with diverse geographical environment. Thus, the causes of dental uorosis in China are more complicated than other countries. For instance, people of some western minority areas in China have the habit of drinking brick tea that contains high content uorine. Long-term intaking high uorine content brick tea causes the endemic uorosis, which manifested as dental uorosis (La et al. 2016).
Additionally, exhaust gas from coal burning may pollute local air, water and crops , Dong et al. 2011. Therefore, it is necessary to gure out the epidemiological characteristics of dental uorosis in China, which would be bene t to allocate the medical resources effectively in China.
Currently, there are no systematic reviews on dental uorosis in adolescents in mainland China.
Therefore, adolescents aged 8-12 years were selected for this study. We analysed the epidemiological characteristics of dental uorosis from all epidemiological survey articles in this meta-analysis, to explore the prevalence trend with time, gender, location, and geography. This may provide useful information for the prevention of DF and implementation of relevant oral health policies for those populations.

Methodology Data sources and searches
Six online databases were comprehensively searched by the rst and second author for the studies published before December, 2020 (PubMed, Embase, CBM, CNKI, Chinese Wan Fang Database, and VIP database). The keywords used for the search included "dental", "tooth or teeth", "enamel", " uoride or uorosis", "China or Chinese", "prevalence or incidence" and "epidemiology". Moreover, potentially relevant articles that may be missed were identi ed through manual searching of reference lists. The systematic review and meta-analysis were reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidance (Moher et al. 2009).

Literature selection
Inclusion criteria: the articles selected in this study were according to the following inclusion criteria; (1) surveys done in mainland China (except for Taiwan, Hong Kong, and Macao); (2) were written in English or Chinese; (3) participants focused on children aged between 8 and 12 ; (4) reported prevalence and risk factors as primary articles; (5) provided full text.
Exclusion criteria: the articles excluded were: (1) letters to editors, conference abstract and presentation, case reports, master and doctoral theses; (2) researches based on the special populations (for example, based on patients); (3) articles were not fully accessed; (4) studies didn't report geographic locations.
Data extraction and quality assessment Study selection and data extraction were carried out by the rst and second author independently. In the rst stage, eligible citations were selected by screening the headlines and abstracts; in the second stage, the full-text analysis was performed. Cohen's Kappa was used to check the consistency between examiners. Disagreements were resolved by consensus and by the third author when necessary. For each study, the items of data extraction included the name of the rst author, year of publication, geographical location and territorial levels, survey date, sampling method, diagnostic criteria, the sample size (and proportion of DF patients), the prevalence of DF and DFI (dental uorosis index). If surveying date was not stated in the article, we assumed it as 2 years before publication.
We used the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) to assess the quality of included studies (Supplementary Checklist 1). This tool consists of 22 items (32 sub-items) that are considered necessary for excellent reporting of observational studies. Quality assessments of all included articles were assessed by answering 'yes or no' to each item in the STROBE checklist. The assessment was judged by two independent authors, and any discrepancy of their assessment results was resolved by consensus or by the third author.

Statistical analysis
Pooled estimates of prevalence and 95% con dence intervals (CI) were calculated using STATA software version 15.1. The heterogeneity was checked using Q-test and I 2test. The level of heterogeneity was low (I 2 <50% or P>0.1), a xed-effects model was adopted. Otherwise, the random-effects model would be selected. We used funnel plots and Begg's test to observe potential publication bias across the studies ( Fig.7), the result was considered to be signi cant when P≤0.05. In addition, subgroup analyses were performed based on gender, survey year, province and region of studies. Cochrane Review Manager (RevMan) version 5.3 was used to estimate the Relative Risk (RR) and 95% CI between different subgroups. In the sensitivity analysis, we eliminated one article individually and the summary results were not affected, indicating that the results were robust enough. Furthermore, studies were sorted into ve periods to observe the temporal trends of DF: 1995DF: -1999DF: , 2000DF: -2004DF: , 2005DF: -2009, 2010-2014, ≥2015. Finally, we created a prevalence map to re ect spatial distribution in mainland China using ArcGIS10.2 software.

Results
A total of 8663 records were initially retrieved by keywords search on Embase (n=1814), PubMed (n=2341), CNKI (n=1241), the Wan Fang (n=1787), the VIP (n=1398) and CBM (n=79). The second, third, fourth National Oral Health Survey of our country were included as well. After the exclusion of duplicate articles (2136) and irrelevant references (6527) from the pooled database by screening the title and abstracts, a full text review was conducted for 827 studies. In total, 786 articles did not meet the eligibility criteria, in which 86 did not state age clearly or age is not within 8-12, 3 were conducted in quota sampling, 30 studies did not have available data, 638 studies reported irrelevant information, 11 studies did not state survey site, 18 publications repeated survey time and sites. In the end, 41 studies are included in our systematic review and meta-analysis (Fig.1) Table S1). The description of the 41 included studies was summarized in Table 1. Among these, prevalence rates of DF varied from 3.1% to 86.5%. The highest prevalence rate (86.5%) was reported in Yunnan and lowest prevalence rate (3.1%) was reported in Guangxi. Dean method de ned by the World Health Organization (WHO) was used as the diagnostic criteria of dental uorosis in 40 studies and all 41 surveys recruited dentists and trained examiners as investigators. As for quality assessment, the number of positive answers ('yes') for the 32 listed items on the Strobe checklist for each study was at least 24 (Supplementary Table S2), which indicating that the quality of the 41 eligible studies was satisfactory.

Prevalence of DF in mainland China
Prevalence of DF over time. The overall estimates of DF prevalence in mainland China was 23.6% (95% CI: 19.3-28.0%, Table 2 Prevalence of DF by gender. A total of 17 articles reported prevalence of DF estimates by gender, yielding an overall prevalence of DF in boys and girls were 15.7% (95% CI: 11.9%-19.5%) and 15.2% (95% CI: 11.6%-18.8%) respectively. The nal pooled meta-analysis indicated no statistically signi cant difference in the prevalence of DF between the sexes (RR=1.05, 95% CI: 1.02-1.07, Fig.3).
Geographical distribution of DF prevalence in mainland China Figure 6 demonstrates a color-coded map of the distribution of the dental uorosis in mainland China (data available from most provinces, except Tibet). Five different color distribution areas were created on the map based on the prevalence of DF. And signi cant variation was observed for geographical region across provinces. The highest prevalence zone showed on the map in the darkest red, including Tianjin (39.9%), Chongqing (41.8%) and Jiangsu (42.5%). And the prevalence of DF in Guangdong (36.5%) and Inner Mongolia (33.0%) was substantial high, as well as in Qinghai, Hebei, Shandong. Other provinces show low prevalence relatively.

Discussion
Dental uorosis acting as the earliest indicator of excess uoride exposure in population was an important public oral health issue around the world (Idon &Enabulele 2018). The meta-analysis is the rst of these kinds to estimate the epidemiological characteristics of dental uorosis in mainland China.
In this meta-analysis, the prevalence of dental uorosis has increased considerably from 2009 to 2018, while a reduction of prevalence during 2005-2009 was shown in our study. The possible reason is more likely due to the implementation of the uorine reduction and water improvement policy in 1997 (Lu et al. 2011). As the result of meta-analysis shows, the prevalence of dental uorosis has decreased signi cantly since 2005. It suggests that the prevalence of dental uorosis decreased obviously after eight years' water improvement, which is consistent with other studies , Sun et al. 2017).
Meanwhile, the overall results of the 7 studies included demonstrated signi cant reduction in DF prevalence after water improvement. Hence, the uoride reduction policy is the most effective measure to reduce the prevalence of dental uorosis. Moreover, thanks to researchers report on dental uorosis, the government has paid more and more attention to oral hygiene and adopts corresponding anti-uorination measures actively.
Nevertheless, it was worth noting that there is an ascent trend of prevalence during 2010-2014. Two possible reasons can explain this trend: the failure of water improvement project and the increase in rate of reports on the prevalence of DF. Firstly, during the implementation of the water improvement project, the water modi cation equipment was damaged gradually, which caused the leakage of high-uorine water (Yin et al. 2017). In a study conducted in Zhoukou (a city in Henan province) (Hao 2010), 57.25% of the water improvement project were damaged or scrapped, which nally resulted in a water uoride exceeding rate of 34.84%. Similar situations have been reported in other studies (Cao et al. 2009, Qiao et al. 2014, Zhao et al. 2013. Secondly, in recent decades, more and more researchers investigated the prevalence of dental uorosis in various regions of China. Relevant reports increased dramatically. Besides, the examiner's diagnosis of dental uorosis might be more precise than before, which can diagnose more patients with DF. Thus, our research results show an ascent trend eventually. The continued increase in uorosis rates in mainland China indicates that additional measures need to be implemented to reduce the prevalence. Of our studies conducted, people in rural areas are more likely to suffer from dental uorosis than those in urban. The possible reason is that the waste gas from coal burning enters into drinking water or crops (Yao &Wang 2016), and another is that rural residents consumed water from the well or groundwater with high uorine content ).
The geographical map shows signi cant regional differences. As the map shows, the prevalence rates of DF in Tianjin, Chongqing and Jiangsu are relatively high. In China, endemic uorosis mainly includes drinking water-type endemic uorosis, coal-burning endemic uorosis and tea-type endemic uorosis. Drinking water-type endemic uorosis keeps main position (Chen &Yu 2013). Research also indicates that dental uorosis in Tianjin is a drinking water-type of endemic uorosis with a serious historical conditions (Sun et al. 2001). In spite of good results got from water improvement, the prevalence of dental uorosis remains high because the water improvement project has not been working properly in recent years (Cui et al. 2017). The disease situation in some areas is rather complicated like Chongqing. Dental uorosis in Chongqing is both drinking water-type and coal-burning type. The combined effect of two different causes exacerbates the prevalence of dental uorosis. In the light of the above reasons, it suggests that the government should control the disease according to local conditions. For drinking-water type endemic uorosis, the policy of water improvement should be implemented actively. In drinking teatype uorosis areas, government should strengthen anti-uoride education to eliminate the harm of drinking brick tea. For coal-burning uorosis areas, the stove should be changed to reduce coal combustion.
Several limitations should be considered into our ndings. Firstly, the literature reporting on dental uorosis in China was insu cient, even though we nally captured 41 articles in our meta-analysis. Secondly, some areas like Tibet lack relevant articles for prevalence of DF. Hence, we are not able to assess the prevalence of DF among the general populations of China. Moreover, a high level of heterogeneity was observed between studies. That might be attributed to eating habits, family economic status and genetic susceptibility, but we did not have enough data to explain. Finally, the diagnosis of dental uorosis was mainly dependent on the investigators' pro ciency. In another word, bias from investigators still exist. For example, mild dental uorosis and enamel hypoplasia are di cult to differentiate. We hope that there will be more epidemiological studies on DF in China.

Conclusion
Although it is not possible to determine the prevalence of dental uorosis in mainland China through these studies; nonetheless, the epidemiological characteristics of dental uorosis in mainland China increased rst and then decreased cannot be ignored. And ndings show that the prevalence of dental uorosis has been greatly reduced after the water improvement policy. Result of this meta-analysis provides evidence enable governments taking effective measures to control dental uorosis. The government should continue to strengthen the uoride reduction and water reform policy, and strengthen the maintenance of water improvement engineering equipment.

Declarations
Authors' contributions Xiaonan Zhang and Yuzhen Zhan contributed to the study conception and design. Study material preparation, data collection and analysis were performed by Yuzhen Zhan, Lianjie Xiong, Zhoujie Gong, Ting Xu, Xiaonan Zhang. All authors have written, read and approved this manuscript for submission.
Funding This work was supported by Chongqing Medical Reserve Talent Studio for Young People, Grant/Award Number: ZQNYXGDRCGZS2019004.
Data availability All data generated or analyzed during this study are included in this published article (and its supplementary information les).

Compliance with ethical standards
Con icts of interest The authors declare that they have no con ict of interest.
Ethical approval Not applicable.
Content to participate Not applicable.
Content to publish Not applicable.

Figure 2
Temporal trends of the prevalence of dental uorosis in mainland China during 1995-2020.

Figure 3
Forest plot of dental uorosis among different gender in mainland China during 1995-2020. Forest plot of dental uorosis among rural and urban areas of mainland China during 1995-2020. Forest plot of dental uorosis among WIB (water improvement before) and WIA (water improvement after) of mainland China during 1995-2020. Figure 6