It has been reported that major studies in China examining tooth loss have been "uncommon and mainly conducted in urban areas," and have generally involved the elderly [3
]. A survey from Chengdu city reported that 19.7% of persons aged 60 years or older were edentulous and only 4.8% had all of their natural 28 teeth [41
]. Although the mean number of missing teeth among rural residents aged 65–74 years was 11.6 in a south China survey, only 3.4% were edentulous [42
]. Findings from the 2nd
Chinese National Oral Health Survey indicated that 10.5% of participants aged 65–74 years were edentulous, with rural residents retaining 16.8 natural teeth compared to 18.8 natural teeth for persons living in urban areas [4
]. Based on 28 natural teeth, tooth retention for this rural Linzhou study group was very similar to the national estimates for rural residents (17.1 vs. 16.8). Although tooth retention was similar in magnitude to national estimates in China, the prevalence of edentulism in the Linzhou study group for 56–67 year-olds was greater (16.3%).
Edentulism among the Linzhou study participants was significantly lower in 40–55 year-olds compared to the 56–67 year-olds (11.7% vs. 24.4%). It has been reported that edentulism in the Beijing area was "rarely seen" in persons younger than 60 years-old, and in Hong Kong it was "not apparent until after age 55" and was seen at a much reduced rate compared to many regions of the developed world [43
]. Complete tooth loss (approximately 17%) was higher in this Chinese study group than recent reports from the United States, which showed that only 5% of persons aged 40–59 years were edentulous [45
Older adults who are cigarette smokers are more likely to have greater tooth loss, fewer functional contacts, and greater tooth loss due to periodontal disease [46
]. In the United States, it is estimated that among current smokers, nearly 75% of periodontal disease is attributable to smoking [49
]. Findings from a 1996 national survey in China indicate that smoking was more prevalent among men (63%) compared to women (3.8%) [50
]. Approximately 20% of the Linzhou study group was identified as current smokers and among those, 98% were men (data not shown). We found that the association between smoking status and the mean number of retained dental roots was statistically significant, but not clinically relevant (1.0 retained roots for smokers vs. 1.4 retained roots for non-smokers). Smoking status was not associated with the calculated mean number of adjusted missing teeth in this study group of 40–67 year-olds. Furthermore, unlike in the United States, we found that smoking was not associated with edentulism and tooth loss in a dentition excluding third molars.
Findings from the Linzhou study group showed that women, older individuals, and persons with lower educational attainment had higher DMFT scores. The mean DMFT scores for men and women with dental caries in this study group were 7.6 and 9.3 respectively. Results for 35–44 year-olds and 65–74 year-olds who participated in the 2nd
National Oral Health Survey in China were 1.7 and 11.6 for men and 2.6 and 13.0 for women respectively [4
]. Overall, our findings suggest that women from the Linzhou study population experience more coronal and root decay compared to men, and women are more likely to have retained residual dental roots and more teeth decayed or filled.
Survey results from the 2nd
National OH Survey and a comprehensive oral health survey conducted in Guangdong Province in 1997 indicated that caries experience was slightly more prevalent for adults in rural areas compared to those living in urban areas [4
]. Overall, findings from the 2nd
National OH Survey indicated that the prevalence of dental caries among rural Chinese was 60% for 35–44 year-olds and 64% for 65–74 year-olds [4
]. The prevalence of dental caries among this rural study group aged 40–67 years was approximately 68%.
Lin and coworkers reported that studies measuring root surface caries are quite uncommon in China [51
]. They reported a prevalence of 38% among the 65–74 year-olds, whereas findings from 56–67 year-olds in our Linzhou study group showed a prevalence of 48.5%. Chinese studies reporting mean residually retained roots are even rarer. Lo and coworkers reported a mean retention of 0.5 residual dental roots for 65–74 year-olds residing in Hong Kong [52
], whereas the findings in 56–67 year-olds in our study showed a mean retention of 1.4 dental roots. It should be noted that differences exist between the WHO and NHANES methodologies for classifying dental roots.
From 1975–76, a study was conducted on 1084 adults aged 21–50 years-old and 1400 school-aged children in Linxian (now known as Linzhou) to examine the relationship between dental caries and urine fluoride levels. They reported that dental caries prevalence was high (80.7%) and that oral hygiene was poor among the adults examined [53
]. Because the methodologies used in this earlier Chinese study are not comparable to the WHO or the NHANES caries assessment protocols, comparisons are very problematic. The caries diagnostic methods described assessed for caries using a 5-point graduated scale with "primary caries" limited only to caries in enamel. As the diagnostic criteria increased along this scale, caries in level 4 were defined as deep dentinal caries with likely pulpal involvement and level 5 caries were limited to retained dental roots.
General observations from the Linzhou study group showed that men, older individuals, persons with higher educational attainment, and smokers had worse periodontal health. Using a CPITN/CPI score of "4" as reflective of an advanced periodontal pocket (i.e., 6 mm or greater), studies conducted in China within the past decade have reported the percentage of individuals with at least one advanced periodontal pocket to range from 4–25% among 65–74 year-olds living in Shanghai, Hong Kong, Hubei, and Guangdong [54
]. Findings from the 2nd
National Oral Health Survey indicated that 3.8% of Chinese aged 65–74 years had advanced periodontal pocketing [4
]. In the Linzhou study group, the prevalence of periodontal pockets measured at 6 mm or greater among persons aged 56–67 years-old was 9.8% (data not shown).
Using the WHO attachment loss categories [2
], Corbet and coworkers reported that among those 65–74 years-of-age, 55% of rural persons and 48% of urban persons had at least one site of attachment loss measured at 6 mm or more [57
]. Moreover, the authors reported that unlike advanced periodontal pocket findings, attachment loss among this older age group was consistent with findings from Hong Kong and Beijing. More than 39% of the CSS2 participants from rural Linzhou aged 56–67 years-old were found to have at least one site with 6 mm or more of attachment loss.
The periodontal health status of rural residents in China has not been extensively researched. Rural adults were included in a comprehensive oral health survey conducted in Guangdong Province in 1997 using WHO criteria [5
]. Results indicated that among rural adults, 42% of 35–44 year-olds had considerable attachment loss for age (minimum 2 sextants w/4+ mm) and 38% of 65–74 year-olds had considerable attachment loss for age (minimum 2 sextants w/6+ mm) [54
]. Moreover, attachment loss was slightly more prevalent for adults in rural areas compared to those living in urban areas. However, an earlier survey conducted in 1990 in Guangdong Province using WHO criteria reported that the prevalence of pocket depth of 4 mm or more was greater among urban residents compared to those living in rural areas [7
]. Results from the 2nd
National Oral Health Survey showed that 14% of rural 35–44 year-olds had "shallow" pocket depths (4–5 mm) and 2.2% had "deep" pockets (6+ mm)[4
]. Moreover, the pocket depth prevalence increased among the 65–74 year-old cohort to approximately 18% and 4% for shallow and deep pockets respectively. There was little difference in pocket depth prevalence between urban and rural adults.
Although clinical loss of attachment is more difficult to measure than pocket depth, attachment loss measures do provide an improved overall estimate of adverse periodontal status [58
]. Attachment loss is a historical reflection of periodontal damage throughout a lifetime and does not necessarily indicate that periodontitis is active at the time of assessment. Moreover, delineations between moderate or severe periodontitis have been somewhat fluid. For instance, using data from NHANES III, a variety of researchers have suggested that severe periodontitis is characterized by having at least one tooth with a pocket depth measure of 6 mm or greater whereas others have defined advanced periodontitis as having at least one tooth with 3 mm or greater attachment loss; moderate periodontitis has been defined as having at least one site with pocket depths between 4–5 mm [59
Using data from NHANES III (1988–94) to investigate acute-phase inflammatory markers and periodontitis, Slade has suggested that pocket depth is most likely associated with active periodontal disease [62
]. Arbes and coworkers classified individuals with at least 1 periodontal site with concomitant measures of 3 mm of attachment loss and 4 mm of pocket depth as having active periodontal disease [40
]. For purposes of this study, we defined "moderate-to-advanced" periodontitis at a lower threshold case definition for disease using Arbes and coworkers' criteria. We also classified individuals with "severe" periodontitis at a higher threshold case definition requiring individuals to have at least one periodontal site with concomitant measures of 4 mm of attachment loss and 5 mm of pocket depth. Although, men were more likely to have more sites of advanced attachment loss compared to women, there was little gender difference in the prevalence of periodontitis at either the lower or upper case definitions for the Linzhou study group.
In this study, NHANES oral health protocols were used in a larger study which was evaluating methods for early detection of esophageal squamous cell cardinoma. ESCC is among the least curable cancers, with 5-year survival rates estimated to be around 10% [63
], whereas 5-year survival rates for oral cancer are approximately 50% [64
]. The relationship between oral health and risk factors for oral-pharyngeal cancers are well known [65
]. Although in many parts of the world the main determinants for both oral cancer and ESCC are tobacco use and alcohol consumption [65
], the linkage between oral health and ESCC is unclear. In this high risk area of China, the primary risk factors for ESCC are poor nutrition, family history, and other unknown risk factors [67
], and oral cancer is uncommon.
More than 20 years ago, a relationship between microorganisms in the oral cavity and those isolated from esophageal carcinomas was suggested [68
]. Advances in the understanding of oral microbial ecology have recently supported the notion that dental plaque is an oral biofilm [69
] and bacterial biofilms are a common source for chronic infections [70
]. Poor oral hygiene contributes to the development of dental plaque, periodontitis, and dental caries [71
]. It has been postulated that poor oral hygiene mediates bacterial load and that "overgrowth" of specific bacterial types on teeth may explain many associations seen between poor dental health and systemic diseases [72
]. Poor oral hygiene may also contribute to the formation of nitrosamines in the oral cavity [73
] and intra-oral nitrate-reducing activity may contribute to the majority of overall nitrosamine exposure in humans [74
]. Because nitrosamines are known carcinogens, untreated caries and periodontal disease should be important considerations when exploring cancer risk factors in this study population, which has very high rates of esophageal cancer. After adjusting for a number of potential confounders, the results from our Linzhou study group indicated that poor oral health was associated with esophageal dysplasia (OR = 1.59; 95% CI 1.06, 2.39).
For this report, poor oral health was derived from periodontal status and caries experience. Caries experience (DMFT) is a composite measure comprising of untreated caries (decay), treated caries (fillings), and tooth loss due to disease. Although preliminary findings indicated that DMFT was related to the presence of esophageal dysplasia, subsequent findings showed that this relationship was driven by the missing component (M) of this index. An earlier report from this same region of China suggested that tooth loss was associated with incident esophageal and gastric cancers [24
]. In a fuller exploration of risk factors for squamous esophageal dysplasia, a multivariate model also suggested that subjects who had lost more, but not all, of their teeth had a higher prevalence of dysplasia [25
]. Tooth loss also has been reported to be associated with oral leukoplakia and oral cancer [26
], with pancreatic cancer [75
], with gastric cancer in smokers [76
], and with esophageal dysplasia [77
There are some study issues to consider that affect the interpretation of our results. The CSS2 study group does not correspond to a random sample of the general population in the Linzhou area. With a nonsystematic random selection of participants, there may have been a potential bias towards the selection of less healthy study participants. Another potential bias may have been the selection of individuals with certain demographic characteristics. Except for a slightly lower male-to-female ratio, the socio-demographic characteristics of the CSS2 study group were comparable with the general population in this region of China. Nevertheless, the external validity of the CSS2 findings is limited.
Partial-mouth exams, like the NHANES periodontal assessment, may underestimate sites with periodontal disease, particularly the more severe conditions [78
]. The magnitude of underestimation of disease using any partial-mouth exam is dependent upon the prevalence of the disease in the group under study. When prevalence is low, the degree of underestimation is greater. However, CPI will also underestimate disease severity[81
] and the use of CPI to assess for periodontal disease status is controversial [82
There are additional assumptions and caveats underlying the statistical methods we elected to use. Although the study group was not a systematic random sample, we used parametric statistical tests to test for differences between subgroups in the CSS2 sample. We made this determination following an evaluation for data quality. We also performed non-parametric statistical tests and found that overall estimates and measures of significance were similar to those produced by parametric testing.
There are numerous advantages to using NHANES oral health assessments protocols in a clinical study: there is enhanced comparability of findings across a variety of study designs; there is improved documentation of baseline oral health status for longitudinal studies investigating the development of oral diseases and conditions; and there are enhanced options for aggregating collected data for testing different concepts of risk indicator exposure/outcomes. For similar reasons, the Women's Interagency HIV Study (WIHS), a multi-center clinical study in the U.S., uses NHANES oral health protocols [83
]. The NHANES methodology may not be suitable for all dental research activities, such as some clinical trials, but these methods should be suitable for exploring many systemic interrelationships.