Introduction

Recently, the effect of oral health on systemic disease has attracted growing attention. Many studies have reported that periodontal disease is a risk factor for cardiovascular disease [1], stroke [2], diabetes mellitus [3], preterm low birth weight [4], osteoporosis [5] and aspiration pneumonia [6].

Meanwhile, deaths from pneumonia have gradually increased since the 1980s in Japan, ranking third among causes of Japanese mortality in 2015 [7]. Because the mortality rate from pneumonia increases rapidly with increasing age [8], the recent rise in mortality from pneumonia is considered to result from aging of Japanese population. In response to the increase in pneumonia deaths in Japan, pneumococcal vaccine was added to the list of routine vaccinations in that country in October 2014 [9]. Additionally, the Japanese Respiratory Society has called for a “stop pneumonia campaign”, which emphasizes oral care by health care workers as a preventive measure for pneumonia [10], together with vaccination and improvement in nutritional status. Many groups, including academic societies, have also publicized the importance of oral care in pneumonia [11], and its significance is now more widely understood. Many facilities for the elderly currently provide oral care not only to improve swallowing and nutrition, but also to prevent pneumonia [12]. The World Health Organization has also reported that respiratory infections including pneumonia are the fourth leading cause of death in the world [13]. Pneumonia is considered an important cause of morbidity and mortality in the USA, Canada, Europe and Latin America [14].

Many intervention studies have been conducted among inpatients and patients in nursing homes investigating the relationship between oral hygiene and pneumonia [15, 16, 17]. The oral bacteria propagated in the periodontal region may drop into the lung and increase the risk of pneumonia [18, 19, 20]. However, oral disorders may have a greater influence on the risk of pneumonia among inpatients or patients in nursing homes (because of dysphagia and deteriorated immune systems) than among community-dwelling people. Nevertheless, few cohort studies have been conducted in community-dwelling individuals [21, 22], and whether the findings among inpatients or patients in nursing homes are applicable to the general population is still unclear. We therefore investigated the association of tooth loss with mortality from pneumonia in a prospective study of Japanese dentists.

Materials and methods

Results

Table 1 shows the baseline characteristics of participants according to the number of teeth lost. Participants with more lost teeth tended to be older and leaner, and more likely to have smoked and to have a history of diabetes. There was a higher proportion of women in the edentulous category.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Baseline characteristics of participants by number of teeth lost.

https://doi.org/10.1371/journal.pone.0195813.t001

During the median follow-up of 9.5 years, we documented 68 deaths from pneumonia. Participants who were edentulous at baseline were at a significantly increased risk of mortality from pneumonia (Table 2): the multivariable-adjusted HRs were 2.07 (95% confidence interval [CI], 1.09–3.95) for the edentulous and 1.60 (95% CI, 0.83–3.10) for loss of 15–27 teeth relative to loss of 0–14 teeth (trend p = 0.026). Comparing loss of 10–19 and > 20 teeth with loss of 0–9 teeth gave multivariable-adjusted HRs of 2.68 (95% CI, 1.31–5.51) and 2.39 (95% CI, 1.22–4.67), respectively (trend p = 0.015). Overall, the risk of mortality from pneumonia was associated with the number of teeth lost, as shown by the HR per tooth (HR, 1.031; 95% CI, 1.004–1.060). When we restricted the analysis to participants aged 50 years or older, the findings were not substantially altered (S1 Table, median follow-up period: 9.5 years).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Hazard ratios (HRs) for mortality from pneumonia by number of teeth lost.

https://doi.org/10.1371/journal.pone.0195813.t002

Discussion

In the present study, the number of teeth lost was positively associated with the risk of pneumonia mortality in a community-dwelling population. The risk remained significant after adjustment for possible confounding factors including BMI, smoking status and diabetes.

Several underlying mechanisms have been proposed regarding associations between tooth loss and pneumonia. The possible pathway most frequently discussed is that the increase in cytokines induced by inflammation in periodontal disease may promote systemic inflammation including the lung [19, 29, 30]. Interleukin-1 and interleukin-6 levels have been assessed as inflammatory markers in patients with periodontal disease [31, 32, 33]. In the Japanese population, periodontal disease is one of the main causes of tooth extraction and accounts for 41.8% of extractions. In the middle-aged group (55–64 years), the disease accounted for nearly 60% of teeth extracted [34]. Thus, periodontal disease may explain the association of tooth loss with pneumonia risk. Many studies have reported that some bacteria involved in periodontal disease are aspirated into the lung [29]. Periodontal disease bacteria such as Porphyromonas gingivalis and Actinobacillus actinomycetemcomitans have been identified in the lungs of patients with pneumonia [32]. Additionally, periodontal disease-associated enzymes in saliva might modify the mucosal surface of the lung, and destroy its pellicles [35]. Furthermore, the reduced masticatory function caused by tooth loss may result in a limited selection of foods [36], and insufficient nutrition might increase the risk of pneumonia [37, 38, 39]. In our previous study, we found that the intake of nutrients and food groups (such as carotene, vitamins A and C, milk and dairy products, and vegetables) declined and carbohydrate intake increased as the number of teeth decreased. This finding suggested that tooth loss leads to nutritional imbalance [40]. In this study, the pathway through masticatory function and nutrition might have more strongly influenced the risk of pneumonia than the inflammation pathway because the HRs were increased in edentulous participants, who would be less affected by inflammation. Retaining teeth might be important for the prevention of pneumonia. Alternatively, tooth loss could serve as a major indicator of aging that is linked to pneumonia risk [21, 34, 41].

Our findings suggest that tooth loss is a predictor of mortality from pneumonia. Because tooth loss is the most apparent indicator of oral health for the general public, it may be useful to presume that individuals with tooth loss may be predisposed to severe pneumonia, enabling preventive action to be taken. Although the dose–response relationship was not clearly shown between the number of teeth lost and the risk of death from pneumonia, we found an overall positive association. The increase in the risk of pneumonia mortality roughly corresponds to an increase by 15.6 per 100,000 population per year in the edentulous compared to those with 0–14 teeth lost ([2.07–1] × [26/177,988] × 100,000). Although this absolute increase may not be large, it was substantial in the elderly aged 70 years or older: 209 per 100,000 per year ([1.99–1] × [17/8044] × 100,000. The HR for the edentulous compared with those with 0–14 teeth lost was 1.99 [95% CI 1.01–3.90]).

Although several investigations reported associations between indicators of periodontal disease and pneumonia [6, 42, 43], only a few examined the number of present or lost teeth in relation to pneumonia [6, 22]. In a cohort study by Awano et al., the risk of mortality from pneumonia was 3.9 times higher in participants with 10 or more teeth with a probing depth (PD) exceeding 4 mm [22]. However, the number of teeth was not significantly correlated with mortality from pneumonia, which is not consistent with our results. Possible reasons are that their participants were much older than ours (all participants were 80 years old at baseline) and may have retained fewer teeth. For individuals retaining a certain number of teeth, there may be an inverse correlation between number of teeth and risk of pneumonia mortality through infection by periodontal disease. This risk, however, might be decreased in the edentulous, and complete dentures could make mastication easier than several wobbly teeth affected by periodontal disease. It is therefore possible that their study did not detect the association of tooth loss with pneumonia mortality. Additionally, they included participants with a history of stroke and cancer. Pneumonia deaths associated with underlying diseases and aging might have attenuated the association of tooth loss with pneumonia mortality.

The present study has several strengths, including the prospective cohort design and adjustment for possible confounding factors. Many previous studies reported associations of oral conditions with aspiration pneumonia [6, 44]. By excluding aspiration pneumonia from endpoints in this study, we concluded that mortality from community-acquired pneumonia was also related to oral health, even in a healthy population.

However, some limitations should be considered. First, we did not establish the cause of tooth loss (e.g., periodontal disease, dental caries). Nevertheless, the number of teeth lost is the most obvious indicator of oral condition. Second, the number of teeth lost was self-reported. However, self-reported counting of teeth lost is considered to be sufficiently valid [45]. In the present study, we expected the number reported by dentists to be more accurate than in the general population. Third, we did not consider temporal changes in the number of teeth lost and covariates. Fourth, no information was available on the history of lung disease, so this could not be considered as a variable in the analysis. Finally, the participants consisted of members of a dental association and the response rate was relatively low (36.2%), which may limit the generalizability of the results. The cohort of dentists may not be representative of a community population. The meaning of losing teeth may differ between a dentist and a member of the general population; the loss of teeth in dentists would have occurred despite sufficient oral care and dental treatment. However, the underlying biological mechanisms would be common among human populations, and the population in the current study was homogeneous in terms of socioeconomic status, which may reduce confounding by socioeconomic factors. Although the healthy worker effect should be taken into account [46], cohort studies of health care workers have provided important health information, as in the Nurses’ Health Study and the Health Professionals Follow-Up Study [47, 48].

In conclusion, a high number of teeth lost may indicate an increased risk of mortality from pneumonia in community-dwelling populations.

Supporting information

Acknowledgments

We thank the prefectural Dental Associations of Japan for their cooperation in establishing and following the cohort. We thank Dr Takashi Sakai and Dr Katsunori Okuda (former directors of the JDA) for greatly contributing to the initiation of this cohort study. We also thank Ms Yuko Watanabe and Ms Etsuko Kimura for technical assistance.