Tooth Loss and 12‐Year Mortality Risk in 8494 Older Adults From Ireland

ABSTRACT

Objectives

To investigate whether tooth loss is associated with all‐cause and cause‐specific mortality in older Irish adults.

Methods

A total of 8494 participants from The Irish Longitudinal Study on Ageing (TILDA) were included. Survey data were linked to death registration records, covering individuals who participated in TILDA Wave 1 (2009/2010) and died by 31st January 2022. Cox proportional hazards regressions and competing risk survival analyses were employed to examine the longitudinal relationship between tooth loss and both all‐cause and cause‐specific mortality.

Results

The mean age of participants at baseline was 63.2 years (SD 10.2). Among the cohort, 3951 (46.5%) were categorized as “Dentate, no denture,” 3041 (35.8%) as “Dentate, with denture(s),” and 1502 (17.7%) as “Edentulous.” Over a median follow‐up of 12 years, 1430 (16.8%) participants died. After adjusting for confounders, edentulous participants had a significantly higher hazard ratio (HR) for all‐cause mortality compared to dentate participants with no dentures (HR = 1.42, 95% CI 1.23–1.65, p < 0.001). For cause‐specific mortality, edentulism had the greatest sub‐distribution hazard ratio (SHR) with respiratory mortality (SHR = 1.57, 95% CI 1.03–2.41, p = 0.04), followed by cancer mortality (SHR = 1.31, 95% CI 1.01–1.71, p = 0.04). There was a nonsignificant association with cardiovascular mortality (SHR = 1.25, 95% CI 0.96–1.63, p = 0.10).

Conclusions

Edentulism was independently associated with all‐cause mortality in a cohort of 8494 men and women from Ireland. Edentulism was significantly associated with respiratory and cancer mortality.

Summary.

• Key points

  • ○
    This large prospective cohort study of 8494 older Irish adults provides valuable insights into the impact of tooth loss on mortality risk, highlighting the potential role of oral health in broader health outcomes.
  • ○
    Edentulism was significantly associated with an increased risk of all‐cause mortality, particularly respiratory and cancer mortality.
  • ○
    Despite significant associations with respiratory and cancer mortality, there was a nonsignificant association with cardiovascular mortality, suggesting differential effects of tooth loss on specific causes of death.
• Why does this paper matter?

  • ○
    This study establishes tooth loss, particularly edentulism, as an important risk indicator for all‐cause mortality in older Irish adults, with strong associations found for respiratory and cancer‐related deaths.
  • ○
    It underscores the significance of oral health as a key factor in predicting long‐term health outcomes, contributing critical evidence to the understanding of how tooth loss may influence mortality risk in aging populations.

1. Introduction

Tooth loss represents the culmination of two primary dental diseases: caries and periodontitis. Both conditions are major contributors to oral health decline, leading to functional and aesthetic impairments that significantly affect individuals' quality of life and self‐esteem [1, 2]. Complete tooth loss, or edentulism, is particularly debilitating and is linked to increased socio‐economic burdens and healthcare costs [3]. In Ireland, around 18% of adults aged 55 and older are edentulous [4]. Despite a global decline in edentulism rates, tooth loss remains disproportionately high in older adults, especially those from disadvantaged backgrounds [5]. The Global Burden of Disease study in 2015 attributed 7.6 million disability‐adjusted life years (DALYs) to edentulism, highlighting its impact on health [6].

Tooth loss is increasingly recognized as a predictor of mortality, especially among older populations [7, 8, 9]. The mechanisms underlying the association between tooth loss and mortality are complex. Several studies have suggested that individuals with significant tooth loss may be at an increased risk of mortality due to associations with chronic diseases such as cardiovascular disease, diabetes, and respiratory illnesses or underlying systemic inflammation [10, 11, 12]. Moreover, periodontitis, often the underlying cause of tooth loss, has itself been repeatedly implicated in the pathogenesis of systemic diseases, including atherosclerosis and diabetes, through its chronic inflammatory burden [13, 14]. This sustained inflammatory state may accelerate vascular changes and exacerbate metabolic dysregulation, thereby contributing to an elevated risk of mortality in affected individuals [15]. Compromised masticatory function may also lead to inadequate dietary intake, resulting in malnutrition, which is particularly concerning in older adults [16]. Confounding factors like smoking, a well‐known risk factor for both periodontal disease and tooth loss [17], as well as reduced life expectancy [18], complicate the relationship further. Furthermore, tooth loss is not merely a health issue but also a marker of social determinants of health, such as socio‐economic status (SES), education, and access to healthcare, which are crucial factors in overall mortality risk [19].

Despite evidence linking tooth loss to mortality, gaps remain in understanding this association. While research has demonstrated a relationship between tooth loss and mortality in certain populations, findings cannot be directly generalized to Ireland, where healthcare systems and socio‐economic factors differ. Additionally, many studies fail to adjust adequately for confounders such as SES, comorbidities, and health behaviors, which influence both tooth loss and mortality [7]. Short follow‐up periods may also underestimate the long‐term risks associated with tooth loss. Accurate mortality recording is also important, particularly in capturing cause‐specific outcomes like cardiovascular disease, cancer, and respiratory conditions. Studies relying on broad registries or self‐reported data often miss important nuances. Using standardized methods like International Classification of Diseases (ICD) codes for cause of death can enhance the precision of mortality data.

The aim of this study, therefore, was to investigate the association between self‐reported tooth loss and 12‐year mortality risk in a cohort of 8494 older adults from Ireland. By analyzing a large, representative sample and controlling for a wide range of confounding factors, this research seeks to determine whether self‐reported tooth loss can serve as a reliable predictor of mortality risk, with a focus on cause‐specific mortality outcomes.

2. Methodology

2.1. Population

Subjects were recruited from The Irish Longitudinal Study of Aging (TILDA), a nationally representative, large prospective cohort study investigating the social, economic, and health circumstances of community‐dwelling adults aged 50 years and older in Ireland. Baseline assessments, known as Wave 1, took place between October 2009 and February 2011, with 8507 participants. Detailed descriptions of TILDA's design, including its survey methodology and weighting scheme, have been previously published [20, 21]. In brief, the study uses a clustered, stratified random sample drawn from the Irish Geodirectory, a comprehensive database of all residential addresses in the Republic of Ireland, compiled by the Irish Postal Service and Ordnance Survey Ireland. The sample was selected using the RANSAM procedure [22], a multi‐stage probability sampling method. Data collection includes three components: a computer‐assisted personal interview (CAPI) conducted by trained social interviewers in participants' homes, a self‐completion questionnaire, and a comprehensive health assessment at a dedicated health center, including blood sampling. Successive biennial waves of data collection cover a broad range of topics, including health status, healthcare utilization, and participants' demographic, social, and economic circumstances.

Ethical approval was obtained from the Research Ethics Committee of the Faculty of Health Sciences, Trinity College Dublin. Participation was voluntary, and all participants provided written informed consent. The study adhered to the principles of the Declaration of Helsinki.

2.2. Self‐Reported Tooth Loss

Tooth loss was assessed using a self‐reported measure during the CAPI. Participants were asked the question, “Which best describes the teeth you have?”, with responses categorized into the following groups:

1. I have all my own natural teeth—none missing: Participants reported having all of their natural teeth with no missing teeth.

2. I have my own teeth, no dentures—but some are missing: Participants reported missing some natural teeth but had no dentures.

3. I have dentures as well as some of my own teeth: Participants reported having both some natural teeth and dentures.

4. I have full dentures: Participants reported wearing full dentures and having no remaining natural teeth.

5. I have no teeth or dentures: Participants reported having neither natural teeth nor dentures.

To facilitate statistical analysis and improve model stability, the response categories were recoded into three main groups, reducing the risk of unreliable estimates from smaller categories and increasing statistical power by creating more evenly distributed groups.

• “Dentate, no denture”: Categories 1 and 2 were combined to represent participants who still had natural teeth, either all or with some missing. This grouping reflects participants with preserved natural dentition.

• “Dentate, with denture(s)”: Category 3 remained as a distinct group, representing participants who had both some natural teeth and dentures. This intermediate group reflects partial tooth loss but retained functionality with prosthetics.

• “Edentulous”: Categories 4 and 5 were combined to represent participants with no natural teeth (edentulous), whether or not they used dentures. This grouping reflects complete tooth loss.

A sensitivity analysis using the original categories was also performed in Supporting Information.

2.3. Mortality

In the Republic of Ireland, all deaths must be registered with the General Register Office (GRO), ensuring comprehensive and reliable mortality data. Legal requirements for registration, combined with the necessity of a death certificate for legal purposes, make non‐registration exceptionally rare. TILDA obtained approval from the GRO to link study participants with their corresponding death certificate information, enabling accurate mortality tracking. As there is no unique personal identifier in Ireland, TILDA participants who died were matched to their official death certificates using name, address, and month/year of birth (with age included to allow for potential misreporting). Where these details were insufficient, additional information such as marital status was utilized. This linkage process covered all individuals who attended TILDA Wave 1 (2009/2010) and died up to 31st January 2022. Death certificates provided data on the date and underlying cause of death, with causes classified according to the 10th Revision of the ICD (ICD‐10). More information about the specific procedures for linking the data can be found in a separate description [23].

For this study, the primary outcome was all‐cause mortality. Secondary outcomes focused on cause‐specific mortality, with particular attention to cardiovascular disease, respiratory disease, and cancer. These were selected as they represent the three most common causes of death among older adults in Ireland [24], and each has plausible biological pathways linking them to oral health. Cause‐specific deaths were categorized using ICD‐10 codes: cardiovascular disease (ICD‐10 codes I00‐I99), respiratory disease (codes J00‐J99), cancer (codes C00‐C97), and other causes. Participants who were still alive by 31st January 2022, or who were lost to follow‐up, were right censored in the analysis.

2.4. Covariates

Age (years) was used as a continuous variable in the statistical modeling. Sex as male/female. Body weight and height were measured using standard procedures during the health assessment. Body mass index (BMI) was calculated as weight/height² (kg/m²). Smoking behavior was divided into three categories: never smoked, former smoker and current smoker. Problem alcohol consumption habits were assessed using the CAGE questionnaire [25]. Co‐morbidity variables included the presence of various cardiovascular conditions such as a history of angina, heart attack, heart failure, stroke, or transient ischemic attack. Due to the relatively low frequency of these conditions, the cardiovascular conditions data were collapsed into a single variable representing the “number of cardiovascular diseases,” categorized as 0, 1–2, and ≥ 3. Participants also self‐reported a history of a doctor's diagnosis of diabetes. Use of antihypertensive medication was coded using the anatomical therapeutic chemical classification (ATC), which categorized medications into classes such as antihypertensive medications (ATC C02), diuretics (ATC C03), β‐blockers (ATC C07), calcium channel blockers (ATC C08), and renin‐angiotensin system agents (ATC C09). Education was classified as: primary (some primary/not complete; primary or equivalent); secondary (intermediate/junior/group certificate or equivalent; leaving certificate or equivalent); and tertiary (diploma/certificate; primary degree; postgraduate/higher degree). SES was derived based on respondent's current occupation (or historic occupation—defined as the job title of the highest paying job they ever held—if they had retired). The coding of occupations followed the Irish Central Statistics Office social class schema: professional; managerial; non‐manual; skilled manual; semi‐skilled; unskilled; and all others gainfully occupied but unknown. These were then aggregated into three categories. All covariate data utilized relates to data collected at Wave 1 (baseline).

2.5. Statistical Analysis

Comparisons of baseline characteristics (using either the independent samples t‐test for continuous variables or the chi‐square test for categorical variables) were made between those who died during follow‐up and those still alive. Categorical variables were explored using n (%) and continuous variables by means (standard deviations [SD]). The distributions of C‐reactive protein (CRP) and glycosylated hemoglobin were positively skewed; therefore, these variables were summarized using geometric mean and interquartile range.

A Kaplan–Meier plot was used to display the cumulative survival of participants based on the three categories of self‐report tooth loss: “Dentate, no denture”; “Dentate, with denture(s)”; and “Edentulous.” A log‐rank test was used to compare cumulative survival across categories.

Cox's proportional hazards analysis was used to estimate the hazard ratio (HR) for all‐cause mortality in “Dentate, with denture(s)” and “Edentulous” categories compared to “Dentate, no denture” as the reference category. A series of sequential models were fitted with adjustment for potential confounding variables. Model 1 included adjustment for age and sex; Model 2 included additional adjustment for education, SES, and marital status; Model 3 included additional adjustment for smoking, alcohol, and BMI; and finally, Model 4 included adjustment for diabetes, number of cardiovascular conditions, and use of anti‐hypertensive medication. A test for trend of estimated HRs for all‐cause mortality across categories was performed. Hazard proportionality was assessed using plots of −log(−log(survival)) against time and with tests on Schoenfeld residuals.

To estimate the association between dentate status and cause‐specific mortality, a competing risk survival analysis was conducted [26], focusing on cardiovascular, cancer, and respiratory causes of death. This approach takes account of the fact that there are competing risks to survival when examining cause‐specific mortality. Participants can survive, die from the specific cause of interest, or die from another competing cause. Interpretation of the resulting sub‐distribution hazard ratios (SHR) is similar to that of the Cox HR [27].

The level of statistical significance was set at p < 0.05. Analyses were performed using Stata 15 (StataCorp. 2017. Stata Statistical Software: Release 15. College Station, TX: StataCorp LLC.) and R (R Core Team, Vienna, Austria).

3. Results

A total of 8494 men and women were included in the study with a median follow‐up time of 12.46 years (interquartile range [IQR] 12.2–12.8). The mean age of the men and women at baseline was 63.2 years (standard deviation [SD] 10.2). Of the participants, 3951 (46.5%) were categorized as “Dentate, no denture,” 3041 (35.8%) “Dentate, with denture(s)”; and 1502 (17.7%) as “Edentulous.” Of the 8494 participants, 1430 (16.8%) died during the study observation period.

Baseline characteristics of the participants by mortality status are reported in Table 1. The men and women who died during the study observation period had a mean age of 73.6 (SD 9.9) years at baseline, compared to 61.0 (SD 8.9) years in survivors. There was a significant difference in dentate status; participants who died were proportionally more likely to be edentulous (38.6% vs. 13.4%). Males were proportionally more likely to have died than females: 52.8% of those that died were male compared to 42.8% being male in the survivor group. There was a significant difference across smoking categories; men who died were proportionally more likely to be current smokers (22.1% vs. 17.6%). There was also a significant difference regarding problem alcohol consumption. Problem alcohol consumption was proportionally lower in the group that died (7.3% versus 10.8%); however, there was a higher proportion that preferred not to disclose among those that died (22.4% vs. 16.2%). Participants who died were more likely to have baseline diabetes, hypertension, use antihypertensive medication, have angina, heart failure, stroke, and a higher total number of CVDs. Men and women who died were also proportionally more likely to have a lower education attainment: among those that died, 47.9% had a highest achievement of primary/none versus 25.9% in the group that survived. Those that died were also more likely to be not married/cohabiting: 50.5% of those who died versus 25.7% among survivors. Blood testing showed a significantly greater baseline CRP and HbA1c in those that died, p < 0.001.

TABLE 1.

Baseline characteristics by mortality status n = 8494.

	Died (n = 1430)	Alive (n = 7064)	p a
Age, years, mean (SD)	73.6 (9.9)	61.0 (8.9)	< 0.001
Dentate status, n (%)
Dentate, no dentures	380 (26.6%)	3571 (50.6%)	< 0.001
Dentate, with dentures(s)	498 (34.8%)	2543 (36.0%)
Edentulous	552 (38.6%)	950 (13.4%)
BMI, m/kg2, mean (SD)	28.7 (5.5)	28.7 (5.0)	0.975
Sex, n (%)
Male	755 (52.8%)	3021 (42.8%)	< 0.001
Female	675 (47.2%)	4043 (57.2%)
Smoker, n (%)
Never	491 (34.3%)	3229 (45.7%)	< 0.001
Past	623 (43.6%)	2589 (36.7%)
Current	316 (22.1%)	1245 (17.6%)
Alcohol, n (%)
None	1005 (70.3%)	5159 (73.0%)	< 0.001
Yes	104 (7.3%)	762 (10.8%)
Preferred to not disclose	321 (22.4%)	1143 (16.2%)
Diabetes, n (%)	190 (13.3%)	450 (6.4%)	< 0.001
Hypertension, n (%)	690 (48.3%)	2394 (33.9%)	< 0.001
Hypertensive medication, n (%)	398 (27.8%)	877 (12.4%)	< 0.001
Angina, n (%)	161 (11.3%)	291 (4.1%)	< 0.001
Heart attack, n (%)	151 (10.6%)	228 (3.2%)	< 0.001
Heart failure, n (%)	36 (2.5%)	53 (0.8%)	< 0.001
Stroke, n (%)	60 (4.2%)	74 (1.0%)	< 0.001
Number of CVDs, n (%)
0	1050 (73.4%)	6203 (87.8%)	< 0.001
1–2	338 (23.6%)	827 (11.7%)
≥ 3	42 (2.9%)	34 (0.5%)
Highest education achieved, n (%)
Primary/none	684 (47.9%)	1831 (25.9%)	< 0.001
Secondary	464 (32.5%)	2966 (42.0%)
Third/higher	280 (19.6%)	2265 (32.1%)
Married/cohabiting, n (%)
Not married/cohabiting	722 (50.5%)	1813 (25.7%)	< 0.001
Married/cohabiting	708 (49.5%)	5251 (74.3%)
CRP, mg/L, geo. mean (IQR) b	2.8 (4.1)	2.0 (2.2)	< 0.001
HbA1c, mmol/mol, geo. mean (IQR) b	34.8 (6.0)	32.6 (5.0)	< 0.001

Abbreviations: BMI, body mass index; CRP, C‐reactive protein; CVD, cardiovascular disease; geo., geometric; HbA1c, glycosylated hemoglobin; IQR, interquartile range; n, number of participants; SD, standard deviation.

^a Independent samples t‐test for continuous variables or chi‐square test for categorical variables.

^b Based on n = 5645 that had a blood test.

The most common cause of death was cancer, which accounted for 499 deaths (34.9%). This was followed by circulatory disease with 438 deaths (30.6%), followed by respiratory disease with 178 deaths (12.5%). Other causes accounted for 315 deaths (22.0%). There was a significant difference in the distribution of the causes of death across categories of tooth loss, χ ² = 26.9, df = 9, p < 0.001 (Table 2).

TABLE 2.

Primary cause of death by dentate status, n = 8494.

	Dentate, no dentures (n = 380)	Dentate, with dentures(s) (n = 498)	Edentulous (n = 552)	Total deaths (n = 1430)
Cardiovascular mortality	108 (28.4)	149 (29.9%)	181 (32.8%)	438 (30.6%)
Cancer mortality	157 (41.3%)	183 (36.8%)	159 (28.8%)	499 (34.9%)
Respiratory mortality	39 (10.3%)	47 (9.4%)	92 (16.7%)	178 (12.5%)
Others	76 (20.0%)	119 (23.9%)	120 (21.7%)	315 (22.0%)

Note: Pearson chi² = 26.9, df = 6, p < 0.001.

The Kaplan–Meier plot (Figure 1) shows the cumulative survival of men and women in the three groups of: “Dentate, no denture”; “Dentate, with denture(s)”; and “Edentulous.” As time progressed, a clear trend emerged across the three groups, with highest survival in the “dentate, no denture” and lowest in the “Edentulous” group. There was a significant difference (log rank test, p < 0.001) in survival across the three groups.

FIGURE 1.

Kaplan–Meier survival analysis of all‐cause mortality by dentate status, n = 8494.

Results of the Cox proportional hazards and competing risk survival analyses are shown in Table 3. Compared to participants in the dentate with no denture group, edentulous men had a HR for all‐cause mortality of 1.42 (95% CI 1.23–1.65) p < 0.001 in the fully adjusted analysis (model 4). Trend tests across categories of dentate status were significant in all models, p < 0.001. In the competing risk cause‐specific survival analysis, edentulous participants had a SHR for cardiovascular mortality of 1.25 (95% CI 0.96–1.63) p = 0.10 in the fully adjusted model. For cancer mortality, edentulous participants had a fully adjusted SHR of 1.31 (95% CI 1.01–1.71) p = 0.04. For respiratory mortality, edentulous participants had a fully adjusted SHR of 1.57 (95% CI 1.03–2.41) p = 0.04. Trend tests, in the fully adjusted models, were significant for cancer and respiratory mortality, but not cardiovascular mortality.

TABLE 3.

Summary table of Cox's proportional hazard and competing risk survival analyses for risk of all‐cause and cause‐specific mortality by dentate status, n = 8485. ^a

	Event/n	Model 1 (S)HR (95% CI)	Model 2 (S)HR (95% CI)	Model 3 (S)HR (95% CI)	Model 4 (S)HR (95% CI)
Dentate, no dentures	380/3951	1.00	1.00	1.00	1.00
Dentate, with denture(s)	498/3041	1.08 (0.94–1.24)	1.06 (0.92–1.22)	1.04 (0.90–1.19)	1.02 (0.89–1.17)
Edentulous	552/1502	1.70 (1.48–1.97)	1.53 (1.32–1.78)	1.45 (1.25–1.68)	1.42 (1.23–1.65)
Trend test estimates		1.32 (1.23–1.42)	1.25 (1.16–1.35)	1.22 (1.13–1.31)	1.20 (1.12–1.30)
p value for trend		< 0.001	< 0.001	< 0.001	< 0.001
Cardiovascular mortality
Dentate, no dentures	108/3951	1.00	1.00	1.00	1.00
Dentate, with denture(s)	149/3041	1.04 (0.81–1.34)	1.04 (0.80–1.34)	1.02 (0.79–1.31)	1.02 (0.79–1.32)
Edentulous	181/1502	1.50 (1.15–1.94)	1.32 (1.01–1.72)	1.27 (0.98–1.66)	1.25 (0.96–1.63)
Trend test estimates		1.24 (1.09–1.42)	1.16 (1.01–1.33)	1.14 (0.99–1.30)	1.12 (0.98–1.29)
p value for trend		< 0.001	0.03	0.06	0.09
Cancer mortality
Dentate, no dentures	157/3951	1.00	1.00	1.00	1.00
Dentate, with denture(s)	183/3041	1.10 (0.88–1.37)	1.09 (0.87–1.36)	1.07 (0.85–1.33)	1.06 (0.85–1.32)
Edentulous	159/1502	1.46 (1.14–1.88)	1.40 (1.07–1.82)	1.32 (1.02–1.72)	1.31 (1.01–1.71)
Trend test estimates		1.21 (1.06–1.38)	1.18 (1.03–1.35)	1.15 (1.00–1.31)	1.14 (1.00–1.31)
p value for trend		< 0.001	0.02	0.04	0.05
Respiratory mortality
Dentate, no dentures	39/3951	1.00	1.00	1.00	1.00
Dentate, with denture(s)	47/3041	0.89 (0.57–1.38)	0.88 (0.56–1.38)	0.84 (0.54–1.31)	0.85 (0.55–1.32)
Edentulous	92/1502	2.00 (1.32–3.04)	1.73 (1.11–2.70)	1.52 (1.01–2.45)	1.57 (1.03–2.41)
Trend test estimates		1.52 (1.21–1.92)	1.39 (1.09–1.77)	1.30 (1.03–1.64)	1.32 (1.05–1.67)
p value for trend		< 0.001	0.01	0.03	0.02

Note: Model 1 included adjustment for age and sex; Model 2 included additional adjustment for education, socioeconomic status, and marital status; Model 3 included additional adjustment for smoking, alcohol, and BMI; and finally, Model 4 included adjustment for diabetes, number of cardiovascular conditions, and use of anti‐hypertensive medication.

Abbreviations: CI, confidence interval; HR, hazard ratio (for all‐cause mortality analysis); SHR, sub‐distribution hazard ratio (for cause‐specific mortality analysis).

^a Nine cases were excluded due to missing confounder data.

4. Discussion

The main finding of this prospective cohort study was that self‐report edentulism was an independent risk predictor of all‐cause mortality during a 12‐year follow‐up in a group of 8494 older men and women in Ireland. After adjustment for potential confounders, participants who were edentulous had a 42% increased risk of all‐cause mortality, compared to participants who were dentate, with no denture. A trend test across the categories (“Dentate, no denture”; “Dentate, with denture(s)”; and “Edentulous”) was significant, which suggests a dose‐dependent response relationship in the risk for all‐cause mortality. When investigating cause‐specific mortality, being edentulous posed a 57% greater risk of respiratory mortality, followed by cancer mortality with a 31% greater risk. There was a nonsignificant finding of 25% greater risk for cardiovascular mortality.

The main result of this study corroborates previously published findings that have shown independent associations between edentulism and the risk of all‐cause mortality. A recent meta‐analysis, restricted to studies reporting HRs with ≥ 10 year follow up [28], found that the risk for edentulism and all‐cause mortality was HR = 1.39 (95% CI 1.26–1.54), which is very similar to HR = 1.42 (95% CI 1.23–1.65) found in the current study. In the only other study of mortality in relation to dental status in Ireland, Winning and colleagues reported that edentulism carried a HR = 1.52 (95% CI 1.16–1.99) for all‐cause mortality in a group of 1558 males during a 17 year observation (Belfast, PRIME study) [11]. In a comparable study to the current, conducted among 12,871 older adults in Scotland with an 8‐year follow‐up, self‐reported dental status was categorized into three groups: natural teeth only, a combination of natural teeth and dentures, and no natural teeth (edentulous). Compared to those with natural teeth, the edentulous group demonstrated a 30% higher risk of all‐cause mortality (HR 1.30; 95% CI, 1.12–1.50) [29]. A more recent study carried out in 12,809 adults aged ≥ 70 years in Australia with a median follow up of 6.4 years reported risk of all‐cause mortality was higher among those with edentulism (vs. no edentulism) HR = 1.43 (95% CI = 1.18–1.73). Despite the consistency in results between studies, and also the broad agreement with estimates derived in the recent meta‐analysis [28], a causal link between tooth loss and all‐cause mortality remains questionable. The highest level of evidence of a causal link would be provided by randomized controlled trials. However, there are obvious practical and ethical issues in carrying out such a trial. Therefore, the highest level of evidence available in respect of tooth loss and all‐cause mortality will be from well‐designed prospective cohort studies [30].

Regarding cause‐specific mortality, the greatest association with edentulism was observed with respiratory mortality (SHR = 1.57 [95% CI 1.03–2.41]). The use of dentures has been associated with an increased risk of respiratory mortality, particularly in older adults, due to the potential for aspiration of oral bacteria and debris during sleep or swallowing [31]. This risk is exacerbated by the fact that dentures can accumulate biofilms of pathogenic microorganisms, including respiratory pathogens such as Streptococcus pneumoniae , Haemophilus influenzae , and even Pseudomonas aeruginosa , which may be aspirated into the lungs and lead to infections [32]. Additionally, poorly fitting dentures can cause microaspirations by impairing swallowing function, further increasing the risk of pneumonia, particularly in frail or immunocompromised individuals [33]. After respiratory mortality, cancer mortality provided the next strongest association with edentulism (SHR = 1.31 [95% CI 1.01–1.71]). Edentulism has been associated with an increased risk of cancer mortality, particularly cancers of the gastrointestinal tract, including oral, esophageal, and gastric cancers [34, 35, 36]. The underlying mechanisms are likely multifactorial; poor oral health, which often accompanies tooth loss, can lead to chronic inflammation and persistent oral infections such as periodontitis, conditions that have been linked to carcinogenesis through systemic inflammatory pathways [37]. Additionally, edentulism can impair diet quality, reducing the intake of protective nutrients like fiber, antioxidants, and vitamins, which are crucial in lowering cancer risk [38]. Some studies suggest that individuals with edentulism have a higher prevalence of behaviors such as smoking and alcohol consumption, both of which are well‐established risk factors for cancer, especially oral and pharyngeal cancers [39]. For cardiovascular mortality, a nonsignificant association between edentulism and cardiovascular disease mortality was observed (SHR = 1.25 [95% CI 0.96–1.63], p = 0.10). The sub‐distribution HR of 1.25 indicates a possible increased risk of cardiovascular mortality, with a wide confidence interval that crosses the threshold for statistical significance. While the adjustment for cardiovascular comorbidities in the final model may have led to potential overadjustment, the attenuation of the sub‐distribution HR from model 3 to model 4 was relatively modest (Model 3 SHR = 1.27 [0.98–1.66]), indicating that the observed association was not substantially influenced by the additional covariates. Although the use of ICD‐10 codes I00–I99 likely improved the sensitivity of capturing cardiovascular deaths, the heterogeneity of this classification may have reduced the specificity of associations with particular disease processes. Additionally, temporal variation in cardiovascular disease presentation and the modifying effects of medical interventions may have further attenuated the observed association. These factors may contribute to the lack of a statistically significant association between tooth loss and cardiovascular mortality in the present study.

Strengths of the study include the relatively large sample size (n = 8494) and extended follow‐up period (12 years), positioning it as one of the larger investigations into the associations between tooth loss and mortality [7]. The nationally representative TILDA cohort enhances the generalizability of the findings to the broader population. The longitudinal design, with its long follow‐up, allows for the assessment of long‐term outcomes, offering a clearer understanding of the relationship between tooth loss and mortality over time. Additionally, the study utilized detailed mortality data, with standardized ICD codes ensuring accuracy in cause‐specific mortality analysis. The statistical models accounted for numerous potential confounders, including age, sex, education, SES, marital status, smoking, alcohol consumption, BMI, diabetes, cardiovascular conditions, and use of anti‐hypertensive medication. This thorough adjustment process indicates that these factors did not fully explain the observed associations.

There are several limitations with this study. First, the reliance on self‐reported tooth loss data introduces the potential for reporting bias, as participants may inaccurately recall or misreport their dental status [40]. While self‐reported tooth loss has been shown to correlate reasonably well with clinical assessments [41, 42], it remains a subjective measure and may not capture the full extent of oral health conditions. However, the large sample size of the current study helps mitigate the impact of this potential bias, providing sufficient statistical power to detect associations. Second, tooth loss is a non‐specific outcome of various oral conditions, with periodontitis being one of the main causes of tooth loss in older adults [43]. Periodontitis is often viewed as a proxy for cumulative oral health decline over time, but directly linking periodontitis as the cause of tooth loss in this study is problematic. Tooth loss may result from multiple factors, including dental caries, dental trauma, congenitally missing teeth, and oral cancer [3]. The inability to determine the precise cause of tooth loss introduces complexity into understanding its direct relationship with mortality. Third, the analysis did not account for changes in dental status over time. Participants who were partially dentate at the start of the study may have experienced progressive tooth loss during the follow‐up period, which could have altered their mortality risk. This dynamic nature of dental health may affect the accuracy of the association between baseline tooth loss and long‐term mortality outcomes. Fourth, the study may be subject to non‐differential misclassification bias. Since tooth loss was self‐reported, errors in estimating the number of missing teeth were likely similar across groups, potentially weakening the observed associations and underestimating the true effect. Additionally, as previously mentioned, because dental status was assessed only once, any tooth loss progression during follow‐up was not captured, contributing to possible non‐differential bias. Fifth, although the study controlled for a wide range of potential confounders, residual confounding cannot be entirely ruled out. Unmeasured factors, such as dietary habits, psychosocial factors, or other health‐related behaviors, may still influence both tooth loss and mortality risk. Compromised mastication due to tooth loss can adversely affect nutrient intake and dietary quality by limiting the range of foods that can be comfortably consumed, potentially leading to deficits in essential vitamins, minerals, and dietary fiber, which may in turn moderate the association with mortality [16, 44, 45]. Psychosocial factors also warrant further attention. Tooth loss can significantly affect an individual's self‐esteem, leading to feelings of embarrassment or social anxiety, particularly in social settings. This can result in reduced social interaction, isolation, and depression, which are well‐documented contributors to poor health outcomes in older adults, including increased mortality risk [46]. Furthermore, chronic psychosocial stress linked to tooth loss may exacerbate inflammation and disrupt immune function, adding to the risk of systemic health decline [47]. These psychosocial impacts were not fully captured in the study and may play a critical role in mediating the relationship between tooth loss and mortality. Sixth, despite the 12‐year prospective design of this study, the possibility of reverse causation should be acknowledged. Poorer general health may lead to greater tooth loss due to systemic conditions or limited access to dental care, potentially contributing to the observed association between tooth loss and increased mortality risk. This highlights the need for cautious interpretation of the direction of causality. Finally, whilst the TILDA cohort is nationally representative, the findings may not be generalizable to populations outside Ireland, where healthcare systems, socio‐economic contexts, and oral health practices may differ. Further research in other settings is necessary to confirm the findings in diverse populations.

5. Conclusion

In conclusion, edentulism was independently associated with all‐cause mortality in a cohort of 8494 men and women from Ireland over a 12‐year observation period. Edentulism was significantly associated with respiratory and cancer mortality, and a nonsignificant association with cardiovascular mortality. Our findings support the hypothesis that edentulism may be an important risk indicator for all‐cause mortality, highlighting the systemic implications of poor oral health. These findings underscore the need for further research to explore the underlying mechanisms linking tooth loss to different causes of death, including the role of nutritional deficits, inflammation, and psychosocial factors.

Author Contributions

Lewis Winning: conceptualization, formal analysis, methodology, writing – original draft. Siobhan Scarlett: formal analysis, methodology, writing – review and editing. Mark Ward: formal analysis, methodology, writing – review and editing. Michael Crowe: methodology, writing – review and editing. Rose Anne Kenny: conceptualization, funding acquisition (TILDA study), writing – review and editing. Brian O'Connell: conceptualization, writing – review and editing.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Figure S1. Kaplan–Meier survival analysis of all‐cause mortality by original five‐category dentate status (n = 8494).

Table S1. Cox proportional hazards analysis for all‐cause mortality by original five‐category dentate status (n = 8485).

Winning L., Scarlett S., Ward M., Crowe M., Kenny R. A., and O'Connell B., “Tooth Loss and 12‐Year Mortality Risk in 8494 Older Adults From Ireland,” Journal of the American Geriatrics Society 73, no. 8 (2025): 2431–2440, 10.1111/jgs.19539.