Periodontal Disease, Tooth Loss, and Risk of Esophageal and Gastric Adenocarcinoma: A Prospective Study

We read with great interest the study by Coker et al.¹ that provided supportive evidence for the role of oral microbiota in gastric cancer. A few studies also highlighted the possible link with esophageal cancer^2-4. However, there is a lack of robust epidemiologic data on whether periodontal disease and tooth loss, indicators of oral microbial dysbiosis, are associated with these 2 cancers.

Here, we prospectively examined the association of periodontal disease and tooth loss with risk of esophageal and gastric adenocarcinoma in 98,459 women from the Nurses’ Health Study (1992-2014) and 49,685 men from the Health Professionals Follow-up Study (1988-2016). Dental measures, demographics, lifestyle, and diet were assessed using validated follow-up questionnaires. Self-reported cancer diagnosis was confirmed by review of medical records. We used Cox proportional hazards models to calculate the hazard ratios (HRs) and 95% confidence intervals (CIs). We also examined the independent association of periodontal disease and tooth loss in a joint analysis.

Over 22-28 years of follow-up, we documented 199 cases of esophageal adenocarcinoma and 238 cases of gastric adenocarcinoma. Periodontal disease was associated with a 43% and 52% increased risk of esophageal adenocarcinoma (HR 1.43; 95% CI 1.05-1.96) and gastric adenocarcinoma (HR 1.52; 95% CI 1.13-2.04) (Table 1). Compared to individuals with no tooth loss, the risks of esophageal and gastric adenocarcinoma for those with ≥ 2 tooth loss were also modestly increased (HR 1.42; 95% CI 1.00-2.03; P_trend 0.05 and HR 1.33; 95% CI 0.95-1.86; P_trend 0.09, respectively). Among individuals with a history of periodontal disease, no tooth loss and ≥ 1 tooth loss were equally associated with 59% increased risk of esophageal adenocarcinoma (HR 1.59; 95% CI 1.04-2.41 and HR 1.59; 95% CI 1.04-2.44, respectively) compared to those with no periodontal disease and no tooth loss. Similarly, the same group of individuals had 50% and 68% greater risk of gastric adenocarcinoma (HR 1.50; 95% CI 1.01-2.23 and HR 1.68; 95% CI 1.13-2.50, respectively). (Figure 1).

Table 1.

Age-adjusted and multivariable associations of periodontal disease and tooth loss with esophageal and gastric adenocarcinoma

	Periodontal disease			Number of tooth loss
	No	Yes	P	0	1	≥ 2	Ptrendd
Esophageal adenocarcinoma
Case	134	65		123	25	51
Person-year	11033730	997259		10205784	1037319	787886
Age-adjusted HR (95% CI)a	1 (reference)	1.72 (1.27-2.33)	0.0005	1 (reference)	1.26 (0.81-1.95)	1.72 (1.22-2.42)	0.002
Multivariable-adjusted HR (95% CI)b	1 (reference)	1.64 (1.21-2.24)	0.002	1 (reference)	1.24 (0.80-1.94)	1.64 (1.16-2.33)	0.005
Multivariable- + smoking-adjusted HR (95% CI)c	1 (reference)	1.43 (1.05-1.96)	0.02	1 (reference)	1.16 (0.74-1.81)	1.42 (1.00-2.03)	0.05
Gastric adenocarcinoma
Case	169	69		156	30	52
Person-year	11033734	997265		10205789	1037326	787884
Age-adjusted HR (95% CI)a	1 (reference)	1.63 (1.22-2.18)	0.001	1 (reference)	1.14 (0.76-1.69)	1.43 (1.03-1.99)	0.03
Multivariable-adjusted HR (95% CI)b	1 (reference)	1.60 (1.19-2.14)	0.002	1 (reference)	1.13 (0.76-1.69)	1.40 (1.00-1.94)	0.05
Multivariable- + smoking-adjusted HR (95% CI)c	1 (reference)	1.52 (1.13-2.04)	0.006	1 (reference)	1.12 (0.75-1.67)	1.33 (0.95-1.86)	0.09

Abbreviation: BMI, body mass index; CI, confidence interval; HR, hazard ratio; HPFS, Health Professionals Follow-up Study; MET, metabolic equivalent of task; NHS, Nurses’ Health Study.

^aCox proportional hazards model stratified by age (continuous, month), cohort (NHS, HPFS), and study period (in 2-year intervals).

^bFurther adjusted for race (white, nonwhite), history of diabetes mellitus (no, yes), BMI (<20, 20-24.9, 25-29.9, 30-34.9, ≥35 kg/m²), physical activity (in quintiles, MET-hr/wk), fruits (in quintiles, serving/d), vegetables (in quintiles, serving/d), red/processed meat (in quintiles, serving/d), alcohol (women: never, <3.5, 3.5-6.9, ≥7.0 g/d; men: never, <7.0, 7.0-13.9, ≥14.0 g/d), regular aspirin use (no, yes), anti-acid medication (no, yes), menopausal hormone therapy (premenopausal, postmenopausal never user/men, postmenopausal past user, postmenopausal current user), and physical examination within the past 2 years (no, yes).

^cFurther adjusted for smoking status (never smoker, past smoker, current smoker) and smoking intensity (continuous, pack-year)

^dP for trend was calculated by using the median number of tooth loss in each category as an ordinal variable.

Figure 1. Multivariable-adjusted hazard ratio of esophageal adenocarcinoma and gastric adenocarcinoma across history of periodontal disease and tooth loss.

Cox proportional hazards models were stratified by age (continuous, month), cohort (NHS, HPFS), and study period (in 2-year intervals) and adjusted for the same covariates as the full model in Table 1. P for interaction between periodontal disease and tooth loss was calculated by including a cross-product interaction term between periodontal disease and tooth loss in the model and estimating the significance using the likelihood ratio test.

Prior findings on the relationship of periodontal disease and tooth loss with esophageal and gastric cancer have been inconsistent⁵. This is in part due to great variation in study design, exposure ascertainment, and confounding adjustment. Compared to prior studies, our analysis has the advantage of prospective design, long-term follow-up, examination of validated periodontal disease, a more specific marker for dysbiotic oral microbiome, in conjunction with tooth loss, as well as detailed covariate assessment that allowed us to control vigorously for confounding factors, especially smoking. Increasing number of studies have highlighted the strong scientific rationale for a link between oral microbiome and esophageal and gastric cancer. Tannerella forsythia and Porphyromonas gingivalis, as members of the “red complex” of periodontal pathogens⁶, have been associated with the presence or risk of esophageal cancer^2-4. Coker et al.¹ observed an over-representation of Tannerella forsythia, along with Peptostreptococcus stomatis and Streptococcus anginosus, among other oral microbes, in gastric cancer compared with other precancerous stages. From a mechanistic standpoint, Porphyromonas gingivalis may modulate the risk of esophageal cancer through anergy of activated T cells⁷, inhibition of apoptosis⁸, and dehydrogenation of ethanol to acetaldehyde, causing DNA damage, mutation, and excessive proliferation of epithelial cells⁹. Poor oral hygiene and periodontal disease could also promote the formation of endogenous nitrosamines known to cause gastric cancer through nitrate-reducing bacteria¹⁰.

Together, these data support the importance of oral microbiome in esophageal and gastric cancer. Further prospective studies that directly assess oral microbiome are warranted to identify specific oral bacteria responsible for this relationship. These additional findings may serve as readily accessible, non-invasive biomarkers and help identify individuals at high risk for these cancers.

Abbreviations:

BMI

body mass index

CI

confidence interval

FFQ

food frequency questionnaire

HR

hazard ratio

HPFS

Health Professionals Follow-up Study

MET

metabolic equivalent of task

NHS

Nurses’ Health Study

P. gingivalis

Porphyromonas gingivalis

T. forsythia

Tannerella forsythia