Periodontitis and Alzheimer's: Association vs. Confounding and Causality Challenges

What does "Observed Association" mean for me as a patient?

A common patient question is how to weigh observed association. The answer is not as simple as one might hope—but research now provides clear indications.

The epidemiological association between periodontal disease and cognitive impairment is one of the most consistently replicated findings in oral systemic medicine over the last decade. The most comprehensive current summary of multiple studies by Larvin et al. (2023) included 39 observational studies—13 cross-sectional studies and 26 longitudinal studies—and found pooled relative risks of 1.33 (95% CI 1.13–1.55) for cognitive decline and 1.22 (95% CI 1.14–1.31) for dementia and Alzheimer's disease in individuals with periodontal disease compared to controls. Notably, there was a dose-response effect: moderate gum disease (periodontitis) showed a relative risk of 1.14 (95% CI 1.07–1.22), while severe gum disease (periodontitis) showed a relative risk of 1.25 (95% CI 1.18–1.32). This severity gradient strengthens the plausibility of a real biological signal, even if it does not prove causation on its own. Additionally, Larvin et al. identified a gender effect: for every 10% increase in the proportion of women in the study population, the risk of cognitive decline increased by 34% (RR 1.34; 95% CI 1.16–1.55), suggesting sex-specific vulnerabilities or exposure differences.

The systematic review by Dibello et al. (2024), published in GeroScience and registered on PROSPERO (CRD42023485688), confirmed this directional finding with a broader outcome spectrum and a more extensive data base. The authors searched six databases and identified 46 studies that met the inclusion criteria—a significantly larger basis than most previous reviews. In the analysis of cross-sectional studies, the relative risk for cognitive impairment was 1.25 (95% CI 1.11–1.40). In longitudinal analyses, the association with cognitive impairment was significantly stronger: RR 3.01 (95% CI 1.52–5.95). The overall risk for dementia was RR 1.22 (95% CI 1.10–1.36). Dibello et al. also examined depressive disorders and found no significantly increased association (RR 1.07; 95% CI 0.95–1.21), which strengthens the specificity of the cognitive signal over a non-specific inflammatory sequence. The geographic distribution of the included studies was broad: 36.9% from Asia, 30.5% from Europe, 26.1% from North America, and 6.5% from South America, involving a total of over 3.2 million study participants.

The GRADE assessment of the overall evidence by Dibello et al. (2024) yielded moderate quality—a relevant finding because it shows that the methodological rigor is sufficient to support an associative statement, but not for a causal conclusion. The quality assessment of individual studies using NIH Quality Assessment Toolkits showed a mixed picture: most studies were rated as methodologically acceptable (fair), with few rated as good or poor. Cross-sectional studies dominated the study design (47.8%), followed by prospective cohorts (23.9%), retrospective cohorts (15.2%), and case-control studies (13.1%). This design heterogeneity is methodologically relevant because cross-sectional studies cannot establish a time sequence between exposure and outcome, making them less informative for the question of causality than prospective designs.

Another meta-analysis of multiple studies by Kaliamoorthy et al. (2022) in Medicine and Pharmacy Reports analyzed the association between periodontal disease (periodontitis) and Alzheimer's disease specifically in observational studies, determining a pooled Odds Ratio of 1.67 (95% CI 1.21–2.32). The study was smaller—only five studies met the inclusion criteria for the systematic review and three for the meta-analysis—but the results were in the same direction as larger reviews. The PROSPERO-registered protocol (CRD42020185264) and PRISMA-compliant execution strengthen the methodological credibility, even if the small number of studies limits the precision of the estimate. Sensitivity analyses and tests for statistical heterogeneity were performed and showed moderate heterogeneity, which must be considered during interpretation.

The systematic review by Khoury et al. (2023) in IJERPH supplemented the quantitative scientific evidence with a narrative synthesis of eleven studies—six cohort studies, three cross-sectional studies, and two case-control studies—which could not be quantitatively pooled due to methodological heterogeneity. Particularly relevant are the longitudinal data: Tzeng et al. (2016) found that after ten years of chronic periodontal disease (periodontitis), there was an increased risk of dementia with an adjusted Hazard Ratio of 2.54 (95% CI 1.30–3.35), adjusted for sex, age, income, degree of urbanization, and comorbidities. Chen et al. (2017) showed an adjusted HR of 1.71 (95% CI 1.15–2.53) for Alzheimer's disease after a ten-year diagnosis of periodontitis, independent of comorbidities and degree of urbanization—the mediation analysis identified cerebrovascular disease as a partial mediator. Choi et al. (2019) documented in a large cohort of 262,349 participants an adjusted HR of 1.06 (95% CI 1.01–1.11) for total dementia and 1.05 (95% CI 1.00–1.11) for Alzheimer's in patients with chronic periodontal disease (periodontitis), adjusted for lifestyle factors. The consistency of the direction across different countries (Taiwan, Korea, USA), time periods, and adjustment methods is remarkable.

Stewart et al., with their large prospective long-term observational study (cited in Khoury et al. 2023), provided a nuanced finding: Although most oral health parameters were not associated with later cognitive decline, gingival inflammation was the exception—it was strongly associated with cognitive impairment (OR 1.62; 95% CI 1.09–2.42) and was the only oral factor that predicted cognitive decline. Demmer et al. confirmed in another cohort that Periodontal Profile Class 'severe'—defined by attachment loss, probing depth, and bleeding on probing—was associated with a moderately increased risk of incident dementia (adjusted HR 1.22; 95% CI 1.01–1.47). Shin et al. found in a Korean case-control study that participants with a history of periodontitis had a 2.1-fold increased risk of cognitive impairment (OR 2.14; 95% CI 1.04–4.41), with the interaction effect of smoking and exercise on periodontal disease (periodontitis) modulating the association.

In summary, the association between periodontal disease (periodontitis) and cognitive impairment shows a converging signal across multiple study designs and populations. The pooled risk estimates are predominantly in the range of 1.2 to 1.7 for dementia and Alzheimer's, with stronger signals in studies that examined more severe periodontal disease or longer exposure times. The severity gradient (Larvin et al. 2023), the specificity for cognitive versus depressive outcomes (Dibello et al. 2024), and the consistency across geographically and methodologically diverse studies strengthen the credibility of the signal. It is clinically relevant and warrants increased attention to oral health in older patients—but it does not establish a direct causal relationship, such as Alzheimer's being caused by periodontal disease (periodontitis) on its own.

For dental practice, the associative scientific evidence means that periodontal disease (periodontitis) in older patients is not just an oral problem but a marker for a potentially increased cognitive risk profile. This does not mean that the dentist should diagnose or prevent Alzheimer's—but rather that periodontal health must be understood as part of comprehensive medical care strategy. The consistency of the signal across various definitions and populations makes it unlikely that the association is a mere artifact—even if the exact causal attribution remains unresolved.

What this scientific evidence does not prove: A causal conclusion such as "periodontitis treatment prevents Alzheimer's" is not supported by intervention studies at this time. Similarly, the association does not justify recommending screening for cognitive impairment in the dental office without a validated clinical pathway. The temptation to draw a therapeutic conclusion from a robust association is understandable but methodologically unsupported. The gap between associative scientific evidence and therapeutic recommendation can only be closed by intervention studies, which are currently lacking.

Clinical decisions should not be based on single studies but rather on the overall body of available scientific evidence: Periodontal therapy is indicated for many reasons—the possible risk reduction for cognitive impairment is an additional argument in a comprehensive assessment, but it is not the sole reason for treatment. When discussing benefits and risks with the patient, the cognitive dimension can be mentioned, but it should not serve as the main argument for treatment so as not to create false expectations.

For dental education and continuing professional development, the current research suggests that the topic of oral-systemic connections in geriatrics deserves greater emphasis. The idea that dentistry only treats local problems is increasingly being challenged by growing scientific evidence of periodontal associations with systemic diseases—diabetes, cardiovascular disease, and indeed cognitive impairment. This does not change the therapy, but it changes the perspective on the discipline itself.

What does this mean for you? The text should not present the connection as a simple cause.

You may encounter this topic more often in daily life than you think. What is important here: Not every report you find in the media or online accurately reflects the state of research. The studies show a more nuanced picture than sensational headlines suggest.

Science has intensively investigated this topic in recent years. For this article, more than 11 scientific papers were evaluated. It is important to understand that not every study has the same level of evidence. Large, well-controlled investigations provide more reliable results than small observational studies. The overall picture from these various studies is what we present to you here.

What does "confounding and shared risk" mean for me as a patient?

When it comes to confounding and shared risk, the research situation is clearer than many people think. Here you will learn what current studies actually show.

The central methodological challenge in periodontitis-Alzheimer's research is the massive overlap of risk factors that promote both periodontal disease and neurodegenerative disease. Age is the strongest common risk factor: Both the prevalence of severe periodontal disease (periodontitis) and dementia risk increase sharply after age 65. Epidemiological data show that periodontal disease affects over 50% of the older population (Khoury et al. 2023), while current projections expect a threefold increase in global dementia cases by 2050. This parallel, age-related increase inevitably creates associative signals whose causal proportion is difficult to separate from the common age dependency. Dioguardi et al. (2019) documented age differences of up to 15 years between case and control groups in individual included studies—an indication of insufficient age matching in parts of the primary literature.

Beyond age, both conditions share a wide spectrum of confounding variables. Cardiovascular diseases—especially hypertension, diabetes mellitus, and cerebrovascular diseases—are independent risk factors for both Alzheimer's disease and periodontitis. Chen et al. (2017) identified cerebrovascular disease in their mediation analysis as a partial mediator between chronic periodontitis and Alzheimer's—a methodologically valuable finding showing that part of the associated signal is transmitted via vascular pathways. Choi et al. (2019) adjusted for lifestyle factors, including smoking, alcohol consumption, and physical activity, in their large cohort—the association remained, but the adjustment cannot rule out residual confounding by unmeasured variables. Socioeconomic status and educational level, both strong predictors of dementia and simultaneously determinants of access to dental care, were not fully controlled in most primary studies.

A particular problem is so-called reverse causation: the direction of the association can be reversed. Patients with early cognitive decline progressively neglect their oral hygiene, which accelerates periodontitis and tooth loss. Khoury et al. (2023) discussed this bidirectional relationship in detail and referred to data showing that utilization of dental care significantly decreases after a dementia diagnosis (Fereshtehnejad et al. 2018). The included literature on oral health in Alzheimer's patients—especially the work by Aragón et al. (2017)—confirms that Alzheimer's patients have significantly poorer oral health, with more caries, more periodontal disease, and more mucosal lesions than control subjects. Cross-sectional studies and case-control studies cannot fundamentally resolve this temporality. Even longitudinal long-term observational studies controlling for periodontitis as an exposure before the onset of dementia rarely account for subclinical cognitive decline, which can precede a formal diagnosis by years.

The summary of several studies by Dioguardi et al. (2019) on tooth loss in Alzheimer's patients particularly illustrates the confounding problem. The pooled hazard ratio for tooth loss was 1.52 (95% CI 1.00–2.30), and for edentulism, it was 2.26 (95% CI 1.70–3.01). However, tooth loss is not a specific marker for periodontitis—it can result from caries, trauma, extractions for prosthetic reasons, or lack of care. The authors themselves pointed out significant heterogeneity among the studies (I² = 98% for the primary endpoint) and methodological weaknesses in the included case-control studies, which lost points on the Newcastle-Ottawa Scale for insufficient comparability and incomplete adjustment. When tooth loss is used as a proxy for periodontitis without differentiating the actual cause of tooth loss, the risk of systematic misclassification increases. The seven main points identified by Dioguardi et al. in their review emphasize the multifactorial nature of the association: inflammatory mechanisms, masticatory insufficiency, nutritional deficiencies, genetic overlap, and gaps in care among cognitively impaired patients act simultaneously.

Mendelian randomization analyses were introduced as a methodological approach to bypass the confounding problem by using genetic variants as instrumental variables for the exposure. Zhu et al. (2025) included 56 MR studies in their systematic review, which examined causal associations between Alzheimer's and various systemic diseases. For periodontitis, the work by Sun et al. (2020) was included, showing significant genetic correlations with 4,924 and 12,289 cases, respectively, in European populations. However, the effect sizes in the MR analysis were smaller than in observational studies, suggesting that part of the epidemiological signal is indeed explained by confounding. In the overall picture from the Zhu et al. review, it was shown that several systemic inflammatory conditions—including hypertension (OR 4.30; p = 0.044), coronary heart disease (OR 1.07; p = 0.021), and depression (OR 1.03; p = 0.001)—are genetically associated with an increased risk of Alzheimer's disease. Periodontitis thus joins a pattern of chronically inflammatory exposures that share a common genetic architecture with neurodegeneration.

However, the validity of MR instruments for periodontal disease is not without issues. Genetic variants associated with periodontal disease can have pleiotropic effects on other inflammatory pathways, thereby violating the exclusion restriction—the core assumption that the instrument only acts on the outcome through the studied exposure pathway. With an exposure as heterogeneous as periodontal disease, which has both local and systemic components, this assumption is difficult to defend. The causal question therefore remains unanswered even with MR scientific evidence, and the confounding structure can only be partially, not completely, resolved by the existing methodology.

The clinical implication of this confounding structure is clear: It is important not to present the association as a simple cause-and-effect chain. Residual confounding must remain visible in the professional text. This does not mean that the association is clinically irrelevant—on the contrary. The shared risk factors offer starting points for integrated care strategies that address both oral and systemic health together. Especially because periodontal disease and Alzheimer's share risk factors, better control of these factors—diabetes management, blood pressure control, smoking cessation—can benefit both conditions simultaneously.

For daily dental practice, this means: In older patients with periodontal disease, it is worthwhile to consider the systemic risk constellation—not to diagnose Alzheimer's, but to improve overall care. The shared risk factors (diabetes, hypertension, smoking, lack of physical activity) are the same ones that also worsen periodontal prognosis. An integrated care perspective is more clinically sensible than an isolated focus on causation. The question 'Did my periodontal disease cause my cognitive decline?' is less productive clinically than the question 'What common risk factors can I address in this patient simultaneously?'

What does not follow from confounding evidence: The conclusion that the association is 'only' a statistical artifact. Shared risk factors can reflect real biological pathways—systemic inflammation, vascular damage, nutritional deficiencies—that drive both periodontal and neurodegenerative processes. The question is not if, but to what extent these pathways act independently or through common mechanisms. The shared-risk hypothesis does not rule out causality; it only relativizes the proportion of direct causation.

What does this mean for you? It is important to keep residual confounding open.

You may encounter this topic more often in daily life than you think. What is important here: Not every report you find in the media or online accurately reflects the state of research. The studies show a more nuanced picture than sensational headlines suggest.

How do scientists arrive at these statements? They don't just evaluate a single study; they look at many studies simultaneously. This allows them to determine whether a result was random or if it is consistently confirmed. In this case, the findings are based on 11 scientific papers from different countries and research groups.

What does "Clinical Implication Despite Causal Limitations" mean for me as a patient?

One point that often causes uncertainty is clinical implication despite causal limitations. However, science has made important progress in recent years.

The contribution of the Gwak et al. study (2025) to the exposure complex is indirect: The study primarily examined atopic dermatitis and dementia by summarizing multiple studies and using Mendelian random assignment, thereby adding to the picture of systemic inflammatory conditions as potential risk factors for dementia. Although the focus is not on periodontal disease (periodontitis), the study strengthens the overarching hypothesis that chronic inflammatory diseases—regardless of the organ of origin—may promote neurodegenerative processes. This is relevant for clinically classifying the link between periodontal disease (periodontitis) and Alzheimer's because it places periodontitis within a broader pattern of chronic inflammation rather than treating it as an isolated causal factor. Periodontal disease (periodontitis) may not be the cause, but rather one component in an overall inflammatory environment that accelerates cognitive aging processes.

For practice, this means: Oral health remains systemically relevant—regardless of whether the causal chain to Alzheimer's can be detailed. Periodontal therapy is indicated for periodontal, cardiovascular, and diabetic reasons. The potential risk reduction for cognitive impairment is an additional argument, but not a standalone reason for treatment. The clinical logic must remain inductive here: We treat periodontal disease (periodontitis) because there are good reasons to do so—a potential added cognitive benefit strengthens these reasons but does not create a new indication for therapy.

What is not supported by this scientific evidence: Recommending periodontal therapy as "Alzheimer's prevention" or recommending specific antimicrobial strategies against P. gingivalis with the goal of neuroprotection. Similarly, cognitive screening in the dental office is not justified by current scientific evidence—there is a lack of both validated screening tools for this setting and evidence-based actions that would result from a positive finding.

The most appropriate clinical stance is systemically interested vigilance: Take periodontal disease (periodontitis) in older patients seriously, consider the systemic risk constellation, but do not include exaggerated causal claims in patient communication. The clinical consequence is better overall care, not an overzealous curative logic. Those who take the topic seriously in practice improve care—those who exaggerate it damage the credibility of evidence-based recommendations.

For the interdisciplinary care of older patients, the current research situation implies a stronger connection between dentistry and geriatrics. The S3 guideline for periodontal therapy (Stages I–III) already recommends systemic risk assessment—the cognitive dimension belongs in this context, not as a standalone endpoint, but as an additional aspect of the overall consideration.

What does this mean for you? Practice messages must be systemically interested but causally cautious.

You may encounter this topic more often in daily life than you think. What is important here: Not every report you find in the media or online accurately reflects the state of research. The studies show a more nuanced picture than sensational headlines suggest.

What makes these results reliable? In medical research, the rule is: the more independent studies that reach the same conclusion, the more certain the statement is. The type of study and the number of participants also play an important role. Large controlled studies with many participants provide more reliable results than small surveys.