The global population is rapidly ageing, attributable to increases in life expectancy (Pais et al., 2020). In 2019, approximately 9% of the population in the world was over 65 years old, and this percentage will increase to 16% by 2050 (Huang and Ren, 2022). Age-related cognitive decline is emerging as a significant health concern among the elderly. In the United States, around 36% of the population experiences cognitive impairment, a rate anticipated to double by 2050 (Hebert et al., 2013). Cognitive impairment imposes substantial emotional and financial strains on individuals and healthcare systems.

Socioeconomic status (SES), a crucial determinant of health outcomes, has been shown to positively correlate with cognitive performance (Liu et al., 2022). However, the underlying mechanisms of this relationship remain unclear. One hypothesis is that lower SES is associated with higher levels of systemic inflammation, which may further contribute to cognitive decline (Muscatell et al., 2020). However, the evidence linking SES and inflammation is inconsistent, potentially due to variations in the measures of SES and sources of inflammation (Muscatell et al., 2020).

Periodontitis, a significant oral health condition that manifests throughout the human lifespan, is marked by a polymicrobial dysbiotic infection of the periodontium (Ebersole et al., 2018; Hajishengallis, 2015). Research has identified certain pathogens involved in periodontitis as contributing factors to the pathogenesis of Alzheimer’s disease (AD) (Honjo et al., 2009). Evidence suggested that participants aged 20–59 years with periodontitis had significantly higher scores on the symbol digit substitution test and serial digit learning test (Sung et al., 2019). Recent systematic reviews and meta-analyses further consolidated evidence linking periodontitis to cognitive disorders, reporting pooled odds ratios of approximately 1.23 for general cognitive decline (Dibello et al., 2024). Furthermore, SES has been highlighted as a potential etiological factor for periodontitis (Lenk et al., 2022; Han et al., 2019). Nevertheless, these studies often used a single variable to represent SES, providing an incomplete depiction of its impact. The role of periodontal status in the relationship between SES and cognitive performance remains unclear.

Therefore, we implemented a cross-sectional study to assess the potential correlation of SES with periodontal status and cognitive performance among older adults from the National Health and Nutrition Examination Survey (NHANES) 2011–2014 cohort. Additionally, we conducted further analysis to evaluate whether periodontal status could mediate the relationship between SES and cognitive performance.

From 2,934 participants who had complete and reliable cognitive performance test data, 1,812 with complete periodontal data were included in our analyses (Figure 1). The proportion of missing data for selected variables did not exceed 10% of the total sample. Based on the results of AIC, BIC, and G², participants were categorized into three SES groups (Supplementary Figure 2 and Table 2; Supplementary Table 2).

Table 3 displays the baseline characteristics of the study cohort (n = 1,812). Statistically significant differences were observed among the participants in different groups according to sex (p = 0.01), age (p < 0.01), race (p < 0.01), marital status (p < 0.01), smoking status (p = 0.03), alcohol consumption (p < 0.01), physical activity (p < 0.01), clinical depression (p < 0.01), hypertension (p < 0.01), diabetes (p < 0.01), and stroke (p < 0.01), but not according to BMI.

The associations between SES and cognitive performance were presented in Table 4. Compared with participants in the low SES group, those in the high SES group exhibited a lower likelihood of experiencing low cognitive performance as measured by CERAD (Model 1, OR = 0.17, 95% CI: 0.1 to 0.31; Model 2, OR = 0.18, 95% CI: 0.1 to 0.31; Model 3, OR = 0.22, 95% CI: 0.12 to 0.4), AFT (Model 1, OR = 0.13, 95% CI: 0.08 to 0.22; Model 2, OR = 0.19, 95% CI: 0.12 to 0.29; Model 3, OR = 0.25, 95% CI: 0.15 to 0.41), and DSST (Model 1, OR = 0.03, 95% CI: 0.02 to 0.05; Model 2, OR = 0.06, 95% CI: 0.03 to 0.1; Model 3, OR = 0.08, 95% CI: 0.04 to 0.14) in all three models.

The relationship between periodontal status and cognitive performance was shown in Table 5. In Model 1, mean AL and mean PD were positively associated with all three cognitive tests. However, when all covariates were adjusted, mean AL was only positively associated with AFT [OR with 95% CI of 1.21 (1.02 to 1.43)] and DSST [OR with 95% CI of 1.61 (1.33 to 1.95)]. Mean PD was significantly associated with CERAD [OR with 95% CI of 1.36 (1.05 to 1.77)] and DSST [OR with 95% CI of 2.14 (1.54 to 2.96)] but not with AFT.

Table 6 showed the associations of SES with mean AL and mean PD. Compared with participants in the low SES group, those in the high SES group showed an inverse associated with mean AL (Model 1, OR = −1.06, 95% CI: −1.31 to −0.81; Model 2, OR = −0.85, 95% CI: −1.14 to −0.56; Model 3, OR = −0.81, 95% CI: −1.15 to −0.46) and mean PD (Model 1, OR = −0.39, 95% CI: −0.5 to −0.27; Model 2, OR = −0.29, 95% CI: −0.42 to −0.16; Model 3, OR = −0.3, 95% CI: −0.45 to −0.15) in all three models.

The mediating effect of mean AL on the correlation between SES and cognitive performance is presented in Table 7. In model 1, mean AL could mediate the association of SES with cognitive performance in all three cognitive tests. However, in the fully adjusted model, mean AL could only mediate 11.05% of the association between SES and DSST [β with 95% CI of −0.25 (−0.41 to −0.13)]. Similarly, the mediation effect of mean PD was only significant in the association of SES with DSST in model 3 [β with 95% CI of −0.16 (−0.3 to −0.06)] (Table 8).

The mediation effect of mean AL and mean PD in different subgroups were analyzed. After adjusting for all covariates, mean AL could only mediate the relationship between SES and CERAD in participants with BMI less than 25 and the relationship between SES and AFT in male participants (Supplementary Tables 3, 4). In terms of DSST, mean AL could mediate this association in all participants except for those older than 80 or NHB participants (Supplementary Table 5). Moreover, mean PD could only mediate the relationship between SES and CERAD in participants with BMI less than 25 or those in the vigorous/moderate group (Supplementary Table 6). No significant mediation effect of mean PD was observed in the relationship between SES and AFT (Supplementary Table 7). Regarding DSST, mean PD could mediate this correlation in all participants except for those aged older than 70, those with BMI less than 30, NHB participants, or those with no alcohol consumption (Supplementary Table 8).

In this study, the relationship between SES and periodontal status and the mediation effect of periodontal status in the association between SES and cognitive performance in participants aged 60 years or older in the United States was investigated. The results showed that the association of SES with cognitive performance, particularly in terms of DSST performance, could be partly mediated by periodontal status. Moreover, there were differences in the mediating effects of periodontal status in this association among different subgroups, particularly in terms of age and race.

Socioeconomic inequity in cognitive decline has been widely discussed. A 2025 JAMA Network Open cohort of 7,303 adults demonstrated that higher adulthood SES was associated with slower cognitive decline and more favorable brain MRI markers, including reduced white matter hyperintensities and greater total brain volume (Krueger et al., 2025). Besides, a previous study based on National Longitudinal Study of Adolescent to Adult Health indicated that participants with high SES, particularly those with occupations that use analytical skills and involve forms of social interaction, were more likely to have better cognitive performance (Stebbins et al., 2022). Furthermore, a meta-analysis of 39 prospective studies reported that low-SES participants had a 31% higher risk of cognitive impairment and dementia compared with high-SES counterparts (Wang et al., 2023). However, several potential reasons leading to the important gaps remain. First, SES was a complex concept which cannot be represented by a single indicator (Petrovic et al., 2018). In addition, the characteristics of study populations, design, and data collection methods varied widely. In this study, we constructed a comprehensive SES variable including PIR, educational level, occupation, and health insurance, and confirmed the socioeconomic disparity in cognitive performance. Notably, periodontal measures demonstrated significant associations with all three cognitive outcomes in unadjusted analyses. However, these associations were attenuated for CERAD and AFT following adjustment for potential confounders, while the relationship with DSST remained robust. This attenuation reflects the multifactorial nature of both periodontal disease and cognitive decline. In particular, the mediation effect of periodontal measures was robust for DSST outcomes, suggesting that periodontal health is particularly relevant to cognitive domains related to processing speed, sustained attention, and working memory. Conversely, the mediation effects observed for the CERAD and AFT tests were weaker or only evident in certain subgroups. The differential findings across cognitive tests may be due to the inherent differences in what each cognitive test measures and their sensitivity to inflammation-driven cognitive decline (Sung et al., 2019; Li et al., 2021).

There have been several possible mechanisms to explain the mediation effect of periodontal status in the relationship between SES and cognitive performance in older adults. First, SES could directly affect oral health behaviors and mental health status, and poor mental health status could further exacerbate poor oral health behaviors (Meyer et al., 2014; Holde et al., 2018). Moreover, P. gingivalis periodontal infection could cause cognitive decline by releasing proinflammatory cytokines, including tumor necrosis factor-alpha (TNF-α), interleukin (IL)-6, and IL-1β, in the brain (Ding et al., 2018). Chronic exposure to P. gingivalis lipopolysaccharide could induce peripheral Aβ accumulation in inflammatory monocytes/macrophages (Nie et al., 2019). Huang et al. (2025) systematically demonstrated that P. gingivalis-derived outer membrane vesicles (OMVs) traverse the blood-brain barrier and induce microglial activation through TLR4-dependent signaling pathways, thus establishing a direct connection between periodontal infection and neuroinflammatory processes in the central nervous system. Shawkatova et al. (2025) identified pathogenic gingipains as catalytic mediators of tau protein and amyloid precursor protein proteolysis, thereby delineating specific enzymatic pathways through which periodontal microbiota may contribute to AD progression. Besides, Li et al. (2021) found that white blood cell count could mediate the relationship between periodontal status and cognitive function. In addition, evidence suggested that mitochondrial dysfunction was a mediating factor in the link between periodontal status and cognitive impairment (Li et al., 2022). Meanwhile, the result of this study showed that periodontal status could significantly mediate the association of SES with processing speed, sustained attention, and working memory as measured by DSST. This mediation effect of periodontal status was different among subgroups, particularly those divided by age and race, in terms of both mean AL and mean PD. Mechanistically, age-related immunosenescence and dysregulated inflammatory responses could attenuate the association between chronic periodontal inflammation and systemic/neural changes (Fulop et al., 2017). In addition, a survival bias may be at work—individuals with greater inherent resilience to both oral and systemic health challenges are more likely to live beyond 80, thereby diminishing observable SES-related differences in both periodontal health and cognitive performance (Tsakos et al., 2011). The underlying mechanisms for this race-based difference were intricate. One possible explanation could be the genetic susceptibility of periodontal disease (Goncalves-Anjo et al., 2022). African Americans may have a higher genetic susceptibility to periodontal disease, which leads to more severe symptoms of periodontal disease even under similar socioeconomic conditions (Schuch et al., 2017). This genetic susceptibility may mask the influence of socioeconomic status.

This study presents several advantages. Firstly, the data used in this study had a large sample size and excellent quality control, which increased the statistical power to provide more reliable results. Moreover, we constructed a comprehensive SES variable including four aspects to more completely represent overall SES. However, a few limitations should be noted in this study. First, it was difficult to ascertain causality due to the cross-sectional study design. In addition, the cognitive performance measures in this study could not replace a diagnosis based on a clinical examination due to the limitation of NHANES data measured cognitive performance domains.

This study suggested that SES was negatively associated with periodontal status in participants aged 60 years or older in the United States. The associations between SES and cognitive performance can be partly explained by periodontal status. Moreover, the mediation effect of periodontal status in this association was significant only in non-HSB participants or in younger participants. This study might help in understanding the mediation effect of periodontitis in the correlation between SES and cognitive performance, and further support the need for preventative approaches to reduce periodontitis.