The albumin-globulin ratio is associated with periodontitis in American adults: results from the NHANES 2009–2014

Abstract

It’s not clear whether periodontitis and the albumin-to-globulin ratio (AGR) are related. This study intended to determine the association between AGR and periodontitis using the NHANES database. The National Health and Nutrition Examination Survey (NHANES) data from 2009 to 2014 were used in this investigation. The dependent variable was periodontitis, the independent variable was AGR, and the 2012 Eke criteria were used to classify periodontitis. The association between AGR and periodontitis was examined using a multivariate logistic regression model. A total of 9134 participants were included in this study, which was analyzed using multivariate logistic regression and showed a significant negative association between AGR and periodontitis (OR = 0.56; 95% CI, 0.47–0.67, p < 0.0001). Delineation of AGR into quartiles showed that this result persisted. In addition, a threshold nonlinear association between AGR and periodontitis was found, with AGR considered a protective factor at AGR < 1.79 (OR = 0.42; 95% CI, 0.33–0.54, p < 0.0001). This cross-sectional investigation revealed a possible negative correlation between AGR and periodontitis, suggesting that AGR may be a potential biomarker for periodontitis screening and prevention. Future research employing causal methods and longitudinal designs is required to confirm our observations.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-06416-9.

Introduction

Periodontitis represents a chronic inflammatory disorder affecting oral tissues, marked by gingival inflammation, periodontal pocket development, and progressive degradation of the periodontium¹. In addition, periodontitis has been associated with several systemic diseases, including hypertension, diabetes, and cardiovascular disease^2–4. Globally, approximately 11% of the world’s population may suffer from severe periodontitis, affecting 743 million people⁵.

AGR includes the relationship between albumin and globulin. Serum albumin is commonly used as a surrogate indicator of nutritional status, and serum globulin is used to assess the severity of chronic inflammation⁶. AGR has been employed as a potential biomarker for the screening and prevention of depression and lupus nephritis in Chinese patients with systemic lupus erythematosus^7,8.

Some scholars have examined the connection between AGR and periodontitis have discovered that people with low AGR are more likely to get periodontitis^9,10. The study’s strengths include a larger sample size of the study population and more confounding variables related to systemic or periodontal health. Second, a curve-fit graph was created to calculate AGR and the risk of developing periodontitis.

This study utilized cross-sectional analysis of NHANES data from 2009 to 2014 with the hypothesis of a possible correlation between serum AGR and periodontitis, with the aim of exploring the potential association between AGR and periodontitis.

Methods

Data source and study population

The National Health and Nutrition Examination Survey (NHANES) is a publicly accessible database that utilizes a stratified, multi-stage probability design to recruit a representative sample of the U.S. population. The NHANES database contains information on demographics, diet, examinations, laboratory tests, questionnaires, and access to restricted data. Restricted-access data require a request to be submitted before they can be used. Data used in this study were obtained from NHANES 2009–2014.

The inclusion and exclusion process is shown in Fig. 1. The study initially included 30,468 participants. Participants were excluded if. (1) No data for periodontal condition; (2) No AGR data; and (3) Missing information on covariates (a few participants lacked covariates such as hypertension, education, marriage, poverty income ratio (PIR), diabetes, floss use, smoking, body mass index (BMI)). Finally, 9134 participants were included in the analysis.

Fig. 1.

Flow chart of participant selection.

Definition of periodontitis

NHANES implemented a Full Mouth Periodontal Examination (FMPE) program from 2009 to 2014, utilizing a Mobile Examination Center (MEC) to evaluate participants 30 years of age and older. Periodontists at the Mobile Screening Center performed periodontal probing depth (PD) and clinical attachment loss (CAL) measurements on all teeth (6 positions per tooth) of 28 eligible participants, with the exception of the third molar^11,12.

The Centers for Disease Control and Prevention (CDC) and the American Academy of Periodontology (CDC-AAP) have developed a case definition for periodontitis surveillance that defines the severity of periodontitis¹³. Mild periodontitis was defined as 2 adjacent sites with CAL ≥ 3 mm, ≥ 2 adjacent sites with PD ≥ 4 mm (not on the same tooth), or 1 site with PD ≥ 5 mm. Moderate periodontitis was defined as 2 interstitial sites with CAL ≥ 4 mm (not on the same tooth), or ≥ 2 interstitial sites with PD ≥ 5 mm (not on the same tooth). Severe periodontitis was defined as 2 interstitial sites with CAL ≥ 6 mm (not on the same tooth) and 1 ≥ 1 interstitial site with PD ≥ 5 mm. We grouped none, mild periodontitis into one category: having no periodontitis, and we grouped moderate and severe periodontitis into another category: having periodontitis¹⁴.

The calculation of AGR

The Beckman UniCel^® DxC800 Synchron device is used to quantify serum albumin (g/dL) and serum globulin (g/dL)¹⁵. The albumin concentration (g/dL) is divided by the globulin concentration (g/dL) to determine the AGR.

Confounding factors

The study investigated the relationship between AGR and periodontitis using the following covariates: age (years), sex (male/female), ethnicity (Mexican American/other Hispanic/non-Hispanic black/other race), education level (high school and below, high school and above), marital status (married, widowed, divorced, other), alcohol use (none, moderate, heavy, binge drinking), smoking (never, former, current), body mass index (BMI), poverty-to-income ratio (PIR: low < 1.56, medium 1.57–3.62, high ≥ 3.63)¹⁶, hypertension (yes/no), diabetes mellitus (yes/no), and other covariates. We divided BMI into three categories: normal (less than 25.0 kg/m2), overweight (25.0 ≤ kg/m²<30.0), and obese (more than 30.0 kg/m²)¹⁷. The three categories of smoking consumption were as follows: “never” meant smoking less than 100 cigarettes in a lifetime, “former” meant smoking more than 100 cigarettes and not smoking at all presently, and “current” meant smoking more than 100 cigarettes and smoking occasionally or daily¹⁸. Heavy smokers were defined as current smokers consuming > 10 cigarettes per day¹⁹.There are four types of drinking: binge drinking (four drinks per day for women and five for men), heavy drinking (two to three drinks per day for women and three to four for men), moderate drinking (one drinks per day for women and one or two for men) and none²⁰. Flossing is divided into 4 groups: 0 times per week, rarely 1–2 days per week, moderately 3–5 days per week, and frequently 6–7 days per week²¹. The self-report questionnaire collected information on hypertension and diabetes. The NHANES website has comprehensive details on how to measure and query these factors.

Statistical analysis

Continuous variables were presented as mean ± standard deviation (SD), while categorical variables were expressed as proportions. Baseline characteristics of periodontitis were analyzed using ANOVA for continuous variables and chi-square tests for categorical variables.

Multivariable logistic regression models were employed to evaluate the independent association between AGR and periodontitis. Variance inflation factors (VIFs) were calculated to assess multicollinearity, with VIF > 10 indicating severe collinearity. Three models were constructed: Model I (unadjusted), Model II (adjusted for age, sex, and race), and Model III (fully adjusted for age, sex, race, education level, marital status, body mass index, smoking status, alcohol consumption, poverty-income ratio, hypertension, diabetes, and dental floss use).

AGR was analyzed both as a continuous variable and as categorical quartiles, with trend tests performed to examine linear relationships. Potential nonlinear associations were assessed using smooth curve fitting and generalized additive models (GAM).

Subgroup analyses were conducted using stratified multivariable logistic regression models, which incorporated the same adjustments as the fully adjusted model. Stratification factors included sex, age (< 60 vs. ≥ 60 years), smoking status, alcohol consumption, education level, marital status, poverty-income ratio, body mass index, dental floss use, hypertension, and diabetes. Interaction effects were evaluated using likelihood ratio tests.

Sensitivity analyses were performed specifically for diabetic patients and heavy smokers. All statistical analyses were performed using EmpowerStats 2.0, with graphical presentations created using Adobe Illustrator (2021 version). A two-sided p-value < 0.05 was considered statistically significant.

Results

Baseline characteristics of the population

As shown in Table 1, we analyzed data from 9134 participants in this study. There were 4592 individuals without periodontitis and 4615 individuals with the condition. The difference between periodontally healthy individuals and periodontitis participants was considered statistically significant (P < 0.05) in terms of age, AGR, gender, race, education, marriage, PIR, smoking status, alcohol consumption, BMI, hypertension, diabetes and floss use. Participants with periodontitis were more likely to be older, male, lower AGR, non-Hispanic white, more educated, married, low PIR, obese, drink alcohol appropriately, do not smoke, never use dental floss and do not have hypertension or diabetes compared to participants without periodontitis.

Table 1.

Baseline characteristics of participants (N = 9207).

Periodontitis	No	Yes	P-value
N	4592	4615
Age(years, mean ± SD)	48.22 ± 13.61	55.51 ± 13.94	< 0.001
AGR	1.55 ± 0.29	1.48 ± 0.30	< 0.001
Gender, n (%)			< 0.001
Male	1844 (40.41%)	2683 (58.70%)
Female	2719 (59.59%)	1888 (41.30%)
Race, n (%)			< 0.001
Mexican American	458 (10.04%)	787 (17.22%)
Other Hispanic	429 (9.40%)	428 (9.36%)
Non-Hispanic White	2390 (52.38%)	1751 (38.31%)
Non-Hispanic Black	710 (15.56%)	1100 (24.06%)
Other Race	576 (12.62%)	505 (11.05%)
Education, n (%)			< 0.001
High school and below	1436 (31.47%)	2553 (55.85%)
Above high school	3127 (68.53%	2018 (44.15%)
Marital status, n (%)			< 0.001
Married	2867 (62.83%)	2506 (54.82%)
Widowed	231 (5.06%)	396 (8.66%)
Divorced	559 (12.25%)	624 (13.65%)
Others	906 (19.86%)	1045 (22.86%)
PIR(Poverty Income Ratio), n (%)			< 0.001
Low(<1.56)	1248 (27.35%)	2029 (44.39%)
Middle (1.57–3.62)	1349 (29.56%)	1463 (32.01%)
High (>3.63)	1966 (43.09%)	1079 (23.61%)
Smoking consumption, n (%)			< 0.001
Never	2929 (64.19%)	2149 (47.01%)
Former	1052 (23.06%)	1272 (27.83%)
Current	582 (12.75%)	1150 (25.16%)
Alcohol consumption, n (%)			< 0.001
Never	527 (11.55%)	591 (12.93%)
Moderate	2565 (56.21%)	2626 (57.45%)
Heavy	1095 (24.00%)	806 (17.63%)
Binge	376 (8.24%)	548 (11.99%)
Hypertension, n (%)			< 0.001
Yes	1429 (31.32%)	1972 (43.14%)
No	3134 (68.68%)	2599 (56.86%)
Diabetes, n (%)			< 0.001
Yes	376 (8.24%)	734 (16.06%)
No	4073 (89.26%)	3704 (81.03%)
Borderline	114 (2.50%)	133 (2.91%)
BMI (Body Mass Index, kg/m2), n (%)			0.011
<24.9	1286 (28.18%)	1165 (25.49%)
25–29.9	1557 (34.12%)	1588 (34.74%)
>30	1720 (37.69%)	1818 (39.77%)
Floss use, n%			< 0.001
Never	1113 (24.39%)	1765 (38.61%)
Rarely	843 (18.47%)	663 (14.50%)
Moderately	969 (21.24%)	670 (14.66%)
Frequently	1638 (35.90%)	1473 (32.22%)

The association between AGR and periodontitis

To assess multicollinearity in the covariates used in our multivariable logistic regression models, we computed variance inflation factors (VIFs). As shown in Table 2, all VIF values fell between 1.1 and 1.3—well below the standard thresholds of 5 or 10—indicating no significant multicollinearity existed and supporting the robustness of our regression results.

Table 2.

Variance inflation factor (VIF) results.

Variable	VIF Value
Education	1.3
Marital status	1.1
PIR	1.3
Hypertension	1.2
Gender	1.1
Age	1.3
Race	1.1
Diabetes	1.1
Floss use	1.1
Smoking consumption	1.2
Alcohol consumption	1.2
BMI	1.1

VIF (Variance Inflation Factor) values are used to check multicollinearity between covariates

Higher AGR levels were associated with a lower prevalence of periodontitis (Table 3). After controlling for all covariates, AGR and periodontitis were statistically associated when assessed as a continuous variable (OR = 0.56; 95% CI, 0.47–0.67, p < 0.0001).

After AGR was changed from a continuous to a categorical variable, the relationship remained significant. There was a substantial 44% lower risk of periodontitis for those in the highest thoroughly adjusted group compared to those in the lowest group (OR = 0.57; 95% CI 0.49–0.66, P < 0.0001).

The risk of periodontitis was reduced by 24%, 25%, and 43% in Q2, Q3, and Q4, respectively. These findings suggest a potential nonlinear relationship between the variables, while the trend test (P < 0.0001) indicates relatively robust results.

Table 3.

Association between AGR and periodontitis in different models.

Exposure	Model I	Model II	Model III
	(OR, 95% CI, P-value)	(OR, 95% CI, P-value)	(OR, 95% CI, P-value)
AGR	0.43 (0.37, 0.49) <0.0001	0.46 (0.39, 0.54) <0.0001	0.56 (0.47, 0.67) <0.0001
AGR quartiles
Q1(<1.3)	1	1	1
Q2(1.3–1.5)	0.72 (0.64, 0.81) <0.0001	0.72 (0.63, 0.82) <0.0001	0.76 (0.66, 0.87) <0.0001
Q3(1.5–1.7)	0.68 (0.61, 0.76) <0.0001	0.67 (0.59, 0.76) <0.0001	0.75 (0.65, 0.86) <0.0001
Q4(>1.7)	0.48 (0.43, 0.54) <0.0001	0.50 (0.43, 0.57) <0.0001	0.57 (0.49, 0.66) <0.0001
P for trend	<0.001	<0.001	<0.001

Model I no covariates were adjusted

Model II adjusted for Gender, Age, Race

Model III adjusted for Gender, Age, Race, Educational level, Hypertension,PIR, Alcohol consumption, BMI, Smoking consumption,

 Marital status, Diabetes, and Dental floss

All models use Q1 as reference for AGR quartiles

AGR albumin-globulin ratio, OR odds ratio, CI confidence interval

In a generalized additive model, AGR and periodontitis were found to have a threshold nonlinear connection with a substantial inflection point (by calculation, this inflection point is 1.79) Table 4. The risk of developing periodontitis was found to be 58% lower with every unit increase in AGR when the AGR value was less than 1.79 (OR = 0.42; 95% CI 0.33–0.54, P < 0.0001). When the AGR value was > 1.79 (OR = 1.12; 95% CI 0.74–1.69, P = 0.6016), AGR was considered a risk factor for periodontitis despite not being statistically significant. Additionally, the log-likelihood ratio test of the linear regression model was P < 0.001 in comparison to the two-stage linear regression model. The variables’ smooth curve fit is shown by the solid red line Fig. 2. The fit’s 95% confidence intervals are shown by blue bars.

Table 4.

Analysis of the threshold effects of periodontitis and AGR using piecewise logistic regression.

Periodontitis	Model: saturation effect analysis (OR 95% CI P-value)
Fitting by two-piecewise linear model
Breakpoint
< 1.79	0.42 (0.33, 0.54) <0.0001
＞1.79	1.12 (0.74, 1.69) 0.6016
Log-likelihood ratio test	<0.001

Fig. 2.

Dose-response relationship between AGR and periodontitis.

Subgroup analysis and sensitivity analysis

The association between AGR and periodontitis in various populations was evaluated using a stratified approach. Age group (less than 60 and more than 60 years), gender, race, BMI category (less than 25 kg/m2, 25–30 kg/m2, and more than 30 kg/m2), marital status, alcohol use, smoking status, PIR, floss use and the existence of comorbidities such diabetes and hypertension were all used to categorize the participants. Subgroup analysis and interaction testing were also performed.

The study’s findings demonstrated that in the majority of subgroups, there was a persistent negative correlation between AGR and periodontitis Fig. 3. Interaction tests further showed that age, smoking, alcohol consumption, PIR, and diabetes significantly influenced the relationship between AGR and periodontitis (interaction P < 0.05). Specifically, the effect of AGR on periodontitis was more significant in age < 60 (age < 60: OR 0.43,95% CI 0.34, 0.55, P < 0.0001; age ≥ 60: OR 0.79, 95% CI 0.60–1.06, P = 0.1173; P < 0.0001 for interaction). The effect of AGR on periodontitis was significant in “never/previously/currently” smokers (“never”: OR 0.55,95%CI 0.42–0.71, P < 0.0001; “former “: OR 0.69,95% CI 0.50–0.95, P = 0.0242; “current”: OR 0.45,95% CI 0.29–0.69, P = 0.0003; P = 0.0347 for interaction). The effect of AGR on periodontitis was significant in “none/moderate/heavy/binge” drinkers (“none”: OR 0.46, 95% CI 0.27–0.78, P = 0.0043; ”Moderate”: OR 0.63,95% CI 0.49–0.80, P = 0.0001; “Heavy” OR 0.66, 95% CI 0.43–1.00, P = 0.0483; “Binge drinking” : OR 0.32,95% CI 0.17–0.58, P = 0.0002; P = 0.0153 for interaction). The effect of AGR on periodontitis was more significant in “low/medium” PIR (“low”: OR 0.49,95%CI 0.36–0.67, P < 0.0001; “medium”: OR 0.41,95% CI 0.29–0.57, P < 0.0001; P = 0.0306 for “interaction”). The effect of AGR on periodontitis was more significant in “yes/no” diabetes (“yes”: OR 0.43,95%CI 0.26–0.72, P = 0.0013; “no “: OR 0.55,95% CI 0.45–0.68, P < 0.0001; P = 0.0390 for interaction), and although not statistically significant, AGR was a risk factor for periodontitis in people with pre-diabetes (OR 1.37,95%CI 0.54–3.47, P = 0.5039).

Fig. 3.

Subgroup analysis using Forest Maps to examine how AGR affects periodontitis.

Given that diabetes and heavy smoking are established major risk factors for periodontitis, we conducted sensitivity analyses by sequentially excluding: (1) individuals with prediabetes or diabetes (n = 7,777 remaining), and (2) heavy smokers (n = 8,443 remaining). In the multivariate Logistic regression analysis, AGR was categorized into quartiles (Q1-Q4). The results demonstrated consistent inverse associations between AGR and periodontitis: After excluding prediabetic/diabetic individuals: OR = 0.55, 95% CI: 0.45–0.68, p < 0.0001(Table S1)After excluding heavy smokers: OR = 0.57, 95% CI: 0.47–0.69, p < 0.0001(Table S2).These sensitivity analyses yielded effect estimates remarkably similar to our primary findings, thereby confirming the robustness of the observed associations.

Discussion

According to our research, there may be a link between periodontitis and serum AGR. Even after adjusting for potential confounders, a decreased incidence of periodontitis remained linked to greater AGR levels. Subgroup and sensitivity analyses further reinforced the results’ robustness. A consistent pattern of AGR levels below and above the crucial breakpoint of 1.79 was also identified by analysis of the data, which showed a nonlinear negative correlation.

AGR takes into consideration the relationship between both serum albumin and globulin, two indicators that are readily available in routine testing and can effectively characterize a patient’s inflammatory state and nutritional status⁸.

The synthesis of albumin in hepatocytes begins with the synthesis of prealbumin, which is followed by the production of albuminogen and albumin sequentially through protein hydrolysis. As a water-soluble protein, albumin is predominantly secreted into the blood circulation, with only a small amount stored in liver tissue²². Albumin has many potential functions, including maintaining the integrity of capillary endothelium, modulating inflammation, controlling acid-base balance, and binding endogenous and foreign substances²³. Notably, it has been shown that serum albumin levels are negatively correlated with the development of chronic periodontal disease²⁴. Periodontitis, as a chronic multifactorial inflammatory disease, is characterized by the progressive destruction of periodontal supporting tissues, including the alveolar bone and periodontal ligament, and is closely associated with the accumulation of plaque, or dental biofilm⁵. As a negative acute phase response protein, albumin levels are reduced in inflammatory states, a phenomenon that is independent of the nutritional status of the patient²⁵. Low albumin may reflect elevated levels of pro-inflammatory cytokines (e.g., TNF-α, IL-6), which play a central role in periodontitis, and albumin may modulate the inflammatory response by binding and neutralizing these cytokines^26,27. Relevant studies have shown that higher serum antioxidant concentrations are associated with a lower risk of developing periodontitis²⁸. Albumin can exhibit potent antioxidant properties by scavenging reactive oxygen species (ROS), and overproduction of ROS can damage periodontal tissues leading to periodontitis^29,30. Therefore, reduced albumin levels in individuals with low AGR may impair antioxidant defense and exacerbate periodontal inflammation. Another important function of albumin is its involvement in the transport and distribution of a variety of metabolic substances, including thyroid hormones and fatty acids, and these metabolic processes may indirectly influence the pathogenesis of periodontitis^8,31,32.

Nutritional status may influence the relationship between periodontal disease and serum albumin concentration. Since albumin synthesis rate is simultaneously regulated by both nutritional intake and inflammatory status, hypoalbuminemia in chronic disease patients may result from insufficient protein-calorie intake, chronic inflammatory responses, or a combination of these factors³³. Inadequate nutritional supply not only alters the oral microbial ecological balance, but also affects key pathological processes involved in the development of periodontal disease, including microbial colonization, tissue invasion, inflammatory damage, and the ability to repair. Specifically, malnutrition reduces salivary secretion and compromises its antimicrobial properties, particularly by decreasing the concentration of secretory immunoglobulin A (sIgA). This impairment facilitates plaque accumulation and enhances colonization by oral pathogens such as Porphyromonas gingivalis and Aggregatibacter actinomycetemcomitans; Meanwhile, malnutrition also compromises the host’s immune response to periodontal pathogens, exacerbating the destruction of periodontal tissues. Furthermore, it significantly impairs the synthesis of key cytokines (e.g., IFN-γ, TGF-β, IL-4, and IL-1Ra), thereby adversely affecting the repair processes of periodontal tissues³⁴.

Globulins, as a class of heterogeneous proteins with vital biological functions, are primarily composed of four components (α1, α2, β, and γ). These include key protein molecules involved in inflammation regulation, such as the complement system, immunoglobulins, and acute-phase reactive proteins. Elevated serum globulin levels have been shown to indicate acute or chronic inflammatory states^8,35. In periodontal defense mechanisms, immunoglobulins exert protective effects through multiple pathways: (1) neutralizing toxic substances like proteases produced by periodontal pathogens (e.g., Porphyromonas gingivalis); (2) enhancing phagocytic clearance of pathogens through opsonization; and (3) inhibiting bacterial adhesion to oral epithelial cells. However, persistent inflammation leads to B-lymphocyte activation and increased γ-globulin (primarily immunoglobulin) production. Complement system activation (e.g., C3, C5) further elevates globulin levels³⁶, and this excessive immune response may paradoxically exacerbate inflammatory damage in periodontal tissues.

The vast majority of population subgroups consistently demonstrate a significant inverse correlation between AGR and periodontitis, though this association exhibits notable variations across demographic characteristics. The protective effect is more pronounced in individuals under 60 years old compared to elderly populations, potentially attributable to age-related immunosenescence and the annual 0.08–0.17 g/L decline in serum albumin concentrations^37,38. Smokers show attenuated AGR protection, consistent with tobacco’s dual effects of suppressing immunoglobulin production while enhancing proinflammatory cytokine release³⁹. Notably, the inverse association between AGR and periodontitis was more pronounced in diabetic patients compared with non-diabetic individuals. As diabetes mellitus - through its chronic inflammatory metabolic derangements and malnutrition-inducing effects on protein absorption - profoundly impacts albumin homeostasis and consequently AGR dynamics⁸. Furthermore, poverty-income ratio (PIR) and alcohol consumption indirectly influence periodontal status through their respective effects on nutritional intake and hepatic albumin synthesis^40–43.

In contrast to other studies, this one has the advantage of discussing both the smooth curve fit between AGR and periodontitis as well as the baseline features of people with varying AGRs. Second, the large sample size—9134 participants—offers compelling data for evaluating the relationship between AGR and periodontitis severity.

This study has a number of drawbacks. First off, the study relied on cross-sectional data, which has limitations because it lacks causal proof. To better grasp the research question, future studies should think about utilizing additional experimental techniques or longitudinal designs. Second, Although we adjusted for several potential confounders, residual confounding may persist due to unmeasured factors such as frequency of dental visits, adherence to dental appointments, and medication use. Third, Heterogeneity in dental records may exist despite the use of a standardized periodontal examination protocol. Edentulous patients and patients with unstable health status were excluded. While this ensures data validity, it may limit generalizability to severe cases. Fourthly, the observed decrease in AGR associated with periodontitis could result from either reduced albumin levels or elevated globulin concentrations. This mechanistic ambiguity warrants further investigation in future studies. Fifth, These observational interactions, especially marginal ones, require prospective validation. Last but not least, the 2009–2014 NHANES database has missing data, therefore the results could be biased.

Conclusion

This cross-sectional study revealed AGR and periodontitis were found to be negatively associated within a certain threshold range. AGR can be readily obtained through routine blood tests. By incorporating this metric into risk assessment protocols, dental and healthcare practitioners may enhance early identification of patients at elevated risk for periodontal disease, enabling more proactive preventive and therapeutic interventions. further causal and longitudinal studies of the mechanism are needed to validate our findings.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

AGR

Albumin-globulin ratio

NHANES

National Health and Nutrition Examination Survey

FMPE

Full Mouth Periodontal Examination

MEC

Mobile Examination Center

PD

Periodontal probing depth

CAL

Clinical attachment loss

SD

Standard deviation

SLE

Systemic lupus erythematosus

CDC

Centers for Disease Control and Prevention

CDC-AAP

The Centers for Disease Control and Prevention and the American Academy of Periodontology

VIF

Variance inflation factor

GAM

Generalized additive model

BMI

Body Mass Index

PIR

Poverty Income Ratio

OR

Odds Ratio

CI

Confidence Interval

Author contributions

Jiajia Chen designed the study, analyzed the data, and drafted the manuscript. Xue Jiang and Na Liu contributed to the preparation of the graphs and data processing. Xingjin Chen contributed to the interpretation of the results. Yamei Li, Jukun Song, and Zhu Chen reviewed the article. Each author has given final consent and acknowledged responsibility.

Funding

Guizhou Provincial Science and Technology Fund (Fund No.: Qiankehejichu-ZK [2024] General 598). The Eighth Batch of High level Innovative Talents Project in Guizhou Province (Fund No.: Qianwei Renlingbanfa [2024] No. 3). Guizhou University 2024 Oral Medicine Wound Materials Cross team Project (Fund No.: Guidayan [2024] 28). Guizhou Province Science and Technology Achievement Transformation and Industrialization Plan Project（Fund No.:Qianke Chengguo LC [2025] General 061).

Data availability

The complete dataset is accessible through the NHANES database, which can be located at the following link: https://www.cdc.gov/nchs/nhanes.

Declarations

Conflict of Interest

The authors declare no conflict of interest.

Ethics approval and consent to participate

The NHANES study was approved by the National Center for Health Statistics (NCHS) Research Ethics Review Board. All methods were strictly carried out following relevant guidelines and regulations. This manuscript does not require additional ethical review.

Consent for publication

Not applicable.