Introduction

Non-alcoholic fatty liver disease (NAFLD) is defined as a condition characterized by the presence of fat in the liver of individuals who do not consume alcohol1. NAFLD is associated with hypertension, diabetes, and dyslipidemia1,2. Obesity is one of the factors of these underlying conditions, and being obese significantly increases the risk of developing NAFLD. NAFLD is a condition with a high global prevalence in which healthy liver cells are replaced by scar tissue due to chronic inflammation. This results in liver fibrosis, which can potentially cause cirrhosis and hepatocellular carcinoma1,2. Severe liver disease can cause various adverse effects and significantly reduce a patient’s quality of life. Therefore, it is imperative to identify signs of NAFLD at an early stage. Liver biopsy is the standard diagnostic method for diagnosing NAFLD3. However, since performing biopsies in all patients with liver disease is challenging, various screening tools have been developed4,5,6.

Periodontitis is a chronic inflammatory condition caused by oral bacteria and is characterized by inflammation of periodontal tissues and loss of alveolar bone7. Periodontitis has been demonstrated to be associated with systemic diseases, including but not limited to diabetes, dementia, and malignant tumors8,9,10,11,12. Therefore, the prevention of periodontitis and maintenance of a good oral environment can improve the overall health. Recent studies have established the correlation between periodontitis and NAFLD10,13,14,15. A large-scale survey conducted in China revealed a correlation between the number of missing teeth in male participants and NAFLD evaluation using ultrasonography16. In the United States, research has been conducted on the relationship between NAFLD assessment tools and periodontitis17. Furthermore, studies have been conducted on Porphyromonas gingivalis (P. g), a major periodontal pathogen, and NAFLD. In previous animal experiments, mice were fed a high-fat diet for 12 weeks and then infected with P.g through the dental pulp cavity18. Mice infected with P.g show more marked fatty degeneration and hepatocyte fibrosis than uninfected mice. On the other hand, few studies focus on the relationship between NAFLD and the probing pocket depth (PPD) or bleeding on probing (BOP). Furthermore, previous studies have only evaluated NAFLD as a single index. Although several indices have been reported for evaluating NAFLD, few studies have compared them. In contrast, a comparison of multiple indicators is needed to elucidate the effectiveness of assessing their association with periodontitis. This cross-sectional study aimed to verify the associations between NAFLD, evaluated by multiple indices, and the combination of PPD and BOP, which indicate the progression and inflammatory state of periodontitis, using the Community Periodontal Index (CPI)19.

Methods

Study population

This study examined data from the initial examinations of participants who underwent medical and dental checkups simultaneously at Asahi University Hospital between April 2013 and March 2022. The analysis was conducted on 2,453 participants (1,403 males and 1,050 females, mean age 46.9 years). The exclusion criteria were missing data, those for whom the periodontal status could not be evaluated, and those with diabetes (Fig. 1). This study was conducted with the approval of the Asahi University Ethics Committee (No. 27010 and 30018) and was performed in accordance with the Declaration of Helsinki.

Fig. 1
Fig. 1
Full size image

Flowcharts showing patient selection.

Demographic status

We collected data on the participants’ basic information, including age, sex, body mass index (BMI), waist circumference, regular exercise20, daily alcohol consumption (drinking at least 1 drink/day)21, and current smoking status (smoking at least 1 cigarette/day)22. In view of the absence of precise data pertaining to the type of tobacco (heated or combustible), all tobacco samples were treated as combustible. In addition, the results of the biochemical tests were used to collect the serum values of triglycerides, γ glutamyl transpeptidase (γGTP), aspartate transaminase (AST), alanine transaminase (ALT), platelet (PLT), and albumin (ALB).

Oral health status

The oral cavity was examined by total five dentists who underwent calibration training. The examinations were consistently performed by five dentists. The state of periodontal disease was evaluated based on clinical examinations using the CPI19. The dentists achieved inter-examiner reliability through repeated calibrations until a consensus (kappa value) about periodontitis of over 0.8 was reached. Following this confirmation, each dentist evaluated the oral cavity of all dental examination participants. Subjects were defined as having periodontitis if they exhibited at least one PPD ≥ 4 mm or greater in an index tooth (17, 16, 11, 26, 27, 37, 36, 31, 46, 47) as measured using a CPI probe. BOP was defined as bleeding from the gums during a periodontal examination19. Tooth loss was defined as the loss of one or more teeth. In addition, the patients were questioned on the frequency of tooth brushing per day and regular dental check-ups at least once a year23.

NAFLD assessment tool

In this study, three indicators were used to evaluate the liver function.

Fibrosis-4 (FIB4)

FIB4 is a scoring system that combines age and blood test results to predict the risk of liver fibrosis24,25. The cutoff values were defined as > 1.3 and > 2.67 based on the scores calculated using the age, platelet count, AST, and ALT levels from previous studies6,26.

NAFLD fibrosis score (NFS)

NFS was developed to assess the risk of liver fibrosis in NAFLD27. The score is calculated using the same indicators as FIB4; however, BMI and the presence or absence of diabetes are added as additional factors. The cutoff values are defined as ≥-1.455 and > 0.67526.

Fatty liver index (FLI)

FLI was developed in Italy to assess the risk of fatty liver and is calculated using BMI, waist circumference, triglycerides, and γGTP28. In this study, the cutoff values were defined as ≥ 30 and ≥ 60.

Statistical analysis

First, the participants were stratified according to the presence or absence of periodontitis. The characteristics were then compared using the Fisher’s exact test, and chi-squared test. Several continuous variables exhibited nonparametric distributions, thus the Mann-Whitney U test was used. Finally, the relationship between the thresholds of the NAFLD assessment tool and both PPD and BOP was examined using binary logistic regression analysis, and subgroup analysis was performed to account for periodontitis activity based on the presence or absence of BOP. To examine the association between PPD and NAFLD risk, cutoff values were adopted for each NAFLD tool (FIB-4: >1.3 and > 2.67, NFS: ≥-1.455 and > 0.675, FLI: ≥30 and ≥ 60). The variables were input into the logistic model using the forced input method, with reference to previous research29. The model fit was evaluated using the Hosmer–Lemeshow test30. Analyses were conducted using Stata17 (Stata Corporation, College Station, TX, U.S.A.), and the statistical significance level was set at 5%.

Results

Table 1 shows the characteristics of participants with and without periodontitis. The participants with periodontitis were more likely to be obese, male, and aged ≥ 40 years. Participants with periodontitis were more likely to have BOP and lose teeth than those without periodontitis. Participants with periodontitis were more likely to have lower proportions of brushing frequency ≥ 2 times/day than those without periodontitis. Several participants with periodontitis were more likely to have higher triglyceride, γ-GTP, and ALT values than those without periodontitis. Several participants with periodontitis consumed alcohol daily or were current smokers. Table 2 shows the relationship between the NAFLD assessment tool and PPD. Participants with periodontitis (PPD ≥ 4 mm) were more likely to have higher values for the NAFLD assessment tool than those without periodontitis. Similar results were obtained in individuals with BOP. Table 3 shows the relationship between the NAFLD assessment tool and both PPD and BOP. A PPD ≥ 6 mm was significantly associated with NFS > 0.675 (OR: 4.11; 95% CI: 1.06–15.99) and FLI ≥ 60 (OR: 1.76; 95% CI: 1.12–2.76). A BOP was significantly associated with FIB-4 > 1.3 (OR: 0.71; 95% CI: 0.51–0.96) and FLI ≥ 30 (OR: 1.53; 95% CI: 1.09–2.16). Table 4 shows the results of the subgroup analysis using the BOP. No significant association was found between PPD and the NAFLD assessment tool in groups without BOP. A significant association was found between a PPD ≥ 6 mm and FLI ≥ 60 (OR: 1.68; 95% CI: 1.05–2.70) in groups with BOP.

Table 1 Characteristics of the target population stratified according to periodontitis.
Full size table
Table 2 Characteristics of the target population liver assessment tool stratified according to probing pocket depth.
Full size table
Table 3 Relationship between NAFLD assessments and probing pocket depth using binary logistic regression analysis.
Full size table
Table 4 Relationship between NAFLD assessments and probing pocket depth by subgroup analysis using binary logistic regression analysis.
Full size table

Discussion

This study examined the relationship between NAFLD and periodontitis as assessed by multiple indicators. Periodontitis activity was also evaluated in the presence or absence of BOP, and the relationship between deep periodontal pockets and NAFLD was clarified. The analysis revealed a relationship between FLI, periodontitis, and BOP. The strength of this study is that it used BOP to evaluate the presence or absence of inflammation, making it the first study to examine the association between periodontal inflammation and NAFLD. BOP indicates inflammation of periodontal tissues31,32and serves as a valuable indicator of periodontitis activity. In addition, an association was found between deep periodontal pockets of ≥ 6 mm and FLI and NFS. In subgroup analysis, deep periodontal pockets of ≥ 6 mm were only associated with FLI when periodontitis is in an active state. Therefore, it can be concluded that there is a correlation between the degree of inflammation caused by periodontitis and NAFLD. The FLI formula is characterized by use of BMI and abdominal circumference in addition to triglycerides and γGTP. Kim et al. investigated the relationship between CPI and FLI in 4,272 Koreans and found that the prevalence of periodontitis increased as the FLI increased29. Inflammation caused by periodontitis results in the production of several inflammatory cytokines. These cytokines may exert their effects on the liver through oral intake or bloodstream infections, potentially exacerbating NAFLD13. In addition, a positive correlation has been reported between the severity of periodontitis and triglyceride levels33,34. Therefore, an increase in neutral fat levels triggered by periodontitis may affect its association with FLI.

This study found significant associations between PPD, NFS, and FLI. Similar results have been reported in previous studies17,29. However, no significant association was observed between PPD and FIB4, which may be related to the BMI. Periodontitis has been linked to obesity, and both conditions can worsen each others status35. The FIB4 formula differs from the NFS and FLI in that it does not consider BMI. Therefore, we believe that there is no relationship between periodontitis and FIB4. Conversely, differences in diagnostic accuracy among various tools have been reported based on BMI. Recent studies indicate that while FIB-4 is independent of BMI, NFS showed low sensitivity for lean patients with NAFLD36. BMI significantly impacts NAFLD and periodontitis, necessitating future research that incorporates body composition into analyses. In addition, this study’s subjects were a health-conscious group who visited hospitals for medical checkups. therefore, the prevalence and characteristics of NAFLD differed from those of the target population in previous studies. The global prevalence of NAFLD is estimated to be 25%; the prevalence of NAFLD in this study was 18%, indicating that the study population had a low incidence of the disease2,37. In a previous US study on periodontitis and NAFLD, the prevalence of NAFLD was 24%17. Differences in the association between periodontitis and various assessment tools arise because each tool measures different parameters. While FLI is a screening tool for NAFLD, FIB-4 and NFS are designed to assess liver fibrosis. Therefore, the FLI screens in a manner that encompasses both FIB-4 and NFS, regardless of NAFLD severity. This implies that the FLI may detect more positive cases than FIB-4 or NFS, potentially leading to a stronger tendency to show a significant association with periodontitis.

A systematic review of periodontitis and NAFLD reported an association between P.g, a major periodontal pathogen involved in the disease, and NAFLD38. The release of endotoxins and lipopolysac40charides due to the swallowing of P.g induces the production of inflammatory cytokines and contributes to the development of NAFLD by stimulating the liver39. A longitudinal study examining clinical attachment level (CAL) in relation to NAFLD progression reported that patients with periodontitis with a baseline CAL of ≥ 3 mm were at an increased risk of developing NAFLD40. Patients susceptible to periodontitis lose up to 1 mm of CAL per year41; thus, patients who have lost more than 3 mm of CAL may experience long-term moderate or severe periodontitis. P.g affects the progression of periodontitis, and the number of bacteria increases as the PPD and inflammation increase42. The detection rate is exceptionally high in areas where periodontitis is active43,44,45,46, the reason for the association between deep PPDs of 6 mm or more and NAFLD in this study may be the influence of the large number of P.g. that are hidden in deep periodontal pockets. The detection rate of P.g is 11.2 times higher in patients with periodontitis than in healthy individuals47. The fact that no significant association was found between PPD and NAFLD in subjects without BOP in this study also supports the above discussion. A more detailed examination of the relationship between the number of P.g detected and NAFLD is necessary. This study is cross-sectional and therefore cannot establish causation. Future research should consider longitudinal studies incorporating inflammation due to periodontitis and microbiological investigations.

This study has several limitations. First, as this was a cross-sectional study, causal relationships could not be explained. Although several longitudinal studies on periodontitis and NAFLD have been conducted, the full extent of this relationship remains unclear48. Furthermore, this analysis relies solely on alternative indicators for NAFLD diagnosis, and differences in sensitivity and specificity may lead to variations in the observed association. It is necessary to continue accumulating findings through long-term studies in the future. Moreover, the high-risk group of NAFLD was small, while the high-risk group for FIB4 (> 2.67) and NFS (≥ 0.675) was less than 2% of the total. Regarding the significant association with PPD ≥ 6 mm and BOP, the results should be interpreted with caution, considering the small sample size and wide confidence interval. The group without BOP in the CPI indicates a limited sample size, which may have influenced the analysis outcomes. Thus, the results of the multivariate regression analysis need to be verified in a large population with a high prevalence of NAFLD and periodontitis. In addition, social factors, such as educational background and income significantly affect an individual’s health status49, and have not been adequately adjusted for in this study. In the multivariate regression analysis, these factors were not treated as covariates, limiting the analysis. The exclusion of these factors from the current analysis is likely to have a significant impact on the findings, and subsequent studies should endeavor to incorporate a more comprehensive range of factors into the analysis. Finally, the BOP level was used as an indicator of inflammation. However, this is a binary variable that only indicates whether bleeding is present. therefore, it lacks sufficient information. Since inflammation is assessed solely based on the presence or absence of bleeding, the results of this study may overestimate gingival inflammation. If we could use a quantitative and specific indicator, such as periodontal inflamed surface area (PISA)32, it would be possible to evaluate the relationship between inflammation due to periodontitis and NAFLD. It is necessary to consider a comprehensive diagnosis and evaluation of periodontitis that does not rely on CPI-based indicators.

In conclusion, our findings suggest that periodontitis and inflammation of periodontal tissues are significantly associated with NAFLD, as assessed using the FLI in Japan. These results suggest that using the FLI as a liver assessment instrument may provide more informative insights when examining the association between periodontitis and NAFLD.