Association between Periodontitis and HbA1c Levels in Non-Diabetic Patients: A Systematic Review and Meta-Analysis

Background: A high detection rate of diabetes among dental visitors has been reported recently. This systematic review aimed to evaluate the association between periodontitis and glycated hemoglobin (HbA1c) levels among non-diabetic individuals. Methods: The EMBASE, MEDLINE, Web of Science, Cochrane Library, PubMed, and Open GREY databases were searched, and observational studies published until 1st June 2023 were identified. A methodological quality assessment was conducted based on the original and modified versions of the Newcastle–Ottawa scale. Cohort, case–control, and cross-sectional studies that performed clinical periodontal examinations and measured HbA1c levels in non-diabetic adults were included. A meta-analysis was conducted to estimate the weighted mean difference (WMD) between individuals with and without periodontitis. Results: In total, 29 case–control and 5 cross-sectional studies were selected from 2583 potentially eligible articles. Among them, sixteen case–control and three cross-sectional studies with moderate to high quality were selected for the meta-analyses. The HbA1c levels in periodontitis patients were significantly higher than those in individuals with healthy periodontal conditions (WMD = 0.16; p < 0.001) among the non-diabetic populations. Conclusions: This study reveals a significant association between periodontitis and HbA1c levels in non-diabetic populations. Thus, HbA1c screening may be recommended to detect potential hyperglycemia in non-diabetic periodontitis patients.


Introduction
Periodontitis is one of the most common inflammatory diseases, characterized by the dysbiosis of periodontal bacteria and host immune response [1].Additionally, it is the most common cause of natural tooth loss in adults [2].The prevalence of periodontitis is nearly 50% among adults worldwide, and this proportion is even higher (70%) among those over 65 years of age, which is a cause of concern in terms of global health and the economic burden [3][4][5].Furthermore, periodontitis can elevate systemic inflammatory pressure and impact systemic health [6][7][8][9].
Diabetes mellitus (DM) is a common chronic metabolic disease characterized by sustained hyperglycemia; it is widely accepted as a major complication of periodontitis [6,10].
The underlying mechanisms between periodontitis and increased glycated hemoglobin (HbA1c) levels have been investigated for several decades.The evidence indicates that chronic systemic inflammation contributes to insulin resistance, which leads to increased HbA1c levels and, subsequently, diabetes and its complications [10,11].The levels of several inflammatory mediators are significantly increased in the circulation of periodontitis patients and decreased after effective periodontal debridement in periodontitis patients [12,13].For instance, it has been shown that tumor necrosis factor-α can interrupt insulin signaling mechanisms and reduce the entrance of glucose into cells [14].Therefore, the long-standing systemic inflammation in periodontitis patients may be a biological plausibility that promotes poor glycemic control and contributes to the progression of DM.Furthermore, periodontitis is known to have a negative impact on the outcomes of DM [15,16].Severe periodontitis can contribute to poor glycemic control and compromise the response to diabetes management [17].
The bidirectional relationship between periodontitis and DM has been demonstrated in the past [18].Diabetic patients have an approximately three-fold higher risk of developing periodontitis than non-diabetics [19].Patients with poor glycemic control are more likely to have a higher prevalence of severe periodontitis, which suggests that diabetes management may be a significant predictor of periodontitis development [19,20].
The biomarker HbA1c is used to assess the formation of sugar-coated hemoglobin and determine the presence of excessive glucose in the bloodstream.It has been utilized as a diagnostic tool to evaluate the average glycemic level.Evidence suggests that effective periodontal treatment can lower the HbA1c and circulating C-reactive protein levels [21][22][23][24].
Previous studies mainly focused on the association between periodontitis and the glycemic level in patients with DM [25,26]; however, evidence of this association among those without DM is scarce.Impaired immune response and enhanced systemic inflammation are commonly observed in patients with periodontitis and/or DM [27].Thus, it is vital to investigate the role of periodontitis in the onset and development of DM, especially in the non-diabetic population.The aim of this study was to identify observational studies that performed clinical periodontal examinations and measured the HbA1c levels in non-diabetic adults.A meta-analysis was performed to estimate the weighted mean difference (WMD) in HbA1c levels between individuals with and without periodontitis.The following PECO question was proposed: among individuals without diabetes, are HbA1c levels different between those with and without periodontitis?

Materials and Methods
The systematic review and meta-analysis were conducted based on the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [28].The proposal for the study was registered at the National Institute for Health Research, International Prospective Register of Systematic Reviews (PROSPERO; registration number, CRD42021249010).
Electronic searches of studies published until 1 June 2023 were conducted in five publication databases, including EMBASE, MEDLINE, Web of Science, Cochrane Library, and PubMed.There were no restrictions on the publication date and language, and unpublished articles were identified in Open GREY.One reviewer (SYY) performed the searches based on the search strategies developed by another reviewer (DZ) with the assistance of a university research librarian.The search criteria for all databases are presented in Table S1.References in the included publications were hand-searched to identify additional potential studies.
The two reviewers (DZ and SYY) independently screened the titles and abstracts of all the related publications and grouped the potentially eligible articles for further evaluation.The included papers were then carefully verified in full text, and reviewer disagreements were resolved through discussion.If an agreement could not be reached, a third reviewer (SLW) decided whether to include or exclude the article in this review.
The eligibility criteria for this systematic review were as follows: Patients (P): adults without DM, which was confirmed by self-reporting, medical history, or glycemic tests.Studies comprising non-DM individuals recruited to match the DM patients were not accepted.
Exposure (E): periodontitis diagnosed based on at least one clinical parameter; selfreporting was unacceptable.
Control (C): individuals without periodontitis.Study outcomes (O): HbA1c measurements; other glycemic test results were considered as secondary outcomes.
Study design (S): (i) observational studies or baseline data of interventional studies; (ii) those with no less than 10 subjects with periodontitis; and (iii) all publications, including in-press/online and unpublished studies.
Data extraction was performed independently by the two reviewers (DZ and SYY).It included the study characteristics (first author, publication year, and study design), population characteristics (sample size, demographic data, and description of subjects without DM), the definition of periodontitis, and outcomes (HbA1c levels and other parameters assessing glycemic status).Details about the origin of the subjects, periodontal examinations, covariates, and clinical assessment methods (blinding, training, and calibration) were also collected (Table S2).
The original and modified versions of the Newcastle-Ottawa Scale (NOS) were used to assess the quality of the cohort/case-control and cross-sectional studies (Table S3), respectively [29][30][31].A star was allotted if the study met the criteria for high quality, and the methodological quality of each study was rated based on the proportion of stars obtained.Only moderate-to high-quality papers with scores of 51% or higher were included in the meta-analysis [32].
Stata version 16.0 was used to conduct the meta-analyses based on the study design.Estimates of the pooled WMD were analyzed using the random-effects inverse-variance model.A forest plot was created to summarize the estimated effects and 95% confidence interval (CI) of all the included articles.The statistical heterogeneity was explored using the Chi-square-based Cochrane's Q statistic (p < 0.05) and I 2 (>50%).Egger's [33] and Begg's [34] tests and the funnel plots were used to demonstrate publication bias.Influence analysis was employed using the one-by-one elimination method to estimate the stability of the results.Meta-regression and subgroup analyses were performed to investigate the potential heterogeneity using study design, gender, body mass index (BMI), smoking status, and the severity of periodontitis as the independent variables.

Search Results
A total of 2583 related publications and 4 unpublished studies were identified from the six electronic databases.After removing 930 duplications, the titles and abstracts of the 1470 remaining articles were screened, and 187 were selected for full-text analysis.Finally, 34 studies that met the predetermined eligibility criteria were included in the systematic review (Figure 1).

Study Characteristics
Table 1 summarizes the characteristics of the 5 cross-sectional studies [35][36][37][38][39] and 29 case-control studies published between 2008 and 2023.These studies evaluated the association between periodontitis and HbA1c levels among subjects without diabetes.Most of the data were derived from convenience samples in hospitals.Two studies contained data from the Korean National Health and Nutrition Examination Survey conducted in 2012 and 2013-2015 [37,38], and one recruited the community sample [40].The number of participants in the included studies varied from 28 to 8341; two case-control studies provided details on the power calculation [41,42].The average age of the included subjects ranged from 29.8 to 56.5 years.One study [40] recruited only males, and six studies [43][44][45][46][47] did not provide information about the gender of the participants; the percentage of males ranged from 20% to 87.1% in the remaining twenty-seven studies.Two studies recruited only those with normal BMI [48,49], whereas five studies comprised non-obese patients only [46,[50][51][52][53]. Furthermore, 73.5% (25/34) of the studies excluded smokers or tobacco users.More than half (19/34) of the studies defined non-DM subjects as systematically healthy individuals, and one study categorized participants free of severe systemic diseases as the non-DM group [54].The remaining studies (14/34) used the HbA1c level, fasting plasma glucose (FPG) level, fasting blood sugar level, random blood sugar (RBS) level, and/or oral glucose tolerance test (OGTT) to describe the non-DM group; multiple glycemic testes were simultaneously applied in 8 of these 14 studies [35,43,48,50,53,[55][56][57].

Study Characteristics
Table 1 summarizes the characteristics of the 5 cross-sectional studies [35][36][37][38][39] and 29 case-control studies published between 2008 and 2023.These studies evaluated the association between periodontitis and HbA1c levels among subjects without diabetes.Most of the data were derived from convenience samples in hospitals.Two studies contained data from the Korean National Health and Nutrition Examination Survey conducted in 2012 and 2013-2015 [37,38], and one recruited the community sample [40].The number of participants in the included studies varied from 28 to 8341; two case-control studies provided details on the power calculation [41,42].The average age of the included subjects ranged from 29.8 to 56.5 years.One study [40] recruited only males, and six studies [43][44][45][46][47] did not provide information about the gender of the participants; the percentage of males ranged from 20% to 87.1% in the remaining twenty-seven studies.Two studies recruited only those with normal BMI [48,49], whereas five studies comprised non-obese patients only [46,[50][51][52][53]. Furthermore, 73.5% (25/34) of the studies excluded smokers or tobacco users.More than half (19/34) of the studies defined non-DM subjects as systematically healthy individuals, and one study categorized participants free of severe systemic diseases as the non-DM group [54].The remaining studies (14/34) used the HbA1c level, fasting plasma glucose (FPG) level, fasting blood sugar level, random blood sugar (RBS) level, and/or oral glucose tolerance test (OGTT) to describe the non-DM group; multiple glycemic testes were simultaneously applied in 8 of these 14 studies [35,43,48,50,53,[55][56][57].The definitions of periodontitis differed among the studies included in this review (Table 1).The American Academy of Periodontology (AAP) 1999 classification was used to diagnose periodontitis in 20 (58.8%) studies, whereas 4 studies [35,58,59,61] used the American Academy of Periodontology and European Federation of Periodontology (AAP/EFP) 2017 classification; meanwhile, 2 studies [37,38] used the Community Periodontal Index (CPI), and in 1 study [50], the disease was diagnosed based on the criteria from the Center for Disease Control and Prevention and the American Academy of Periodontology (CDC/AAP).Six sites per tooth were examined in more than half of the studies, whereas four sites per tooth were examined in two studies [47,66].Twelve studies did not describe the details of the specific sites examined.Full-mouth periodontal examinations were performed in more than half of the included studies (19/34), and the third molars 32.4% were skipped in 11 (32.4%)studies.A single examiner conducted the examinations in 55.9% (19/34) of the studies, and the training or calibration method was used in half of these studies.Four studies [36,53,55,66] employed two examiners and reported that the examiners were well-trained or calibrated.The examiners were blinded in four studies [40,48,59,60].The criteria for the minimum number of remaining teeth were mentioned in 82.4% (28/34) of the studies (range, [10][11][12][13][14][15][16][17][18][19][20].Patients who had undergone periodontal therapy within the past 90 days to 1.5 years were excluded in 76.5% (26/34) of the studies (Table S2).

Discussion
This systematic review is the first to summarize the evidence on the association between periodontitis and HbA1c levels in the non-diabetic population based on observational studies.All the studies included in this review were cross-sectional and case-control, suggesting low to moderate levels of evidence in terms of the hierarchy of the sources of evidence [69,70].The majority of the studies compared the HbA1c levels between individuals with and without periodontitis, but most failed to demonstrate significant findings.However, the meta-analysis in the current review shows that despite the high heterogeneity, the HbA1c levels in non-diabetic patients with periodontitis were significantly greater than those in individuals with healthy periodontal conditions.Subgroup analysis suggested that study design and gender as the potential sources of heterogeneity.Thus, two additional meta-analyses were conducted separately.Interestingly, the significant differences in HbA1c levels were observed between those with and without periodontitis remained in the cross-sectional and case-control studies.
Statistical analyses comparing differences between individuals with or without periodontitis were not conducted in two [35,39] out of the five cross-sectional studies.Of the remaining three studies, two from Korea [37,38] found a significant association between periodontitis and HbA1c levels; the data were obtained from the Korean National Health and Nutrition Examination Survey.One of these Korean studies conducted multivariate logistic regression to adjust the confounders, but no independent association was detected [37].Although the remaining single study from Serbia [36] failed to report a significant relationship between periodontitis and HbA1c levels, it showed an association between periodontitis and FPG, which may result from the variability in FPG relative to HbA1c [71].Diagnostic screening can be performed using FPG or HbA1c levels [72].HbA1c reveals the average blood glucose level over the previous three months because it remains stable on a daily basis [73].Therefore, it is not always completely correlated with the average blood glucose level [74].Moreover, the HbA1c level is less impacted by diet when compared to other glycemic biomarkers.Thus, this assay can be conducted without fasting.
Statistical analysis was performed in 96.4% (27/28) of the case-control studies, and 33.3% (9/27) reported statistically significant associations between periodontitis and HbA1c levels.One revealed that covariates, such as age, gender, BMI, and smoking habit, could mask the significant correlation between periodontitis and HbA1c [41].These risk factors, which are common for both periodontitis and DM, could impact the research findings.Notably, the covariates were considered in almost all the studies included in the current review; however, they were addressed in only 13.8% (4/29) of the studies.One study employed the multivariate linear regression model [41], one matched the gender during the recruitment of participants [51], and the other two studies divided the subjects based on the BMI and analyzed the association between periodontitis and HbA1c levels in the subgroups [47,66].More than half of the studies (58.6%; 17/29) did not employ normality tests before performing the parametric tests.Incorrect selection of the statistical method during statistical analyses can impact the internal validity of the results.Furthermore, blinding, training, and calibrating the investigators can reduce information bias [75].However, the investigators were blinded in only four studies [40,48,59,60], and training and calibration were performed in five studies [40,49,60,61,64].Two trained and calibrated investigators were employed in three studies, which may have reduced the measurement bias during the operational process [53,55,66].
The origin of the sample can impact the external validity of this systematic review.The two cross-sectional studies with significant association were conducted in Korea [37,38], and almost all the case-control studies with significant association (7/8) were conducted in India [42][43][44][45][46]50,54].The majority of the studies used convenience samples from hospitals, except for three, which recruited community samples [37,38,40].Therefore, the findings of the review should be interpreted with caution.
This review has some limitations.Firstly, only cross-sectional and case-control studies were included.The lack of a prospective study can weaken the strength of evidence in this study.Thus, the inclusion of cohort studies to detect a potential directional relationship is warranted in the future.Secondly, 44.1% of the studies included in this review were of low quality.The exclusion of these articles from the quantitative analysis decreased the sample size and the diversity of the demographic characteristics, which may have impacted the external validity of the results of the current meta-analysis.Thirdly, the high heterogeneity caused by variations in the definitions of non-diabetes and periodontitis may have impacted the real glycemic and periodontal condition of subjects and, consequently, influenced the comparability and confidence in the pooled estimates.The recruitment of subjects with various periodontal statuses is required to validate the findings of this review and provide evidence on the potential dose effect of periodontitis on HbA1c levels.Finally, the majority of studies recruited convenience samples.Thus, additional community samples, including subjects from various races, are expected to make the findings more representative of the broader population.

Conclusions
The results from the meta-analyses identified significantly higher levels of HbA1c among non-diabetic periodontitis patients compared to those in non-diabetic individuals with healthy periodontal conditions.Periodontitis and DM can act as comorbidities; therefore, prospective studies using representative community samples should be conducted to improve the level of evidence and clarify whether the association is directional or nondirectional.Nonetheless, the present meta-analysis shows that a healthy periodontal status has a beneficial effect on the glycemic condition in healthy individuals.The findings of this study may help general dental practitioners and periodontists to focus on general health and encourage them to collaborate closely with medical professionals.The emerging co-management care scheme will help improve the oral health and the general well-being of patients who visit the dental clinic.

Figure 1 .
Figure 1.Flowchart of the study selection.

Figure 1 .
Figure 1.Flowchart of the study selection.

Figure 3 .
Figure 3. Subgroup analysis of HbA1c levels between non-DM subjects with and without periodontitis based on study design, smoking status, gender, severity of periodontitis, and BMI.BMI: body mass index; WMD: weighted mean difference.

Table 1 .
Data collection of cross-sectional and case-control studies.
* Study design was based on the parts we focused on.AAP 1999: The 1999 International Workshop for a Classification of Periodontal Diseases and Conditions; AAP/EFP: The American Academy of Periodontology and European Federation of Periodontology; ADA: American Diabetes Association; ANOVA: analysis of variance; BL: bone loss; BMI: body mass index; BOP: bleeding on probing; CAL: clinical attachment level; CDC: the Centers for Disease Control and Prevention; CEJ: cemento-enamel junction; CPI: community periodontal index; DM: diabetes mellitus; FBS: fasting blood sugar; FPG: fasting plasma glucose; GI: gingival index; HbA1c: glycated hemoglobin; IQR: interquartile range; KNHANES: Korean National Health and Nutritional Examination Survey; mSBI: modified sulcus bleeding index; OGTT: oral glucose tolerance test; OHI-S: oral hygiene index-simplified; PD: probing depth; PI: plaque index; RBS: random blood sugar; SD: standard deviation