Introduction
Periodontitis is a multifactorial chronic inflammatory condition caused by an imbalanced interaction between periodontal microbiota and host inflammatory response, which seems to have a bidirectional link with systemic inflammatory diseases [
1‐
3]. The host immune response to periodontal infection can be modified by many genetic and environmental factors, among which gene polymorphisms (e.g.,
IL1) and chronic non-infectious diseases (e.g., diabetes, obesity, etc.) play a significant role [
1,
4,
5]. Severe periodontitis affects 23.6% of the global population [
6]. In addition to causing tooth loss and a decline in quality of life, periodontitis imposes an enormous socioeconomic burden [
7]. Impaired immune and metabolic response induced by periodontal pathogens is the critical feature in the pathogenesis of periodontitis [
8,
9]. However, the underlying mechanisms remain unclear, making early prevention challenging.
Oxidative stress has been considered as an important mechanism in the progression of periodontitis [
10]. It is formed when the reactive oxygen species overproduced by immune cells when infections cannot be neutralized by the antioxidant defense system, causing lipid, protein and DNA damage [
11]. Historically, uric acid (UA) was regarded as an important radical scavenger among the antioxidant pools. However, its anti-oxidative roles appear to be restricted to hydrophilic environments only [
12]. Recent studies have identified hyperuricemia (i.e., blood UA levels > 6.5 mg/dL) as risk factors for a variety of inflammatory conditions such as gout, diabetes, metabolic syndrome and cardiovascular diseases [
13‐
15]. Specifically, UA can exhibit pro-oxidative effects in certain environments. For instance, UA can react with peroxynitrite to form radicals [
16,
17]. It can also enhance intracellular superoxide production by elevating nicotinamide adenine dinucleotide phosphate (NADPH) oxidase activity [
18]. Hence, the current consensus acknowledges that a pathological elevation of UA levels (i.e., hyperuricemia) represents a pro-inflammatory, -oxidative and -osteoclastic state [
12].
UA was identified as an anti-oxidative parameter in previous research [
19,
20] on periodontitis. Alterations of UA levels in blood have been associated with the presence or severity of periodontitis [
21,
22]. Interventions with urate-lowering drugs have shown beneficial effects on animals with periodontitis [
23,
24]. However, a re-assessment of their relationship is urgently required because the results of relevant studies are highly contradictory. Firstly, it was discovered that blood UA levels in periodontitis patients were either upregulated, downregulated or unchanged in comparison to controls [
25‐
27]. UA levels in the saliva of periodontitis patients were either reduced or unchanged [
28,
29]. Secondly, the presence of periodontitis was associated with an increase in UA levels in the blood but a decrease in saliva UA levels [
30]. Thirdly, periodontal treatment increased salivary UA levels but decreased blood UA levels compared to baseline [
31]. Integrating the contradictory data would therefore necessitate a systematic review of previous findings. In addition, it would be useful to answer an unresolved question of whether hyperuricemia and periodontitis may be linked [
24,
31,
32]. The present systematic review and meta-analysis focuses on the question of whether periodontitis patients have altered UA levels in blood, saliva, and gingival crevicular fluid (GCF) compared to controls.
Materials and methods
The systematic review and meta-analysis were registered in PROSPERO (no. CRD42020172535) and prepared in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) statement [
33]. The present study adhered to the PECO principles: P (population) was participants in systemic health or with gout/hyperuricemia but without systemic complications/comorbidities; E (exposure) was patients with periodontitis; C (comparison) was periodontally healthy controls; and O (outcome) was UA levels in blood/saliva/GCF.
Eligibility criteria
The studies should be cohort/cross-sectional/case–control study, or randomized/non-randomized controlled trials. For interventional clinical trials, the baseline data before intervention were deemed as information originated from observational studies. Case reports/series, animal/in-vitro studies, narrative/systematic reviews, conference abstracts, editorials, letters and comments were excluded. The following were the selection criteria for full-text analysis.
Definition of periodontitis and periodontal health control
Plaque-induced periodontal destruction should be measured by periodontal probing or radiographs to diagnose periodontitis. From an epidemiological standpoint, the periodontal healthy control concerned in the present study would include not only clinical periodontal health but also mild (Stage I) or localized periodontitis.
Inclusion criteria
-
a) Periodontitis should be clearly defined or stated;
-
b) periodontitis and control groups should have UA levels in blood, saliva or GCF measured;
-
c) participants in systemic health or with gout/hyperuricemia but without systemic complications (e.g., renal dysfunction including abnormal glomerular filtration rate);
-
and d) participants in interventional studies should have baseline/pre-treatment UA levels recorded.
Exclusion criteria
-
a) studies without periodontitis patients or control groups;
-
b) female subjects during pregnancy;
-
c) subjects with potential comorbidities shared by hyperuricemia and periodontitis, such as cardiovascular diseases, diabetes, osteoporosis, chronic kidney disease, metabolic syndrome and obesity [
31];
-
d) subjects with other potential conditions associated with altered purine/UA metabolism such as inflammatory bowel diseases [
34,
35], hyperparathyroidism [
36] and vitamin D deficiency [
37];
-
e) subjects with significant systemic diseases or conditions such as cancers, liver cirrhosis, organ transplantation, etc.;
-
and f) subjects receiving antibiotics or anti-inflammatory or urate-lowering drugs or have had periodontal treatment in the past three months.
Highly sensitive electronic search was conducted in four databases, including PubMed, Scopus, Web of Science and Embase with no language restriction (update to January 28, 2023). The following search model was constructed using Boolean operators. For exposure, “periodontal diseases” and “periodontitis” were used, while “apical periodontitis” was excluded. Regarding UA-tested samples, the keywords “blood”, “serum”, “plasma”, “circulation”, “GCF”, “gingival crevicular fluid” and “saliva” were used. As for outcome, the keywords “UA”, “urate”, “purine”, “hyperuricemia”, “gout” and “antioxidant” were used. The advanced search was based on each database-specific search strategies (see the Supporting Information). The search terms employed were either medical subject headings (MeSH) terms or keywords classified under general category (title/abstract/keywords).
Study selection
Initial assessment of titles and abstracts was conducted by two independent reviewers (L.Y. and L. Z.), followed by full-text screening of the eligible articles. The disagreement was discussed until consensus was reached or the supervisor arbitrated (T. Y.). During the process, any study that failed to meet the eligibility criteria was excluded and the reason was formally recorded in detail. Inter-examiner agreement for abstract review and full-text screening was assessed using the
κ test (
κ values > 0.75 and < 0.4 indicated high and low consistency, respectively) [
38].
Quality assessment and quality of evidence
Newcastle–Ottawa Scale (NOS) and the Agency for Healthcare Research and Quality (AHRQ) methodology checklist were used to assess the methodological quality of case–control studies and cross-sectional studies, respectively. As for randomized controlled trials, only the baseline data before intervention were collected and the research design was deemed as an observational one. Namely, RCTs were also evaluated by NOS during quality assessment. Quality assessment was conducted by two reviewers independently (L. Y. and L. Z.). Inter-examiner agreement for quality assessment were assessed by the κ test.
The total quality score on the Newcastle–Ottawa Scale ranged from 0 to 9 stars (7 to 9, 5 to 6 and < 5 stars indicated high, moderate and low quality, respectively) [
39]. Regarding ARHQ, studies with 8 to 11 points, 4 to 7 points, and 0 to 3 points, respectively, were deemed to be of high, moderate and low quality [
40]. Disputes would be discussed with a third reviewer (T. Y.).
The certainty of evidence was evaluated following the Grade of Recommendations Assessment, Development and Evaluation (GRADE) method with GRADEprofiler (v 3.6, the GRADE working group) [
41,
42]. The evidence quality of each outcome was rated as high, moderate, low and very low.
Once these studies were identified, two reviewers independently (L. Y. and L. Z.) extracted the information including bibliometric information (the names, e-mail addresses and institutions of authors, publication date and journals of articles, etc.), characteristics of study design (study place/type, diagnostic criteria of periodontitis/control groups and sample size, etc.), demographics of participants (sex, age, race and smoking habit [
43,
44], periodontal parameters (probing pocket depth, gingival/plaque index, bleeding on probing, clinical attachment loss, radiological alveolar bone loss, etc.), collection methods for saliva (resting/stimulated) or blood (plasma/serum) or GCF (paper strip/point), detection methods for UA (enzyme-based colorimetric method or gas chromatography/mass spectrometer, etc.), statistics for UA levels (means and standard deviations (
SDs)/errors (
SEs)) in periodontitis and control groups. Data extraction forms were cross-checked to verify accuracy and consistency of the extracted data. All data were checked by the third author (T. Y.) and disagreements were resolved by discussion. Three emails were sent to the corresponding authors of the included articles to request the raw data, which include the gender and age of the participants. Unanswered or undelivered emails were regarded as having no response.
Data conversion and preprocessing
The UA concentrations may be recorded in mg/dL, µmol/L, mmol/L or relative units. The unit reported in this systematic review was mg/dL (1 mg/dL = 59.48 µmol/L) [
45]. According to the new classification for periodontal diseases [
46], chronic and aggressive periodontitis are no longer distinguished from one another, and the data presented here were compiled in accordance with this classification. The data of UA in periodontitis at different stages or in a single type of sample collected using different methods were also combined based on the means,
SDs and sample size of the subgroups. If some studies reported
SEs instead of
SDs, the former was converted to the latter (
SD =
SE ×
\(\surd n\)) [
47]. If the data of UA in a study were shown in bar charts without detailed values, the heights of three repeats were measured using a digital ruler (v1.8.0, Image J, National Institutes of Health, Bethesda, MD, USA) to determine the absolute values.
Statistical analyses
Statistical analyses were performed using a commercial software (v 14.0, STATA, Stata Corporation, College Station, TX). The UA levels are presented as means ±
SDs. The distribution of potential confounding variables (such as gender and age) that may influence the UA levels was compared [
48]. Meta-analyses are displayed as forest plots. It would be calculated as weighted mean differences (
WMDs) or standardized mean differences (
SMDs) for continuous outcomes. Statistical heterogeneity was estimated by Cochrane’s
Q test and Higgins’s
I2 test (
I2 = 0, no heterogeneity;
I2 < 50%, low heterogeneity;
I2 < 75%, moderate heterogeneity;
I2 > 75%, high heterogeneity) [
49]. If a statistically significant heterogeneity was found, a random-effect model was used; otherwise, a fixed-effect model was applied. A subgroup analysis was conducted based on the collection methods of testing samples (i.e., serum vs. plasma).
P < 0.05 was considered statistically significant. Publication bias was assessed by Egger’s tests [
50]. If publication bias was detected, the Duval and Tweedie trim-and-fill method was used to make adjustments [
51]. If statistical heterogeneity was detected, a sensitivity analysis was conducted by excluding each study individually to determine whether the heterogeneity changed significantly.
Discussion
The present study explored the association between oral/blood UA and periodontitis using systematic review and meta-analysis. We discovered a positive correlation between periodontitis and blood UA content. The increased blood UA might result from accelerated purine degradation in both periodontal tissues and systemic organs. Patients with periodontitis exhibit accelerated purine catabolism and enhanced xanthine oxidoreductase expression in the periodontium [
30,
67]. Enhanced secretion of UA has been observed in immune cells stimulated by periodontal pathogens [
67,
68], and in the gingiva of mice with periodontitis [
9]. Given that ubiquitous xanthine oxidoreductase is sensitive to inflammation and oxidative stress [
69], periodontitis-induced low-grade systemic inflammation may accelerate purine catabolism in distant organs. For instance, periodontal infection is associated with increased UA levels in the liver and feces in rodents [
70,
71]. However, it should be noted that the present study only supported a positive association between blood UA and periodontitis when normouricemia was present. UA has been considered as a potent antioxidant in blood, especially at a physiological level [
12,
72,
73]. Nonetheless, such a theory has not been rigorously examined in the field of periodontology. It is unknown whether elevated UA levels in the circulation or periodontal vessels in a state of normouricemia are beneficial or detrimental for periodontitis.
Almost no direct evidence has investigated the relationship between periodontitis and hyperuricemia or gout. One cross-sectional study identified hyperuricemia as a protecting factor for periodontitis [
32]. However, the outcome was measured by questionnaire and only represented a retrospective history of periodontitis. Gouty patients generally have increased abundance of periodontal pathogens (e.g.,
Prevotella Intermedia) [
74]. Given that the highly prevalent hyperuricemia is considered an emerging risk factor for many inflammatory comorbidities of periodontitis (e.g., cardiovascular disease, diabetes, and chronic kidney disease) [
75], the association between hyperuricemia and periodontitis would make for an intriguing research topic. Recent research indicates that UA plays a pathological role in periodontitis. For instance, systemic injection of UA aggravated alveolar bone loss in mice with periodontitis [
24]. The urate-lowering drug febuxostat alleviated experimental periodontitis induced by molar ligation in rats [
23]. Moreover, non-surgical periodontal treatment for periodontitis appeared to reduce urate levels in the circulation [
76]. Taken together, pathologically elevated blood UA appears to contribute to the progression of periodontitis.
Interestingly, the change of UA in saliva shows an opposite trend compared to that in blood in periodontitis patients versus controls. UA has been considered as a major antioxidant in saliva, accounting for ~ 70% of the total antioxidant capacity [
77]. However, the levels of antioxidants, including UA, in saliva do not appear to correspond to those in blood [
19]. A reduced level of salivary UA in periodontitis may be due to either increased consumption or decreased production. Increased consumption of salivary UA may be a result of the enhanced oxidative stress caused by periodontal infection. In the presence of oxidative stress, UA may be oxidized by reactive oxygen species into allantoin in the absence of uricase [
78]. Additionally, UA can serve as a substrate for the synthesis of bacterial components [
31], a process that may be accelerated by an increase in dental plaque accumulation. Substances that inhibit the purine oxidation activity of the UA-producing enzyme (i.e., xanthine oxidoreductase) may lead to a decrease in salivary UA production. For instance, there is an increase in the demand of nitrate and its metabolites nitrite and nitric oxide in saliva for bactericidal and anti-inflammatory purposes in periodontitis [
79]. These substances may competitively inhibit the purine-oxidizing activity of xanthine oxidoreductase, considering that xanthine oxidoreductase also has a nitrate/nitrite reductase activity [
80,
81]. In summary, the mechanisms underlying the reduction of salivary UA in periodontitis are unclear, and further investigation is warranted.
The present systematic review included only one article that found a decreased level of UA in GCF in periodontitis patients. The result was consistent with the findings from some studies that were not included in the review [
30,
82]. The changes of UA in GCF resembled those in saliva. Different from saliva that comes from salivary glands, GCF is serum transudate in healthy periodontium or from inflammatory exudate during periodontal diseases [
83]. In the context of periodontitis with an increased UA content in the blood, a decrease in the exudation of UA from circulation into GCF was unlikely. A more probable scenario, similar to that of saliva, would be that subgingival microbiota could enhance UA consumption. Moreover, it should be noted that some urate transporters (SLC2A9 and SLC22A12) showed increased gene expressions in gingival tissues from periodontitis patients [
84], which could be a potential contributor or confounder for an altered UA concentration in periodontal pockets.
There were several potential confounding variables that could have affected the applicability of this study. The data of sex, age and smoking, which have been associated with hyperuricemia or altered tissular UA in the past studies [
85‐
87], were partly missing. We failed to obtain the original data from the corresponding authors in relevant studies. However, the distribution of age and sex was even in most or all of the studies (age, 6/11; sex, 11/11). Smoking was unlikely to have a significant impact on the results given that most (12/15) of the studies excluded smokers. The data of body mass index (a common cofounding factor), which has been reported to significantly impact blood UA [
88,
89], were completely missing. Consequently, it was unclear how much and to what extent body mass index influences the results.
The definition and control of periodontitis could be another source of bias. Periodontitis is frequently defined differently in clinical studies and systematic reviews. It may diminish the comparability of the studies included in the present systematic review. No studies can make a diagnosis without probing depth, clinical attachment loss, or radiographic bone loss, despite the fact that diagnostic thresholds for periodontitis vary. In the majority of the included studies in this review, the diagnosis was based on these periodontal parameters. Specifically, a clinical misdiagnosis regarding whether or not it was periodontitis rarely occurred. Instead, the discrepancies mainly occurred due to the severity and extent of periodontitis. Indeed, the severity of periodontitis appears to be correlated with UA levels in blood or saliva [
19,
21,
31,
90]. For instance, patients with severe periodontitis had higher blood UA levels than those with mild or moderate periodontitis [
21]. In this context, the meta-analysis tended to homogenize the data, but did not cover up the effect of periodontitis. Another concern involves the definitions of controls. In many studies, the control groups included not only healthy periodontium, but also gingivitis, mild periodontitis and even localized moderate-to-severe periodontitis. This is a common practice in case definitions when conducting clinical research, which could lead to an underestimation of the differences in UA levels between periodontitis and controls. However, the direction of the disparities would not be changed. Future research should utilize a uniform criterion on periodontitis and controls, i.e., the 2018 classifications of periodontal diseases, in order to enhance comparability between included studies.
Methods of sample collection and analysis may be potential confounding variables. Some studies found no difference between plasma and serum UA levels as measured by enzymatic colorimetric methods [
91,
92]. As determined by gas chromatography–mass spectrometry in a separated study, the UA concentration in plasma was 1.59 times that in serum [
93]. The discrepancy could be attributed to the differences in technological sensitivity [
94]. The present study did not calibrate the UA levels in the two types of blood samples because all of the included studies detected the UA content using enzymatic colorimetric methods. Notably, subgroup analysis on studies involving blood UA showed that the statistical heterogeneity was significantly lower in the plasma subgroup than that in serum. A better homogeneity in the plasma subgroup may be partly due to those anticoagulants (e.g., ethylenediamine tetraacetic acid) inhibit xanthine oxidoreductase activity and thus reduce UA production from undesired sources [
95], whereas sustained purine degradation may still occur in serum. Another anticoagulant heparin sodium, however, does not appear to affect xanthine oxidoreductase activity [
96]. Therefore, plasma with specific anticoagulants may be preferrable to serum for comparing the results of blood UA content across studies. The combined data from studies using plasma would be more reliable from a heterogeneity standpoint than those using serum. In addition, the effect size from studies using plasma was clearly greater than that from all relevant studies using blood regardless of plasma or serum. Another concern would be whether collection methods influence the results of UA levels in saliva. Resting saliva appears to contain more UA than stimulated saliva [
77]. The ratios of resting to stimulated salivary UA was found to be approximately 2:1 in both periodontitis patients and healthy controls. In addition, the concluding meta-analysis excluded studies involving stimulated saliva. Thus, the method of saliva collection may not be a significant confounder in the present meta-analysis.
The present systematic review and meta-analysis had some limitations. First, the number and sample size of included studies were limited, particularly those involving saliva and GCF. Even if the 4 studies (all involving saliva) with small sample sizes(i.e., n < 20) were excluded from the meta-analysis [
29,
53,
54,
57], the main findings regarding forest plots did not change significantly. The findings should be interpreted with caution until they are confirmed by large-scale studies. Second, the majority of the findings were derived from retrospective studies, which must be confirmed by prospective and interventional studies. Lastly, the raw data for some potential confounding variables (i.e., age, smoking, and body mass index) were unavailable and their effects (especially body mass index) on the results were unknown. It may contribute to statistical heterogeneity in the present meta-analysis. Taken together, high-quality studies, particularly prospective cohort studies and interventional (e.g., periodontal or urate-lowering treatments) studies, are required to elucidate the association between periodontitis and UA in blood and oral fluids.
Acknowledgements
The authors would like to thank Dr Salima Senouci (Ahmed Ben Bella Faculty of Medicine, Department of Pharmacy, University of Oran1, Algeria), Dr Gen-Min Lin (Department of Internal Medicine, Tri-Service General Hospital, National Defense Medical Center, Taiwan), Dr Meenakshi Sreeram (Department of Physiology and Biochemistry, MGV's KBH Dental College and Hospital, India), Dr Shaheena Banu (Sri Jayadeva Institute of Cardiovascular Science and Research, Department of Biochemistry, India) and Dr Antonio Canabarro (Department of Periodontology, School of Dentistry, Veiga de Almeida University, Brazil) for providing the raw data of their clinical studies during the stage of quality assessment.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.