1 INTRODUCTION

Degradation of periodontal tissue is related to the activity of proteases involved in the inflammatory process.1, 2 One of the most important families of proteases associated with periodontal disease are the matrix metalloproteinases (MMPs).3 In particular, MMP-8, MMP-9, and, to a lesser extent, MMP-14 have been studied in relation to periodontitis.4 These MMPs are not only responsible for the degradation of the extracellular matrix during periodontitis but are also key factors in tissue remodeling processes.5, 6 The activity of MMPs is regulated by tissue inhibitors of metalloproteinases (TIMPs) which are produced and secreted by many cell types. Their production is regulated by various cytokines and growth factors. Besides MMPs, TIMPs also regulate the activity of other families such as the disintegrin metalloproteinases (ADAM and ADAMTS).7 Therefore, TIMPs play a crucial role in important biological processes like the formation of the extracellular matrix and cell proliferation.

Upon binding to MMPs, TIMPs act like a wedge which connects to the active site of the MMP and thereby blocking the binding of substrate to MMP, resulting in reduced MMP activity.7 An imbalance between MMP activity and regulation by TIMP has been associated with progression of periodontal disease. This imbalance results in the degradation of matrix proteins, and thereby contributes to the destruction of periodontal tissue.8-11

The strong relation between TIMPs and MMPs suggests that TIMPs might potentially serve as a biomarker to diagnose periodontitis and monitor disease progression in oral fluids.12, 13 Of the four types of TIMPs identified in humans, TIMP-1, an inhibitor of MMP-9, has most often been associated with periodontal disease. However, so far the diagnostic value of TIMP-1 in periodontal disease has not been systematically reviewed.7

In this context, the aim of this systematic review was to analyze the validity of TIMP-1 solely as a biomarker to diagnose periodontal disease in saliva and gingival crevicular fluid (GCF).

2 MATERIALS AND METHODS

This systematic review was elaborated according to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines.14 The PRISMA checklist is included in Table S1.15 The protocol was registered at the International Prospective Register of Systematic Reviews (PROSPERO) under registration number CRD42021246024.16

2.1 Research strategy, selection, and inclusion and exclusion criteria

An electronic database search was performed until December 31th 2020 in the database of the National Library of Medicine (MEDLINE by PubMed) and Web of Science using a combination of medical subject headings (MeSH) terms and free text words (Appendix S1).

The resulting articles were reviewed independently by title, abstract, and full text by two reviewers (PdB and WEK). Any disagreements during the review process were resolved by discussion. Articles that met the following inclusion criteria were retrieved: studies including patients with chronic periodontitis or gingivitis diagnosed based on clinical parameters, publications in English, and studies investigating TIMP-1 concentrations in oral fluids. Publications that did not present a compatible methodology for a systematic analysis were excluded (e.g., reviews, opinions, book chapters, abstracts, and editorial letters). In vitro studies, animal studies, experiments that interfered with the expression of TIMP-1 through therapeutic methods, studies that evaluated patients with systemic diseases, studies that investigated other types of periodontitis than chronic periodontitis or gingivitis, studies investigating the systemic effect of proteases, studies that evaluated pregnant patients, and studies that evaluated children were also excluded. In addition, studies without a control group were also excluded. Where possible, sample sizes, mean-values, and standard deviations were retrieved from the publications or calculated based on the available data. In case limited data were available, study investigators were contacted to retrieve the missing information. The whole process of literature selection was executed according to the PRISMA guidelines and is summarized in Figure 1.

Schematic PRISMA diagram for procedural methodology

2.2 Data extraction

Information retrieved from all studies involved: authors, year of publication, number of patients diagnosed with periodontitis and number of controls, severity of the periodontal disease, criteria for diagnosis used for inclusion, TIMP-1 detection method, study results, and relevant conclusions.

2.3 Assessment of risk of bias

The selected studies were analyzed with tools from the National Heart, Lung, and Blood Institute (NHLBI) to assess their quality.17 First, the selected studies were classified by research design.18 Depending on the research design, the following three risk assessment questionnaire tools were used: Controlled Intervention Studies, Observational Cohorts and Cross-Sectional Studies, and Case–Control Studies. All articles were independently assessed by two reviewers (PdB and WEK) rating each domain as ‘yes’, ‘no’, ‘not applicable’, or ‘not reported’. The overall rating of each study could be ‘good’, ‘fair’, or ‘poor’. Any disagreement on the bias risk assessment between the two reviewers was resolved by discussion.

2.4 Statistical analysis

Statistical analyses were performed using the Cochrane Collaboration's software for preparing and maintaining Review Manager 5.4.1.. A quantitative synthesis (meta-analysis) for generating an estimate on the effect size was possible. This meta-analysis was conducted to the primary outcome: TIMP-1 concentration (ng/mL) (mean ± SD) compared between periodontitis/ gingivitis patients and healthy individuals. In case in a study varying degrees of periodontal disease were monitored, the most severe condition was included. When in a study both chronic and acute periodontitis patients were monitored, data from the chronic patients were included in the analysis. Because of lack of identity between the included studies, the random-effects model was used to perform the meta-analysis.19 I2-values higher than 50% were considered as indicative of substantial heterogeneity. P-values less than 0.05 were considered as statistically significant.

3 ETHICAL REVIEW

This study was approved by the ACTA Ethics Committee (registration number 2020113).

4 RESULTS 4.1 Summary of the literature search and description of the included studies

The literature screening and selection process is presented in Figure 1. The search strategy retrieved a total of 77 studies using the PubMed database and 245 studies upon searching the Web of Science database. After removal of duplicate records, the titles and abstracts of the remaining 252 records were screened on inclusion and exclusion criteria. In total, 219 records were removed from the study. Detailed reading of the full text of the remaining 33 articles led to the additional removal of 19 records. Six articles only determined MMP/ TIMP-1 ratio, three articles did not investigate TIMP-1, one did not investigate periodontal disease in combination with TIMP-1, in three articles the measurement of TIMP-1 concentrations was discontinued during the study, three articles investigated interventions, and three articles were rejected for other reasons. The authors of six of the remaining 14 articles were approached for additional data. Two of the corresponding authors provided extra data. The author of another study reported that the records of the studies no longer existed, and the authors of the remaining three studies did not respond to the request for additional data. Therefore, these four articles were also excluded, based on missing data. The 10 remaining articles were included in the study and used in the meta-analyses.

The main characteristics of the included studies are described in Table 1. All selected studies were published between 2006 and 2019 and accounted for 1336 participants with a mean of 128 participants per study and an age range between 15 and 64 years. Of 597 patients suffering from periodontal disease salivary TIMP-1 levels were measured. The included studies were executed in Turkey, Brazil, Finland, Sweden, Denmark, and the USA. Five studies used stimulated saliva as clinical fluid, three unstimulated saliva, and two studies GCF.

TABLE 1. List of the included studies Author, year Country

n

(Total)

n

(P/NP)

Gender

(P/NP)

Age (mean or range)

(P/NP)

Used reference standard criteria Biological sample

Detection method

(TIMP−1)

Study design

Emingil et al.,

2006 21

Turkey 60 40/20

♂ 25/9

♀ 15/11

41.5/27.4

Periodontitis: CAL ≥5 mm, PD ≥6 mm in multiple sites of all four quadrants of the mouth.

Gingivitis: BOP ≥10%

Control: no BOP, no ABL observed in radiographs, PD <3 mm.

GCF ELISA Case–Control Study Marcaccini et al., 201023 Brasil 42 27/15

♂ 10/5

♀ 17/10

44.1/44.9

Periodontitis: ≥2 teeth with PD ≥5 mm, CAL ≥6 mm, evidence of ABL observed in radiographs.

Control: no BOP, no ABL observed in radiographs, PD <3 mm.

GCF ELISA Controlled Intervention Study

Gürsoy et al.,

2010 5

Finland 106 40/66

♂ 27/22

♀ 13/44

50.7/48.6

Periodontitis: ≥14 teeth with PPD ≥4 mm.

Control: No teeth with PPD ≥4 mm.

SS ELISA Cross-Sectional Study

Buduneli et al.,

2011 36

Turkey 32 15/17 NR NR/35.4

Periodontitis: ≥1 pocket in each quadrant with PD ≥5 mm and CAL ≥6 mm.

Control: systemic and periodontal healthy.

SS ELISA Cross-Sectional Study Rathnayake et al., 201222 Sweden 352 49/303 NR 64.4/42.6

Periodontitis: Mild (ABL >one-third of the root length in <30% of sites) to severe (ABL >one-third of the root length in >30% of sites)

Control: No ABL observed in radiographs.

SS ELISA Cross-Sectional Study Meschiari et al., 201310 Brazil 42 23/19 NR NR

Periodontitis: ≥2 teeth with PPD ≥5 mm, CAL ≥6 mm and evidence of ABL observed in radiographs.

Control: periodontal healthy subjects.

SS ELISA Controlled Intervention Study

Nizam et al.,

2014 20

Turkey 36 18/18

♂ 10/11

♀ 8/7

42-61/26–63

Periodontitis: ≥4 teeth in each jaw PD ≥5 mm, CAL ≥4 mm, ABL ≥50% in at least in two quadrants, BOP >80%.

Control: no BOP, no ABL observed in radiographs.

US ELISA Case–Control Study

Morelli et al.,

2014 37

USA 67 34/33

♂ 14/10

♀ 20/23

34.4/30.3

Periodontitis: ≥1 site with PD >3 mm, BOP >50%.

Gingivitis: all PD <3 mm, BOP ≥10%.

Control: all PD <3 mm, BOP <10%.

US Bioplex Multiplex system Observational Cohort Study Lahdentausta et al., 20189 Finland 481 285/196

♂ 194/120

♀ 91/76

64.1/62.4

Periodontitis: Mild (ABL in cervical third of the root) to severe (ABL in the apical third of the root) and PPD ≥4 mm in ≥4 sites.

Control: periodontal healthy, gingivitis and edentulous patients.

SS ELISA Cross-Sectional Study Nascimento et al., 20198 Denmark 84 42/42 NR 18–35 Experimental gingivitis: ≥20 teeth in each jaw, PD 4 mm, CAL 2 mm. US ELISA Observational Cohort Study Abbreviations: ABL: Alveolar bone loss, BOP: bleeding on probing, CAL: clinical attachment level, GCF: gingival crevicular fluid, NP: non-periodontal disease, NR: not reported, P:Periodontal disease, PD: pocket depth, PPD: periodontal probing depth, SS: Stimulated saliva, US: unstimulated saliva.

The reported TIMP-1 outcome and main conclusions of the included studies are described in Table 2. The range of TIMP-1 concentrations varied considerably between the included studies, from 0.32 ± 0.15 ng/mL to 719 ± 24 ng/mL in periodontitis/ gingivitis patients, and from 0.37 ± 0.20 ng/mL to 721 ± 24 ng/mL in healthy individuals (mean ± SD) (Table 2). This variance was not directly related to type of oral fluid investigated, sample handling, or study population (Table 1).

TABLE 2. Reported outcome for TIMP-1 and periodontal disease Author, year Biological sample TIMP−1 (ng/mL) Periodontal disease (mean ± SD (n))

Control

(mean ± SD (n))

Results and conclusions on TIMP−1 as biomarker for periodontal disease Emingil et al., 200621 GCF

P: 0.56 ± 0.33 (20)

G: 0.32 ± 0.15 (20)

0.37 ± 0.20 (20) Total amounts of TIMP−1 in GCF were significantly higher in the periodontitis and gingivitis group compared to the healthy group (p < 0.0001). The concentration of TIMP−1 in GCF was comparable to that of the healthy group (p = .074). Marcaccini et al., 201023 GCF 103 ± 63 (27)* 74 ± 47 (15)* No difference in TIMP−1 levels between the groups at baseline, or after therapy. MMP−8/ TIMP−1 ratio was significantly higher in the periodontitis group compared to the healthy controls at baseline (p = .03). Periodontal treatment of the periodontitis patients resulted in a significantly lower MMP−8/TIMP−1 ratio (p = .001). Gürsoy et al., 20105 SS 61 ± 68 (40) 110 ± 72 (66) TIMP−1 concentration in stimulated saliva is significantly lower (p = .001) in the periodontitis group than in the control group. Buduneli et al., 201136 SS 11 ± 5 (15) 9,6 ± 2,8 (17) TIMP−1 levels between healthy controls, non-Acute Myocardial Infarction (AMI) and AMI patients significantly different (p = .001). No statements on the comparison of healthy controls with non-AMI periodontitis patients. Rathnayake et al., 201222 SS 264 ± 175 (49) 268 ± 206 (303) The difference in TIMP−1 concentrations between healthy controls and periodontitis patients is not significant. MMP−8/TIMP−1 ratio is significantly higher in periodontitis patients than in the controls. Meschiari et al., 201310 SS 70 ± 111 (23)* 83 ± 127 (19)* TIMP−1 concentration in stimulated whole saliva is not significantly different between healthy patients and periodontitis patients. Nizam et al., 201420 US 82 ± 62 (18) 298 ± 208 (18) The salivary TIMP−1 concentration was significantly lower in the periodontitis group than in the control group (p<0.001). The ratio of MMP−8/ TIMP−1 was significantly higher in the periodontitis group than in the control group (p < .001). Morelli et al., 201437 US

P: 717 ± 24 (34)

G: 719 ± 24 (34)

721 ± 24 (33) A significant increase in salivary TIMP−1 concentrations from baseline to peak induction in all groups (p < .001). No significant change in MMPs/ TIMPs ratio. No significant difference between the healthy group and the periodontitis or gingivitis group at baseline (p = .15). Lahdentausta et al., 20189 SS 177 ± 116 (285) 212 ± 122 (196) No significant difference in salivary TIMP−1 concentrations between healthy controls and periodontitis patients without acute coronary syndrome (ACS). Nascimento et al., 20198 US G: 452 ± 300 (42) 543 ± 430 (42) TIMP−1 levels in unstimulated saliva are positively associated with gingival inflammation to the similar magnitude as MMP−8. TIMP−1 concentrations were lower on day 35 of the gingivitis study than at the start of the experimental gingivitis study but no significant difference was found. Abbreviations: G: Gingivitis patients, GCF: gingival crevicular fluid, P: Periodontitis patients, SS: Stimulated saliva, US: Unstimulated saliva. * Data provided by authors.

Among the 10 included studies, a wide variety of conclusions was presented. In seven studies, TIMP-1 concentrations were lower in patients with periodontal disease than in healthy individuals, of which two found a significant difference.5, 20 In three studies, the TIMP-1 values were higher in patients with periodontal disease compared to the healthy individuals, of which one study found a significant difference.21

Three studies found that the MMP-8/ TIMP-1 ratio was significant higher in periodontitis patients20, 22, 23 (Table 2). The increase in MMP-8/ TIMP-1 ratio in these studies was predominantly related to increased salivary MMP-8 levels in periodontitis patients and not necessarily to decreased TIMP-1 concentrations in saliva. Only Nizam and co-workers observed a significant decrease in TIMP-1 level, whereas all three articles found a significant increase in MMP-8 concentration (Table 2).

4.2 Quality assessment

The methodological quality of the 10 included studies was analyzed through the use of tools from the National Heart, Lung, and Blood Institute (NHLBI). Some of the items of the quality assessment tool were defined ‘not reported’. Because not all these items had a relation to the focus of this study, TIMP-1 as biomarker for periodontal disease, the outcome of these items weighted less in the assessment of study quality. Among the six observational cohort studies and studies with a cross-sectional design, five were rated good and one was judged fair (Table 3). Both control intervention studies were rated fair (Table 4), due to the high number of ‘not reported’ items. None of the two control intervention studies applied randomization of the study population. Of the two case control studies, one rated good whereas the other study was judged fair (Table 5). The difference in quality is mainly due to lack of correction for potential confounders and differences in recruitment populations between the periodontal disease patients and control group participants.

TABLE 3. Quality assessment tool for Observational Cohort and Cross-Sectional studies Lahdentausta et al., 20189

Gürsoy et al.,

20105

Buduneli et al., 201136

Nascimento et al.,

20198

Morelli et al.,

201437

Rathnayake et al., 201222 1. Was the research question or objective in this paper clearly stated? YES YES YES YES YES YES 2. Was the study population clearly specified and defined? YES YES YES YES YES YES 3. Was the participation rate of eligible persons at least 50%? YES NO NR NR NR NO 4. Were all the subjects selected or recruited from the same or similar populations (including the same time period)? Were inclusion and exclusion criteria for being in the study pre-specified and applied uniformly to all participants? YES YES YES YES YES YES 5. Was a sample size justification, power description, or variance and effect estimates provided? NR NR NR NR YES NR 6. For the analyses in this paper, were the exposure(s) of interest measured prior to the outcome(s) being measured? YES YES YES YES YES YES 7. Was the timeframe sufficient so that one could reasonably expect to see an association between exposure and outcome if it existed? NA NA NA YES YES NA 8. For exposures that can vary in amount or level, did the study examine different levels of the exposure as related to the outcome (e.g., categories of exposure, or exposure measured as continuous variable)? NO NO YES YES YES YES 9. Were the exposure measures (independent variables) clearly defined, valid, reliable, and implemented consistently across all study participants? YES YES YES YES YES YES 10. Was the exposure(s) assessed more than once over time? NO NO NO YES YES NO 11. Were the outcome measures (dependent variables) clearly defined, valid, reliable, and implemented consistently across all study participants? YES YES YES YES YES YES 12. Were the outcome assessors blinded to the exposure status of participants? NO NO NO NO NO NO 13. Was loss to follow-up after baseline 20% or less? NA NA NA YES YES NA 14. Were key potential confounding variables measured and adjusted statistically for their impact on the relationship between exposure(s) and outcome(s)? YES YES YES NO YES YES Results GOOD FAIR GOOD GOOD GOOD GOOD Abbreviations: NA: Not applicable, NR: Not reported. TABLE 4. Quality assessment tool for Controlled Intervention Studies Meschiari et al., 201310