Expression of gingival crevicular fluid markers during early and late healing of intrabony defects after surgical treatment: a systematic review

Surgical treatments such as guided tissue regeneration (GTR) and access flap surgery are widely employed for the treatment of intrabony defects. However, little is known regarding the postoperative expression of gingival crevicular fluid (GCF) markers. The aim of this systematic review was to compare the expression of GCF markers following treatment of periodontal intrabony defects with guided tissue regeneration or access surgery. The association of the markers’ expression with the clinical outcome was also assessed. An electronic literature search was conducted in MEDLINE, EMBASE, OpenGrey, LILACS and Cochrane Library up to December 2018 complemented by a manual search. Human, prospective clinical studies were identified. The changes from baseline up to 30 days (early healing) and 3 months (late healing) were assessed. A total of 164 publications were identified and reviewed for eligibility. Of these, 10 publications fulfilled the inclusion criteria. The included studies evaluated 15 different GCF markers with a follow-up time between 21 and 360 days postoperatively. PDGF, VEGF and TIMP-1 changes were often investigated in the included studies; however, contrasting results were reported. Two studies agreed that both GTR and OFD lead to similar OPG level changes. TGF-β1 is increased early postoperatively, irrespective of the surgical technique employed. There is limited evidence available on the expression of GCF markers after surgical interventions of intrabony periodontal defects. However, OPG and TGF-β1 tend to increase early post-operatively, irrespective of the surgical technique employed, irrespective of the surgical technique employed. More well-designed, powered studies with sampling periods reflecting the regenerative process are needed, and future research should focus on employing standardised protocols for collecting, storing and analysing GCF markers.


Introduction
Periodontitis is a chronic inflammatory disease caused by bacterial biofilm that leads to a progressive destruction of the supporting apparatus of a tooth and eventually to tooth loss. The prevalence of periodontitis, according to the 2009-2010 data from the National Health and Nutrition Examination Survey (NHANES), reaches 46% in US adults [1].
As periodontal disease progresses, it results in bone loss that can be horizontal or vertical or a combination of both. The loss of supporting bone vertically results in the formation of intrabony defects that progressively worsen and are associated with an increased probability of tooth loss [2]. While non-surgical periodontal therapy is effective in improving the clinical parameters, such as probing pocket depth (PPD) and clinical attachment levels (CAL) [3], surgical approaches are more effective-in particular for PPD of more than 6 mm [4,5]. Currently, intrabony defects are identified as sites favourable for periodontal regeneration [6,7] with the most commonly used techniques being guided tissue regeneration (GTR) and enamel matrix derivatives (EMD) presenting with similar clinical outcomes which are superior to open flap debridement (OFD) and osseous surgery (OS) [8][9][10][11].
However, irrespective of the regenerative modality employed for the treatment of intrabony defects, little is known regarding the processes and sequences involved in the periodontal regeneration and consequently, in the postoperative expression of angiogenesis, regeneration and inflammation markers in the gingival crevicular fluid (GCF) that accompany these processes [12]. The expression of such markers postoperatively may define whether the healing process moves towards a regenerative or a reparative direction [12]. Understanding the cellular and biological events in periodontal wound healing can possibly provide useful information in identifying predictable regenerative treatment for the periodontium.
The aim of this systematic review was to investigate the healing patterns of intrabony defects after surgical interventions (GTR, OS, OFD, EMD) by means of angiogenesis, regeneration and inflammation markers detected in the GCF before and early (≤ 30 days) or late (3 months) after the surgical intervention. Furthermore, the association of the expression of the GCF markers with the clinical outcome was investigated.

Protocol and Registration
The present systematic review followed the PRISMA (Preferred Reporting Items for Systematic Review and Meta-Analyses) guidelines [13] (Supplemental Material 1) and was registered with PROSPERO under the ID number CRD42018115794.

PICO question
The PICO question (patient, intervention, comparison and outcome) formulated was: "In patients with periodontal intrabony defects, does the expression of GCF markers for angiogenesis, regeneration and inflammation differ when treated with GTR employing a membrane and/or bone graft and/or biologics (e.g. EMD) (test group(s)) compared with intrabony defects treated with access surgery [OFD or OS or apically positioned flap (APF)] (control group) early (≤ 30 days) and late (3 months) after the surgical intervention?"

Types of studies
Human, prospective clinical studies assessing the expression of angiogenesis, regeneration and inflammation markers in the GCF were considered. Only studies with at least ten patients per group were included. No language restriction was set.

Population
Systemically healthy individuals with chronic periodontitis (CP) with at least one tooth with PPD ≥ 5 mm, CAL and evidence of radiographic bone loss or aggressive periodontitis [14,15] or periodontitis stages III or IV [16] and contributing a minimum of 1 intrabony defect.

Intervention and comparison
Intrabony defects treated with GTR employing a membrane and/or bone graft and/or with biologics (e.g. Emdogain) (test group(s)) and intrabony defects treated with access flap surgery (OFD or OS or APF) (control group). No restriction related to the flap technique (minimally invasive or not) was applied to avoid omitting potentially relevant data. Intrabony defects treated with adjunct growth factors e.g. EMD were included in the test group(s).

Outcome measures
The primary outcome of this review was the change in the expression of angiogenesis, regeneration and inflammation markers in the GCF during early healing (from baseline up to 30 days) and during late healing (from baseline to at least 3 months postoperatively). Secondary outcomes considered were the association of the expression of GCF markers (early and/or late healing) with the clinical outcome, assessed with the use of surrogate measures such as PPD and/or CAL.

Information sources and electronic search
An electronic search was conducted by two independent reviewers (VK and GC) in MEDLINE, EMBASE, Cochrane Library, LILACS and OpenGrey for publications up to 10 December 2018. Combinations of controlled terms (MeSH and EMTREE) and keywords were utilised: ("infrabony" or "intrabony" or "infra-bony" or "intrabony" or "angular defect" or "periodontal defect") and ("guided tissue regeneration" or "GTR" or "periodontal regeneration" or "periodontal surgery" or "open flap debridement" or "OFD" or "access surgery") and ("gingival crevicular fluid" or "crevicular fluid" or "GCF" or "inflammatory marker" or "marker" or "growth factor" or "inflammatory mediator" or "biomarker") Additionally, a manual search of periodontology-related journals including Journal of Dental Research, Journal of Clinical Periodontology, Journal of Periodontal Research and the Journal of Periodontology was performed from 2015 to 2018. The list of references in the publications included in this review as well as the list of references in relevant reviews were screened for potential additional publications fulfilling the inclusion criteria.

Study selection
The search results were initially screened for relevancy by means of title, keywords and abstract, independently and in duplicate by two reviewers (VK, GC). Irrelevant records were excluded at this stage. Any conflict was resolved with discussion. At the second round of screening, the full text of the publications remaining after the first round was assessed for inclusion in this review against the eligibility criteria described previously. The level of agreement between the two reviewers was calculated using Kappa statistics.

Data collection process/data items
The characteristics of the included publications were extracted by two reviewers (VK, GC). Among the details extracted were study characteristics (authors, journal of publication, year, country), number of patients, their demographics and risk factors (age, gender, smoking), diagnosis, number of intrabony defects, history of non-surgical treatment of the sites and time elapsed, characteristics of the included defects, surgical procedure employed (GTR, OFD), biomaterials used in the test group(s), postoperative care protocol, exposure rate, followup period, expression levels of the GCF markers, clinical outcomes (PPD, CAL), details of the methodology employed for the GCF sampling, storage, processing and detection of the markers, information regarding the main study outcome and power calculation of the study. When data from the included studies were missing, the authors of the publication were contacted through email.

Risk of bias assessment
The risk of bias of the included publications was assessed by the two reviewers independently and in duplicate. For the RCTs included, the quality of the selected publications was assessed according to the Cochrane Collaboration's tool for assessing risk of bias [17]. The selected publications were assessed for seven domains: sequence generation, allocation concealment, blinding of the participants and personnel, blinding of the outcome assessment, incomplete outcome data, free of selective outcome reporting and other sources of bias. For each of the individual domains, studies were classified as low, unclear or high risk of bias. Observational studies were assessed using the MINORS tool [18]. Studies were assessed in 12 items including clarity of the aim, inclusion of consecutive patients, prospective data collection, appropriateness of end points, unbiased assessment of study end points, appropriateness of follow-up time, inclusion of loss to follow-up rate, prospective calculation of the study size, comparable control group, contemporary control groups, baseline equivalence of groups on several factors and adequate statistical analysis. Each study may receive 0-2 points for each item and the total score ranges from 0 to 24 points. Studies with fewer than 16 points are considered of low quality, while high-quality studies need to have a score of greater than or equal to 16.

Study selection
The flowchart of the study selection and inclusion process is shown in Fig. 1. The initial search identified 68 MEDLINE, 110 EMBASE, 59 Cochrane database and 1 LILACS titles, with a total of 163 after duplicates' removal. One additional title was identified through hand search for a total of 164 titles. Following the screening of titles and abstracts by the two reviewers, 10 articles qualified for full text screening and all 10 met the inclusion criteria. The kappa value for interreviewer agreement was 0.99 at first round and 1.00 at second round.
The follow-up of the expression of GCF markers ranged from 21days [12] to 360 days [24], while the follow-up of the clinical parameters after treatment ranged from 90 [23] to 360 days [24]. However, there was rarely coincidence of the sampling times for the GCF markers with the clinical assessments postoperatively.
Finally, only 2 studies [21,24] reported the postoperative occurrence of exposures. Gamal et al excluded the exposed sites from the study [21], while Rakmanee et al reported that 13 out of the 18 sites presented exposure of the membrane that was treated either with removal of the membrane (2 sites, classified as major exposure with size > 4 mm) or with administration of antibiotics (2 sites, classified as minor) [24].

Synthesis of results
The results and conclusions of the individual studies included are presented in Table 2. Due to the significant heterogeneity of the included studies, in relation to the methodology employed, a meta-analysis was not performed.

GTR
Regarding GTR, 7 studies reported on the expression of GCF markers postoperatively [11,12,21,22,[24][25][26]. Both Gamal, 2011 [11] and Gamal 2016 [21] employed the same GCF sampling method using a micropipette inserted at 2mm depth in the sulcus and filled with 5μL of GCF. The samples were subsequently stored at -76C and analysed with ELISA. The concentrations of platelet-derived growth factor-BB (PDGF-BB) peaked during the early stages of healing (< 14 days) and decreased to baseline values by 30 days. Similarly, Rakmanee et al employing a different methodology, using pre-cut chromatography strips at the entrance of the gingival crevice for 2 min and stored at − 70C, found again increased PDGF-AB amounts 7 days postoperatively that decreased to baseline levels after 42 days [24].  Furthermore, Rakmanee et al found similar PDGF-AB levels both after GTR and after OFD that were accompanied by a similar clinical response. However, the sites subjected to GTR were associated with high rates of exposure (13/18) that may have significantly affected the regenerative process and thus the clinical response observed. Furthermore, Rakmanee et al reported that GCF osteoprotegerin (OPG) amounts significantly increased 2-3 days postoperatively and subsequently declined [24]. No significant differences were noted between sites treated with GTR and sites treated with OFD. Pellegrini et al. using Periopaper inserted in the gingival crevice for 30 s found OPG levels to decrease following GTR and OFD; however, no comparison by treatment was reported for the change of the marker [12].
The expression levels of vascular endothelial growth factor (VEGF) were investigated by Rakmanee et al. [24] and Gamal et al. [21]. The former did not detect any significant difference in the change of VEGF GCF levels between sites treated with GTR and sites treated with access surgery using pre-cut chromatography strips [24]. However, the study by Gamal and co-workers, which used micropipettes, found that VEGF concentrations measured statistically significant higher concentrations in defects treated with OFD and GTR using a perforated membrane during the early postoperative period (days 1, 3 and 7) compared to defects treated using the occlusive membrane [21]. Kuru et al. 2004, using pre-cut chromatography strips at the entrance of the gingival crevice for 2 min, found increased transforming growth factor βi (TGF-β1 levels 2 weeks postoperatively, that however were not statistically significant and declined to below baseline levels by 4 weeks [26]. The change in the TGF-β1 levels was similar both after GTR and after OFD and accompanied a similar clinical response 6 months postoperatively.

EMD
Regarding EMD, 3 studies reported on the expression of GCF markers [19,20,23]. Ribeiro et al., using Periopaper in the gingival crevice until resistance was felt and for 30 s, reported that TGF-β1 levels in sites treated with MIST and EMD significantly increased by 15 days postoperatively and the levels decreased after 3 months [20]. Furthermore, the changes for TGF-β1 levels were similar for sites treated with MIST and MIST with EMD and accompanied a similar clinical and radiographic response for both treatments. In contrast, Agrali et al. using Periopaper, inserted in the gingival crevice for unspecified amount of time, reported significantly higher TGF-β1 levels for EMD-treated defects compared with OFD-treated defects 7 and 14 days postoperatively [19]. In the same line, the authors concluded that defects treated with EMD presented a superior clinical and radiographic improvement compared with defects treated with OFD. It is however worth noting that the majority of the defects treated with EMD   were localised at anterior teeth, compared with the majority of the sites treated with OFD that were localised at molar teeth. Okuda et al. described an increase for matrix metalloproteinase-8 (MMP-8) and metallopeptidase inhibitor 1 (TIMP-1) GCF levels 2 weeks postoperatively for defects treated with EMD and OFD that thereafter declined, more dramatically for EMD-treated defects [23]. MMP-1 levels significantly decreased from 2 to 4 weeks postoperatively for defects treated with EMD [23]. Ribeiro et al. using Periopaper in the gingival crevice for 30 s concluded that OPG levels increase 15 days postoperatively and similar changes are noted after both MIST and MIST with EMD [20]. Interestingly, Rakmanee et al. concluded that the OPG amount significantly increased 2-3 days postoperatively and subsequently declined with no significant differences between sites treated with GTR and OFD [24]. Consequently, similar changes were noted for the OPG levels after MIST and MIST with EMD or after GTR and access flap. Furthermore, both studies reported similar clinical and radiographic improvements for both treatment groups [20,24]; thus, the similar expression patterns in the levels of expression of OPG accompanied a similar clinical response, irrespective of the surgical technique employed.

GTR and EMD
Regarding the combination of GTR and EMD, Agrali et al. reported on the levels of TGF-β1. The combination-treated defects, similarly distributed to anterior and posterior teeth, presented similar changes in the TGF-β1 levels as the EMDtreated defects in the first 2 postoperative weeks [19]. The similar changes of TGF-β1 levels accompanied a similar clinical response for EMD and EMD with GTR, that was superior to OFD. However, as discussed previously the majority of the defects treated with OFD were localised in posterior teeth.

OFD
Finally, regarding OFD alone, there is agreement between two investigations that an increase in TGF-β1 levels is initially observed, accompanied by a return to baseline levels by 14 [19,26]. For PDGF, conflicting results are presented; Gamal and co-workers reported an initial increase and a decrease by 7 days postoperatively to below baseline levels [11]. In contrary, another investigation from the same group reported a decrease for the PDGF levels that continued until 30 days postoperatively [21]. Rakmanee et al. noted an increase in the PDGF amount after OFD that continued until 3 months postoperatively [24]. For the remaining markers and for more detail regarding the expression of the investigated GCF markers after the surgical treatments, the reader is referred to the detailed Table 2.

Risk of bias assessment
The risk of bias assessment is presented in Fig. 2 and Table 3. Seven of the included studies (RCTs) were assessed using the Cochrane Collaboration tool [11,[19][20][21][22][23][24]. Four of the seven studies were of low risk of bias in all but one domain [11,20,22,24], two were of low risk of bias in five domains [21,24] and one [19] was of high risk of bias. The remaining three studies [12,25,26] were prospective cohort studies and were assessed using the MINORS tool. These studies were rated with 16 to 18, indicating high quality of the included studies.

Discussion
This systematic review identified 15 GCF markers expressed after surgical treatment of intrabony defects (GTR, OS, OFD). For 7 of those, most of which are related with the healing of connective tissue, TGF-β1, PDGF, VEGF, FGF, MMP-1, TIMP-1 and OPG, data was available from more than one investigation. While for the majority of factors a definitive conclusion cannot be reached, robust suggestions can be drawn regarding the OPG levels in regenerative surgeries. In two investigations, employing different GCF sampling and storing techniques, the OPG levels after MIST or MIST with EMD and after GTR or access flap similarly increased within the two postoperative weeks and thereafter declined [20,24]. Furthermore, both studies reported similar clinical and radiographic improvements; thus, the similar expression patterns of OPG likely accompanied a similar clinical response. OPG acts as a soluble decoy receptor, binding to the receptor activator of nuclear factor-kappa B (RANKL) and inhibiting the osteoclastogenic action [27]. Therefore, OPG has been identified as a critical factor in bone formation and the regulation of bone resorption.
The finding of this review however comes in contrast with a human polymerase chain reaction (PCR) study assessing the Fig. 2 Risk of bias assessment of RCTs using the Cochrane Collaboration tool gene modulation 21 days following treatment of intrabony defects with either GTR with an expanded polytetrafluorethylene (ePTFE) membrane or flap surgery [28]. Among others, OPG mRNA levels were significantly higher in GTR sites, compared with access flap sites. Furthermore, in an investigation of the gene expression profile of cells derived from GTR subjected defects (regenerating-tissue derived cells-RTCs), a differential and highlighted expression of the gene encoding OPG (TNFRSF11B) was found compared with matched periodontal ligament mesenchymal cells (PLCs) [29]. These contrasting results may be due to the high rate of exposures (13/18 GTR sites) in the study by Rakmanee et al. that may have significantly affected the regenerative process and thus the OPG expression [24]. In addition, the Ribeiro et al. investigation, as most of the included studies, was not powered for GCF and could therefore lack statistical power to detect differences in the expression levels between treatments [20].
Interestingly, Okuda et al. found that the use of EMD in intrabony defects resulted in an early postoperative increase for MMP-8 and TIMP-1, followed by an accelerated return to baseline levels, when compared to OFD [23]. This highlighted reduction may associate with an EMD-induced accelerated pattern of wound healing and resolution of inflammation moving towards regeneration rather than repair.
With respect to PDGF, three isoforms exist (AA, AB, BB). Two studies included in this review [11,24] demonstrated that GTR and access flap lead to similar changes: an initial increase of PDGF-BB [11] and PDGF-AB [24] during the early healing period (up to 7 days), accompanied by a decrease to baseline levels. Two PDGF receptors exist, the PDGF-Rα and the PDGF-Rβ who binds PDGF-AB with low and PDGF-BB with high affinity [30]. In contrast with the included in this review studies, a significant upregulation of PDGF-Rβ in regenerating periodontal tissues has been observed [31] suggesting that the ligands are involved in the early cascade of events involved in regeneration.
Furthermore, in the only study powered for GCF markers [21], perforated PTFE membranes were shown to result in significantly higher VEGF during the early healing period (1, 3 and 7 days) when compared to occlusive membranes. In an animal model of GTR using porcine extracellular matrix (ECM), membrane cells recruited early postoperatively into the membrane compartment result in highlighted expression of, among other factors, VEGF at the RNA level. The VEGF expression was significantly highlighted 3 days postoperatively and thereby decreased by 28 days [32]. This VEGF upregulation, along with other regenerative molecules, early postoperatively in the mRNA and the protein level, may suggest that the membrane itself acts as a bioactive compartment guiding the regenerative process and not solely as an active barrier.
TGF-β1 is a connective tissue cell signalling protein that plays a critical role in several stages of wound healing, as it promotes the mitogenic activity of gingival and periodontal ligament cells and the upregulation of extracellular matrix components [33,34]. With regards to TGF-β1, the existing literature is conflicting. The studies included in this review studies suggest that the clinical and radiographic outcome may be related to the TGF-β1 level changes [20,26]. Ribeiro et al and Kuru et al showed that GTR or OFD and MIST or EMD treatments exhibited similar TGF-β1 increase early post-operatively, as well as similar clinical and radiographic improvements [20,26]. However, the significant clinical and radiographic improvement following EMD in Agrali's study was associated with a significant TGF-β1 increase for EMD at anterior teeth,  18 16 in contrast with OFD at mainly posterior teeth [19]. When OFD was employed alone for the treatment of intrabony defects, two investigations agreed that an increase in TGF-β1 levels is initially observed, accompanied by a return to baseline levels by 14 days [19,26]. In the same line, in an immunocytochemistry investigation in biopsies, a highlighted increase was noted for TGF-β1 receptor in regenerating periodontal tissues (6 weeks), while the receptor was almost undetectable in healthy tissues [31]. The highlighted receptor presence in regenerating tissues may suggest that the corresponding TGF-β1 plays a pivotal role in the early healing. As it became evident, an important limitation was that only one of the ten included studies [21] was powered to detect significant differences in the GCF markers, whereas the remaining nine studies were either powered for the clinical outcomes or did not report any power calculation. Furthermore, inclusion of intrabony defects with varying number of defect walls does not allow for meaningful conclusions as regeneration is more likely to occur in 3-walled defects and to be accompanied by a different array of GCF markers compared to a 1-wall defect. In addition, reportedly, a large variation across investigations was observed in the methodology employed for the GCF sampling, storage and detection. These variations would introduce confounders if a meta-analysis was attempted. For example, the sampling methods (Periopaper, micropipette), the duration of collection (30 s, 2 min, until a specific volume is collected), the depth of strip insertion (entrance of the pocket or full depth), the storage (temperature) or the preparation of the samples (processing individual or pooled samples) introduce variations that would affect the conclusions drawn and their generalisability. Furthermore, it becomes imperative that more well-designed, powered studies with sampling periods reflecting the regenerative process are needed. Future investigations should employ standardised protocols for GCF sampling, processing and storage.
In conclusion: There is limited evidence available on the expression of markers of angiogenesis, regeneration and inflammation in the GCF in the early and late healing after surgical interventions of intrabony periodontal defects OPG is increased early postoperatively, irrespective of the surgical technique employed A trend is noted for TGF-β1 increase early postoperatively, irrespective of the surgical technique employed. A highlighted increase is noted after use of EMD at anterior teeth that may relate with an improved clinical outcome.
More well-designed, powered studies with sampling periods reflecting the regenerative process are needed Future research should focus on employing standardised protocols for collecting, storing and analysing GCF markers and establishing adequate statistical power to reach conclusions that may shed light in the biological events involved in the early periodontal wound healing and thus facilitate the development of predictable regenerative treatments

Compliance with ethical standards
Conflict of interest The authors declare that they have no conflict of interest.
Ethical approval This article does not contain any studies with human participants or animals performed by any of the authors.
Informed consent For this type of study, formal consent is not required.
Clin Oral Invest (2020) 24:487-502 501 Open Access This article is distributed under the terms of the Creative Comm ons Attribution 4.0 International License (http:// creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.