Socket seal surgery techniques in the esthetic zone: a systematic review with meta-analysis and trial sequential analysis of randomized clinical trials

This systematic review was registered in the PROSPERO database under number CRD42018094314 and was written following the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) statement [22]. AMSTAR 2 checklist was followed to ensure the quality and transparency of the search [23].

What is the effect of different socket seal surgery techniques on alveolar ridge preservation in the esthetic zone?

Only RCTs were included in the analysis. PICO elements were used for ordinate reporting of information [24].

The search strategy involves both electronic and manual searches. Electronic searches were performed in three databases: The National Library of Medicine (MEDLINE via PubMed), the Cochrane Central Register of Controlled Trials (CENTRAL), and Web of Science. The search strategy included terms related to the intervention and used the following combinations of keywords: ((((((((tooth extraction) OR socket) OR tooth removal) OR alveoli)) OR immediate placement) OR postextraction)) AND ((((((((((free gingival punch graft) OR free gingival graft) OR porcine collagen matrix) OR autogenous soft tissue graft) OR socket seal) OR collagen matrix seal) OR collagen sponge) OR subepithelial connective tissue)) AND ((((tooth extraction) OR socket) OR tooth removal) OR alveoli)). The results were limited to human studies. Also, an electronic screening of grey literature through Literature Report [25] and OpenGrey databases [26], as well as the consulting of references list of included studies, was conducted to detect potential eligible titles. The following journals were also screened up to June 2020: Clinical Oral Implants Research, Clinical Implant Dentistry and Related Research, European Journal of Oral Implantology, Implant Dentistry, International Journal of Oral and Maxillofacial Implants, International Journal of Periodontics and Restorative Dentistry, Journal of Clinical Periodontology, Journal of Dental Research, Journal of Oral and Maxillofacial Surgery, Journal of Periodontology. A manual search was also made in the bibliographies of the articles included. Only articles in English were included.

It was conducted independently and in duplicate by two reviewers (A.L.P and M.G). According to selection criteria, titles and abstracts of search results were screened, and potential articles, or those with insufficient data to make a clear decision, were analyzed in full text for the eligibility criteria. Disagreements are resolved by discussion and consultation with a third author (M.A.A). The reasons for exclusion at this or subsequent stages were recorded. The level of agreement between reviewers against title eligibility was done using kappa scores (Cohen’s ĸ coefficient) and interpreted according to Landis and Koch scale [27].

The following data are extracted in predefined Excel spreadsheets by two authors (A.L.P, M.G) and considering: author, year, country, reporting of a priori sample size estimation, sample size (sites), socket seal surgery approach, type of bone graft material, socket integrity, method of dimension measurement, follow-up, radiographic and clinical outcome measurements (alveolar ridge width, alveolar ridge height of the buccal plate, and buccal gingival thickness), study setting, and funding sources. The data extraction was ascertained for adequacy by a third author (M.A.A); disagreements were solved by consensus.

Assessment of the risk of bias was carried out by two reviewers (A.L.P, MG) independently. The assessment was carried out in duplicate and based on guidance from a modification of the Cochrane tool [28] used to evaluate the methodological quality of the included studies.

The following six parameters—random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, and selective reporting—were evaluated as low risk, moderate risk, or high risk of bias.

Statistical data handling was performed by one author (D.S.P.). Random-effects meta-analysis is conducted for the primary outcome (ridge width changes) with at least 6 months follow-up after SSS. Forest plots were created to illustrate the effects of the meta-analysis results. The effect size between test and control groups are summarized as mean difference (MD) in millimeters, and its respective 95% confidence intervals (CI).

Because the results of ridge preservation procedures might be affected by either clinical (e.g., age, sex, surgery, clinical expertise) and methodological aspects inherent to clinical trials conduction, a Sidik-Jonkman (Hartung-Knapp-Sidik-Jonkman) random effects model is carried out because it provides adequate type I error rates [29]. This approach is more robust to changes in the heterogeneity variance estimates, especially in meta-analyses that contain few studies [30]. Between-study heterogeneity was visually inspected in the forest plots and by calculating the τ2 (absolute heterogeneity) and the I² statistics (relative heterogeneity), and the corresponding nullity statistical Q test was calculated. The I² index defines the proportion of total variability in the result explained by heterogeneity, but no chance. Heterogeneity was roughly categorized as low, moderate, and high to I² values of 25%, 50%, and 75% [31]. Publication bias was investigated by visual inspection of the funnel plot and employing the Egger test only if at least 10 studies are included. In the case of high heterogeneity, a sensitivity analysis is conducted to test the robustness of estimates excluding risk populations or effect modifiers. A two-sided level of significance of 5% (α = 0.05) was established. The Stata/SE version 16.1 for Mac (StataCorp LP, College Station, TX, USA) was used for quantitative synthesis.

The meta-analytic output is tested for the propensity of type I (false positives) and type II (false genitives) statistical errors and to analyze the power of available evidence using the Trial Sequential Analysis (TSA) open-access software (Trial Sequential Analysis v0.9 β, Copenhagen Trial Unit, Center for Clinical Intervention Research, Denmark). A type 1 error of 5% and a power of 80% (type 2 error = 80%) were set to calculate trial sequential monitoring boundaries, futility boundaries, and the required information size (RIS). A “model variance-based” approach is defined for the inconsistency correction and I² values established according to meta-analytic data output if the heterogeneity observed is zero; a lower inconsistency (I² = 25%) is assumed for RIS estimation based on a random-effects model as previously reported [32]. The anticipated mean difference between intervention groups is setting for an expected minimal biological plausible difference (lower bound of 95% CI) of − 0.18, − 0.33, and − 0.38 mm between DBBM and DBBM-C, for horizontal ridge preservation at 1, 3, and 5 mm from the bone crest [33]. A graphical evaluation was performed to determine whether the Z curve (showing the treatment effect) crossed either monitoring or futility boundaries and to obtain the RIS threshold.

The initial search found a total of 741 records in the electronic and manual searches. After removal of duplicates and the title and abstract screening, a total of 14 articles remained for full-text assessment (Fig. 1). Six studies were finally included for qualitative analysis [34‐39] and two for quantitative analysis [35, 37]. The reviewers showed an almost perfect level of agreement (k = 0.92). The most common reason for exclusion was those studies that compare different bone filling materials per group. The excluded papers and the reasons for exclusion are listed in Table 1, and the main characteristics of the included studies are summarized in Table 2.

Changes in the primary outcomes were assessed by clinical and radiographic examinations. Bone changes in width and height were examined in four studies using CBCT [35‐38]. Clinical measurement was reported in three studies [34, 38, 39]; one study evaluated the ridge width using digital models [39]; one study recorded the soft tissue thickness using stents [34]. Moreover, only one study registers the width of keratinized tissue band at the baseline and the endpoint of the investigation [38].

The height of the buccal and lingual crest was analyzed in a split-mouth study [36] comparing sockets that had healed spontaneously with those treated with FGG without additional bone graft. After 3 months of healing, the control sockets had lost height in the buccal crestal bone − 1.42 mm; however, the height in the buccal crestal bone was preserved at the test sites − 0.01 mm. This difference between the two groups was statistically significant (p ≤ 0.05). In contrast, both the control and test groups lost width in the buccal and lingual crestal bone; the difference between the control and test groups was not statistically significant (p ≥ 0.05).

Socket seal surgery using FGG or CM was compared by CBCT analysis in two studies [35, 37]. After 5 [37] and 6 months [35], they found width changes at 1 mm below the crest, between − 0.54 and − 1.4 mm respectively, for sockets sealed with FGG, and − 0.67 and − 1.2 mm respectively in sockets sealed with a CM. This difference between the two groups was not statistically significant (p > 0.05).

Natto et al. [38] evaluated the results of CM seal in similar hard and soft tissue remodeling compared to that with CS used as barriers at 4 months following ARP, in combination with freeze-dried bone allograft. They found that reduction in coronal ridge width (− 1.21 mm CM and − 1.47 mm CS) and vertical buccal bone resorption (− 0.30 mm CMS and − 0.79 mm CS) was not statistically significant. A slight increase in buccal gingival thickness at the coronal part was observed in both groups (0.9 mm CM and 0.5 mm CS). The width of keratinized tissue decreased by an average of 0.08 mm in both groups.

The use of resorbable and non-resorbable membranes in ARP was compared by Fotek et al. [34]. The PTFEm exfoliated prematurely by 28 days, whereas the ADM appeared to be incorporated into the tissues. After 4 months, they found ridge width changes between − 0.06 ± 0.24 mm for sockets sealed with ADM and − 0.17 ± 0.24 mm in sockets sealed with a PTFEm. No statistically significant difference was observed between the two treatment modalities (p = 0.280).

Blinding of participants and personnel was not comprehensively reported, resulting in an unclear risk of bias for these domains in all studies. However, random sequence generation, allocation concealment, attrition, and reporting were of less concern in most studies (Fig. 2).

Meta-analysis was able only for the primary outcome of the review. Only 2 studies allow a statistical comparison for RW reduction at − 1, − 3, and − 5 mm from bone crest after 6 months of socket sealing surgery [35, 37]. The results of meta-analyses are depicted visually throughout forest plots (Fig. 3). It was observed that mean differences of RW reduction did not reach statistical significance among data subsets (p ≥ 0.05). The horizontal bone reduction tended to be more accentuated within collagen matrix group at 1 mm (MD − 0.09; 95% CI [− 0.37; 0.19]; I² = 13%) and 3 mm (MD − 0.06; 95% CI [− 0.28; 0.15]; I² = 0%) cutoff points from alveolar bone crest (Fig. 3a, b). These findings occur in a context of low heterogeneity. The analysis at 5 mm from bone crest showed an increase of RW dimension in favor of FGG group (MD 0.15; 95% CI [− 0.35; 0.64]; I² = 68%), but in a context of high heterogeneity. Egger’s test for publication bias is not applicable due to the lack of information (≤ 10 studies); indeed, it is visually explored using funnel plots, which showed a slight asymmetry. Though, these observations must be considered as merely exploratory and interpreted with caution (Additional file 1).

It was observed how the z curve does not cross the statistical significance threshold (brown lines) nor the monitoring boundaries’ (red lines). A sample of 297 grafted sites (individuals) is required for an anticipated horizontal ridge width reduction of − 0.33 mm at 3 mm, for an alpha error (0.05%), and power (1-beta = 0.20%) in a two-tailed statistical test. This required information size is adjusted, assuming a low heterogeneity (I² = 25%) because the observed inconsistency in this meta-analysis was zero. The RIS obtained surpassed the accrued data in the present meta-analysis (Fig. 4b).

The RIS was estimated for the ridge width reduction at − 1 and − 5 mm from the bone crest. The RIS estimation was set according to anticipated mean differences as mentioned above, and it was observed that accrued meta-analytical data for RW reduction was not reached (Fig. 4b, c). Thus, more studies are warranted to determine which alveolar socket seal technique provides better results on ridge preservation, due to the sparse data available to make clinical significance inferences.

To our knowledge, this is the first systematic review focused on specifically evaluating the different SSS techniques. The use of FGG as SSS can have the benefit of preserving the height of the buccal bone following tooth extraction according to the study of Karaca et al. [36]. Sockets with spontaneous healing had lost height in the buccal bone (− 1.42 mm) after 3 months; conversely, the height was preserved at sockets using SSS without bone filling (− 0.01 mm). This seems to be an adequate alternative for those cases in which the buccal table is greater than 1 mm, and the use of bone filling may not be imperative [42].

Additional benefits of SSS are the effects on soft tissue conditions by maintaining and increasing the keratinized tissue using FGG or CM. Three studies compared the two techniques [35, 37, 39] and found them to be effective and predictable for preserving the alveolar dimensions and attaining a band of keratinized tissue which at the same time affords benefits on peri-implant health [43, 44]. Nevertheless, the use of a CM reduces the risk of necrosis in the FGG. No significant differences were found between the two groups; however, CM was associated with a significantly lower patient morbidity [45], avoiding graft donor site involvement. Patient comfort should be an important factor in considering the appropriate treatment alternative for the patient [46].

Since ARP is incapable of entirely avoiding the alveolar ridge reduction [5], SSS surgery procedures may optimize esthetic results by limiting the postoperative external contour shrinkage [14]. Schneider et al. [39] evaluated the horizontal volume change in sockets with ARP and SSS as compared with unassisted sockets clinically. Six months after the extraction, they found a reduction on dimensional changes of the buccal contour when performing ARP with the use of FGG or CM as SSS. These results are consistent with Thalmair et al. [40], in which they determined that regardless of the use of bone, soft tissue cover minimizes the contraction of the buccal contour.

The use of CS material was suggested not only to protect the graft materials but also to induce blood clot formation and to stabilize the wound [47]. Collagen dressing materials are preferable due to their inherent properties. The material is a hemostatic agent and possesses the ability to stimulate platelet aggregation and enhance fibrin linkage, which may lead to initial clot formation, stability, and maturation [48]. A limited number of studies evaluated this technique in combination with ARP [17, 37, 49‐51]. One included study compared it to the use of CM seal and indicated that CS could form an effective barrier for the bone graft in ARP procedures [37]. They found a reduction in coronal ridge width, vertical buccal bone resorption, and a slight increase in buccal gingival thickness compared to CM. The width of keratinized tissue decreased by an average of 0.08 mm in both groups, which shows minimal loss of tissue. The use of CS could be an alternative; however, despite the benefits, these results should be viewed with caution, due to the small number of patients and studies using this alternative.

Even though, most articles recommend the use of absorbable materials for socket seal; intentionally exposed PTFE membrane at post-extraction sites can predictably lead to an increase of keratinized tissue [52] and the preservation of gingival architecture [21]. Although PTFEm is exposed to the oral cavity, they are not permeable to bacteria. Histological data showed that directly after the removal of the membrane there were no endothelial cells or bacterial contamination, on the contrary, initial granulation tissue was found, so PTFE membranes do not seem to interrupt the healing process of the newly formed tissue [53]. However, despite the benefits, the results of the study by Fotek et al. [34] did not find a statistically significant difference when comparing the width of the keratinized tissue and the thickness of the soft tissue compared to an ADM.

Additional techniques have been found on the literature as the ice cream cone technique using a collagen membrane for socket seal [18, 19], the use of PRF as a socket plug [54], or a saddle connective tissue graft [55]. Although these are also techniques commonly used in clinical practice and according to the previously mentioned studies show favorable results, to date, we have not found randomized clinical trials that, according to our inclusion criteria, allow us to add in these or other techniques in this study.

The efforts of the present work rely on the comprehensive literature search, the implementation of critical appraisal of literature methodology and the implementation of quantitative synthesis methods to reduce the propensity of type I error, as the use of a “trial sequential analysis” to determine the adequacy of information size.

Due to seeking the best quality of studies and avoiding bias, only randomized clinical trials were included in this study. Note to mention, the need to find studies comparing SS materials based on the same bone filler between groups had a significant impact on the number of included studies. However, it is important to highlight the importance of this criterion, since having homologous comparative groups will allow us to objectively evaluate the effects of socket seal, which is the main objective of this study.

Among the limitations, the variety of materials, the small number of participants per group, the lack of long-term data, and the lack of homogeneity in particulate bone grafts between studies, are sources of clinical and methodological heterogeneity. Moreover, the sparse data on meta-analysis increased the risk of infraestimating or overestimating the effect size and variance of interventions.

A slight trend for lessening the ridge width loss is observed in favor of the FGG approach, but based on sparse data that did not meet the required information size after a sequential trial analysis. This systematic review supports that socket sealing surgery is a technique that can afford benefits than conventional approaches for alveolar preservation. For this reason, its importance should be underscored, to promote further investigations to enhance the body of the evidence, allowing clinicians the draw of evidence-based reliable recommendations.