Impact of surface decontamination and systemic antimicrobials for surgical treatment of peri-implantitis: A systematic review and meta-analysis of randomized clinical trials

Background: Efficient control of infection is essential to achieve desired outcomes in the surgical treatment of peri-implantitis lesions, although methods employed are largely heterogeneous. Purpose: To compare the impact of different decontamination protocols and adjunctive systemic antimicrobials on the outcomes of surgical treatment of peri-implantitis. Materials and methods: Randomized clinical trials (RCTs) on surgical treatment of peri-implantitis were selected through an electronic search on Medline, Embase, Scopus, and Central databases. Only studies comparing two or more anti-infective strategies were included. Following data extraction, two different sets of meta-analyses were performed. Firstly, overall impact of different implant surface decontamination methods was assessed by comparing baseline values with outcomes at 6-12 months. Secondly, pairwise comparisons evaluated the potential benefit of adjunctive systemic antimicrobials over placebo. Results were expressed as weighted mean effect (WME), weighed mean difference (WMD) or risk ratio (RR). Results: Sixteen RCTs were included. No pairwise comparisons were available for different surface decontamination methods. Use of curettes resulted in improved probing depth (PD) (WME = 2.13 mm), but the results in terms of marginal bone levels (MBL) and percentage of disease resolution were unsatisfactory. Moreover, the adjunctive benefit of systemic antimicrobials over placebo was evaluated in two studies, representing a total of 178 implants. The meta-analyses identified a higher disease resolution (RR = 1.50) for test procedures which approached statistical significance. In terms of overall outcome, systemic antimicrobials with open flap debridement resulted in improved MBL (WME = 0.44 mm), reduced PD (WME = 2.46 mm) and 51.4% of disease resolution. Conclusions: There is not enough evidence to support adjunctive usage of systemic antimicrobials together with the surgical treatment of peri-implantitis. Moreover, higher consistency is required to support the superiority of a surface decontamination protocol over another (PROSPERO CRD42020182303).


Introduction
Peri-implantitis is a rather frequent biological complication negatively affecting the success and survival of dental implants that is characterized by inflammation in the adjacent connective tissue and progressive resorption of supporting bone. 1 Its estimated prevalence is around 22%, although there is a wide variation in case definitions and diagnostic criteria used across the studies. 2,3 Periimplantitis has a bacterial etiology, and the success of treatment mostly depends on arresting the inflammatory process through efficient control of infection and removal of dysbiotic biofilm from the implant surface. 4 Treatment of peri-implantitis traditionally involves a first phase of supramucosal plaque control and a subsequent deep debridement of the implant surface. 5 Nonetheless, clinical success of non-surgical methods is limited in resolving peri-implantitis, and surgical intervention is usually recommended for better mechanical access to contaminated implant surfaces and efficient decontamination. 6,7 Although surgical procedures demonstrated favorable treatment outcomes in terms of reductions in probing depth (PD) and inflammation, 8 complete disease resolution (DR) still remains unattainable for most cases. 9,10 Insufficient implant decontamination may be regarded as a major explanation for these limited outcomes. 11 Due to poor mechanical access to the bottom of bony defects and differences in micro and macro topography of titanium interfaces, appropriate surface decontamination continues 7 There are systematic reviews investigating the efficacy of surgical techniques and biomaterials in improving the clinical parameters of peri-implantitis lesions. 18,19 However, the clinical impact of available implant surface decontamination procedures has not been systematically evaluated so far. Proving the superiority of a specific method would be of uttermost relevance in enhancing the consistency of outcomes in peri-implantitis treatment and research. Accordingly, the evidence supporting usage of peri-surgical systemic antimicrobials is limited, and further data is required to justify their prescription. Therefore, the aims of the present systematic review were: 1) to evaluate and compare the efficacy of different implant surface decontamination protocols for the surgical treatment of peri-implantitis; 2) to investigate whether adjunctive systemic antimicrobials provide additional clinical benefits.

Report and protocol
This systematic review was reported according to the PRISMA statement 20 and the protocol was registered on PROSPERO (CRD42020182303).

Focused questions
This systematic review was designed to answer the following focused questions:

Eligibility criteria
Inclusion criteria were determined a priori and organized according to the PICOST acronym.
2.3.1 PICO 1 (P) Population. Patients in good general health requiring surgical treatment of peri-implantitis.
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted March 1, 2021

Search methods for the identification of studies
The reviewing authors were calibrated before each phase of the study. Electronic search (Supplementary Table 1) was conducted on Medline (via PubMed), Embase, Scopus, and Cochrane Library electronic databases independently by two authors (GB and FC) with no restrictions on language, date of publication or publication status up to December 2020. In addition, hand searching (GB and NB) was performed on periodontics/implantology-related journals.

Study selection
Titles and abstracts (when available) of all identified studies were screened by two independent reviewers (NB and GB). Any disagreement was resolved by discussion with a third reviewer (FC).
Full text of studies of possible relevance or for which there was insufficient data in the title and abstract were assessed independently by two reviewers (NB and FC). Differences between them were settled by a third reviewing author (GB). The reasons for exclusion after the full text analysis were recorded.

Data extraction and management
Data from included studies were extracted by two reviewers (GB and FC) independently using predefined data extraction forms (Table 1). If necessary, corresponding authors of the included studies were contacted for clarification of any missing information. If no reply was received within three months, the study was excluded.
Data on general information (first author, year of publication and setting); methods (study design, diagnostic criteria for peri-implantitis, sample size, follow-up period, population); interventions and controls (pre-treatment phase, type of surgery, type/combination of decontaminating agents, biomaterials, post-surgical care) and outcomes (changes in radiographic MBL, CAL, PD, REC percentage of implants with BoP and/or suppuration, implant survival, DR, PROMs) were tabled.
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) 1 0

Assessment of risk of bias in the included studies
The risk of bias in the included studies was assessed independently and in duplicate by the two reviewing authors (GB and FC) according to RoB.2 tool. 22

Data analyses
For continuous outcomes at 6-12 months (changes of MBL and PD reduction), mean values and standard deviations were combined and analyzed with weighted mean effect (WME) and 95% confidence intervals (CIs). Dichotomous data were pooled as weighted mean percentage and 95% CIs. In pairwise comparison, the estimates of the effect were expressed as weighted mean differences (WMD) and 95% CIs for continuous outcomes and as risk ratio (RR) and 95% CIs for dichotomous outcomes. In order to account for within-patient correlation in studies which failed to adjust for it, an intracluster correlation coefficient of 0.07 was assumed for the calculation of the effective sample size and CIs. Study-specific estimates were pooled with and random-effects models to account for between studies heterogeneity. 23 . The Mantel-Haenszel method was performed to combine the dichotomous outcomes, and the inverse of variance method to combine the continuous outcomes.
Two different sets of analyses were conducted. First, the outcome of different protocols for the decontamination of the implant surface was assessed by comparing baseline values with outcomes obtained with follow-up longer than 6 months. Only protocols reported exactly the same procedures, including the type of surgery [i.e. open flap debridement (OFD), resective or regenerative], in two or more studies were included. For studies with multiple arms, each decontaminating intervention was considered separately. Secondly, RCTs were used to evaluate the potential benefit of adjunctive systemic antimicrobials in OFD. For this analysis, no discrimination was made between the different methods of surface decontamination. OFD without systemic antimicrobials was considered as control.
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) 1 1 Statistical heterogeneity among studies was explored using the I 2 index 24,25 and the Cochrane's Q statistic (p < 0.1). For each meta-analysis forest plots were generated. Statistical significance was set to p < 0.05. Statistical analyses were performed using statistical software package OpenMeta [Analyst]. 26 3. Results

Search
As outlined in Supplementary Figure 1 Ultrasonic scalers were adopted in 10 arms of five studies; powder spray in five arms of four studies; titanium brushes in four arms of three studies, while implantoplasty was carried out in six arms of four studies. Among chemical and physical decontaminations, chlorhexidine (CHX) at 0.12 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted March 1, 2021. ; https://doi.org/10.1101/2021.02.26.21252213 doi: medRxiv preprint or 0.2 formulations was applied in four arms of three studies; lasers were used in three test groups of three different studies; a gel of metronidazole followed by a solution of tetracycline hydrochloride was rubbed on implant surfaces in both arms of the study by Romeo et al.; 45 while enamel matrix derivatives (EMD); 46 ozone therapy; 42 photodynamic therapy; 38 electrolytic current; 40 and minocycline ointments 36 were part of only one study arm. The main outcomes of the included studies are presented in Table 2.

Risk of bias in individual studies
The risk of bias assessment for the included RCTs is summarized in Figure 1. Ten papers had a low risk of bias. Three studies exhibited selection bias, 32,33,42 one attrition bias, 38 and one missing outcome data. 44 One study was considered to have a high risk of bias due to the selection of reported results. 40

Outcomes of different decontamination protocols in OFD
A total of 452 implants in 339 patients were treated by OFD procedures. The MBL change ranged from a reduction of -0.96 mm in a control arm in which only titanium curettes plus gauze soaked in saline were employed, 21 to a gain of 1.12 mm with the use of titanium brushes. 32 The single use of curettes for surface decontamination resulted in a negligible MBL reduction (WME = -0.05 mm, 95% CI: -0.90/0.81) in three studies (three arms), with high heterogeneity (I 2 = 94.82 p < 0.001). In terms of PD reduction, the analysis included four studies (4 arms) yielding a WME of 2.14 mm (95% CI: 1.07/3.20); while the weighed mean percentage of DR amounted to 25% (95% CI: 16/33) with no heterogeneity (Figure 2A,B,C).
None of the studies gave information on defect fill or PROMs, and implant survival was described in only two with results ranging from 90.9% to 100% on the implant level.

Outcomes of different decontamination protocols in reconstructive surgery
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted March 1, 2021. ; https://doi.org/10.1101/2021.02.26.21252213 doi: medRxiv preprint A total of 192 implants in 156 patients were treated using reconstructive or combined procedures.
None of the study arms provided comparable decontamination protocols, thus data meta-analysis was not possible.

Outcomes of different decontamination protocols in resective surgery
A total of 222 implants in 156 patients were treated by resective procedures. The use of CHX and cetylpyridinium chloride (CPC) together with ORS was investigated in two arms of two different studies by the same research group and no systemic antibiotics were used. 34,35 WME in terms of post-treatment MBL was -0.49 mm (95% CI: -1.02/0.04) with significant heterogeneity (I 2 = 85.73; p = 0.008), WME for PD reduction was 2.05 mm (95% CI: 1.722.38) with no heterogeneity (I 2 = 0; p = 0.633). Eighteen out of 85 implants were successfully treated, and this accounted for a weighted mean percentage of DR of 17% with significant heterogeneity (I 2 = 92.72; p < 0.001) ( Figure   2D,E,F).

Potential advantages of adjunctive systemic antimicrobials
Only two studies directly compared OFD with and without adjunctive systemic antibiotics. 17,21 , while no data were available for the other surgical treatment approaches. The analysis evaluated 178 implants for PD reduction as outcome and a total of implants ranging from 178 to 209 for the secondary outcomes. Implant survival rate was comprised between 93.3% and 100%. is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted March 1, 2021. ; 1 5 RR of 1.50 (95% CI: 0.94/2.38; p = 0.086) favoring the antibiotics administration with low heterogeneity. None of the studies reported on defect fill or PROMs, while implant survival over one year or more was described in two. ORS was performed with or without systemic antibiotics but no study provided a direct comparison, thus, it was not possible to perform a pair-wise meta-analysis. As depicted in Figure 5, only two arms of the same study reported data on MBL changes 12 months after ORS combined with systemic antibiotics. The WME was −0.27 mm (95% CI: −0.82/0.29) with significant heterogeneity. 45 Four arms in two different studies from the same research group used ORS without systemic antibiotics. 34,35 WME for MBL accounted for -0.40 mm (95% CI: -0.67/0.13) with significant heterogeneity (I 2 = 81.76; p < 0.001), WME for PD reduction was 1.81 mm (95% CI: 1.59/2.03) with non-significant heterogeneity (I 2 = 23.05, p = 0.273). The same studies reported DR in a total of 26 successfully treated implants out of 171. This accounted for a weighted DR percentage of 11% (95% CI: 1/22) with significant heterogeneity (I 2 = 83; p < 0.001).
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

Discussion
Peri-implantitis has become a rather prevalent complication with the increasing placement of dental implants for the rehabilitation of missing teeth. 47,48 Infection control has a pivotal role in the treatment of peri-implantitis, and there is a need for reliable and replicable methods. The present systematic review aimed at evaluating the efficacy of different anti-infective protocols for the surgical treatment of dental implants with clinical diagnosis of peri-implantitis. Overall, 16 RCTs were included in the qualitative and 9 in the quantitative analysis. Cohort studies or case series were excluded to obtain a higher quality of data and decrease the risk of bias.

Surface decontamination
Implant surface decontamination represents a fundamental step to achieve resolution of periimplantitis lesions. However, the available clinical protocols show great variation and there is no evidence for the most effective procedure. 5 Generally, there is a distinction between mechanical and chemical/physical decontamination methods, although a combination is usually recommended. 12 Unfortunately, due to a significant heterogeneity within and among RCTs in terms of study design (definition of peri-implantitis, outcome, decontamination protocols, supportive care) and characteristics of the population (patient, implant and prosthetic features), a pairwise comparison between decontamination strategies was not possible and it remains unclear whether one protocol would provide greater benefits than others. However, the performances of different methods have still been assessed pooling data across studies.
OFD represents the best model for studies investigating clinical impact of different surface decontamination protocols at dental implants as it enables evaluation of the individual effect of the particular decontamination method on the final outcomes. Curettes are the most commonly used instruments for mechanical decontamination. Four study arms employed curettes alone to clean the implant surface during OFD. Improved PD values were reported (WME = 2.14 mm), while no significant differences were found in MBL changes (WME = -0.05 mm), and the effect in disease . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted March 1, 2021. ; https://doi.org/10.1101/2021.02.26.21252213 doi: medRxiv preprint resolution was very modest (25%). It has to be kept in mind that the type of curettes (titanium or plastic) may account for a small variability in the findings. 49 The RCT by Toma et al. 32 offers a direct comparison between three different methods of mechanical decontamination. The titanium brush and glycine air-polishing devices were more effective than the plastic curettes, but treatment success remained low for all three treatment modalities (33%, 29% and 22%, respectively).
The two RCTs by de Waal et al. 34,35 indicated no additional significant benefit for the adjunctive employment of 0.12% CHX + 0.05% CPC with ORS. It is likely that implant surface decontamination may have less prominent effect when ORS is combined with apically positioned flap as a submucosal part becomes usually exposed by the surgical procedure. Moreover, there is controversial evidence on the role of implantoplasty. Romeo et al. 27,45 reported superior 3-year clinical and radiographic results with this intervention. On the contrary, no clinical benefit of implantoplasty on implant survival rate was observed in a long-term retrospective study, with other factors such as the initial defect depth and the frequency of recalls accounting for a larger influence on the outcomes. 50 However, maintenance may be easier for smooth surfaces than surfaces with micro-and macro-plaque-retentive elements.
Ideally, implant surface decontamination should remove biofilm without causing surface damage not to render surfaces more conducive to bacterial colonization. 14 In vitro studies indicated that non-metal curettes and rubber cups were minimally traumatic but ineffective to clean contaminated titanium surfaces; while ultrasonic scalers, metal curettes and rotating titanium brushes were effective particularly on modified titanium surfaces. 51 The air abrasive system was effective in all types of implant surfaces, with glycine powder causing less alterations than sodium bicarbonate. 11,52 Nevertheless, in vitro studies do not properly simulate accessibility to the implant surface during surgical intervention. Anatomical factors and thread geometry can limit access of the decontamination devices, and implants with lower thread pitch and thread depth values may cause insufficient cleaning especially for instruments such as plastic curettes and air powder inserts with less flexibility. 53 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted March 1, 2021. ;

8
Removal of bacterial biofilm from contaminated implant surfaces with air-powder abrasion has been reported to provide significantly superior microbiological outcomes than chemical decontamination with hydrogen peroxide and chlorhexidine gluconate. 54 Coupling mechanic instruments with chemical/physical agents may improve the overall cleaning ability as the chemical agent may reach niches inaccessible for the instruments. 12  On the other hand, current evidence shows that laser or photodynamic therapy in combination with surgical/non-surgical treatment of peri-implant diseases provided minimal benefit in PD reduction, CAL gain, and MBL improvements. 59,60 However, properly controlled clinical trials with low risk of bias are warranted to better clarify the issue.
Within the context of surgical regenerative treatment of peri-implantitis, several decontamination techniques have been described, but the large heterogeneity in flap design, biomaterials, bone defect anatomy and post-operative care restricts the conclusion on their efficacy. Overall, the findings of the present systematic review support the use of systemic antibiotics in the surgical treatment of peri-implantitis, even though a careful risk/benefit analysis is mandatory considering PROMs and the growing issue of antibiotic resistances. The available RCTs included in this review fail to adequately address these aspects.
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. To the best of our knowledge, this systematic review is the first attempting to provide a qualitative and quantitative evaluation of the clinical outcomes of anti-infective protocols across different surgical procedures reported in published RCTs. This systematic review had a rather wide scope including RCTs using regenerative, resective, and combined surgery. This approach and the information outlined in Table 2  is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted March 1, 2021. ; https://doi.org/10.1101/2021.02.26.21252213 doi: medRxiv preprint 1 of bias are warranted in order to establish an evidence-based protocol for decontamination during the surgical treatment of peri-implantitis.

Conclusion
There is lack of evidence to support one intra-surgical protocol of implant surface decontamination over others. The use of curettes during OFD is able to produce PD reduction but has limited effect on MBL and DR; while different CHX formulations do not offer any benefits. Lack of replication was observed for all other chemical and mechanical decontamination methods and future research should aim for higher consistency. In parallel, the use of adjunctive systemic antimicrobials with the surgical treatment of peri-implantitis seems to be supported by a tendency for a clinical improvement at short-term, even though the risk/benefit ratio should be cautiously assessed by the clinicians.
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.  . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted March 1, 2021. is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint  .

CC-BY 4.0 International license
It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint .

CC-BY 4.0 International license
It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted March 1, 2021. .

CC-BY 4.0 International license
It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint      is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted March 1, 2021. ; https://doi.org/10.1101/2021.02.26.21252213 doi: medRxiv preprint