The use of dental implants for replacement of missing teeth represents a widely accepted treatment option in oral rehabilitation. Implants have shown satisfactory results in rehabilitation of patient’s function and long-term survival rates of approximately 89% after 10 years [1]. Nevertheless dental implants are not free of complications and long-term prognosis can be altered by technical difficulties or infections [2]. Peri-implant disease can be classified as peri-implant mucositis and peri-implantitis. Whereas peri-implant mucositis is characterized by a reversible inflammation limited to the peri-implant mucosa, peri-implantitis is a non-reversible inflammation of periimplant tissues leading to a reduction of the peri-implant bone level [3]. If no successful treatment is performed, peri-implantitis can bring to implant loss [4-6].

Peri-implant disease is an important pathological entity due to its prevalence and incidence. Peri-implant mucositis has been reported to affect 80% of the subjects and 50% of the implants, while periimplantitis was identified in percentages ranging between 28% and 56% of subjects and 12% and 43% of implant sites, respectively [7].

Peri-implant disease is infectious in nature and the main etiological factor is the development of a bacterial biofilm on implant surface [8]. After being inserted in bone, the supragingival aspect of implant surface is covered by an organic layer formed by proteins, glycoproteins and lipids and this represents the stratum on which bacterial colonization occurs [9].

A study on composition of the microbiota in peri-implantitis revealed the presence of periodontopathogen microorganisms as Prevotella intermedia, Porphyromonas gingivalis, Treponema denticola and Aggregatibacter actinomycetemcomitans [10]. On the other hand, a recent review concluded that even if the microbiota associated with peri-implant disease is similar to that of chronic periodontitis, there are cases specifically associated with other bacterial species, as Peptostreptococcus spp . or Staphylococcus spp. [11].

Although bacterial contamination is the primary etiological factor, there are multiple risk indicators that must be taken into account in peri-implant disease manifestation and progression, as a previous history of periodontitis and cigarettes smoking [3]. Further factors, as the influence of keratinized mucosa, genetic traits and implant surface are under investigation [12].

The diagnosis of peri-implantitis is based on clinical and radiographic findings. Changes in probing depth associated with bleeding or suppuration on probing indicate the presence of periimplant inflammation and radiographs are used to confirm periimplant bone loss [13,14].

The target for the treatment of peri-implant disease is the removal of bacterial biofilm and the disinfection of implant surface. However the presence of screw threads and surface roughness makes implant decontamination a difficult issue. Non-surgical treatment is the firstline intervention to control peri-implant disease.

Different studies on non-surgical approach have reported promising results for inflammation control in terms of reduction of bleeding on probing but whilst peri-implant mucositis can completely heal, there are unpredictable results for the treatment of peri-implantitis [3].

The aim of the present systematic review was to evaluate all the scientific evidence on non-surgical treatment of peri-implantitis and to systematically assess the efficacy of this therapy on implant failure, radiographic margin bone level, complications and peri-implantitis recurrence.

The present systematic review was developed and conducted following the PRISMA statement [15] and, before the start; a protocol was set including these definitions [16].

Focused question

Study population

Types of intervention

Types of comparisons

Outcome

Search strategy

Eligibility criteria for study inclusion

Outcome measures

Screening methods and data extraction

Quality assessment and data synthesis

Assessment of heterogeneity and drawing of conclusions

Focused question

The focused question was: “In patients with diagnosed periimplantitis what are the most effective non-surgical treatment on implant failure, changes in radiographic margin bone level, complications and peri-implantitis recurrence?”

Population, interventions and outcome

The population of interest for this review was represented by humans in good systemic conditions, older than 18 years with at least one implant with diagnosed peri-implantitis.

The definitions used for peri-implantitis were: “plaque-induced progressive marginal bone loss observed on radiographs with clinical signs of infection of the peri-implant soft tissues” [17] for clinical sign and “detectable bone loss from 1-year examination and bone level loss ≥ 1.8 mm” [18] for radiographic evaluation.

The interventions examined were any non-surgical mechanical therapy for biofilm debridement in association or not with any adjunctive measure.

The primary outcomes analyzed were drawn from a previous systematic review of Cochrane Collaboration Group [17].

- Implant failure defined as implant mobility of a previously clinically osteointegrated implant and removal of non-mobile implants because of progressive marginal bone loss or infection.

- Radiographic margin bone level measured as the change on intraoral radiographs taken with paralleling technique.

- Complications and side effects.

- Peri-implantitis recurrence.

The secondary outcomes were probing pocket depth (PPD), clinical attachment level (CAL) and bleeding on probing (BOP).

Search strategy

Two reviewers (GLS and GP) conducted a literature search independently in three electronic databases: PubMed, Embase and Central up to and including 10 February 2017. The following key words and MeSH terms were used:

Population: ("peri implantitis" or periimplantitis or perimplantitis or peri-implantitis or peri-implant* or "peri implant*" or perimplant* or periimplant* or "Peri-Implantitis" [Mesh]).

Intervention: (Treatment or therapy or prevention or management or maintenance or "non-surgical" or "Therapeutics" [Mesh] or ("Tertiary Prevention" [Mesh] or "Secondary Prevention" [Mesh] or "Primary Prevention" [Mesh]) or "Disease Management" [Mesh])).

In addition, the following journals were hand-searched: Journal of Clinical Periodontology, Journal of Periodontology, Clinical Implant Dentistry, Clinical Oral Implants Research, European Journal of Oral Implantology, International Journal of Oral and Maxillofacial Implants, Journal of Oral Implantology, International Journal of Periodontics and Restorative Dentistry. The reference list of the selected studies was scanned for cross-references. In addition, a search in the grey literature was performed on Opengrey (http://www.opengrey.eu) using the words “peri-implantitis” and “treatment” for this specific database. If it was necessary, authors were contacted to provide missing information. There was no language restriction.

Inclusion and exclusion criteria

Inclusion criteria used were defined as follows:

• Non-surgical professional treatments of peri-implantitis.

• Randomized controlled trials either split-mouth or parallel group design.

• Patients in good systemic conditions who have ≥ 1 implant with peri-implantitis.

Exclusion criteria were, instead, defined as follows:

• Less than 10 patients/group.

Study selection

In the first stage, two independent reviewers (GLS and GP) screened titles and abstracts and did the primary search.

The management of the articles was achieved using commercially available software (Endnote X7, Thomson, London, UK).

Subsequently, the studies that could meet the inclusion criteria and articles with unclear information in the title and abstract were selected for full-text analysis, which was performed independently by the same two reviewers (GLS and GP). Any disagreement was discussed with a third member (RB) until consensus was reached.

Finally full-text analysis was performed, excluding the studies that did not meet inclusion criteria and the reasons of this choice were recorded at any step.

In case of discrepancy between the reviewers, disagreements were discussed with a third person (RB) until the reaching of the consensus.

Data extraction and data analysis

Data from the studies were extracted by two independent reviewers (GLS and GP) in specific forms based on type of study design, number of implants treated, durations of follow-up, dimensions of the populations, mean ages, smoking habits, periodontal health, type of implants, treatments performed, measurement methods, outcomes evaluated, sites and sources of funding. Any disagreement was discussed with a third reviewer (RB) when necessary.

Authors of the manuscripts were contacted when data were incomplete or missing. When the results of a study were published more than once, the most complete dataset was included only one time.

To compare the selected studies, primary and secondary outcomes (probing pocket depth, clinical attachment level and bleeding on probing) were pooled and analyzed using weighted mean differences (WMD) and 95% confidence intervals.

Data were calculated using commercially available software (STATA^® 13, StataCorp LP, and Lakeway Drive, College Station, TX, USA).

Statistical significance was defined as a P-value <0.05.

Quality assessment

The quality assessment was conducted in duplicate by two reviewers (RB and GLS). The Cochrane collaboration’s tool for assessing risk of bias [19] was used to determine the quality of the full-text articles included, performed in two steps.

Firstly it was considered if the studies met the criteria for the following domains:

• Sequence generation. Describe the method used to generate the allocation sequence in sufficient detail to allow an assessment of whether it should produce comparable groups.

• Allocation concealment. Describe the method used to conceal the allocation sequence in sufficient detail to determine whether intervention allocations could have been foreseen in advance of, or during, enrolment.

• Blinding of participants, personnel and outcome assessors. Describe all measures used, if any, to blind study participants and personnel from knowledge of which intervention a participant received. Provide any information relating to whether the intended blinding was effective.

• Incomplete outcome data. Describe the completeness of outcome data for each main outcome, including attrition and exclusions from the analysis. State whether attrition and exclusions were reported, the numbers in each intervention group (compared with total randomized participants), reasons for attrition/exclusions where reported, and any re-inclusions in analyses performed by the review authors.

• Selective outcome reporting. State how the possibility of selective outcome reporting was examined by the review authors, and what was found.

• Other sources of bias. State any important concerns about bias not addressed in the other domains in the tool.

Subsequently, a judgement was assigned answering to the previous domains. An answer “+” indicates a low risk of bias and an answer “-” indicates a high risk of bias. The answer “?” indicates an unclear risk of bias. Publications were considered as low risk of bias when all the five criteria were low.

Articles were considered as high risk of bias if at least two of the five criteria were unclear or at least one of the criteria was high or if the randomization was not done in a proper way. When one of the criteria was unclear, the protocol was considered as unclear risk of bias.

Screening

The search strategy returned 4105 results. Of all those, 4103 articles were considered after duplicates removal. During the first stage of study selection 4080 publications were excluded based on titles and abstracts.

In the second stage, 22 potentially relevant publications were evaluated. A total of 10 papers had to be excluded after reading the full-text because they did not fulfill the inclusion criteria of the present review (Table 1). The flow diagram describes the results of search queries (Figure 1). A total of 12 studies were included for the analyses. The pooled analysis comprised 334 patients for 536 implants affected by peri-implantitis.

Table 1: A total of 10 papers had to be excluded after reading the fulltext because they did not fulfill the inclusion criteria of the present review.

Figure 1: The flow diagram describes the results of search queries.

Study population and study design

Variables of study population and study design are presented in Table 2.

Table 2: Variables of study population, study design and comparison were either combined treatments or treatments alone.

The follow-up of the included studies ranged from 3 to 12 months. Three studies presented a split-mouth design with 3 and 6 months of follow-up. Smoking status was reported in 8 studies and ranged from 0% to 38% [30,31].

The periodontal status of the involved patients was expressed in 8 papers [32-39]. The study population ranged from 10 to 63 subjects and the number of treated implants ranged between 22 and 100.

Six studies specified the implant surface, which were TPS [31,33,40] and SLA [30,31,33,36,40] and 1 study [35] didn’t treat TPS or HA surfaces.

In the others six studies, 4 specified only the type of implants used [32,37,38,41] 1 described only the surface as “machined” [34] and the last one didn’t report anything [39].

Type of intervention and comparison

The study interventions and comparison were either combined treatments or treatments alone (Table 2).

Therapies employed for mechanical debridement were sonic/ ultrasonic scalers [39], plastic/carbon/titanium courettes [30,32,38,39] and the VECTOR^® system [32].

Jansaker et al. [37] performed a submucosal mechanical debridement with ultrasonic scaler to all patients and then employed Chloramine perisolve in test group. VECTOR^® system was compared with the efficacy of carbon courettes in the study of Karring et al. [32].

To treat the peri-implants infections Schwarz et al. [31,33] performed a professional cleaning with rubber cups and polishing paste, then then they compared the efficacy of Er:YAG laser with plastic courettes with Chlorhexidine 0.2% irrigations and Chlorhexidine 0.2% Gel. Renvert et al. [34] used Er:YAG laser in the test group and PERIO-FLOW in control group. Bassetti et al. [36] first of all treat patients with carbon courettes in combination with Glycine Air Powder, then use or Diode Laser or Minocycline Hydrochloride Microspheres.

Arisan et al. [40] used plastic courettes associated, in test group, with Diode Laser. In the study by Abduljabbar et al. [41] an Nd:YAG laser was employed as an adjunctive measure to plastic courettes in the test group. Antibiotics and antiseptics were used as chemical agents for treatment of peri-implantitis.

In the study by Gomi et al. [39] systemic azithromycin was prescribed starting 3 days before non-surgical therapy with a doses of 500 mg/day for 3 days, in the test group, as an adjunctive measure to ultrasonic Scaler with plastic tips and plastic courettes. Büchter et al. [30] subgingival irrigations with 8.5% doxycycline were performed in adjunction of plastic courettes and Chlorhexidine 0.2% irrigations in test group.

Antiseptic therapy was performed using 0%, 1% and 0%, 2% subgingival irrigations in association with mechanical treatments in 4 studies [30,31,33,38] while in one paper 0%, 2% or 1% chlorhexidine gel was applied [31]. In one study [36], 0%, 5% chlorhexidine chips were employed in association with ultrasonic debridement and compared with placebo.

John et al. [38] performed a study in which given supramucosal/ gingival professional implant/tooth cleaning with rubber cups and polishing paste to all patients and treated the test group with carbon curettes and Chlorhexidine 0.1% submucosal irrigation with Chlorhexidine 1% gel submucosal application, while the control group received submucosal application of Amino acid Glycine powder.

In 5 studies patients have received treatments before their allocation in test and control groups [30,31,36,38,39] while in the remaining protocols no previous interventions were performed [32-35,37,40,41].

Method of measurement

Three of the selected studies [32,34,41] used intraoral radiographs to evaluate changes in peri-implant marginal bone levels. In one study [40] orthopantomography was used to assess peri-implant marginal bone levels.

In the study by Karring et al. [32] and by Abduljabbar et al. [41] marginal bone changes were calculated as the distance in mm between the bone crest and the most apical bone to implant contact point.

In a similar manner, Renvert et al. [34] calculated the distance between a reference points to the deepest point of the bone lesion in mm.

In the study by Arisan et al. [40] the distance in mm between implant shoulder and the bottom of the defect was assessed by calibrated software.

Changes in the secondary outcomes were assessed by probing with calibrated periodontal probes. In two studies the type of periodontal probe used was not specified [39,40]. Measurements were performed either at four [30,32,34,37,40] or six sites/implant [31,33,36,38,39,41].

In 3 papers probing was performed with calibrated force ranging from 0.15 N to 0.25 N [34,36,37].

Probing pocket depth was assessed in all studies as the distance in millimetres between the gingival margin and the bottom of the sulcus/ pocket. Clinical attachment level was assessed in seven of the selected studies [30,31,33,35-38].

While Bassetti et al. [36] and Machtei et al. [35] calculated the CAL the sum of PPD and REC, other authors measured the CAL as the distance in millimetres from the implant neck to the bottom of the sulcus/pocket.

All of the selected studies evaluated bleeding on probing. Büchter et al. [30], Schwarz et al. [31], Karring et al. [32], Schwarz et al. [33], John et al. [38], Gomi et al. [39] and Arisan et al. [40] evaluated if BOP was evident in the first 30 sec after probing, while Bassetti et al. [36] evaluated the presence of bleeding in the first 10-15 sec after probing.

Four studies [34,35,37,41] evaluated the presence/absence of bleeding on probing after probing pocket depth assessment. Three [30,31,39] studies don’t specify if they were sponsored. Three [33,40,41] studies received funding from university. The remaining six groups [32,34-38] were sponsored by companies of the product tests in their trials.

Interestingly the group of Renvert et al. [34] was sponsored by both companies whose products were compared.

Quality assessment

Data from the quality assessment are reported in and in Tables 3 and 4. All studies were considered to have a high risk of bias except one, which presents a low risk [35].

Table 3: Data from the quality assessment.

Table 4: Descriptive analysis of primary variables.

Study outcome, descriptive analysis of primary variables: Among the included studies, no implant failures are reported at the end of follow-up. Peri-implantitis recurrence was reported in three studies. The study by Schwarz et al. [33], reported disease recurrence in all the treated implants at the 6 months follow-up control, so all patients had to be treated with surgical therapy.

The study by Bassetti et al. [36] reported 5 recurrences, 3 in the control group treated with minocycline and 2 in the test group, treated with photodynamic therapy. The study of Renvert et al. [34] reported that the 11% of the patients in the test and in the control group still presented pus on probing.

Only one of the selected studies Gomi et al. [39] reported 11 complications, 10 made up analgesic administration to treat pain given by non-surgical procedures and 1 case of diarrhoea following the consumption of azithromycin. Four trials [32,34,40,41] reported the changes in radiographic bone level.

The group of Renvert [34], instead, found a mean difference of 0.3 ± 0.9 mm in the laser group and of 0.1 ± 0.8 mm in the air-flow group, failing to found a clinical relevance in this outcome.

In the study by Arisan et al. [40] statistically significant mean bone changes were found in both groups between baseline and the 6 months follow-up (F=38.34, p<0.0001). A statistically significant difference was found also comparing the test and control groups at the end of the follow-up period (0.5227 mm, adjusted p=0.0013) in favour of the laser group.

The group of Karring et al. [32] showed no statistical significant differences between baseline and the 6- month control regarding bone changes, neither within each treatment group nor between treatments, with a difference, after 6 months, of -0.3 ± 1 mm in test group and -0.3 ± 0.8 mm in control group.

In the study of Abduljabbar et al. [41] at baseline and at 6-month follow-up, the mean crestal bone loss was comparable among individuals in groups 1 and 2, with a base-line 1.8 mm of mean loss and 1.7 mm at the follow-up in group 1 and 2.1 mm of mean crestal bone loss at the base-line, that change in 2.2 mm at follow-up in group 2.

Study outcome, descriptive analysis of the changes in probing pocket depth: Data about secondary outcomes are presented in Tables 5 (PPD), 6 (CAL) and 7 (BOP).

Table 5: Differences in PPD between baseline and the end of the investigation for test and control groups.

Table 6: There is a general tendency toward better results when treatment combinations are performed for non-surgical mechanical therapies in terms of PPD reduction and CAL gain.

Table 7: Descriptive analysis of the changes in bleeding on probing.

In the study conducted by Büchter et al. [30] treatment in the test group was performed using plastic courettes in association with Chlorhexidine 0.2% and Doxycycline 8.5% irrigations whilst in the control group plastic courettes with Chlorhexidine 0.2% irrigations were employed.

Between the start and the end the test group showed a mean PPD reduction of 1.15 ± 0.23 mm, while in the control group was 0.56 ± 0.30 mm, with a statistically significant difference between groups of 0.59 ± 0.53 mm (p=0.046).

In the study by Schwarz et al. [31] all the patients were initially treated with supramucosal/gingival professional implant/tooth cleaning with rubber cups and polishing paste. The test group was treated with Er:YAG laser, and the control group with plastic courettes in association with 0.2% chlorhexidine irrigations and gel.

The difference of PPD between baseline and 12 months follow-up in the test group was 0.8 ± 0.1 mm, and 0.6 ± 0.1 mm in the control group. A statistically significant difference between groups of 0.2 ± 0.2 mm was found (p<0.001).

Karring et al. [32] performed a split-mouth study comparing the Vector^® system in the test group with carbon courettes in the control group, finding a difference of 0 ± 0.1 mm and 0.1 ± 0.6 mm respectively. No significant differences between group could be found (0.1 ± 0.7 mm, p>0.1).

In another study, Bassetti et al. [36] used carbon courettes and Glycine air powder on all patients. In this protocol, photodynamic therapy (test) and Minocycline Hydrochloride microspheres (control) were compared. The Authors found a difference of 0.36 ± 0.03 mm and 0.49 ± 0.01 mm, respectively. There was a difference of 0.07 ± 0.04 mm between groups with a p value>0.05.

Schwarz et al. [33] evaluated the efficacy of Er:YAG laser, compared to plastic courettes, 0.2% chlorhexidine irrigations and the use of 0.2% chlorhexidine gel. They found a difference between baseline and 12 months follow-up amounting to 0.33± 2.3 mm in the test group and 0 ± 2.64 mm in the control group. Mean inter-group differences were 0 ± 3.22 mm.

Renvert et al. [34] analyzed the effect of Er:YAG laser in comparison to Perioflow. They found a PPD reduction of 2.13 ± 0.22 mm in the test group and 1.73 ± 0.19 mm in the control group. The inter-group difference amounted to 0.40 ± 0.31 mm. No statistical differences are reported between baseline and 6 months follow-up.

Machtei et al. [35] studied the efficacy of the association of ultrasonic debridement and a matrix containing 0.5 mg of chlorhexidine compared to the association of ultrasonic debridement and a gelatin matrix. The test group showed a reduction of 2.13 ± 0.22 mm, while the control groups a reduction of 1.73 ± 0.19 mm. The inter-group difference amounted to 0.40 ± 0.31 mm. No statistical differences could be found.

John et al. [38] performed a study in which given supramucosal/ gingival professional implant/tooth cleaning with rubber cups and polishing paste to all patients and treated the test group with carbon curettes and Chlorhexidine 0.1% submucosal irrigation with Chlorhexidine 1% gel submucosal application, while the control group received submucosal application of Amino acid Glycine powder. Differences were of 0.5 ± 0.9 mm in the test group, 0.4 ± 0.9 mm in the control group, 0.1 ± 1.8 mm between groups and P value was >0.05.

In the study included by Gomi et al. [39], all patients received Supragingival Scaling using ultrasonic scaler with plastic tips and plastic courettes. In the test group systemic subministration of 500 mg/day of Azithromycin was performed, starting 3 days before. The control group was treated with mechanical debridement alone. There was a difference of PPD of 1.19 ± 0.39 mm in the test group, 0.23 ± 0.01 mm in the control group and 0.96 ± 0.4 mm between groups, with a p value of 0.01.

Arisan et al. [40] performed a study in which the test group was treated with plastic courettes plus one application of diode laser, while in the control group only plastic courettes were employed.

A mean PPD reduction of 0.17 ± 0.41 mm was found in the test group, and of 0.21 ± 0.83 mm in the control group. Between test and control group there was a difference of 0.04 ± 2.41 mm with a p value <0.001.

Study outcome, descriptive analysis of the changes in clinical attachment level: Only six of the selected studies reported clinical attachment level values and in one of them [36] the method of measurement was not clear.

The largest reduction in CAL was found by Machtei et al. [35] with a mean value of 2.21 ± 0.23 mm in the test group and 1.56 ± 0.25 mm in the control group. The difference between groups was statistically significant and amounted to 0.65 ± 0.34 mm (P=0.05).

Instead, Bassetti et al. [36] found the lowest mean CAL reduction with an average value of 0.16 ± 0.04 in the test group, 0.19 ± 0.07 mm in the control group and a difference between groups of 0.03 ± 0.11 mm.

In this protocol no statistically significant differences between groups could be found at the end of the follow-up period (P>0.05). The study by Büchter et al. [30] reported a gain of 1.15 ± 0.03 mm in the test group using plastic courettes, submucosal irrigations of 0%, 2% chlorhexidine and 8,5 doxycycline, and a gain of 0.33 ± 0.06 mm in the control group using plastic courettes and submucosal irrigations with 0%, 2% chlorhexidine. The inter group difference was statistically significant (P=0.024) and amounted to 0.82 ± 0.09 mm.

Schwarz et al. [31] reported a mean gain of 0.6 ± 0.1 mm in the test group and 0.7 ± 0 mm in the control group, with a difference of 0.1 ± 0.1 mm between groups (P>0.05). In a another study Schwarz et al. [33] found a gain in CAL of 0.25 ± 2 mm in the test group and 0.15 ± 2.2 mm in the control group with a mean intergroup difference of 0.1 ± 2.1 mm employing the same treatment modalities.

John et al. [38] reported a gain of 0.5 ± 1.1 mm and 0.6 ± 1.3 mm in the test and control groups respectively, with a non-significant difference between groups amounting to 0.1 ± 0.3 mm (P>0.05).

Study outcome, descriptive analysis of the changes in bleeding on probing: All the included studies reported the BOP outcome. Four of the selected papers [31,32,40,41] didn’t report standard deviations for this parameter so were not included in the resuming Table 7.

Büchter et al. [30] reported a mean reduction of 0.27 ± 0.01 in the test group and of 0.13 ± 0.01 in the control group, with a difference of 0.14 ± 0.02 between groups and a significant P value of 0.01.

Schwarz et al. [33] found a reduction of 45.83 ± 38.7 in the test group and 25.33 ± 22.6 in the control group, with an inter-group difference of 16 ± 30.7 (P<0.01).

In the study by Bassetti et al. [36] a reduction of 2.52 ± 0.25 was found in the test group while the control group showed a reduction of 2.31 ± 0.08, with a difference between groups amounting to 0.21 ± 0.33 and a statistically considerable P value (<0.05).

Machtei et al. [35] found a reduction of 57.5 ± 7.92 with ultrasonic debridement and a matrix containing 0.5 mg of chlorhexidine and a reduction of 45.5 ± 8.8 using ultrasonic debridement and a matrix gelatin. The inter-group difference was 12.1 ± 6.7 in favor of the test group but no statistically significant differences could be found between groups (P=0.3125).

John et al. [38] reported a mean reduction value of 41.2 ± 29.5 in the control group, 16.6 ± 33.4 in the test group, with an relevant difference of 24.6 ± 63.9 and a P value <0.05.

In the study by Gomi et al. [39] a reduction of 24.5 ± 3.9 was found in the test group and of 6.1 ± 1.5 in the control group, setting a difference between groups of 18.4 ± 5.4 with a very low P value (<0.001).

In a more recent study, Jansaker et al. [37] reported a mean reduction of 0.67 ± 0.58 in the test group, which was treated with the association of ultrasonic mechanical debridement and chloramine perisolve and a decrease of 0.64 ± 0.54 in the control group using mechanical debridement alone. An inter-group difference of 0.03 ± 1.12 was found (P=0.001).

The limitations of the present review consist in the paucity of available data and reliable data from studies included. In fact we found a great heterogeneity in the protocols utilized, indeed they have evaluated different endpoints with different follow-up and most of the studies have a high risk of bias.

Non-surgical therapy of peri-implantitis has the primary aim to decontaminate implant surface by the bacterial biofilm, which can reduce implant failure, disease recurrence and parameters such as clinical attachment level, probing pocket depth and bleeding on probing, possibly without causing side effects to the patient. This systematic review was conducted to evaluate the present literature and to provide scientific evidence on the existing RCTs evaluating different therapeutic protocols for non-surgical treatment of peri-implantitis.

The primary endpoints selected were implant failure, changes in radiographic marginal bone level, presence of complications and recurrence of peri-implantitis while the secondary variables were the change in probing pocket depth, gains in clinical attachment level and reduction of bleeding on probing.

In none of the included studies implant failure was adopted as a primary variable, but Authors reported data about this parameter. Among the papers included in the present review, no implant failure was reported at the end of the follow-up period.

Disease recurrence was reported in three trials. In particular, Bassetti et al. [36] reported disease recurrence in 5 of the 39 included implants, treated either with photodynamic therapy or with the application of minocycline microspheres. In a previous trial [33], the test group was treated with Er:YAG laser and the control group with subgingival manual debridement. After six months, the Authors found deterioration in clinical parameters of all patients and had to re-treat the patients and performed bone augmentation procedures.

In another trial in which Er:YAG laser and air-abrasives were employed [34] after six months of follow-up, 11% of the patients in the test and in the control group still presented pus on probing.

These data underline the fact that in short-term evaluation, poor conclusions can be drawn on the efficacy of non-surgical therapy of peri-implantitis to arrest disease progression and stabilize clinical parameters. Studies with longer follow-up are needed.

Moreover, in some of the studies included in this systematic review, treatments were performed only one time, while in other papers the therapies were applied multiple times so the effect of repeatedapplication versus single application of therapies needs to be further investigated.

One paper reported about complications related to treatments. In this study, systemic Azithromycin in association to subgingival debridement was administered to test group patients, while in the control group only subgingival debridement was performed. The majority of complications consisted in pain related to mechanical therapy (10 subjects from both groups) that was controlled with the administration of analgesics, and only one patient in the test group presented diarrhoea.

Among the included studies, four trials [32,34,40,41] reported about radiographic bone changes before and after therapy. The authors were unable to find clinically significant differences both in the test and in the control group between baseline and the end of the study.

There is a general tendency toward better results when treatment combinations are performed for non-surgical mechanical therapies in terms of PPD reduction (Table 5) and CAL gain (Table 6). These results are in agreement with a recent systematic review by Faggion et al. [42], also assessing the influence of additional treatments in association with mechanical debridement alone for non-surgical therapy of periimplantitis.

However, these data must be carefully evaluated, because all except one of the included studies [35] presented a high risk of bias.

Among the different treatment combinations, it seems that the association of mechanical debridement and local antiseptics can bring to better clinical results.

Data by Machtei et al. [35] show reduction of 0.40 ± 0.31 mm greater in PPD when a 0.5 mg chlorhexidine matrix is associated to mechanical debridement with an ultrasonic device compared with the control group in which a gelatin matrix was employed with mechanical debridement.

The study of Gomi et al. [39] is the only study in which systemic antibiotics were employed. The greater results in PPD reduction (1.19 ± 0.39 mm of reduction in the test group and 0.23 ± 0.01 mm in the control group) if compared with topic antibiotics agents, can be explained by the difference in drugs administration. In fact, while with topic agent’s higher concentrations of the drug can be reached in a pocket; the crevicular fluid is able to wash it out rapidly, as demonstrated in the study by Goodson et al. [43].

Instead, using systemic antibiotics agents is possible to obtain lower concentrations in a periodontal/peri-implant pocket, but a continuous drug delivery.

These data have to be evaluated taking into account that studies presented different methods of implant surface decontamination and different combinations of treatments. When further interpreting the results of the qualitative analysis, it could be also noted that combination of treatments result in greater reductions of BOP. Decrease in this parameter could be seen both in test and control groups, but the test groups showed statistically greater improvement than the controls. Only six of the selected studies reported data on Clinical Attachment Level changes from baseline to the end of the follow-up period.

In terms of CAL gain, it seems that the use of antiseptic agents in association to mechanical therapy is able to provide better results. In particular, Machtei et al. [35] found a statistically significant greater CAL gain of 0.65 ± 0.34 mm better in the test group, when 0.5 mg chlorhexidine matrix is associated to mechanical debridement.

The other studies reporting on CAL changes were unable to find statistically significant differences between test and control groups.

In general, there is a lack of consistency of data and this could be due to the following factors:

1. Although all the included studies are RCTs on non-surgical treatment of peri-implantitis, a high heterogeneity in study nature and concept can be seen. In some protocols, the application of treatment was performed one time, while other RCTs are based on repeated applications. Combination of treatments is based on the association of different mechanical and chemical agents. Since most of the trials are based on the association of different treatments, it is difficult to discriminate which can be the most effective.

2. RCTs included are different in their design, either split-mouth or parallel groups design.

3. There is a lack of standardization in the control groups. This means that nowadays no gold-standard therapy is established for nonsurgical treatment of peri-implantitis. Often, the same therapies are used both as a test or a control group. Furthermore comparisons between completely different types of interventions are made.

These problems were treated in a recent systematic review in which the authors pointed out that the quality of evidence on peri-implantitis is low, even if the most recent publications show a higher level of report [44].

Even if statistically significant improvements in clinical parameters can be found with the different treatments alone or in combination, one may argue if these results are clinically significant. Within the limitations of the present systematic review, it seems that combinations of treatments result in a greater reduction in PPD when compared with treatments alone.

In order to better define which therapy can provide better clinical outcomes, researches should follow some important features. First of all therapies should be defined and clearly distinguished for test group, while control group should not receive any adjunctive treatment. This is because no gold-standard therapy is now eligible for peri-implantitis. The sample of population should be well described reporting data about sex, age and ethnicity and about risk factors such as smoke and diabetes. Number and type of implants per patient and their sites should be reported for each group involved. The same trained assessor should measure the outcome variables and the methods of recording and then register measurements over time, detailing every intervention or drop out. Follow-up measurements should continue every two months for at least 12 months, conforming to set methods.

Non-surgical therapy of peri-implantitis seems to have a limited efficacy. The association of mechanical debridement and adjunctive measures seems to provide slightly greater benefits. In particular the use of systemic azithromycin associated to mechanical debridement can give better clinical outcomes; indeed in the study of Gomi et al. [39] was reported a greater reduction of 0.96 ± 0.4 mm in the test group compared with the data of the control group after 1 year of follow up.

Clinical trials on this topic are characterized by a great heterogeneity and nowadays there is no gold-standard therapy that can be used as control. Better and standard designed clinical trials and with longer follow-up are needed.

The authors declare no conflict of interest concerning the contents of the study. The study was self-founded by the authors.