J Clin Periodontol. 2019;1–17. wileyonlinelibrary.com/journal/jcpe  | 1© 2019 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd

Received:5June2018  |  Revised:7January2019  |  Accepted:9January2019DOI: 10.1111/jcpe.13070

S U P P L E M E N T A R T I C L E

Efficacy of reconstructive surgical therapy at peri- implantitis- related bone defects. A systematic review and meta- analysis

Cristiano Tomasi1  | Erik Regidor2 | Alberto Ortiz-Vigón2,3  | Jan Derks1

1DepartmentofPeriodontology,InstituteofOdontology,TheSahlgrenskaAcademyatUniversityofGothenburg,Gothenburg,Sweden2ClínicaOrtiz-Vigón,PerioCentrum,Bilbao,Spain3ETEPResearchGroup,FacultyofOdontology,UniversityComplutenseofMadrid,Madrid,Spain

CorrespondenceJanDerks,DepartmentofPeriodontology,InstituteofOdontology,TheSahlgrenskaAcademyatUniversityofGothenburg,Gothenburg,Sweden.Email:jan.derks@odontologi.gu.se

Funding informationThisConsensusMeetingwassupportedbytheOsteologyFoundationandbytheEuropeanFederationofPeriodontology.

AbstractObjectives:  The present systematic review aimed at evaluating the efficacy ofreconstructivesurgicaltherapyatperi-implantitis-relatedbonedefects.Methods: Studiesreportingonoutcomesofreconstructivesurgeryatperi-implantitis-related bone defects at 12months were identified through an electronic search.Following data extraction, two different sets of meta-analyses were performed.Primarily, controlled studies were used to evaluate the potential benefit ofreconstructive surgical therapy over controls. Secondly, overall outcome ofreconstructive surgical therapy was assessed by comparing baseline values withoutcomes at 12months. Results were expressed as weighted mean differences(WMD)orriskratios(RR).HeterogeneitywasdescribedbyI2andpredictionintervals.Results: Thepotentialbenefitofreconstructivetechniquesovercontrolprocedureswasevaluatedinthreestudies,representingatotalof116implants.Altogether,16studies reported on the outcome of reconstructive measures at 12months aftersurgery.Themeta-analysesidentifiedalargerimprovementinmarginalbonelevels(MBL, WMD=1.7mm) and in defect fill (WMD=57%) for test procedures, butfound no differences for clinical measures (reduction of probing depth (PD) andbleedingonprobing(BOP).Changesofclinicalattachmentandsofttissuelevelswerenot considered. In terms of overall outcome, therapy resulted in improved MBL(WMD=2.0mm) and CAL (WMD=1.8mm), in recession (WMD=0.7mm), inreduced PD (WMD=2.8mm) and in reduced BOP (Implants: RR=0.4/Sites:RR=0.2). None of the included studies addressed patient-reported outcomemeasures.Conclusions: Theavailable evidenceon reconstructive therapyatperi-implantitis-relateddefectsislimitedby(a)thelownumberofcontrolledstudies,(b)thelackofcontrolled studies for commonly used procedures, (c) the heterogeneity betweenstudies and (d) the choice of outcomemeasures. A high variability for predictedoutcomesat12monthswasnoted.The interpretationof thedemonstrated largerMBLgainfortestproceduresisdifficultasgraftmaterialmaynotbedistinguishablefrom newly formed bone. Potential aesthetic and patient-reported advantagesremaintobedemonstrated.

K E Y WO RD S

boneregeneration,dentalimplant,peri-implantitis,reconstructivetherapy

2  |     TOMASI eT Al.

1  | INTRODUCTION

Peri-implantitis is a pathological condition occurring in tissuesaround dental implants. It is characterized by inflammation in theperi-implant connective tissue and progressive loss of supportingbone (Schwarz,Derks,Monje,&Wang, 2018). In a systematic re-view,aweightedmeanpatientprevalenceofperi-implantitisof22%wasreported(Derks&Tomasi,2015).Theprevalencerangedfrom1%to47%indifferentreports,mainlyduetothevariationofcasedefinitionsamongstudies(Tomasi&Derks,2012).Itwassuggestedthatperi-implantitis-associatedbone losswastime-dependentandacceleratedovertime(Derksetal.,2016;Franssonetal.,2010).

The goal of peri-implantitis treatment is resolution of softtissue inflammation and, subsequently, the prevention of fur-ther marginal bone loss. Non-surgical treatment modalities arefrequently insufficient to achieve this objective (Faggion, Listl,Frühauf,Chang,&Tu,2014;John,Becker,Schmucker,&Schwarz,2017),whilesurgicalproceduresareconsideredmoreefficaciousinthetreatmentofperi-implantitis(Faggion,Chambrone,Listl,&Tu,2013;Lindhe&Meyle,2008).Thefeasibilityofasurgicalap-proachhasbeenextensivelydemonstratedinpre-clinicalresearch(Albouy, Abrahamsson, Persson, & Berglundh, 2011; Carcuac,Abrahamsson, Charalampakis, & Berglundh, 2015; Persson,Araújo, Berglundh, Gröndahl, & Lindhe, 1999), and clinical ef-ficacy has been documented in studieswith substantial periodsof follow-up (Berglundh,Wennström,&Lindhe,2018;Roccuzzo,Pittoni, Roccuzzo, Charrier, & Dalmasso, 2017; Schwarz, John,Schmucker,Sahm,&Becker,2017).

Outcomesofsurgicaltherapyofperi-implantitisarereportedtobeinfluencedbyimplantsurfacecharacteristics(Carcuacetal.,2017;Roccuzzoetal.,2017)andbytheconfigurationoftheperi-implantbonedefect (Schwarz, Sahm, Schwarz,&Becker, 2010).Bonede-fectsassociatedwithperi-implantitiscommonly involve thewholecircumferenceoftheaffected implantandhaveanangularoutlineonthemesialanddistalaspects(Schwarzetal.,2007).Angularbonedefectsatteethmayhaveanoveralldifferentmorphology,andstud-ieshaveindicatedthatreconstructivetechniquesaresuccessful intermsofclinicalattachmentgainandpreventionofsofttissuereces-sion(Reynoldsetal.,2015).Whilethedocumentationofreconstruc-tivesurgeryintheperiodontalliteratureisextensive,evidenceonitsuseatperi-implantitissitesisonlyemerging.Theaimofthepresentsystematic review was to evaluate the efficacy of reconstructivesurgicaltherapyatperi-implantitis-relatedbonedefects.

2  | MATERIALS AND METHODS

2.1 | Protocol and eligibility

Preferred Reporting Items for Systematic Review and Meta-Analyses (PRISMA)guidelineswereconsidered (Moher,Liberati,Tetzlaff, & Altman, 2009), and the protocol was registered atProspero (CRD42018089236). The focused question of this

systematicreviewwas:Whatisthebenefitofusingareconstruc-tivetechniqueasadjuncttosurgicaltherapyofperi-implantitis?

2.2 | Inclusion criteria (PICOS)

2.2.1 | Population

Patientsingoodgeneralhealthrequiringtreatmentofperi-implantitis.

2.2.2 | Intervention

Reconstructive technique as adjunction to surgical therapy ofperi-implantitis.

2.2.3 | Comparison

Surgicaltherapyofperi-implantitisalone(open-flapdebridement).

2.2.4 | Outcomes

Changes of radiographic marginal bone level, clinical attachmentlevel and soft tissue level. Reduction of probing depth and peri-implant bleeding on probing. Implant survival (in studies with afollow-upof≥5years).

2.2.5 | Study design

Randomized clinical trials (RCT), controlled clinical trials orprospectivecaseserieswithatleast12monthsoffollow-upwithaminimumof10patients(5pergroupincontrolledstudies).

2.3 | Exclusion criteria

• Pre-clinicalstudies.• Surgicaltherapyofperi-implantitisapplyingpocketelimination.• ArticlespublishedinlanguagesotherthanEnglish.

Clinical Relevance

Scientific rationale for review:Peri-implantitis isacommoncomplicationinpatientsprovidedwithimplant-supportedrestorativetherapy,andtreatmentcommonlyrequiressur-gicalaccess.Principal findings:Evidenceonreconstructivetherapyatperi-implantitis-relatedbonedefectsislimited.Abenefitforreconstructivetechniquesovercontrolpro-cedures was observed for radiographic outcomes only.Practical implications:Cliniciansshouldbeawareofthelackof evidence suggesting improved aesthetic or patient-reported outcomes following reconstructive therapy atperi-implantitis-relatedbonedefects.

     |  3TOMASI eT Al.

2.4 | Interventions and comparisons

Studies on reconstructive techniques in the surgical therapy ofperi-implantitis were considered. The following procedures andpossiblecombinationswereaccepted:(a)bonegrafting(autologous,allogeneicorxenogeneic);(b)guidedboneregeneration;and(c)useofbiologicalagents/growthfactors.

Primarily, the outcomes of reconstructive therapywere com-paredtoresultsafteridenticalsurgicalinterventionsomittingthereconstructive technique (control: open-flap debridement). In asecond step, the overall outcome of reconstructive therapywasevaluated by comparing pre-surgical findings with outcomes at12months.

2.5 | Outcome measures

Outcomes at 12months following surgical therapy of peri-implantitiswereextracted.Theprimaryoutcomewas changeofradiographicmarginalbonelevel(MBL)expressedinmm.Wefur-therconsideredfillofthebonedefectexpressedasapercentage.Additional secondary outcomeswere changes of clinical attach-ment level (CAL), changes of soft tissue level (REC), changes ofbleeding on probing (BOP), changes of suppuration on probing(SUP) and changes of probing depth (PD). Combinations of out-comes, that is composite outcomes (absence of additional boneloss,absenceof inflammationandshallowprobing), andpatient-reported outcome measures (PROMs) were also considered.Implant survival was assessed in studies with a follow-up of≥5years.

2.6 | Search strategy

Three electronic databases were searched for relevant articles inFebruary2018usingthefollowingsearchalgorithms.

2.6.1 | MEDLINE via PubMed (2018- 02- 20)

(peri-implantitis OR periimplantitis OR peri implantitis) AND(surgical treatment OR surgery OR surgical OR reconstructiveOR regenerative OR regeneration) NOT (retrospective OR re-viewOR invitroORcase reportORorthopedicORanimalORexperimental)

Filter:English

2.6.2 | Web of science (2018- 02- 20)

((peri-implantitis OR periimplantitis OR peri implantitis) AND(surgical treatmentORsurgeryORsurgicalOR reconstructiveORregenerativeORregeneration))

Refined by: WEB OF SCIENCE CATEGORIES:(DENTISTRYORALSURGERYMEDICINE)ANDDOCUMENTTYPES:(ARTICLE)ANDLANGUAGES:(ENGLISH)

2.6.3 | Cochrane central register of controlled trials (2018- 02- 20)

(peri-implantitis OR periimplantitis OR peri implantitis) AND(surgical treatmentORsurgeryORsurgicalOR reconstructiveORregenerativeORregeneration)

In addition, a hand searchwas performed including referencelistsandprevioussystematicreviews.Titlesofallidentifiedarticleswere screened for eligibility.Abstractswere then studied and se-lectedindependentlybytworeviewers(CT&JD).Relevantarticleswereanalysedinfulltextand,again,independentlyselectedforin-clusion.Thelevelofagreementforthetwoselectionswasexpressedby k-scores.Disagreementwasresolvedbydiscussion.

2.7 | Data extraction

Tworeviewers(ER&JD)extracteddatafromincludedarticlesandentered relevant information into pre-defined evidence tables.Studies were sorted according to design (controlled studies vs.studieswithoutcontrols), andoutcomesat12monthswere illus-tratedforallstudyarms.Forcontrolledstudies,potentialbenefitsoftestprocedureswerehighlighted.Wespecificallyfocusedonin-clusionandexclusioncriteriaforeachofthestudysamplesinorderto evaluate potential heterogeneity among the sampled popula-tions.Incaseofmissingdata/information,authorswerecontacted.

2.8 | Quality assessment

Onereviewer(AOV)assessedtheriskofbiasfor includedstudies.Forrandomizedtrials,criteriadescribedintheCochraneHandbook(Higgins etal., 2011) were used. In six categories (sequencegeneration, allocation concealment, detection bias, attrition bias,selectivereportingbiasandotherpotentialriskofbias),aratingoflow,unclearorhighriskofbiaswasperformed.

Forstudieslackingcontrols,thatisobservationalstudies,riskofbiaswasassessedusingamodifiedversionoftheNewcastle-OttawaScale for cohort studies (Wells etal., 2014).Thus, five items (rep-resentativenessofcohort, ascertainmentofexposure,outcomeatstartofstudy,comparabilityofcohortsandassessmentofoutcome)werescoredaspresent,notpresentornotapplicable.

2.9 | Risk of bias across studies

The publication bias was evaluated using funnel plots for theoutcomes: (a) changes ofMBL and (b) PD reduction. A sensitivityanalysisofthemeta-analysisresultswasalsoperformed,ifplausible,byselectivelyexcludingstudiesfromthedifferentanalyses.

2.10 | Data analysis

Forcontinuousoutcomesat12months(changesofMBL,defectfill,CAL,RECandPD),meanvaluesandstandarddeviationswerepooled

4  |     TOMASI eT Al.

andanalysedwithweightedmeandifferences(WMD)and95%con-fidence intervals (CIs). Fordichotomousdata (BOP), theestimatesoftheeffectwereexpressedasriskratios(RR)and95%CIs.Study-specific estimateswere pooledwith both the fixed- and random-effectsmodels(DerSimonian&Laird,1986),andtherandom-effectsmodelresultswerepresented.

We performed two different sets of analyses. Primarily, con-trolledstudieswereusedtoevaluatethepotentialbenefitofrecon-structivesurgicaltherapy.Forthisanalysis,nodistinctionwasmadebetweenthedifferentsurgicaltechniques.Open-flapdebridementwasconsideredascontrol.Secondly,theoutcomeofreconstructive

surgical therapy was assessed by comparing baseline values withoutcomes at 12months. In this analysis, controlled studies andstudieswithout controlswere included. For studieswithmultiplearms,each reconstructive interventionwasconsidered separately.Further, sensitivity analysiswasperformedby considering studieswithandwithoutcontrolsseparately.

StatisticalheterogeneityamongstudieswasexploredbytheI2 index(Higgins,Thompson,Deeks,&Altman,2003)andCochrane'sQ statistic (p < 0.1). Forest plotswere used to illustrate the out-comesoftheanalyses.Resultswerecombinedwithrandom-effectsmeta-analysis, reporting tau-squared (between studies variance

F IGURE  1 PRISMAflowdiagramdepictingtheselectionprocess

     |  5TOMASI eT Al.

TABLE 1 Includedstudies:Outcomesreportedincontrolledstudies(n=4articles,n=3studies)

Study

Arm

s

Number of 

obse

rvat

ions

at

12

mon

ths

Outcomes

at

12

mon

ths

Bene

fit o

f tes

t pro

cedu

re

at 1

2 m

onth

sCo

mm

ents

Isehedetal.(2016)

Open-flap

debridement

13patients

13implants

ΔMBL:−0.2±1.1mm

ΔPD:−4.0±2.9mm

ΔBOP%(implants):−20

ΔSUP%(implants):−36

ΔMBL:0.5mm

ΔPD:1.5mm(infavourofcontrol

group)

ΔBOP%(implants):0

ΔSUP%(implants):17

Mea

n ΔMBLandΔPDkindlyprovidedbytheauthors.

BOP/SUPpositiveifpresentatanysiteofimplant.

RECandPROMsnotreported.

Nosystemicantibioticsprescribed.

Enamelmatrix

derivative

12patients

12implants

ΔMBL:−0.7±1.1mm

ΔPD:−2.5±2.0mm

ΔBOP%(implants):−20

ΔSUP%(implants):−53

Jepsenetal.(2016)

Open-flap

debridement

26patients

26implants

ΔMBL:−1.0±1.4mm(mesial),

−0.8±1.1mm(distal)

%Defectfill:23.1±46.3

(mesial),21.9±30.2(distal)

ΔPD:−2.6±1.4mm

ΔBOP%(sites):−44.9±38.2

ΔBOP%(implants):−30.8

ΔSUP%(implants):−25.6±32.7

ΔMBL:2.6mm(mesial),

2.7mm(distal)

%Defectfill:55.9(mesial),52.3

(distal)

ΔPD:0.2mm

ΔBOP%(sites):11.2

ΔBOP%(implants):0

ΔSUP%(implants):2.4(infavour

ofcontrolgroup)

Compositeoutcome(BOP-,PD<5mmandabsence

ofadditionalboneloss)

Control:23.0%ofimplants

Test:30.3%ofimplants

BOP%basedon6sites/implant.

BOP/SUPpositiveifpresentatanysiteofimplant.

RECandPROMsnotreported.

Systemicantibioticsprescribed

Poroustitanium

granules

33patients

33implants

ΔMBL:−3.6±2.1mm(mesial),

−3.5±2.2mm(distal)

%Defectfill:79.0±29.9(mesial),

74.2±36.3(distal)

ΔPD:−2.8±1.3mm

ΔBOP%(sites):−56.1±30.5

ΔBOP%(implants):−30.3

ΔSUP%(implants):−23.2±32.8

Wohlfahrtetal.(2012)

12

mon

ths

alsoreportedin:

Andersenetal.(2017)

7

year

s

Open-flap

debridement

16patients

16implants

ΔMBL:−0.1±1.9mm

%Defectfill:−14.8±83.4

ΔPD:−2.0±2.3mm

ΔBOP(sites):−0.56±2.9

ΔMBL:1.9mm

%Defectfill:71.8

ΔPD:0.3mm(infavourofcontrol

group)

ΔBOP(sites):0.18(infavourof

controlgroup)

BOPexpressedasmeannumberofpositivesites(out

of6)perimplant.

SUP,RECandPROMsnotreported.

Systemicantibioticsprescribed.

Submergedhealingfor6monthsfollowingsurgery.

17/32patientsattendedthe7-yearexamination:3

implantsin3patients(allinthetestgroup)werelost

duringfollow-up.

Poroustitanium

granules

16patients

16implants

ΔMBL:−2.0±1.7mm

%Defectfill:57.0±45.1

ΔPD:−1.7±1.7mm

ΔBOP(sites):−0.38±2.1

Not

e.BOP:bleedingonprobing;MBL:marginalbonelevel;PD:probingdepth;REC:softtissuelevel;PROMs:patient-reportedoutcomemeasures;SUP:suppurationonprobing.

6  |     TOMASI eT Al.

TABLE 2 Includedstudies:Outcomesreportedinstudieswithoutcontrols(n=21articles,n=13studies)

Study

Gro

ups

Number of observations

at

12

mon

ths

Outcomes

at

12

mon

ths

Com

men

ts

Aghazadehetal.(2012)

Group1:

Autologousbonegraft

(particulate),resorbablemembrane

22patients

36implants

ΔMBL:−0.2mm(S

E: 0

.3)

ΔPD:−2.0mm(S

E: 0

.2)

ΔBOP%(sites):−44.8(S

E:6.3)

ΔSUP%(sites):−11.5(S

E: 5

.2)

Compositeoutcome1(BOP-,PD<5mmandabsence

ofadditionalboneloss)

Group1:11.1%ofimplants

Group2:20.5%ofimplants

Compositeoutcome2(BOP+at≤1site,PD<5mm

andabsenceofadditionalboneloss)

Group1:13.9%ofimplants

Group2:38.5%ofimplants

BOP/SUP%basedon4sites/implant.

RECandPROMsnotreported.

Systemicantibioticsprescribed

Group2:

Bovinebonemineral(particles),

resorbablemembrane

23patients

39implants

ΔMBL:−1.1mm(S

E: 0

.3)

ΔPD:−3.1mm(S

E: 0

.2)

ΔBOP%(sites):−50.4(S

E: 5

.3)

ΔSUP%(sites):−25.2(S

E:4.2)

Behnekeetal.(2000)

Autologousbonegraft(blockor

particulate),supportingscrewsor

fibringlue

Numberofpatientsnot

reported

18implants

ΔMBL:−3.9mm

ΔPD:−2.7mm

ΔCAL:−2.2mm

SE o

r SDofmeanchangesnotreported.

BOP,SUP,RECandPROMsnotreported.

ΔRECcalculated:0.5mm

Outcomesat12monthsreportedforsubsample.

Systemicantibioticsprescribed.

KhouryandBuchmann

(200

1)Group1:

Autologousbonegraft(blockand

particulate)

7patients

12implants

ΔMBL:−2.4mm

ΔPD:−5.4mm

ΔProbingbonelevel:−2.4mm

SE o

r SDofmeanchangesnotreported.

BOP,SUP,RECandPROMsnotreported.

Systemicantibioticsprescribed

Group2:

Autologousbonegraft(blockand

particulate),ePTFEmembrane

11patients

20implants

ΔMBL:−3.3mm

ΔPD:−4.8mm

ΔProbingbonelevel:−3.7mm

Group3:

Autologousbonegraft(blockand

particulate),collagenmembrane

7patients

9implants

ΔMBL:−2.5mm

ΔPD:−3.3mm

ΔProbingbonelevel:−2.5mm

Matarassoetal.(2014)

Bovinebonemineral(particles),

collagenmembrane

11patients

11implants

ΔMBL:−2.8mm

%Defectfill:93.3±13.0

ΔREC:−1.3mm

ΔPD:−4.1mm

ΔCAL:−3.0mm

ΔBOP%(sites):−13.6

SE o

r SDofmeanchangesnotreported.

BOP%basedon6sites/implant.

6of11implantsfreeofBOPpriortosurgery.

SUPandPROMsnotreported.

Systemicantibioticsprescribed

Nartetal.(2017)

Allograft(impregnatedwith

vancomycinandtobramycine),

collagenmembrane

13patients

17implants

ΔMBL:−3.6mm

%Defectfill:87.0±18.2

ΔREC:−1.3±0.5mm

ΔPD:−4.2±1.5mm

ΔBOP%(sites):−70.6

ΔSUP%(sites):−100.0

SE o

r SDofmeanchangesnotreportedforall

outcomes.

BOP/SUP%basedon6sites/implant.

PROMsnotreported.

Nosystemicantibioticsprescribed

(Continued)

     |  7TOMASI eT Al.

Study

Gro

ups

Number of observations

at

12

mon

ths

Outcomes

at

12

mon

ths

Com

men

ts

Roccuzzoetal.(2011)

12

mon

ths

alsoreportedin:

Roccuzzoetal.(2017)

7

year

s

Group1:TPSimplants

Bovinebonemineral(block)

14patients

14implants

ΔMBL:−1.6±0.7mm

ΔPD:−2.1±1.2mm

ΔBOP%(sites):−33.9

ΔSUP%(implants):−60.0

BOP%basedon4sites/implant.

SUPexpressedas%ofimplantsdemonstrating

suppuration.

RECandPROMsnotreported.

Systemicantibioticsprescribed.

24/26patientsattendedthe7-yearexamination:4

implantsin4patients(2ineachgroup)werelost

duringfollow-up

Group2:SLAimplants

Bovinebonemineral(block)

12patients

12implants

ΔMBL:−1.9±1.3mm

ΔPD:−3.4±1.7mm

ΔBOP%(sites):−60.4

ΔSUP%(implants):−100.0

Roccuzzoetal.(2016)

Bovinebonemineral(block)

71patients

71implants

ΔPD:−2.9±1.7mm

ΔBOP%(sites):−53.2

ΔSUP%(implants):−35.5

Compositeoutcome1(BOP-andPD≤5mm)

49.3%ofimplants

Compositeoutcome2(BOP-andPD≤6mm)

56.0%ofimplants

BOP%basedon4sites/implants.

SUPexpressedas%ofimplantsdemonstrating

suppuration.

MBL,RECandPROMsnotreported.

ΔRECfordifferentdefectcategoriesvariedfrom

0.5mmto0.9mm.Estimatedmean:0.69±0.76mm.

Systemicantibioticsprescribed

Roos-Jansåkeretal.(2007a)

Hydroxyapatite,resorbable

mem

bran

e12patients

16implants

ΔMBL:−2.3±1.2mm

ΔREC:−2.8±1.4mm

ΔPD:−4.2±1.5mm

ΔCAL:−1.4±1.7mm

ΔBOP%(implants):−62.5

BOPexpressedas%ofbleedingatdeepestsiteof

implant.

SUPandPROMsnotreported.

Systemicantibioticsprescribed.

Submergedhealingfor6monthsfollowingsurgery.

Roos-Jansåkeretal.(2007b)

12 m

onth

salsoreportedin:

Roos-Jansåker,Lindahl,

Persson,andRenvert(2011)

3

year

sRoos-Jansåkeretal.(2014)

5

year

s

Group1:Hydroxyapatite,

resorbablemembrane

17patients

29implants

ΔMBL:−1.5±1.2mm

ΔREC:−1.3±1.5mm

ΔPD:−2.9±2.0mm

ΔCAL:−1.6±2.0mm

ΔBOP%(sites):−57.7

BOP%basedon4sites/implant.

SUPandPROMsnotreported.

Systemicantibioticsprescribed.

25/38patientsattendedthe5-yearexamination:no

implantswerelostduringfollow-up

Group2:Hydroxyapatite

19patients

36implants

ΔMBL:−1.4±1.3mm

ΔREC:−1.6±1.6mm

ΔPD:−3.4±1.6mm

ΔCAL:−1.8±1.4mm

ΔBOP%(sites):−67.9

TABLE 2  (Continued)

(Continued)

8  |     TOMASI eT Al.

Study

Gro

ups

Number of observations

at

12

mon

ths

Outcomes

at

12

mon

ths

Com

men

ts

Schwarzetal.(2008)

2

year

salsoreportedin:

Schwarz,Bieling,Latz,

Nuesry,andBecker(2006)

6 m

onth

sSchwarz,Sahm,Bieling,and

Becker(2009)

4

year

s

Group1:Hydroxyapatite

9patients

9implants

ΔREC:−0.4±0.2mm

ΔPD:−2.0±0.5mm

ΔCAL:−1.6±0.3mm

ΔBOP%(sites):−44

BOP%basedon6sites/implant.

MBL,SUPandPROMsnotreported.

Nosystemicantibioticsprescribed

Group2:Bovinebonemineral

(particles),collagenmembrane

11patients

11implants

ΔREC:−0.3±0.4mm

ΔPD:−2.7±0.6mm

ΔCAL:−2.4±0.7mm

ΔBOP%(sites):−49

Schwarzetal.(2010)

Bovinebonemineral(particles),

collagenmembrane

27patients

27implants

DefectClassIb(n=9)

ΔREC:−0.4±0.7mm

ΔPD:−1.6±0.9mm

ΔCAL:−1.2±1.1mm

ΔBOP%(sites):−39.9±16.6

DefectClassIc(n=9)

ΔREC:−0.5±0.5mm

ΔPD:−1.6±0.7mm

ΔCAL:−1.1±0.9mm

ΔBOP%(sites):−25.9±14.7

DefectClassIe(n=9)

ΔREC:−0.3±0.6mm

ΔPD:−2.7±0.7mm

ΔCAL:−2.4±1.0mm

ΔBOP%(sites):−61.1±16.7

BOP%basedon6sites/implant.

MBL,SUPandPROMsnotreported.

Nosystemicantibioticsprescribed

Schwarzetal.(2012)

2

year

salsoreportedin:

Schwarz,Sahm,Iglhaut,and

Becker(2011)

6

mon

ths

Schwarz,Hegewald,John,

Sahm,andBecker(2013)

4 ye

ars

Schwarzetal.(2017)

7

year

s

Group1:Implantsurfacedecon-

taminationwithplasticcurettes,

cottonpelletsandsterilesaline

Bovinebonemineral(particles),

collagenmembrane

14patients

ΔREC:−0.5±0.4mm

ΔPD:−2.0±1.6mm

ΔCAL:−1.5±1.6mm

ΔBOP%(sites):−60.1±26.6

BOP%basedon6sites/implant.

MBL,SUPandPROMsnotreported.

Nosystemicantibioticsprescribed.

15/32patientsattendedthe7-yearexamination:4

patientswereexcludedbetweenyears2and3dueto

severere-infection.Noinformationonimplantloss

Group2:Implantsurfacedecon-

taminationwithEr:YAGlaser

Bovinebonemineral(particles),

collagenmembrane

10patients

ΔREC:−0.4±0.2mm

ΔPD:−1.7±1.2mm

ΔCAL:−1.3±1.2mm

ΔBOP%(sites):−55.0±28.4

Wiltfangetal.(2012)

Autologousbonegraft(particulate),

xenogeneicbonegraft

22patients

36implants

ΔDefectdepth:−3.5mm(95%CI:

−4.3/−2.7)

ΔREC:−1.3mm

ΔPD:−4.0mm(95%CI:−4.6/−3.3)

ΔBOP%(implants):−36

ΔSUP%(implants):−72

SE,S

DorCIofmeanchangesnotreportedforall

outcomes.

BOP/SUPpositiveifpresentatanysiteofimplant.

39%implantsfreeofBOPpriortosurgery.

MBLandPROMsnotreported.

Systemicantibioticsprescribed.

Not

e.BOP:bleedingonprobing;CAL:clinicalattachmentlevel;CI:confidenceinterval;MBL:marginalbonelevel;PD:probingdepth;PROMs:patient-reportedoutcomemeasures;REC:softtissuelevel;

SUP:suppurationonprobing;S

E:standarderror;

SD:standarddeviation.

TABLE 2  (Continued)

     |  9TOMASI eT Al.

includedintheanalysis),whichwasusedtocalculatepredictionin-tervals(Borenstein,Higgins,Hedges,&Rothstein,2017).Statisticalsignificancewassettop < 0.05.AllanalyseswereperformedwithReviewManager (RevMan version 5.3. Copenhagen: The NordicCochraneCentre,TheCochraneCollaboration,2014).

3  | RESULTS

3.1 | Search

ResultsofthesearchareillustratedinFigure1.Theelectronicsearchyielded754titles.Oneadditionalarticle (Hamzacebi,Oduncuoglu,& Alaaddinoglu, 2015) was identified through the hand search,renderinganinitialselectionof755records.Followingscreeningoftitlesandabstracts,35articleswereselectedfor full-textanalysis(kappa=0.86).Anadditional10articleswereexcluded,resultingina finalselectionof25articlesdescribing16differentstudies.ThereasonsforexclusionaregiveninTableA-1.Theagreementonthefinalselectionwaskappa=0.87.

3.2 | Description of selected studies

3.2.1 | Design

TheincludedstudiesaredescribedinTables1and2.Outofthe16relevantstudies,threeincludedcontrolsandweredesignedasrandomizedcontrolledtrials(Isehedetal.,2016;Jepsenetal.,2016;Wohlfahrtetal.,2012).Eachstudyincludedonesingleexperimen-talgroup,andopen-flapdebridementwasprovidedtocontrols.

Of the 13 studies without controls, three were designed asrandomizedtrials (Aghazadeh,Persson,&Renvert,2012;Schwarz,John,Mainusch,Sahm,&Becker,2012;Schwarzetal.,2008),whiletheremainingwereobservationalstudies.Atotaloffourstudiesin-cludedmultipleinterventionalarms(Aghazadehetal.,2012;Khoury& Buchmann, 2001; Roos-Jansåker, Renvert, Lindahl, & Renvert,2007b; Schwarz etal., 2008), comparing different reconstructivetechniqueswitheachother.Oneadditionalstudycomparedtheout-comesofonereconstructivetechniqueattwodifferenttypesofim-plants(Roccuzzo,Bonino,Bonino,&Dalmasso,2011),whileanothertworeportedoutcomesoftreatmentatdifferentdefectconfigura-tions(Roccuzzo,Gaudioso,Lungo,&Dalmasso,2016;Schwarzetal.,2010).Schwarzetal.(2012)assessedtwomethodsofsurfacedecon-taminationpriortoonereconstructiveapproach.Theremainingfivestudiesdidnotincludeanycomparativeanalyses(Behneke,Behneke,&d'Hoedt,2000;Matarasso,IorioSiciliano,Aglietta,Andreuccetti,&Salvi,2014;Nart,deTapia,Pujol,Pascual,&Valles,2017;Roos-Jansåker,Renvert,Lindahl,&Renvert,2007a;Wiltfangetal.,2012)

3.2.2 | Study samples

Sample sizes varied from 29 to 63 patients for the controlledstudiesandfrom11to75patientsforthestudieslackingcontrols.Roughly, every second study included selectedparticipants,while

the remaining studies described consecutive patient enrolment.In 10 studies, patientswithmedical conditions (e.g., uncontrolleddiabetes)wereexcluded.Heavysmokingwasanexclusioncriterionin5studies.

All studies were based exclusively on samples from Europeanpopulations.Fourstudiesreportedonpatientstreatedinaprivatepracticesetting,whiletheothersreportedonpatientstreatedinaspecialist or university setting.One study (Jepsenetal., 2016) in-cluded multiple centres, while the remaining investigations wereperformedatsingleclinicalcentres.Aformalpowercalculationwasdescribedinsix(Aghazadehetal.,2012;Isehedetal.,2016;Jepsenetal., 2016; Schwarz etal., 2010, 2012;Wohlfahrt etal., 2012) ofthe16includedstudies.

Meanageofincludedpatientsrangedfrom48to71years,andtheratiobetweenmalesandfemalesincludedvariedfrom0.9to0.1.Theproportionofsmokersrangedfrom15%to70%. In5studies,onlyonesingletypeofimplantwasincluded,while2to11differentimplantbrandsweretreatedintheremainingstudies.Prerequisitesfor theperi-implantdefects tobe treatedvaried considerablybe-tweenstudies.Eightstudies,forinstance,didnotspecifythepres-enceof signs of inflammation (e.g., BOP) for inclusion.All studiesbuttwo(Roos-Jansåkeretal.,2007a,b),however,requiredthepres-enceofanangularbonedefect.Requirementsintermsofdepthofthebonedefectrangedfromaminimumof3to4mm.Twostudies(Jepsen etal., 2016; Nart etal., 2017) further specified the peri-implantdefectsintermsofconfiguration(fordetails,seeTableA-2).

3.2.3 | Interventions

Intwoofthethreecontrolledstudies,poroustitaniumgranuleswereusedintheinterventiongroups(Jepsenetal.,2016;Wohlfahrtetal.,2012), while the third evaluated the potential benefit of enamelmatrixderivate(Isehedetal.,2016).

Allofthestudieswithoutcontrolsusedbonereplacementgrafts,either alone (Behneke etal., 2000; Khoury & Buchmann, 2001;Roccuzzo etal., 2011, 2016; Roos-Jansåker etal., 2007b; Schwarzetal.,2008;Wiltfangetal.,2012)orincombinationwithmembranes(Aghazadeh etal., 2012; Khoury & Buchmann, 2001; Matarassoetal.,2014;Nartetal.,2017;Roos-Jansåkeretal.,2007a,b;Schwarzetal., 2008, 2010, 2012). The most commonly used graft mate-rialwas bovinebonemineral as particles (Aghazadeh etal., 2012;Matarasso etal., 2014; Schwarz etal., 2008, 2010, 2012) or as ablock(Roccuzzoetal.,2011,2016).Hydroxyapatiteandautologousbonegraftswereappliedin3(Roos-Jansåkeretal.,2007a,b;Schwarzetal.,2008)andfour(Aghazadehetal.,2012;Behnekeetal.,2000;Khoury&Buchmann,2001;Wiltfangetal.,2012)studies, respec-tively.Onestudydescribedtheuseofanallograft(Nartetal.,2017).

3.3 | Risk of bias in individual studies

Theassessmentof riskofbias in the included randomized trials isillustrated inTableA-3.A low riskofbiaswasnoted inonestudyonly (Isehedetal.,2016),basedontheapparentdifficultytoblind

10  |     TOMASI eT Al.

theexaminerduringradiologicalassessmentsinthestudiesevaluat-ingbonereplacementgrafts(detectionbias).Inaddition,asubstan-tialpatientdropoutwasnotedforthelong-termfollow-up(7years)fortwostudies(Andersen,Aass,&Wohlfahrt,2017;Schwarzetal.,2017).This,however,wasnotrelevantforoutcomesat12months.

Thequalityofreportingintheselectedobservationalstudiesisdepicted inTableA-4.Threeof thestudies (Roccuzzoetal.,2016;

Schwarz etal., 2010;Wiltfang etal., 2012) met all of the qualitycategories.

3.4 | Risk of bias across studies

NosignificantpublicationbiaswasobservedforthethreecontrolledstudiesintermsofchangeofMBLandPDreduction(FigureA-1aand

F IGURE  2  (a)Forestplot.PotentialbenefitintermsofchangeofMBL(studieswithcontrols).Unitofanalysis:implant.AdditionaldatakindlyprovidedbyIsehedetal.(2016).ForJepsenetal.(2016):averageofmesialanddistal.(b)Forestplot.Potentialbenefitintermsofdefectreduction/defectfill[%](studieswithcontrols).Unitofanalysis:implant.ForJepsenetal.(2016):averageofmesialanddistal.(c)Forestplot.PotentialbenefitintermsofPDreduction(studieswithcontrols).Unitofanalysis:implant.AdditionaldatakindlyprovidedbyIsehedetal.(2016).ForJepsenetal.(2016):averageofmesialanddistal.(d)Forestplot.PotentialbenefitintermsofBOPreductionatimplants(studieswithcontrols).Unitofanalysis:implant.Numberofeventsestimatedbasedonreportedpercentages.(e).Forestplot.PotentialbenefitintermsofBOPreductionatimplantsites(studieswithcontrols).Unitofanalysis:implantsite.Numberofeventsestimatedbasedonreportedpercentagesornumbers

     |  11TOMASI eT Al.

b). Considering only reconstructive interventions from controlledanduncontrolled studies, the funnelplots indicatedno significantpublicationbiasforchangeofMBLbut indicatedasignificantbiasforPDreduction(FigureA-2aandb).

3.5 | Potential benefit of reconstructive surgical therapy

Theanalysisincludedatotalof116implantsfortheprimaryoutcome.Forthesecondaryoutcomes,thenumberofincludedimplantsvariedfrom84to116.HeterogeneityamongstudiesexpressedasI2 varied from0%to88%,dependingontheoutcomemeasure.

3.5.1 | Primary outcome: change of radiographic marginal bone level

Figure2a illustrates the results of the meta-analysis for changesof MBL. All three studies with controls reported on the primaryoutcome. A statistically significant benefit (WMD=1.7mm; 95%CI: 0.3/3.1; p=0.02)wasobserved infavourofthereconstructiveinterventions.

3.5.2 | Secondary outcomes

Defect fill was reported in two of the studies with controls(Figure2b), indicating a statistically significant greater fill at testsites (WMD=56.5%; 95% CI: 39.3/73.8; p < 0.001). The analysisfailed to identify a significant benefit in terms of PD reduction,based on data from all three studies (Figure2c). Similar findingsweremadeforBOPreduction.TheRRforBOPreductionwas1.0(95% CI: 0.8/1.3; p = 0.97) and 0.9 (95% CI: 0.6/1.4; p = 0.70) forimplants and implant sites, respectively (Figures2d and e). Noneof the studieswith controls reported on changes of CAL, REC orPROMs.Implantsurvivaloveratleast5yearswasdescribedinoneoftheinterventionalstudies(Andersenetal.,2017).Overa7-yearfollow-up,3of17implantswerelostinthreepatients,allbelongingtothetestgroup(poroustitaniumgranules).

3.6 | Outcome of reconstructive surgical therapy

The analysis included a total of 433 implants for the primaryoutcome. For the secondary outcomes, the number of includedimplants varied from 147 to 821. Heterogeneity among studies,namelythepercentageofvariationduetoatruevariationbetweentreatmenteffectsinrelationtothevariationduetosamplingerror,variedfrom51%to95%.

3.6.1 | Primary outcome: change of radiographic marginal bone level

Figure3aillustratesthereductionofMBLfrombaselineto12monthspostsurgery.Basedon11armsinsevenstudies,theWMDamountedto2.0mm(95%CI:1.3/2.7;p < 0.001).Thepredictioninterval,that

istheexpectedrangeoftheoutcomeoftreatmentappliedtoaran-dom subject of the overall studied population, was −0.4/4.4mm.Estimateswereconsistent,regardlessofstudydesign.

3.6.2 | Secondary outcomes

ACALgainof1.8mm(95%CI:1.3/2.2;p < 0.001)withapredictionintervalof0.6/3.0(Figure3b)wasobservedbasedonfindingsfromfour studies.ChangeofRECwasassessed in six studies (10arms)andamountedto−0.7mm(95%CI:−1.0/−0.3;p < 0.001)(Figure3c).Theprediction intervalwas−1.7/0.4.PDreduction is illustrated inFigure3d.Theweightedmeaneffectwas2.8mm(95%CI:2.3/3.4;p < 0.001) at 12months, based on 21 arms in 13 studies. Theprediction intervalwas estimated to be0.4/5.3mm.Reduction ofinflammationat12monthswasstatisticallysignificant.TheRRforBOPwas0.4(95%CI:0.2/0.8;p = 0.004)and0.2(95%CI:0.2/0.4;p < 0.001) for implants and implant sites, respectively (Figure3eandf).Therespectivepredictionintervalswere0.1/2.3and0.0/1.7.ResultsofthesensitivityanalysisrevealedthattheRRforBOPforimplantswasconsiderablylowerinstudieswithcontrols(0.07;95%CI:0.0/0.5)thanwithout(0.51;95%CI:0.3/0.8).

NoneofthestudieswithoutcontrolsreportedondefectfillorPROMs,whileimplantsurvivaloveratleast5yearswasdescribedintwo.Resultsrangedfrom83%(Roccuzzoetal.,2017)to100%(Roos-Jansåker,Persson,Lindahl,&Renvert,2014)ontheimplantlevel.

4  | DISCUSSION

Thepresentsystematic reviewaimedatevaluatingtheefficacyofreconstructive surgical therapy at peri-implantitis-related bonedefects. The potential benefit of reconstructive techniques overcontrol procedures was evaluated in three studies, representinga total of 116 implants. Altogether, 16 studies reported on theoutcomeofreconstructivemeasuresat12monthsaftersurgery.Themeta-analyses identified a statistically significant largerMBL gain(WMD=1.7mm)anddefectfill(WMD=57%)fortestprocedures,butfoundnodifferencesforclinicalmeasures(PDreduction,BOPreduction). Changes of clinical attachment and soft tissue levelswerenotconsidered.

In terms of overall outcome, therapy resulted in improvedMBL (WMD=2.0mm) and CAL (WMD=1.8mm), in recession(WMD=0.7mm), in reducedPD (WMD=2.8mm)and in reducedBOP(Implants:RR=0.4/Sites:RR=0.2).Noneoftheincludedstud-iesaddressedpatient-reportedoutcomemeasures.

Resultsofthepresentmeta-analysesareinlinewithcalculationspresented in previously published systematic reviews on the topic(Chan, Lin, Suárez, MacEachern, & Wang, 2014; Khoshkam etal.,2013,2016;Sahrmann,Attin,&Schmidlin,2011).Thepresentanal-ysis suggests that reconstructive therapy at peri-implantitis-relateddefectsisafeasibleconcept,butitisalsoobviousthattheavailableevidence is limited.Themajorityofstudies identified in thepresentreviewwasobservationalandhad thecharacteristicsofcaseseries.

12  |     TOMASI eT Al.

Onlythreeoftheincludedstudieswererandomizedcontrolledtrialsaddressingthescientificquestionofthepresentreview,thatisthepo-tentialbenefitofreconstructivemeasuresoveropen-flapdebridementalone.Noneoftheincludedstudysamplesoriginatedfromoutsideof

Europe,which represents a limitation in termsofgeneralizability. Inaddition,moststudieswereconductedatspecialistoruniversityclin-ics.Thesignificantheterogeneitywithinandbetweensamplesmayinpartbeattributedtothedifferenttechniquesandmaterialsusedand

F IGURE  3  (continued)

     |  13TOMASI eT Al.

to inconsistent inclusion/exclusion criteria.The gender ratio amongselectedstudypopulationsvariedconsiderably,asdidtheproportionof smokers and patients with systemic conditions. There was alsoa noteworthy discrepancy in terms of selection of graftingmaterialamongtheincludedarticles.Themostcommonlyusedmaterialintheuncontrolled, observational studieswas bovine bonemineral,whilenoneofthecontrolledstudiesevaluatedthisspecificproduct.

The variation of outcomes between study samples was ex-plored by I2, illustrating the proportion of variance exceeding thesamplingerror(truevariance)(Higgins&Thompson,2002;Higginsetal.,2003).Thepredictionintervalsobservedinthepresentmeta-analysesshowedawiderangeofexpectedeffects,also indicatinga significant variability between the included studies (Borensteinetal., 2017; IntHout, Ioannidis, Rovers, & Goeman, 2016). For

F IGURE  3  (a)Forestplot.ReductionofMBL(reconstructiveinterventionsfromcontrolledanduncontrolledstudies).Unitofanalysis:implant.AdditionaldatakindlyprovidedbyIsehedetal.(2016).Notconsidered:Baselineand/or12-monthbonelevels(mean)notreported:Roccuzzoetal.(2016),Roos-Jansåkeretal.(2007a,b),Schwarzetal.(2008,2010,2012),Wiltfangetal.(2012),Wohlfahrtetal.(2012).SE or SDofmeanvaluesnotreported:Behnekeetal.(2000).(b)Forestplot.CALgain(reconstructiveinterventionsfromcontrolledanduncontrolledstudies).Unitofanalysis:implant.Notconsidered:Baselineand/or12-monthclinicalattachmentlevels(mean)notreported:Aghazadehetal.(2012),Isehedetal.(2016),Jepsenetal.(2016),KhouryandBuchmann(2001),Nartetal.(2017),Roccuzzoetal.(2011,2016),Roos-Jansåkeretal.(2007a,b),Wiltfangetal.(2012),Wohlfahrtetal.(2012).SE or SDofmeanvaluesnotreported:Behnekeetal.(2000).(c)Forestplot.Changeofsofttissuelevel(REC)(reconstructiveinterventionsfromcontrolledanduncontrolledstudies).Unitofanalysis:implant.Notconsidered:Softtissuelevels(mean)atbaselineand/or12monthsnotreported:Aghazadehetal.(2012),Behnekeetal.(2000),Isehedetal.(2016),Jepsenetal.(2016),KhouryandBuchmann(2001),Roccuzzoetal.(2011,2016),Roos-Jansåkeretal.(2007a,b),Wohlfahrtetal.(2012).(d)Forestplot.PDreduction(reconstructiveinterventionsfromcontrolledanduncontrolledstudies).Unitofanalysis:implant.AdditionaldatakindlyprovidedbyIsehedetal.(2016).Notconsidered:Baselineand/or12-monthPD(mean)notreported:Roos-Jansåkeretal.(2007a,b).SE or SDofmeanvaluesnotreported:Behnekeetal.(2000).(e)Forestplot.RiskforBOPatimplants(reconstructiveinterventionsfromcontrolledanduncontrolledstudies).Unitofanalysis:implant.Notconsidered:BOP%(implants)atbaselineand/or12monthsnotreported:Aghazadehetal.(2012),Behnekeetal.(2000),KhouryandBuchmann(2001),Nartetal.(2017),Roccuzzoetal.(2011,2016),Roos-Jansåkeretal.(2007b),Schwarzetal.(2008,2010,2012),Wohlfahrtetal.(2012).(f)Forestplot.RiskforBOPatimplantsites(reconstructiveinterventionsfromcontrolledanduncontrolledstudies).Unitofanalysis:implantsite.Notconsidered:BOP%(sites)atbaselineand/or12monthsnotreported:Behnekeetal.(2000),Isehedetal.(2016),KhouryandBuchmann(2001),Roos-Jansåkeretal.(2007a),Wiltfangetal.(2012)

14  |     TOMASI eT Al.

example, in termsofMBLchanges, theexpectedoutcome rangedfromnoeffectatall(0isincluded)orbonelosstoimprovementofmarginal bone level of4.4mm. It shouldbekept inmind that the95%CIoftheestimatesimplyreflectstheprecisionoftheestima-tionofthemeanvalue.Thepredictioninterval,however,isanindexofdispersion,indicatinghowwidelytheeffectvariesacrossagivenpopulation(Borensteinetal.,2017).

In theconsensus report fromthe8thEuropeanWorkshoponPeriodontology,itwassuggestedtousecompositeoutcomestode-scribe results of peri-implantitis therapy (Sanz&Chapple, 2012).These should ideally include clinical measures of inflammationandradiographicassessmentsofbone-levelalterations.Followingtheserecommendations,studiesapplyingpocketeliminationtech-niques (Carcuac etal., 2016, 2017) or reconstructive techniques(Aghazadehetal.,2012;Jepsenetal.,2016)havereportedresultsaccordingly. An additional goal of reconstructive therapy is themaintenanceofsofttissueheight.Noneofthecontrolledstudies,however, reported data on soft tissue alterations. The recessionobserved in observational studies (WMD=0.7mm) was statisti-cally significant and correspondedwell with findings reported instudiesonperiodontalreconstructivetherapy(Heden,Wennström,&Lindhe,1999;S.Jepsenetal.,2008;Tonettietal.,2004,2002;Trombelli,Simonelli,Minenna,Rasperini,&Farina,2017).Itremainsunclear, however,whether the reconstructive techniques appliedat peri-implant defects resulted in a better aesthetic outcome ascomparedtocontrols.

Studieswithcontrolsindicatedabenefitof1.7mminfavourofreconstructivemeasuresforchangesofmarginalbonelevel.It shouldbekept inmind,however, that twoof the threecon-trolledstudies(Jepsenetal.,2016;Wohlfahrtetal.,2012)usedporoustitaniumgranulesasbonereplacementgraft inthetestgroups.Itisobviousthatblindingoftheexaminerduringradio-graphicassessments in such studieswasnotpossible.Further,itisunclearhowthepresenceofaradiopaquegraftingmaterialaffectedtheassessmentofmarginalbonelevelsasneitherstudyreported measurement errors. Isehed etal. (2016), who usedenamelmatrixderivativeinthetestgroup,reportedamoderatebenefit of 0.5mm. Considering all different reconstructive in-terventionsincludedinthepresentreview,marginalbonelevelswereimprovedby2.0mmonaverage.However,thepredictioninterval was wide and included 0, indicating a high degree ofvariability.

Despite theobservedpotential benefit in radiographic ap-pearance,reconstructivetherapydidnotprovideanybenefitintermsofreductionofPDandBOP.Consideringoverallchangesfrom baseline to 12months, however, peri-implant inflamma-tion was significantly reduced by the surgical treatment. Theanalysis showed that it was more likely to arrest bleeding ata single site (RR=0.2) than to achieve peri-implant health atallaspectsoftheaffectedimplant(RR=0.4).DatareportedbyJepsen etal. (2016) further illustrate this observation. Whilethepercentageofbleedingsiteswasreducedby45%and56%in control and test groups, respectively, the proportion of

implants freeof anybleeding at12monthswas30% forbothgroups. Similar figures have been described in studies apply-ing pocket elimination techniques. Thus,Carcuac etal. (2017)and Heitz-Mayfield etal. (2018) reported 40% of implants tobe completely free of bleeding at 3 and 5years, respectively.Again,ahighvariation intermsofexpectedchangeofBOPatsiteandimplantlevelswastestified,asillustratedbywidepre-dictionintervals.

Thepresentworksuffersfromanumberofshortcomings.Datafrompre-clinical(e.g.,Albouyetal.,2011;Carcuacetal.,2015)andclinicalstudies(e.g.,Berglundhetal.,2018;Roccuzzoetal.,2017)onsurgical therapy of peri-implantitis pointed towards the impact ofimplantsurfacecharacteristicsontreatmentoutcomes.Further,in-clusioncriteriainregardtotheconfigurationofperi-implantdefectsdiffered between studies,whichmay have influenced subsequenthealing(Schwarzetal.,2010).Andfinally,thefrequencyandqualityof maintenance therapy following reconstructive procedures mayalsohavebeenof importanceashasbeendemonstrated forperi-odontitis patients (Cortellini, Pini Prato, & Tonetti, 1994). Neitherof these factors was, however, considered in the present meta-analyses.Inthesecondarycalculationoftheoverallchanges,mea-sures from baseline and 12monthswere handled as independentdatasets.Itmaybearguedthattreatmentoutcomesare,infact,cor-relatedtotheinitialsituation.Inaddition,differenttreatmentarmsoriginatingfromthesamestudywereconsideredtobeindependentofeachother.Thismayhaveaffectedassessmentofheterogeneity.

In conclusion, the available evidence on reconstructive ther-apy at peri-implantitis-related defects is limited by (a) the lownumberofcontrolledstudies, (b) the lackofcontrolledstudies forcommonly used procedures, (c) the heterogeneity between stud-ies and (d) the choiceof outcomemeasures.Ahigh variability forpredictedoutcomesat12monthswasnoted.TheinterpretationofthedemonstratedlargerMBLgainfortestproceduresisdifficultasgraftmaterialmaynotbedistinguishablefromnewlyformedbone.Potential aesthetic and patient-reported advantages remain to bedemonstrated.

ACKNOWLEDGEMENTS

TheauthorsaregratefultoDr.DanielleLaytonforherkindadviceinpreparingthemanuscript.

CONFLICT OF INTEREST

Theauthorsdeclarenoconflictof interestwithrespecttotheau-thorshipand/orpublicationofthisarticle.

ORCID

Cristiano Tomasi https://orcid.org/0000-0002-3610-6574

Alberto Ortiz-Vigón https://orcid.org/0000-0002-1863-5907

Jan Derks https://orcid.org/0000-0002-1133-6074

     |  15TOMASI eT Al.

REFERENCES

Aghazadeh,A.,Persson,G.R.,&Renvert,S.(2012).Asingle-centreran-domized controlled clinical trial on the adjunct treatmentof intra-bony defects with autogenous bone or a xenograft: Results after12months.Journal of Clinical Periodontology,39(7),666–673.https://doi.org/10.1111/j.1600-051X.2012.01880.x

Albouy, J. P., Abrahamsson, I., Persson, L. G., & Berglundh, T.(2011). Implant surface characteristics influence the outcomeof treatment of peri-implantitis: An experimental study in dogs.Journal of Clinical Periodontology, 38(1), 58–64. https://doi.org/10.1111/j.1600-051X.2010.01631.x

Andersen,H., Aass, A.M., &Wohlfahrt, J. C. (2017). Porous titaniumgranules in the treatmentofperi-implantosseousdefects-a7-yearfollow-up study. International Journal of Implant Dentistry,3(1), 50.https://doi.org/10.1186/s40729-017-0106-2

Behneke, A., Behneke, N., & d'Hoedt, B. (2000). Treatment of peri-implantitisdefectswithautogenousbonegrafts:Six-monthto3-yearresultsofaprospectivestudyin17patients.International Journal of Oral and Maxillofacial Implants,15(1),125–138.

Berglundh,T.,Wennström,J.L.,&Lindhe,J.(2018).Long-termoutcomeof surgical treatment of peri-implantitis. A 2-11-year retrospectivestudy. Clinical Oral Implants Research, 29(4), 404–410. https://doi.org/10.1111/clr.13138

Borenstein,M.,Higgins,J.P.T.,Hedges,L.V.,&Rothstein,H.R.(2017).Basicsofmeta-analysis:I2isnotanabsolutemeasureofheterogeneity.Research Synthesis Methods,8(1),5–18.https://doi.org/10.1002/jrsm.1230

Carcuac, O., Abrahamsson, I., Charalampakis, G., & Berglundh, T.(2015). The effect of the local use of chlorhexidine in surgi-cal treatment of experimental peri-implantitis in dogs. Journal of Clinical Periodontology, 42(2), 196–203. https://doi.org/10.1111/jcpe.12332

Carcuac,O.,Derks, J.,Abrahamsson, I.,Wennström, J. L.,Petzold,M.,& Berglundh, T. (2017). Surgical treatment of peri-implantitis: 3-year results from a randomized controlled clinical trial. Journal of Clinical Periodontology,44(12),1294–1303.https://doi.org/10.1111/jcpe.12813

Carcuac,O.,Derks,J.,Charalampakis,G.,Abrahamsson,I.,Wennström,J.L.,&Berglundh,T.(2016).Adjunctivesystemicandlocalantimicro-bialtherapyinthesurgicaltreatmentofperi-implantitis:Arandom-izedcontrolledclinicaltrial.Journal of Dental Research,95(1),50–57.https://doi.org/10.1177/0022034515601961

Chan,H.-L.,Lin,G.-H.,Suárez,F.,MacEachern,M.,&Wang,H.-L.(2014).Surgical management of peri-implantitis: A systematic review andmeta-analysis of treatment outcomes. Journal of Periodontology,85(8),1027–1041.https://doi.org/10.1902/jop.2013.130563

Cortellini, P., PiniPrato,G.,&Tonetti,M. (1994).Periodontal regener-ationof infrabonydefects (V).Effectoforalhygieneon long-termstability. Journal of Clinical Periodontology,21(9), 606–610. https://doi.org/10.1111/j.1600-051x.1994.tb00751.x

Derks, J., Schaller, D., Håkansson, J.,Wennström, J. L., Tomasi, C., &Berglundh,T.(2016).Peri-implantitis-onsetandpatternofprogres-sion. Journal of Clinical Periodontology,43(4), 383–388. https://doi.org/10.1111/jcpe.12535

Derks,J.,&Tomasi,C.(2015).Peri-implanthealthanddisease.Asystem-aticreviewofcurrentepidemiology.Journal of Clinical Periodontology,42(Suppl16),S158–S171.https://doi.org/10.1111/jcpe.12334

DerSimonian, R., & Laird, N. (1986). Meta-analysis in clinical trials.Controlled Clinical Trials,7(3),177–188.

Faggion,C.M. Jr,Chambrone,L., Listl,S.,&Tu,Y.K. (2013).Networkmeta-analysis for evaluating interventions in implant den-tistry: The case of peri-implantitis treatment. Clinical Implant Dentistry and Related Research, 15(4), 576–588. https://doi.org/10.1111/j.1708-8208.2011.00384.x

Faggion,C.M.Jr,Listl,S.,Frühauf,N.,Chang,H.J.,&Tu,Y.K.(2014).AsystematicreviewandBayesiannetworkmeta-analysisofrandom-ized clinical trials on non-surgical treatments for peri-implantitis.Journal of Clinical Periodontology, 41(10), 1015–1025. https://doi.org/10.1111/jcpe.12292

Fransson, C., Tomasi, C., Pikner, S. S., Gröndahl, K.,Wennström, J. L.,Leyland,A.H.,&Berglundh,T.(2010).Severityandpatternofperi-implantitis-associated bone loss. Journal of Clinical Periodontology,37(5),442–448.https://doi.org/10.1111/j.1600-051X.2010.01537.x

Hamzacebi,B.,Oduncuoglu,B.,&Alaaddinoglu,E.E.(2015).Treatmentof peri-implant bone defects with platelet-rich fibrin. International Journal of Periodontics and Restorative Dentistry, 35(3), 415–422.https://doi.org/10.11607/prd.1861

Heden,G.,Wennström,J.L.,&Lindhe,J. (1999).Periodontaltissueal-terations following Emdogain treatment of periodontal sites withangular bone defects. A series of case reports. Journal of Clinical Periodontology,26(12),855–860.

Heitz-Mayfield,L. J.A.,Salvi,G.E.,Mombelli,A.,Loup,P. J.,Heitz,F.,Kruger,E.,&Lang,N.P.(2018).Supportiveperi-implanttherapyfol-lowinganti-infectivesurgicalperi-implantitistreatment:5-yearsur-vivalandsuccess.Clinical Oral Implants Research,29(1),1–6.https://doi.org/10.1111/clr.12910

Higgins,J.P.,Altman,D.G.,Gotzsche,P.C.,Juni,P.,Moher,D.,Oxman,A. D., … Cochrane Statistical Methods, G. (2011). The CochraneCollaboration's tool for assessing risk of bias in randomised trials.BMJ,343,d5928.https://doi.org/10.1136/bmj.d5928

Higgins,J.P.,&Thompson,S.G.(2002).Quantifyingheterogeneityinameta-analysis.Statistics in Medicine,21(11),1539–1558.https://doi.org/10.1002/sim.1186

Higgins, J. P., Thompson, S. G., Deeks, J. J., & Altman, D. G. (2003).Measuring inconsistency in meta-analyses. BMJ, 327(7414), 557–560.https://doi.org/10.1136/bmj.327.7414.557

IntHout,J., Ioannidis,J.P.,Rovers,M.M.,&Goeman,J.J. (2016).Pleaforroutinelypresentingpredictionintervalsinmeta-analysis.British Medical Journal Open, 6(7), e010247. https://doi.org/10.1136/bmjopen-2015-010247

Isehed, C., Holmlund, A., Renvert, S., Svenson, B., Johansson, I., &Lundberg,P.(2016).Effectivenessofenamelmatrixderivativeontheclinicalandmicrobiologicaloutcomesfollowingsurgicalregenerativetreatmentofperi-implantitis.Arandomizedcontrolledtrial.Journal of Clinical Periodontology,43(10),863–873.https://doi.org/10.1111/jcpe.12583

Jepsen,K.,Jepsen,S.,Laine,M.L.,AnssariMoin,D.,Pilloni,A.,Zeza,B.,&Renvert,S.(2016).Reconstructionofperi-implantosseousdefects:Amulticenterrandomizedtrial.Journal of Dental Research,95(1),58–66.https://doi.org/10.1177/0022034515610056

Jepsen, S., Topoll, H., Rengers, H., Heinz, B., Teich, M., Hoffmann,T., & Jervoe-Storm, P. M. (2008). Clinical outcomes after treat-ment of intra-bony defects with an EMD/synthetic bone graft orEMD alone: A multicentre randomized-controlled clinical trial.Journal of Clinical Periodontology, 35(5), 420–428. https://doi.org/10.1111/j.1600-051X.2008.01217.x

John,G.,Becker, J., Schmucker,A.,&Schwarz,F. (2017).Non-surgicaltreatment of peri-implant mucositis and peri-implantitis at two-piecezirconiumimplants:Aclinicalfollow-upobservationafterupto3years.Journal of Clinical Periodontology,44(7),756–761.https://doi.org/10.1111/jcpe.12738

Khoshkam,V.,Chan,H.,Lin,G.,MacEachern,M.P.,Monje,A.,Suárez,F.,…Wang,H..(2013).Reconstructiveproceduresfortreatingperi-implantitis: A systematic review. Journal of Dental Research, 92(12 Suppl),131S–138S.https://doi.org/10.1177/0022034513509279

Khoshkam,V.,Suárez,F.,Monje,A.,Lin,G.-H.,Chan,H.-L.,&Wang,H.-L.(2016).Long-termradiographicandclinicaloutcomesofaregenera-tiveapproachfortreatingperi-implantitis:Asystematicreviewand

16  |     TOMASI eT Al.

meta-analysis. International Journal of Oral & Maxillofacial Implants,31(6),1303–1310.https://doi.org/10.11607/jomi.4691

Khoury, F., & Buchmann, R. (2001). Surgical therapy of peri-implantdisease:A3-yearfollow-upstudyofcasestreatedwith3differenttechniques of bone regeneration. Journal of Periodontology,72(11),1498–1508.https://doi.org/10.1902/jop.2001.72.11.1498

Lindhe, J., & Meyle, J. (2008). Peri-implant diseases: ConsensusReport of the Sixth European Workshop on Periodontology.Journal of Clinical Periodontology,35(8Suppl),282–285.https://doi.org/10.1111/j.1600-051X.2008.01283.x

Matarasso,S.,IorioSiciliano,V.,Aglietta,M.,Andreuccetti,G.,&Salvi,G.E. (2014).Clinicalandradiographicoutcomesofacombinedresec-tiveandregenerativeapproachinthetreatmentofperi-implantitis:Aprospectivecaseseries.Clinical Oral Implants Research,25(7),761–767.https://doi.org/10.1111/clr.12183

Moher,D.,Liberati,A.,Tetzlaff,J.,&Altman,D.G.(2009).Preferredre-portingitemsforsystematicreviewsandmeta-analyses:ThePRISMAstatement.PLoS Medicine,6(7),e1000097.https://doi.org/10.1371/journal.pmed.1000097

Nart,J.,deTapia,B.,Pujol,A.,Pascual,A.,&Valles,C.(2017).Vancomycinand tobramycin impregnatedmineralized allograft for the surgicalregenerative treatmentofperi-implantitis:A1-year follow-upcaseseries. Clinical Oral Investigations, 22(6), 2199–2017. https://doi.org/10.1007/s00784-017-2310-0

Persson, L.G.,Araújo,M.G.,Berglundh,T.,Gröndahl,K.,&Lindhe, J.(1999).Resolutionofperi-implantitisfollowingtreatment.Anexperi-mentalstudyinthedog.Clin Oral Implants Res,10(3),195–203.

Reynolds,M. A., Kao, R. T., Camargo, P.M., Caton, J. G., Clem,D. S.,Fiorellini, J. P., &Nevins,M. L. (2015). Periodontal regeneration -intrabonydefects:AconsensusreportfromtheAAPRegenerationWorkshop.Journal of Periodontology,86(2Suppl),S105–S107.https://doi.org/10.1902/jop.2015.140378

Roccuzzo,M.,Bonino,F.,Bonino,L.,&Dalmasso,P.(2011).Surgicalther-apyofperi-implantitis lesionsbymeansofabovine-derivedxeno-graft:Comparativeresultsofaprospectivestudyontwodifferentimplant surfaces. Journal of Clinical Periodontology,38(8), 738–745.https://doi.org/10.1111/j.1600-051X.2011.01742.x

Roccuzzo,M.,Gaudioso,L.,Lungo,M.,&Dalmasso,P.(2016).Surgicalther-apyof singleperi-implantitis intrabonydefects,bymeansofdepro-teinized bovine bone mineral with 10% collagen. Journal of Clinical Periodontology,43(3),311–318.https://doi.org/10.1111/jcpe.12516

Roccuzzo, M., Pittoni, D., Roccuzzo, A., Charrier, L., & Dalmasso, P.(2017). Surgical treatment of peri-implantitis intrabony lesions bymeans of deproteinized bovine bone mineral with 10% collagen:7-year-results. Clinical Oral Implants Research, 28(12), 1577–1583.https://doi.org/10.1111/clr.13028

Roos-Jansåker,A.M., Lindahl,C.,Persson,G.R.,&Renvert, S. (2011).Long-term stability of surgical bone regenerative procedures ofperi-implantitis lesions in a prospective case-control study over3years.Journal of Clinical Periodontology,38(6),590–597.https://doi.org/10.1111/j.1600-051X.2011.01729.x

Roos-Jansåker,A.M.,Persson,G.R., Lindahl,C.,&Renvert, S. (2014).Surgical treatment of peri-implantitis using a bone substitutewithorwithout a resorbablemembrane: A 5-year follow-up. Journal of Clinical Periodontology,41(11), 1108–1114. https://doi.org/10.1111/jcpe.12308

Roos-Jansåker, A. M., Renvert, H., Lindahl, C., & Renvert, S. (2007a).Submergedhealingfollowingsurgicaltreatmentofperi-implantitis:Acaseseries.Journal of Clinical Periodontology,34(8),723–727.https://doi.org/10.1111/j.1600-051X.2007.01098.x

Roos-Jansåker, A. M., Renvert, H., Lindahl, C., & Renvert, S. (2007b).Surgical treatment of peri-implantitis using a bone substitutewithor without a resorbable membrane: A prospective cohort study.Journal of Clinical Periodontology, 34(7), 625–632. https://doi.org/10.1111/j.1600-051X.2007.01102.x

Sahrmann,P.,Attin,T.,&Schmidlin,P.R.(2011).Regenerativetreatmentofperi-implantitisusingbonesubstitutesandmembrane:Asystem-aticreview.Clinical Implant Dentistry and Related Research,13(1),46–57.https://doi.org/10.1111/j.1708-8208.2009.00183.x

Sanz, M., & Chapple, I. L. (2012). Clinical research on peri-implantdiseases: Consensus report of Working Group 4. Journal of Clinical Periodontology, 39(Suppl 12), 202–206. https://doi.org/10.1111/j.1600-051X.2011.01837.x

Schwarz,F.,Bieling,K.,Latz,T.,Nuesry,E.,&Becker,J.(2006).Healingof intrabony peri-implantitis defects following application of ananocrystalline hydroxyapatite (Ostim (TM)) or a bovine-derivedxenograft(Bio-Oss(TM))incombinationwithacollagenmembrane(Bio-Gide(TM)).Acaseseries.Journal of Clinical Periodontology,33(7),491–499.https://doi.org/10.1111/j.1600-051X.2006.00936.x

Schwarz,F.,Derks,J.,Monje,A.,&Wang,H.-L.(2018).Peri-implantitis.Journal of Clinical Periodontology, 45(6), S246–S266. https://doi.org/10.1111/jcpe.12954

Schwarz, F., Hegewald, A., John, G., Sahm, N., & Becker, J. (2013).Four-year follow-up of combined surgical therapy of advancedperi-implantitis evaluating two methods of surface decontamina-tion. Journal of Clinical Periodontology,40(10),962–967.https://doi.org/10.1111/jcpe.12143

Schwarz, F., Herten, M., Sager, M., Bieling, K., Sculean, A., & Becker,J. (2007). Comparison of naturally occurring and ligature-induced peri-implantitis bone defects in humans and dogs.Clinical Oral Implants Research, 18(2), 161–170. https://doi.org/10.1111/j.1600-0501.2006.01320.x

Schwarz, F., John, G., Mainusch, S., Sahm, N., & Becker, J. (2012).Combinedsurgicaltherapyofperi-implantitisevaluatingtwometh-odsofsurfacedebridementanddecontamination.Atwo-yearclini-calfollowupreport.Journal of Clinical Periodontology,39(8),789–797.https://doi.org/10.1111/j.1600-051x.2012.01867.x

Schwarz, F., John, G., Schmucker, A., Sahm, N., & Becker, J. (2017).Combined surgical therapy of advanced peri-implantitis evaluatingtwomethodsofsurfacedecontamination:A7-yearfollow-upobser-vation.Journal of Clinical Periodontology,44(3),337–342.https://doi.org/10.1111/jcpe.12648

Schwarz, F., Sahm,N., Bieling, K., & Becker, J. (2009). Surgical regen-erative treatment of peri-implantitis lesions using a nanocrys-talline hydroxyapatite or a natural bone mineral in combinationwith a collagen membrane: A four-year clinical follow-up report.Journal of Clinical Periodontology, 36(9), 807–814. https://doi.org/10.1111/j.1600-051X.2009.01443.x

Schwarz, F., Sahm, N., Iglhaut, G., & Becker, J. (2011). Impact of themethodofsurfacedebridementanddecontaminationontheclinicaloutcomefollowingcombinedsurgical therapyofperi-implantitis:Arandomizedcontrolledclinicalstudy.Journal of Clinical Periodontology,38(3),276–284.https://doi.org/10.1111/j.1600-051X.2010.01690.x

Schwarz,F.,Sahm,N.,Schwarz,K.,&Becker,J.(2010).Impactofdefectconfigurationontheclinicaloutcomefollowingsurgicalregenerativetherapy of peri-implantitis. Journal of Clinical Periodontology,37(5),449–455.https://doi.org/10.1111/j.1600-051X.2010.01540.x

Schwarz, F., Sculean, A., Bieling, K., Ferrari, D., Rothamel, D., &Becker, J. (2008). Two-year clinical results following treatmentof peri-implantitis lesions using a nanocrystalline hydroxyapatiteor a natural bone mineral in combination with a collagen mem-brane. Journal of Clinical Periodontology, 35(1), 80–87. https://doi.org/10.1111/j.1600-051X.2007.01168.x

Tomasi, C., & Derks, J. (2012). Clinical research of peri-implant dis-eases–quality of reporting, case definitions andmethods to studyincidence, prevalence and risk factors of peri-implant diseases.Journal of Clinical Periodontology,39(Suppl12),207–223.https://doi.org/10.1111/j.1600-051X.2011.01831.x

Tonetti, M. S., Cortellini, P., Lang, N. P., Suvan, J. E., Adriaens, P.,Dubravec, D., … Zybutz, M. (2004). Clinical outcomes following

     |  17TOMASI eT Al.

treatmentofhumanintrabonydefectswithGTR/bonereplacementmaterialoraccess flapalone.Amulticenter randomizedcontrolledclinicaltrial.Journal of Clinical Periodontology,31(9),770–776.https://doi.org/10.1111/j.1600-051x.2004.00562.x

Tonetti, M. S., Lang, N. P., Cortellini, P., Suvan, J. E., Adriaens, P.,Dubravec,D.,&Wallkamm,B.(2002).Enamelmatrixproteinsintheregenerative therapy of deep intrabony defects. Journal of Clinical Periodontology,29(4),317–325.

Trombelli,L.,Simonelli,A.,Minenna,L.,Rasperini,G.,&Farina,R.(2017).Effectofaconnectivetissuegraftincombinationwithasingleflapapproach in the regenerative treatment of intraosseous defects.Journal of Periodontology,88(4), 348–356. https://doi.org/10.1902/jop.2016.160471

Wells,G.A.,Shea,B.,O'Connell,D.,Peterson,J.,Welch,V.,Losos,M.,&Tugwell,P. (2014).TheNewcastle-OttawaScale(NOS)forassessingthe quality of nonrandomised studies in meta-analyses. Retrievedfromhttp://www.ohri.ca/programs/clinical_epidemiology/oxford.asp.

Wiltfang, J., Zernial, O., Behrens, E., Schlegel, A., Warnke, P. H., &Becker, S. T. (2012). Regenerative treatment of peri-implantitisbone defects with a combination of autologous bone and a de-mineralized xenogenic bone graft: A series of 36 defects. Clinical Implant Dentistry and Related Research,14(3), 421–427. https://doi.org/10.1111/j.1708-8208.2009.00264.x

Wohlfahrt,J.C.,Lyngstadaas,S.P.,Ronold,H.J.,Saxegaard,E.,Ellingsen,J. E., Karlsson, S., & Aass, A.M. (2012). Porous titanium granulesin the surgical treatment of peri-implant osseous defects: A ran-domized clinical trial. International Journal of Oral and Maxillofacial Implants,27(2),401–410.

SUPPORTING INFORMATION

Additional supporting information may be found online in theSupportingInformationsectionattheendofthearticle.

How to cite this article:TomasiC,RegidorE,Ortiz-VigónA,DerksJ.Efficacyofreconstructivesurgicaltherapyatperi-implantitis-relatedbonedefects.Asystematicreviewandmeta-analysis.J Clin Periodontol. 2019;00:1–17. https://doi.org/10.1111/jcpe.13070